@article{davis_howard_ellis_jonas_carey_morrissey_thomas_2022, title={Feasibility of a best–worst scaling exercise to set priorities for autism research}, volume={25}, ISSN={1369-6513 1369-7625}, url={http://dx.doi.org/10.1111/hex.13508}, DOI={10.1111/hex.13508}, abstractNote={Abstract}, number={4}, journal={Health Expectations}, publisher={Wiley}, author={Davis, Scott A. and Howard, Kirsten and Ellis, Alan R. and Jonas, Daniel E. and Carey, Timothy S. and Morrissey, Joseph P. and Thomas, Kathleen C.}, year={2022}, month={Jun}, pages={1643–1651} }
 @article{richmond_adams_annis_ellis_perryman_sikich_thomas_2022, title={Rapid and Deferred Help Seeking Among African American Parents of Children With Emotional and Behavioral Difficulties}, volume={73}, ISSN={1075-2730 1557-9700}, url={http://dx.doi.org/10.1176/appi.ps.202100553}, DOI={10.1176/appi.ps.202100553}, abstractNote={OBJECTIVE
Little is known about the factors African American parents consider when seeking care for their child after emotional and behavioral difficulties emerge. This study aimed to examine factors associated with seeking professional care within 30 days after identifying a child's need (i.e., rapid care seeking) and with deferring care for ≥1 year.


METHODS
This cross-sectional study surveyed African American parents raising a child with emotional or developmental challenges (N=289). Logistic regression was used to examine associations of parent activation, medical mistrust, and care-seeking barriers with two outcomes: rapidly seeking care and deferring care seeking.


RESULTS
About 22% of parents rapidly sought care, and 49% deferred care for 1 year or longer. Parents were more likely to rapidly seek care if they had higher parent activation scores; lived with other adults with mental health challenges; or, contrary to the authors' hypothesis, mistrusted doctors. Parents were less likely to rapidly seek care if the challenge did not initially bother them much or if their health insurance would not cover the service. Parents were more likely to defer care if they feared involuntary hospitalization for their child or if their health insurance would not cover the service. Parents were less likely to defer care if they had at least some college education or lived with other adults with mental health challenges.


CONCLUSIONS
Community-based pediatric and child welfare professionals should be informed about facilitators and barriers to mental health care seeking as part of efforts to develop interventions that support African American families.}, number={12}, journal={Psychiatric Services}, publisher={American Psychiatric Association Publishing}, author={Richmond, Jennifer and Adams, Leslie B. and Annis, Izabela E. and Ellis, Alan R. and Perryman, Twyla and Sikich, Linmarie and Thomas, Kathleen C.}, year={2022}, month={Dec}, pages={1359–1366} }
 @article{swietek_domino_grove_beadles_ellis_farley_jackson_lichstein_dubard_2021, title={Duration of medical home participation and quality of care for patients with chronic conditions}, volume={56}, ISSN={0017-9124 1475-6773}, url={http://dx.doi.org/10.1111/1475-6773.13710}, DOI={10.1111/1475-6773.13710}, abstractNote={Abstract}, number={S1}, journal={Health Services Research}, publisher={Wiley}, author={Swietek, Karen E. and Domino, Marisa Elena and Grove, Lexie R. and Beadles, Chris and Ellis, Alan R. and Farley, Joel F. and Jackson, Carlos and Lichstein, Jesse C. and DuBard, C. Annette}, year={2021}, month={Aug}, pages={1069–1079} }
 @article{nagavelli_mariano_krishnamoorthy_ray_hsia_ellis_memtsoudis_bryan_pepin_raghunathan_2021, title={Evaluation of trends in continuous peripheral nerve block utilization for total knee arthroplasty within and outside the Veterans Affairs Healthcare System}, volume={47}, ISSN={1098-7339 1532-8651}, url={http://dx.doi.org/10.1136/rapm-2021-102731}, DOI={10.1136/rapm-2021-102731}, abstractNote={To cite: Nagavelli H, Mariano ER, Krishnamoorthy V, et al. Reg Anesth Pain Med 2022;47:62–63. Department of Anesthesiology, Duke University Medical Center, Durham, North Carolina, USA Anesthesiology Service, Durham Veterans Affairs Medical Center, Durham, North Carolina, USA Anesthesiology and Perioperative Care Service, VA Palo Alto Health Care System, Palo Alto, California, USA Department of Anesthesiology, Perioperative and Pain Medicine, Stanford University School of Medicine, Stanford, California, USA School of Social Work, North Carolina State University, Raleigh, North Carolina, USA Department of Anesthesiology, Critical Care & Pain Management, Hospital for Special Surgery, New York, New York, USA Departments of Anesthesiology and Public Health, Weill Cornell Medical College, New York, New York, USA Pharmacy Service, Durham Veterans Affairs Medical Center, Durham, North Carolina, USA}, number={1}, journal={Regional Anesthesia & Pain Medicine}, publisher={BMJ}, author={Nagavelli, Harika and Mariano, Edward R and Krishnamoorthy, Vijay and Ray, Neil D. and Hsia, Hung-Lun and Ellis, Alan R. and Memtsoudis, Stavros G and Bryan, William E and Pepin, Marc J and Raghunathan, Karthik}, year={2021}, month={Jun}, pages={62–63} }
 @book{ellis_thomas_howard_ryan_bekker-grob_lancsar_2021, title={Improving Methods for Discrete Choice Experiments to Measure Patient Preferences}, url={https://doi.org/10.25302/03.2021.ME.160234572}, DOI={10.25302/03.2021.ME.160234572}, abstractNote={Results Summary Download Summary Results Summary What was the project about? Researchers can use experiments to learn about what patients prefer. Discrete choice experiments, or DCEs, describe treatments with different features, such as out-of-pocket costs or wait times. Patients fill out surveys about which treatments they prefer. From their choices, researchers learn what is most important to patients and how they think about the different features. DCEs can be hard to design and analyze. When surveys are complex, patients may ignore information or take shortcuts, which leads to inaccurate results. To make DCE results more accurate, researchers can Change the design of the DCE Apply statistical methods But current knowledge of how to do this is limited. In this project, the research team looked at improving methods to design and analyze DCEs. What did the research team do? First, the research team looked at how changes to the design of the DCE affected results. Using a computer program and data from two DCEs, the team created test data for 100,000 patients. The team used the test data to see how changes in DCE design, such as the number of patients taking part, affected results. DCEs are complex, so researchers often test the design in a small pilot study, which informs the design of the main study. The team also looked at how the changes in pilot study designs affected the accuracy of results from the main studies. Next, the research team looked at one type of statistical method used in DCEs called random parameter logit estimation with Halton draws. This method lets researchers measure what patients prefer while accounting for different preferences across patients. The team tested the method under different conditions, such as how much preferences vary from patient to patient. Then they looked at how many Halton draws were needed to get accurate results in a DCE study. The research team worked with other DCE researchers to design this study. What were the results? When the DCE design included more patients, the results were more accurate for assessing patient preferences. If the pilot study had design errors, results from the main study were less accurate. In random parameter logit estimation with Halton draws, the research team figured out the number of Halton draws needed to improve the accuracy of DCE results. What were the limits of the project? The research team used two DCEs and varied a few study design aspects. Results may differ for other data sets and design changes. Future research could test the random parameter logit estimation with Halton draws with other data sets and designs. How can people use the results? Researchers can use the results to improve how they design and analyze DCEs. Professional AbstractProfessional AbstractBackground Researchers use discrete choice experiments (DCEs) to measure individual patient preferences. In DCEs, researchers give patients a survey describing scenarios with different options from which to choose. For example, a hypothetical DCE offers two healthcare interventions. The two interventions differ in their features, like out-of-pocket costs and wait times. Patients choose the intervention they prefer. Patients' choices help researchers understand which features are most important to patients. Researchers also learn how patients think about the different options for each feature, such as different out-of-pocket costs. Designing and analyzing a DCE is challenging. For example, in DCEs with complex options, patients may ignore information, which may lead to inconsistent responses and inaccurate analyses. Altering DCE design features and statistical model assumptions may increase accuracy of DCE results. Objective To improve understanding of the effects of selected DCE design features and statistical model assumptions on DCE results. Study Design Design Element Description Design Simulations, empirical analysis Data Sources and Data Sets Empirical data on 2 DCEs: Study 1 examined preferences for organ allocation among adults (N=2,051) in the Australian general public. Study 2 examined preferences for labor induction among women (N=362) who were participating in a randomized trial of labor induction alternatives in South Australia. Simulated data for 100,000 participants based on results from empirical data sets Analytic Approach Simulations Random parameter logit estimation Outcomes Estimates of bias, relative standard error, and D-error (measure of overall error in parameter estimation) Methods First, the research team examined how different DCE designs affect study estimates. DCEs have two parts: a pilot study and a main study. The team created simulated DCE pilot and main studies by replicating two empirical DCEs in a simulated population of 100,000 individuals. They generated 864 simulations representing variations in DCE design such as sample size and the prevalence, correlations, and interactions of different variables. Using different analytic models, the team assessed estimation errors due to DCE design. Next, the research team examined the effects of using Halton draws on estimates from a random parameter logit model. Halton draws are a sampling technique that generates random data points simulating the overall population. The random parameter logit model assumes that parameters, such as the strength of preference for a certain healthcare feature, are random and vary across individuals. The team identified the number of Halton draws and the number of parameters for generating accurate results. DCE researchers helped design the study. Results In simulations, increasing sample size decreased random error. Random errors due to small sample size in the main study increased if the pilot study had a small sample size (n=30), unmeasured interactions, and selection bias. Random parameter logit analysis estimates had greater bias when model parameters were highly correlated. With correlations of 0.1, 0.2, and 0.3, bias reached 8%, 16%, and 24%, respectively. Too few Halton draws or a greater number of random parameters violated model assumptions and produced inaccurate results. Estimates were more accurate with fewer random parameters (less than 10). Up to 20,000 Halton draws were needed when the random parameters exceeded 15. Limitations Simulation scenarios did not cover the full range of study designs. Random parameters followed normal distributions. Results may differ for other parameter distributions and data sets. Conclusions and Relevance Improving methods for designing and analyzing DCEs can help researchers study patient preferences. Using more Halton draws when more random parameters are present may increase accuracy of random parameter logit models for DCEs. Future Research Needs Future research could examine additional DCE design features with other data sets.}, author={Ellis, Alan R. and Thomas, Kathleen and Howard, Kirsten and Ryan, Mandy and Bekker-Grob, Esther and Lancsar, Emily}, year={2021}, month={Mar} }
 @article{thomas_annis_ellis_adams_davis_lightfoot_perryman_wheeley_sikich_morrissey_2021, title={Parent Activation and Child Mental Health Service use in African American Families in a Large Cross-Sectional Study}, volume={25}, ISSN={1552-5775}, url={http://dx.doi.org/10.7812/TPP/20.003}, DOI={10.7812/TPP/20.003}, abstractNote={OBJECTIVES
1) To describe activation skills of African American parents on behalf of their children with mental health needs. 2) To assess the association between parent activation skills and child mental health service use.


METHODS
Data obtained in 2010 and 2011 from African American parents in North Carolina raising a child with mental health needs (n = 325) were used to identify child mental health service use from a medical provider, counselor, therapist, or any of the above or if the child had ever been hospitalized. Logistic regression was used to model the association between parent activation and child mental health service use controlling for predisposing, enabling, and need characteristics of the family and child.


RESULTS
Mean parent activation was 65.5%. Over two-thirds (68%) of children had seen a medical provider, 45% had seen a therapist, and 36% had seen a counselor in the past year. A quarter (25%) had been hospitalized. A 10-unit increase in parent activation was associated with a 31% higher odds that a child had seen any outpatient provider for their mental health needs (odds ratio = 1.31, confidence interval = 1.03-1.67, p = 0.03). The association varied by type of provider. Parent activation was not associated with seeing a counselor or a therapist or with being hospitalized.


CONCLUSION
African American families with activation skills are engaged and initiate child mental health service use. Findings provide a rationale for investing in the development and implementation of interventions that teach parent activation skills and facilitate their use by practices in order to help reduce disparities in child mental health service use.}, number={1}, journal={The Permanente Journal}, publisher={The Permanente Federation}, author={Thomas, Kathleen C and Annis, Izabela and Ellis, Alan R and Adams, Leslie B and Davis, Scott A and Lightfoot, Tywanda and Perryman, Twyla and Wheeley, Madeline and Sikich, Linmarie and Morrissey, Joseph P}, year={2021}, month={Mar}, pages={1–1} }
 @article{krishnamoorthy_motika_ohnuma_mclean_ellis_raghunathan_2021, title={Perioperative colloid choice and bleeding in patients undergoing musculoskeletal surgery: An observational administrative database study}, volume={11}, ISSN={2229-5151}, url={http://dx.doi.org/10.4103/IJCIIS.IJCIIS_178_20}, DOI={10.4103/IJCIIS.IJCIIS_178_20}, abstractNote={Background: The synthetic colloid hydroxyethyl starch (HES) received a black box warning, issued by the US Food and Drug Administration (FDA) in June 2013, in patients with sepsis, due to increased risk of bleeding, renal injury, and death. Risks of HES in populations undergoing noncardiac surgery are unclear. Here, we examine the association of colloid choice – human-derived albumin versus HES – with bleeding in musculoskeletal surgery. Methods: Inpatient musculoskeletal surgical patients who received colloids on the day of surgery were included during a time period before the FDA warning on HES using the Premier Healthcare database. The exposure was type of colloids administered on the day of surgery: HES versus albumin. The primary outcome was major perioperative bleeding, measured on the 1st postoperative day through hospital discharge. The secondary outcomes included acute renal failure and postoperative length of stay >75th percentile. Results: We identified 41,211 patients who received albumin (n = 12,803) and HES (n = 28,408) on the day of surgery. The propensity-weighted multivariable analysis demonstrated a reduced risk of major perioperative bleeding on the day after surgery following treatment with albumin versus HES (relative risk: 0.89 [95% confidence interval, 0.84–0.93]). No significant differences were observed in the secondary outcomes. Conclusion: When compared with albumin, treatment with HES on the day of musculoskeletal surgery was associated with an increased risk of major perioperative bleeding on subsequent days. Given that HES continues to be used as a colloid in multiple patient populations worldwide, further studies examining the safety of HES versus albumin solutions are needed.}, number={4}, journal={International Journal of Critical Illness and Injury Science}, publisher={Medknow}, author={Krishnamoorthy, Vijay and Motika, CalvinO and Ohnuma, Tetsu and McLean, Duncan and Ellis, AlanR and Raghunathan, Karthik}, year={2021}, pages={223} }
 @article{stürmer_webster-clark_lund_wyss_ellis_lunt_rothman_glynn_2021, title={Propensity Score Weighting and Trimming Strategies for Reducing Variance and Bias of Treatment Effect Estimates: A Simulation Study}, volume={190}, ISSN={0002-9262 1476-6256}, url={http://dx.doi.org/10.1093/aje/kwab041}, DOI={10.1093/aje/kwab041}, abstractNote={Abstract}, number={8}, journal={American Journal of Epidemiology}, publisher={Oxford University Press (OUP)}, author={Stürmer, Til and Webster-Clark, Michael and Lund, Jennifer L and Wyss, Richard and Ellis, Alan R and Lunt, Mark and Rothman, Kenneth J and Glynn, Robert J}, year={2021}, month={Feb}, pages={1659–1670} }
 @article{conover_rothman_stürmer_ellis_poole_jonsson funk_2021, title={Propensity score trimming mitigates bias due to covariate measurement error in inverse probability of treatment weighted analyses: A plasmode simulation}, volume={40}, ISSN={0277-6715 1097-0258}, url={http://dx.doi.org/10.1002/sim.8887}, DOI={10.1002/sim.8887}, abstractNote={BackgroundInverse probability of treatment weighting (IPTW) may be biased by influential observations, which can occur from misclassification of strong exposure predictors.}, number={9}, journal={Statistics in Medicine}, publisher={Wiley}, author={Conover, Mitchell M. and Rothman, Kenneth J. and Stürmer, Til and Ellis, Alan R. and Poole, Charles and Jonsson Funk, Michele}, year={2021}, month={Feb}, pages={2101–2112} }
 @article{ohnuma_raghunathan_fuller_ellis_johnbull_bartz_stefan_lindenauer_horn_krishnamoorthy_2021, title={Trends in Comorbidities and Complications Using ICD-9 and ICD-10 in Total Hip and Knee Arthroplasties}, volume={103}, ISSN={0021-9355 1535-1386}, url={http://dx.doi.org/10.2106/JBJS.20.01152}, DOI={10.2106/JBJS.20.01152}, abstractNote={
            Background:
            The transition to the new ICD-10 (International Classification of Diseases, Tenth Revision) coding system in the U.S. poses challenges to the ability to consistently and accurately measure trends in comorbidities and complications. We examined the prevalence of comorbidities and postoperative medical complications before and after the transition from ICD-9 to ICD-10 among patients who underwent primary total hip or knee arthroplasty (THA or TKA). We hypothesized that the transition to ICD-10 codes was associated with discontinuity and slope change in comorbidities and medical complications.
          }, number={8}, journal={Journal of Bone and Joint Surgery}, publisher={Ovid Technologies (Wolters Kluwer Health)}, author={Ohnuma, Tetsu and Raghunathan, Karthik and Fuller, Matthew and Ellis, Alan R. and JohnBull, Eric A. and Bartz, Raquel R. and Stefan, Mihaela S. and Lindenauer, Peter K. and Horn, Maggie E. and Krishnamoorthy, Vijay}, year={2021}, month={Feb}, pages={696–704} }
 @article{raghunathan_cobert_ellis_krishnamoorthy_mccartney_nathanson_stefan_lindenauer_2020, title={A clinical investigation into the benefits of using charge codes in perioperative and critical care epidemiology: A retrospective cohort database study}, volume={10}, ISSN={2229-5151}, url={http://dx.doi.org/10.4103/IJCIIS.IJCIIS_47_19}, DOI={10.4103/IJCIIS.IJCIIS_47_19}, abstractNote={Context: Epidemiologic studies in critical care routinely rely on the codes listed in International Classification of Diseases (ICD) manuals which are primarily intended for reimbursement of claims to payers. Standardized billing codes may minimize the measurement error when used in conjunction with ICD codes. Aims: The aim was to examine the impact of using charge codes in addition to ICD codes for ascertaining two common procedures in surgical intensive care unit (ICU) settings: hemodialysis (HD) and red blood cell (RBC) transfusions. Settings and Design: This was a retrospective cohort study of Premier Inc. Database. Subjects and Methods: Elective surgical patients aged >18 years treated in the ICU postoperatively were included in this study. This includes the ascertainment of HD and RBC transfusions in the population using a standard “ICD code” versus an “either ICD code or charge code” approach. Statistical Analysis Used: Descriptive analysis using t-tests, Chi-square tests as appropriate was used. Results: A total of 40,357 patients were identified as having undergone elective surgery, followed by admission to an ICU across 520 US hospitals. The use of “ICD codes only” uniformly underestimated rates of HD or RBC transfusions when compared to “Charge Codes only” and “ICD Codes or Charge Codes” (% increase of 15.4%–45.6% and 50.8%–93.1%, respectively). Differences varied with specific surgical populations studied. Patients identified using the “ICD code” approach had more comorbidities, were more likely to be female, and more likely to be Medicare beneficiaries. Conclusions: Epidemiologic studies in critical care should consider using multiple independent data sources to improve ascertainment of common critical care interventions.}, number={5}, journal={International Journal of Critical Illness and Injury Science}, publisher={Medknow}, author={Raghunathan, Karthik and Cobert, Julien and Ellis, AlanR and Krishnamoorthy, Vijay and McCartney, SharonL and Nathanson, BrianH and Stefan, MihaelaS and Lindenauer, Peter}, year={2020}, pages={39} }
 @article{chen_rasouli_ellis_ohnuma_bartz_krishnamoorthy_haines_raghunathan_2020, title={Associations Between Perioperative Crystalloid Volume and Adverse Outcomes in Five Surgical Populations}, volume={251}, ISSN={0022-4804}, url={http://dx.doi.org/10.1016/j.jss.2019.12.013}, DOI={10.1016/j.jss.2019.12.013}, abstractNote={Background Optimal administration of fluids is an important part of enhanced recovery after surgery (ERAS) protocols. We sought to examine the relationship between perioperative crystalloid volume and adverse outcomes in five common types of surgical procedures with ERAS fluid guidelines in place where large randomized controlled trials have not been conducted: breast reconstruction, bariatric, major urologic, gynoncologic, and head and neck oncologic procedures. Methods This retrospective cohort study included patients who had undergone any one of the aforementioned procedures within any facility in a large multihospital alliance (Premier, Inc, Charlotte, NC) between 2008 and 2014. We used multivariable generalized additive models to examine relationships between the total crystalloid volume (TCV) on the day of surgery and a composite adverse outcome of prolonged (>75th percentile) hospital or intensive care unit stay or in-hospital mortality. Models were constructed separately within each surgical category and adjusted for demographic, clinical, and hospital characteristics. Informed consent requirements were waived because deidentified data were used. Results We identified 83,685 patients within 312 US hospitals undergoing breast reconstruction (n = 8738), bariatric surgery (n = 8067), major urologic surgery (n = 28,654), gynoncologic surgery (n = 34,559), and head/neck oncology surgery (n = 3667). There was significant patient-independent variation in TCV. Probabilities of adverse outcomes increased at a TCV below 3 L and above 6 L for all types of surgeries except bariatric surgery, where larger volumes were associated with progressively better outcomes. Conclusions and Relevance Relationships between TCV and adverse outcomes were generally J shaped with higher volumes (>6 L) associated with increased risk. As per current ERAS guidelines, it is important to avoid excessive crystalloid volume in most surgical procedures except for bariatric surgery.}, journal={Journal of Surgical Research}, publisher={Elsevier BV}, author={Chen, Fangyu and Rasouli, Mohammad R. and Ellis, Alan R. and Ohnuma, Tetsu and Bartz, Raquel R. and Krishnamoorthy, Vijay and Haines, Krista L. and Raghunathan, Karthik}, year={2020}, month={Jul}, pages={26–32} }
 @article{krishnamoorthy_ellis_mclean_stefan_nathanson_cobert_lindenauer_brookhart_ohnuma_raghunathan_2020, title={Bleeding After Musculoskeletal Surgery in Hospitals That Switched From Hydroxyethyl Starch to Albumin Following a Food and Drug Administration Warning}, volume={131}, ISSN={0003-2999}, url={http://dx.doi.org/10.1213/ANE.0000000000004942}, DOI={10.1213/ANE.0000000000004942}, abstractNote={
            BACKGROUND:
            While US Food and Drug Administration (FDA) black box warnings are common, their impact on perioperative outcomes is unclear. Hydroxyethyl starch (HES) is associated with increased bleeding and kidney injury in patients with sepsis, leading to an FDA black box warning in 2013. Among patients undergoing musculoskeletal surgery in a subset of hospitals where colloid use changed from HES to albumin following the FDA warning, we examined the rate of major perioperative bleeding post- versus pre-FDA warning.
          }, number={4}, journal={Anesthesia & Analgesia}, publisher={Ovid Technologies (Wolters Kluwer Health)}, author={Krishnamoorthy, Vijay and Ellis, Alan R. and McLean, Duncan J. and Stefan, Mihaela S. and Nathanson, Brian H. and Cobert, Julien and Lindenauer, Peter K. and Brookhart, M. Alan and Ohnuma, Tetsu and Raghunathan, Karthik}, year={2020}, month={Jun}, pages={1193–1200} }
 @article{ohnuma_raghunathan_ellis_whittle_pyati_bryan_pepin_bartz_krishnamoorthy_2020, title={Effects of Acetaminophen, NSAIDs, Gabapentinoids, and Their Combinations on Postoperative Pulmonary Complications After Total Hip or Knee Arthroplasty}, volume={21}, ISSN={1526-2375 1526-4637}, url={http://dx.doi.org/10.1093/pm/pnaa017}, DOI={10.1093/pm/pnaa017}, abstractNote={Abstract}, number={10}, journal={Pain Medicine}, publisher={Oxford University Press (OUP)}, author={Ohnuma, Tetsu and Raghunathan, Karthik and Ellis, Alan R and Whittle, John and Pyati, Srinivas and Bryan, William E and Pepin, Marc J and Bartz, Raquel R and Krishnamoorthy, Vijay}, year={2020}, month={Feb}, pages={2385–2393} }
 @article{grove_domino_farley_swietek_beadles_ellis_jackson_dubard_2020, title={Medical Home Effects on Enrollees With Mental and Physical Illness}, volume={26}, ISSN={["1088-0224"]}, DOI={10.37765/ajmc.2020.43153}, abstractNote={OBJECTIVES
To assess the effect of medical home enrollment on acute care use and healthcare spending among Medicaid beneficiaries with mental and physical illness.


