Journal of General Internal Medicine最新文献

Background and objective: Polypharmacy is common in older adults due to multimorbidity and is associated with frailty, falls, decreased function, and mortality. However, the association between older age and polypharmacy has never been studied in the context of limited English proficiency (LEP). We investigated whether older adults with LEP were more likely to experience polypharmacy than older adults who are English proficient.

Design, participants, and main measures: We conducted a cross-sectional analysis of pooled 2013-2018 Medical Expenditure Panel Survey (MEPS), a nationally representative US household survey. Participants were community-dwelling adults aged 65 and older. Polypharmacy was defined as using five or more prescription medications. LEP was defined as participants who reported speaking English "not well" or "not at all" and spoke a non-English language at home. We used multivariable logistic regression models adjusted for demographic, socioeconomic, and health characteristics.

Key results: There were 27,697 MEPS respondents representing 50.5 million community-dwelling older adults. In total, 57.7% of participants were 65-74 years old; 55.2% female; 76.4% white; and 9.8% living below the federal poverty line. The median number of chronic conditions was 3 (IQR, 2-5). A total of 57.8% (95% CI, 56.7-58.9%) used five or more medications, and 4.6% (95% CI, 4.1-5.2%) had LEP. In unadjusted analyses, polypharmacy was less common among older adults with LEP than English-proficient adults (51.7% vs. 58.1%; p < 0.001), which remained significant after adjusting for demographic, socioeconomic, and health characteristics (aOR = 0.79, 95% CI 0.67-0.97; p = 0.03).

Conclusions: In a representative sample of community-dwelling older adults, older adults with LEP were less likely to experience polypharmacy compared to older adults who are English proficient. While polypharmacy is associated with safety concerns, lower prevalence among individuals with LEP may reflect barriers to care rather than better prescribing practices, underscoring the need for language-concordant interventions to improve medication adherence and accessibility while avoiding inappropriate medications.

Importance: Alcohol use disorder (AUD) has been associated with reduced adherence to and discontinuation of HIV preexposure prophylaxis (PrEP), potentially compromising its effectiveness.

Objective: This study examines the relationship between AUD and PrEP adherence and continuation.

Design: Retrospective cohort study using MarketScan Commercial Claims data.

Participants: We included individuals aged 16-64 who initiated PrEP between January 1, 2014, and December 31, 2021, and had continuous insurance coverage.

Main measures: AUD diagnosis was identified within six months before PrEP initiation. Adherence was measured using the proportion of days covered (PDC) over 180 days and categorized by clinically relevant thresholds (≥ 85%, ≥ 80%, ≥ 57%). Continuation was assessed based on uninterrupted PrEP supply without a ≥ 30-day gap. Statistical analyses included propensity-score matching and regression modeling.

Key results: Among 43,913 eligible individuals, 1,245 (2.84%) had an AUD diagnosis prior to PrEP initiation. In a matched sample of 1,153 individuals, those with AUD had lower mean PDC (59.54% [34.14] vs. 65.85% [33.18]; p < 0.001) and fewer mean days of continuous PrEP use (107 [67.20] days vs. 119.90 [65.21] days; p < 0.001) compared to the Without AUD group. Regression analyses showed individuals with AUD had 6.31% lower mean PDC (95% CI: -9.05% to -3.57%; p < 0.001) and 12.93 fewer days of PrEP continuity (95% CI: -18.34 to -7.52 days; p < 0.001) compared to the Without AUD group. Findings may not be generalizable beyond commercially insured individuals, and we may have not captured all factors that influence PrEP adherence.

Conclusion: Individuals with AUD before PrEP initiation exhibited lower adherence and continuation compared to those without AUD. Targeted interventions may be required to enhance PrEP adherence and continuity in this population.

Background: Ohio ranks among the highest US states for cardiovascular disease (CVD) morbidity and mortality. Although interventions exist for managing CVD risk factors, adoption in primary care is often limited. The Agency for Healthcare Research and Quality funded four states to develop scalable, statewide models for implementing evidence-based practices to address these gaps.

Objective: To evaluate the effectiveness of the Heart Healthy Ohio Initiative (HHOI), a statewide quality improvement (QI) initiative focused primarily on improving blood pressure (BP) control DESIGN: Pre-post, repeated cross-sectional QI study using electronic health record (EHR) data to compare patient outcomes 6 months pre- and post-intervention PARTICIPANTS: A total of 293,638 adult patients (aged ≥ 18 years) from 48 primary care clinics across 21 counties, of whom 107,216 (37%) had hypertension.

Interventions: Practices received structured QI support to implement evidence-based strategies for hypertension management, including BP measurement, timely follow-up, treatment protocols, and outreach. Monthly QI coaching, peer learning, and data feedback supported implementation. Smoking cessation strategies were also encouraged.

