BMJ Quality & Safety最新文献

{"title":"AI mirage: medical algorithms and the vanishing autopsy.","authors":"Qi Wang","doi":"10.1136/bmjqs-2026-020083","DOIUrl":"10.1136/bmjqs-2026-020083","url":null,"abstract":"","PeriodicalId":9077,"journal":{"name":"BMJ Quality & Safety","volume":" ","pages":"576-578"},"PeriodicalIF":6.8,"publicationDate":"2026-07-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147353793","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"When emergency physicians meet patients displaying irritable behaviours: a randomised vignette-based experiment investigating physicians' emotions and clinical reasoning.","authors":"Linda M Isbell, Nathan R Huff, Guanyu Liu, Howard Bessen, Vincent Kan, Isabella Levesque, Matthew Bird, Peter Smulowitz","doi":"10.1136/bmjqs-2025-019061","DOIUrl":"10.1136/bmjqs-2025-019061","url":null,"abstract":"<p><strong>Background: </strong>Clinical encounters with patients who display challenging behaviours (eg, rude, demanding) are increasingly common in emergency departments and can heighten negative emotions, compromise healthcare quality and impact physician well-being. This study investigated the influence of patients' irritable behaviours on physicians' emotions, clinical assessments, decisions and behaviours, and the extent to which physicians' predispositions to experience anxiety and concern about bad clinical outcomes (ie, stress from uncertainty) may magnify these effects. We also explored whether these effects differed if patients had a stigmatised condition (mental illness history).</p><p><strong>Methods: </strong>Experienced emergency physicians (n=134) completed a randomised vignette-based experiment between June and August 2022. Physicians assessed four patients in computer-simulated, interactive clinical encounters in which patient behaviour (irritable vs calm) and mental illness history (present vs absent) were varied within subjects. Physicians viewed each patient's health record, a video-recorded encounter and physical exam results, and provided patient assessments, clinical decisions, ordered tests and reported a diagnosis. Physicians also reported their emotional reactions to each patient and completed the Stress from Uncertainty Scale (SUS) after assessing all patients. We adhered to the Guideline for Reporting Vignette Experiments.</p><p><strong>Results: </strong>Physicians reported greater anger, anxiety and fatigue and less empathy, happiness and engagement during encounters with patients displaying irritable (vs calm) behaviours (all ps<0.001) and evaluated those with irritable behaviours as poorer historians, more likely to be exaggerating pain, less cooperative, engaged, likeable, likely to adhere to treatment and likely to return to work (all ps<0.001). Interactions revealed that SUS amplified the impact of patients' irritable behaviours with high SUS associated with greater physician anger, anxiety and fatigue (all ps≤0.04) and with physicians evaluating these patients as poorer historians, less likely to adhere to treatment and less likeable (all ps≤0.03). Effects were similar regardless of patient mental illness history (all ps>0.60). No effects emerged for clinical decisions, behaviour or diagnosis.</p><p><strong>Conclusion: </strong>Although patients' irritable behaviours adversely influenced physicians' emotions, engagement and patient assessments, they did not affect the quality of care provided (ie, clinical decisions, behaviour, diagnostic accuracy). Despite this, our findings have important implications. Physicians' reactions to patients displaying irritable behaviours may reduce the quality of physician-patient relationships (eg, reduce trust, effective communication) and contribute to physician burnout-both of which can compromise patient care. Additional research is needed to investigate the multiple ways in","PeriodicalId":9077,"journal":{"name":"BMJ Quality & Safety","volume":" ","pages":""},"PeriodicalIF":6.8,"publicationDate":"2026-07-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13536421/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148429702","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Mapping Theories, Models and Frameworks from implementation science for evaluating quality improvement initiatives: a scoping review.","authors":"Klaudia A Arizmendi-Barrera, Jafet Arrieta, Tamara Nevarez Tinoco, Priyanka Chahal, Pierre Barker","doi":"10.1136/bmjqs-2025-019903","DOIUrl":"10.1136/bmjqs-2025-019903","url":null,"abstract":"<p><strong>Background: </strong>Improvement science has supported the methodological foundations for the application of quality improvement (QI) in healthcare. However, concerns persist regarding the rigour of QI evaluation. Theories, Models and Frameworks (TMFs) from implementation science support the uptake of evidence-based practices but their use to guide the evaluation of QI initiatives remains insufficiently characterised.