Health Services Research最新文献

Objective: To characterize health insurance gap patterns related to age-19 Medicaid and age-26 commercial age-eligibility cutoffs.

Study setting and design: This descriptive analysis spans 2014-2018, after Affordable Care Act implementation, but before COVID-19 emergency provisions. We defined insurance gaps as ≥3 consecutive months without observed enrollment, preceded and followed by ≥1 month of enrollment and stratified results by insurance source and clinical severity (e.g., with chronic, complex, or disabling conditions or not).

Data sources and analytic sample: The Colorado all-payer claims database provided data for enrollees aged 10-29, 52% (649,346) of whom were initially Medicaid insured, whereas 47% (576,596) were commercially insured.

Principal findings: The percent of enrollees with insurance gaps peaks within six months of turning age-19 and age-26-at 8.9% Medicaid and 8.7% commercial, respectively. The percentage point difference between ages 27-28 and 11-18 was 3.3 percentage points higher for prior Medicaid recipients (p < 0.001) and 2.2 percentage points greater for prior commercial enrollees (p < 0.001). Relative to the other clinical severity groups, young adults with disabling health conditions who were initially Medicaid insured had the lowest peak gap rate, 5.7%, compared with 10.5% among the previously commercially insured; this latter finding was sensitive to gap specification.

Conclusions: Young adults would likely benefit from greater attention to age-19 and age-26 health insurance "unwinding."

Objective: To examine the extent of segregation between hospitals for Medicare beneficiaries by race, ethnicity, and dual-eligible status over time.

Data sources and study setting: We used Medicare inpatient hospital provider data for fee-for-service (FFS) beneficiaries, and the Dartmouth Atlas of Health Care from 2013 to 2021 nationwide, for hospital referral regions (HRRs), and for and hospital service areas (HSAs).

Study design: We conducted time trend analysis with dissimilarity indices (DIs) for Black (DI-Black), Hispanic (DI-Hispanic), non-White (including Black, Hispanic, and other non-White) (DI-non-White), and dual-eligible (DI-Dual) beneficiaries. DIs between hospitals were contextualized and correlated with population compositions and residential DIs.

Data collection/extraction methods: We included 3177 hospitals with more than 250 Medicare FFS beneficiaries discharged per year. We cross-linked data on hospital-level patient race, ethnicity, and dual-eligible status with geographic data and examined time trends using linear mixed models.

Principal findings: Nationwide DIs ranged from 0.23 to 0.53. HRRs and HSAs generally had low segregation (DI medians: 0.08-0.19, highest among Black, then non-White, Hispanic, and dual-eligible beneficiaries). However, some HRRs and HSAs had moderate or high segregation (DI-Black >0.30 in 19.1% of HRRs and 5.8% of HSAs; DI-non-White >0.30 for two HRRs with high American Indian/Alaska Native populations). Time trends indicated small declines in segregation from 2013 to 2021 (0.15%-0.30% per year; all p < 0.001). DI-Dual correlated moderately with non-White populations.

Conclusions: For Medicare FFS, we observe generally low and slightly declining levels of segregation across HRRs and HSAs, with notable exceptions. Improving race reporting and contextualizing select areas of higher segregation with their hospital and residential population compositions can help frame and understand health inequities. Interpretation of HRR-level DI may require additional historical, demographic, and spatial context due to its potential to oversimplify, overstate, or obscure segregation. Future work should identify drivers and mitigators of segregation, including sorting patterns among health systems.

Objective: To examine the effects of data breach incidents on healthcare quality and to explore potential mechanisms.

Data source: Hospital-level data breach reports from the U.S. Department of Health and Human Services and patient-level hospitalization records from Florida State Inpatient Database during 2013-2017.

Study design: We employ a propensity score matching difference-in-difference model to estimate changes in a patient's emergency department (ED) door-to-hospital admission hours, days to undergo principal procedure after admission, length of stay days, and in-hospital mortality rates following data breaches. We compare the health information technology (HIT) functionalities of breached and non-breached hospitals during both pre and post periods.

Data collection/extraction methods: Our primary analysis covers 1,295,537 records of inpatients admitted through the EDs of 12 hospitals.

