Clinician perspectives on how situational context and augmented intelligence design features impact perceived usefulness of sepsis prediction scores embedded within a simulated electronic health record.

OBJECTIVE: Obtain clinicians' perspectives on early warning scores (EWS) use within context of clinical cases. MATERIAL AND METHODS: We developed cases mimicking sepsis situations. De-identified data, synthesized physician notes, and EWS representing deterioration risk were displayed in a simulated EHR for analysis. Twelve clinicians participated in semi-structured interviews to ascertain perspectives across four domains: (1) Familiarity with and understanding of artificial intelligence (AI), prediction models and risk scores; (2) Clinical reasoning processes; (3) Impression and response to EWS; and (4) Interface design. Transcripts were coded and analyzed using content and thematic analysis. RESULTS: Analysis revealed clinicians have experience but limited AI and prediction/risk modeling understanding. Case assessments were primarily based on clinical data. EWS went unmentioned during initial case analysis; although when prompted to comment on it, they discussed it in subsequent cases. Clinicians were unsure how to interpret or apply the EWS, and desired evidence on its derivation and validation. Design recommendations centered around EWS display in multi-patient lists for triage, and EWS trends within the patient record. Themes included a "Trust but Verify" approach to AI and early warning information, dichotomy that EWS is helpful for triage yet has disproportional signal-to-high noise ratio, and action driven by clinical judgment, not the EWS. CONCLUSIONS: Clinicians were unsure of how to apply EWS, acted on clinical data, desired score composition and validation information, and felt EWS was most useful when embedded in multi-patient views. Systems providing interactive visualization may facilitate EWS transparency and increase confidence in AI-generated information.

Information displays for automated surveillance algorithms of in-hospital patient deterioration: a scoping review.

OBJECTIVE: Surveillance algorithms that predict patient decompensation are increasingly integrated with clinical workflows to help identify patients at risk of in-hospital deterioration. This scoping review aimed to identify the design features of the information displays, the types of algorithm that drive the display, and the effect of these displays on process and patient outcomes. MATERIALS AND METHODS: The scoping review followed Arksey and O'Malley's framework. Five databases were searched with dates between January 1, 2009 and January 26, 2022. Inclusion criteria were: participants-clinicians in inpatient settings; concepts-intervention as deterioration information displays that leveraged automated AI algorithms; comparison as usual care or alternative displays; outcomes as clinical, workflow process, and usability outcomes; and context as simulated or real-world in-hospital settings in any country. Screening, full-text review, and data extraction were reviewed independently by 2 researchers in each step. Display categories were identified inductively through consensus. RESULTS: Of 14 575 articles, 64 were included in the review, describing 61 unique displays. Forty-one displays were designed for specific deteriorations (eg, sepsis), 24 provided simple alerts (ie, text-based prompts without relevant patient data), 48 leveraged well-accepted score-based algorithms, and 47 included nurses as the target users. Only 1 out of the 10 randomized controlled trials reported a significant effect on the primary outcome. CONCLUSIONS: Despite significant advancements in surveillance algorithms, most information displays continue to leverage well-understood, well-accepted score-based algorithms. Users' trust, algorithmic transparency, and workflow integration are significant hurdles to adopting new algorithms into effective decision support tools.