</p><p><strong>Objectives: </strong>To identify which implementation science TMFs have been used to evaluate QI initiatives in healthcare, assess how these TMFs were applied in the evaluation, and compare the TMFs' intrinsic methodological attributes.</p><p><strong>Methods: </strong>We conducted a scoping review following Arksey and O'Malley's framework. Eligible studies were published between 2010 and 2025 and evaluated a QI initiative using at least one TMF. Peer-reviewed literature was searched in PubMed, Scopus and Web of Science. Grey literature was retrieved through targeted Google searches. TMFs were evaluated at the article level according to their application in each study and at the TMF level according to their intrinsic methodological attributes.</p><p><strong>Results: </strong>Out of 1824 records screened, 80 studies were included. 23 unique TMFs were identified. Of these, the most frequently used were the Reach, Effectiveness, Adoption, Implementation and Maintenance framework (27%) and the Consolidated Framework for Implementation Research (20%). Although 94% of studies used TMFs to guide data collection or analysis, only 43% explicitly tested construct relationships or causal pathways. TMFs demonstrated high conceptual clarity and structure but limited equity and sociocultural responsiveness.</p><p><strong>Conclusions: </strong>Application of TMFs in the evaluation of QI initiatives was heterogeneous. No single TMF captured the full range of evaluative functions. More rigorous evaluation requires an integrative approach that combines complementary TMFs to examine context, mechanisms of improvement and outcomes. Prospective operationalisation of TMF constructs, particularly equity, within study design and evaluation is essential to strengthen reproducibility and improve the transferability of findings across settings.</p>","PeriodicalId":9077,"journal":{"name":"BMJ Quality & Safety","volume":" ","pages":""},"PeriodicalIF":6.8,"publicationDate":"2026-07-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148315711","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Differences in physician task load are associated with gender variance in physician well-being metrics.","authors":"Elizabeth Harry, Julie Evans, Hanhan Wang, Jane M Liebschutz, Jennifer Berliner, Chen-Tan Lin, Mariah Quinn, Tait D Shanafelt","doi":"10.1136/bmjqs-2025-019297","DOIUrl":"https://doi.org/10.1136/bmjqs-2025-019297","url":null,"abstract":"<p><strong>Introduction: </strong>Physician burnout is related to patient safety and quality of care. Female physicians face unique challenges which may contribute to higher physician task load (PTL). We investigated the association of gender and PTL, and if gender differences in PTL contribute to burnout, professional fulfilment and intent to leave in medicine.</p><p><strong>Methods: </strong>A retrospective observational study using secondary data was conducted in June 2019-June 2021. Physicians at eight large US academic-affiliated medical centres completed a survey assessing PTL using four NASA-TLX items, burnout and professional fulfilment using the Professional Fulfilment Index (PFI), and intent to leave using a standardised single-item measure; demographic items were included. Descriptive statistics characterised the participant sample. χ² tests examined differences across participant characteristics and PTL, PFI and ITL by gender. Logistic regressions assessed associations between PTL and well-being variables. Analysis of indirect association explored the association of PTL and the relationship between gender and well-being variables.</p><p><strong>Results: </strong>Females had higher mean PTL as compared with males (6.67 (1.82) vs 6.45 (1.96), p<0.001). Higher levels of PTL were associated with lower PFI by PTL quintile (p<0.001), higher burnout by PTL quintile (p<0.001) and greater ITL in the highest PTL quintile (p<0.001). Despite higher burnout and PTL, females had lower odds of ITL than males. Analyses of indirect associations suggested PTL may partially account for the relationship between gender and PFI and burnout, with an opposite-direction (suppression) effect for ITL.