Effect of mental health staffing inputs on initiation of care among recently separated Veterans.

OBJECTIVE: To estimate a causal relationship between mental health staffing and time to initiation of mental health care for new patients. DATA SOURCES AND STUDY SETTING: As the largest integrated health care delivery system in the United States, the Veterans Health Administration (VHA) provides a unique setting for isolating the effects of staffing on initiation of mental health care where demand is high and out-of-pocket costs are not a relevant confounder. We use data from the Department of Defense and VHA to obtain patient and facility characteristics and health care use. STUDY DESIGN: To isolate exogenous variation in mental health staffing, we used an instrumental variables approach-two-stage residual inclusion with a discrete time hazard model. Our outcome is time to initiation of mental health care after separation from active duty (first appointment) and our exposure is mental health staffing (standardized clinic time per 1000 VHA enrollees per pay period). DATA COLLECTION/EXTRACTION METHODS: Our cohort consists of all Veterans separating from active duty between July 2014 and September 2017, who were enrolled in the VHA, and had at least one diagnosis of post-traumatic stress disorder, major depressive disorder, and/or substance use disorder in the year prior to separation from active duty (N = 54,209). PRINCIPAL FINDINGS: An increase of 1 standard deviation in mental health staffing results in a higher likelihood of initiating mental health care (adjusted hazard ratio: 3.17, 95% confidence interval: 2.62, 3.84, p < 0.001). Models stratified by tertile of mental health staffing exhibit decreasing returns to scale. CONCLUSIONS: Increases in mental health staffing led to faster initiation of care and are especially beneficial in facilities where staffing is lower, although initiation of care appears capacity-limited everywhere.

Physicians and Specialties in the Veterans Health Administration's Community Care Network.

This cross-sectional study of data from the US Veterans Health Administration examines the availability of services provided through community care networks by specialty and clinical characteristics.

Restrictiveness of Medicare Advantage provider networks across physician specialties.

OBJECTIVE: The objective was to measure specialty provider networks in Medicare Advantage (MA) and examine associations with market factors. DATA SOURCES AND STUDY SETTING: We relied on traditional Medicare (TM) and MA prescription drug event data from 2011 to 2017 for all Medicare beneficiaries in the United States as well as data from the Area Health Resources File. STUDY DESIGN: Relying on a recently developed and validated prediction model, we calculated the provider network restrictiveness of MA contracts for nine high-prescribing specialties. We characterized network restrictiveness through an observed-to-expected ratio, calculated as the number of unique providers seen by MA beneficiaries divided by the number expected based on the prediction model. We assessed the relationship between network restrictiveness and market factors across specialties with multivariable linear regression. DATA COLLECTION/EXTRACTION METHODS: Prescription drug event data for a 20% random sample of beneficiaries enrolled in prescription drug coverage from 2011 to 2017. PRINCIPAL FINDINGS: Provider networks in MA varied in restrictiveness. OB-Gynecology was the most restrictive with enrollees seeing 34.5% (95% CI: 34.3%-34.7%) as many providers as they would absent network restrictions; cardiology was the least restrictive with enrollees seeing 58.6% (95% CI: 58.4%-58.8%) as many providers as they otherwise would. Factors associated with less restrictive networks included the county-level TM average hierarchical condition category score (0.06; 95% CI: 0.04-0.07), the county-level number of doctors per 1000 population (0.04; 95% CI: 0.02-0.05), the natural log of local median household income (0.03; 95% CI: 0.007-0.05), and the parent company's market share in the county (0.16; 95% CI: 0.13-0.18). Rurality was a major predictor of more restrictive networks (-0.28; 95% CI: -0.32 to -0.24). CONCLUSIONS: Our findings suggest that rural beneficiaries may face disproportionately reduced access in these networks and that efforts to improve access should vary by specialty.

Measuring restrictiveness of Medicare Advantage networks: A claims-based approach.