Main measures: The primary outcome was BP control, defined as < 140/90 mmHg. Process measures included repeat BP measurement, timely follow-up, and medication intensification. Regression analyses evaluated the impact of process measures on BP control improvement. Secondary measures on smoking cessation included screening, quit advice, resource referrals, and medications prescribed.

Key results: BP control improved from 67.7% to 70.7% post-intervention. Greater improvements were observed among rural and uninsured patients (> 6%) compared to smaller gains among Medicaid enrollees, younger patients, and women (~ 2%). All three process measures were significantly associated with better BP control in multivariable models. Smoking cessation measures were maintained or declined by ~ 2%, although only five sites actively addressed smoking cessation.

Conclusions: HHOI demonstrates the feasibility and early success of a statewide, cooperative QI infrastructure to improve BP control. This model may be replicable in other states and offers insights for addressing hypertension control through targeted, scalable strategies.

Background: Diagnostic errors are a large source of harm among hospitalized patients. Understanding how medical team composition influences diagnostic safety is needed to design care models that support accurate and effective diagnostic processes.

Objective: To evaluate whether direct attending care, resident teaching teams, or advanced practice provider-led (APP) services are associated with differences in the risk of diagnostic errors.

Design: Retrospective observational study using average treatment effect and inverse probability of treatment weighting to estimate marginal rate ratios (mRRs) for diagnostic errors across care models. Diagnostic errors were previously adjudicated by two-physician chart review methods.

Participants: One thousand five hundred forty-four general medicine patients who transferred to the ICU or died in hospital between January 1 and December 31, 2019, at 29 hospitals across the USA.

Interventions: None.

Main measures: Primary outcomes were any diagnostic error and harmful diagnostic error. Any diagnostic error was defined as a missed opportunity for a timely and accurate diagnosis, regardless of whether harm occurred, whereas harmful diagnostic errors were limited to those resulting in temporary harm requiring intervention, permanent harm, or death.

Key results: Of 1544 patients, 969 (63%) were cared for by teaching teams, 442 (29%) by direct care, and 133 (9%) by APP services. Direct care was associated with a higher risk of any diagnostic error compared with teaching teams (mRR, 1.36; 95% CI, 1.04-1.68).

Conclusion: Care team structure may influence diagnostic error risk. Models involving residents, often working with attending physicians, may offer diagnostic safety advantages through team-based diagnostic decision-making and workload distribution.

Background: Like the US population, the Veteran population is aging rapidly and has increasing current and future needs for skilled nursing facilities (SNF).

Objectives: We examined trends in SNF use among Veterans after hospitalization before and after the 2018 MISSION Act and the onset of the COVID-19 pandemic. We assessed factors influencing SNF placement and outcomes across three SNF types.

Design: Retrospective cohort study of hospitalized Veterans with subsequent SNF stays from January 1, 2015, to December 31, 2022, using data from both VA and non-VA sources.

Participants: Veterans aged 66 + with an acute hospitalization followed by SNF admission within 1 day were included. We excluded long-term nursing home residents, psychiatric discharges, and users of other institutional post-acute care.

Main measures: We analyzed trends in SNF use over time, examined factors affecting SNF type, and compared outcomes across three settings: VA-operated SNFs (Community Living Centers, CLCs), VA-paid non-VA SNFs (Contract Nursing Homes, CNHs), and Medicare-paid non-VA SNFs.

Key results: Among 1,405,701 qualifying hospital-to-SNF transitions, 90.0% were to Medicare-paid non-VA SNFs, 5.8% to CLCs, and 4.2% to CNHs. CNH use increased after mid-2019, while overall SNF use declined sharply with the start of COVID-19 and remained below pre-pandemic levels through 2020. The hospital type was the strongest predictor of SNF destination: VA hospitals sent Veterans to VA SNFs, while non-VA hospitals used non-VA SNFs. CLCs had the highest 30-day readmission rates (22.1%) but the lowest 30-day mortality (4.3%). CNHs had the longest median stays (29 days; IQR 14-63) and the highest rates of stays over 100 days (16.2%).

Conclusions: Veteran SNF use shifted after the MISSION Act and COVID-19, with notable differences in utilization and outcomes by SNF type. These findings highlight the need to better understand how hospital and policy factors affect post-acute care access, quality, and cost for older Veterans.

Background: Workforce development programs aim to address the disparity in the number of primary care physicians (PCPs) practicing in rural U.S. areas. Both publicly and privately funded rural recruitment and retention programs have worked for decades to enhance healthcare access. However, little is known about their comparative effectiveness in retaining PCPs long term. This scoping review assessed the success of programs in retaining rural PCPs.