</p><p><strong>Conclusions: </strong>Higher mean PTL scores were observed in females compared with males. Findings suggest that both gender-specific and gender-agnostic interventions for addressing cognitive burden may be relevant for supporting well-being for all physicians.</p>","PeriodicalId":9077,"journal":{"name":"BMJ Quality & Safety","volume":" ","pages":""},"PeriodicalIF":6.8,"publicationDate":"2026-07-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148403539","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Paradox of engagement: precautionary public health alerts and gender inequality in cancer diagnosis-a quasi-experimental study.","authors":"I-Hua Chen, Juifen Rachel Lu, Ya-Chen Tina Shih, Chang-Xi Kung, Chun-Ying Wu, Li-Lin Liang","doi":"10.1136/bmjqs-2026-020450","DOIUrl":"https://doi.org/10.1136/bmjqs-2026-020450","url":null,"abstract":"<p><strong>Background: </strong>Emergency policies designed to protect populations can inadvertently cause diagnostic harm when essential services are suspended. How such policies reshape gender gaps in cancer detection, and whether either sex faces disproportionate harm, remains underexplored.</p><p><strong>Objectives: </strong>We leveraged Taiwan's 2021 level 3 alert-a precautionary directive implemented during minimal COVID-19 transmission (0.07%) that mandated hospital capacity reductions and school closures-as a natural experiment to test whether the alert produced gendered diagnostic disruptions.</p><p><strong>Methods: </strong>Using Taiwan Cancer Registry and National Health Insurance claims (N=312 522; 2015-2021), we employed interrupted time series and difference-in-differences models to estimate changes in cancer detection rates and tumour size at diagnosis; robustness was confirmed through sensitivity and falsification tests.</p><p><strong>Results: </strong>The alert disproportionately reduced detection of early-stage cancers, consistent with a collapse of preventive and opportunistic detection pathways. Women experienced larger declines in detection than men (-43.9% vs -32.0%; p=0.001) and greater increases in tumour size (17.3% vs 11.5%; p=0.026). These gaps were most pronounced among women under 50 (-50.7% vs -25.5%; p=0.001) and disadvantaged women, whose tumour size increase was nearly double that of disadvantaged men (16.6% vs 8.8%; p=0.049). The gap was particularly large for opportunistically detected cancers (lung: -47.1% women vs -22.6% men; p=0.011).</p><p><strong>Conclusions: </strong>These findings reveal a system-level patient safety failure. Women's higher engagement with routine care, ordinarily a health advantage, became a structural vulnerability when these pathways were suspended, possibly amplified by intensified caregiving, employment shocks and infection-risk avoidance. As health systems prepare for recurring shocks, resilience planning must incorporate gender-sensitive safeguards that prevent diagnostic harm. Protecting early detection pathways is both a clinical imperative and a patient safety requirement for equitable emergency response.</p>","PeriodicalId":9077,"journal":{"name":"BMJ Quality & Safety","volume":" ","pages":""},"PeriodicalIF":6.8,"publicationDate":"2026-06-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148359411","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Priorities for improving paediatric diagnosis: findings from a modified Delphi study.","authors":"Irit R Rasooly, Katherine Wu, Meghan Galligan, Morgan Congdon, Anna Costello, Yasaman Fatemi, Courtney Gidengil, Joseph A Grubenhoff, Harry Hoar, Cara T Jefferies, Amarpreet Kaur, Jillian Krause, Prashant Mahajan, Sanjiv D Mehta, Jonathan G Sawicki, Trisha Marshall, Kenneth A Michelson, Ursula Nawab, Andrew P J Olson, Michael Perry, Grant Joseph Dat Chiu Shafer, Kathy N Shaw, Hardeep Singh, Kathleen Walsh, Christina L Cifra","doi":"10.1136/bmjqs-2025-020052","DOIUrl":"10.1136/bmjqs-2025-020052","url":null,"abstract":"<p><strong>Background: </strong>Efforts to improve diagnosis should reflect the unique needs of children; however, there is no consensus on paediatric-specific priority areas.</p><p><strong>Methods: </strong>An expert panel of 25 United States (US) paediatric diagnostic excellence researchers, patient safety leaders and family partners from 15 institutions participated in a modified Delphi panel. Panellists participated in generating a literature-derived list of topics relevant to improving diagnosis. Panellists then rated agreement on each topic as a research or operational improvement priority. Topics that achieved consensus (≥75% of panellists rating them as high priority) were subsequently rated based on feasibility for implementation at the respondent's institution.