Josiah D. Rich

Background and aims: Artificial intelligence (AI) may improve addiction medicine workflows and health outcomes but raises questions about data privacy and unintended harms for people who use drugs (PWUD). This qualitative study aimed to understand the perspectives of clinicians and PWUD regarding perceived benefits and concerns about clinical AI in addiction medicine. Design: A qualitative study with one‐on‐one semi‐structured interviews about AI. Setting: Addiction medicine clinicians in Rhode Island ( n = 7), Massachusetts ( n = 2) and Pennsylvania ( n = 3), USA. PWUD ( n = 25) residing in Rhode Island. Participants: In 2025, a purposive sample of 12 clinicians and 25 PWUD participated. Eligible clinicians were English‐speaking, ≥18 years old, licensed and provided care to PWUD exposed to xylazine in Rhode Island, Massachusetts or Pennsylvania. PWUD recruited from four Rhode Island social service organizations were eligible if they reported past‐month illicit fentanyl and xylazine use, lived in Rhode Island or Massachusetts and were English‐speaking and ≥18 years old. Eight (67%) clinicians were women and 10 (83%) were white non‐Hispanic. Fifteen (60%) PWUD were men and 17 (68%) were white non‐Hispanic. Measurements: Transcribed audio‐recorded interviews exploring benefits and concerns about clinical AI in addiction medicine. Clinician interviews were analyzed using rapid qualitative analysis to assess the breadth of clinical experience, whereas PWUD interviews were analyzed following inductive thematic analysis to assess depth of lived experience. Five themes were produced. Results: Five themes were generated: (1) transparently developed and accurate clinical AI tools offer practical benefits; (2) clinician comfort with clinical AI hinges on transparent development and assurances against the perpetuation of drug‐related stigma; (3) clinical AI may not keep pace with the volatile and regionally specific drug supply; (4) PWUD have limited AI knowledge and some characterized AI as untrustworthy; and (5) PWUD are altruistically motivated to release medical data for clinical AI development but want control over identifiable data usage. Summary: Clinicians and patients feel that artificial intelligence in addiction medicine holds promise, but they are concerned about clinical relevance, stigma perpetuation and the privacy of potentially incriminating data about illicit drug use.

Background: Law enforcement is a common yet complicated presence in the emergency department (ED). Studies have found emergency medicine (EM) attending physicians have varied interactions with and attitudes toward law enforcement. EM resident physicians are in a uniquely formative professional period. However, the experiences and perspectives of EM resident physicians have not been previously explored in depth. Methods: We recruited EM resident physicians from three geopolitically distinct urban academic EDs in the United States to participate in semi‐structured interviews regarding their experiences with law enforcement in the ED. We analyzed the transcripts thematically, looking for unifying concepts across institutions and latent meaning. A codebook was created and iteratively refined based on review of transcripts, after which salient themes were developed through ongoing discussions with the research team. Results: Twenty‐four resident physicians across three sites participated. Three salient themes emerged from our study: ambiguity of authority between emergency physicians and law enforcement, concerns about the impact of law enforcement presence on patients, and disruptions to the ED team workflow. First, as trainees learning professional norms in the ED, participants reported difficulty navigating situations in which law enforcement attempted to assert authority over patient care. Second, participants perceived negative impacts of law enforcement presence on patients in custody and non‐incarcerated patients of color receiving care in the ED. Finally, participants described frequent disruptions to the ED team workflow due to the presence of law enforcement. Participants relied on hospital security, rather than law enforcement, to ensure the safety of patients and staff. Conclusion: Emergency resident physicians described ambiguous boundaries and complex interactions among law enforcement, patients, and themselves in the ED. Although our study found several concerning impacts of law enforcement in the ED, these represent opportunities to clarify roles and responsibilities in the emergency physician‐law enforcement relationship and improve patient care. Further research is needed, including eliciting perspectives of law enforcement and patients, particularly those with substance use disorders, trauma, and serious mental illness.

Older adults comprise the fastest growing demographic in correctional facilities in the United States. Studies have shown older adults in prison report difficulty accessing healthcare and navigating the prison environment. Little is known about how older adults navigate aging and access healthcare in the jail setting. We sought to explore experiences of older adults incarcerated in a county jail. We conducted semi-structured qualitative interviews with 20 older adults (> 50 years old) who were incarcerated at a local county jail and were referred to the emergency department for acute care needs. Interviews focused on experiences of aging while incarcerated and access to healthcare. We used an inductive approach to code data and identify overarching themes. Participant demographics were representative of the broader jail population demographics. Two major themes emerged. First, participants reported difficulty navigating the jail environment related to aging and unaddressed disabilities. Examples included climbing to use the top bunk of a bunk bed, difficulty grasping objects due to a Parkinsonian tremor, and significant visual impairments. Participants relied on other inmates for assistance with limited systemic accommodations. Second, participants faced numerous challenges when attempting to access healthcare, including lack of clarity around procedures for accessing care and mutual distrust sometimes marking their relationship with medical staff. Participants described difficulty using a digital kiosk system to submit medical requests, and successful healthcare experiences involved an element of serendipity such as advocacy from a sympathetic officer. Despite a constitutional right to healthcare and high rates of medical comorbidities, older prisoners experienced a challenging physical environment in jail and difficulty accessing healthcare. Further work is needed to identify potential age-friendly modifications to the jail environment and improved healthcare delivery.

As health systems move to make electronic health records (EHRs) accessible to patients, there is a need to examine if, and the extent to which, bias toward patients may be evident in these records. This scoping review aimed to summarize the scientific literature on methods used to detect biased language about patients in the EHR and the nature and object of the biases detected. A comprehensive literature search was conducted of PubMed, CINAHL, Web of Science, APA PsycInfo, and SOCIndex for peer-reviewed English language studies conducted in the United States published on or before December 22, 2022. Seven studies were included in this review. Four methods were identified: natural language processing methods including machine learning-based (n = 3), Linguistic Inquiry and Word Count (n = 2), and exploratory vocabulary analysis (n = 1), and manual content analysis (n = 2). In four studies, the EHR of Black patients contained significantly greater bias relative to the EHR of White patients. Bias about health conditions (i.e., diabetes, substance use disorder, and chronic pain), women, and preexposure prophylaxis-a medication that prevents HIV infection-were identified. Machine-based learning methods may be best to (a) analyze robust data sampling frames, (b) detect a rare outcome like bias, (c) facilitate inferential analysis, and (d) transcend limitations of manual content analysis. Findings provide an overview of methods that can be used by investigators to analyze EHR records for bias to inform clinical interventions, health policies, and procedures to reduce bias among health care providers.