Doctors' perception on the ethical use of AI-enabled clinical decision support systems for antibiotic prescribing recommendations in Singapore.

The increased utilization of Artificial intelligence (AI) in healthcare changes practice and introduces ethical implications for AI adoption in medicine. We assess medical doctors' ethical stance in situations that arise in adopting an AI-enabled Clinical Decision Support System (AI-CDSS) for antibiotic prescribing decision support in a healthcare institution in Singapore. We conducted in-depth interviews with 30 doctors of varying medical specialties and designations between October 2022 and January 2023. Our interview guide was anchored on the four pillars of medical ethics. We used clinical vignettes with the following hypothetical scenarios: (1) Using an antibiotic AI-enabled CDSS's recommendations for a tourist, (2) Uncertainty about the AI-CDSS's recommendation of a narrow-spectrum antibiotic vs. concerns about antimicrobial resistance, (3) Patient refusing the "best treatment" recommended by the AI-CDSS, (4) Data breach. More than half of the participants only realized that the AI-enabled CDSS could have misrepresented non-local populations after being probed to think about the AI-CDSS's data source. Regarding prescribing a broad- or narrow-spectrum antibiotic, most participants preferred to exercise their clinical judgment over the AI-enabled CDSS's recommendations in their patients' best interest. Two-thirds of participants prioritized beneficence over patient autonomy by convincing patients who refused the best practice treatment to accept it. Many were unaware of the implications of data breaches. The current position on the legal liability concerning the use of AI-enabled CDSS is unclear in relation to doctors, hospitals and CDSS providers. Having a comprehensive ethical legal and regulatory framework, perceived organizational support, and adequate knowledge of AI and ethics are essential for successfully implementing AI in healthcare.

Huang Z ，Lim HY ，Ow JT ，Sun SH ，Chow A ... -  《Frontiers in Public Health》

Are physicians ready for precision antibiotic prescribing? A qualitative analysis of the acceptance of artificial intelligence-enabled clinical decision support systems in India and Singapore.

Huang Z ，George MM ，Tan YR ，Natarajan K ，Devasagayam E ，Tay E ，Manesh A ，Varghese GM ，Abraham OC ，Zachariah A ，Yap P ，Lall D ，Chow A ... -  《-》

Patient and physician predictors of patient receipt of therapies recommended by a computerized decision support system when initially prescribed broad-spectrum antibiotics: a cohort study.

Antibiotic computerized decision support systems (CDSSs) were developed to guide antibiotic decisions, yet prescriptions of CDSS-recommended antibiotics have remained low. Our aim was to identify predictors of patients' receipt of empiric antibiotic therapies recommended by a CDSS when the prescribing physician had an initial preference for using broad-spectrum antibiotics. We conducted a prospective cohort study in a 1 500-bed tertiary-care hospital in Singapore. We included all patients admitted from October 1, 2011 through September 30, 2012, who were prescribed piperacillin-tazobactam or carbapenem for empiric therapy and auto-triggered to receive antibiotic recommendations by the in-house antibiotic CDSS. Relevant data on the patient, prescribing and attending physicians were collected via electronic linkages of medical records and administrative databases. To account for clustering, we used multilevel logistic regression models to explore factors associated with receipt of CDSS-recommended antibiotic therapy. One-quarter of the 1 886 patients received CDSS-recommended antibiotics. More patients treated for pneumonia (33.2%) than sepsis (12.1%) and urinary tract infections (7.1%) received CDSS-recommended antibiotic therapies. The prescribing physician - but not the attending physician or clinical specialty - accounted for some (13.3%) of the variation. Prior hospitalization (odds ratio [OR] 1.32, 95% CI, 1.01-1.71), presumed pneumonia (OR 6.77, 95% CI, 3.28-13.99), intensive care unit (ICU) admission (OR 0.38, 95% CI, 0.21-0.66), and renal impairment (OR 0.70, 95% CI, 0.52-0.93) were factors associated with patients' receipt of CDSS-recommended antibiotic therapies. We observed that ICU admission and renal impairment were negative predictors of patients' receipt of CDSS-recommended antibiotic therapies. Patients admitted to ICU and those with renal impairment might have more complex clinical conditions that require a physician's assessment in addition to antibiotic CDSS.

Chow AL ，Lye DC ，Arah OA 《-》

Clinicians' roles and necessary levels of understanding in the use of artificial intelligence: A qualitative interview study with German medical students.