</p><p><strong>Results: </strong>Literature review identified 24 topics relevant to paediatric diagnostic safety and quality, which were expanded to 62 topics during the initial panel discussion and grouped into 25 survey topics. Consensus was reached on 6 topics as research priorities and 11 as operational improvement priorities. Of these, three were deemed highly feasible for research and three for operational improvement. Research priorities included: identifying paediatric conditions at high risk for diagnostic error, developing communication methods to enhance the diagnostic process and establishing diagnostic process feedback mechanisms. Operational improvement priorities included: identifying care delivery or health services scenarios at high risk for diagnostic error, establishing interdisciplinary, structured review of cases of diagnostic error and encouraging reporting of missed opportunities for improving diagnostic safety.</p><p><strong>Conclusion: </strong>Experts successfully prioritised important and feasible topics in research and practice for improving paediatric diagnosis in US academic medical centres.</p>","PeriodicalId":9077,"journal":{"name":"BMJ Quality & Safety","volume":" ","pages":""},"PeriodicalIF":6.8,"publicationDate":"2026-06-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13450823/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148337838","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Brief mindfulness-based intervention to improve clinician mindful behaviours and well-being at two US hospitals: a randomised controlled, mixed methods trial.","authors":"M Todd Greene, Nathan Houchens, Virginia Sheffield, Prathit A Kulkarni, Karen E Fowler, David Clive, Molly Harrod, Martha Quinn, Barbara Trautner, Sanjay Saint","doi":"10.1136/bmjqs-2025-019980","DOIUrl":"10.1136/bmjqs-2025-019980","url":null,"abstract":"<p><strong>Background: </strong>Clinician burnout is prevalent and associated with poorer patient safety outcomes. Although mindfulness-based interventions can improve clinician resilience, most are time-intensive and not easily embedded into clinical workflows. We thus evaluated a brief mindfulness-based intervention designed to enhance mindful awareness and clinician well-being among physicians and nurses.</p><p><strong>Setting: </strong>Two Veterans Affairs academic medical centres in (Ann Arbor, MI, USA) and (Houston, TX, USA).</p><p><strong>Design: </strong>A randomised controlled, mixed methods study was conducted. Physicians and nurses were randomised to intervention or control at the team/unit level. Quantitative data were collected at baseline, 1 month and 7 months; qualitative interviews and group discussions explored acceptability and perceived impact.</p><p><strong>Participants: </strong>A total of 456 clinicians were randomised and 343 consented to participate (170 intervention and 173 control). Participants included attending and resident physicians and registered nurses working on inpatient general medicine wards.</p><p><strong>Intervention: </strong>The intervention integrated mindfulness practice into routine hand hygiene moments. Components included self-directed online training and a brief facilitated group discussion led by site mindfulness champions. Control participants received no intervention.</p><p><strong>Main outcome measures: </strong>Primary outcomes were self-reported mindfulness (Five Facet Mindfulness Questionnaire (FFMQ)-possible range 1-5, higher scores better) and well-being (Well-Being Index (WBI)-possible range -2 to 9, lower scores better).</p><p><strong>Results: </strong>Clinicians highly engaged (completed modules and discussion) with the intervention showed significant increases in mindfulness (FFMQ) at 1 month (β=0.23; p< 0.001) and 7 months (β=0.23; p<0.001) versus controls. Well-being (WBI) improved modestly over time, with non-significant trends towards reduced distress in the intervention group. One hundred forty-eight clinicians participated in a group discussion and 16 in a semistructured interview. Qualitative findings indicated high acceptability and perceived benefits, including enhanced focus and calm, but identified time pressures and cultural barriers to sustained practice.