Responsible AI in action: Planning through implementation of a mortality model for palliative care

ImportanceInterest in the use of prediction models to support referrals to palliative care is surging. Few high-performing models have been developed, implemented, and pilot-tested following responsible artificial intelligence (AI) principles and transparent reporting guidelines. ObjectiveCollaborate with physicians and clinical informaticists to plan, develop, validate, implement, and pilot-test a high-performing, statistically fair prediction model--the Serious Illness Clinical Indicator (SICLI)--to support palliative care referrals following responsible AI principles. DesignCohort study of hospitalizations occurring from January 1st, 2022, to December 31st, 2022. SettingKaiser Permanente Southern California ParticipantsPatients 18+ years, not admitted to maternity or psychiatry departments and hospitalized for at least 36 hours. ExposuresWe augmented a commercial prediction model, the Epic End-of-Life-Care Index (EOLCI), by adding locally available predictors, including co-morbidity and laboratory acuity risk scores, nursing flow sheet and health care utilization data. Main Outcomes and MeasuresWe predicted 12-month mortality risk using a split-sample design, least absolute shrinkage, and selection operator (LASSO) using logistic regression, and five-fold cross-validation. Performance was assessed via the Area Under the Receiver Operating Curve (AUC) and calibration plots. The final model was chosen by clinicians and informaticists for its balance of performance and parsimony. Responsible AI principles guided each development step. ResultsTwelve-month mortality rate among 133,043 hospitalizations of at least 36 hours was 23%. Patients who died were older (76{+/-}13 vs. 64{+/-}17 years) and had higher co-morbidity burden (Charlson, 4.49{+/-}3.14 vs. 2.06{+/-}2.44). SICLI was very well calibrated and outperformed the EOLCI (AUC of 0.87 (95% CI:0.86, 0.87) vs. 0.81, (95% CI: (0.80, 0.81)). SICLIs high and very-high risk group (probability cut-off 0.6-<95 and >=0.95), produced positive predictive values (PPVs) of 72.0% and 94.0%, respectively. SICLI was implemented in KPSCs Electronic Health Record. Post-implementation validation by a palliativist against the 12-month surprise question via review of 25 charts yielded a PPV of 100% and 80% for the very-high and high-risk groups. We report on model documentation, socialization, and governance. Conclusion and RelevanceSICLI is an equitable, high-performing mortality prediction model that builds on and outperforms the EOLCI. It was implemented and pilot-tested in a large health care system. Key pointsO_ST_ABSQuestionC_ST_ABSCan we build on a commercially available mortality risk score to develop, implement, and pilot-test a high-performing prediction model to support palliative care decision-making in a large integrated health care system? FindingsThe Serious Illness Clinical Indicator (SICLI) augments and outperforms the Epic End-of-Life Care Score (EOLCI) and was implemented and pilot-tested using a responsible AI framework. MeaningPredictive models can be valuable tools to support palliative care referrals.