Bruce Kinosian, Susan Schmitt, Matthew Augustine, Scotte Hartronft, Rajesh Makineni, Kimberly Judon, Gregory Krautner, Cheryl Schmitz, Mary K. Goldstein, Ciaran S. Phibbs, Orna Intrator
Objective
To identify risk of long-term institutionalization (LTI) among Veterans receiving care in the Veterans Health Administration (VA).
Study Setting and Design
We developed the “Predicted Long-term Institutionalization” (PLI) risk model for Veterans alive in the community at the end of fiscal-year (FY) 2017 followed for LTI in nursing home (cumulative NH days allowing any acute care and up to 7 days in community > 90 days) during FY2018-FY2019.
Data Sources and Analytic Sample
PLI used demographics, diagnoses, prior hospital and nursing home (NH) use, and risk indices for death and frailty from VA and Medicare claims and Minimum Data Set data. Development of PLI used multiple iterations to maximize sensitivity, constrained by achieving a number needed to screen (≤ 8), including age normalization to minimize algorithmic bias. We combined the elevated risk (ER) and common risk (CR) strata-specific predictions from the logistic regression models to identify three tiers of PLI: low risk, moderate risk, and high risk. We describe Veterans' outcomes in FY2018/2019 (LTI, death, hospitalization and VA cost) across the three PLI tiers.
Principal Findings
For identifying Veterans in LTI, compared to a baseline model that used only VA data as predictors (sensitivity 23%, specificity 98%), calibrating separate ER and CR strata increased sensitivity to 30%, the addition of Medicare data increased sensitivity to 33%, and age-normalization with differential risk strata thresholds increased sensitivity to 41% (specificity 96.6%). The final PLI model (c-statistic = 0.87) identified 3.5% of Veterans in PLI-high risk (13% LTI rate), who accounted for 41% of new LTI, 22% of decedents, 19% of VA cost, and 11% of hospitalizations in FY2018–2019.
Conclusions
The PLI score identifies Veterans at high risk of LTI for further assessment and targeting of resources to support continued community residence.
{"title":"Predicting Risk of Long-Term Institutionalization Among Community Dwelling Veterans Before the COVID-19 Pandemic","authors":"Bruce Kinosian, Susan Schmitt, Matthew Augustine, Scotte Hartronft, Rajesh Makineni, Kimberly Judon, Gregory Krautner, Cheryl Schmitz, Mary K. Goldstein, Ciaran S. Phibbs, Orna Intrator","doi":"10.1111/1475-6773.70016","DOIUrl":"10.1111/1475-6773.70016","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Objective</h3>\u0000 \u0000 <p>To identify risk of long-term institutionalization (LTI) among Veterans receiving care in the Veterans Health Administration (VA).</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Study Setting and Design</h3>\u0000 \u0000 <p>We developed the “Predicted Long-term Institutionalization” (PLI) risk model for Veterans alive in the community at the end of fiscal-year (FY) 2017 followed for LTI in nursing home (cumulative NH days allowing any acute care and up to 7 days in community > 90 days) during FY2018-FY2019.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Data Sources and Analytic Sample</h3>\u0000 \u0000 <p>PLI used demographics, diagnoses, prior hospital and nursing home (NH) use, and risk indices for death and frailty from VA and Medicare claims and Minimum Data Set data. Development of PLI used multiple iterations to maximize sensitivity, constrained by achieving a number needed to screen (≤ 8), including age normalization to minimize algorithmic bias. We combined the elevated risk (ER) and common risk (CR) strata-specific predictions from the logistic regression models to identify three tiers of PLI: low risk, moderate risk, and high risk. We describe Veterans' outcomes in FY2018/2019 (LTI, death, hospitalization and VA cost) across the three PLI tiers.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Principal Findings</h3>\u0000 \u0000 <p>For identifying Veterans in LTI, compared to a baseline model that used only VA data as predictors (sensitivity 23%, specificity 98%), calibrating separate ER and CR strata increased sensitivity to 30%, the addition of Medicare data increased sensitivity to 33%, and age-normalization with differential risk strata thresholds increased sensitivity to 41% (specificity 96.6%). The final PLI model (c-statistic = 0.87) identified 3.5% of Veterans in PLI-high risk (13% LTI rate), who accounted for 41% of new LTI, 22% of decedents, 19% of VA cost, and 11% of hospitalizations in FY2018–2019.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Conclusions</h3>\u0000 \u0000 <p>The PLI score identifies Veterans at high risk of LTI for further assessment and targeting of resources to support continued community residence.</p>\u0000 </section>\u0000 </div>","PeriodicalId":55065,"journal":{"name":"Health Services Research","volume":"61 1","pages":""},"PeriodicalIF":3.2,"publicationDate":"2025-07-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12857474/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144719181","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Sungchul Park, Joseph L. Dieleman, Rockli Kim, S. V. Subramanian
Objectives
To examine the associations of health and social spending with health outcomes, including Disability-Adjusted Life Years (DALY), Years of Life Lost (YLL), Years Lived with Disability (YLD), death, and life expectancy at birth among Organization for Economic Cooperation and Development (OECD) member countries from 2000 to 2019.
Study Setting and Design
We conducted a retrospective longitudinal study.
Data Sources and Analytical Sample
Our sample included 36 OECD member countries as of 2019 using data from the Global Burden of Disease Study 2021, the OECD, and the World Bank.
Principal Findings
Fixed-effect analysis revealed significant associations of health and social spending with health outcomes, but the patterns varied. Specifically, a one-percentage-point increase in health spending was associated with a 1.43% (95% CI: −1.86, −1.01) decrease in the death rate per 100,000 population and a 0.68% (0.56, 0.79) increase in YLD per 100,000 population. In contrast, a one-percentage-point increase in social spending was associated with a 0.29% (−0.45, −0.12) reduction in DALYs, primarily driven by a 0.30% (−0.37, −0.23) decrease in YLDs and a 0.07% (0.03, 0.12) increase in life expectancy. No significant associations were found for the remaining outcomes. These associations remained robust when incorporating one- and two-year lagged effects.