</p><p><strong>Conclusions: </strong>Embedding mindfulness into routine behaviours such as hand hygiene is feasible and improved clinicians' mindful awareness, although it did not significantly improve well-being. Microinterventions that integrate mindfulness into clinical workflows may offer scalable approaches to support clinician presence, focus and mental resilience.</p><p><strong>Trial registration number: </strong>NCT05261282.</p>","PeriodicalId":9077,"journal":{"name":"BMJ Quality & Safety","volume":" ","pages":""},"PeriodicalIF":6.8,"publicationDate":"2026-06-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13531388/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148367036","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Do patient safety incident investigations align with systems thinking? An analysis of contributing factors and recommendations.","authors":"Lorelle Bowditch, Charlotte Molloy, Brandon King, Masoumeh Abedi, Samantha Jackson, Mia Bierbaum, Yinghua Yu, Louise Raggett, Paul Salmon, Jeffrey Braithwaite, Johanna I Westbrook, Robyn Clay-Williams, Raghu Lingam, Sandy Middleton, Farah Magrabi, Virginia Mumford, Peter Hibbert","doi":"10.1136/bmjqs-2025-019063","DOIUrl":"10.1136/bmjqs-2025-019063","url":null,"abstract":"<p><strong>Background: </strong>Globally, up to 17% of hospitalised people suffer a patient safety incident. Learning from adverse events through patient safety investigation is critical to prevention; however, their utility is still questioned. Two key investigation outputs include identifying contributing factors (CFs) and proposing recommendations to prevent future occurrences. Criticisms of current methods include incomplete analysis of CFs and weak incident prevention strategies. A proposed solution is systems thinking analysis, which recognises healthcare complexity. However, it is not clear whether such methods are being applied in practice.</p><p><strong>Objective: </strong>This study aimed to assess current use of systems thinking-based strategies by examining a set of Australian patient safety incident investigations.</p><p><strong>Methods: </strong>Investigations (n=300) from 56 different Australian health services were deductively analysed. Identified CFs were classified by healthcare system level using a framework combining Systems Engineering Initiative for Patient Safety (SEIPS) principles and AcciMap's hierarchical structure. Recommendation sustainability and effectiveness were classified as weak, medium or strong using US Department of Veteran Affairs' criteria.</p><p><strong>Results: </strong>51% of incidents were issues with clinical processes and procedures. The investigations identified CFs that disproportionally focused on the people involved in those processes (n=677, 47%) rather than other system levels and as a consequence, most recommendations were of medium (n=665, 51%) and weak (n=560, 43%) strength. Notably, 10% of investigations lacked any CFs or recommendations.</p><p><strong>Conclusion: </strong>The focus on individual actions highlighted that simple linear thinking persists in patient safety incident investigations. This study proposes five key areas of effective incident analysis and investigation: a sociotechnical focus; improved data collection techniques; investigative independence; the professionalisation of investigators; and the aggregation of data. Learning from incidents is key to maximising their preventative effectiveness, especially in an increasingly complex healthcare system.</p>","PeriodicalId":9077,"journal":{"name":"BMJ Quality & Safety","volume":" ","pages":"443-455"},"PeriodicalIF":6.8,"publicationDate":"2026-06-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13311954/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145051789","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Impact of COVID-19 on incidence and trends of adverse events among hospitalised patients in Calgary, Canada: a retrospective chart review study.","authors":"Guosong Wu, Cathy A Eastwood, Cheligeer Cheligeer, Danielle A Southern, Yong Zeng, William A Ghali, Jeffrey A Bakal, Bastien Boussat, Ward Flemons, Alan Forster, Yuan Xu, Hude Quan","doi":"10.1136/bmjqs-2024-018182","DOIUrl":"10.1136/bmjqs-2024-018182","url":null,"abstract":"<p><strong>Background: </strong>While the incidence of hospital adverse events appeared to be declining before 2019, the COVID-19 pandemic may have changed its course. This study aimed to evaluate adverse event incidence rates and trends during the pandemic and analyse differences in patient outcomes.