Using nudges with machine learning algorithms to increase serious illness conversations

Opportunity

Patients with cancer often receive care that is discordant with their wishes, resulting in inappropriate and costly utilization.  Evidence demonstrates that proactive serious illness conversations (SICs) can help clinicians align care with patient goals. However, many patients with advanced cancer die without ever having a SIC.  

Approach

We piloted an intervention that paired nudges with a machine learning algorithm that predicts the risk of six-month mortality to see if we could increase SICs between clinicians and cancer patients. Each week, clinicians were provided with a list of patients they were scheduled to see who had a high mortality risk. Then, using an opt-out approach, we asked clinicians to pre-commit to having SICs and sent a text message reminder on the day of the visit. To assess the impact of the intervention, we designed a stepped-wedge cluster randomized controlled trial.

Impact

We found that combining machine learning mortality estimates with behavioral nudges led to a fourfold increase in SICs. It also increased the overall rate of SICs among all patients. This is one of the first clinical trials to demonstrate the potential of combining machine learning with behavioral nudges. The intervention from this study has been implemented across all practices at the Penn Abramson Cancer Center.

Collaborators

Penn Medicine Abramson Cancer Center

Funding

Penn Center for Precision Medicine