Health Services Research最新文献

Objective: To investigate women's perceptions of telemedicine for reproductive health care services, focusing on how perceived benefits and privacy risks influence their intentions to adopt telemedicine and their willingness to disclose personal health information.

Study setting and design: A cross-sectional survey was conducted. The study applied the privacy calculus theory to the context of telemedicine for reproductive health, using adapted, validated variables to develop the survey. Outcome variables included intentions to adopt telemedicine and willingness to disclose accurate personal health information.

Data sources and analytic sample: Data were collected in May and June 2023 using Qualtrics online panel services, targeting women across the United States who had not used telemedicine for reproductive health. The sample comprised 847 women aged 18 and older. Structural equation modeling was employed using AMOS v28.0 to test the hypothesized relationships between perceived benefits, perceived risks, and adoption intentions. The analysis controlled for age, household income, political affiliation, religious views, and prior births.

Principal findings: Perceived benefits were positively related to intention to adopt telemedicine for reproductive care (β: 0.600, p < 0.001), and willingness to disclose accurate personal health information (β: 0.453, p < 0.001). Unexpectedly, perceived privacy risks were positively related to adoption intentions (β: 0.128, p < 0.001), but negatively related to willingness to disclose (β: -0.282, p < 0.001). Intentions to adopt were positively associated with willingness to disclose (β: 0.089, p < 0.05). Lastly, older women and women located in states with abortion restrictions expressed lower intentions to adopt. The model explained 40.2% of variance in intention to adopt and 38.3% of variance in willingness to disclose.

Conclusions: The study demonstrates the importance of perceived benefits and privacy risks in driving telemedicine adoption and disclosure intentions among women in the reproductive health context. These findings suggest the need for targeted strategies to address privacy concerns and support telemedicine adoption, particularly in restrictive regulatory environments.

Objective: To characterize the landscape of policies that determine eligibility for family planning services through Medicaid programs and describe trends in eligibility and its determinants over time.

Data sources and study setting: Secondary data were collected for all states in the United States for the years 2008 through 2023. Data on economic and demographic characteristics came from the American Community Survey (ACS).

Study design: Our descriptive study characterized state adoptions of Medicaid family planning section 1115 waivers and state plan amendments (SPA) and their eligibility criteria. We then estimated the proportion of women aged 19-44 years who were eligible for family planning services through Medicaid and identified the key determinants of changes in eligibility, by state and year.

Data collection/extraction methods: Information on state Medicaid policies was extracted from documentation on the Centers for Medicare & Medicaid Services website. When estimating the eligible population sizes, the denominator was women aged 19-44 years, the group most likely to be eligible for Medicaid family planning programs. Supplemental data on program enrollment or utilization were collected from states' websites and reports.

Principal findings: Though eligibility limits for family planning through Medicaid generally increased over time, the proportion of women aged 19-44 years eligible for at least limited benefits decreased from 45.0% in 2012 to 39.4% in 2022, largely because of increases in household income. Trends varied considerably across states and by eligibility pathway. Among women with incomes below the poverty level, the proportion who were not eligible for Medicaid family planning services decreased from 6.3% in 2013 to 1.5% in 2022.

Conclusions: Our data demonstrated substantial geographic and temporal variation in eligibility for family planning services through Medicaid. We identified key drivers of eligibility changes that may have important implications for health services analyses of means-tested public programs.

Objective: To examine how private Medicare Advantage (MA) plans responded to slower growth in federal payments after the Affordable Care Act (ACA).

Data sources and study setting: We used publicly available data from the Centers for Medicare and Medicaid Services on MA plan subsidies ("benchmarks"), asking prices ("bids"), plan premiums, cost-sharing, and covered benefits. Data covered all counties with MA plans between 2006 through 2019.

Study design: We examined plan responses to changes in benchmark subsidies by comparing changes in bids, rebates, and other outcomes between counties experiencing larger changes in benchmarks and counties with smaller changes, pre- and post-ACA. We used longitudinal fixed effects regression models to assess heterogeneity in how plans adjusted premiums and benefits across more and less competitive markets.