</p><p><strong>Methods: </strong>This retrospective electronic chart review included a random sample of adult patients admitted to four acute care hospitals in Calgary between 2017 and 2022. 18 adverse events and patient information were extracted. We calculated the observed and risk-standardised incidence rates of adverse events. Interrupted time series analysis was employed to determine the impact of COVID-19 on adverse events trends. Outcome differences were evaluated using mixed-effects logistic regression and negative binomial models.</p><p><strong>Results: </strong>Among 10 673 patient admissions, 2310 adverse events were identified, resulting in an incidence rate of 21.64 (95% CI 20.77 to 22.54) per 100 patient admissions, or 26.85 (95% CI 25.77 to 27.97) per 1000 patient days. After adjusting for patient characteristics, seasonal variations and overall trends, the adverse event incidence rate increased by 14% (incidence rate ratio (IRR) 1.14, 95% CI 1.01 to 1.29) during the COVID-19 pandemic. In multivariable mixed-effects models, adverse events were associated with significantly longer hospital stays (IRR 3.13, 95% CI 2.97 to 3.30), increased odds of 30-day readmission (OR 1.4, 95% CI 1.17 to 1.68) and in-hospital death (OR 1.72, 95% CI 1.43 to 2.08).</p><p><strong>Conclusion: </strong>The incidence of adverse events was high but relatively stable in acute healthcare settings before the COVID-19 pandemic and increased during the pandemic. Strengthening healthcare resilience and prioritising patient safety initiatives are crucial as we transition into the post-pandemic era.</p>","PeriodicalId":9077,"journal":{"name":"BMJ Quality & Safety","volume":" ","pages":"464-473"},"PeriodicalIF":6.8,"publicationDate":"2026-06-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146059698","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Artificial intelligence chain-of-thought reasoning in nuanced medical scenarios: mitigation of cognitive biases through model intransigence.","authors":"Jonathan Wang, Donald A Redelmeier","doi":"10.1136/bmjqs-2025-019299","DOIUrl":"10.1136/bmjqs-2025-019299","url":null,"abstract":"<p><strong>Background: </strong>Artificial intelligence large language models (LLMs) are increasingly used to inform clinical decisions but sometimes exhibit human-like cognitive biases when facing nuanced medical choices.</p><p><strong>Methods: </strong>We tested whether new chain-of-thought reasoning LLMs might mitigate cognitive biases observed in physicians. We presented medical scenarios (n=10) to models released by DeepSeek, OpenAI and Google. Each scenario was presented in two versions that differed according to a specific bias (eg, surgery framed in survival vs mortality statistics). Responses were categorised and the extent of bias was measured by the absolute discrepancy between responses to different versions of the same scenario. The extent of intransigence (also termed dogma or inflexibility) was measured by Shannon entropy. The extent of deviance in each scenario was measured by comparing the average model response to the average practicing physician response (n=2507).</p><p><strong>Results: </strong>DeepSeek-R1 mitigated 6 out of 10 cognitive biases observed in practicing physicians by generating intransigent all-or-none responses. The four biases that persisted were post hoc fallacy (34% vs 0%, p<0.001), decoy effects (44% vs 5%, p<0.001), Occam's razor fallacy (100% vs 0%, p<0.001) and hindsight bias (56% vs 0%, p<0.001). In every scenario, the average model response deviated substantially from the average response of practicing physicians (p<0.001 for all). Similar patterns of persistent specific biases, intransigent responses and substantial deviance from practicing physicians were also apparent in OpenAI and Google.</p><p><strong>Conclusion: </strong>Some biases persist in chain-of-thought reasoning LLMs, and models tend to produce intransigent recommendations. These findings highlight the role of clinicians to think broadly, respect diversity and remain vigilant when interpreting chain-of-thought reasoning artificial intelligence LLMs in nuanced medical decisions for patients.</p>","PeriodicalId":9077,"journal":{"name":"BMJ Quality & Safety","volume":" ","pages":"456-463"},"PeriodicalIF":6.8,"publicationDate":"2026-06-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145595914","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}