OBJECTIVE: To develop and validate a measure of provider network restrictiveness in the Medicare Advantage (MA) population. DATA SOURCES: Prescription drug event data and beneficiary information for Part D enrollees from the Center for Medicare and Medicaid Services, along with prescriber identifiers; geographic variables from the Area Health Resources Files. STUDY DESIGN: A prediction model was used to predict the unique number of primary care providers that would have been seen by MA beneficiaries absent network restrictions. The model was trained and validated on Traditional Medicare (TM) beneficiaries. A pseudo-Poisson and a random forest model were evaluated. An observed-to-expected (O/E) ratio was calculated as the number of unique providers seen by MA beneficiaries divided by the number expected based the TM prediction model. Multivariable linear models were used to assess the relationship between network restrictiveness and plan and market factors. DATA COLLECTION/EXTRACTION METHODS: Prescription drug event data were obtained for a 20% random sample of beneficiaries enrolled in prescription drug coverage from 2011 to 2017. PRINCIPAL FINDINGS: Health Maintenance Organization plans were more restrictive (O/E = 55.5%; 95% CI 55.3%-55.7%) than Health Maintenance Organization-Point of Service plans (67.2%; 95% CI 66.7%-67.8%) or Preferred Provider Organization plans (74.7%; 95% CI 74.3%-75.1%), and rural areas had more restrictive networks (31.6%; 95% CI 29.0%-34.2%) than metropolitan areas (61.5%; 95% CI 61.3%-61.7%). Multivariable results confirmed these findings, and also indicated that increased provider supply was associated with less restrictive networks. CONCLUSIONS: We developed a means of estimating provider network restrictiveness in MA from claims data. Our results validate the approach, providing confidence for wider application (e.g., for other markets and specialties) and use for regulation.

Use of High-Risk Medications Among Older Adults Enrolled in Medicare Advantage Plans vs Traditional Medicare.

Importance: Limiting the use of high-risk medications (HRMs) among older adults is a national priority to provide a high quality of care for older beneficiaries of both Medicare Advantage and traditional fee-for-service Medicare Part D plans. Objective: To evaluate the differences in the rate of HRM prescription fills among beneficiaries of traditional Medicare vs Medicare Advantage Part D plans and to examine the extent to which these differences change over time and the patient-level factors associated with higher rates of HRMs. Design, Setting, and Participants: This cohort study used a 20% sample of Medicare Part D data on filled drug prescriptions from 2013 to 2017 and a 40% sample from 2018. The sample comprised Medicare beneficiaries aged 66 years or older who were enrolled in Medicare Advantage or traditional Medicare Part D plans. Data were analyzed between April 1, 2022, and April 15, 2023. Main Outcomes and Measures: The primary outcome was the number of unique HRMs prescribed to older Medicare beneficiaries per 1000 beneficiaries. Linear regression models were used to model the primary outcome, adjusting for patient characteristics and county characteristics and including hospital referral region fixed effects. Results: The sample included 5 595 361 unique Medicare Advantage beneficiaries who were propensity score-matched on a year-by-year basis to 6 578 126 unique traditional Medicare beneficiaries between 2013 and 2018, resulting in 13 704 348 matched pairs of beneficiary-years. The traditional Medicare vs Medicare Advantage cohorts were similar in age (mean [SD] age, 75.65 [7.53] years vs 75.60 [7.38] years), proportion of males (8 127 261 [59.3%] vs 8 137 834 [59.4%]; standardized mean difference [SMD] = 0.002), and predominant race and ethnicity (77.1% vs 77.4% non-Hispanic White; SMD = 0.05). On average in 2013, Medicare Advantage beneficiaries filled 135.1 (95% CI, 128.4-142.6) unique HRMs per 1000 beneficiaries compared with 165.6 (95% CI, 158.1-172.3) HRMs per 1000 beneficiaries for traditional Medicare. In 2018, the rate of HRMs had decreased to 41.5 (95% CI, 38.2-44.2) HRMs per 1000 beneficiaries in Medicare Advantage and to 56.9 (95% CI, 54.1-60.1) HRMs per 1000 beneficiaries in traditional Medicare. Across the study period, Medicare Advantage beneficiaries received 24.3 (95% CI, 20.2-28.3) fewer HRMs per 1000 beneficiaries per year compared with traditional Medicare beneficiaries. Female, American Indian or Alaska Native, and White populations were more likely to receive HRMs than other groups. Conclusion and Relevance: Results of this study showed that HRM rates were consistently lower among Medicare Advantage than traditional Medicare beneficiaries. Higher use of HRMs among female, American Indian or Alaska Native, and White populations is a concerning disparity that requires further attention.

Excess Mortality at Veterans Health Administration Facilities During the COVID-19 Pandemic.

