Evaluating a Clinical Decision Support Tool for Cancer Risk Assessment in Primary Care: Simulation Study of...

Early cancer detection is crucial, but recognizing the significance of associated symptoms such as unintended weight loss in primary care remains challenging. Clinical decision support systems (CDSSs) can aid cancer detection but face implementation barriers and low uptake in real-world settings. To address these issues, simulation environments offer a controlled setting to study CDSS usage and improve their design for better adoption in clinical practice.

This study aimed to evaluate a CDSS integrated within general practice electronic health records aimed at identifying patients at risk of undiagnosed cancer.

The evaluation of a CDSS to identify patients with unintended weight loss was conducted in a simulated primary care environment where general practitioners (GPs) interacted with the CDSS in simulated clinical consultations. There were four possible clinical scenarios based on patient gender and risk of cancer. Data collection included interviews with GPs, cancer survivors (lived-experience community advocates), and patient actors, as well as video analysis of GP-CDSS interactions. Two theoretical frameworks were employed for thematic interpretation of the data.

We recruited 10 GPs and 6 community advocates, conducting 20 simulated consultations with 2 patient actors (2 consultations per GP: 1 high-risk consultation and 1 low-risk consultation). All participants found the CDSS acceptable and unobtrusive. GPs utilized CDSS recommendations in three distinct ways: as a communication aid when discussing follow-up with the patient, as a reminder for differential diagnoses and recommended investigations, and as an aid to diagnostic decision-making without sharing with patients. The CDSS's impact on patient-doctor communication varied, facilitating and hindering interactions depending on the GP's communication style.

We developed and evaluated a CDSS for identifying cancer risk in patients with unintended weight loss in a simulated environment, revealing its potential to aid clinical decision-making and communication while highlighting implementation challenges and the need for context-sensitive application.