Development of a Preliminary Patient Safety Classification System for Generative AI.

Generative artificial intelligence (AI) technologies have the potential to revolutionise healthcare delivery but require classification and monitoring of patient safety risks. To address this need, we developed and evaluated a preliminary classification system for categorising generative AI patient safety errors. Our classification system is organised around two AI system stages (input and output) with specific error types by stage.

Artificial intelligence related safety issues associated with FDA medical device reports.

The Biden 2023 Artificial Intelligence (AI) Executive Order calls for the creation of a patient safety program. Patient safety reports are a natural starting point for identifying issues. We examined the feasibility of this approach by analyzing reports associated with AI/Machine Learning (ML)-enabled medical devices.

Using Z Codes to Document Social Risk Factors in the Electronic Health Record: A Scoping Review.

Introduction: Individual-level social risk factors have a significant impact on health. Social risks can be documented in the electronic health record using ICD-10 diagnosis codes (the "Z codes"). This study aims to summarize the literature on using Z codes to document social risks.

Inbox message prioritization and management approaches in primary care.

Objectives: Patient messaging to clinicians has dramatically increased since the pandemic, leading to informatics efforts to categorize incoming messages. We examined how message prioritization (as distinct from categorization) occurs in primary care, and how primary care clinicians managed their inbox workflows.

Materials And Methods: Semi-structured interviews and inbox work observations with 11 primary care clinicians at MedStar Health.

Does one size fit all? Developing an evaluation strategy to assess large language models for patient safety event report analysis.

Objective: Collecting and analyzing patient safety event (PSE) reports is a key component to the improvement of patient safety yet report analysis has been challenging. Large language models (LLMs) may support analysis; however, PSE reports tend to be a hybrid of clinical and general language.

Materials And Methods: We propose a data-driven evaluation strategy to assess LLM fit for report analysis.

Using Patient Experience Surveys to Identify Potential Diagnostic Safety Breakdowns: A Mixed Methods Study.

Objectives: One in 20 outpatients in the United States experiences a diagnostic error each year, but there are no validated methods for collecting feedback from patients on diagnostic safety. We examined patient experience surveys to determine whether patients' free text comments indicated diagnostic breakdowns. Our objective was to evaluate associations between patient-perceived diagnostic breakdowns reported in free text comments and patients' responses to structured survey questions.

"What Else Could It Be?" A Scoping Review of Questions for Patients to Ask Throughout the Diagnostic Process.

Background: Diagnostic errors are a global patient safety challenge. Over 75% of diagnostic errors in ambulatory care result from breakdowns in patient-clinician communication. Encouraging patients to speak up and ask questions has been recommended as one strategy to mitigate these failures.

Falling short in bariatric surgery: An exploration of key barriers and motivators of attrition.

Background: In the United States, obesity-related diseases pose significant healthcare challenges, with bariatric surgery offering a potential solution. However, bariatric surgery completion rates, particularly among Black and Hispanic populations, remain low.

Objective: This study applied the Theoretical Domains Framework (TDF) to explore behavioral factors influencing bariatric surgery program attrition among a majority Black participant population to inform interventions for improving attrition.

Understanding Perceptions of Care Coordination and Chronic Illness Management among Black Breast and Prostate Cancer Survivors and Providers: Findings from a Quality Improvement Study.

Navigating cancer care is complex and is exacerbated by pre-existing comorbidities managed by multiple providers. In this quality improvement study, we evaluated changes in perceived care coordination, navigation, and chronic illness care with community health worker (CHW) and mHealth support among Black breast cancer and prostate cancer patients with hypertension and/or diabetes. We collected patient and provider surveys on chronic illness care coordination at baseline and six months and found improvements in multiple domains.

Accounting for taste: preferences mediate the relationship between documentation time and ambulatory physician burnout.

Objectives: Physician burnout in the US has reached crisis levels, with one source identified as extensive after-hours documentation work in the electronic health record (EHR). Evidence has illustrated that physician preferences for after-hours work vary, such that after-hours work may not be universally burdensome. Our objectives were to analyze variation in preferences for after-hours documentation and assess if preferences mediate the relationship between after-hours documentation time and burnout.

