Journal of General Internal Medicine最新文献

Background: People with HIV (PWH), particularly women, have a high cardiovascular disease (CVD) burden compared to the general population. There is little evidence describing statin adherence among PWH, which could inform interventions to reduce CVD disparities.

Objective: Observational cohort of privately insured PWH under age 65 who initiated statin therapy during 2015-2022 in MarketScan data.

Main measures: We used outpatient pharmacy claims to examine (1) statin discontinuation, defined as a gap > 90 days, and (2) proportion of days covered (PDC) by a statin in 90-day intervals. We estimated hazard ratios (HRs) using Cox models to compare discontinuation rates and prevalence ratios (PRs) from log-binomial regression to compare the probability of having low adherence (PDC < 80%) between women and men. We adjusted for potential confounding by demographic and clinical factors and accounted for repeated PDC measures.

Key results: We included 9522 PWH who initiated a statin (median age 52 years, 17.3% women). Overall, 50.0% of PWH had statin discontinuation within 2 years, and 26.5% had low adherence during statin therapy. Within 2 years, 59.0% of women and 48.1% of men had experienced discontinuation (adjusted HR 1.30 (95% CI, 1.20-1.41)). Among PWH remaining on statins, low adherence (PDC < 80%) was more common among women (34.1%) than men (25.2%) (adjusted PR 1.29 (1.22-1.35)).

Conclusions: PWH had high statin discontinuation rates, and a quarter had low adherence. Compared to men, women were both more likely to discontinue and have lower adherence. Efforts are needed to address statin adherence barriers to prevent CVD in PWH, especially in women with HIV.

Background/aim: Celiac disease (CD) is frequently underdiagnosed in adults due to the absence of classic gastrointestinal symptoms, leading to delayed diagnosis and potential long-term complications. This study aimed to identify subtle clinical and laboratory indicators preceding CD diagnosis in young adults (18-40 years).

Methodology: A retrospective cohort study was conducted using data from over 430,000 members of Clalit Health Services. Incident CD cases were identified by an ICD-9 code and a positive serological test. Laboratory values were analyzed for trends up to 7 years prior to diagnosis, and "relative" abnormalities were defined based on medians of diagnosed CD patients. Cox proportional hazards regression was used to assess the association between various pre-diagnostic clinical, laboratory, factors and the risk of CD diagnosis for 10 years.

Results: Relative lower hemoglobin and MCV in males and females were associated with CD with HR of 2.15 (1.43-3.22) and 5,519 (4.41-7.41) respectively. Liver enzymes (ALT more than 22 U/L in females and 27 U/L in males), and lower BMI even within the normal range, were significantly associated with an increased risk of later CD diagnosis with HR of 1.881 (1.29-2.75) for women and 2.919 (1.81-4.71) for males. These subtle laboratory and BMI findings were observable as early as 7 years before formal diagnosis. Additionally, chronic gastrointestinal symptoms and chromosomal risk factors significantly preceded CD diagnosis.

Conclusion: Early recognition of subtle indicators like low normal hemoglobin, elevated liver enzymes, and lower BMI, alongside persistent GI symptoms and genetic predispositions, can enable timelier screening and diagnosis of CD in young adults, potentially mitigating long-term health consequences.

Background: Advance directive completion and collection is a complex process including collecting documents, checking them for accuracy, and uploading them into a hospital or health system's electronic health record (EHR). Hospitals and health systems need mechanisms to facilitate the process.

Objective: To use the EHR patient portal and Health Informatics & Information Management System (HIIMS) personnel to collect and check the quality of advance care planning (ACP) documents and surrogate decision makers, and integrate them into the EHR.

Design: Single arm time series quality improvement project.

Intervention: Health IT, HIIMS, and the Advance Care Planning Program collaboratively developed a mechanism for ACP documents and surrogates submitted through the patient portal to be evaluated by HIIMS personnel who received training to process information entered by patients. HIIMS personnel integrate documents and surrogate information into the EHR or return a rejection message to the patient.

Main measures: Number of documents submitted, accepted, rejected, and accepted after resubmission, and the number of surrogates entered, updated, and removed.

Key results: Over 41 months from May 2021 to September 2024, 7274 ACP documents were uploaded via the patient portal (mean 177 documents per month): 4700 (65%) accepted on first submission and 2574 (35%) rejected. Of rejected documents, 397 (15%) were resubmitted and accepted within the month. Over time, the proportion of accepted documents increased. During the study period, 12,509 surrogates were added (mean 305 per month), 1790 were updated, and 911 were removed.

Conclusion: Engaging HIIMS personnel can facilitate ACP document capture through a common EHR portal, including quality checks and patient feedback.