Data collection: Analyses included all counties with at least one MA plan available to individual beneficiaries. Plans targeting special populations were excluded.

Principal findings: Average plan benchmarks fell by $89 per month post-ACA, adjusted for inflation. Plans responded similarly to benchmark subsidy decreases and increases, increasing bids by 62 cents for every dollar increase in subsidies pre-ACA (95% confidence interval [CI]: 0.56 to 0.67) and decreasing them by 57 cents for every dollar reduction in subsidies post-ACA (95% CI: 0.49-0.65). However, post-ACA, plans altered less salient benefits, such as cost-sharing, by about twice as much as they had pre-ACA. Premiums changed by similar amounts before and after the ACA (-$0.07, 95% CI: from -$0.09 to -$0.06). Plans in more competitive markets responded less to payment changes than plans did in less competitive markets, suggesting the former are operating closer to marginal costs. Finally, payments to plans declined far less than projected due in part to other changes in MA policy.

Conclusions: Plans used partial pass-through of benchmark subsidy decreases to shield beneficiaries from cuts and allocated benchmark decreases to benefits that were less salient to the average enrollee. These findings, combined with higher-than-projected payments post-ACA, may explain the continued growth in MA enrollment.

Objective: To examine the association of Medicaid home- and community-based services (HCBS) expenditures on the home care workforce.

Data sources/study setting: We use two national, secondary data sources from 2008 to 2019: state-level Medicaid HCBS expenditures and the American Community Survey, in which we identify direct care workers in the home (i.e., home care workers), defined as nursing, psychiatric, and home health aides or personal care aides working in home health care services, individual and family services, and private households.

Study design: Our key explanatory variable is HCBS expenditures per state per year. To estimate the association between changes in Medicaid HCBS expenditures and the workforce size, hourly wages and hours worked, we use negative binomial, linear, and generalized ordered logit regression, respectively. All models include demographic and socioeconomic characteristics, the number of potential HCBS beneficiaries (individuals with a disability and income under the federal maximum income eligibility limits), indicators for minimum wage and/or overtime protections for direct care workers, wage pass-through policies, and state and year fixed effects.

Data collection/extraction methods: We exclude states with incomplete reporting of expenditures.

Principal findings: States' HCBS expenditures increased between 2008 and 2019 after adjusting for inflation and the number of potential HCBS beneficiaries. Yet, home care workers' wages remained stagnant at $11-12/h. We find no association between changes in Medicaid HCBS expenditures and wages. For every additional $1 million in Medicaid HCBS expenditures, the expected number of workers increases by 1.2 and the probability of working overtime increased (0.0015% points; p < 0.05). Results are largely robust under multiple sensitivity analyses.

Conclusions: We find no evidence of a statistically significant relationship between changes in state-level changes in Medicaid HCBS expenditures and worker wages but do find a significant, but small, association with hours worked and workforce size.

Objective: To evaluate whether having previously disenrolled from Medicare Advantage (MA) is associated with lower hazards of future MA enrollment.

Data sources and study setting: Secondary data from Medicare.

Study design: We examined beneficiaries with baseline FFS enrollment from 2017-2019 using a 20% sample of Medicare claims. Cox proportional hazard models were used to examine the association of prior MA enrollment (in the three years prior to baseline FFS enrollment) with MA re-enrollment, and whether this association is modified by Alzheimer's Disease and Related Dementias (ADRD), prior nursing home use, chronic illness, dual eligible status, and availability of MA plans and quality.

Data collection: Not applicable.