IMPORTANCE: The COVID-19 pandemic resulted in excess mortality among the general US population and at Veterans Health Administration (VHA) facilities. It is critical to understand the characteristics of facilities that experienced the highest and lowest pandemic-related mortality to inform future mitigation efforts. OBJECTIVE: To identify facility-level excess mortality during the pandemic and to correlate these estimates with facility characteristics and community-wide rates of COVID-19 burden. DESIGN: We used pre-pandemic data to estimate mortality risk prediction models using 5-fold cross-validation and Poisson quasi-likelihood regression. We then estimated excess mortality and observed versus expected (O/E) mortality ratios by the VHA facility from March to December 2020. We examined facility-level characteristics by excess mortality quartile. PARTICIPANTS: Overall, there were 11.4 million VHA enrollees during 2016 and 2020. MAIN MEASURES: Facility-level O/E mortality ratios and excess all-cause mortality. RESULT: VHA-enrolled veterans experienced 52,038 excess deaths from March to December 2020, equating to 16.8% excess mortality. Facility-specific rates ranged from -5.5% to +63.7%. Facilities in the lowest quartile for excess mortality experienced fewer COVID-19 deaths (0.7-1.51, P <0.001) and cases (52.0-63.0, P =0.002) per 1,000 population compared with the highest quartile. The highest quartile facilities had more hospital beds (276.7-187.6, P =0.024) and a higher percent change in the share of visits conducted via telehealth from 2019 to 2020 (183%-133%, P <0.008). CONCLUSIONS: There was a large variation in mortality across VHA facilities during the pandemic, which was only partially explained by the local COVID-19 burden. Our work provides a framework for large health care systems to identify changes in facility-level mortality during a public health emergency.

Racial and ethnic disparities in excess mortality among U.S. veterans during the COVID-19 pandemic.

OBJECTIVE: The COVID-19 pandemic disproportionately affected racial and ethnic minorities among the general population in the United States; however, little is known regarding its impact on U.S. military Veterans. In this study, our objectives were to identify the extent to which Veterans experienced increased all-cause mortality during the COVID-19 pandemic, stratified by race and ethnicity. DATA SOURCES: Administrative data from the Veterans Health Administration's Corporate Data Warehouse. STUDY DESIGN: We use pre-pandemic data to estimate mortality risk models using five-fold cross-validation and quasi-Poisson regression. Models were stratified by a combined race-ethnicity variable and included controls for major comorbidities, demographic characteristics, and county fixed effects. DATA COLLECTION: We queried data for all Veterans residing in the 50 states plus Washington D.C. during 2016-2020. Veterans were excluded from analyses if they were missing county of residence or race-ethnicity data. Data were then aggregated to the county-year level and stratified by race-ethnicity. PRINCIPAL FINDINGS: Overall, Veterans' mortality rates were 16% above normal during March-December 2020 which equates to 42,348 excess deaths. However, there was substantial variation by racial and ethnic group. Non-Hispanic White Veterans experienced the smallest relative increase in mortality (17%, 95% CI 11%-24%), while Native American Veterans had the highest increase (40%, 95% CI 17%-73%). Black Veterans (32%, 95% CI 27%-39%) and Hispanic Veterans (26%, 95% CI 17%-36%) had somewhat lower excess mortality, although these changes were significantly higher compared to White Veterans. Disparities were smaller than in the general population. CONCLUSIONS: Minoritized Veterans experienced higher rates excess of mortality during the COVID-19 pandemic compared to White Veterans, though with smaller differences than the general population. This is likely due in part to the long-standing history of structural racism in the United States that has negatively affected the health of minoritized communities via several pathways including health care access, economic, and occupational inequities.

Effect of mental health staffing inputs on suicide-related events.