Conclusions
These findings highlight the distinct mechanisms through which health and social spending impact health outcomes. Health spending predominantly influenced mortality, while social spending was more closely associated with improvements in quality-of-life measures. Policymakers should consider these complementary effects when designing interventions to optimize health outcomes.
{"title":"Association of Health and Social Spending With Health Outcomes in OECD Countries","authors":"Sungchul Park, Joseph L. Dieleman, Rockli Kim, S. V. Subramanian","doi":"10.1111/1475-6773.14660","DOIUrl":"10.1111/1475-6773.14660","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Objectives</h3>\u0000 \u0000 <p>To examine the associations of health and social spending with health outcomes, including Disability-Adjusted Life Years (DALY), Years of Life Lost (YLL), Years Lived with Disability (YLD), death, and life expectancy at birth among Organization for Economic Cooperation and Development (OECD) member countries from 2000 to 2019.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Study Setting and Design</h3>\u0000 \u0000 <p>We conducted a retrospective longitudinal study.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Data Sources and Analytical Sample</h3>\u0000 \u0000 <p>Our sample included 36 OECD member countries as of 2019 using data from the Global Burden of Disease Study 2021, the OECD, and the World Bank.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Principal Findings</h3>\u0000 \u0000 <p>Fixed-effect analysis revealed significant associations of health and social spending with health outcomes, but the patterns varied. Specifically, a one-percentage-point increase in health spending was associated with a 1.43% (95% CI: −1.86, −1.01) decrease in the death rate per 100,000 population and a 0.68% (0.56, 0.79) increase in YLD per 100,000 population. In contrast, a one-percentage-point increase in social spending was associated with a 0.29% (−0.45, −0.12) reduction in DALYs, primarily driven by a 0.30% (−0.37, −0.23) decrease in YLDs and a 0.07% (0.03, 0.12) increase in life expectancy. No significant associations were found for the remaining outcomes. These associations remained robust when incorporating one- and two-year lagged effects.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Conclusions</h3>\u0000 \u0000 <p>These findings highlight the distinct mechanisms through which health and social spending impact health outcomes. Health spending predominantly influenced mortality, while social spending was more closely associated with improvements in quality-of-life measures. Policymakers should consider these complementary effects when designing interventions to optimize health outcomes.</p>\u0000 </section>\u0000 </div>","PeriodicalId":55065,"journal":{"name":"Health Services Research","volume":"60 6","pages":""},"PeriodicalIF":3.2,"publicationDate":"2025-07-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1111/1475-6773.14660","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144709972","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
To compare outcomes between patients admitted to different clinicians within skilled nursing facilities for post-acute care, leveraging the plausibly random distribution of patients to admitting clinicians in the case of clinicians who specialize in nursing facility care (SNFists). We also compare patient outcomes between SNFists who are physicians versus advanced practice providers (APPs).
Study Setting and Design
We used multi-level modeling to estimate within-SNF variation in the characteristics and outcomes of patients admitted to different SNFists and linear regression to compare patient characteristics and outcomes between physician and APP SNFists. Our main outcomes were 30-day hospitalizations, 30-day mortality, and antipsychotic use.
Data Sources and Analytic Sample
We analyzed claims data for a 20% sample of traditional Medicare beneficiaries admitted to a SNF for post-acute care from 2016 to 2019.
Principal Findings
The sample included 81,789 post-acute patients seen by 6273 SNFists at 1479 facilities between 2016 and 2019. Within-facility variation in patient characteristics across admitting SNFists was modest and substantially greater across admitting clinicians who were not SNFists, consistent with our key assumption that patients are distributed in a more balanced fashion across admitting clinicians who are SNFists. With patient-level confounding limited by this focus on SNFists, there was minimal to modest variation in the rates of mortality (adjusted standard deviation: −0.14), hospitalization (0.40), and antipsychotic use (1.10) across admitting clinicians. Outcomes also did not differ between APP and physician admitting SNFists (mortality: 0.001 [95% CI: −0.001, 0.003]; hospitalization: 0.004 [95% CI: −0.001, 0.010], antipsychotic use: −0.001 [95% CI: −0.006, 0.003]). In contrast, outcomes varied substantially across admitting clinicians who were not SNFists.
Conclusions
Quasi-experimental assignment of patients to clinicians in SNFs reveals that the admitting clinician appears to have little influence on key outcomes in the post-acute setting, in contrast with similar research conducted in other care settings. An analysis of non-SNFists might falsely conclude that the impact of clinician factors is large because of evident non-random sorting of patients to non-SNFist clinicians in SNFs.