Principal findings: Overall, 3.3% of beneficiaries switched to MA annually. Of those with prior MA enrollment, MA switching percentages were 9.0%, 4.6%, and 6.8% for those whose most recent MA enrollments were 1, 2, and 3 years prior to their baseline FFS year. Comparatively, the switching percentages was 3.2% for those with no prior MA enrollment. The hazards of switching to MA were 2.73 (p < 0.001), 1.29 (p < 0.001), and 1.97 (p < 0.001) times greater than remaining in FFS for beneficiaries whose most recent MA enrollments were one, two, and three years prior to their baseline FFS year. Hazards of switching were generally similar between those with and without ADRD, stratified by recency in prior MA experience, except those with dual eligibility. Among those with ADRD, switching hazards were greatest for 3 years prior MA enrollees in counties with the fewest available (HR: 3.84, p < 0.001) and lowest-rated plans (HR: 4.02, p < 0.001).

Conclusions: Recency of switching from MA to FFS was the strongest predictor of a FFS-to-MA switch, identifying a population of beneficiaries who multiply switch regardless of health status or MA access. Future health policy considerations should more closely examine the vulnerabilities and long-term outcomes of this population.

Objective: To decompose the mental health disparities between breast cancer patients and survivors (hereafter survivors) of racial and ethnic minority groups and non-Hispanic White survivors into the contributions of individual-, interpersonal-, community-, and societal-level determinants.

Data sources and study setting: We used data from the 2010-2020 Medical Expenditure Panel Survey Household Component (MEPS-HC). Our primary outcome was whether the person had mental health conditions or not.

Study design: We employed the Kitagawa-Oaxaca-Blinder (KOB) method to understand to what extent the differences in outcomes were explained by the differences in the determinants between non-Hispanic Black or Hispanic breast cancer survivors and non-Hispanic White survivors. We also bifurcated the Hispanic sample analysis by the US-born status (and county of origin).

Data collection/extraction methods: Confidential geographic identifiers are utilized to supplement the MEPS-HC data with information on community characteristics and local healthcare resources.

Principal findings: The prevalence of mental health conditions among non-Hispanic Black and Hispanic breast cancer survivors was 26.1% (95% CI: 20.4, 31.7) and 28.3% (95% CI: 21.9, 34.6), respectively. These rates were higher than those for their non-Hispanic White counterparts, 19.7% (95% CI: 17.4, 21.9). In our KOB model, the disparity between non-Hispanic Black and White survivors was fully explained by differences in education, health, and family structure, with community- and societal-level determinants playing no significant role. Conversely, our KOB model did not explain any of the overall differences between Hispanic and non-Hispanic White survivors. However, for foreign-born Hispanic survivors, the disparity was fully explained by a combination of individual- and societal-level determinants.

Conclusions: Our findings, which identify specific individual-, interpersonal-, and societal- determinants that were associated with racial and ethnic differences in mental health, can be used by clinicians and policymakers to proactively address racial and ethnic disparities in health.

Objective: The 2020 Family First Coronavirus Response Act (FFCRA) extended Medicaid coverage for enrollees without rechecking eligibility. Before that, women were eligible for Medicaid coverage up to 60 days postpartum. We examine the FFCRA's effect on Medicaid postpartum coverage, depression symptoms, and birth control use beyond 60 days after delivery.

Data sources and setting: We use data from the 2016-2021 Pregnancy Risk Assessment Monitoring System (PRAMS). The primary sample includes 56,828 women who were income eligible up to 60 days postpartum or beyond in their state.

Study design: We employ a difference-in-differences design comparing outcomes before and after the FFCRA between women who were not income eligible for Medicaid coverage beyond 60 days postpartum in their state and those who were income eligible. We estimate models without and with controlling for state-specific trends in outcomes over time.

Data collection/extraction methods: N/A.

Principal findings: There is an 8.1 percentage-point increase in the likelihood of having Medicaid coverage without controlling for state-specific trends, and 5.4 percentage-points when controlling for state-specific trends (both p < 0.05). There is a decline in likelihood of being uninsured by 3 percentage-points (p < 0.05) without state-specific trends and a smaller and non-significant decline when including state-specific trends. Estimated effects on depression symptoms and birth control use are small and statistically non-significant.