Principal findings: Data breaches are associated with long-term improvements in healthcare quality, particularly in the timeliness of patient care and acute myocardial infarction (AMI) mortality. Over time, patients experience a reduction of 0.56 h in ED door-to-hospital admission time (95% confidence interval [CI]: -1.04 to -0.06 h) and a decrease of 0.18 days in time to undergo the principal procedure after hospital admission (95% CI: -0.23 to -0.13 days). Additionally, AMI patients experience a one percentage point reduction in in-hospital mortality (95% CI: -2 to -0.06 percentage points), while mortality rates for other patient groups remain unchanged. Hospitals affected by data breaches show long-term advancements in their HIT functionalities.

Conclusions: Hospital data breach incidents are associated with improved healthcare quality. This improvement may be attributed to hospitals' enhanced functionality of HIT.

Objective: To examine the association of Massachusetts Medicaid Accountable Care Organization (ACO) implementation with changes in mental health care utilization in the postpartum period.

Study setting and design: We examine care for people with a birth covered by Medicaid or private insurance. We used a difference-in-differences design to compare differences before and after Medicaid ACO implementation for those with Medicaid versus those with private insurance. The primary outcome was a binary measure of having at least one outpatient mental health care visit in the 6 months postpartum. We estimated linear probability models controlling for age, prenatal mental illness, pregnancy complications, birth mode, and ZIP code characteristics.

Data sources and analytic sample: Data are from the Massachusetts All-Payer Claims Database. The analytic sample included Massachusetts residents with a live birth between July 1, 2016, and September 30, 2019, with complete data.

Principal findings: 107,813 births were included (53.0% Medicaid, 47.0% private). 7.8% of these had at least one outpatient mental health visit in the 6 months postpartum, with similar rates among those with Medicaid versus those with private insurance pre-ACO implementation (7.9% Medicaid versus 7.7% private). An increase in utilization among privately insured individuals and a decrease among Medicaid beneficiaries post-ACO implementation was observed. Regression-adjusted difference-in-differences estimates indicate that Medicaid ACO implementation was associated with a 1.3 percentage point [pp] decrease (95% confidence interval: 1.3 pp, -0.5 pp; p < 0.01) in the probability of having an outpatient mental health visit for those with Medicaid.

Conclusions: Medicaid ACO implementation was associated with decreases in use of outpatient mental health care in the postpartum period among people with Medicaid, overall and compared to those with private insurance. Future research should determine whether this increased disparity in mental health care utilization persists with maturation of the ACO delivery model.

Objective: To assess mental health related outcomes of Recipe4Health, a multisectoral social care partnership implementing produce prescriptions with or without group medical visits (GMVs).

Study setting and design: Recipe4Health was implemented at five community health centers from 2020 to 2023. Primary care teams referred patients with food insecurity and/or nutrition-sensitive chronic conditions (e.g., diabetes, depression) to 16 weeks of Food Farmacy (produce prescriptions) with the option of GMV participation. We used a convergent mixed-methods design including survey and interview data.

Data sources and analytic sample: We conducted (1) participant surveys pre- and post-intervention and (2) semi-structured interviews with Recipe4Health participants and partner organization staff. Linear mixed effects models examined changes in mental health and related outcomes. Interviews were analyzed using codebook thematic analysis.

Principal findings: Program participants were middle-aged, primarily women, and from diverse racial/ethnic backgrounds (majority Latine and Black). At baseline, moderate or severe depression and/or anxiety symptoms were reported by 77/188 (41%) of Food Farmacy-only participants, and 113/284 (40%) of Food Farmacy +GMV participants. Among Food Farmacy-only participants, post-intervention depression and anxiety symptoms significantly improved only among those who did not have baseline depression/anxiety (PHQ9: -1.7 [95% CI: -2.8, -0.6]; GAD7: -1.8 [95% CI: -2.9, -0.8]). Among Food Farmacy +GMV participants, mental health symptoms improved regardless of baseline mental health; among those with baseline depression/anxiety: PHQ9: -2.4 (95% CI: -3.6, -1.2); GAD7: -0.9 (95% CI: -2.0, 0.1); among those without: PHQ9: -2.2 (95% CI: -3.2, -1.2); GAD7: -2.2 (95% CI: -3.1, -1.2). Improvements in social needs (food insecurity, loneliness) and health-related behaviors (fruit/vegetable intake, physical activity) varied by intervention arm and baseline depression/anxiety symptom level. In interviews, staff and patients endorsed produce prescriptions for improving nutrition and food insecurity, and GMVs for increasing social support.

Conclusion: Social care interventions providing vegetables and fruit, with or without group medical visits, may concurrently address mental health symptoms and social needs.