Barriers and facilitators to screen for and address social needs in primary care practices in Maryland: a qualitative study.

Background: Social needs screening can help modify care delivery to meet patient needs and address non-medical barriers to optimal health. However, there is a need to understand how factors that exist at multiple levels of the healthcare ecosystem influence the collection of these data in primary care settings.

Methods: We conducted 20 semi-structured interviews involving healthcare providers and primary care clinic staff who represented 16 primary care practices.

Consistency is key: documentation distribution and efficiency in primary care.

Objectives: We analyzed the degree to which daily documentation patterns in primary care varied and whether specific patterns, consistency over time, and deviations from clinicians' usual patterns were associated with note-writing efficiency.

Materials And Methods: We used electronic health record (EHR) active use data from the Oracle Cerner Advance platform capturing hourly active documentation time for 498 physicians and advance practice clinicians (eg, nurse practitioners) for 65 152 clinic days. We used k-means clustering to identify distinct daily patterns of active documentation time and analyzed the relationship between these patterns and active documentation time per note.

Development and Evaluation of Patient Safety Interventions: Perspectives of Operational Safety Leaders and Patient Safety Organizations.

Objectives: The purpose of this study is to understand how patient safety professionals from healthcare facilities and patient safety organizations develop patient safety interventions and the resources used to support intervention development.

Methods: Semistructured interviews were conducted with patient safety professionals at nine healthcare facilities and nine patient safety organizations. Interview data were qualitatively analyzed, and findings were organized by the following: patient safety solutions and interventions, use of external databases, and evaluation of patient safety solutions.

Understanding social needs screening and demographic data collection in primary care practices serving Maryland Medicare patients.

Background: Health outcomes are strongly impacted by social determinants of health, including social risk factors and patient demographics, due to structural inequities and discrimination. Primary care is viewed as a potential medical setting to assess and address individual health-related social needs and to collect detailed patient demographics to assess and advance health equity, but limited literature evaluates such processes.

Methods: We conducted an analysis of cross-sectional survey data collected from n = 507 Maryland Primary Care Program (MDPCP) practices through Care Transformation Requirements (CTR) reporting in 2022.

Interoperability in the Wild: Comparison of Real-World Electronic C-CDA Documents from Two Sources.

Although health information exchange (HIE) networks exist in multiple nations, providers still require access multiple sources to obtain medical records. We sought to measure and compare differences in data presence and concordance across regional HIE and EHR vendor-based networks. Using 1,054 randomly selected patients from a large health system in the US, we generated consolidated clinical document architecture (C-CDA) documents from each network.

Guidance for reporting analyses of metadata on electronic health record use.

Introduction: Research on how people interact with electronic health records (EHRs) increasingly involves the analysis of metadata on EHR use. These metadata can be recorded unobtrusively and capture EHR use at a scale unattainable through direct observation or self-reports. However, there is substantial variation in how metadata on EHR use are recorded, analyzed and described, limiting understanding, replication, and synthesis across studies.

Pharmacogenomic testing in oncology: a health system's approach to identify oncology provider perspectives.

Identify oncology healthcare providers' attitudes toward barriers to and use cases for pharmacogenomic (PGx) testing and implications for prescribing anticancer and supportive care medications. A questionnaire was designed and disseminated to 71 practicing oncology providers across the MedStar Health System. 25 of 70 (36%) eligible oncology providers were included.

A text mining approach to categorize patient safety event reports by medication error type.

Patient safety reporting systems give healthcare provider staff the ability to report medication related safety events and errors; however, many of these reports go unanalyzed and safety hazards go undetected. The objective of this study is to examine whether natural language processing can be used to better categorize medication related patient safety event reports. 3,861 medication related patient safety event reports that were previously annotated using a consolidated medication error taxonomy were used to develop three models using the following algorithms: (1) logistic regression, (2) elastic net, and (3) XGBoost.