{"title":"Influence of Admitting Clinician on Outcomes in Post-Acute Facilities","authors":"Amanda C. Chen, J. Michael McWilliams","doi":"10.1111/1475-6773.70017","DOIUrl":"10.1111/1475-6773.70017","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Objective</h3>\u0000 \u0000 <p>To compare outcomes between patients admitted to different clinicians within skilled nursing facilities for post-acute care, leveraging the plausibly random distribution of patients to admitting clinicians in the case of clinicians who specialize in nursing facility care (SNFists). We also compare patient outcomes between SNFists who are physicians versus advanced practice providers (APPs).</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Study Setting and Design</h3>\u0000 \u0000 <p>We used multi-level modeling to estimate within-SNF variation in the characteristics and outcomes of patients admitted to different SNFists and linear regression to compare patient characteristics and outcomes between physician and APP SNFists. Our main outcomes were 30-day hospitalizations, 30-day mortality, and antipsychotic use.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Data Sources and Analytic Sample</h3>\u0000 \u0000 <p>We analyzed claims data for a 20% sample of traditional Medicare beneficiaries admitted to a SNF for post-acute care from 2016 to 2019.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Principal Findings</h3>\u0000 \u0000 <p>The sample included 81,789 post-acute patients seen by 6273 SNFists at 1479 facilities between 2016 and 2019. Within-facility variation in patient characteristics across admitting SNFists was modest and substantially greater across admitting clinicians who were not SNFists, consistent with our key assumption that patients are distributed in a more balanced fashion across admitting clinicians who are SNFists. With patient-level confounding limited by this focus on SNFists, there was minimal to modest variation in the rates of mortality (adjusted standard deviation: −0.14), hospitalization (0.40), and antipsychotic use (1.10) across admitting clinicians. Outcomes also did not differ between APP and physician admitting SNFists (mortality: 0.001 [95% CI: −0.001, 0.003]; hospitalization: 0.004 [95% CI: −0.001, 0.010], antipsychotic use: −0.001 [95% CI: −0.006, 0.003]). In contrast, outcomes varied substantially across admitting clinicians who were not SNFists.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Conclusions</h3>\u0000 \u0000 <p>Quasi-experimental assignment of patients to clinicians in SNFs reveals that the admitting clinician appears to have little influence on key outcomes in the post-acute setting, in contrast with similar research conducted in other care settings. An analysis of non-SNFists might falsely conclude that the impact of clinician factors is large because of evident non-random sorting of patients to non-SNFist clinicians in SNFs.</p>\u0000 </section>\u0000 </div>","PeriodicalId":55065,"journal":{"name":"Health Services Research","volume":"61 1","pages":""},"PeriodicalIF":3.2,"publicationDate":"2025-07-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144700420","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Christopher S. Brunt, John R. Bowblis, Robert Applebaum
Objective
To assess cost-effective strategies to reduce nurse aide turnover, this study examines the relationship between turnover and compensation, including wage rates, spending on fringe benefits, and specific fringe benefit offerings.
Study Setting and Design
The study uses national data from 2022 and 2023, a period following major COVID-19 labor market disruptions. The analysis uses regression models to assess the impact of wages and fringe benefits on turnover, with additional subgroup analyses by ownership type (for-profit, not-for-profit, and government).
Data Sources and Analytic Sample
Data were sourced from Medicare Cost Reports, the Payroll-Based Journal Public Use Employee Detail File, and Care Compare archives. After excluding nursing homes with missing observations and applying exclusions for outliers, the final analytic sample included 19,238 nursing home-year observations from 12,116 unique nursing homes.
Principal Findings
The results indicate that higher wages and fringe benefit spending are both associated with slightly lower nurse aide turnover. A 10% increase in wages was linked to a 0.28 (95% CI: 0.04, 0.53) to 0.39 (95% CI: 0.09, 0.70) percentage point reduction in turnover, an effect primarily driven by for-profit nursing homes. Fringe benefit spending was significantly associated with lower turnover among for-profits and not-for-profits, with a 1-percentage-point increase in fringe rates reducing turnover by 0.08 (95% CI: 0.01, 0.15) to 0.28 (95% CI: 0.23, 0.34) percentage points. Specific fringe benefits, such as daycare assistance and accident/disability insurance, were associated with lower turnover. A simulation analysis suggests that investments in fringe benefits are more effective at reducing turnover than equivalent investments in wages.
Conclusions
Nursing homes seeking to reduce nurse aide turnover should consider enhancing fringe benefits in addition to increasing wages. Given the higher cost-effectiveness of fringe benefits in reducing turnover, policymakers and nursing home administrators should refine these strategies to improve workforce stability and care quality.
{"title":"An Assessment of the Association Between Wages and Fringe Benefits on Nurse Aide Turnover in Nursing Homes","authors":"Christopher S. Brunt, John R. Bowblis, Robert Applebaum","doi":"10.1111/1475-6773.70019","DOIUrl":"10.1111/1475-6773.70019","url":null,"abstract":"<div>\u0000 <section>\u0000 <h3> Objective</h3>\u0000 <p>To assess cost-effective strategies to reduce nurse aide turnover, this study examines the relationship between turnover and compensation, including wage rates, spending on fringe benefits, and specific fringe benefit offerings.</p>\u0000 </section>\u0000 <section>\u0000 <h3> Study Setting and Design</h3>\u0000 <p>The study uses national data from 2022 and 2023, a period following major COVID-19 labor market disruptions. The analysis uses regression models to assess the impact of wages and fringe benefits on turnover, with additional subgroup analyses by ownership type (for-profit, not-for-profit, and government).</p>\u0000 </section>\u0000 <section>\u0000 <h3> Data Sources and Analytic Sample</h3>\u0000 <p>Data were sourced from Medicare Cost Reports, the Payroll-Based Journal Public Use Employee Detail File, and Care Compare archives. After excluding nursing homes with missing observations and applying exclusions for outliers, the final analytic sample included 19,238 nursing home-year observations from 12,116 unique nursing homes.</p>\u0000 </section>\u0000 <section>\u0000 <h3> Principal Findings</h3>\u0000 <p>The results indicate that higher wages and fringe benefit spending are both associated with slightly lower nurse aide turnover. A 10% increase in wages was linked to a 0.28 (95% CI: 0.04, 0.53) to 0.39 (95% CI: 0.09, 0.70) percentage point reduction in turnover, an effect primarily driven by for-profit nursing homes. Fringe benefit spending was significantly associated with lower turnover among for-profits and not-for-profits, with a 1-percentage-point increase in fringe rates reducing turnover by 0.08 (95% CI: 0.01, 0.15) to 0.28 (95% CI: 0.23, 0.34) percentage points. Specific fringe benefits, such as daycare assistance and accident/disability insurance, were associated with lower turnover. A simulation analysis suggests that investments in fringe benefits are more effective at reducing turnover than equivalent investments in wages.</p>\u0000 </section>\u0000 <section>\u0000 <h3> Conclusions</h3>\u0000 <p>Nursing homes seeking to reduce nurse aide turnover should consider enhancing fringe benefits in addition to increasing wages. Given the higher cost-effectiveness of fringe benefits in reducing turnover, policymakers and nursing home administrators should refine these strategies to improve workforce stability and care quality.</p>\u0000 </section>\u0000 </div>","PeriodicalId":55065,"journal":{"name":"Health Services Research","volume":"61 1","pages":"1-9"},"PeriodicalIF":3.2,"publicationDate":"2025-07-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144700419","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Megan E. Vanneman, Ciaran S. Phibbs, Michael K. Ong, Yue Zhang, Adam Chow, Jean Yoon
Objective
To compare outcomes for Department of Veterans Affairs (VA) enrollees' behavioral health (BH) hospitalizations by source (VA-direct, VA-purchased community care (CC), Medicaid, Medicare, private insurance, and other payers).