STUDY DESIGN
Retrospective cohort analysis of administrative data.


METHODS
We used 2007-2010 Medicaid claims and state psychiatric hospital data from a sample of 83,819 individuals diagnosed with schizophrenia or depression and at least 1 comorbid physical condition. We performed fixed-effects regression analysis at the person-month level to examine the effect of medical home enrollment on the probabilities of emergency department (ED) use, inpatient admission, and outpatient care use and on amount of Medicaid spending.


RESULTS
Medical home enrollment had no effect on ED use in either cohort and was associated with a lower probability of inpatient admission in the depression cohort (P <.05). Medical home enrollees in both cohorts experienced an increase in the probability of having any outpatient visits (P <.05). Medical home enrollment was associated with an increase in mean monthly spending among those with schizophrenia ($65.8; P <.05) and a decrease among those with depression (-$66.4; P <.05).


CONCLUSIONS
Among Medicaid beneficiaries with comorbid mental and physical illness, medical home enrollment appears to increase outpatient healthcare use and has mixed effects on acute care use. For individuals in this population who previously had no engagement with the healthcare system, use of the medical home model may represent an investment in providing improved access to needed outpatient services with cost savings potential for beneficiaries with depression.}, number={5}, journal={AMERICAN JOURNAL OF MANAGED CARE}, author={Grove, Lexie R. and Domino, Marisa Elena and Farley, Joel F. and Swietek, Karen E. and Beadles, Christopher and Ellis, Alan R. and Jackson, Carlos T. and DuBard, C. Annette}, year={2020}, month={May}, pages={218–223} }
 @article{webster‐clark_stürmer_wang_man_marinac‐dabic_rothman_ellis_gokhale_lunt_girman_et al._2020, title={Using propensity scores to estimate effects of treatment initiation decisions: State of the science}, volume={40}, ISSN={0277-6715 1097-0258}, url={http://dx.doi.org/10.1002/sim.8866}, DOI={10.1002/sim.8866}, abstractNote={Abstract}, number={7}, journal={Statistics in Medicine}, publisher={Wiley}, author={Webster‐Clark, Michael and Stürmer, Til and Wang, Tiansheng and Man, Kenneth and Marinac‐Dabic, Danica and Rothman, Kenneth J. and Ellis, Alan R. and Gokhale, Mugdha and Lunt, Mark and Girman, Cynthia and et al.}, year={2020}, month={Dec}, pages={1718–1735} }
 @article{ohnuma_krishnamoorthy_ellis_yan_ray_hsia_pyati_stefan_bryan_pepin_et al._2019, title={Association ‘Between Gabapentinoids on the Day of Colorectal Surgery and Adverse Postoperative Respiratory Outcomes}, volume={270}, ISSN={0003-4932 1528-1140}, url={http://dx.doi.org/10.1097/SLA.0000000000003317}, DOI={10.1097/SLA.0000000000003317}, abstractNote={
            Objective:
            The aim of this study was to determine the association between gabapentinoids on the day of surgery and adverse postoperative outcomes in patients undergoing colorectal surgery in the United States.
          }, number={6}, journal={Annals of Surgery}, publisher={Ovid Technologies (Wolters Kluwer Health)}, author={Ohnuma, Tetsu and Krishnamoorthy, Vijay and Ellis, Alan R. and Yan, Rosalie and Ray, Neil D. and Hsia, Hung-Lun and Pyati, Srinivas and Stefan, Mihaela and Bryan, William E. and Pepin, Marc J. and et al.}, year={2019}, month={Dec}, pages={e65–e67} }
 @article{ohnuma_raghunathan_moore_setoguchi_ellis_fuller_whittle_pyati_bryan_pepin_et al._2019, title={Dose-Dependent Association of Gabapentinoids with Pulmonary Complications After Total Hip and Knee Arthroplasties}, volume={102}, ISSN={0021-9355 1535-1386}, url={http://dx.doi.org/10.2106/JBJS.19.00889}, DOI={10.2106/JBJS.19.00889}, abstractNote={
            Background:
            Gabapentinoids are commonly prescribed in perioperative multimodal analgesia protocols. Despite widespread use, the optimal dose to reduce opioid consumption while minimizing risks is unknown. We assessed dose-dependent effects of gabapentinoids on opioid consumption and postoperative pulmonary complications following total hip or knee arthroplasty (THA or TKA). We hypothesized that use of a gabapentinoid on the day of THA or TKA is associated with an increased risk of postoperative pulmonary complications in a dose-response fashion compared with the risk for patients who did not receive the drug.
          }, number={3}, journal={Journal of Bone and Joint Surgery}, publisher={Ovid Technologies (Wolters Kluwer Health)}, author={Ohnuma, Tetsu and Raghunathan, Karthik and Moore, Sean and Setoguchi, Soko and Ellis, Alan R. and Fuller, Matthew and Whittle, John and Pyati, Srinivas and Bryan, William E. and Pepin, Marc J. and et al.}, year={2019}, month={Dec}, pages={221–229} }
 @article{ellis_fraser_2019, title={Effortful Control in Childhood: Dimensionality and Validation Through Associations With Sex, Aggression, and Social Information Processing Skills}, volume={10}, ISSN={2334-2315 1948-822X}, url={http://dx.doi.org/10.1086/704211}, DOI={10.1086/704211}, abstractNote={Objective: Many school-based prevention programs seek to strengthen effortful control in students. However, the concept of effortful control has been defined and measured inconsistently in prior studies. We identified 3 theory-based, practice-relevant dimensions of children’s effortful control and developed a teacher-rated measure to evaluate effects of programs to prevent or reduce aggressive behavior in elementary school children. Method: Using data from 690 3rd-grade students in 10 Southeastern U.S. schools, we conducted exploratory factor analysis (EFA) of 16 teacher-rated items in a training sample (n = 350). Results: The final EFA model included 13 items in an obliquely rotated solution that included factors of inhibitory control, attention control, and impulsivity. Confirmatory factor analysis revealed acceptable model fit. Effortful control was positively associated with female sex and negatively associated with aggression and conduct problems. Conclusion: Our effortful-control measure showed acceptable reliability and validity and could be used to evaluate outcomes such as aggression in tests of preventive interventions for elementary school children.}, number={3}, journal={Journal of the Society for Social Work and Research}, publisher={University of Chicago Press}, author={Ellis, Alan R. and Fraser, Mark W.}, year={2019}, month={Sep}, pages={423–439} }
 @article{zhang_mcgrath_ellis_wyss_lund_stürmer_2019, title={Restriction of Pharmacoepidemiologic Cohorts to Initiators of Medications in Unrelated Preventive Drug Classes to Reduce Confounding by Frailty in Older Adults}, volume={188}, ISSN={0002-9262 1476-6256}, url={http://dx.doi.org/10.1093/aje/kwz083}, DOI={10.1093/aje/kwz083}, abstractNote={Abstract}, number={7}, journal={American Journal of Epidemiology}, publisher={Oxford University Press (OUP)}, author={Zhang, Henry T and McGrath, Leah J and Ellis, Alan R and Wyss, Richard and Lund, Jennifer L and Stürmer, Til}, year={2019}, month={Mar}, pages={1371–1382} }
 @article{ellis_2018, title={A conceptual framework for preventing aggression in elementary schools}, volume={36}, ISSN={1536-5581}, url={http://dx.doi.org/10.1002/crq.21231}, DOI={10.1002/crq.21231}, abstractNote={Pervasive physical conflict generates negative outcomes. This paper (a) thoroughly describes the problem of early aggression, (b) identifies emotion regulation (ER) and social information processing (SIP) skills as targets for aggression prevention, and (c) locates skills training within a new conceptual framework. According to this framework, prevention programs should teach ER and SIP skills early and should target contextual factors. Multiple professions are well positioned to intervene using existing tools. Aggression prevention research should consider both emotion and cognition, improve measurement and study design, and incorporate theories that address skill development as well as the social justice implications of aggression prevention.}, number={3}, journal={Conflict Resolution Quarterly}, publisher={Wiley}, author={Ellis, Alan R.}, year={2018}, month={Aug}, pages={183–206} }
 @article{soekhai_de bekker-grob_ellis_vass_2018, title={Discrete Choice Experiments in Health Economics: Past, Present and Future}, volume={37}, ISSN={1170-7690 1179-2027}, url={http://dx.doi.org/10.1007/s40273-018-0734-2}, DOI={10.1007/s40273-018-0734-2}, abstractNote={Discrete choice experiments (DCEs) are increasingly advocated as a way to quantify preferences for health. However, increasing support does not necessarily result in increasing quality. Although specific reviews have been conducted in certain contexts, there exists no recent description of the general state of the science of health-related DCEs. The aim of this paper was to update prior reviews (1990–2012), to identify all health-related DCEs and to provide a description of trends, current practice and future challenges. A systematic literature review was conducted to identify health-related empirical DCEs published between 2013 and 2017. The search strategy and data extraction replicated prior reviews to allow the reporting of trends, although additional extraction fields were incorporated. Of the 7877 abstracts generated, 301 studies met the inclusion criteria and underwent data extraction. In general, the total number of DCEs per year continued to increase, with broader areas of application and increased geographic scope. Studies reported using more sophisticated designs (e.g. D-efficient) with associated software (e.g. Ngene). The trend towards using more sophisticated econometric models also continued. However, many studies presented sophisticated methods with insufficient detail. Qualitative research methods continued to be a popular approach for identifying attributes and levels. The use of empirical DCEs in health economics continues to grow. However, inadequate reporting of methodological details inhibits quality assessment. This may reduce decision-makers’ confidence in results and their ability to act on the findings. How and when to integrate health-related DCE outcomes into decision-making remains an important area for future research.}, number={2}, journal={PharmacoEconomics}, publisher={Springer Science and Business Media LLC}, author={Soekhai, Vikas and de Bekker-Grob, Esther W. and Ellis, Alan R. and Vass, Caroline M.}, year={2018}, month={Nov}, pages={201–226} }
 @article{swietek_domino_beadles_ellis_farley_grove_jackson_dubard_2018, title={Do Medical Homes Improve Quality of Care for Persons with Multiple Chronic Conditions?}, volume={53}, ISSN={0017-9124 1475-6773}, url={http://dx.doi.org/10.1111/1475-6773.13024}, DOI={10.1111/1475-6773.13024}, abstractNote={ObjectiveTo examine the association between medical home enrollment and receipt of recommended care for Medicaid beneficiaries with multiple chronic conditions (MCC).}, number={6}, journal={Health Services Research}, publisher={Wiley}, author={Swietek, Karen E. and Domino, Marisa Elena and Beadles, Christopher and Ellis, Alan R. and Farley, Joel F. and Grove, Lexie R. and Jackson, Carlos and DuBard, C. Annette}, year={2018}, month={Aug}, pages={4667–4681} }
 @article{domino_lin_morrissey_ellis_fraher_richman_thomas_prinstein_2018, title={Training Psychologists for Rural Practice: Exploring Opportunities and Constraints}, volume={35}, ISSN={0890-765X}, url={http://dx.doi.org/10.1111/jrh.12299}, DOI={10.1111/jrh.12299}, abstractNote={Abstract}, number={1}, journal={The Journal of Rural Health}, publisher={Wiley}, author={Domino, Marisa Elena and Lin, Ching-Ching Claire and Morrissey, Joseph P. and Ellis, Alan R. and Fraher, Erin and Richman, Erica L. and Thomas, Kathleen C. and Prinstein, Mitchell J.}, year={2018}, month={Apr}, pages={35–41} }
 @article{zhang_mcgrath_wyss_ellis_stürmer_2017, title={Controlling confounding by frailty when estimating influenza vaccine effectiveness using predictors of dependency in activities of daily living}, volume={26}, ISSN={1053-8569}, url={http://dx.doi.org/10.1002/PDS.4298}, DOI={10.1002/PDS.4298}, abstractNote={Abstract}, number={12}, journal={Pharmacoepidemiology and Drug Safety}, publisher={Wiley}, author={Zhang, Henry T. and McGrath, Leah J. and Wyss, Richard and Ellis, Alan R. and Stürmer, Til}, year={2017}, month={Aug}, pages={1500–1506} }
 @article{olesiuk_farley_domino_ellis_morrissey_lichstein_beadles_jackson_dubard_2017, title={Do Medical Homes Offer Improved Diabetes Care for Medicaid Enrollees with Co-occurring Schizophrenia?}, volume={28}, ISSN={1548-6869}, url={http://dx.doi.org/10.1353/hpu.2017.0094}, DOI={10.1353/hpu.2017.0094}, abstractNote={Purpose. To determine whether Medicaid recipients with co-occurring diabetes and schizophrenia that are medical-home-enrolled are more likely to receive guideline-concordant diabetes care than those who are not medical-home-enrolled, controlling for confounders. Methods. We used administrative data on adult, non-dually eligible North Carolina Medicaid beneficiaries with diagnoses of both diabetes and schizophrenia (N= 3,897) for fiscal years 2008–2010. We evaluated the relationship between medical-home-enrollment and receipt of recommended diabetes care reimbursed by Medicaid (lipid profiles, HbA1c tests, medical attention for nephropathy, and eye exams for those over 30), using fixed-effects regression models on person-month level data. Results. There was a statisti-cally significant, positive effect of medical home enrollment on receipt of Medicaid-funded eye exams, HbA1c tests, and medical attention for nephropathy, but not receipt of lipid profiles. Conclusions. For Medicaid enrollees with diabetes and schizophrenia, medical home enrollment is generally associated with greater likelihood of receiving guideline-concordant diabetes care.}, number={3}, journal={Journal of Health Care for the Poor and Underserved}, publisher={Project Muse}, author={Olesiuk, William J. and Farley, Joel F. and Domino, Marisa Elena and Ellis, Alan R. and Morrissey, Joseph P. and Lichstein, Jesse C. and Beadles, Christopher A. and Jackson, Carlos T. and DuBard, C. Annette}, year={2017}, pages={1030–1041} }
 @article{grove_olesiuk_ellis_lichstein_dubard_farley_jackson_beadles_morrissey_domino_2017, title={Evaluating the potential for primary care to serve as a mental health home for people with schizophrenia}, volume={47}, ISSN={0163-8343}, url={http://dx.doi.org/10.1016/j.genhosppsych.2017.03.002}, DOI={10.1016/j.genhosppsych.2017.03.002}, abstractNote={Primary care-based medical homes could improve the coordination of mental health care for individuals with schizophrenia and comorbid chronic conditions. The objective of this paper is to examine whether persons with schizophrenia and comorbid chronic conditions engage in primary care regularly, such that primary care settings have the potential to serve as a mental health home.We examined the annual primary care and specialty mental health service utilization of adult North Carolina Medicaid enrollees with schizophrenia and at least one comorbid chronic condition who were in a medical home during 2007-2010. Using a fixed-effects regression approach, we also assessed the effect of medical home enrollment on utilization of primary care and specialty mental health care and medication adherence.A substantial majority (78.5%) of person-years had at least one primary care visit, and 17.9% had at least one primary care visit but no specialty mental health services use. Medical home enrollment was associated with increased use of primary care and specialty mental health care, as well as increased medication adherence.Medical home enrollees with schizophrenia and comorbid chronic conditions exhibited significant engagement in primary care, suggesting that primary-care-based medical homes could serve a care coordination function for persons with schizophrenia.}, journal={General Hospital Psychiatry}, publisher={Elsevier BV}, author={Grove, Lexie R. and Olesiuk, William J. and Ellis, Alan R. and Lichstein, Jesse C. and DuBard, C. Annette and Farley, Joel F. and Jackson, Carlos T. and Beadles, Christopher A. and Morrissey, Joseph P. and Domino, Marisa Elena}, year={2017}, month={Jul}, pages={14–19} }
 @article{wyss_hansen_ellis_gagne_desai_glynn_stürmer_2017, title={The “Dry-Run” Analysis: A Method for Evaluating Risk Scores for Confounding Control}, volume={185}, ISSN={0002-9262 1476-6256}, url={http://dx.doi.org/10.1093/aje/kwx032}, DOI={10.1093/aje/kwx032}, abstractNote={A propensity score (PS) model's ability to control confounding can be assessed by evaluating covariate balance across exposure groups after PS adjustment. The optimal strategy for evaluating a disease risk score (DRS) model's ability to control confounding is less clear. DRS models cannot be evaluated through balance checks within the full population, and they are usually assessed through prediction diagnostics and goodness-of-fit tests. A proposed alternative is the "dry-run" analysis, which divides the unexposed population into "pseudo-exposed" and "pseudo-unexposed" groups so that differences on observed covariates resemble differences between the actual exposed and unexposed populations. With no exposure effect separating the pseudo-exposed and pseudo-unexposed groups, a DRS model is evaluated by its ability to retrieve an unconfounded null estimate after adjustment in this pseudo-population. We used simulations and an empirical example to compare traditional DRS performance metrics with the dry-run validation. In simulations, the dry run often improved assessment of confounding control, compared with the C statistic and goodness-of-fit tests. In the empirical example, PS and DRS matching gave similar results and showed good performance in terms of covariate balance (PS matching) and controlling confounding in the dry-run analysis (DRS matching). The dry-run analysis may prove useful in evaluating confounding control through DRS models.}, number={9}, journal={American Journal of Epidemiology}, publisher={Oxford University Press (OUP)}, author={Wyss, Richard and Hansen, Ben B. and Ellis, Alan R. and Gagne, Joshua J. and Desai, Rishi J. and Glynn, Robert J. and Stürmer, Til}, year={2017}, month={Mar}, pages={842–852} }
 @article{freburger_ellis_wang_butler_kshirsagar_winkelmayer_brookhart_2016, title={Comparative Effectiveness of Iron and Erythropoiesis-Stimulating Agent Dosing on Health-Related Quality of Life in Patients Receiving Hemodialysis}, volume={67}, ISSN={0272-6386}, url={http://dx.doi.org/10.1053/J.AJKD.2015.09.011}, DOI={10.1053/J.AJKD.2015.09.011}, abstractNote={BACKGROUND
The potential effects of iron-dosing strategies and erythropoiesis-stimulating agents (ESAs) on health-related quality of life (HRQoL) in the dialysis population are unclear. We examined the independent associations of bolus versus maintenance iron dosing and high versus low ESA dosing on HRQoL.


STUDY DESIGN
Retrospective cohort design.


SETTING & PARTICIPANTS
Clinical data (2008-2010) from a large dialysis organization merged with data from the US Renal Data System. 13,039 patients receiving center-based hemodialysis were included.


PREDICTOR
Iron and ESA dosing were assessed during 1-month (n=14,901) and 2-week (n=15,296) exposure periods.


OUTCOMES
HRQoL was measured by the Kidney Disease Quality of Life (KDQOL) instrument (0-100 scale) during a 3-month follow-up period.


MEASUREMENTS
Generalized linear mixed models, adjusting for several covariates, were used to estimate associations between iron and ESA dosing and HRQoL overall and for clinically relevant subgroups.


RESULTS
For the 1-month exposure period, patients with lower baseline hemoglobin levels who received higher ESA dosing had higher physical health and kidney disease symptom scores (by 2.4 [95% CI, 0.6-4.2] and 5.6 [95% CI, 2.8-8.4] points, respectively) in follow-up than patients who received lower ESA dosing. For the 2-week exposure period, patients with low baseline hemoglobin levels who received bolus dosing had higher mental health scores (by 1.9 [95% CI, 0.0-3.8] points) in follow-up. Within the low-baseline-hemoglobin subgroup, individuals with a catheter or dialysis vintage less than 1 year who received higher ESA dosing had higher HRQoL scores in follow-up (by 5.0-9.9 points) and individuals with low baseline transferrin saturations who received bolus dosing had higher HRQoL scores in follow-up (by 2.6-5.8 points).


LIMITATIONS
Observational design; short duration of observation.


CONCLUSIONS
For individuals with low baseline hemoglobin levels, higher ESA dosing and bolus iron dosing were associated with slightly higher HRQoL scores in follow-up. These differences became more pronounced and clinically relevant for specific subgroups.}, number={2}, journal={American Journal of Kidney Diseases}, publisher={Elsevier BV}, author={Freburger, Janet K. and Ellis, Alan R. and Wang, Lily and Butler, Anne M. and Kshirsagar, Abhijit V. and Winkelmayer, Wolfgang C. and Brookhart, M. Alan}, year={2016}, month={Feb}, pages={271–282} }
 @article{brookhart_freburger_ellis_winkelmayer_wang_kshirsagar_2016, title={Comparative Short-term Safety of Sodium Ferric Gluconate Versus Iron Sucrose in Hemodialysis Patients}, volume={67}, ISSN={0272-6386}, url={http://dx.doi.org/10.1053/J.AJKD.2015.07.026}, DOI={10.1053/J.AJKD.2015.07.026}, abstractNote={BACKGROUND
Despite different pharmacologic properties, little is known about the comparative safety of sodium ferric gluconate versus iron sucrose in hemodialysis patients.


