SHM Converge Scientific Abstract ePoster Gallery-From Algorithm to Action: A Rapid Qualitative Analysis of How Clinicians Use Mortality Risk Prediction to Guide Prognosis and Care at a Tertiary Care Center

From Algorithm to Action: A Rapid Qualitative Analysis of How Clinicians Use Mortality Risk Prediction to Guide Prognosis and Care at a Tertiary Care Center

Back to course

Pdf Summary

This qualitative study explored how 25 hospitalists at an academic health center interpret and use mortality risk prediction models in real clinical care. The goal was to understand how clinicians trust these tools, integrate them into workflow, and use them to support prognosis and goals-of-care discussions.<br /><br />The findings show that mortality risk scores are most useful when they complement, rather than replace, clinical judgment. Clinicians valued models that matched bedside intuition, improved situational awareness, and supported anticipatory planning. However, adoption depended on several factors: the score had to be transparent and explainable, fit smoothly into workflow, and align with clinician reasoning. Trust was strengthened when clinicians could understand “why” a risk score was generated and when the tool demonstrated concordance with their own assessment.<br /><br />The study also highlighted that communication is a central part of mortality risk use. Prognosis conversations were emotionally difficult, but structured communication frameworks such as REMAP and SPIKES helped clinicians deliver serious news more effectively. Participants emphasized that communication is not separate from care—it is care.<br /><br />A key theme was that mortality prediction should be dynamic, not static. Reassessment over time, rather than reliance on a single score, better supports decision-making. The authors conclude that successful implementation requires explainable outputs, workflow usability, interdisciplinary engagement, communication training, and clear linkage between risk estimates and clinical action.

Asset Subtitle

Neetu Mahendraker

Meta Tag

Author List Amy Johnson, Ann Cottingham, Neetu Mahendraker, Titus Schleyer

Category Research

Concept Mortality Prediction Model

Concept Prognostication

Concept Patient-Centered Decision Making

Concept Clinician Trust

Concept Workflow Integration

Distinguished Non-Finalist

Presenter Organization Indiana University

Presenting Author Neetu Mahendraker

Track Technology in Hospital Medicine

Keywords

mortality risk prediction

hospitalists

clinical judgment

prognosis communication

goals of care

risk scores

workflow integration

explainable models

decision support

serious illness conversations

Mortality Prediction Model

Prognostication

Patient-Centered Decision Making

Clinician Trust

Workflow Integration