Artificial intelligence-driven Clinical Decision Support Systems (AI-CDSS) are being increasingly introduced into various domains of health care for diagnostic, prognostic, therapeutic and other purposes. A significant part of the discourse on ethically appropriate conditions relate to the levels of understanding and explicability needed for ensuring responsible clinical decision-making when using AI-CDSS. Empirical evidence on stakeholders' viewpoints on these issues is scarce so far. The present study complements the empirical-ethical body of research by, on the one hand, investigating the requirements for understanding and explicability in depth with regard to the rationale behind them. On the other hand, it surveys medical students at the end of their studies as stakeholders, of whom little data is available so far, but for whom AI-CDSS will be an important part of their medical practice. Fifteen semi-structured qualitative interviews (each lasting an average of 56 min) were conducted with German medical students to investigate their perspectives and attitudes on the use of AI-CDSS. The problem-centred interviews draw on two hypothetical case vignettes of AI-CDSS employed in nephrology and surgery. Interviewees' perceptions and convictions of their own clinical role and responsibilities in dealing with AI-CDSS were elicited as well as viewpoints on explicability as well as the necessary level of understanding and competencies needed on the clinicians' side. The qualitative data were analysed according to key principles of qualitative content analysis (Kuckartz). In response to the central question about the necessary understanding of AI-CDSS tools and the emergence of their outputs as well as the reasons for the requirements placed on them, two types of argumentation could be differentiated inductively from the interviewees' statements: the first type, the clinician as a systemic trustee (or "the one relying"), highlights that there needs to be empirical evidence and adequate approval processes that guarantee minimised harm and a clinical benefit from the employment of an AI-CDSS. Based on proof of these requirements, the use of an AI-CDSS would be appropriate, as according to "the one relying", clinicians should choose those measures that statistically cause the least harm. The second type, the clinician as an individual expert (or "the one controlling"), sets higher prerequisites that go beyond ensuring empirical evidence and adequate approval processes. These higher prerequisites relate to the clinician's necessary level of competence and understanding of how a specific AI-CDSS works and how to use it properly in order to evaluate its outputs and to mitigate potential risks for the individual patient. Both types are unified in their high esteem of evidence-based clinical practice and the need to communicate with the patient on the use of medical AI. However, the interviewees' different conceptions of the clinician's role and responsibilities cause them to have different requirements regarding the clinician's understanding and explicability of an AI-CDSS beyond the proof of benefit. The study results highlight two different types among (future) clinicians regarding their view of the necessary levels of understanding and competence. These findings should inform the debate on appropriate training programmes and professional standards (e.g. clinical practice guidelines) that enable the safe and effective clinical employment of AI-CDSS in various clinical fields. While current approaches search for appropriate minimum requirements of the necessary understanding and competence, the differences between (future) clinicians in terms of their information and understanding needs described here can lead to more differentiated approaches to solutions.

Funer F ，Tinnemeyer S ，Liedtke W ，Salloch S ... -  《BMC Medical Ethics》

Psychosocial determinants of physicians' acceptance of recommendations by antibiotic computerised decision support systems: A mixed methods study.

Antibiotic computerised decision support systems (CDSSs) were developed to facilitate optimal prescribing, but acceptance of their recommendations has remained low. We aimed to evaluate physicians' perceptions and attitudes toward antibiotic CDSSs and determine psychosocial factors associated with acceptance of CDSS recommendations for empirical therapy. A mixed methods study was conducted in an adult tertiary-care hospital in Singapore, with its in-house antibiotic CDSS that integrates antimicrobial stewardship with electronic prescribing. Focus group discussions were conducted among purposively sampled physicians and data were analysed using the framework approach. Emerging themes were included in the questionnaire with newly developed scales for the subsequent cross-sectional survey involving all physicians. Principal components analysis was performed to derive the latent factor structure that was later applied in multivariate analyses. Physicians expressed confidence in the credibility of CDSS recommendations. Junior physicians accepted CDSS recommendations most of the time, whilst senior physicians acknowledged overriding recommendations in complex patients with multiple infections or allergies. Willingness to consult the CDSS for common and complex infections (OR=1.68, 95% CI 1.16-2.44) and preference for personal or team decision (OR=0.61, 95% CI 0.43-0.85) were associated with acceptance of CDSS recommendations. Cronbach's α for scales measuring physicians' attitudes and perceptions towards acceptance of CDSS recommendations ranged from 0.64 to 0.88. Physicians' willingness to consult an antibiotic CDSS determined acceptance of its recommendations. Physicians would choose to exercise their own or clinical team's decision over CDSS recommendations in complex patient situations when the antibiotic prescribing needs were not met.

Chow A ，Lye DC ，Arah OA 《-》