STUDY DESIGN
Retrospective cohort study using the clinical database of a large dialysis provider (2004-2005) merged with administrative data from the US Renal Data System.


SETTING & PARTICIPANTS
66,207 patients with Medicare coverage who received center-based hemodialysis.


PREDICTORS
Iron formulation use assessed during repeated 1-month exposure periods (n=278,357).


OUTCOMES
All-cause mortality, infection-related hospitalizations and mortality, and cardiovascular-related hospitalizations and mortality occurring during a 3-month follow-up period.


MEASUREMENTS
For all outcomes, we estimated 90-day risk differences between the formulations using propensity score weighting of Kaplan-Meier functions, which controlled for a wide range of demographic, clinical, and laboratory variables. Risk differences were also estimated within various clinically important subgroups.


RESULTS
Ferric gluconate was administered in 11.4%; iron sucrose, in 48.9%; and no iron in 39.7% of the periods. Risks for most study outcomes did not differ between ferric gluconate and iron sucrose; however, among patients with a hemodialysis catheter, use of ferric gluconate was associated with a slightly decreased risk for both infection-related death (risk difference, -0.3%; 95% CI, -0.5% to 0.0%) and infection-related hospitalization (risk difference, -1.5%; 95% CI, -2.3% to -0.6%). Bolus dosing was associated with an increase in infection-related events among both ferric gluconate and iron sucrose users.


LIMITATIONS
Residual confounding and outcome measurement error.


CONCLUSIONS
Overall, the 2 iron formulations studied exhibited similar safety profiles; however, ferric gluconate was associated with a slightly decreased risk for infection-related outcomes compared to iron sucrose among patients with a hemodialysis catheter. These associations should be explored further using other data or study designs.}, number={1}, journal={American Journal of Kidney Diseases}, publisher={Elsevier BV}, author={Brookhart, M. Alan and Freburger, Janet K. and Ellis, Alan R. and Winkelmayer, Wolfgang C. and Wang, Lily and Kshirsagar, Abhijit V.}, year={2016}, month={Jan}, pages={119–127} }
 @article{domino_jackson_beadles_lichstein_ellis_farley_morrissey_dubard_2016, title={Do primary care medical homes facilitate care transitions after psychiatric discharge for patients with multiple chronic conditions?}, volume={39}, ISSN={0163-8343}, url={http://dx.doi.org/10.1016/J.GENHOSPPSYCH.2015.11.002}, DOI={10.1016/J.GENHOSPPSYCH.2015.11.002}, abstractNote={Primary-care-based medical homes may facilitate care transitions for persons with multiple chronic conditions (MCC) including serious mental illness. The purpose of this manuscript is to assess outpatient follow-up rates with primary care and mental health providers following psychiatric discharge by medical home enrollment and medical complexity. Using a quasi-experimental design, we examined data from North Carolina Medicaid-enrolled adults with MCC hospitalized with an inpatient diagnosis of depression or schizophrenia during 2008–2010. We used inverse-probability-of-treatment weighting and assessed associations between medical home enrollment and outpatient follow-up within 7 and 30 days postdischarge. Medical home enrollees (n= 16,137) were substantially more likely than controls (n= 11,304) to receive follow-up care with any provider 30 days post discharge. Increasing patient complexity was associated with a greater probability of primary care follow-up. Medical complexity and medical home enrollment were not associated with follow-up with a mental health provider. Hospitalized persons with MCC including serious mental illness enrolled in a medical home were more likely to receive timely outpatient follow-up with a primary care provider but not with a mental health specialist. These findings suggest that the medical home model may be more adept at linking patients to providers in primary care rather than to specialty mental health providers.}, journal={General Hospital Psychiatry}, publisher={Elsevier BV}, author={Domino, Marisa E. and Jackson, Carlos and Beadles, Christopher A. and Lichstein, Jesse C. and Ellis, Alan R. and Farley, Joel F. and Morrissey, Joseph P. and DuBard, C. Annette}, year={2016}, month={Mar}, pages={59–65} }
 @article{la_lich_wells_ellis_swartz_zhu_morrissey_2016, title={Increasing Access to State Psychiatric Hospital Beds: Exploring Supply-Side Solutions}, volume={67}, ISSN={1075-2730 1557-9700}, url={http://dx.doi.org/10.1176/appi.ps.201400570}, DOI={10.1176/appi.ps.201400570}, abstractNote={OBJECTIVE
The objective of this study was to identify supply-side interventions to reduce state psychiatric hospital admission delays.


METHODS
Healthcare Enterprise Accounts Receivable Tracking System (HEARTS) data were collected for all patients admitted between July 1, 2010, and July 31, 2012, to one of North Carolina's three state-operated psychiatric hospitals (N=3,156). Additional information on hospital use was collected at nine meetings with hospital administrators and other local stakeholders. A discrete-event simulation model was built to simulate the flow of adult nonforensic patients through the hospital. Hypothetical scenarios were used to evaluate the effects of varying levels of increased capacity on annual number of admissions and average patient wait time prior to admission.


RESULTS
In the base case, the model closely approximated actual state hospital utilization, with an average of 1,251±65 annual admissions and a preadmission wait time of 3.3±.1 days across 50 simulations. Results from simulated expansion scenarios highlighted substantial capacity shortfalls in the current system. For example, opening an additional 24-bed unit was projected to decrease average wait time by only 6%. Capacity would need to be increased by 165% (356 beds) to reduce average wait time below 24 hours.


CONCLUSIONS
Without more robust community-based hospital and residential capacity, major increases in state psychiatric hospital inpatient capacity are necessary to ensure timely admission of people in crisis.}, number={5}, journal={Psychiatric Services}, publisher={American Psychiatric Association Publishing}, author={La, Elizabeth M. and Lich, Kristen Hassmiller and Wells, Rebecca and Ellis, Alan R. and Swartz, Marvin S. and Zhu, Ruoqing and Morrissey, Joseph P.}, year={2016}, month={May}, pages={523–528} }
 @article{akizawa_kurita_mizobuchi_fukagawa_onishi_yamaguchi_ellis_fukuma_alan brookhart_hasegawa_et al._2016, title={PTH-dependence of the effectiveness of cinacalcet in hemodialysis patients with secondary hyperparathyroidism}, volume={6}, ISSN={2045-2322}, url={http://dx.doi.org/10.1038/srep19612}, DOI={10.1038/srep19612}, abstractNote={Abstract}, number={1}, journal={Scientific Reports}, publisher={Springer Science and Business Media LLC}, author={Akizawa, Tadao and Kurita, Noriaki and Mizobuchi, Masahide and Fukagawa, Masafumi and Onishi, Yoshihiro and Yamaguchi, Takuhiro and Ellis, Alan R. and Fukuma, Shingo and Alan Brookhart, M. and Hasegawa, Takeshi and et al.}, year={2016}, month={Apr} }
 @article{mcgrath_ellis_brookhart_2015, title={Controlling Time-Dependent Confounding by Health Status and Frailty: Restriction Versus Statistical Adjustment}, volume={182}, ISSN={1476-6256 0002-9262}, url={http://dx.doi.org/10.1093/aje/kwu485}, DOI={10.1093/aje/kwu485}, abstractNote={Nonexperimental studies of preventive interventions are often biased because of the healthy-user effect and, in frail populations, because of confounding by functional status. Bias is evident when estimating influenza vaccine effectiveness, even after adjustment for claims-based indicators of illness. We explored bias reduction methods while estimating vaccine effectiveness in a cohort of adult hemodialysis patients. Using the United States Renal Data System and linked data from a commercial dialysis provider, we estimated vaccine effectiveness using a Cox proportional hazards marginal structural model of all-cause mortality before and during 3 influenza seasons in 2005/2006 through 2007/2008. To improve confounding control, we added frailty indicators to the model, measured time-varying confounders at different time intervals, and restricted the sample in multiple ways. Crude and baseline-adjusted marginal structural models remained strongly biased. Restricting to a healthier population removed some unmeasured confounding; however, this reduced the sample size, resulting in wide confidence intervals. We estimated an influenza vaccine effectiveness of 9% (hazard ratio = 0.91, 95% confidence interval: 0.72, 1.15) when bias was minimized through cohort restriction. In this study, the healthy-user bias could not be controlled through statistical adjustment; however, sample restriction reduced much of the bias.}, number={1}, journal={American Journal of Epidemiology}, publisher={Oxford University Press (OUP)}, author={McGrath, Leah J. and Ellis, Alan R. and Brookhart, M. Alan}, year={2015}, month={Apr}, pages={17–25} }
 @article{beadles_farley_ellis_lichstein_morrissey_dubard_domino_2015, title={Do Medical Homes Increase Medication Adherence for Persons With Multiple Chronic Conditions?}, volume={53}, ISSN={0025-7079}, url={http://dx.doi.org/10.1097/MLR.0000000000000292}, DOI={10.1097/MLR.0000000000000292}, abstractNote={Background:Medications are an integral component of management for many chronic conditions, and suboptimal adherence limits medication effectiveness among persons with multiple chronic conditions (MCC). Medical homes may provide a mechanism for increasing adherence among persons with MCC, thereby enhancing management of chronic conditions. Objective:To examine the association between medical home enrollment and adherence to newly initiated medications among Medicaid enrollees with MCC. Research Design:Retrospective cohort study comparing Community Care of North Carolina medical home enrollees to nonenrollees using merged North Carolina Medicaid claims data (fiscal years 2008–2010). Subjects:Among North Carolina Medicaid-enrolled adults with MCC, we created separate longitudinal cohorts of new users of antidepressants (N=9303), antihypertensive agents (N=12,595), oral diabetic agents (N=6409), and statins (N=9263). Measures:Outcomes were the proportion of days covered (PDC) on treatment medication each month for 12 months and a dichotomous measure of adherence (PDC>0.80). Our primary analysis utilized person-level fixed effects models. Sensitivity analyses included propensity score and person-level random-effect models. Results:Compared with nonenrollees, medical home enrollees exhibited higher PDC by 4.7, 6.0, 4.8, and 5.1 percentage points for depression, hypertension, diabetes, and hyperlipidemia, respectively (P’s<0.001). The dichotomous adherence measure showed similar increases, with absolute differences of 4.1, 4.5, 3.5, and 4.6 percentage points, respectively (P’s<0.001). Conclusions:Among Medicaid enrollees with MCC, adherence to new medications is greater for those enrolled in medical homes.}, number={2}, journal={Medical Care}, publisher={Ovid Technologies (Wolters Kluwer Health)}, author={Beadles, Christopher A. and Farley, Joel F. and Ellis, Alan R. and Lichstein, Jesse C. and Morrissey, Joseph P. and DuBard, C. Annette and Domino, Marisa E.}, year={2015}, month={Feb}, pages={168–176} }
 @article{beadles_ellis_lichstein_farley_jackson_morrissey_domino_2015, title={First Outpatient Follow-Up After Psychiatric Hospitalization: Does One Size Fit All?}, volume={66}, ISSN={1075-2730 1557-9700}, url={http://dx.doi.org/10.1176/appi.ps.201400081}, DOI={10.1176/appi.ps.201400081}, abstractNote={OBJECTIVE
Claims-based indicators of follow-up within seven and 30 days after psychiatric discharge have face validity as quality measures: early follow-up may improve disease management and guide appropriate service use. Yet these indicators are rarely examined empirically. This study assessed their association with subsequent health care utilization for adults with comorbid conditions.


METHODS
Postdischarge follow-up and subsequent utilization were examined among adults enrolled in North Carolina Medicaid who were discharged with claims-based diagnoses of depression or schizophrenia and not readmitted within 30 days. A total of 24,934 discharges (18,341 individuals) in fiscal years 2008-2010 were analyzed. Follow-up was categorized as occurring within 0-7 days, 8-30 days, or none in 30 days. Outcomes in the subsequent six months included psychotropic medication claims, adherence (proportion of days covered), number of hospital admissions, emergency department visits, and outpatient visits.


RESULTS
Follow-up within seven days was associated with greater medication adherence and outpatient utilization, compared with no follow-up in 30 days. This was observed for both follow-up with a mental health provider and with any provider. Adults receiving mental health follow-up within seven days had equivalent, or lower, subsequent inpatient and emergency department utilization as those without follow-up within 30 days. However, adults receiving follow-up with any provider within seven days were more likely than those with no follow-up to have an inpatient admission or emergency department visit in the subsequent six months. Few differences in subsequent utilization were observed between mental health follow-up within seven days versus eight to 30 days.