Study Setting and Design
We conducted a retrospective, longitudinal study with VA enrollees from 2015 to 2017 to examine differences in BH hospitalization outcomes by source. We used generalized linear models with clustered standard errors to predict length of stay (LOS), cost, and 30-day readmission.
Data Sources and Analytic Sample
We studied 124,609 BH hospitalizations of 77,299 VA enrollees in 11 geographically diverse states.
Principal Findings
Predicted mean LOS (9.03 days, 95% CI 8.92–9.14 days; p < 0.001) and cost ($17,608, 95% CI $17,347–$17,870; p < 0.001) were highest for VA-direct hospitalizations, while the mean readmission rate was lowest for VA-direct hospitalizations (17.36%, 95% CI 17.03%–17.69%; p < 0.001). Average marginal effects for each non-VA hospitalization source were statistically significantly different from VA-direct hospitalizations (p < 0.001): between 2.13 and 2.90 days less for LOS, $11,141 to $12,144 less for cost, and 2.71% to 5.18% higher for readmission rate.
Conclusions
The majority of BH hospitalizations were in VA-direct care (56%), with 44% provided in locations outside VA hospitals: Medicare (19%), CC (7%), private insurance (7%), other payers (6%), and Medicaid (5%). There are trade-offs between BH hospitalizations provided in VA-direct care (lowest readmission rate, highest LOS and costs) and other sources.
目的：比较退伍军人事务部（VA）入选者按来源（VA直接、VA购买的社区护理（CC）、Medicaid、Medicare、私人保险和其他支付者）的行为健康（BH）住院治疗的结果。研究设置和设计：我们对2015年至2017年VA入组者进行了一项回顾性纵向研究，以检查不同来源的BH住院结果的差异。我们使用具有聚类标准误差的广义线性模型来预测住院时间（LOS）、费用和30天再入院。数据来源和分析样本：我们研究了11个地理位置不同的州77,299名VA注册者的124,609例BH住院。主要发现：预测平均LOS(9.03天，95% CI 8.92-9.14天；p结论：大多数BH住院是VA直接护理（56%），44%在VA医院以外的地方提供：医疗保险（19%），CC(7%)，私人保险（7%），其他付款人（6%）和医疗补助（5%）。在va直接护理中提供的BH住院治疗（再入院率最低，LOS和费用最高）与其他来源之间存在权衡。
{"title":"Veterans' Behavioral Health Hospitalizations and Outcomes in VA Versus Non-VA Hospitals","authors":"Megan E. Vanneman, Ciaran S. Phibbs, Michael K. Ong, Yue Zhang, Adam Chow, Jean Yoon","doi":"10.1111/1475-6773.70013","DOIUrl":"10.1111/1475-6773.70013","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Objective</h3>\u0000 \u0000 <p>To compare outcomes for Department of Veterans Affairs (VA) enrollees' behavioral health (BH) hospitalizations by source (VA-direct, VA-purchased community care (CC), Medicaid, Medicare, private insurance, and other payers).</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Study Setting and Design</h3>\u0000 \u0000 <p>We conducted a retrospective, longitudinal study with VA enrollees from 2015 to 2017 to examine differences in BH hospitalization outcomes by source. We used generalized linear models with clustered standard errors to predict length of stay (LOS), cost, and 30-day readmission.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Data Sources and Analytic Sample</h3>\u0000 \u0000 <p>We studied 124,609 BH hospitalizations of 77,299 VA enrollees in 11 geographically diverse states.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Principal Findings</h3>\u0000 \u0000 <p>Predicted mean LOS (9.03 days, 95% CI 8.92–9.14 days; <i>p</i> < 0.001) and cost ($17,608, 95% CI $17,347–$17,870; <i>p</i> < 0.001) were highest for VA-direct hospitalizations, while the mean readmission rate was lowest for VA-direct hospitalizations (17.36%, 95% CI 17.03%–17.69%; <i>p</i> < 0.001). Average marginal effects for each non-VA hospitalization source were statistically significantly different from VA-direct hospitalizations (<i>p</i> < 0.001): between 2.13 and 2.90 days less for LOS, $11,141 to $12,144 less for cost, and 2.71% to 5.18% higher for readmission rate.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Conclusions</h3>\u0000 \u0000 <p>The majority of BH hospitalizations were in VA-direct care (56%), with 44% provided in locations outside VA hospitals: Medicare (19%), CC (7%), private insurance (7%), other payers (6%), and Medicaid (5%). There are trade-offs between BH hospitalizations provided in VA-direct care (lowest readmission rate, highest LOS and costs) and other sources.</p>\u0000 </section>\u0000 </div>","PeriodicalId":55065,"journal":{"name":"Health Services Research","volume":"61 1","pages":""},"PeriodicalIF":3.2,"publicationDate":"2025-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144692505","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Kameswari Potharaju, Laura M. Gottlieb, Holly E. Wing, Alejandra Gonzalez-Rocha, Amanda L. Brewster, Danielle Hessler Jones, Andrea Quiñones-Rivera
Objective
To identify key factors that define patient experiences of social care in healthcare settings.