Clinical reasoning is a fundamental aspect of providing effective and safe clinical care, yet it remains a challenging concept to teach in the clinical setting. Drawing parallels to Sherlock Holmes' deductive reasoning, this article explores how Holmes' techniques can make clinical reasoning more tangible for learners. Unlike algorithmic steps, clinical reasoning involves complex cognitive processes. By highlighting key concepts through Sherlockian maxims, this article aims to enhance the understanding and teaching of clinical reasoning.

The Association of Chiefs and Leaders of General Internal Medicine (ACLGIM), a group of division chiefs and other division leaders within the Society for General Internal Medicine (SGIM), convened the Hess Initiative Work Group on Learner Experience in Primary Care in 2023 to (1) identify the gaps between the current state of resident continuity clinic and current standards for a high-functioning primary care setting, (2) review the evidence to identify solutions, and (3) develop a set of evidence-based recommendations. These recommendations align with the current Accreditation Council for Graduate Medical Education (ACGME) IM milestones, with the goal to improve the experience of resident continuity clinic and meet the needs of the three key stakeholders of GME: residents, patients, and the healthcare system. We recruited experts in internal medicine ambulatory training from across the country, representing a variety of US regions, membership in professional societies and organizations related to internal medicine, community and academic settings, and residency program sizes during national and regional meetings at SGIM between 2023 and 2025. Using standard frameworks of high-functioning primary care practice, we performed an extensive review of the literature to identify discrepancies between these standards and resident continuity clinics, then, using the expertise of the group, produced a set of recommendations, which were reviewed and endorsed by the SGIM Council in September 2025.

Background: Guidelines endorse specific life expectancy (LE) cutoffs for consideration of definitive management of localized prostate cancer. Previous work has shown that patients prefer numeric, patient-specific LE estimates in consultations but that physicians often do not communicate LE in such detail. The persistent overtreatment of men with limited longevity despite these guidelines may in part be related to physician-mediated factors.

Objective: To determine physician barriers to the integration of LE in prostate cancer management, confidence in LE estimates, preferred method for LE calculation, and mode of LE communication in consultations.

Design: Structured interview study.

Participants: Multidisciplinary cohort of 20 urologists, radiation oncologists, and medical oncologists who regularly counsel patients with prostate cancer.

Approach: Interviews were transcribed verbatim and coded using an open coding approach. Saliency analysis was used to describe emergent concepts.

Key results: Physician-identified barriers to integration of LE in consultations included concerns about patient receptiveness (75%), the accuracy of LE prediction tools (50%), and lack of knowledge about how to effectively communicate LE (30%). There was significant heterogeneity in the sources of information used to calculate LE, with 85% using methods not incorporating a quantitative assessment of health status (55% gestalt method, 30% life tables). Most physicians had low (35%) or moderate (60%) confidence in LE predictions, which they cite as due to high variability in estimates (65%) and the need for further validation of predictive models (20%). To improve confidence in estimates, physicians wished for validated prediction tools (45%), incorporation of more variables (45%), reporting of the variability of LE estimates (30%), and molecular/genetic biomarkers (25%).

Conclusions: Addressing physician concerns about patient receptiveness, lack of knowledge about effective LE communication, and lack of confidence in LE predictions may be essential to reducing overtreatment of men with limited LE.

Background: Medical aid in dying (MAID) is legal in an increasing number of U.S. jurisdictions, yet public understanding of its legal status remains limited. Prior studies often collapse incorrect and "don't know" responses, obscuring potentially distinct ideological and structural correlates.

Objective: To examine predictors of uncertainty and misinformation about MAID legality among U.S. adults.

Design: Cross-sectional survey.

Participants: We used a convenience sample of U.S. adults recruited through Cloud Research's Prime Panels (N = 3,222), oversampling racial and ethnic minoritized groups, older adults (≥ 60 years), and residents of MAID-legal states. Analyses focused on two subsamples restricted to respondents who were uncertain or incorrect about MAID legality: national-level knowledge (n = 2,278) and state level knowledge among residents of MAID-legal states (n = 1,711).

Main measures: Respondents indicated whether MAID was legal (a) anywhere in the US and (b) in their state. Knowledge was categorized as "correct," "incorrect" (misinformation), or "don't know." Logistic regressions identified predictors of misinformation versus not knowing at state and national levels.

Key results: In the national sample, misinformed respondents (vs those who answered "don't know") were more likely to believe MAID should not be legal compared to respondents who were unsure about legality (average marginal effects (AME) = 0.22; CI: 0.14, 0.29; p < 0.001). They were also more likely to say that MAID is morally acceptable compared to respondents unsure about its morality (AME = 0.11; CI: 0.06, 0.17; p < 0.001). In both subsamples, structural factors (lower educational attainment and greater financial insecurity) were associated with higher likelihood of uncertainty rather than misinformation.

Conclusions: Public knowledge about MAID legality reflects two distinct patterns: misinformation appears ideologically motivated, while uncertainty is linked to structural factors. Efforts to improve health literacy may reduce uncertainty, but misinformation may require value-aligned communication and trusted messengers to support informed decision-making.