Objective: To evaluate the completeness and quality of Medicaid comprehensive managed care (CMC) data in national MAX/TAF research files.

Study setting and design: This observational study compared CMC with fee-for-service (FFS) enrollee data in 2001-2019 Medicaid MAX/TAF inpatient, outpatient, and pharmacy files. Completeness was assessed as the proportion of enrollees with any claim and mean claims per enrollee with any claim. Quality was assessed as the proportion of inpatient and outpatient claims with primary diagnosis and procedure codes and the proportion of prescription drug claims with fill dates, National Drug Codes (NDC), days supplied, and quantity dispensed. Acceptable ranges for each study measure were defined as the national FFS mean ± 2 standard deviations.

Data sources and analytic sample: We analyzed secondary data on 45 states from 2001 to 2013 (MAX) and 50 states and DC from 2014 to 2019 (TAF). The sample included adults aged 18-64 with continuous calendar-year enrollment who were eligible for full Medicaid benefits and ineligible for Medicare. We determined CMC enrollment rates and assessed data completeness and quality among state-years with ≥10% CMC penetration, comparing CMC with FFS enrollees.

Principal findings: Across 891 state-years, 194,364,647 enrollees met inclusion criteria. Of 540 state-years (60.6%) with ≥10% CMC enrollment, CMC data were largely comparable to national FFS distributions for all inpatient (n = 430; 79.6%), outpatient (n = 467, 86.5%), and prescription (n = 459, 85.0%) completeness criteria and for all inpatient (n = 449, 83.1%), outpatient (n = 511, 94.6%), and prescription (n = 528, 97.8%) quality criteria. Overall completeness (92.3%) and quality (84.6%) improved substantially by 2019.

Conclusions: Completeness and quality of CMC data were largely comparable to FFS data, with increasing state-years meeting criteria over time. Further research on national Medicaid populations should assess and address differences in data completeness and quality by plan type across states, over time, and in relation to specific study samples and measures of interest.

Objective: To compare healthcare services utilization across the healthcare system between frequent and non-frequent emergency department (ED) users among Medicaid enrollees in South Carolina.

Study setting and design: We conducted a retrospective, longitudinal study of individuals with at least one ED visit in 2017 in South Carolina and identified their healthcare services visits over 730 days (2 years) after their first ED visit. We classified individuals based on intensity of ED use: superfrequent (≥9 ED visits/year), frequent (4-8 ED visits/year), and non-frequent ED users (≤3 visits/year). We estimated differences between the three groups of ED users and non-ED hospital and office-based visits using multivariable two-part regression models.

Data sources and analytic sample: We used statewide Medicaid claims from January 2017 to December 2019 for ED users aged 18-64 years with continuous Medicaid enrollment. We analyzed data on all frequent and superfrequent users and selected a 4:1 random sample among all non-frequent users (~half of all non-frequent users).

Principal findings: The study included 52,845 ED users, of whom 42,764 were non-frequent, 7677 frequent, and 2404 superfrequent users. Within 2 years from the date of their first ED visit, superfrequent ED users averaged 38.3 ED visits, frequent ED users 10.9 ED visits, and non-frequent ED users 2.6 ED visits (p < 0.001). Compared with non-frequent users, frequent and superfrequent ED users had more comorbidities and chronic conditions on average (1.6 vs. 3.5 vs. 6.4, p < 0.001). Both frequent and superfrequent users had more hospital visits beyond the ED overall (marginal effects: 0.23, 95% CI 0.18-0.27; 0.40, 95% CI 0.29-0.50), and more outpatient office visits overall (marginal effects: 4.39, 95% CI 2.52-6.27; 9.23, 95% CI 5.66-12.81), including primary care and most specialists' visits, compared with non-frequent users.

Conclusions: Frequent ED users utilized non-ED hospital and outpatient office-based healthcare services significantly more than non-frequent ED users. These findings can guide tailored interventions using data across the healthcare system to efficiently coordinate care, contain costs, and improve health outcomes for these individuals.

Objective: To evaluate the effect of enrolling in Supportive Services for Veteran Families (SSVF) on short- and long-term housing outcomes among Veterans experiencing housing instability.

Study setting and design: We analyzed data from the Department of Veterans Affairs (VA) electronic health record (EHR) between October 2015 and December 2018 using the target trial emulation framework. Veterans were included in one or more trials if they were 18 years or older, had recent evidence of housing instability, had received care in VA for at least 1 year, and had never before enrolled in SSVF. Patients who enrolled in SSVF after meeting eligibility were assigned to the treatment group, while patients who did not enroll in SSVF were assigned to a control group. We extracted patients' housing outcomes from the EHR and modeled the probability of being unstably housed each day while accounting for confounders and irregular visit times.