Study Setting and Design
This is a qualitative study using interviews from participants recruited by collaborators of a social care research group from across the United States.
Data Sources and Analytic Sample
We conducted 30 semi-structured interviews between September 2023 and February 2024. Participants were 18 or older, English- or Spanish-speaking, and had received social care in a healthcare setting within the last 12 months. Interview transcripts were dually coded and analyzed using a mixed inductive-deductive approach.
Principal Findings
Patient experience was defined by elements of social care delivery that fell into two categories: the functional and relational domains of social care. Participants reported that operational or “functional” aspects of social care, including screening, resource connections, and other forms of follow-up, represented an important part of their experiences of social care. Experiences of social care were also defined by relational factors, for example, demonstrations of empathy, positive perceptions of screening intentions, linguistic concordance, and longitudinal relationships with the care team. Many participants felt that these functional and relational factors were inextricably linked.
Conclusions
The impressive role that relational factors—that is, interactions and relationships with social care providers—play in defining patient experiences highlights the need to include these factors in efforts to evaluate social care interventions. Discussions about social needs may retain value even in the absence of available resources if healthcare teams attend to the relational factors that drive patients' social care experiences. In the future, measures of social care quality should account for both the functional and relational dimensions of social care.
{"title":"Drivers of Patient Experiences With Healthcare-Based Social Care","authors":"Kameswari Potharaju, Laura M. Gottlieb, Holly E. Wing, Alejandra Gonzalez-Rocha, Amanda L. Brewster, Danielle Hessler Jones, Andrea Quiñones-Rivera","doi":"10.1111/1475-6773.70020","DOIUrl":"10.1111/1475-6773.70020","url":null,"abstract":"<div>\u0000 <section>\u0000 <h3> Objective</h3>\u0000 <p>To identify key factors that define patient experiences of social care in healthcare settings.</p>\u0000 </section>\u0000 <section>\u0000 <h3> Study Setting and Design</h3>\u0000 <p>This is a qualitative study using interviews from participants recruited by collaborators of a social care research group from across the United States.</p>\u0000 </section>\u0000 <section>\u0000 <h3> Data Sources and Analytic Sample</h3>\u0000 <p>We conducted 30 semi-structured interviews between September 2023 and February 2024. Participants were 18 or older, English- or Spanish-speaking, and had received social care in a healthcare setting within the last 12 months. Interview transcripts were dually coded and analyzed using a mixed inductive-deductive approach.</p>\u0000 </section>\u0000 <section>\u0000 <h3> Principal Findings</h3>\u0000 <p>Patient experience was defined by elements of social care delivery that fell into two categories: the functional and relational domains of social care. Participants reported that operational or “functional” aspects of social care, including screening, resource connections, and other forms of follow-up, represented an important part of their experiences of social care. Experiences of social care were also defined by relational factors, for example, demonstrations of empathy, positive perceptions of screening intentions, linguistic concordance, and longitudinal relationships with the care team. Many participants felt that these functional and relational factors were inextricably linked.</p>\u0000 </section>\u0000 <section>\u0000 <h3> Conclusions</h3>\u0000 <p>The impressive role that relational factors—that is, interactions and relationships with social care providers—play in defining patient experiences highlights the need to include these factors in efforts to evaluate social care interventions. Discussions about social needs may retain value even in the absence of available resources if healthcare teams attend to the relational factors that drive patients' social care experiences. In the future, measures of social care quality should account for both the functional and relational dimensions of social care.</p>\u0000 </section>\u0000 </div>","PeriodicalId":55065,"journal":{"name":"Health Services Research","volume":"61 1","pages":"1-9"},"PeriodicalIF":3.2,"publicationDate":"2025-07-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144692504","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Arthur S Hong, Lesi He, Pranathi Pilla, Joshua M Liao, D Mark Courtney, Navid Sadeghi, Ethan A Halm
Objective: To assess changes in publicly reported, potentially avoidable hospital visits after chemotherapy since the introduction of a Medicare quality measure.
Study setting and design: Retrospective analysis of avoidable emergency department (ED) and inpatient admission (ADM) rates after chemotherapy between 2018 and 2022, across absolute visit rates and relative hospital performance ("better than", "no different than", "worse than" the national rate). We stratified hospitals into quartiles of visit rates in 2018 and used this to model the change in visit rates from 2018 to 2022 with generalized linear regression.
Data sources and analytic sample: A longitudinal cohort of hospitals from the Medicare Outpatient Quality Reporting Program.
Principal findings: We analyzed 1179 hospitals (94.3% non-profit, 22.9% teaching). National avoidable ED visit rates were 6.0% in 2018, 5.4% in 2022; ADM rates were 12.5% in 2018, 10.3% in 2022. Nearly all hospitals were deemed to have performed "no different" than the national rate each year in ED (≥ 95.3%) and ADM (≥ 91.1%). In adjusted analyses, visit rates for hospitals in the lowest 2018 visit rate quartiles declined the least by 2022 (ED: -0.44% 95% CI: -0.58 to -2.94; ADM: -0.91%, 95% CI: -1.14 to -0.69), and declined the most for hospitals in the highest 2018 quartiles (ED: -1.72%, 95% CI: -1.85 to -7.73; ADM: -3.03%, 95% CI: -3.27 to -2.81). We estimated that the tendency for extreme baseline values to approach the average over time accounted for up to one-tenth of the decline among the worst-performing 2018 quartiles (ED: 10.6% of rate change, 95% CI: 9.8 to 11.5; ADM: 9.0%, 95% CI: 8.2 to 9.8).