OBJECTIVE: To estimate the effects of changes in Veterans Health Administration (VHA) mental health services staffing levels on suicide-related events among a cohort of Veterans. DATA SOURCES: Data were obtained from the VHA Corporate Data Warehouse, the Department of Defense and Veterans Administration Infrastructure for Clinical Intelligence, the VHA survey of enrollees, and customized VHA databases tracking suicide-related events. Geographic variables were obtained from the Area Health Resources Files and the Centers for Medicare and Medicaid Services. STUDY DESIGN: We used an instrumental variables (IV) design with a Heckman correction for non-random partial observability of the use of mental health services. The principal predictor was a measure of provider staffing per 10,000 enrollees. The outcome was the probability of a suicide-related event. DATA COLLECTION/EXTRACTION METHODS: Data were obtained for a cohort of Veterans who recently separated from active service. PRINCIPAL FINDINGS: From 2014 to 2018, the per-pay period probability of a suicide-related event among our cohort was 0.05%. We found that a 1% increase in mental health staffing led to a 1.6 percentage point reduction in suicide-related events. This was driven by the first tertile of staffing, suggesting diminishing returns to scale for mental health staffing. CONCLUSIONS: VHA facilities appear to be staffing-constrained when providing mental health care. Targeted increases in mental health staffing would be likely to reduce suicidality.

Geographic Variation in Appointment Wait Times for US Military Veterans.

Importance: Timely access to medical care is an important determinant of health and well-being. The US Congress passed the Veterans Access, Choice, and Accountability Act in 2014 and the VA MISSION (Maintaining Systems and Strengthening Integrated Outside Networks) Act in 2018, both of which allow veterans to access care from community-based clinicians, but geographic variation in appointment wait times after the passage of these acts have not been studied. Objective: To describe geographic variation in wait times experienced by veterans for primary care, mental health, and other specialties. Design, Setting, and Participants: This is a cross-sectional study using data from the Veterans Health Administration (VHA) Corporate Data Warehouse. Participants include veterans who sought medical care from January 1, 2018, to June 30, 2021. Data analysis was performed from February to June 2022. Exposures: Referral to either VHA or community-based clinicians. Main Outcomes and Measures: Total appointment wait times (in days) for 3 care categories: primary care, mental health, and all other specialties. VHA medical centers are organized into regions called Veterans Integrated Services Networks (VISNs); wait times were aggregated to the VISN level. Results: The final sample included 22 632 918 million appointments for 4 846 892 unique veterans (77.3% male; mean [SD] age, 61.6 [15.5] years). Among non-VHA appointments, mean (SD) VISN-level appointment wait times were 38.9 (8.2) days for primary care, 43.9 (9.0) days for mental health, and 41.9 (5.9) days for all other specialties. Among VHA appointments, mean (SD) VISN-level appointment wait times were 29.0 (5.5) days for primary care, 33.6 (4.6) days for mental health, and 35.4 (2.7) days for all other specialties. There was substantial geographic variation in appointment wait times. Among non-VHA appointments, VISN-level appointment wait times ranged from 25.4 to 52.4 days for primary care, from 29.3 to 65.7 days for mental health, and from 34.7 to 54.8 days for all other specialties. Among VHA appointments, wait times ranged from 22.4 to 43.4 days for primary care, from 24.7 to 42.0 days for mental health, and from 30.3 to 41.9 days for all other specialties. There was a correlation between wait times across care categories and setting (VHA vs community care). Conclusions and Relevance: This cross-sectional study found substantial variation in wait times across care type and geography, and VHA wait times in a majority of VISNs were lower than those for community-based clinicians, even after controlling for differences in specialty mix. These findings suggest that liberalized access to community care under the Veterans Access, Choice, and Accountability Act and the VA MISSION Act may not result in lower wait times within these regions.

Dually-enrolled patients choose providers with lower wait times: Budgetary implications for the VHA.

OBJECTIVE: To estimate the effect of wait times on patients' choice of provider and simulate changes in choice of provider due to compliance with VA MISSION Act wait time targets. DATA SOURCES: We use nationwide administrative data (2014-2017) on Veterans who are enrolled in Medicare and the Veterans Health Administration (VHA), the Survey of VHA Enrollees, Area Health Resource Files, and other data provided by the Centers for Medicare & Medicaid Services. STUDY DESIGN: We use an instrumental variables approach to identify the effect of VHA wait times on the proportion of total (Medicare and VHA) services that are paid for by the VHA ("reliance"). We exploit shocks to VHA provider supply to isolate supply-driven changes in wait times and estimate the effect on VHA reliance. We control for market and time fixed effects and local demand factors. DATA COLLECTION/EXTRACTION METHODS: We use monthly aggregated data on 140 markets (groups of counties). VHA reliance is computed among patients aged 65 years or older who are dually enrolled in VHA and Medicare. VHA wait times and reliance are calculated for multiple specialties: cardiology, gastroenterology, orthopedics, urology, dermatology, and ophthalmology/optometry. PRINCIPAL FINDINGS: A 10% increase in the mean wait time (+2.8 days) reduces VHA reliance by 2.3 percentage points (95% CI: 2.3, 2.7), or 7.9% of the sample mean. This implies that meeting the MISSION Act wait time targets may have multi-billion-dollar budgetary impacts. Effects vary across specialties. For example, a 10% increase in the mean wait time for cardiology services (+2.0 days) reduces reliance by 1.8 percentage points (95% CI: 1.6, 2.1), or 6.3% of the sample mean for cardiology services. CONCLUSIONS: Meeting statutory wait time targets may have substantial unforeseen impacts on federal health care spending as patients sort to providers who have lower wait times.