CONCLUSIONS
For patients not readmitted within 30 days, follow-up within 30 days appeared to be beneficial on the basis of subsequent service utilization.}, number={4}, journal={Psychiatric Services}, publisher={American Psychiatric Association Publishing}, author={Beadles, Christopher A. and Ellis, Alan R. and Lichstein, Jesse C. and Farley, Joel F. and Jackson, Carlos T. and Morrissey, Joseph P. and Domino, Marisa Elena}, year={2015}, month={Apr}, pages={364–372} }
 @article{wyss_ellis_brookhart_jonsson funk_girman_simpson_stürmer_2015, title={Matching on the disease risk score in comparative effectiveness research of new treatments}, volume={24}, ISSN={1053-8569}, url={http://dx.doi.org/10.1002/PDS.3810}, DOI={10.1002/PDS.3810}, abstractNote={Abstract}, number={9}, journal={Pharmacoepidemiology and Drug Safety}, publisher={Wiley}, author={Wyss, Richard and Ellis, Alan R. and Brookhart, M. Alan and Jonsson Funk, Michele and Girman, Cynthia J. and Simpson, Ross J., Jr and Stürmer, Til}, year={2015}, month={Jun}, pages={951–961} }
 @article{la_zhu_lich_ellis_swartz_kosorok_morrissey_2015, title={The Effects of State Psychiatric Hospital Waitlist Policies on Length of Stay and Time to Readmission}, volume={42}, ISSN={0894-587X 1573-3289}, url={http://dx.doi.org/10.1007/S10488-014-0573-1}, DOI={10.1007/S10488-014-0573-1}, abstractNote={This study examined the effects of a waitlist policy for state psychiatric hospitals on length of stay and time to readmission using data from North Carolina for 2004–2010. Cox proportional hazards models tested the hypothesis that patients were discharged “quicker-but-sicker” post-waitlist, as hospitals struggled to manage admission delays and quickly admit waitlisted patients. Results refute this hypothesis, indicating that waitlists were associated with increased length of stay and time to readmission. Further research is needed to evaluate patients’ clinical outcomes directly and to examine the impact of state hospital waitlists in other areas, such as state hospital case mix, local emergency departments, and outpatient mental health agencies.}, number={3}, journal={Administration and Policy in Mental Health and Mental Health Services Research}, publisher={Springer Science and Business Media LLC}, author={La, Elizabeth Holdsworth and Zhu, Ruoqing and Lich, Kristen Hassmiller and Ellis, Alan R. and Swartz, Marvin S. and Kosorok, Michael R. and Morrissey, Joseph P.}, year={2015}, month={May}, pages={332–342} }
 @article{christian_gaynes_saavedra_sheitman_wines_jonas_viswanathan_ellis_woodell_carey_2015, title={Use of Antipsychotic Medications in Pediatric and Young Adult Populations}, volume={21}, ISSN={1538-1145}, url={http://dx.doi.org/10.1097/01.pra.0000460619.10429.4c}, DOI={10.1097/01.pra.0000460619.10429.4c}, abstractNote={The use of antipsychotics, particularly second generation antipsychotics, among children and adolescents has increased markedly during the past 20 years. Existing evidence gaps make this practice controversial and hinder treatment decision-making. This article describes and prioritizes future research needs regarding antipsychotic treatment in youth, focusing on within-class and between-class drug comparisons with regard to key population subgroups, efficacy and effectiveness outcomes, and adverse event outcomes. Using as a foundation a recent systematic review of antipsychotic treatment among youth, which was completed by a different Evidence-based Practice Center, we worked with a diverse group of 12 stakeholders representing researchers, funders, health care providers, patients, and families to identify and prioritize research needs. From an initial list of 16 evidence gaps, we enumerated 6 high-priority research needs: 1) long-term comparative effectiveness across all psychiatric disorders; 2) comparative long-term risks of adverse outcomes; 3) short-term risks of adverse events; 4) differentials of efficacy, effectiveness, and safety for population subgroups; 5) comparative effectiveness among those with attention-deficit/hyperactivity disorder and disruptive behavior disorders and common comorbidities; 6) comparative effectiveness among those with bipolar disorder and common comorbidities. In this article, we describe these future research needs in detail and discuss study designs that could be used to address them. (Journal of Psychiatric Practice 2015;21:26–36)}, number={1}, journal={Journal of Psychiatric Practice}, publisher={Ovid Technologies (Wolters Kluwer Health)}, author={Christian, Robert B. and Gaynes, Bradley N. and Saavedra, Lissette M. and Sheitman, Brian and Wines, Roberta and Jonas, Daniel E. and Viswanathan, Meera and Ellis, Alan R. and Woodell, Carol and Carey, Timothy S.}, year={2015}, month={Jan}, pages={26–36} }
 @article{gaynes_christian_saavedra_wines_jonas_viswanathan_ellis_woodell_carey_2014, title={Attention-Deficit/Hyperactivity Disorder}, volume={20}, ISSN={1538-1145}, url={http://dx.doi.org/10.1097/01.pra.0000445245.46424.25}, DOI={10.1097/01.pra.0000445245.46424.25}, abstractNote={With onset often occurring before 6 years of age, attention-deficit/hyperactivity disorder (ADHD) involves attention problems, impulsivity, overactivity, and sometimes disruptive behavior. Impairment usually persists into adulthood, with an estimated worldwide prevalence in adults of 2.5%. Existing gaps in evidence concerning ADHD hinder decision-making about treatment. This article describes and prioritizes future research needs for ADHD in three areas: treatment effectiveness for at-risk preschoolers; long-term treatment effectiveness; and variability in prevalence, diagnosis, and treatment. Using a recent systematic review concerning ADHD completed by a different evidence-based practice center as a foundation, we worked with a diverse group of 12 stakeholders, who represented researchers, funders, healthcare providers, patients, and families, to identify and prioritize research needs. From an initial list of 29 evidence gaps, we enumerated 8 high-priority research needs: a) accurate, brief standardized diagnosis and assessment; b) comparative effectiveness and safety of pharmacologic treatments for children under 6 years of age; c) comparative effectiveness of different combinations of psychosocial and pharmacologic treatments for children under 6 years of age; d) case identification and measurement of prevalence and outcomes; e) comparative effectiveness of psychosocial treatment alone versus pharmacologic and combination treatments for children under 6 years of age; f) comparative long-term treatment effectiveness for people 6 years of age and older; g) relative efficacy of specific psychosocial program components for children under 6 years of age; and h) identification of person-level effect modifiers for people 6 years of age and older. In this article, we describe these future research needs in detail and discuss study designs that could be used to address them. (Journal of Psychiatric Practice 2014;20:104–117)}, number={2}, journal={Journal of Psychiatric Practice}, publisher={Ovid Technologies (Wolters Kluwer Health)}, author={Gaynes, Bradley N. and Christian, Robert and Saavedra, Lissette M. and Wines, Roberta and Jonas, Daniel E. and Viswanathan, Meera and Ellis, Alan R. and Woodell, Carol and Carey, Timothy S.}, year={2014}, month={Mar}, pages={104–117} }
 @article{freburger_ellis_kshirsagar_wang_brookhart_2014, title={Comparative short-term safety of bolus versus maintenance iron dosing in hemodialysis patients: a replication study}, volume={15}, ISSN={1471-2369}, url={http://dx.doi.org/10.1186/1471-2369-15-154}, DOI={10.1186/1471-2369-15-154}, abstractNote={Recent research has reported that patients receiving bolus (frequent large doses to achieve iron repletion) versus maintenance dosing of iron have an increased short-term risk of infection, but a similar risk of cardiovascular events. We sought to determine whether these findings could be replicated using the same methods and a different data source. Clinical data from 6,605 patients of a small U.S. dialysis provider merged with Medicare claims data were examined. Iron dosing patterns (bolus, maintenance, no iron) were identified during 1-month exposure periods and cardiovascular and infection-related outcomes were assessed during 3-month follow-up periods. The effects of bolus versus maintenance dosing were assessed using Cox proportional hazards regression analyses to estimate hazard ratios and semiparametric additive risk models to estimate hazard rate differences, controlling for demographic and clinical characteristics, laboratory values and medications, and comorbidities. 48,050 exposure/follow-up periods were examined. 13.9 percent of the exposure periods were bolus dosing, 49.3 percent were maintenance dosing, and the remainder were no iron use. All of the adjusted hazard ratios were >1.00 for the infection-related outcomes, suggesting that bolus dosing increases the risk of these events. The effects were greatest for hospitalized for infection of any major organ system (hazard ratio 1.13 (1.03, 1.24)) and use of intravenous antibiotics (hazard ratio 1.08 (1.02, 1.15). When examining the subgroup of individuals with catheters, the hazard ratios for the infection-related outcomes were generally greater than in the overall sample. There was little association between type of dosing practice and cardiovascular outcomes. Results of this study provide further evidence of the association between bolus dosing and increased infection risk, particularly in the subgroup of patients with a catheter, and of the lack of an association between dosing practices and cardiovascular outcomes.}, number={1}, journal={BMC Nephrology}, publisher={Springer Science and Business Media LLC}, author={Freburger, Janet K and Ellis, Alan R and Kshirsagar, Abhijit V and Wang, Lily and Brookhart, M Alan}, year={2014}, month={Sep} }
 @article{domino_beadles_lichstein_farley_morrissey_ellis_dubard_2014, title={Heterogeneity in the Quality of Care for Patients With Multiple Chronic Conditions by Psychiatric Comorbidity}, volume={52}, ISSN={0025-7079}, url={http://dx.doi.org/10.1097/MLR.0000000000000024}, DOI={10.1097/MLR.0000000000000024}, abstractNote={Background:Little is known about the quality of care received by Medicaid enrollees with multiple chronic conditions (MCCs) and whether quality is different for those with mental illness. Objectives:To examine cancer screening and single-disease quality of care measures in a Medicaid population with MCC and to compare quality measures among persons with MCC with varying medical comorbidities with and without depression or schizophrenia. Research Design:Secondary data analysis using a unique data source combining Medicaid claims with other administrative datasets from North Carolina’s mental health system. Subjects:Medicaid-enrolled adults aged 18 and older with ≥2 of 8 chronic conditions (asthma, chronic obstructive pulmonary disease, diabetes, hypertension, hyperlipidemia, seizure disorder, depression, or schizophrenia). Medicare/Medicaid dual enrollees were excluded due to incomplete data on their medical care utilization. Measures:We examined a number of quality measures, including cancer screening, disease-specific metrics, such as receipt of hemoglobin A1C tests for persons with diabetes, and receipt of psychosocial therapies for persons with depression or schizophrenia, and medication adherence. Results:Quality of care metrics was generally lower among those with depression or schizophrenia, and often higher among those with increasing levels of medical comorbidities. A number of exceptions to these trends were noted. Conclusions:Cancer screening and single-disease quality measures may provide a benchmark for overall quality of care for persons with MCC; these measures were generally lower among persons with MCC and mental illness. Further research on quality measures that better reflect the complex care received by persons with MCC is essential.}, number={Supplement 2}, journal={Medical Care}, publisher={Ovid Technologies (Wolters Kluwer Health)}, author={Domino, Marisa E. and Beadles, Christopher A. and Lichstein, Jesse C. and Farley, Joel F. and Morrissey, Joseph P. and Ellis, Alan R. and Dubard, C. Annette}, year={2014}, month={Mar}, pages={S101–S109} }
 @article{wyss_ellis_lunt_brookhart_glynn_stürmer_2014, title={Model Misspecification When Excluding Instrumental Variables from PS Models in Settings Where Instruments Modify the Effects of Covariates on Treatment}, volume={0}, ISSN={2194-9263 2161-962X}, url={http://dx.doi.org/10.1515/em-2013-0012}, DOI={10.1515/em-2013-0012}, abstractNote={Abstract}, number={0}, journal={Epidemiologic Methods}, publisher={Walter de Gruyter GmbH}, author={Wyss, Richard and Ellis, Alan R. and Lunt, Mark and Brookhart, M. Alan and Glynn, Robert J. and Stürmer, Til}, year={2014}, month={Jan} }
 @article{wyss_ellis_brookhart_girman_jonsson funk_locasale_stürmer_2014, title={The Role of Prediction Modeling in Propensity Score Estimation: An Evaluation of Logistic Regression, bCART, and the Covariate-Balancing Propensity Score}, volume={180}, ISSN={0002-9262 1476-6256}, url={http://dx.doi.org/10.1093/aje/kwu181}, DOI={10.1093/aje/kwu181}, abstractNote={The covariate-balancing propensity score (CBPS) extends logistic regression to simultaneously optimize covariate balance and treatment prediction. Although the CBPS has been shown to perform well in certain settings, its performance has not been evaluated in settings specific to pharmacoepidemiology and large database research. In this study, we use both simulations and empirical data to compare the performance of the CBPS with logistic regression and boosted classification and regression trees. We simulated various degrees of model misspecification to evaluate the robustness of each propensity score (PS) estimation method. We then applied these methods to compare the effect of initiating glucagonlike peptide-1 agonists versus sulfonylureas on cardiovascular events and all-cause mortality in the US Medicare population in 2007-2009. In simulations, the CBPS was generally more robust in terms of balancing covariates and reducing bias compared with misspecified logistic PS models and boosted classification and regression trees. All PS estimation methods performed similarly in the empirical example. For settings common to pharmacoepidemiology, logistic regression with balance checks to assess model specification is a valid method for PS estimation, but it can require refitting multiple models until covariate balance is achieved. The CBPS is a promising method to improve the robustness of PS models.}, number={6}, journal={American Journal of Epidemiology}, publisher={Oxford University Press (OUP)}, author={Wyss, Richard and Ellis, Alan R. and Brookhart, M. Alan and Girman, Cynthia J. and Jonsson Funk, Michele and LoCasale, Robert and Stürmer, Til}, year={2014}, month={Aug}, pages={645–655} }
 @article{lichstein_domino_beadles_ellis_farley_morrissey_gauchat_dubard_jackson_2014, title={Use of Medical Homes by Patients With Comorbid Physical and Severe Mental Illness}, volume={52}, ISSN={0025-7079}, url={http://dx.doi.org/10.1097/MLR.0000000000000025}, DOI={10.1097/MLR.0000000000000025}, abstractNote={Background:Patients with comorbid severe mental illness (SMI) may use primary care medical homes differently than other patients with multiple chronic conditions (MCC). Objective:To compare medical home use among patients with comorbid SMI to use among those with only chronic physical comorbidities. Research Design:We examined data on children and adults with MCC for fiscal years 2008–2010, using generalized estimating equations to assess associations between SMI (major depressive disorder or psychosis) and medical home use. Subjects:Medicaid and medical home enrolled children (age, 6–17 y) and adults (age, 18–64 y) in North Carolina with ≥2 of the following chronic health conditions: major depressive disorder, psychosis, hypertension, diabetes, hyperlipidemia, seizure disorder, asthma, and chronic obstructive pulmonary disease. Measures:We examined annual medical home participation (≥1 visit to the medical home) among enrollees and utilization (number of medical home visits) among participants. Results:Compared with patients without depression or psychosis, children and adults with psychosis had lower rates of medical home participation (−12.2 and −8.2 percentage points, respectively, P<0.01) and lower utilization (−0.92 and −1.02 visits, respectively, P<0.01). Children with depression had lower participation than children without depression or psychosis (−5.0 percentage points, P<0.05). Participation and utilization among adults with depression was comparable with use among adults without depression or psychosis (P>0.05). Conclusions:Overall, medical home use was relatively high for Medicaid enrollees with MCC, though it was somewhat lower among those with SMI. Targeted strategies may be required to increase medical home participation and utilization among SMI patients.}, number={Supplement 2}, journal={Medical Care}, publisher={Ovid Technologies (Wolters Kluwer Health)}, author={Lichstein, Jesse C. and Domino, Marisa E. and Beadles, Christopher A. and Ellis, Alan R. and Farley, Joel F. and Morrissey, Joseph P. and Gauchat, Gordon W. and DuBard, C. Annette and Jackson, Carlos T.}, year={2014}, month={Mar}, pages={S85–S91} }
 @article{ellis_brookhart_2013, title={Approaches to inverse-probability-of-treatment–weighted estimation with concurrent treatments}, volume={66}, ISSN={0895-4356}, url={http://dx.doi.org/10.1016/J.JCLINEPI.2013.03.020}, DOI={10.1016/J.JCLINEPI.2013.03.020}, abstractNote={In a setting with two concurrent treatments, inverse-probability-of-treatment weights can be used to estimate the joint treatment effects or the marginal effect of one treatment while taking the other to be a confounder. We explore these two approaches in a study of intravenous iron use in hemodialysis patients treated concurrently with epoetin alfa (EPO).We linked US Renal Data System data with electronic health records (2004-2008) from a large dialysis provider. Using a retrospective cohort design with 776,203 records from 117,050 regular hemodialysis patients, we examined a composite outcome: mortality, myocardial infarction, or stroke.With EPO as a joint treatment, inverse-probability-of-treatment weights were unstable, confidence intervals for treatment effects were wide, covariate balance was unsatisfactory, and the treatment and outcome models were sensitive to omission of the baseline EPO covariate. By handling EPO exposure as a confounder instead of a joint treatment, we derived stable weights and balanced treatment groups on measured covariates.In settings with concurrent treatments, if only one treatment is of interest, then including the other in the treatment model as a confounder may result in more stable treatment effect estimates. Otherwise, extreme weights may necessitate additional analysis steps.}, number={8}, journal={Journal of Clinical Epidemiology}, publisher={Elsevier BV}, author={Ellis, Alan R. and Brookhart, M. Alan}, year={2013}, month={Aug}, pages={S51–S56} }
 @article{ellis_dusetzina_hansen_gaynes_farley_stürmer_2013, title={Confounding control in a nonexperimental study of STAR*D data: logistic regression balanced covariates better than boosted CART}, volume={23}, ISSN={1047-2797}, url={http://dx.doi.org/10.1016/J.ANNEPIDEM.2013.01.004}, DOI={10.1016/J.ANNEPIDEM.2013.01.004}, abstractNote={Propensity scores (PSs), a powerful bias-reduction tool, can balance treatment groups on measured covariates in nonexperimental studies. We demonstrate the use of multiple PS estimation methods to optimize covariate balance.We used secondary data from 1292 adults with nonpsychotic major depressive disorder in the Sequenced Treatment Alternatives to Relieve Depression trial (2001-2004). After initial citalopram treatment failed, patient preference influenced assignment to medication augmentation (n = 565) or switch (n = 727). To reduce selection bias, we used boosted classification and regression trees (BCART) and logistic regression iteratively to identify two potentially optimal PSs. We assessed and compared covariate balance.After iterative selection of interaction terms to minimize imbalance, logistic regression yielded better balance than BCART (average standardized absolute mean difference across 47 covariates: 0.03 vs. 0.08, matching; 0.02 vs. 0.05, weighting).Comparing multiple PS estimates is a pragmatic way to optimize balance. Logistic regression remains valuable for this purpose. Simulation studies are needed to compare PS models under varying conditions. Such studies should consider more flexible estimation methods, such as logistic models with automated selection of interactions or hybrid models using main effects logistic regression instead of a constant log-odds as the initial model for BCART.}, number={4}, journal={Annals of Epidemiology}, publisher={Elsevier BV}, author={Ellis, Alan R. and Dusetzina, Stacie B. and Hansen, Richard A. and Gaynes, Bradley N. and Farley, Joel F. and Stürmer, Til}, year={2013}, month={Apr}, pages={204–209} }
 @article{carey_crotty_morrissey_jonas_thaker_ellis_woodell_wines_viswanathan_2013, title={Future Research Needs for Evaluating the Integration of Mental Health and Substance Abuse Treatment with Primary Care}, volume={19}, ISSN={1538-1145}, url={http://dx.doi.org/10.1097/01.pra.0000435034.37685.ce}, DOI={10.1097/01.pra.0000435034.37685.ce}, abstractNote={Research needs are many in the current health care environment. In this article, we describe a novel method developed by the Agency for Healthcare Research and Quality (AHRQ) Evidence-based Practice Center Program for pri- oritizing areas for future research. Using a recent- ly published systematic review as a foundation, investigators worked with a diverse group of 10 stakeholders to identify and prioritize research needs. We enumerate 13 high-priority research needs, as determined by stakeholders who repre- sented researchers, funders, health care providers, and patients and families, and discuss considerations for specific study designs. Our findings suggest that future research on inte- grating mental health and primary care should focus first on a) identifying methods of integrat- ing primary care into specialty mental health settings, b) identifying cross-cutting strategies for integration across multiple mental health diagnostic categories as opposed to a separate strategy for each diagnostic category, and c) examining the use of information technology for integrating mental and general medical health care. Other priorities for consideration include examining the economic and organizational sus- tainability of successful integration models, identifying dissemination methods for various settings, examining the business case for inte- gration as well as methods of payment, assessing the cost-effectiveness of integration, and identi- fying key components of successful strategies. The importance of sustainability and economic justification for integrated care strategies was a recurring theme in discussions with the stake- holders. The ability to sustain integrated care in everyday practice remains to be proved and will depend in part on the level of incentives and sup- port provided through payment system reform, as well as the ability of practices to provide care efficiently. (Journal of Psychiatric Practice 2013; 19:345–359)}, number={5}, journal={Journal of Psychiatric Practice}, publisher={Ovid Technologies (Wolters Kluwer Health)}, author={Carey, Timothy S. and Crotty, Karen A. and Morrissey, Joseph P. and Jonas, Daniel E. and Thaker, Samruddhi and Ellis, Alan R. and Woodell, Carol and Wines, Roberta C and Viswanathan, Meera}, year={2013}, month={Sep}, pages={345–359} }
 @article{brookhart_freburger_ellis_wang_winkelmayer_kshirsagar_2013, title={Infection Risk with Bolus versus Maintenance Iron Supplementation in Hemodialysis Patients}, volume={24}, ISSN={1046-6673}, url={http://dx.doi.org/10.1681/ASN.2012121164}, DOI={10.1681/ASN.2012121164}, abstractNote={Intravenous iron may promote bacterial growth and impair host defense, but the risk of infection associated with iron supplementation is not well defined. We conducted a retrospective cohort study of hemodialysis patients to compare the safety of bolus dosing, which provides a large amount of iron over a short period of time on an as-needed basis, with maintenance dosing, which provides smaller amounts of iron on a regular schedule to maintain iron repletion. Using clinical data from 117,050 patients of a large US dialysis provider merged with data from Medicare's ESRD program, we estimated the effects of iron dosing patterns during repeated 1-month exposure periods on risks of mortality and infection-related hospitalizations during the subsequent 3 months. Of 776,203 exposure/follow-up pairs, 13% involved bolus dosing, 49% involved maintenance dosing, and 38% did not include exposure to iron. Multivariable additive risk models found that patients receiving bolus versus maintenance iron were at increased risk of infection-related hospitalization (risk difference [RD], 25 additional events/1000 patient-years; 95% confidence interval [CI], 16 to 33) during follow-up. Risks were largest among patients with a catheter (RD, 73 events/1000 patient-years; 95% CI, 48 to 99) and a recent infection (RD, 57 events/1000 patient-years; 95% CI, 19 to 99). We also observed an association between bolus dosing and infection-related mortality. Compared with no iron, maintenance dosing did not associate with increased risks for adverse outcomes. These results suggest that maintenance iron supplementation may result in fewer infections than bolus dosing, particularly among patients with a catheter.}, number={7}, journal={Journal of the American Society of Nephrology}, publisher={Ovid Technologies (Wolters Kluwer Health)}, author={Brookhart, M. Alan and Freburger, Janet K. and Ellis, Alan R. and Wang, Lily and Winkelmayer, Wolfgang C. and Kshirsagar, Abhijit V.}, year={2013}, month={Jul}, pages={1151–1158} }
 @article{kshirsagar_freburger_ellis_wang_winkelmayer_brookhart_2013, title={Intravenous Iron Supplementation Practices and Short-Term Risk of Cardiovascular Events in Hemodialysis Patients}, volume={8}, DOI={10.1371/journal.pone.0078930.g001}, number={11}, journal={PLoS One}, publisher={Public Library of Science (PLoS)}, author={Kshirsagar, A.V. and Freburger, J.K. and Ellis, A.R. and Wang, L. and Winkelmayer, W.C. and Brookhart, M.A.}, year={2013}, month={Nov}, pages={e78930} }
 @article{ellis_dusetzina_hansen_gaynes_farley_stürmer_2013, title={Investigating differences in treatment effect estimates between propensity score matching and weighting: a demonstration using STAR*D trial data}, volume={22}, ISSN={1053-8569}, url={http://dx.doi.org/10.1002/pds.3396}, DOI={10.1002/pds.3396}, abstractNote={ABSTRACT}, number={2}, journal={Pharmacoepidemiology and Drug Safety}, publisher={Wiley}, author={Ellis, Alan R. and Dusetzina, Stacie B. and Hansen, Richard A. and Gaynes, Bradley N. and Farley, Joel F. and Stürmer, Til}, year={2013}, month={Feb}, pages={138–144} }
 @article{thomas_ellis_2013, title={Patterns of healthcare use and employment among people with disabilities}, volume={6}, ISSN={1936-6574}, url={http://dx.doi.org/10.1016/j.dhjo.2012.11.008}, DOI={10.1016/j.dhjo.2012.11.008}, abstractNote={Employment rates among people with disabilities are low. Poor health is often cited as a barrier to work. Disability or a lack of disability-related resources may interfere with the ability to secure and maintain work. This paper presents an exploratory examination of the association between variation in service use and employment. The paper uses data from North Carolina Medicaid recipients age 18–64 who were eligible in fiscal year 2007 due to receipt of Supplemental Security Income (n = 60,190). Logistic regression was used to model employment as a function of variation in healthcare use, with conditional models stratifying by days of service use and unconditional models run by quantile of service use. People with the least service use (<12 days) had the highest employment rate (over 20%); those with the most service use (≥54 days) had the lowest employment rate (7.8%). Those in between displayed remarkably little variation in employment rate by level of service use. The amount of week-to-week variation in service use was positively associated with the probability of employment. Among Medicaid enrollees with disabilities who use outpatient services, amount of service use is negatively associated with employment and variation in use is positively associated with employment. Future research involving more extensive administrative data, primary data collection, and the use of mixed methods would improve understanding of these findings.}, number={2}, journal={Disability and Health Journal}, publisher={Elsevier BV}, author={Thomas, Kathleen C. and Ellis, Alan R.}, year={2013}, month={Apr}, pages={133–140} }
 @article{kshirsagar_freburger_ellis_wang_winkelmayer_brookhart_2013, title={The Comparative Short-term Effectiveness of Iron Dosing and Formulations in US Hemodialysis Patients}, volume={126}, ISSN={0002-9343}, url={http://dx.doi.org/10.1016/J.AMJMED.2012.11.030}, DOI={10.1016/J.AMJMED.2012.11.030}, abstractNote={Intravenous iron is used widely in hemodialysis, yet there are limited data on the effectiveness of contemporary dosing strategies or formulation type.We conducted a retrospective cohort study using data from the clinical database of a large dialysis provider (years 2004-2008) merged with administrative data from the US Renal Data System to compare the effects of intravenous iron use on anemia management. Dosing comparisons were bolus (consecutive doses ≥100 mg exceeding 600 mg during 1 month) versus maintenance (all other iron doses during the month); and high (>200 mg over 1 month) versus low dose (≤200 mg over 1 month). Formulation comparison was administration of ferric gluconate versus iron sucrose over 1 month. Outcomes were hemoglobin, epoetin dose, transferrin saturation, and serum ferritin during 6 weeks of follow-up.We identified 117,050 patients for the dosing comparison, and 66,207 patients for the formulation comparison. Bolus dosing was associated with higher average adjusted hemoglobin (+0.23 g/dL; 95% confidence interval [CI], 0.21-0.26), transferrin saturation (+3.31%; 95% CI, 2.99-3.63), serum ferritin (+151 μg/L; 95% CI, 134.9-168.7), and lower average epoetin dose (-464 units; 95% CI, -583 to -343) compared with maintenance. Similar trends were observed with high-dose iron versus low-dose. Iron sucrose was associated with higher adjusted average hemoglobin (+0.16 g/dL; 95% CI, 0.12-0.19) versus ferric gluconate.Strategies favoring large doses of intravenous iron or iron sucrose lead to improved measures of anemia management. These potential benefits should be weighed against risks, which currently remain incompletely characterized.}, number={6}, journal={The American Journal of Medicine}, publisher={Elsevier BV}, author={Kshirsagar, Abhijit V. and Freburger, Janet K. and Ellis, Alan R. and Wang, Lily and Winkelmayer, Wolfgang C. and Brookhart, M. Alan}, year={2013}, month={Jun}, pages={541.e1–541.e14} }
 @article{katz_dusetzina_farley_ellis_gaynes_castillo_stürmer_hansen_2012, title={Distressing Adverse Events After Antidepressant Switch in the Sequenced Treatment Alternatives to Relieve Depression (STAR*D) Trial: Influence of Adverse Events During Initial Treatment with Citalopram on Development of Subsequent Adverse Events with an Alternative Antidepressant}, volume={32}, ISSN={0277-0008}, url={http://dx.doi.org/10.1002/j.1875-9114.2011.01020.x}, DOI={10.1002/j.1875-9114.2011.01020.x}, abstractNote={Study ObjectiveTo determine whether distressing adverse events (DAEs) experienced during initial antidepressant treatment are associated with subsequent DAEs after switching to a second antidepressant.}, number={3}, journal={Pharmacotherapy: The Journal of Human Pharmacology and Drug Therapy}, publisher={Wiley}, author={Katz, Aaron J. and Dusetzina, Stacie B. and Farley, Joel F. and Ellis, Alan R. and Gaynes, Bradley N. and Castillo, Wendy C. and Stürmer, Til and Hansen, Richard A.}, year={2012}, month={Feb}, pages={234–243} }
 @article{thomas_ellis_konrad_morrissey_2012, title={North Carolina’s Mental Health Workforce: Unmet Need, Maldistribution, and No Quick Fixes}, volume={73}, ISSN={0029-2559 0029-2559}, url={http://dx.doi.org/10.18043/ncm.73.3.161}, DOI={10.18043/ncm.73.3.161}, abstractNote={Recent data show a maldistribution of psychiatrists in North Carolina and critical shortages in some areas. However, only 11 entire counties have official mental health professional shortage designation.This paper presents estimates of the adequacy of the county-level mental health professional workforce. These estimates build on previous work in 4 ways: They account for mental health need as well as provider supply, capture adequacy of the prescriber and nonprescriber workforce, consider mental health services provided by primary care providers, and account for travel across county lines by providers and consumers. Workforce adequacy is measured at the county level by the percentage of rieed for mental health visits that is met by the current supply of prescribers and nonprescribers.Ninety-five of North Carolina's 100 counties have unmet need for prescribers. In contrast, only 7 have unmet need for nonprescribers, and these counties have inadequate numbers of prescribers as well. To eliminate the deficit under current national patterns of care, the state would need about 980 more prescribers.Data limitations constrain findings to focus on percentage of met need rather than supplying exact counts of additional professionals needed. Estimates do not distinguish between public and private sectors of care, nor do they embody a standard of care.North Carolina is working to develop its mental health prescriber workforce. The Affordable Care Act provides new opportunities to develop the mental health workforce, innovative practices involving an efficient mix of professionals, and financing mechanisms to support them.}, number={3}, journal={North Carolina Medical Journal}, publisher={North Carolina Institute of Medicine}, author={Thomas, Kathleen C. and Ellis, Alan R. and Konrad, Thomas R. and Morrissey, Joseph P.}, year={2012}, month={May}, pages={161–168} }
 @article{hansen_dusetzina_ellis_stürmer_farley_gaynes_2012, title={Risk of adverse events in treatment-resistant depression: propensity-score-matched comparison of antidepressant augment and switch strategies}, volume={34}, ISSN={0163-8343}, url={http://dx.doi.org/10.1016/j.genhosppsych.2011.10.001}, DOI={10.1016/j.genhosppsych.2011.10.001}, abstractNote={The objective was to assess differences in adverse events between major depressive patients augmented with a second medication and patients switched to an alternative monotherapy after failing first-step treatment with citalopram.Adverse event profiles for second-step switch and augment medication strategies were compared using public data files from the Sequenced Treatment Alternatives to Relieve Depression (STAR*D) trial. In the STAR*D trial, participants failing citalopram selected acceptable next-step strategies and were randomized within acceptable strategies. This design resulted in clinically important differences when comparing across strategies, so a propensity-score-matched sample was created to compare switch (n=269) and augment (n=269) strategies.Incidence proportions of any adverse event and specific adverse events were similar between the augment and switch groups. The overall incidence proportion of any distressing event was 0.78 [95% confidence interval (CI) 0.72-0.84] in the augment group and 0.80 (95% CI 0.74-0.85) in the switch group. This contrasts unmatched analyses where distressing adverse events were less common in the augment group than the switch group (risk ratio 0.85, 95% CI 0.81-0.90).After adjusting for selection bias inherent in the STAR*D comparison of augment with switch, clinically meaningful differences in the adverse event profiles between these treatment strategies were not observed.}, number={2}, journal={General Hospital Psychiatry}, publisher={Elsevier BV}, author={Hansen, Richard A. and Dusetzina, Stacie B. and Ellis, Alan R. and Stürmer, Til and Farley, Joel F. and Gaynes, Bradley N.}, year={2012}, month={Mar}, pages={192–200} }
 @article{gaynes_dusetzina_ellis_hansen_farley_miller_stürmer_2012, title={Treating Depression After Initial Treatment Failure}, volume={32}, ISSN={0271-0749}, url={http://dx.doi.org/10.1097/JCP.0b013e31823f705d}, DOI={10.1097/JCP.0b013e31823f705d}, abstractNote={Objective Augmenting and switching antidepressant medications are the 2 most common next-step strategies for depressed patients failing initial medication treatment. These approaches have not been directly compared; thus, our objectives are to compare outcomes for medication augmentation versus switching for patients with major depressive disorder in the Sequenced Treatment Alternatives to Relieve Depression (STAR*D) clinical trial. Methods We conducted a retrospective analysis of participants aged 18 to 75 years with DSM-IV nonpsychotic depression who failed to remit with initial treatment in the STAR*D clinical trial (N = 1292). We compared depressive symptom remission, response, and quality of life among participants in each study arm using propensity score matching to minimize selection bias. Results The propensity-score-matched augment (N = 269) and switch (N = 269) groups were well balanced on measured characteristics. Neither the likelihood of remission (risk ratio, 1.14; 95% confidence level, 0.82–1.58) or response (risk ratio, 1.14; 95% confidence level, 0.82–1.58), nor the time to remission (log-rank test, P = 0.946) or response (log-rank test, P = 0.243) differed by treatment strategy. Similarly, quality of life did not differ. Post hoc analyses suggested that augmentation improved outcomes for patients tolerating 12 or more weeks of initial treatment and those with partial initial treatment response. Conclusions For patients receiving and tolerating aggressive initial antidepressant trials, there is no clear preference for next-step augmentation versus switching. Findings tentatively suggest that those who complete an initial treatment of 12 weeks or more and have a partial response with residual mild depressive severity may benefit more from augmentation relative to switching.}, number={1}, journal={Journal of Clinical Psychopharmacology}, publisher={Ovid Technologies (Wolters Kluwer Health)}, author={Gaynes, Bradley N. and Dusetzina, Stacie B. and Ellis, Alan R. and Hansen, Richard A. and Farley, Joel F. and Miller, William C. and Stürmer, Til}, year={2012}, month={Feb}, pages={114–119} }
 @article{gaynes_farley_dusetzina_ellis_hansen_miller_stürmer_2011, title={Does the presence of accompanying symptom clusters differentiate the comparative effectiveness of second-line medication strategies for treating depression?}, volume={28}, ISSN={1091-4269}, url={http://dx.doi.org/10.1002/da.20898}, DOI={10.1002/da.20898}, abstractNote={Background: We explored whether clinical outcomes differ by treatment strategy following initial antidepressant treatment failure among patients with and without clinically relevant symptom clusters. Methods: The Sequenced Treatment Alternatives to Relieve Depression (STAR*D) trial was used to examine depression remission and response in patients with coexisting anxiety, atypical features, insomnia, and low energy. We applied propensity scoring to control for selection bias that precluded comparisons between augmentation and switch strategies in the original trial. Binomial regressions compared the likelihood of remission or response among patients with and without symptom clusters for switch versus augmentation strategies (n = 269 per arm); augmentation strategy type (n = 565); and switch strategy type (n = 727). Results: We found no statistically significant difference in remission or response rates between augmentation or switch strategies. However, symptom clusters did distinguish among augmentation and switch strategies, respectively. For patients with low energy, augmentation with buspirone was less likely to produce remission than augmentation with bupropion (remission Risk Ratio (RR): 0.54, 95% CI: 0.35–0.85, response RR: 0.67, 95% CI: 0.43, 1.03). Also, for patients with low energy, switching to venlafaxine or bupropion was less likely to produce remission than switching to sertraline (RR: 0.59, 95% CI: 0.36–0.97; RR: 0.63, 95% CI: 0.38–1.06, respectively). Conclusions: Remission and response rates following initial antidepressant treatment failure did not differ by treatment strategy for patients with coexisting atypical symptoms or insomnia. However, some second‐step treatments for depression may be more effective than others in the presence of coexisting low energy. Subsequent prospective testing is necessary to confirm these initial findings. Depression and Anxiety, 2011. © 2011 Wiley Periodicals, Inc.}, number={11}, journal={Depression and Anxiety}, publisher={Wiley}, author={Gaynes, Bradley N. and Farley, Joel F. and Dusetzina, Stacie B. and Ellis, Alan R. and Hansen, Richard A. and Miller, William C. and Stürmer, Til}, year={2011}, month={Sep}, pages={989–998} }
 @article{fraser_guo_ellis_thompson_wike_li_2011, title={Outcome Studies of Social, Behavioral, and Educational Interventions}, volume={21}, ISSN={1049-7315 1552-7581}, url={http://dx.doi.org/10.1177/1049731511406136}, DOI={10.1177/1049731511406136}, abstractNote={This article describes the core features of outcome research and then explores issues confronting researchers who engage in outcome studies. Using an intervention research perspective, descriptive and explanatory methods are distinguished. Emphasis is placed on the counterfactual causal perspective, designing programs that fit culture and context, and developing nuanced explanations for program outcomes. Five emerging challenges are discussed: (a) adapting interventions to the contexts and cultures in which programs are to be implemented, (b) avoiding potentially false attributions of program failure due to differential implementation, (c) making causal inferences from observational data with propensity score analysis (PSA), (d) examining person-centered outcomes in program evaluation, and (e) adjusting for rater effects in longitudinal research.}, number={6}, journal={Research on Social Work Practice}, publisher={SAGE Publications}, author={Fraser, Mark W. and Guo, Shenyang and Ellis, Alan R. and Thompson, Aaron M. and Wike, Traci L. and Li, Jilan}, year={2011}, month={Apr}, pages={619–635} }
 @article{ellis_2010, title={The Administration of Psychotropic Medication to Children Ages 0–4 in North Carolina: An Exploratory Analysis}, volume={71}, ISSN={0029-2559 0029-2559}, url={http://dx.doi.org/10.18043/ncm.71.1.9}, DOI={10.18043/ncm.71.1.9}, abstractNote={The increasing use of psychotropic medication among preschool children raises concern because there are insufficient clinical guidelines and possible disparities.This study explored published administrative data (2007-2006) on the receipt of psychotropic medication by North Carolina Medicaid enrollees ages 0-4 by mental health catchment area. Quarterly prevalence statistics were examined and potential predictors of receipt were identified for future study.During the study period the state's quarterly prevalence ranged from 2.3 to 3.0 recipients per 1,000 enrollees (range in catchment areas: 0.5 to 9.8). The state rate peaked at 3.0 in the third quarter of 2002 and at 2.9 in the third quarter of 2004.The data are aggregated to a large area level and limited to Medicaid enrollees. The small number of catchment areas (36) limits the utility of statistical associations.Prevalence rates are high enough to deserve further exploration. Geographic variation exists. Psychotropic medication prescriptions for preschool children should be included as the state's mental health practitioners, policymakers, and planners discuss the service system and the mental health of children in our communities.}, number={1}, journal={North Carolina Medical Journal}, publisher={North Carolina Institute of Medicine}, author={Ellis, Alan R.}, year={2010}, month={Jan}, pages={9–14} }
 @article{ellis_morrissey_2009, title={Assessing Multiple Outcomes for Women with Co-Occurring Disorders and Trauma in a Multi-Site Trial: A Propensity Score Approach}, volume={36}, ISSN={0894-587X 1573-3289}, url={http://dx.doi.org/10.1007/S10488-009-0204-4}, DOI={10.1007/S10488-009-0204-4}, abstractNote={The current study assesses the ability of two promising propensity scoring methods to reduce selection bias in a set of secondary data from the women with co-occurring disorders and violence study (WCDVS), whose purpose was to evaluate the effect of integrated treatment for women with mental health, substance use, and trauma issues (N = 2,729). Weighting, the more successful method, is demonstrated in a re-analysis of 6- and 12-month WCDVS outcomes. In addition to demonstrating propensity score weighting, the current study increases confidence in earlier findings by considering multiple time points simultaneously and by controlling more completely for pre-treatment differences.}, number={2}, journal={Administration and Policy in Mental Health and Mental Health Services Research}, publisher={Springer Science and Business Media LLC}, author={Ellis, Alan R. and Morrissey, Joseph P.}, year={2009}, month={Jan}, pages={123–132} }