Conclusion: Hospitals reduced their potentially avoidable hospital visit rates, though Medicare deemed that nearly all hospitals performed "no different" than the national average each year. It remains unclear if the reductions were driven by this quality measure.
{"title":"Early Examination of Hospital-Level Performance on Unplanned, Potentially Avoidable Hospital Visits After Chemotherapy, 2018-2022.","authors":"Arthur S Hong, Lesi He, Pranathi Pilla, Joshua M Liao, D Mark Courtney, Navid Sadeghi, Ethan A Halm","doi":"10.1111/1475-6773.70014","DOIUrl":"https://doi.org/10.1111/1475-6773.70014","url":null,"abstract":"<p><strong>Objective: </strong>To assess changes in publicly reported, potentially avoidable hospital visits after chemotherapy since the introduction of a Medicare quality measure.</p><p><strong>Study setting and design: </strong>Retrospective analysis of avoidable emergency department (ED) and inpatient admission (ADM) rates after chemotherapy between 2018 and 2022, across absolute visit rates and relative hospital performance (\"better than\", \"no different than\", \"worse than\" the national rate). We stratified hospitals into quartiles of visit rates in 2018 and used this to model the change in visit rates from 2018 to 2022 with generalized linear regression.</p><p><strong>Data sources and analytic sample: </strong>A longitudinal cohort of hospitals from the Medicare Outpatient Quality Reporting Program.</p><p><strong>Principal findings: </strong>We analyzed 1179 hospitals (94.3% non-profit, 22.9% teaching). National avoidable ED visit rates were 6.0% in 2018, 5.4% in 2022; ADM rates were 12.5% in 2018, 10.3% in 2022. Nearly all hospitals were deemed to have performed \"no different\" than the national rate each year in ED (≥ 95.3%) and ADM (≥ 91.1%). In adjusted analyses, visit rates for hospitals in the lowest 2018 visit rate quartiles declined the least by 2022 (ED: -0.44% 95% CI: -0.58 to -2.94; ADM: -0.91%, 95% CI: -1.14 to -0.69), and declined the most for hospitals in the highest 2018 quartiles (ED: -1.72%, 95% CI: -1.85 to -7.73; ADM: -3.03%, 95% CI: -3.27 to -2.81). We estimated that the tendency for extreme baseline values to approach the average over time accounted for up to one-tenth of the decline among the worst-performing 2018 quartiles (ED: 10.6% of rate change, 95% CI: 9.8 to 11.5; ADM: 9.0%, 95% CI: 8.2 to 9.8).</p><p><strong>Conclusion: </strong>Hospitals reduced their potentially avoidable hospital visit rates, though Medicare deemed that nearly all hospitals performed \"no different\" than the national average each year. It remains unclear if the reductions were driven by this quality measure.</p>","PeriodicalId":55065,"journal":{"name":"Health Services Research","volume":" ","pages":"e70014"},"PeriodicalIF":3.1,"publicationDate":"2025-07-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144683622","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Maricela Cruz, Susan M. Shortreed, Gregory E. Simon, Yates Coley
Objective
To compare alternative Difference-in-Differences (DID) methods for evaluating the effect of risk-stratified interventions, or interventions targeting at-risk groups, on binary outcomes.
Study Setting and Design
In simulations, we compared operating characteristics of recycled prediction estimators for common average treatment effect on the treated (ATT) estimands across three DID models: the traditional two groups and two periods model, a risk score adjusted model, and a model adjusting for risk score and its interactions with risk group and period. We compared DID ATT estimates to randomized evaluation estimates of a risk-stratified intervention implemented at Kaiser Permanente Washington (KPWA), delivering additional text-message reminders to reduce missed clinic visits.
Data Sources and Analytic Sample
Our study included 588,503 KPWA visits, with 285,814 (49%) visits pre-evaluation (05/01/2018–10/30/2018) and 302,689 (51%) visits during the evaluation (02/01/2019–09/30/2019). Pre-evaluation, 120,350 visits were classified as high-risk. During the evaluation, 125,076 visits were labeled as high-risk, with 62,557 (50%) randomized to the intervention. We generated data in simulations based on this setting.
Principal Findings
In simulations, the traditional DID and risk score adjusted models had smaller bias and standard errors, and better coverage probabilities. DID estimates closest to randomized evaluation estimates (−0.007, 95% CI [−0.010, −0.004]) were from the traditional DID model assuming the identity link (−0.008, 95% CI [−0.011, −0.005]) or the risk adjusted model with any link (−0.006, 95% CI [−0.008, −0.003] identity; −0.007, 95% CI [−0.011, −0.003] logit; −0.007, 95% CI [−0.012, −0.003] log) for the ATT on the absolute difference scale (usual DID ATT estimand), and the risk score adjusted model with log or logit links for all other estimands.
Conclusions
Compared with randomized evaluation results, the traditional DID model is appropriate for the ATT on the absolute difference scale, while the risk score adjusted model with log or logit links is appropriate for all ATT estimands considered.
