Journal of General Internal Medicine最新文献

Background: Buprenorphine is the most commonly used medication for treating opioid use disorder (OUD). It may also be beneficial for individuals with chronic pain who are physically dependent on opioids but do not meet the criteria for OUD. Therefore, the buprenorphine patient population may be heterogeneous in terms of OUD and comorbidities.

Objectives: To characterize the heterogeneity of patient profiles among those initiating buprenorphine treatment in a large multi-state healthcare system, focusing on comorbid pain, psychiatric, and substance use disorder diagnoses, and other patient characteristics associated with profiles.

Design: Retrospective cohort study using electronic health record data.

Partcipipants: Patients with at least 12 months of data before a new buprenorphine prescription (OUD formulation) and at least 6 months of follow-up.

Main measures: Presence vs absence of opioid use disorder diagnoses and latent class profiles of comorbid diagnoses and clinical characteristics. Post-index outcomes, including the number of buprenorphine orders and incidence of drug-related poisoning, were assessed for each class.

Key results: In a cohort of 5726 individuals (mean age 38.6 years, 56.2% female, 84.0% white), four latent classes were identified: (1) high multimorbidity burden; (2) moderate multimorbidity burden; (3) high rates of comorbid pain diagnoses but lower rates of psychiatric and substance use disorder diagnoses, with most patients lacking opioid use disorder diagnoses; (4) lower rates of pain diagnoses but high rates of comorbid psychiatric and substance use disorder diagnoses. Post-index measures aligned well with comorbidity profiles.

Conclusion: Patients receiving buprenorphine in general healthcare settings likely form a heterogeneous group. This includes a subgroup with low rates of opioid use disorder diagnoses (class 2, pain comorbidity), who may resemble chronic pain patients more than conventional OUD patients.

Clinical documentation is a cornerstone of physician training, not only as a record of care but as a catalyst for clinical reasoning. For medical trainees, writing notes compels them to prioritize information, justify decisions, and synthesize fragmented data into coherent narratives. With the emergence of artificial intelligence (AI) scribes that auto-generate clinical notes from ambient audio, the task of composing notes is increasingly outsourced, raising questions about its impact on education. At our internal medicine residency program, we piloted a 6-month implementation of an AI scribe tool with 48 residents, generating nearly 1000 notes. We propose seven best practices-mapped to the Accreditation Council for Graduate Medical Education (ACGME) Core Competencies-with the goal that AI scribes support, rather than erode, the development of reflective practice and diagnostic thinking. These include establishing baseline documentation skills, structured AI training, critical review of AI-generated notes, and new opportunities for feedback. In this formative moment, educators must guide learners to use AI as a scaffold for reasoning, not a substitute for it.

Artificial intelligence (AI) decision support tools (DSTs) are increasingly used across clinical settings to improve efficiency and support decision-making. However, these tools risk perpetuating existing healthcare disparities if not designed and implemented with transparency, equity, and cultural sensitivity. This review explores how racial and ethnic biases manifest within AI-driven DSTs and evaluates the role of governance frameworks in mitigating such harms. It examines the implications of biased algorithms, presents case examples highlighting disparities in tool performance, and critically assesses the adequacy of current national and international regulatory guidance. The review reports that bias can stem from unrepresentative training datasets, exclusion of equity auditing in design, and the absence of mandated transparency in reporting. Although several frameworks exist to guide development and reporting, few are mandatory, and most do not include equity as a core criterion. The current UK and US regulatory models are decentralized and lack mechanisms to systematically detect or prevent bias. To prevent biased tools from entering practice, equity must be structurally embedded across the AI lifecycle. Embedding equity into AI tools requires standardized subgroup performance reporting, mandating fairness assessments, and establishing national and global governance standards to ensure AI tools serve all populations equitably.

Background: Clinical champions are a known strategy for implementing evidence-based practices; however, their application in de-implementing potentially inappropriate medications (PIMs) among persons with dementia is underexplored. We conducted a pragmatic cluster-randomized clinical trial of a champion-led deprescribing intervention in accountable care organization (ACO) primary care settings.

Objective: To (1) understand clinical champions' perspectives of their deprescribing projects and (2) identify key contextual factors that influenced champions' deprescribing PIMs for persons with dementia within ACO primary care.

Design: A qualitative study guided by the de-implementation outcomes framework.

Participants: Clinician champions who engaged in deprescribing projects.

Approach: Data sources included transcripts from monthly learning calls and semi-structured interviews. All champions were invited to participate in learning calls and semi-structured interviews. The 30- to 60-min interviews were conducted using a semi-structured guide. We coded transcripts and performed thematic analysis to identify overarching themes.

Key results: Eleven champions participated. Feasibility and fidelity of deprescribing were commonly undermined by external and organizational disruptions. Five contextual factors influenced champions' experiences: limited organizational readiness, lack of information technology infrastructure to support data access and patient identification, the importance of relationship-building and care coordination, the dyadic nature of deprescribing involving care partners, and the pharmacist's role as a multidisciplinary liaison. Champions employed adaptive, communication- and relationship-centered strategies to support deprescribing efforts.

Conclusions: Champion-led deprescribing for persons with dementia is shaped by key contextual factors within ACO primary care settings. Dementia-specific training helped clinicians tailor deprescribing to local needs, but sustained efforts require supportive organizational structures, including ongoing education, accessible clinical data, and multidisciplinary collaboration. Primary care clinicians are uniquely positioned to lead personalized deprescribing conversations while pharmacists serve as liaisons with providers, patients, and care partners to coordinate deprescribing. Interventions that align with value-based care principles may strengthen system-level coordination and promote safer medication management in dementia care.

Trial registration: ClinicalTrials.gov ID: NCT05359679.

Introduction: Patients admitted for hip fracture surgery may receive care from a team led either by the operating surgeon or a hospitalist.

Objective: To describe the prevalence of the hospitalist care model for hip fracture admissions and its association with patient outcomes.

Design: We conducted a retrospective cohort study of patients admitted in 2018-2019 for fracture of the head and neck of the femur (ICD10 S72.0x-2x). We compared outcomes at hospitals with low versus high use of the hospitalist care model. Our exposure was the hospital-level adoption of hospitalist care, categorized into quartiles.

Participants: Fee-for-service Medicare patients 66 years old or greater.

Main measures: Length of stay (LOS), professional services (Part B) inpatient spending, specialty consultation, discharge to home, all-cause 7- and 30-day readmissions, and 30-day mortality.

Key results: A total of 294,150 patients with hip fracture were admitted to 2466 hospitals. Patients cared for in low-use (Q1) versus high-use (Q4) hospitals did not differ meaningfully in demographic characteristics or comorbidities. Hospitals ranged in use of the hospitalist care model from 12% in low-use (Q1) hospitals to 81% in high-use (Q4) hospitals. Low-use hospitals had significantly higher inpatient consult use (unadjusted: Q1 vs Q4, 1.06 vs 0.63 consults, p < 0.0001; adjusted: -0.36, p < 0.001) and length of stay (unadjusted: Q1 vs Q4, 6.04 vs 5.94 days, p < 0.0001; adjusted: -0.09 days, p < 0.05), but no significant difference in adjusted analyses for spending, likelihood of discharge home, 7- and 30-day readmission, or 30-day mortality.

Conclusions: Hospitalist care for older adults admitted for hip fracture surgery is both common and associated with slightly shorter length of stay.