 @article{thomas_ellis_konrad_holzer_morrissey_2009, title={County-Level Estimates of Mental Health Professional Shortage in the United States}, volume={60}, ISSN={1075-2730 1557-9700}, url={http://dx.doi.org/10.1176/ps.2009.60.10.1323}, DOI={10.1176/appi.ps.60.10.1323}, abstractNote={Back to table of contents Previous article Next article ArticleFull AccessCounty-Level Estimates of Mental Health Professional Shortage in the United StatesKathleen C. Thomas M.P.H., Ph.D.Alan R. Ellis M.S.W.Thomas R. Konrad Ph.D.Charles E. Holzer Ph.D.Joseph P. Morrissey Ph.D.Kathleen C. Thomas M.P.H., Ph.D.Search for more papers by this authorAlan R. Ellis M.S.W.Search for more papers by this authorThomas R. Konrad Ph.D.Search for more papers by this authorCharles E. Holzer Ph.D.Search for more papers by this authorJoseph P. Morrissey Ph.D.Search for more papers by this authorPublished Online:1 Oct 2009https://doi.org/10.1176/ps.2009.60.10.1323AboutSectionsPDF/EPUB ToolsAdd to favoritesDownload CitationsTrack Citations ShareShare onFacebookTwitterLinked InEmail The shortage of mental health professionals has been a persistent concern for decades ( 1 , 2 , 3 , 4 ). Most recently, the President's New Freedom Commission on Mental Health report ( 5 ) and the Institute of Medicine's report Crossing the Quality Chasm: A New Health System for the 21st Century ( 6 ) raised concerns about workforce inadequacies. The Annapolis Coalition ( 7 , 8 ) has also called for workforce development to address shortages and poor distribution of behavioral health professionals across the United States. Most previous reports do not quantify the extent of shortages within professional groups. Quantifying mental health professional shortages can aid in describing systemic problems that are affected by shortages; such problems include bottlenecks in referral for mental health services, involvement of persons with mental illness with the justice system, homelessness, and widespread unmet need for community mental health services. Moving from identifying a shortage as a problem to quantifying it can help small areas and states in planning to meet constituents' needs. Documenting shortages may also motivate states and communities to improve recruitment, training, licensure, and retention. Even documenting an adequate supply of providers in a community can focus attention on the distribution of mental health providers between sectors. For example, mental health providers may gravitate toward employment in the private sector, leaving the public sector in extreme shortage, or they may be restricted to residential institutions (state hospitals or prisons) and may not be available to other community residents.For a number of reasons, it has been difficult to quantify and address mental health professional shortages. National small-area data on the workforce and need have not been collected regularly. Moreover, there has been a lack of consensus about what constitutes adequate supply. This article presents an approach to quantifying shortages of mental health professionals at a small-area level for the entire United States by using nationally representative secondary data. The goals of the study were to provide a comprehensive picture of current shortages and to motivate a renewed discussion of the data improvements and practice standards required to ultimately develop an adequate workforce of mental health professionals.MethodsConceptualizing shortage Mental health professional shortage was conceptualized as the percentage of need for mental health visits that was unmet within a county as of 2006. This measure improves on earlier measures ( 3 , 9 , 10 ) by taking into account variation in need as well as supply of professionals ( 11 , 12 ). Need was considered across the entire adult community, and we took into account the fact that individuals with and without serious mental illness have different levels of need. Children's needs and the mental health workforce to serve those needs have unique challenges ( 13 , 14 ) and are not addressed in this study. Supply was measured for the mental health professional workforce: psychiatrists, psychologists, advanced practice psychiatric nurses, social workers, licensed professional counselors, and marriage and family therapists. These six professions were categorized as prescribers (psychiatrists) and nonprescribers (all others) in an effort to simplify findings and still reflect this primary functional difference. Although it was recognized that other groups of professionals, such as personal aides, hypnotherapists, and registered counselors, serve people with mental health needs, we focused on major professional groups that are educated at the master's or doctoral level and licensed in most states to diagnose and treat mental disorders. Shortage was determined at the county level across the United States. The goal was to choose a unit of analysis that reflected local planning responsibility and mental health service use as well as one for which national secondary data were available. Mental health catchment areas developed in the 1970s once served this function, but they are rarely used by states now. Because census data and workforce counts could be measured at a county level, the county was chosen as the unit of analysis.Compiling the data Need was estimated as provider full-time equivalents needed in each of 3,140 counties. Details are provided in a companion article in this issue ( 11 ). Separate county-level need estimates were developed for prescribers and nonprescribers, for each of two county subpopulations: adults with and without serious mental illness. County prevalence of serious mental illness was developed with the use of a synthetic estimation procedure ( 11 , 15 , 16 ). These provider need estimates were based on utilization data from the National Comorbidity Survey Replication ( 15 ) and the Medical Expenditure Panel Survey ( 17 ). For people with serious mental illness, need was measured on the basis of actual utilization among users; for those without serious mental illness, need was measured on the basis of actual utilization by the entire population. Utilization was recorded in terms of outpatient visit minutes, which were translated into provider full-time equivalents. Need estimates were deflated to adjust for the portion of need met by primary care providers, which was based on county-level area scores indicating shortage of primary care health professionals; these scores were proposed recently to the Health Resources and Services Administration (HRSA) ( 18 ). This deflation adjustment ensures that the need estimates reflect need for mental health professionals only. In summary, need estimates reflect adult need for visits to mental health professionals for every U.S. county and account for the different levels of need among people with and without serious mental illness. Supply was estimated as provider full-time equivalents available in each county (N=3,140). Details are provided in a companion article in this issue ( 12 ). As with measures of need, separate estimates of visit minutes available were developed for prescribers (psychiatrists) and nonprescribers (psychologists, social workers, advanced practice psychiatric nurses, marriage and family therapists, and professional counselors). The data were compiled from professional associations, state licensure boards, and national certification boards in order to count all mental health professionals in both the public and private sectors for every county. Provider counts were translated into full-time equivalents based on professional practice patterns. In summary, the supply estimates reflect the full volume of met need for mental health professional visits with prescribers and nonprescribers for each U.S. county. Each county-level need and supply estimate was adjusted with a smoothing technique to account for travel across county boundaries for mental health services. The maximum amount of time that people travel for care is about 60 minutes ( 19 , 20 ). Therefore, for a given index county, the need and supply estimates of counties within a 60-minute radius were weighted and added to the estimates for the index county. The weights were generated with an exponential distance decay function e β d , such that a county zero minutes from the index county would receive a weight of 1, and a county 60 minutes from the index county would receive a weight close to 0 (.1). Travel times were measured on the intercounty distance matrix developed at the Oak Ridge National Laboratories, which estimates the travel times between county population centroids ( 21 ). Once weighted, the estimates were further scaled so that the national need and supply totals for prescribers and nonprescribers were unchanged by the smoothing process. The choice of an hour's travel time is supported by the literature and minimizes masking of shortage (which would occur with a larger radius). Each county's unmet need was calculated as the difference between its need and supply estimates. The shortage score represents unmet need as a proportion of total need in a county. Three scores were calculated: for prescribers, nonprescribers, and both groups combined. Prescriber and nonprescriber scores could be negative, indicating a surplus. The overall shortage score was based on the sum of positive prescriber and nonprescriber shortages. To describe the distribution of shortages across the United States, ordinary least-squares regression of overall shortage was estimated as a function of county characteristics.Results Over three-quarters (77%) of U.S. counties had a severe shortage of mental health prescribers or nonprescribers, with over half their need unmet. Eight percent of U.S. counties had a severe shortage of nonprescribers, with over half of their need unmet. Almost one in five counties (18%) in the nation had at least some unmet need for nonprescribers. Seventy-seven percent of U.S. counties had a severe shortage of prescribers, with over half of their need unmet. Nearly every county (96%) had at least some unmet need for prescribers. Table 1 provides statistics on the percentage of need unmet at the county level. Ordinary least-squares regression of the percentage of county overall need unmet as a function of county characteristics indicated that rurality and per capita income were the best predictors of unmet need (R 2 =.34). A 1-point increase in rurality on the 9-point Rural-Urban Continuum Code corresponded to an increase in unmet need of 3.3 percentage points. A $1,000 increase in per capita income corresponded to a decrease in unmet need of 1.3 percentage points. Table 1 Counties with unmet need for mental health professionals, by provider categoryTable 1 Counties with unmet need for mental health professionals, by provider categoryEnlarge tableFigure 1 shows the distribution of counties with unmet need across the United States. [The map can be viewed in closer detail as an online supplement to this article at ps.psychiatryonline.org .] The percentage of need that was unmet for prescribers, nonprescribers, and overall mental health professional workforce, respectively, was grouped into quartiles. As shown, counties with a high percentage of unmet need (darkest shading) are most pronounced in a north-south strip down the middle of the country and on the eastern side of the Rocky Mountains. Figure 1 Unmet need for mental health professionals among counties with an overall shortageDiscussionLimitations A number of compromises had to be made in order to generate a national picture of mental health professional shortages. The HRSA, which funded this work, required a current and rational method to designate mental health professional shortage areas across the United States ( 22 ). Because of cost and time considerations, the method was based on currently available secondary data. The county-level percentage of need that is unmet provides a relative measure of shortage across counties that meets HRSA's need for a shortage designation method to target limited resources on a national level. Because our estimates of "minutes needed per person" were somewhat arbitrary and because unusual distributions can occur across dimensions other than geography (such as in residential areas and with community services), this measure of relative shortage is not well suited for measuring the absolute shortfall in a particular county. The estimates of need were built from model-based estimates (not actual counts) of the number of people with and without serious mental illness in each county. The model predicting the prevalence of serious mental illness was derived from National Comorbidity Survey Replication (NCS-R) data ( 11 , 15 ). These data are the most current and in-depth data on mental disorders across the United States, but a number of refinements, such as a larger sample of individuals with serious mental illness and a broader sampling frame, would improve these estimates. This issue is discussed in more detail elsewhere ( 11 ). The estimates of need also rest on estimates of provider time needed by people with and without serious mental illness ( 11 ). Because standards quantifying the need for mental health treatment do not yet exist ( 23 ), the study goal was to identify a moderate standard of need for mental health services. Provider time needed was measured on the basis of utilization, and in an effort to avoid drastically underestimating time needed by individuals with serious mental illness, their need was measured on the basis of average use among service users (versus overall). Although not all people with serious mental illness require treatment at a given point in time, the estimate of provider time needed is still likely to be conservative because of suboptimal rates of service use in this group ( 24 ). The shortage measure is very sensitive to changes in the estimate of provider time needed, underscoring the importance of using shortage as a relative versus absolute measure among counties. The supply data for mental health professionals constitute the most comprehensive, detailed, and up-to-date information on the mental health workforce of which we are aware ( 12 ). Numerous validity checks lend confidence that these data reflect the geographic distribution of mental health professionals. However, to the extent that the compilation of supply data failed to generate correct counts for counties, some counties may be ranked incorrectly in terms of unmet need. Also, it should be noted that the study excludes the fields of pastoral counseling, clinical sociology, and psychosocial rehabilitation. Many providers in these fields may be licensed in one of the included mental health professions, but some are not. Furthermore, the study was unable to distinguish providers who work only in inpatient psychiatric care. To the extent that the time of these providers is not available to community-based clients, the estimated unmet need will be too low. The measures of unmet need are based on need and supply estimates that were smoothed with a 60-minute travel radius in order to account for travel of providers and consumers across county boundaries. A possible limitation of this smoothing technique is that it does not account for the fact that tertiary medical centers may draw people from a wider radius than other provider organizations. A larger travel radius might be more realistic in areas with such centers. The study compromised detail for the sake of parsimony by collapsing the mental health professions into prescribers and nonprescribers. Clearly, each profession brings a unique philosophy and focus to the care and treatment of mental illness. Local treatment modalities—assertive community treatment, for example ( 25 )—can capitalize on the strengths of each profession in order to provide individualized multidisciplinary treatment. On a national scale, practice patterns and supply data are too fluid to allow the precise matching of roles with actors in each county. It is clear, however, that prescribers and nonprescribers are not functionally substitutable. This implies that need for a nonprescriber may be filled by any type of nonprescriber, whereas the need for prescribers cannot be filled by nonprescribers. In the overall measure of unmet need, prescribers and nonprescribers have equal weight, to reflect that each group plays an equally critical role in treatment. Refinements in standards of care as well as more extensive epidemiological and workforce data may shift our estimates of the relative shortages of prescribers and nonprescribers for each county and the country as a whole. It is important to consider several issues about the delivery of mental health services that were not addressed by the estimated shortage scores. First, the need estimates and provider counts do not distinguish between public-sector services (where the greatest impairment exists) and private practice (where many consumers have less serious disorders). It is likely that the actual shortages in the public sector are greater than those in the private sector. Second, the scores do not assess the extent to which an appropriate continuum of care exists for mental health or the degree to which other needs, such as housing, are being met. Third, simply comparing estimated need with estimated supply does not address the quality of care or the degree to which professional practice is evidence based. All of these issues are important in determining whether an area's mental health services are adequate to meet the needs of residents.The regression of shortage as a function of county characteristics is subject to mild collinearity between the independent variables, because rural areas tend to have lower per capita income (r=-.49). The equation system is also subject to endogeneity because people with serious mental illness may be attracted to live in inexpensive areas (where other low earners live), but because people with serious mental illness have low employment rates and levels of earnings, they may also contribute to the low per capita earnings of an area. Although the regression results presented here make sense intuitively, these shortcomings should be addressed in future work that explores the factors leading to professional shortage.Implications These findings underscore the importance of the Annapolis Coalition's ( 7 , 8 ) call for workforce development. In particular, the widespread prescriber shortage and uneven distribution of nonprescribers identified here might be lessened by following the coalition's suggestions to strengthen the workforce through development of strategies to improve recruitment, retention, education, and leadership. The coalition has also discussed the need for better mental health professional workforce data that would improve shortage estimates and support evaluation of workforce development efforts. Small-area management entities can use this shortage information to distinguish between service gaps arising from problems in system organization and those resulting from actual shortage. Quantifying mental health professional shortage should help small areas target resources to fill workforce gaps. Quantifying mental health professional shortage should help states justify investing in efforts to alleviate shortage and distribution problems. States are already working on a number of fronts to address mental health professional shortage. For example, consumer-centered and peer-run care increases access, continuity of care, and satisfaction ( 26 ). Telepsychiatry to support rural mental health providers in developing treatment plans, managing medications, and following best-practice guidelines has been shown to be a way to promote distance learning and to stretch the mental health workforce to better meet needs in shortage areas ( 27 ). Integrating mental health with primary care through colocation of providers, expansion of primary care provider treatment roles, and increased opportunities for consultations between primary care clinicians and psychiatrists can improve access to mental health services, continuity, and quality of care ( 28 ). Expanded prescriptive authority for advanced practice psychiatric nurses or psychologists also enlarges the prescriber workforce ( 29 ). Psychiatric nurses can prescribe in many states; New Mexico and Louisiana have also extended prescribing privileges to psychologists. To date, the challenges involved in acquiring additional training, setting up an independent (likely rural) practice, and establishing referral linkages have limited the impact and spread of these policies. Many states are trying to bridge gaps through existing Medicaid policy—for example, by authorizing home- and community-based services and rehabilitation waivers ( 30 , 31 ) or by initiating Medicaid reimbursement for care management and mental health consultations ( 32 ) that expand the scope of covered services for individuals with mental illness. States are also expanding Medicaid policy through Medicaid buy-in programs that allow working adults with mental illness or other disabilities to keep their Medicaid coverage even as their earnings grow ( 33 ). A major challenge of the buy-in programs is establishing and maintaining enrollment. Next steps for future researchFuture efforts to identify mental health professional shortages would benefit from improvements in assessing both the need and the supply sides. The measurement of need should address additional populations such as persons with co-occurring substance use disorders, children, homeless adults, and those who are linguistically isolated. The measurement of supply should address the expanded practice scope of nurses and psychologists and should include additional professions (such as psychosocial rehabilitation). Although we recognize that data refinements come at a high cost, there are several that would greatly improve the data for use in generating shortage scores, even staying within the data framework currently available ( 11 , 12 ). The NCS-R data are the most current and in-depth data on mental disorders across the United States, but a number of refinements are desirable. These include a larger sample of individuals with serious mental illness, a broader sampling frame that includes people who are homeless or living in institutions, consistent handling of psychotic and other disorders, and a more thorough measurement of race and income. The supply data on mental health professionals constructed for this report represent the most comprehensive, detailed, and up-to-date information on the mental health workforce now available. Yet to the extent that the data compilation failed to generate correct counts for counties, some counties may receive inaccurate shortage scores. Better supply data would be comprehensive, consistent, and without duplication. ConclusionsThree-quarters of U.S. counties were estimated to have a severe shortage of prescribers, with over half their need unmet, ten times the number with a severe shortage of nonprescribers. Nearly all (96%) U.S. counties were found to have at least some prescriber shortage, whereas only 18% were found to have any nonprescriber shortage. Comprehensive standards of care, more extensive epidemiological data, and investment in a national workforce database would greatly improve these shortage estimates. But the big challenge for future work is to progress from simply describing shortages to overcoming them.Acknowledgments and disclosuresThis work was supported by contract HHSH-230200532038C from the HRSA. The authors acknowledge the help of the project officer, Andy Jordan, M.S.P.H.; their advisory board, which included Michael Almog, Ph.D., David Bergman, J.D., Tim Dall, M.S., Sheron R. Finister, Ph.D., John C. Fortney, Ph.D., Nancy P. Hanrahan, Ph.D., R.N., Sharon M. Jackson, M.S.W., L.C.S.W., Nina Gail Levitt, Ed.D., Ronald W. Manderscheid, Ph.D., Noel A. Mazade, Ph.D., Bradley K. Powers, Psy.D., Richard M. Scheffler, Ph.D., Laura Schopp, Ph.D., Lynn Spector, M.P.A., Marvin S. Swartz, M.D., and Joshua E. Wilk, Ph.D.; and the following individuals: Rick Harwood, Marlene Wicherski, Jessica Kohout, Ph.D., Lynn Bufka, Ph.D., Becky Corbett, A.C.S.W., Charles Housen, Tracy Whitaker, Ph.D., Paul Wing, Ph.D., Jim Fitch, Scott Barstow, Emily Wisniewski, Mark Holmes, Ph.D., Tom Ricketts, Ph.D., Jennifer Groves, M.B.A., Randy Randolph, M.P.R., Olivia Silber Ashley, Dr.P.H., Bob Bray, Ph.D., J. Valley Rachal, Ph.D., Tina McRee, M.A., Harold Goldsmith, Ph.D., Barbara Van Horne, M.B.A., Ph.D, Edward Norton, Ph.D., Gary Koch, Ph.D., Robert McConville, Sarah Curtis, Ph.D., Bruce Peterson, M.S., Susan Shafer, M.Ed., Susanne Phillips, M.S.N., F.N.P., Linda Beeber, Ph.D., R.N., Victoria Soltis-Jarrett, Ph.D., A.P.R.N.-B.C., and Cheryl Jones, Ph.D., R.N. The views expressed in this report do not necessarily reflect the official policies of the U.S. Department of Health and Human Services, nor does mention of organizations imply endorsement by the U.S. Government.The authors report no competing interests.Dr. Thomas, Mr. Ellis, Dr. Konrad, and Dr. Morrissey are affiliated with the Cecil G. Sheps Center for Health Services Research, University of North Carolina at Chapel Hill, 725 Martin Luther King Jr. Blvd., Campus Box 7590, Chapel Hill, NC 27599 (e-mail: [email protected]). Dr. Holzer is with the Department of Psychiatry and Behavioral Sciences, University of Texas Medical Branch, Galveston. Preliminary findings from this study were presented at a session on mental health workforce and needs assessment at the annual meeting of American Public Health Association, November 3–7, 2007, Washington, D.C.References1. Moritz T: A state perspective on psychiatric manpower development. Hospital and Community Psychiatry 30:775–777, 1979Google Scholar2. Tucker G, Turner J, Chapman R: Problems in attracting and retaining psychiatrists in rural areas. Hospital and Community Psychiatry 32:118–120, 1981Google Scholar3. Sierles F, Taylor M: Decline of US medical student career choice of psychiatry and what to do about it. American Journal of Psychiatry 152:1416–1426, 1995Google Scholar4. Goldman W: Is there a shortage of psychiatrists? Psychiatric Services 52:1587–1589, 2001Google Scholar5. Achieving the Promise: Transforming Mental Health Care in America. Pub no SMA-03-3832. Rockville, Md, Department of Health and Human Services, President's New Freedom Commission on Mental Health, 2003Google Scholar6. Institute of Medicine: Crossing the Quality Chasm: A New Health System for the 21st Century. Washington, DC, National Academies Press, 2001Google Scholar7. An Action Plan for Behavioral Health Workforce Development. Cincinnati, Ohio, Annapolis Coalition on the Behavioral Health Workforce, 2007. Available at www.annapoliscoalition.org Google Scholar8. Hoge MA, Morris JA, Stuart GW, et al: A national action plan for workforce development in behavioral health. Psychiatric Services 60:883–887, 2009Google Scholar9. Baldwin LM, Patanian MM, Larson EH, et al: Modeling the mental health workforce in Washington State: using state licensing data to examine provider supply in rural and urban areas. Journal of Rural Health 22:50–58, 2006Google Scholar10. Holzer C, Goldsmith H, Ciarlo J: The availability of health and mental health providers by population density. Journal of the Washington Academy of Sciences 86:25–33, 2000Google Scholar11. Konrad TR, Ellis AR, Thomas KC, et al: County-level estimates of need for mental health professionals in the United States. Psychiatric Services 60:1307–1314, 2009Google Scholar12. Ellis AR, Konrad TR, Thomas KC, et al: County-level estimates of mental health professional supply in the United States. Psychiatric Services 60:1315–1322, 2009Google Scholar13. Koppelman J: The provider system for children's mental health: workforce capacity and effective treatment. National Health Policy Forum Issue Brief 26:1–18, 2004Google Scholar14. Thomas CR, Holzer CE: The continuing shortage of child and adolescent psychiatrists. Journal of the American Academy of Child and Adolescent Psychiatry 45:1023–1031, 2006Google Scholar15. Kessler RC, Berglund P, Chiu WT, et al: The US National Comorbidity Survey Replication (NCS-R): design and field procedures. International Journal of Methods in Psychiatric Research 13:69–92, 2004Google Scholar16. Census 2000 Public Use Microdata Sample (PUMS). Washington, DC, US Census Bureau, 2003Google Scholar17. Cohen J: Methodology Report #1: Design and Methods of the Medical Expenditure Panel Survey Household Component. Rockville, Md, Agency for Health Care Policy and Research, 1997. Available at www.meps.ahrq.gov/data_files/publications/mr1mr1.shtml Google Scholar18. Ricketts TC, Goldsmith LJ, Holmes GM, et al: Designating places and populations as med}, number={10}, journal={Psychiatric Services}, publisher={American Psychiatric Association Publishing}, author={Thomas, Kathleen C., M.P.H., Ph and Ellis, Alan R., M.S.W. and Konrad, Thomas R., Ph.D. and Holzer, Charles E., Ph.D. and Morrissey, Joseph P., Ph.D.}, year={2009}, month={Oct}, pages={1323–1328} }
 @article{ellis_konrad_thomas_morrissey_2009, title={County-Level Estimates of Mental Health Professional Supply in the United States}, volume={60}, ISSN={1075-2730 1557-9700}, url={http://dx.doi.org/10.1176/ps.2009.60.10.1315}, DOI={10.1176/appi.ps.60.10.1315}, abstractNote={Back to table of contents Previous article Next article ArticleFull AccessCounty-Level Estimates of Mental Health Professional Supply in the United StatesAlan R. Ellis M.S.W.Thomas R. Konrad Ph.D.Kathleen C. Thomas M.P.H., Ph.D.Joseph P. Morrissey Ph.D.Alan R. Ellis M.S.W.Search for more papers by this authorThomas R. Konrad Ph.D.Search for more papers by this authorKathleen C. Thomas M.P.H., Ph.D.Search for more papers by this authorJoseph P. Morrissey Ph.D.Search for more papers by this authorPublished Online:1 Oct 2009https://doi.org/10.1176/ps.2009.60.10.1315AboutSectionsView articleSupplemental MaterialPDF/EPUB ToolsAdd to favoritesDownload CitationsTrack Citations ShareShare onFacebookTwitterLinked InEmail View article Of approximately $100 billion spent annually on U.S. mental health care, about 70% pays for the labor of mental health professionals ( 1 ). Yet we lack valid and reliable workforce data, and academic research rarely focuses on the mental health workforce ( 2 ). A workforce crisis currently affects diverse areas—recruitment, retention, training and technical assistance, compensation, career advancement, and geographic distribution ( 2 )—making the need for comprehensive workforce data even more critical. Various workforce reports can be found in the literature, but none provides a detailed national picture of the mental health professions. Prior studies have described the characteristics, needs, and practice patterns of the national mental health workforce and compared the professions ( 3 ; also unpublished documents: "Practitioner Research Network: Summary of Initiative and Findings," Substance Abuse and Mental Health Services Administration [SAMHSA], Center for Substance Abuse Treatment [CSAT]; "Practitioner Services Network II Initiative: Summary of Findings," SAMHSA, CSAT, 2003), discussed how rural workforce needs have been and could be addressed ( 4 ), assessed the effects of licensure laws on workforce availability ( 5 ), examined cross-sectional data on individual professions ( 6 , 7 , 8 , 9 ), and conducted within-state, small-area analyses ( 10 , 11 ). This study built on this literature by compiling national county-level data to examine the geographic distribution of providers in six mental health professions and the correlates of county-level provider supply. Our main goal was to present profiles that would be useful for workforce planning at local, state, and national levels. A secondary goal was to provide information about the availability and comprehensiveness of existing workforce data to the research and practice communities. Further information is provided in two companion articles in this issue exploring county-level need for and shortages of mental health professionals in the United States ( 12 , 13 ). MethodsData sources Because this study was part of a project involving the designation of shortage in the mental health profession ( 14 ), which is a responsibility of the Health Resources and Services Administration (HRSA), we used HRSA's definition of "mental health professionals": advanced practice psychiatric nurses, licensed professional counselors, marriage and family therapists, psychiatrists, psychologists, and social workers. Although other professionals and nonprofessionals contribute significantly to mental health services, these six groups constitute a majority of mental health professionals, and information about them is critically important for mental health policy and planning. Our goal was to count clinically active providers (specifically, those who are actively engaged in the diagnosis and treatment of mental disorders) rather than the larger population of clinically trained providers (those who have been trained at the master's or doctoral level to perform these functions). We explored several potential data sources (see below). Their advantages and disadvantages are summarized in a table available as an online supplement to this article at ps.psychiatryonline.org . The typical tradeoff is between coverage (for example, national scope or inclusion of multiple professions) and identification of the correct group of providers. The Bureau of Labor Statistics has employer-reported data on psychiatric nurses, family therapists, psychiatrists, psychologists, and social workers, but these data are limited by aggregation to the state or metropolitan statistical area (MSA) level, lack of information on professional degree, failure to distinguish among professions, and exclusion of self-employed providers. Census data and the Area Resource File ( 15 ) are easily accessible national data sets that contain counts of nurses, psychologists, and social workers. However, they do not cover areas with populations under 50,000, indicate professional degree, or distinguish between clinical and other specialties. For most professions, state licensing data would yield the best counts of clinically active providers, because licensure is usually required for clinical practice and is not trivial to maintain. However, licensing data are difficult to obtain because they are not centrally collected, are often confidential, and are maintained by state boards, many of which have few resources. Also, licensing data are not standardized, may not include provider specialty, and may include the same individual in multiple professions or states.Certification and professional association membership data are national in scope but yield undercounts of clinically active providers because membership is voluntary and certification is not required for most professions and states (especially where licensure is required). Also, membership data often do not indicate provider specialty.Licensing, certification, and especially membership data include some inactive practitioners, who generally cannot be distinguished from clinically active providers. Licensing data may be less affected by this limitation because of renewal and continuing education requirements. Most data sets from any source lack consistent, up-to-date information on practice locations, do not incorporate multiple practice locations, and do not distinguish between home and work addresses.Data collectionConsidering the data source characteristics, we preferred licensing data where available, then membership data, then certification data. Therefore, we combined licensing counts from state boards, certification counts from national credentialing organizations, and membership counts from professional associations, always choosing the most preferred data source available for a given state and profession. These data were difficult to obtain but allowed us to estimate with reasonable accuracy the number of clinically active providers in each profession at the county level. Also, we were able to use some multistate licensing data previously assembled by others.Even when counts were available at the zip code level, they were aggregated to the county level because a zip code could be associated with either a practice location or a home address, likely making the county-level counts a less error-prone approximation of practice locations. Aggregation also made the counts comparable across professions, because counts of marriage and family therapists were not available below the county level. Furthermore, whereas zip codes were designed for mail delivery, county boundaries are a meaningful basis for mental health service planning, which is often done for counties or county groups. Although zip code areas are often nested within counties, this is not always the case; therefore, a table of approximate zip-to-county conversions was used.For nurses we used psychiatric nursing certification data provided in 2003 by the American Nurses Credentialing Center. Zip-level counts were generated and were converted to county-level counts by using the table of approximate zip-to-county associations. Membership data were not used for nursing because the American Nurses Association does not record specialty and the American Psychiatric Nurses Association has data for only a subset of psychiatric nurses.For licensed professional counselors, the American Counseling Association (ACA) provided licensing information for 38 states. For the other 13 states, certification data from the National Board of Certified Counselors Web site were used. Zip-level counts were converted to county-level counts.Similarly, for marriage and family therapists, the American Association of Marriage and Family Therapists provided county-level counts based on licensing data where available (26 states) and on clinical membership otherwise (25 states). For psychiatrists, data from the American Medical Association's ( 16 ) Physician Masterfile in regard to individual general psychiatrists were used. Residents and those not treating patients were excluded, and office (versus home) address was used where available. Zip-level counts were converted to county-level counts. For psychologists, the American Psychological Association (APA) provided data sufficient to generate zip-level counts of licensed clinically active members, which were converted to county-level counts.For social workers, the National Association of Social Workers (NASW) provided zip-level counts of members at the master's level (M.S.W.); these were converted to county-level counts.The University of North Carolina's Public Health Institutional Review Board determined that this study did not require board approval.Data management Data cleaning and validity checks were performed, and the data were scaled to match the best available state-level counts. Data cleaning began with the exclusion of inactive, suspended, and nonclinical providers and the correction of discrepancies between address components. County-level counts from multiple sources were compared where possible (for example, for counselors, National Board for Certified Counselors certification counts versus ACA licensing counts; for marriage and family therapists, 2006 membership versus 2003 membership and licensing counts; for social workers, NASW membership counts versus approximate licensing counts provided by the Center for Health Workforce Studies at the State University of New York at Albany). State-level counts were compared across professions. As a further check, state-level counts were compared with those in Mental Health, United States, 2004 ( 17 ). For psychology and social work, state totals were also compared with state-level counts collected from state licensing boards. Comparisons between databases were made with correlations, plots, and regression diagnostics (such as residual plots and influence statistics). Discrepancies and extreme counts were investigated and were corrected where possible. For example, when NASW state totals for social workers were initially compared with state-level licensing counts, the correlation was .95, with the NASW membership total amounting to about 57% of the state licensing count on average (as expected), but diagnostic analysis suggested further investigation of counts for five states, resulting in the correction of two errors in the state licensing data. Similarly, when the state totals for marriage and family therapy were compared to Mental Health, United States, 2004 counts, diagnostic data suggested further investigation of counts for three states, resulting in the correction of one error from the Mental Health, United States, 2004 chartbook (for New Hampshire) through consultation via the state licensing board's Web site. For professions requiring multiple data sources, the cleaned counts were scaled so that state totals matched the best available state-level counts of clinically active providers. For counselors we used state-level counts from the ACA's annual survey of state licensing boards where available (47 states). In the other four states, we inflated by 3% the state-level counts from Mental Health, United States, 2004 ( 17 ). (On average the ACA counts exceeded the chartbook counts by 3%. We assumed that the ACA counts, which were more recent, reflected real increases in the number of licensees.) For marriage and family therapists we scaled membership counts to match Mental Health, United States, 2004 (except for New Hampshire, as mentioned above) because the state-level counts in the chartbook reflect a consensus among experts. Because membership data yield undercounts, the psychology counts were increased by a factor of 1.896, which is the estimated ratio of licensed clinically active psychologists to licensed clinically active APA members based on APA data and on estimates reported in Mental Health, United States, 2004 . Social work membership counts required no scaling because only 52% of licensed social workers were specialized in mental health ( 8 ), and we estimated that our counts represented approximately that proportion of licensed social workers. Table 1 summarizes the data sources used for each profession, the results of comparison with the chartbook, and the scaling factors used. In most cases our data source was the same as that used in the chartbook; slight deviations from a perfect correlation were due to our use of more recent data where available. Table 1 Data source selected by provider typeTable 1 Data source selected by provider typeEnlarge tableStatistical analysisThe geographic distribution of professionals was examined with descriptive statistics and a national choropleth map. County-level counts for the six professions were correlated with each other. Provider-to-population ratios (based on 2006 population) were correlated with the 2006 population and with county characteristics based on census 2000 data: population density (population per square mile of total area), MSA status (1 if a county is in an MSA, otherwise 0), rurality (2003 Rural-Urban Continuum Code, which ranges from 1 to 9, with 1 indicating the most urban and 9 the most rural), frontier status (1 if population density is less than 7, otherwise 0), indicator variables for census region (Northeast, Midwest, South, or West), per capita income, and percentage of population in poverty. Several variables related to population size and to density were included because theory did not suggest a preferred indicator. The District of Columbia was included as one of 3,140 counties and one of 51 states.ResultsFigure 1 and Table 2 summarize the county-level distribution of the 353,398 mental health professionals. [ Figure 1 can be viewed in closer detail as an online supplement to this article at ps.psychiatryonline.org .] Table 2 classifies counties into three groups on the basis of the Rural-Urban Continuum Code. The concentration of providers (per 10,000 population) varied greatly across counties, both within professions and overall. For every profession, the highest concentrations of providers were in metropolitan areas, especially in the Northeast and West, and the lowest concentrations were in rural areas. Rural counties that were not adjacent to metropolitan areas typically had slightly lower concentrations of providers than did other rural counties, based on comparisons of median provider-to-population ratios ( Table 2 ). Figure 1 Number of mental health professionals, by county, among counties with mental health professionalsTable 2 Distribution of mental health professionals among 3,140 countiesTable 2 Distribution of mental health professionals among 3,140 countiesEnlarge table We were interested in describing the extent to which providers in different mental health professions were distributed similarly across the country. Table 3 shows correlations among county-level counts. When all counties were included (values above the diagonal), most of the associations were strong. Marriage and family therapists stood out as the exception. Because over half of marriage and family therapists were in California, the correlations were reexamined with California excluded. Here (values below the diagonal in Table 3 ) the marriage and family therapist counts were much more strongly associated with the others. Table 3 Correlations among county-level provider counts, with and without California countiesTable 3 Correlations among county-level provider counts, with and without California countiesEnlarge table The correlations between provider-to-population ratio and county-level characteristics ( Table 4 ) showed, not surprisingly, that providers in each profession—especially psychiatrists, psychologists, and social workers—tended to be concentrated in high-population and urban areas. This was true even when taking into account characteristics such as county area (in the case of population density and frontier status) or adjacency to metropolitan areas (in the case of MSA status and rurality). Table 4 Correlations between number of mental health care providers per 10,000 population and county characteristicsTable 4 Correlations between number of mental health care providers per 10,000 population and county characteristicsEnlarge tableThe clearest regional effect was the concentration of providers in the Northeast, which was strongest among social workers and weak to nonexistent among licensed professional counselors and marriage and family therapists. For each profession there was a weak negative association between provider-to-population ratio and percentage of population in poverty (r=-.06 to -.22 with California included), and there was a moderate positive association between provider-to-population ratio and per capita income (r=.25 to .52 with California included). The correlations for marriage and family therapists (r=-.06 and .28, respectively) and licensed professional counselors (r=-.09 and .25, respectively) were smaller than those for other professions. For marriage and family therapists, excluding California counties had the effect of lowering the correlations between provider-to-population ratio and population (from .18 to .10) and between provider-to-population ratio and per capita income (from .28 to .22), in addition to changing the correlations with regional variables. In general, marriage and family therapists, counselors, and psychiatric nurses appeared to be less concentrated in higher-income urban areas, compared with providers in the other three professions.DiscussionWe identified about 350,000 clinically active mental health providers in six professions. Social workers and licensed professional counselors formed the largest groups; psychiatrists and advanced practice psychiatric nurses constituted the smallest. Marriage and family therapists are unique in that 54% of them are located in California, but otherwise there were fairly strong positive associations among county-level provider counts across the six professions.Providers in all six groups tended to be in urban, high-population, high-income counties; aside from marriage and family therapists they were concentrated in the Northeast. Based on our descriptive results, much of the variation in provider location is probably explained by region, county-level income variables, and variables related to population size and density. The unique distribution of marriage and family therapists is likely largely a result of the concentration of marriage and family therapy graduate programs in California. However, there are a few notable differences in geographic distribution among the professions. For example, social workers were especially concentrated in the Northeast. Also, psychiatrists, psychologists, and social workers had especially similar distributions across counties and were more heavily concentrated than the other professions in population-dense, metropolitan, higher-income counties. A partial explanation may