Identifying Oncology Patients at High Risk for Potentially Preventable Emergency Department Visits Using a Novel...

Area of Science:

• Oncology
• Healthcare Management
• Health Services Research

Background:

• Cancer patients frequently utilize emergency departments (EDs).
• A subset of these visits may be potentially preventable (PPEDs), impacting care quality and costs.
• Clear definitions and risk factors for PPEDs in oncology are lacking.

Purpose of the Study:

• To define and characterize oncology-related PPEDs.
• To identify patient subgroups at high risk for PPEDs.
• To inform targeted interventions for reducing avoidable ED visits.

Main Methods:

• Retrospective analysis of oncology patient ED visits (April 2019-April 2021).
• Defined PPEDs as visits with immediate discharge or admission <48 hours.
• Employed descriptive statistics, logistic regression, and machine learning (ML) for trend and risk analysis.

Main Results:

• Over 13,000 ED visits by 6,689 oncology patients were analyzed.
• 62.1% of visits were classified as PPEDs.
• High-risk groups for PPEDs included early-stage (I-III) breast cancer patients and those on systemic therapy. Stage IV GI/lung cancer patients and proximity to ED were linked to non-PPEDs.

Conclusions:

• A novel definition of PPEDs shows promise for identifying at-risk oncology patients.
• Early-stage breast cancer patients on systemic therapy are prime candidates for interventions to prevent ED visits.
• ML models offer a promising avenue for robust PPED prediction, though validation is needed.