Methods: We searched PubMed, Embase, and Health Business Elite for peer-reviewed literature, published between 2013 and 2023, focusing on PCPs' rural recruitment and retention in the U.S. The gray literature search included sources like the Rural Health Information Hub and the Rural Medical Training Collaborative, followed by a Google search. Articles were screened by two authors, with discrepancies resolved by a third.

Results: From 2227 articles identified, only 10 met the inclusion criteria, with 7 additional programs found through gray literature, totaling 17 programs. Financial incentives, such as state loan repayment programs (n = 2) and scholarships (n = 2), showed the highest reported retention rates (50-100%) in rural areas; however, these results should be interpreted cautiously due to small sample sizes and substantial variability in follow-up periods across programs.

Discussion: Most rural PCP workforce development programs did not report retention outcomes. Among those that did, financial incentive programs had higher retention rates than rural education and residency programs. These findings are limited by the heterogeneity of results reported and the variability in program sample sizes. Additionally, many education programs reported the number of clinicians in rural areas but did not specify whether they were in primary care. Future program reports and research should standardize reporting to include the number of individuals who completed the program, their specialty, duration of rural practice post service time commitment, and current practice status in rural areas.

Background: There is a lack of diversity in the Internal Medicine (IM) workforce with respect to race, ethnicity, sex, and socioeconomic background. Evaluating sociodemographic predictors and student interest in IM residency may provide insights for support at the medical school level to promote diverse interest in IM.

Objective: To examine associations between student ethnoracial identity, sex, and socioeconomic background with IM interest from matriculation to graduation.

Design: Retrospective cohort study.

Participants: US MD matriculants from 2014-2015 and 2015-2016 cohorts.

Main measures: A sustained interest was defined as IM interest at matriculation and subsequent IM residency placement. A cultivated interest was defined as initial non-IM interest and placement into IM residency. Poisson regressions estimated associations between ethnoracial identity, sex, and family income with sustained or cultivated interest in IM, adjusting for IM clerkship satisfaction, USMLE Step 1 and Step 2 score.

Key results: Among 18,765 matriculants, 51.7% identified as female, 61.5% as White, and 24.3% as low-income. IM clerkship satisfaction was associated with IM residency placement (p < 0.001). Among students initially interested in IM, females (aRR 0.87, 95% CI [0.80-0.94]) were less likely to sustain interest. Among initially uninterested students, Asian (aRR 1.65, 95% CI [1.54-1.76]), Hispanic (aRR 1.31, 95% CI [1.16-1.48]), and Black (aRR 1.22, 95% CI [1.07-1.39]) students were more likely to report cultivated interest compared to White students, while females were less likely compared to male students (aRR 0.77, 95% CI [0.72-0.82]).

Conclusion: Disparities in IM workforce diversity may originate in medical school, where female students were less likely to sustain IM interest, while Asian, Black, and Hispanic students were more likely to develop interest. Such results necessitate targeted support by medical schools and graduate medical education programs.

Background: Opioid use disorder (OUD) is responsible for significant morbidity and mortality in the USA. Hospitalization rates for patients with OUD have increased over the recent decades. Those with OUD have a substantially higher rate of patient-directed discharge (PDD) than those without OUD. There have been mixed results when examining the association between inpatient MOUD and PDD.

Objective: To determine the association between inpatient MOUD and the rate of PDD among patients without evidence of MOUD treatment prior to hospitalization.

Design: Retrospective study comparing admissions receiving inpatient MOUD and propensity score-matched control admissions who did not receive MOUD.

Subjects: Two thousand seven hundred seventy-one admissions with a diagnosis of OUD and without evidence of prior MOUD treatment were compared to 2771 propensity-matched admissions.

Intervention: Provision of inpatient MOUD, either buprenorphine or methadone during admission.

Main measures: Primary outcome was patient-directed discharge. Secondary outcomes were buprenorphine prescription at discharge, buprenorphine prescription within 60 days of discharge, admission into an outpatient methadone program within 30 days of discharge, 30-day readmission, and 30-day post-discharge ED visit.

Key results: Among 5542 admissions with OUD and no evidence of MOUD prior to admission, those that received inpatient MOUD were significantly less likely to have a PDD (11.9% vs 14.4%; OR 0.80 [CI 0.67-0.96]) and significantly more likely to receive a discharge prescription for buprenorphine (8.6% vs 1.2%; OR 8.04 [CI 5.52-11.71]) and another buprenorphine prescription within 60 days of discharge (5.5% vs 1.1%; OR 5.09 [CI 3.35-7.74]), compared with control admissions who did not receive MOUD. Inpatient MOUD was not significantly associated with admission into an outpatient methadone program within 30 days, 30-day readmission, and 30-day post-discharge ED visit.

Conclusions: Receipt of inpatient MOUD was associated with a statistically significant reduction in PDD among those with OUD and without evidence of MOUD before admission when compared with propensity-matched admissions which did not receive inpatient MOUD.