be that these three professions serve consumers with greater average severity of illness and that areas with higher population density have the resources to provide services to these consumers. Other contributing factors may be the locations of graduate programs for each profession, the unique history of each profession, and the fact that psychiatry, clinical psychology, and clinical social work are relatively older professions. However, the weaker associations for other professions ( Tables 3 and 4 ) may also be due in part to the facts that multiple data sources were used for counseling and psychology and older certification data were used to count nurses. The variation in size among the professions is important. For example, psychiatrists and nurses had the smallest numbers by far, and many members of the psychiatry profession are reaching retirement age ( 3 ). Aside from a negligible number of psychologists, only psychiatrists and some nurses can prescribe psychotropic medications. Assuming that prescription medication continues to be a key component of mental health treatment, factors such as these will need to be included in careful workforce planning to maintain or increase the supply of prescribers (for example, by expanding these professions or extending prescriptive authority more widely). Also, each profession's relative size (along with variables such as professional status and income) is likely to affect its level of influence on county, state, and federal policy. This should be considered by policy makers interested in balancing consumer needs with the wants of professional stakeholders. Limitations This study has limitations that deserve mention, including several involving study scope. We excluded pastoral counseling, clinical sociology, and psychosocial rehabilitation, which account for over 5,000 providers certified for clinical work ( 17 , 18 , 19 ). We also excluded providers with lower levels of licensure or certification, primary care practitioners, and peer providers. Ideally each provider population and its unique role and focus (such as prescribing versus psychosocial therapy, an individual versus a systems approach, and adult versus child clientele) would be considered in workforce assessment and planning. Provider supply is considered in the absence of information about need, demand, or typical utilization or about the breadth and quality of services—all important factors in workforce planning. Finally, we have only approximated practice locations and have not accounted for travel across county boundaries. There are also data quality limitations. Licensing data would be preferred, but licensing is not required for all professions, and for the licensed professions we did not have the resources and authorization necessary to obtain data from all states. Therefore we combined eight data sets, each of which probably has different sources of random error and systematic bias. The determination of practice location was subject to error. We did not always have information about clinical specialty or clinically active status; because this information was not available for social workers and their job responsibilities vary widely, this issue affects our counts for that profession especially. We believe that we adjusted our psychology and social work estimates appropriately by using national scaling factors, but at the county level, random error makes our counts for these professions unstable, especially for counties with small populations. This limits the utility of the data for local workforce planning. Finally, our provider counts do not reflect important information related to provider availability, such as service sector (public or private), hours worked per week, and hours per week in direct contact with clients.The correlations presented here should be interpreted with caution. Some of the variables involved are dichotomous (MSA status, regional variables, and frontier status), and the others have distributions that are skewed or that otherwise deviate from normality. In particular, our measure of rurality is not strictly a ratio-level variable. Also, our analysis did not control for the clustering of counties within states. In a thorough investigation of the relationships between provider count and county characteristics, these issues would need to be addressed, perhaps in part through the use of a nonlinear random-effects model. Nonetheless, Pearson correlations serve the purpose of this study by providing a simple summary that broadly describes the geographic distribution of mental health professionals.ImplicationsWe compiled, cleaned, and calibrated data from national certification, state licensure, and national professional association membership records in order to develop a comprehensive, current, nationwide county-level profile of the mental health professional workforce. Despite the limitations discussed above, these data yield simple but valuable descriptive information and allowed us to discuss some of the data gaps and issues that need to be addressed in order to facilitate mental health workforce planning. One reasonable inference from our data is that rural, low-income counties have relatively few mental health professionals and therefore are likely candidates for interventions such as the training of local clinicians or the provision of incentives and infrastructure to facilitate clinical practice. Another is that there is important geographic variation in the relative contribution of each mental health profession to the total pool of providers. (The concentration of marriage and family therapists in California is an extreme example.) Workforce planning and policy analysis should consider the unique combination of professions in each area.It is difficult to tease out the relationships between provider counts and population-related variables. For example, provider-to-population ratios for some professions had a weaker relationship with population density than with MSA status or rurality, even though all three of these population-related variables were correlated. Understanding these complex relationships may require more in-depth investigation, such as single-state analyses of geographic distribution or qualitative analyses of each profession's practice patterns and location preferences.The limitations of our data highlight important national data needs. National workforce planning efforts would benefit from the central collection of standardized practice information from clinically active providers in all mental health professions, including specific practice information such as location, specialty, service sector, hours worked per week, and hours per week in direct contact with clients. In addition to simplifying the assembly of national small-area data, centralized data collection could provide state boards with more efficient mechanisms and improved reporting, help state and local governments to identify underserved areas, increase the efficiency of state-level workforce planning, and potentially serve as a mechanism for representative surveys of practitioners.ConclusionsRural, low-income counties have the fewest mental health professionals per capita, and these counties would be appropriate targets for interventions such as the training of local clinicians or the provision of incentives and infrastructure to facilitate clinical practice. Because there is substantial variation across counties in the proportion of mental health professionals belonging to each profession, workforce planning and policy analysis should consider the unique combination of professions in each area. National workforce planning efforts and state licensing boards would both benefit from the central collection of standardized practice information from clinically active providers in all mental health professions.Acknowledgments and disclosuresThis work was supported by contract HHSH-230200532038C from the HRSA. The authors acknowledge the help of the project officer, Andy Jordan, M.S.P.H.; their advisory board, which included Michael Almog, Ph.D., David Bergman, J.D., Tim Dall, M.S., Sheron R. Finister, Ph.D., John C. Fortney, Ph.D., Nancy P. Hanrahan, Ph.D., R.N., Sharon M. Jackson, M.S.W., L.C.S.W., Nina Gail Levitt, Ed.D., Ronald W. Manderscheid, Ph.D., Noel A. Mazade, Ph.D., Bradley K. Powers, Psy.D., Richard M. Scheffler, Ph.D., Laura Schopp, Ph.D., Lynn Spector, M.P.A., Marvin S. Swartz, M.D., and Joshua E. Wilk, Ph.D.; and the following individuals: Marlene Wicherski, Jessica Kohout, Ph.D., Lynn Bufka, Ph.D., Becky Corbett, A.C.S.W., Charles Housen, Tracy Whitaker, Ph.D., Paul Wing, Ph.D., David Bergman, J.D., Nancy Hanrahan, Ph.D., Jim Fitch, Scott Barstow, Emily Wisniewski, Olivia Silber Ashley, Ph.D., Bob Bray, Ph.D., J. Valley Rachal, Ph.D., Tina McRee, M.A., Barbara Van Horne, M.B.A., Ph.D., Robert McConville, Susan Shafer, M.Ed., Linda Beeber, Ph.D., R.N., Victoria Soltis-Jarrett, Ph.D., A.P.R.N.-B.C., and Cheryl Jones, Ph.D., R.N. The views expressed in this report do not necessarily reflect the official policies of the U.S. Department of Health and Human Services, nor does mention of organizations imply endorsement by the U.S. Government.The authors report no competing interests.The authors are affiliated wi}, number={10}, journal={Psychiatric Services}, publisher={American Psychiatric Association Publishing}, author={Ellis, Alan R., M.S.W. and Konrad, Thomas R., Ph.D. and Thomas, Kathleen C., M.P.H., Ph and Morrissey, Joseph P., Ph.D.}, year={2009}, month={Oct}, pages={1315–1322} }
 @article{konrad_ellis_thomas_holzer_morrissey_2009, title={County-Level Estimates of Need for Mental Health Professionals in the United States}, volume={60}, ISSN={1075-2730 1557-9700}, url={http://dx.doi.org/10.1176/ps.2009.60.10.1307}, DOI={10.1176/appi.ps.60.10.1307}, abstractNote={Back to table of contents Previous article Next article ArticleFull AccessCounty-Level Estimates of Need for Mental Health Professionals in the United StatesThomas R. Konrad Ph.D.Alan R. Ellis M.S.W.Kathleen C. Thomas M.P.H., Ph.D.Charles E. Holzer Ph.D.Joseph P. Morrissey Ph.D.Thomas R. Konrad Ph.D.Search for more papers by this authorAlan R. Ellis M.S.W.Search for more papers by this authorKathleen C. Thomas M.P.H., Ph.D.Search for more papers by this authorCharles E. Holzer Ph.D.Search for more papers by this authorJoseph P. Morrissey Ph.D.Search for more papers by this authorPublished Online:1 Oct 2009https://doi.org/10.1176/ps.2009.60.10.1307AboutSectionsPDF/EPUB ToolsAdd to favoritesDownload CitationsTrack Citations ShareShare onFacebookTwitterLinked InEmail The report of the President's New Freedom Commission called for the U.S. mental health care system to be transformed into a consumer-centered system that is focused on recovery and delivers excellent care without disparities ( 1 ). Such a transformation will require large-scale workforce development efforts, such as training and redistribution initiatives, that are informed by a national perspective on county-level need. Although there is evidence of widespread unmet need in this country ( 2 , 3 ), we lack a current and accurate assessment of need for mental health professionals on a county-specific basis. In the absence of national minimum standards for mental health services (based on mental disorder, symptoms, and functional limitations, for example), need estimation is challenging for at least four reasons. First, the prevalence of mental disorders (and therefore the appropriate level of utilization) varies across demographic and socioeconomic groups ( 4 , 5 , 6 , 7 ). Second, not everyone needs mental health services, and among those who do, the level of need varies greatly ( 8 , 9 , 10 ). For example, people with diagnosed mental disorders are more likely to require mental health services than are those without ( 10 ). Third, some need for mental health services is met not by mental health professionals but by primary care physicians, who see about half of those with mental health needs and cover about 21% of all mental health visits ( 11 ). This complicates the measurement of utilization. Fourth, there is often a mismatch between the level of need and the amount of services received. This may be related to the fact that perceived need correlates very weakly with measured morbidity ( 12 ). Among people with serious mental illness, 54% do not receive timely care ( 2 , 3 , 13 ). On the other hand, about half of those who receive mental health care do not have serious mental illness, in part because decreased stigma among the more affluent segments of the population has turned mental health services into something of a consumer good ( 3 , 14 , 15 ). The goal of this study was to formulate the best estimates possible, given current data, to aid in workforce planning and to call attention to the need for better data. We address the lack of a national picture of local need and the challenges of utilization-based need estimation by using a synthetic estimation technique combined with U.S. census data to develop county-level estimates of the prevalence of serious mental illness throughout the United States. This method was developed in response to an invitation by the Health Resources and Services Administration (HRSA) to update the process for designating shortages in the supply of mental health professionals ( 16 ); the process is used for resource allocation, including the placement of specific types of mental health professionals throughout the country by the National Health Service Corps in response to the needs of local communities. Assessments of health workforce requirements typically use one of three methods: need estimation based on disease prevalence and the time required for treatment, economic forecasting of demand for services, or benchmarking to a known population (for example, a health maintenance organization) ( 17 ). Our approach was to combine the need and demand methods. Estimates of disease prevalence and current utilization were generated from nationally representative survey data, but the utilization estimates for adults with serious mental illness were adjusted in order to include people who need treatment and are not receiving it. A full report describes the method and results in more detail ( 18 ). MethodsEstimating county prevalence of serious mental illness Following several earlier estimates of the prevalence of serious mental illness ( 19 , 20 ), we generated county-level prevalence estimates by applying the predicted probabilities from demographic models to cross-tabulations of census population data. The predictive models were based on the most recent nationally representative data on the distribution of psychiatric disorders in the general population: the 2001 National Comorbidity Survey Replication (NCS-R) ( 21 ). The NCS-R (N=9,282) contains information about diagnoses, functional limitations, mental health service utilization, and demographic descriptors. Defining serious mental illness operationally is a complex task involving diagnosis, functional impairment, and duration ( 22 ). Although a conceptual definition is published ( 23 ), no commonly accepted computational algorithm exists ( 24 ). We classified NCS-R respondents as having serious mental illness if they met three requirements: first, Composite International Diagnostic Interview diagnosis of bipolar I, bipolar II, mania, major depressive disorder, agoraphobia, generalized anxiety disorder, hypomania, panic disorder, posttraumatic stress disorder, social phobia, or specific phobia; second, a high level of disability as indicated by either the inability to carry out normal activities as a result of mental health problems for at least 120 days in the past year or a mean self-rated impairment level of 7 or higher on a 10-point scale across four dimensions (home, work, relationships, and social life); and third, age at onset at least two years less than the respondent's age (in order to exclude those whose disorders had lasted less than 12 months). Our diagnosis, disability, and duration criteria were designed to identify the group of people of primary concern to the mental health service system—those who have significant mental health service needs. Therefore, although we included a relatively broad range of diagnoses, these criteria reflect a higher level of impairment than does the traditional definition of serious mental illness. Because we planned to compare our need estimates with workforce estimates, and because data on the substance abuse treatment workforce are quite limited, our criteria did not include substance use disorders. We also excluded some typically less severe mental disorders: adult separation anxiety disorder, attention-deficit disorder, subthreshold bipolar symptoms, conduct disorder, dysthymia, intermittent explosive disorder, oppositional defiant disorder, and panic attack. Unfortunately we were unable to incorporate schizophrenia and other psychotic disorders into the inclusion criteria because the NCS-R does not elicit enough information to diagnose psychotic disorders, its psychosis screening questions are subject to false positives, and the psychosis section does not include questions about functional impairment. Analyses of clinical data not in the public release files suggest that 79% of people with nonaffective psychosis would be identified through other diagnoses as having serious mental illness ( 25 , 26 ). NCS-R data were used to model the probability of having serious mental illness in relation to demographic predictors. We used a two-stage logit model to generate the predicted probabilities, because poverty level, an important predictor of serious mental illness, was available only for the 61% of NCS-R respondents who completed part 2 of the survey instrument, and participation in part 2 was strongly associated with having serious mental illness. (Among NCS-R respondents with serious mental illness, 99% responded to part 2.) In order to maximize use of the available information, we predicted the probability of part 2 participation as a function of age, sex, race, marital status, and education level. Then, given part 2 participation, we predicted the probability of having serious mental illness as a function of poverty level along with the other demographic predictors. We combined the predicted probabilities from the two models to yield overall predicted probabilities of serious mental illness (based on all six predictors) that were not conditional on part 2 participation. The set of predictors was limited to demographic variables in order to avoid reinforcing any disparities related to other predictors (such as region). The small cell sizes for many combinations of predictor values did not support interaction terms in the models. Using synthetic estimation procedures ( 27 , 28 ), we then applied the predicted probabilities to county subpopulations, defined by all the permutations of the same demographic variables used in the logit models. We used demographic data from the U.S. census for 2006 ( 29 , 30 ) to yield estimates of the number of people with serious mental illness for each county subpopulation. The county subpopulation estimates were then aggregated to obtain an estimate of the number of people with serious mental illness in each county. Because these prevalence estimates were derived from NCS-R sample data rather than measured in the entire U.S. household population, they are subject to sampling error. Taking into account the NCS-R sampling design, we used balanced repeated replication to assess the effect of sampling variation on the prevalence estimates. Prevalence estimates were generated from 88 randomly selected half-samples from the NCS-R data, and a 95% confidence interval was constructed around the whole-sample prevalence estimate for each county and state. Estimating need for mental health servicesWe estimated the need for mental health professionals separately for individuals with and without serious mental illness based on each group's use of mental health services. For people with serious mental illness, we chose the NCS-R as the most appropriate data source; for those without serious mental illness, we used the 2000 Medical Expenditure Panel Survey (MEPS). Both data sets provide detailed information on service use and types of mental health providers.A total of 377 persons met the above three criteria for serious mental illness. Using data on outpatient services from the 356 individuals in the NCS-R sample with adequate data, we estimated mental health service need. We assumed that everyone with serious mental illness should receive outpatient mental health services at some time over the course of a year and estimated their need for mental health services as the mean number of visit minutes among service users. Using provider categories defined in the NCS-R instrument, we generated separate estimates of need for prescribers (psychiatrists, general practitioners, or family physicians) and for nonprescribers (psychologists, social workers, counselors, or other mental health professionals such as psychotherapists or mental health nurses). Visits to primary care providers for mental health care were included in order to estimate the full extent of need for mental health services. The role of primary care providers is addressed below. Because the NCS-R survey obtained utilization data only from individuals who met diagnostic criteria, need among people without serious mental illness was estimated with the 2000 MEPS ( 31 ), which has a large sample of noninstitutionalized civilians. Individuals who appeared to have serious mental illness (that is, those with ICD-9 codes 295–301 or 312 who rated their conditions as "serious" in any data collection round; weighted proportion 2.1% of respondents) were excluded to avoid duplication with the NCS-R sample, yielding a very large sample of persons without serious mental illness (N=16,418). Based on respondents' self-report data, we defined mental health visits as outpatient visits involving a possible mental health provider (physician, nurse, nurse practitioner, psychologist, social worker, or other) and one of the following: psychotherapy, a psychotherapeutic drug, or diagnosis of a mental disorder. Because it was not appropriate to assume that every individual without serious mental illness should receive outpatient mental health services over the course of a year, the estimate of need for providers was the mean number of visit minutes overall (not only among service users). Again we generated separate estimates of need for prescribers and need for nonprescribers. As with the population with serious mental illness, we included primary care providers because the goal was to estimate the total need for mental health services. Our estimates exclude inpatient services, general health care providers other than physicians and nurses, hotlines, religious or spiritual advisors, support groups, self-help, and complementary and alternative medical professionals.We converted minutes of services needed for the populations with and without serious mental illness to full-time-equivalent (FTE) estimates of prescribers and nonprescribers, using practice pattern data from various sources (32; also unpublished data: American Psychiatric Association's National Survey of Psychiatric Practice, 2002; Center for Substance Abuse Treatment's Practitioner Services Network II, 2003). Conversion factors reflected that 71% of nonprescribers' time (range 64%–79% across professions) and 60% of psychiatrists' time is spent in direct contact with patients. (These factors imply that psychiatrists typically provide 1,208 hours of direct patient care per year, whereas other mental health professionals typically provide 1,410 hours.) For each provider category (prescribers and nonprescribers), the sum of the need estimates for people with and without serious mental illness was used as a preliminary county-level estimate of total need. To account for the portion of need that is met by primary care providers, we adjusted the preliminary need estimates. Primary care providers account for about 21% of all mental health visits; however, the scope of their mental health practice is constrained by their mental health training and the physical health issues competing for their limited time ( 11 , 33 , 34 , 35 , 36 , 37 ). Also, primary care providers tend to see a mix of patients with less severe mental health problems than the problems of clients seen by mental health professionals ( 38 , 39 ). In order to acknowledge the role of primary care providers in addressing mental health need while estimating conservatively the appropriate size of their contribution, the need estimate was reduced by 15% in counties where there is a sufficient primary care workforce (in other words, there is no shortage). We chose 15% as the factor because the metric of need here was visit minutes rather than number of visits; because we found no empirical guide in choosing a cutoff point, the choice was arbitrary. This percentage was prorated on the basis of the proportion of the county's primary care need that is met, according to the shortage score proposed by Ricketts and colleagues ( 40 ) (calculated with 1998 data). The University of North Carolina's Public Health Institutional Review Board determined that this study did not require board approval.ResultsPrevalence of serious mental illness The results of the two-stage logit model to predict the probability of having serious mental illness in relation to demographic predictors are presented in Table 1 . The two-stage logit model had 65% sensitivity and 64% specificity with a predicted probability threshold of .04. The area under the receiver operating characteristic curve was .71, which indicates acceptable fit ( 41 ). Table 1 Two-stage logit model predicting serious mental illness among respondents to the 2001 National Comorbidity Survey ReplicationTable 1 Two-stage logit model predicting serious mental illness among respondents to the 2001 National Comorbidity Survey ReplicationEnlarge table Our method yielded a prevalence estimate of serious mental illness in the NCS-R sample of 3.9% (model-based prevalence 3.7%), which is in line with national estimates derived from other sources ( 13 , 42 , 43 , 44 , 45 ) ( Table 2 ). The highest estimates counted persons with any mental disorder in the past 12 months; measures with shorter recall periods or narrower frequency and severity criteria yielded lower estimates. Our criteria were among the strictest in that they did not include substance use disorders or psychosis screen information and required significant functional impairment. Thus we attempted to identify a group of people who have both serious disorders and a level of functional impairment that necessitates significant service use. Table 2 Estimates of the prevalence of serious mental illness in the United States, by sourceTable 2 Estimates of the prevalence of serious mental illness in the United States, by sourceEnlarge tableSynthetic estimation resulted in state-level prevalence estimates (N=51) that ranged from 3.2% to 4.5% (mean±SD=3.8%±.3%); county-level estimates (N=3,140) of prevalence of serious mental illness ranged from 2.5% to 9.0% (4.0%±.5%). The width of the 95% confidence interval (CI) around the prevalence estimate for each county or state derived from the half-sample replications gives an indication of the effect of sampling variation on the estimates. At the state level, the width of the CI ranged from .6 to 1.8, with a mean and median of .9. At the county level, it ranged from .5 to 4.1, with a mean and median of 1.2. (The mean relative standard errors were .06±.01 and .08±.01 at the state and county levels respectively.)Service use and requirements for health professionalsWe calculated the proportion of respondents who were service users and the mean total hours of visits along with CIs. About half of adults with serious mental illness used services; they typically spent 10.54 hours per year (CI=5.46–15.63) with nonprescriber mental health professionals and 4.38 hours per year (CI=3.40–5.37) with primary care physicians or prescriber mental health professionals. Less than 10% of adults without serious mental illness used specialized mental health services. Overall, adults without serious mental illness spent only 7.8 minutes (CI=5.4–9.6) with nonprescriber mental health professionals and 12.6 minutes (CI=10.8–14.4) with primary care physicians or prescriber mental health professionals in the reference year.Table 3 displays our estimates with CIs of the mental health professional FTEs required to treat the U.S. adult household population (calculated as described above). For 2006 we estimated that the U.S. adult household population with serious mental illness was 8,138,223, and another 210,106,179 adults without serious mental illness required a much lower amount of mental health services. County-level adjustments for the contribution of primary care physicians reduced the national estimates of mental health professional FTEs needed by 14.5%. Under our assumptions (detailed above), approximately 56,462 FTE prescribers and 68,581 FTE nonprescribers are needed to provide services to the U.S. adult household population. Figure 1 shows the distribution of need among counties after considering prevalence and primary care availability. The total number of FTE mental health professionals needed per county has a wide range, from near zero to 4,000 (40±124, median=12). [The map can be viewed in closer detail as an online supplement to this article at ps.psychiatryonline.org .] Table 3 National 2006 estimates of full-time-equivalent (FTE) mental health professionals required to serve the U.S. adult household population, by type of provider and mental illness status of the populationTable 3 National 2006 estimates of full-time-equivalent (FTE) mental health professionals required to serve the U.S. adult household population, by type of provider and mental illness status of the populationEnlarge tableFigure 1 Number of full-time-equivalent mental health professionals needed in the United States, by countyDiscussion Clearly, the quality of our estimate of the national level of need for mental health services is only as good as the data and assumptions we used. It is important to note, for example, that the estimates presented here do not reflect the needs of children or the needs of adults who are homeless, in the military, or living in institutions. Because of the lack of data for people living in institutions and the difficulty of distinguishing long and short hospital stays, the estimates presented here are limited to need for outpatient visits. They do not reflect local neighborhood, community, or personal factors that affect individual need, such as stressors and environmental stigma ( 46 , 47 ), nor do they speak to the most appropriate mix of prescribers and nonprescribers. Although we believe that we treated race and ethnicity appropriately, variation in the prevalence of serious mental illness associated with race, ethnicity, and linguistic isolation is not well understood. Further, although we believe that the county is the appropriate unit of analysis given the available data and the goals of the effort, there may be within-county variation in need not reflected here. Because our focus was on licensed mental health professionals, our estimates do not reflect need for providers such as registered counselors or registered hypnotherapists, nor do they address need for frontline workers such as hospital- or community-based aides, who are an important foundation of mental health services ( 48 ). Although these estimates were adjusted to account for the role of primary care providers, the validity of our adjustment factor is untested. Furthermore, there are three major sources of error associated with our estimates of need. First, there is statistical error associated with modeling serious mental illness status in the NCS-R. Our two-stage logit model had acceptable fit but was nonetheless subject to prediction error. Second, although our county-level prevalence estimates had relative standard errors that met the usual reliability criterion of less than .30, there was sampling error associated with estimating the prevalence of serious mental illness by applying the NCS-R model to census data. Our NCS-R and MEPS utilization estimates were subject to sampling error as well. (These estimates also had relative standard errors less than .30.) Third, there is error associated with collecting census data. To the extent that the census data were inaccurate (for example, undercounting people who are likely to have serious mental illness or incorrectly assessing low income, transient residence, or household composition), our prevalence estimates will be imprecise. Estimating need in stages compounds the error from these three sources. We are aware of no contemporary standard by which to validate our need estimates. However, it is worth noting that our estimates of need for psychiatrists, which translated to about 25.9 psychiatrists per 100,000 adult population (corrected for primary care substitution), substantially exceed the need-based standard developed almost 30 years ago by the Graduate Medical Education National Advisory Committee ( 49 ) of 15.4 psychiatrists per 100,000 population. This may not be surprising, given changes in mental health treatment patterns over the past several decades. For example, about 38% of the mental health professional FTEs in our estimate would be required to treat adults with serious mental illness who now live in the community, although such individuals might have been long-stay hospital patients in the 1970s. Conclusions This article presents a method to estimate the need for mental health professionals in individual counties throughout the United States and reports preliminary national estimates of need. Our estimates are probably most useful when taken as an expression of relative rather than absolute need. Our continuous measure of county-level need could be used in conjunction with other information, such as mental health professional supply or met need ( 50 , this issue), in order to target resources based on a given threshold or mental health professional shortage measure ( 51 , this issue) to guide policy decisions. Assessing local need in absolute terms would require a more detailed classification of levels of need, specific population estimates for different levels of functioning, and comprehensive standards for level of service use. We view our preliminary estimates of the national need for mental health professionals as a starting point that mental health planners, educators, and workforce analysts can improve on in future work. Acknowledgments and disclosuresThis work was supported by contract HHSH-230200532038C from the Health Resources and Services Administration. The authors acknowledge the help of the project officer, Andy Jordan, M.S.P.H.; their advisory board, which included Michael Almog, Ph.D., David Bergman, J.D., Tim Dall, M.S., Sheron R. Finister, Ph.D., John C. Fortney, Ph.D., Nancy P. Hanrahan, Ph.D., R.N., Sharon M. Jackson, M.S.W., L.C.S.W., Nina Gail Levitt, Ed.D., Ronald W. Manderscheid, Ph.D., Noel A. Mazade, Ph.D., Bradley K. Powers, Psy.D., Richard M. Scheffler, Ph.D., Laura Schopp, Ph.D., Lynn Spector, M.P.A., Marvin S. Swartz, M.D., and Joshua E. Wilk, Ph.D.; and the following individuals: Rick Harwood, Marlene Wicherski, Jessica Kohout, Ph.D., Lynn Bufka, Ph.D., Olivia Silber Ashley, Dr.P.H., Bob Bray, Ph.D., J. Valley Rachal, Ph.D., Mark Holmes, Ph.D., Edward Norton, Ph.D., and Gary Koch, Ph.D. The views expressed in this report do not necessarily reflect the official policies of the U.S. Department of Health and Human Services, nor does mention of organizations imply endorsement by the U.S. Government.The authors report no competing interests.Dr. Konrad, Mr. Ellis, Dr. Thomas, and Dr. Morrissey are affiliated with the Cecil G. Sheps Center for Health Services Research, University of North Carolina at Chapel Hill, 725 Martin Luther King Jr. Blvd., Campus Box 7590, Chapel Hill, NC 27599 (e-mail: [email protected]). Dr. Holzer is with the Department of Psychiatry and Behavioral Sciences, University of Texas Medical Branch, Galveston. Preliminary findings from this study were presented at a session on mental health workforce and needs assessment at the annual meeting of American Public Health Association, November 3–7, 2007, Washington, D.C.References1. Achieving the Promise: Transforming Mental Health Care in America. Pub no SMA-03-3832. Rockville, Md, Department of Health and Human Services, President's New Freedom Commission on Mental Health, 2003Google Scholar2. Wang P, Berglund P, Kessler R: Recent care of common mental disorders in the United States. Journal of General and Internal Medicine 15:284–292, 2001Google Scholar3. Katz S, Kessler R, Frank R, et al: Mental health care use, morbidity, and socioeconomic status in the United States and Ontario. Inquiry 34:38–49, 1997Google Scholar4. Kessler R, Berglund P, Demler O, et al: Lifetime prevalence and age-of-onset distributions of DSM-IV disorders in the National Comorbidity Survey Replication. Archives of General Psychiatry 62:593–602, 2005Google Scholar5. Breslau J, Aguilar-Gaxiola S, Kendler K, et al: Specifying race-ethnic differences in risk for psychiatric disorder in a USA national sample. Psychological Medicine 35:1–12, 2005Google Scholar6. Goldsmith H, Holzer C, Manderscheid R: Neighborhood characteristics and mental illness. Evaluation and Program Planning 21:211–225, 1998Google Scholar7. Muntaner C, Eaton W, Diala C, et al: Social class, assets, organizational control and the prevalence of common groups of psychiatric disorders. Social Science and Medicine 47:2043–2053, 1998Google Scholar8. Mechanic D: Is the prevalence of mental disorders a good measure of the need for services? Health Affairs 22(5):8–20, 2003Google Scholar9. Ciarlo J, Tweed D: Implementing indirect needs-assessment models for planning state mental health and substance abuse services. Evaluation and Program Planning 15:195–210, 1992Google Scholar10. Shapiro S, Skinner E, Kramer M, et al: Measuring need for mental health services in a general population. Medical Care 23:1033–1043, 1985Google Scholar11. Regier D, Shap}, number={10}, journal={Psychiatric Services}, publisher={American Psychiatric Association Publishing}, author={Konrad, Thomas R., Ph.D. and Ellis, Alan R., M.S.W. and Thomas, Kathleen C., M.P.H., Ph and Holzer, Charles E., Ph.D. and Morrissey, Joseph P., Ph.D.}, year={2009}, month={Oct}, pages={1307–1314} }
 @article{isett_ellis_topping_morrissey_2009, title={Managed Care and Provider Satisfaction in Mental Health Settings}, volume={45}, ISSN={0010-3853 1573-2789}, url={http://dx.doi.org/10.1007/S10597-008-9171-6}, DOI={10.1007/S10597-008-9171-6}, abstractNote={We assess the satisfaction of mental health providers using four dimensions from the medical practice literature--degree of autonomy, relationship with patients, compensation, and administrative burden--and extend current work on professional satisfaction to include frontline service providers rather than only psychiatrists or other physicians. In contrast to results reported for primary care settings, we find that the impact of managed care on satisfaction is minimal for the mental health providers in our study of a Medicaid capitation demonstration in the southeastern US. Instead, variables relevant to everyday working conditions have an important effect on job satisfaction.}, number={3}, journal={Community Mental Health Journal}, publisher={Springer Science and Business Media LLC}, author={Isett, Kimberley R. and Ellis, Alan R. and Topping, Sharon and Morrissey, Joseph P.}, year={2009}, month={Jun}, pages={209–221} }
 @article{ellis_2009, title={Using SAS to calculate betweenness centrality}, volume={29}, number={1}, journal={CONNECTIONS: Official Journal of the International Network for Social Network Analysis}, author={Ellis, A.R.}, year={2009}, pages={26–32} }
 @article{cusack_morrissey_ellis_2008, title={Targeting Trauma-related Interventions and Improving Outcomes for Women with Co-occurring Disorders}, volume={35}, ISSN={0894-587X 1573-3289}, url={http://dx.doi.org/10.1007/S10488-007-0150-Y}, DOI={10.1007/S10488-007-0150-Y}, abstractNote={National attention to the effects of interpersonal trauma has led mental health systems to adopt policies on trauma-related services; however, there is a lack of clarity regarding targeting of these services. Data from the Women, Co-occurring Disorders and Violence Study (WCDVS) were reanalyzed by grouping women on their baseline PTSD and substance abuse presentation and assessing the differential response to an integrated mental health/substance abuse intervention. Treatment effects were largest for subgroups characterized by high levels of PTSD, whereas the effects for those in the low symptom group were near zero. These findings underscore the need for clinicians to conduct careful assessments of trauma-related symptoms and to target the most intensive trauma-related interventions to individuals with PTSD symptoms.}, number={3}, journal={Administration and Policy in Mental Health and Mental Health Services Research}, publisher={Springer Science and Business Media LLC}, author={Cusack, Karen J. and Morrissey, Joseph P. and Ellis, Alan R.}, year={2008}, month={May}, pages={147–158} }
 @article{thomas_ellis_mclaurin_daniels_morrissey_2007, title={Access to Care for Autism-Related Services}, volume={37}, ISSN={0162-3257 1573-3432}, url={http://dx.doi.org/10.1007/S10803-006-0323-7}, DOI={10.1007/S10803-006-0323-7}, abstractNote={This paper identifies family characteristics associated with use of autism-related services. A telephone or in-person survey was completed during 2003-2005 by 383 North Carolina families with a child 11 years old or younger with ASD. Access to care is limited for racial and ethnic minority families, with low parental education, living in nonmetropolitan areas, and not following a major treatment approach. Service use is more likely when parents have higher stress. Families use a broad array of services; the mix varies with child ASD diagnosis and age group. Disparities in service use associated with race, residence and education point to the need to develop policy, practice and family-level interventions that can address barriers to services for children with ASD.}, number={10}, journal={Journal of Autism and Developmental Disorders}, publisher={Springer Science and Business Media LLC}, author={Thomas, Kathleen C. and Ellis, Alan R. and McLaurin, Carolyn and Daniels, Julie and Morrissey, Joseph P.}, year={2007}, month={Mar}, pages={1902–1912} }
 @article{landis_gaynes_morrissey_vinson_ellis_domino_2007, title={Generalist care managers for the treatment of depressed medicaid patients in North Carolina: A pilot study}, volume={8}, ISSN={1471-2296}, url={http://dx.doi.org/10.1186/1471-2296-8-7}, DOI={10.1186/1471-2296-8-7}, abstractNote={Abstract}, number={1}, journal={BMC Family Practice}, publisher={Springer Science and Business Media LLC}, author={Landis, Suzanne E and Gaynes, Bradley N and Morrissey, Joseph P and Vinson, Nina and Ellis, Alan R and Domino, Marisa E}, year={2007}, month={Mar} }
 @article{lee_morrissey_thomas_craig carter_ellis_2006, title={Assessing the Service Linkages of Substance Abuse Agencies with Mental Health and Primary Care Organizations}, volume={32}, ISSN={0095-2990 1097-9891}, url={http://dx.doi.org/10.1080/00952990500328620}, DOI={10.1080/00952990500328620}, abstractNote={Fragmentation of substance abuse treatment represents a major barrier to effective treatment for individuals with cooccurring substance abuse and mental and physical health disorders. Linkages of substance abuse treatment organizations with primary care and mental health agencies are widely considered to be a feasible way to integrate services. In this study, we analyzed information collected from a national sample of 62 outpatient substance abuse treatment units (OSATs) to understand the extent of services linkages in these organizations and to identify facilitators and barriers to service linkages. Results showed that OSATs had limited service linkages with primary care and mental health providers. The cited barriers to linkages included clients' financial problems, managed care restrictions, and limited organizational capacity. Onsite service provision was implemented in some OSATs. The pattern of service linkages in OSATs appeared to reflect the health needs of substance abuse clients.}, number={1}, journal={The American Journal of Drug and Alcohol Abuse}, publisher={Informa UK Limited}, author={Lee, Shoou-Yih D. and Morrissey, Joseph P. and Thomas, Kathleen C. and Craig Carter, W. and Ellis, Alan R.}, year={2006}, month={Jan}, pages={69–86} }
 @article{morrissey_ellis_gatz_amaro_reed_savage_finkelstein_mazelis_brown_jackson_et al._2005, title={Outcomes for women with co-occurring disorders and trauma: Program and person-level effects}, volume={28}, ISSN={0740-5472}, url={http://dx.doi.org/10.1016/j.jsat.2004.08.012}, DOI={10.1016/j.jsat.2004.08.012}, abstractNote={Six-month outcomes are evaluated from a 9-site quasi-experimental study of women with mental health and substance use disorders who have experienced physical or sexual abuse who enrolled in either comprehensive, integrated, trauma-informed, and consumer/survivor/recovering person-involved services (N = 1023) or usual care (N = 983). Mental health, post-traumatic stress symptoms, and substance use outcomes are assessed with multilevel regression models, controlling for program and personal characteristics. Person-level variables predict outcomes independent of intervention condition and, to a small extent, moderate intervention and program effects. In sites where the intervention condition provided more integrated counseling than the comparison condition, there are increased effects on mental health and substance use outcomes; these effects are partially mediated by person-level variables. These results encourage further research to identify the longer-term effects of integrated counseling for women with co-occurring disorders and trauma histories.}, number={2}, journal={Journal of Substance Abuse Treatment}, publisher={Elsevier BV}, author={Morrissey, Joseph P. and Ellis, Alan R. and Gatz, Margaret and Amaro, Hortensia and Reed, Beth Glover and Savage, Andrea and Finkelstein, Norma and Mazelis, Ruta and Brown, Vivian and Jackson, Elizabeth W. and et al.}, year={2005}, month={Mar}, pages={121–133} }
 @article{morrissey_jackson_ellis_amaro_brown_najavits_2005, title={Twelve-Month Outcomes of Trauma-Informed Interventions for Women With Co-occurring Disorders}, volume={56}, ISSN={1075-2730 1557-9700}, url={http://dx.doi.org/10.1176/appi.ps.56.10.1213}, DOI={10.1176/appi.ps.56.10.1213}, abstractNote={OBJECTIVE
Women with co-occurring mental health and substance use disorders frequently have a history of interpersonal violence, and past research has suggested that they are not served effectively by the current service system. The goal of the Women, Co-occurring Disorders, and Violence Study was to develop and test the effectiveness of new service approaches specifically designed for these women.