目的：比较不同的差分法（DID）来评估风险分层干预或针对高危人群的干预对二元结果的影响。研究设置和设计：在模拟中，我们比较了三种DID模型（传统的两组两期模型、风险评分调整模型和风险评分及其与风险组和时期的相互作用调整模型）中回收预测估计器的共同平均处理效果（ATT）估计的运行特征。我们比较了DID ATT估计值与Kaiser Permanente Washington （KPWA）实施的风险分层干预的随机评估估计值，提供额外的短信提醒以减少错过的诊所就诊。数据来源及分析样本：我们的研究包括588,503次KPWA访问，其中评估前（2018/01/05 - 2018/10/30）访问285,814次（49%），评估期间（2019年1月02 - 2019年9月30日）访问302,689次（51%）。预评估中，120,350次就诊被归为高风险。在评估期间，125,076次就诊被标记为高风险，其中62,557次（50%）随机分配到干预组。我们基于这个设置在模拟中生成数据。主要发现：在模拟中，传统的DID和风险评分调整模型具有较小的偏差和标准误差，并且具有更好的覆盖概率。最接近随机评价估计的DID估计（-0.007,95% CI[-0.010, -0.004]）来自传统的DID模型，假设存在身份关联（-0.008,95% CI[-0.011, -0.005]）或具有任何联系的风险调整模型（-0.006,95% CI [-0.008, -0.003]）；-0.007, 95% CI [-0.011, -0.003] logit；-0.007, 95% CI [-0.012, -0.003] log)用于绝对差量表上的ATT（通常的DID ATT估计），风险评分调整模型与log或logit链接用于所有其他估计。结论：与随机评价结果相比，传统DID模型适用于绝对差量表上的ATT，而带有log或logit链接的风险评分调整模型适用于所有考虑的ATT估计值。
{"title":"Evaluating Clinical Implementation of Risk Prediction Based Interventions Using Difference-In-Differences","authors":"Maricela Cruz, Susan M. Shortreed, Gregory E. Simon, Yates Coley","doi":"10.1111/1475-6773.70015","DOIUrl":"10.1111/1475-6773.70015","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Objective</h3>\u0000 \u0000 <p>To compare alternative Difference-in-Differences (DID) methods for evaluating the effect of risk-stratified interventions, or interventions targeting at-risk groups, on binary outcomes.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Study Setting and Design</h3>\u0000 \u0000 <p>In simulations, we compared operating characteristics of recycled prediction estimators for common average treatment effect on the treated (ATT) estimands across three DID models: the traditional two groups and two periods model, a risk score adjusted model, and a model adjusting for risk score and its interactions with risk group and period. We compared DID ATT estimates to randomized evaluation estimates of a risk-stratified intervention implemented at Kaiser Permanente Washington (KPWA), delivering additional text-message reminders to reduce missed clinic visits.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Data Sources and Analytic Sample</h3>\u0000 \u0000 <p>Our study included 588,503 KPWA visits, with 285,814 (49%) visits pre-evaluation (05/01/2018–10/30/2018) and 302,689 (51%) visits during the evaluation (02/01/2019–09/30/2019). Pre-evaluation, 120,350 visits were classified as high-risk. During the evaluation, 125,076 visits were labeled as high-risk, with 62,557 (50%) randomized to the intervention. We generated data in simulations based on this setting.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Principal Findings</h3>\u0000 \u0000 <p>In simulations, the traditional DID and risk score adjusted models had smaller bias and standard errors, and better coverage probabilities. DID estimates closest to randomized evaluation estimates (−0.007, 95% CI [−0.010, −0.004]) were from the traditional DID model assuming the identity link (−0.008, 95% CI [−0.011, −0.005]) or the risk adjusted model with any link (−0.006, 95% CI [−0.008, −0.003] identity; −0.007, 95% CI [−0.011, −0.003] logit; −0.007, 95% CI [−0.012, −0.003] log) for the ATT on the absolute difference scale (usual DID ATT estimand), and the risk score adjusted model with log or logit links for all other estimands.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Conclusions</h3>\u0000 \u0000 <p>Compared with randomized evaluation results, the traditional DID model is appropriate for the ATT on the absolute difference scale, while the risk score adjusted model with log or logit links is appropriate for all ATT estimands considered.</p>\u0000 </section>\u0000 </div>","PeriodicalId":55065,"journal":{"name":"Health Services Research","volume":"61 1","pages":""},"PeriodicalIF":3.2,"publicationDate":"2025-07-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144676613","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Stephanie B Wheeler, Jason Rotter, Lisa P Spees, Caitlin B Biddell, Justin G Trogdon, Catherine M Alfano, Deborah K Mayer, Michaela A Dinan, Larissa Nekhlyudov, Sarah A Birken
Objective: To develop and validate a clinical risk prediction algorithm to identify breast cancer survivors at high risk for adverse outcomes.
Study setting and design: Our national retrospective analysis used cross-validated random forest machine learning models to separately predict the risk of all-cause death, cancer-specific death, claims-derived risk of recurrence, and other adverse health outcomes within 3 and 5 years following treatment completion.
Data sources and analytic sample: Our study used the Surveillance and Epidemiology End Results (SEER) registry-Consumer Assessment of Healthcare Providers and Systems (CAHPS) survey (SEER-CAHPS) linked data for survivors diagnosed between 2003 and 2011, with follow-up claims data to 2017.
Principal findings: Within the 3-year follow-up period, 372/4516 survivors (mean age 75.1; 81.7% white) in the primary cohort (8.2%) died, 111 from cancer (2.5%), 665 (14.7%) experienced cancer recurrence, and 488 (10.8%) were hospitalized for adverse health outcomes. The algorithm's prediction resulted in 91.9% out-of-sample accuracy (the percent of observations classified correctly) and a 37.6% Cohen's Kappa (i.e., improvement over an uninformed model). Out-of-sample accuracy was 97.5% (44% improvement) for predicting cancer-specific death, 85% (26% improvement) for recurrence, and 89% (28% improvement) for other adverse health outcomes. Important predictors across outcomes included geographic region, age, frailty, comorbidity, time since diagnosis, and out-of-pocket cost responsibility.
Conclusions: Machine learning models accurately predicted relevant adverse survivorship outcomes, driven primarily by non-cancer specific factors. Breast cancer survivors at high risk for adverse outcomes may benefit from more intensive care, whereas those at low risk may be more appropriately managed by primary care.
