Navigating specific targets of breast cancer symptoms: An innovative computer-simulated intervention analysis

Abstract

Purpose

To pinpoint optimal interventions by dissecting the complex symptom interactions, encompassing both their static and temporal dimensions.

Methods

The study incorporated a cross-sectional survey utilizing the MD Anderson Symptom Inventory. Participants with breast cancer undergoing chemotherapy were recruited from the “Be Resilient to Breast Cancer” from April 2023 to June 2024. Static symptom interrelationships were elucidated using undirected and Bayesian network models, complemented by an exploration of their dynamic counterparts through computer-simulated interventions.

Results

The study included 602 patients with breast cancer. Both undirected networks and computer-simulated interventions concurred on the symptoms of distress and fatigue as optimal alleviation targets. The Bayesian network and computer-simulated interventions both emphasized “shortness of breath” as preventive care. Notably, Distress appeared to be the most effective target for interventions, and compared to fatigue (decreasing score = 1.84–2.20, decreasing prevalence = 14.2–16.7%). Conversely, disturbed sleep, despite its high position in Bayesian network, had no propelling effects on increasing the network's overall symptom activity levels (increasing score<1).

Conclusions

Computer-simulated intervention integrating with traditional network analysis can improve intervention precision and efficacy by prioritizing individual symptom impacts, both statically and dynamically.