Data sources and analytic sample: We extracted housing status and covariates from the VA Corporate Data Warehouse. Housing instability was ascertained using a combination of structured data elements and natural language processing.

Principal findings: We identified 238,059 unique patients who met the eligibility criteria for one or more trials. The risk of housing instability decreased in both arms over the three years after initiating a trial but was lower among SSVF enrollees, with a risk difference of -12.9% (95% confidence band -14.6%, -11.2%) on Day 120 and an attenuated difference of -2.4% (-4.0%, -0.7%) on Day 1095.

Conclusions: SSVF is one of the largest rapid rehousing and homelessness prevention programs in the nation. We found that SSVF improves housing outcomes over the three years following enrollment, but the effect reduces over time. These findings can inform policy and program design for improving housing outcomes for homeless-experienced individuals.

Objective: To examine the impact of medical and recreational cannabis laws on inpatient visits for asthma and by payer-type.

Study setting and design: Quasi-experimental difference-in-differences regression analysis was conducted while accounting for variations in cannabis laws implementation timing by states. Inpatient visits for asthma in states with a given type of cannabis law were compared with those in states that did not implement the specific law. Four different cannabis laws were examined in the study-initial passage of medical cannabis law, opening of a medical cannabis dispensary, home cultivation of medical cannabis, and recreational cannabis legalization.

Data sources and analytic sample: State-level quarterly inpatient visit data for asthma patients were utilized from the Healthcare Cost and Utilization Project Fast Stats database. The primary analysis included inpatient visits for asthma by all payer adult patients aged 19 and above in 38 states from 2005 to 2017, and the secondary analysis included inpatient visits for asthma by payer-type (i.e., private, Medicare, Medicaid, uninsured).

Principal findings: States with medical cannabis dispensaries and legalized recreational cannabis experienced 14.12% (2.14; 95% CI, 0.74-3.53; p < 0.01) and 20.45% (3.08; 95% CI, 1.47-4.69; p < 0.001) increases in inpatient visits for asthma compared with states without these policies, respectively. These increases in inpatient visits for asthma were primarily driven by populations covered by Medicare and private insurance, with Medicare population showing larger effects of both recreational cannabis laws and medical cannabis dispensaries.

Conclusions: States with medical cannabis dispensaries and legalized recreational cannabis experienced higher rate of inpatient visits for asthma compared with states without these policies. Clinicians and policymakers should consider strategies to curb adverse health outcomes of cannabis, that is likely to result in increased costs of healthcare.

Objective: To test whether enrolling in traditional Medicare (TM) or Medicare Advantage (MA) at age 65 reduces mental healthcare utilization among individuals with mental health symptoms and low or moderate family incomes.

Study setting and design: We employ a fuzzy regression discontinuity design, comparing the likelihood of having an outpatient mental health visit or a psychotropic drug fill among individuals younger than or older than the age 65 Medicare eligibility threshold.

Data sources and analytic sample: We analyze 2014-2021 Medical Expenditure Panel Survey data. Our primary sample is restricted to individuals with probable mental health symptoms as indicated by their score on the Kessler K6 psychological distress scale (K6) and Patient Health Questionnaire-2 instrument (PHQ-2) and who have incomes less than 400% of the federal poverty level.

Principal findings: Among individuals with probable mental health symptoms and low or moderate incomes, enrolling in Medicare (combining the effect of MA and TM) is associated with a 24.9 percentage point reduction (95% CI -49.1 to -0.8; p = 0.043) in the likelihood of having any type of outpatient mental health visit and a 31.3 percentage point reduction (95% CI -54.2 to -8.4; p = 0.008) in the likelihood of having a prescription drug fill for a psychotropic drug. Effects of MA and TM on mental healthcare utilization are not statistically different from each other. We observe no impact of enrolling in Medicare on the likelihood of having a visit to a primary care provider, having a visit to a non-mental healthcare specialist, or having a fill for a prescribed non-psychotropic drug.

Conclusions: Enrolling in Medicare is associated with a reduction in the use of mental healthcare among individuals with probable mental health symptoms and low or moderate family incomes. Our findings suggest that the program poses access barriers specific to mental healthcare.