County-Level Impact of the COVID-19 Pandemic on Excess Mortality Among U.S. Veterans: A Population-Based Study.

BACKGROUND: As the novel coronavirus (COVID-19) continues to impact the world at large, Veterans of the US Armed Forces are experiencing increases in both COVID-19 and non-COVID-19 mortality. Veterans may be more susceptible to the pandemic than the general population due to their higher comorbidity burdens and older age, but no research has examined if trends in excess mortality differ between these groups. Additionally, individual-level data on demographics, comorbidities, and deaths are provided in near-real time for all enrolees of the Veterans Health Administration (VHA). These data provide a unique opportunity to identify excess mortality throughout 2020 at a subnational level, and to validate these estimates against local COVID-19 burden. METHODS: We queried VHA administrative data on demographics and comorbidities for 11.4 million enrolees during 2016-2020. Pre-pandemic data was used to develop and cross-validate eight mortality prediction models at the county-level including Poisson, Poisson quasi-likelihood, negative binomial, and generalized estimating equations. We then estimated county-level excess Veteran mortality during 2020 and correlated these estimates with local rates of COVID-19 confirmed cases and deaths. FINDINGS: All models demonstrated excellent agreement between observed and predicted mortality during 2016-2019; a Poisson quasi-likelihood with county fixed effects minimized median squared error with a calibration slope of 1.00. Veterans of the U.S. Armed Forces faced an excess mortality rate of 13% in 2020, which corresponds to 50,299 excess deaths. County-level estimates of excess mortality were correlated with both COVID-19 cases (R2=0.77) and deaths per 1,000 population (R2=0.59). INTERPRETATION: We developed sub-national estimates of excess mortality associated with the pandemic and shared our data as a resource for researchers and data journalists. Despite Veterans' greater likelihood of risk factors associated with severe COVID-19 illness, their excess mortality rate was slightly lower than the general population. Consistent access to health care and the rapid expansion of VHA telemedicine during the pandemic may explain this divergence. FUNDING: This work was supported by grants from the Department of Veterans Affairs Quality Enhancement Research Initiative [PEC 16-001]. Dr. Griffith's effort was supported in part by the Agency for Healthcare Research & Quality [K12 HS026395].

Appointment wait time data for primary & specialty care in veterans health administration facilities vs. community medical centers.

The datasets summarized in this article include more than 38 million appointment wait times that U.S. military veterans experienced when seeking medical care since January 2014. Our data include both within Veterans Health Administration (VHA) facilities and community medical centers, and wait times are stratified by primary/specialty care type. Deidentified wait time data are reported at the referral-level, at the VHA facility-level, and at the patient's 3-digit ZIP code-level. As of this writing, no other U.S. health care system has made their wait times publicly available. Our data thus represent the largest, national, and most representative measures of timely access to care for patients of both VHA and community providers. Researchers may use these datasets to identify variations in appointment wait times both longitudinally and cross-sectionally, conduct research on policies and interventions to improve access to care, and to incorporate fine-grained measures of wait times into their analyses.

County-level data on U.S. opioid distributions, demographics, healthcare supply, and healthcare access.