{"title":"Machine Learning Risk Stratification for Older Breast Cancer Survivors: Clinical Care Implications.","authors":"Stephanie B Wheeler, Jason Rotter, Lisa P Spees, Caitlin B Biddell, Justin G Trogdon, Catherine M Alfano, Deborah K Mayer, Michaela A Dinan, Larissa Nekhlyudov, Sarah A Birken","doi":"10.1111/1475-6773.70005","DOIUrl":"https://doi.org/10.1111/1475-6773.70005","url":null,"abstract":"<p><strong>Objective: </strong>To develop and validate a clinical risk prediction algorithm to identify breast cancer survivors at high risk for adverse outcomes.</p><p><strong>Study setting and design: </strong>Our national retrospective analysis used cross-validated random forest machine learning models to separately predict the risk of all-cause death, cancer-specific death, claims-derived risk of recurrence, and other adverse health outcomes within 3 and 5 years following treatment completion.</p><p><strong>Data sources and analytic sample: </strong>Our study used the Surveillance and Epidemiology End Results (SEER) registry-Consumer Assessment of Healthcare Providers and Systems (CAHPS) survey (SEER-CAHPS) linked data for survivors diagnosed between 2003 and 2011, with follow-up claims data to 2017.</p><p><strong>Principal findings: </strong>Within the 3-year follow-up period, 372/4516 survivors (mean age 75.1; 81.7% white) in the primary cohort (8.2%) died, 111 from cancer (2.5%), 665 (14.7%) experienced cancer recurrence, and 488 (10.8%) were hospitalized for adverse health outcomes. The algorithm's prediction resulted in 91.9% out-of-sample accuracy (the percent of observations classified correctly) and a 37.6% Cohen's Kappa (i.e., improvement over an uninformed model). Out-of-sample accuracy was 97.5% (44% improvement) for predicting cancer-specific death, 85% (26% improvement) for recurrence, and 89% (28% improvement) for other adverse health outcomes. Important predictors across outcomes included geographic region, age, frailty, comorbidity, time since diagnosis, and out-of-pocket cost responsibility.</p><p><strong>Conclusions: </strong>Machine learning models accurately predicted relevant adverse survivorship outcomes, driven primarily by non-cancer specific factors. Breast cancer survivors at high risk for adverse outcomes may benefit from more intensive care, whereas those at low risk may be more appropriately managed by primary care.</p>","PeriodicalId":55065,"journal":{"name":"Health Services Research","volume":" ","pages":"e70005"},"PeriodicalIF":3.1,"publicationDate":"2025-07-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144651272","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The objective of this study is to update estimates of comprehensive service availability among CCBHCs and compare them to other settings serving individuals with severe mental illness, including community mental health centers (CMHCs) and federally qualified health centers (FQHCs).
Study Design and Setting
This study is a cross-sectional secondary data analysis.
Data Sources and Analytic Sample
Using 2022 National Substance Use and Mental Health Services Survey (N-SUMHSS) data, logistic regression examined associations between service setting (CCBHC, CMHC, FQHC) and the availability of psychiatric, health management, and navigation, and social care services.
Principle Findings
Compared to CCBHCs, FQHC designation was associated with a decreased likelihood of offering psychiatric rehabilitation services, including ACT (marginal effect = −0.26, 95% CI: −0.33 to −0.19) and peer coaching (marginal effect = −0.36, 95% CI: −0.43 to −0.29), and psychiatric crisis intervention (marginal effect = −0.14, 95% CI: −0.22 to −0.07). Rates of health management services were comparable to those at CCBHCs. CMHCs were also less likely to offer health management services (marginal effect = −0.26, 95% CI: −0.32 to −0.21) and a range of psychiatric rehabilitation services relative to CCBHCs.
Conclusions
CCBHC certified clinics were more likely to offer psychiatric and social services than FQHC or CMHC clinics serving individuals with severe mental illness.
{"title":"Enhanced Service Capacity for Severe Mental Illness: A Comparative Analysis of Certified Community Behavioral Health Centers, Community Mental Health Centers, and Federally Qualified Health Centers","authors":"Elizabeth B. Matthews, Victoria Stanhope","doi":"10.1111/1475-6773.70010","DOIUrl":"10.1111/1475-6773.70010","url":null,"abstract":"<div>\u0000 \u0000 \u0000 <section>\u0000 \u0000 <h3> Objectives</h3>\u0000 \u0000 <p>The objective of this study is to update estimates of comprehensive service availability among CCBHCs and compare them to other settings serving individuals with severe mental illness, including community mental health centers (CMHCs) and federally qualified health centers (FQHCs).</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Study Design and Setting</h3>\u0000 \u0000 <p>This study is a cross-sectional secondary data analysis.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Data Sources and Analytic Sample</h3>\u0000 \u0000 <p>Using 2022 National Substance Use and Mental Health Services Survey (N-SUMHSS) data, logistic regression examined associations between service setting (CCBHC, CMHC, FQHC) and the availability of psychiatric, health management, and navigation, and social care services.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Principle Findings</h3>\u0000 \u0000 <p>Compared to CCBHCs, FQHC designation was associated with a decreased likelihood of offering psychiatric rehabilitation services, including ACT (marginal effect = −0.26, 95% CI: −0.33 to −0.19) and peer coaching (marginal effect = −0.36, 95% CI: −0.43 to −0.29), and psychiatric crisis intervention (marginal effect = −0.14, 95% CI: −0.22 to −0.07). Rates of health management services were comparable to those at CCBHCs. CMHCs were also less likely to offer health management services (marginal effect = −0.26, 95% CI: −0.32 to −0.21) and a range of psychiatric rehabilitation services relative to CCBHCs.</p>\u0000 </section>\u0000 \u0000 <section>\u0000 \u0000 <h3> Conclusions</h3>\u0000 \u0000 <p>CCBHC certified clinics were more likely to offer psychiatric and social services than FQHC or CMHC clinics serving individuals with severe mental illness.</p>\u0000 </section>\u0000 </div>","PeriodicalId":55065,"journal":{"name":"Health Services Research","volume":"61 1","pages":""},"PeriodicalIF":3.2,"publicationDate":"2025-07-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146088096","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号