METHODS
A quasi-experimental treatment outcome study was conducted from 2001 to 2003 at nine sites. Although intervention specifics such as treatment length and modality varied across sites, each site used a comprehensive, integrated, trauma-informed, and consumer-involved approach to treatment. Substance use problem severity, mental health symptoms, and trauma symptoms were measured at baseline, and follow-up data were analyzed with prospective meta-analysis and hierarchical linear modeling.


RESULTS
A total of 2,026 women had data at the 12-month follow-up: 1,018 in the intervention group and 1,008 in the usual-care group. For substance use outcomes, no effect was found. The meta-analysis demonstrated small but statistically significant overall improvement in women's trauma and mental health symptoms in the intervention relative to the usual-care comparison condition. Analysis of key program elements demonstrated that integrating substance abuse, mental health, and trauma-related issues into counseling yielded greater improvement, whereas the delivery of numerous core services yielded less improvement relative to the comparison group. A few person-level characteristics were associated with increases or decreases in the intervention effect. These neither moderated nor supplanted the effects of integrated counseling.


CONCLUSIONS
Outcomes for women with co-occurring disorders and a history of violence and trauma may improve with integrated treatment.}, number={10}, journal={Psychiatric Services}, publisher={American Psychiatric Association Publishing}, author={Morrissey, Joseph P. and Jackson, Elizabeth W. and Ellis, Alan R. and Amaro, Hortensia and Brown, Vivian B. and Najavits, Lisa M.}, year={2005}, month={Oct}, pages={1213–1222} }
 @article{morrissey_stroup_ellis_merwin_2002, title={Service Use and Health Status of Persons With Severe Mental Illness in Full-Risk and No-Risk Medicaid Programs}, volume={53}, ISSN={1075-2730 1557-9700}, url={http://dx.doi.org/10.1176/appi.ps.53.3.293}, DOI={10.1176/appi.ps.53.3.293}, abstractNote={OBJECTIVE
The service use patterns and health status outcomes of Medicaid recipients with severe mental illness in a system that assigned full financial risk to managed care organizations through capitation and a system that paid for mental health care on a no-risk fee-for-service basis were compared.