The dataset summarized in this article is a combination of several of U.S. federal data resources for the years 2006-2013, containing county-level variables for opioid pill volumes, demographics (e.g. age, race, ethnicity, income), insurance coverage, healthcare demand (e.g. inpatient and outpatient service utilization), healthcare infrastructure (e.g. number of hospital beds or hospices), and the supply of various types of healthcare providers (e.g. medical doctors, specialists, dentists, or nurse practitioners). We also include indicators for states which permitted opioid prescribing by nurse practitioners. This dataset was originally created to assist researchers in identifying which factors predict per capita opioid pill volume (PCPV) in a county, whether early state Medicaid expansions increased PCPV, and PCPV's association with opioid-related mortality. Missing data were imputed using regression analysis and hot deck imputation. Non-imputed values are also reported. Taken together, our data provide a new level of precision that may be leveraged by scholars, policymakers, or data journalists who are interested in studying the opioid epidemic. Researchers may use this dataset to identify patterns in opioid distribution over time and characteristics of counties or states which were disproportionately impacted by the epidemic. These data may also be joined with other sources to facilitate studies on the relationships between opioid pill volume and a wide variety of health, economic, and social outcomes.

Implications of county-level variation in U.S. opioid distribution.

BACKGROUND: Prescription opioids accounted for the majority of opioid-related deaths in the United States prior to 2010, and continue to contribute to opioid misuse and mortality. We used a novel dataset to investigate the distributional patterns of prescription opioids, whether opioid pill volume was associated with opioid-related mortality, and whether early state Medicaid expansions were associated with either pill volume or opioid-related mortality. METHODS: Data on opioid shipments to retail pharmacies for 2006-2013 were obtained from the U.S. Drug Enforcement Administration, and opioid-related deaths (ORDs) were obtained from the Centers for Disease Control and Prevention. We first compared characteristics of counties in the highest and lowest quartiles for per capita pill volume (PCPV). We used adjusted difference-in-differences regression models to identify factors associated with PCPV or ORDs, and whether early state Medicaid expansions were associated with either outcome. All models were estimated as linear regressions with standard errors clustered by county, and weighted by county population. RESULTS: We found large geographic variations in opioid distribution, and this variation appears to be driven by differences in demographics, healthcare access, and healthcare supply. In adjusted models, a one-pill increase in PCPV was associated with a 0.20 increase in ORDs per 100,000 population (95 % CI 0.11-0.30). Early Medicaid expansions were associated with lower PCPV (-2.20, 95 % CI -2.97 to -1.43). CONCLUSIONS: Our findings validate the relationship between PCPV and ORDs, identify important environmental drivers of the opioid epidemic, and suggest early state Medicaid expansions were beneficial in reducing opioid pill volume.

Association of Childcare Facility Closures With Employment Status of US Women vs Men During the COVID-19 Pandemic.

This cross-sectional study assesses the association of closures of childcare facilities with the employment status of women and men with children in the US during the COVID-19 pandemic.

The persistence of medicare advantage spillovers in the post-Affordable Care Act era.

Spillovers can arise in markets with multiple purchasers relying on shared producers. Prior studies have found such spillovers in health care, from managed care to nonmanaged care populations-reducing spending and utilization, and improving outcomes, including in Medicare. This study provides the first plausibly causal estimates of such spillovers from Medicare Advantage (MA) to Traditional Medicare (TM) in the post-Affordable Care Act era using an instrumental variables approach. Controlling for health status and other potential confounders, we estimate that a one percentage point increase in county-level MA penetration results in a $64 (95% CI: $18 to $110) (0.7%) reduction in standardized per-enrollee TM spending. We find evidence for reductions in utilization both on the intensive and extensive margins, across a number of health care services. Our results complement and extend prior work that found spillovers from MA to TM in earlier years and under different payment policies than are in place today.

Effectiveness of COVID-19 shelter-in-place orders varied by state.

State "shelter-in-place" (SIP) orders limited the spread of COVID-19 in the U.S. However, impacts may have varied by state, creating opportunities to learn from states where SIPs have been effective. Using a novel dataset of state-level SIP order enactment and county-level mobility data form Google, we use a stratified regression discontinuity study design to examine the effect of SIPs in all states that implemented them. We find that SIP orders reduced mobility nationally by 12 percentage points (95% CI: -13.1 to -10.9), however the effects varied substantially across states, from -35 percentage points to +11 percentage points. Larger reductions were observed in states with higher incomes, higher population density, lower Black resident share, and lower 2016 vote shares for Donald J. Trump. This suggests that optimal public policies during a pandemic will vary by state and there is unlikely to be a "one-size fits all" approach that works best.