Conclusion: Following the FFCRA, there was an increase in Medicaid coverage beyond 60 days postpartum among women who would have been income ineligible for Medicaid after 60 days. However, there is a less pronounced effect on likelihood of being uninsured, which might be partly due to some switching from private to Medicaid coverage. There were no discernable effects on depression symptoms and birth control use. Examining additional health and health care utilization measures over a longer postpartum period is needed to further understand the FCCRA effects on these outcomes.

Objective: To examine whether Medicaid managed care plan (MCP) utilization management policies for buprenorphine-naloxone and injectable naltrexone are related to key state Medicaid program policy decisions.

Data sources and study setting: We abstracted data on state Medicaid regulatory and policy information from publicly available sources and publicly available insurance benefit documentation from all 241 Medicaid MCPs operating in 2021.

Study design: In this cross-sectional study, we used bivariate and multivariate analyses to examine whether Medicaid MCP prior authorization and quantity limits on receipt of buprenorphine and injectable naltrexone were associated with key state Medicaid choices to leverage federal funds to expand coverage and eligibility (Medicaid expansion, 1115 waivers) and to regulate Medicaid MCPs (uniform preferred drug lists, medical loss ratio remittance). Models were adjusted for MCP characteristics, including profit status, behavioral health contracting arrangement, National Committee for Quality Assurance accreditation, size, market share, and state opioid overdose death rates. Average marginal effects (AME) were reported.

Principal findings: Utilization management was common among MCPs, and restrictions were more commonly applied to buprenorphine than injectable naltrexone, despite its higher cost. States that required MCPs to comply with utilization management policies stipulated in a uniform preferred drug list were more likely to require prior authorization for buprenorphine (AME: 0.29, 95% CI: 0.15-0.42) and injectable naltrexone (AME: 0.25, 95% CI: 0.12-0.38). MCPs in states that required plans to pay back earnings above a certain threshold were less likely to require prior authorization for buprenorphine (AME: -0.30, 95% CI: -0.43 to -0.18).

Conclusions: Restrictions on medications for opioid use disorder are widespread among MCPs and vary by medication. State Medicaid regulatory and policy characteristics were strongly linked to MCPs' utilization management approaches. State Medicaid policy and contracting approaches may be levers to eliminate utilization management restrictions on medications for opioid use disorder.

Objective: To understand why American Indian and Alaskan Native (AIAN) populations have had exceptionally high COVID-19 mortality, we compare patterns of healthcare utilization and outcomes for two serious infectious respiratory diseases-Influenza-like-illness (ILI) and coronavirus disease 2019 (COVID-19)-between American Indian and Alaskan Native (AIAN) populations (as identified in Medicaid data) and non-Hispanic Whites over the 2009-2021 period.

Study setting and design: We select all people under the age of 65 years identified as non-Hispanic White or AIAN in the New York State Medicaid claims data between 2009 and 2021. We analyze data across 10 ILI cohorts (between September 2009 and August 2020) and 4 COVID-19 cohorts (March-June 2020, July-September 2020, October-December 2020, and January-June 2021). We examine mortality and utilization rates using logistic regressions, adjusting for demographic characteristics, prior chronic conditions, and geographic location (including residence near a reservation). We stratify the analysis by rural vs. nonrural counties.

Data sources and analytic sample: We use the New York State Medicaid claims data for the analysis.

Principal findings: We find that even among Medicaid beneficiaries, who are similar in socioeconomic status and identical in health insurance coverage, AIAN populations have much lower rates of use of outpatient services and much higher rates of acute (inpatient and emergency room) service utilization for both ILI and COVID-19 than non-Hispanic Whites. Prior to COVID-19, demographic and health status-adjusted all-cause mortality rates, including from ILI, were lower among American Indians than among non-Hispanic Whites on New York State Medicaid, but this pattern reversed during the COVID-19 pandemic. Both findings are driven by nonrural counties. We did not observe significant differences in all-cause mortality and acute service utilization comparing AIAN to non-Hispanic Whites in rural areas.

Conclusion: The utilization and mortality disparities we identify within the Medicaid population highlight the need to move beyond insurance in addressing poor health outcomes in the American Indian population.