METHODS
With use of a quasi-experimental design, initial interviews (time 1) and follow-up interviews six months later (time 2) were conducted among 92 clients in the full-risk group and 112 clients in the no-risk group. Regression models were used to compare self-reported service use and health status between the two groups.


RESULTS
Service use patterns differed between the two groups. When symptom severity at time 1 was controlled for, clients in the full-risk group were more likely to have received case management but less likely to report contact with a psychiatrist or to have received counseling than clients in the no-risk group. When health status at time 1 was controlled for, clients in the full-risk group reported poorer mental health at time 2 than clients in the no-risk group. When physical health status at time 1 was controlled for, clients in the full-risk group reported poorer physical health at time 2 than clients in the no-risk group.


CONCLUSIONS
Capitation was associated with lower use of costly services. Clients with serious mental illness in the full-risk managed care system had poorer mental and physical health outcomes than those in the no-risk system.}, number={3}, journal={Psychiatric Services}, publisher={American Psychiatric Association Publishing}, author={Morrissey, Joseph P. and Stroup, T. Scott and Ellis, Alan R. and Merwin, Elizabeth}, year={2002}, month={Mar}, pages={293–298} }
 @article{stroup_morrissey_ellis_blank_2001, volume={29}, ISSN={0894-587X}, url={http://dx.doi.org/10.1023/A:1014384413652}, DOI={10.1023/A:1014384413652}, abstractNote={This study examined predictors of family burden (assistance in daily living, supervision, and subjective concern) for family members of Medicaid recipients with severe mental illness in two regions of Virginia. In the Richmond area, mental health services were provided on a no-risk fee-for-service basis, while in Tidewater these services were provided through a risk-based capitated contract with a managed care organization. No differences in family burden were attributable to the risk-based payment system. Predictors of increased family burden were (a) more reported client symptoms and disruptive behaviors, (b) status as a parent, and (c) living with the client.}, number={2}, journal={Administration and Policy in Mental Health}, publisher={Springer Science and Business Media LLC}, author={Stroup, T. Scott and Morrissey, Joseph P. and Ellis, Alan R. and Blank, Michael}, year={2001}, pages={117–128} }
 @article{fried_topping_morrissey_ellis_stroup_blank_2000, title={Comparing provider perceptions of access and utilization management in full-risk and no-risk medicaid programs for adults with serious mental illness}, volume={27}, ISSN={1094-3412 1556-3308}, url={http://dx.doi.org/10.1007/BF02287802}, DOI={10.1007/BF02287802}, abstractNote={This article compares provider perceptions of access to services and utilization management (UM) procedures in two Medicaid programs in the same state: a full-risk capitated managed care (MC) program and a no-risk, fee-for-service (FFS) program. Survey data were obtained from 198 mental health clinicians and administrators. The only difference found between respondents in the FFS and MC sites was that outpatient providers in the MC site reported significantly lower levels of access to high-intensity services than did providers in the FFS site (p < .001). Respondents in the two sites reported similar attitudes toward UM procedures, including a strong preference for internal over external UM procedures. These findings support the conclusion that through diffusion of UM procedures, all care in the Medicaid program for persons with a serious mental illness is managed, regardless of risk arrangement. Implications for mental health services and further research are discussed.}, number={1}, journal={The Journal of Behavioral Health Services & Research}, publisher={Springer Science and Business Media LLC}, author={Fried, Bruce J. and Topping, Sharon and Morrissey, Joseph P. and Ellis, Alan R. and Stroup, Scott and Blank, Michael}, year={2000}, month={Feb}, pages={29–46} }