Validating a Patient-Reported Outcomes-Derived Algorithm for Classifying Symptom Complexity Levels Among Patients...

Area of Science:

• Oncology
• Health Services Research
• Patient-Reported Outcomes

Background:

• Patient-reported outcomes (PROs) symptom complexity algorithm requires extensive validation.
• The algorithm uses Edmonton Symptom Assessment System and Canadian Problem Checklist data.

Purpose of the Study:

• To validate the PRO-derived symptom complexity algorithm.
• To assess the algorithm's accuracy and screening capacity in patients with cancer.

Main Methods:

• Retrospective chart review of 1,466 cancer patients in Alberta, Canada (2016-2017).
• Utilized Alberta Cancer Registry and electronic medical records.
• Validated against Karnofsky performance status and compared with a 2-step cluster analysis algorithm.

Main Results:

• The algorithm demonstrated good accuracy (77.7%) and a high area under the receiver operating characteristic curve (0.824).
• It effectively classified patients into distinct health status subgroups with a large effect size (d=1.2).
• Outperformed a 2-step cluster analysis algorithm (AUC=0.721).

Conclusions:

• The PRO-derived symptom complexity algorithm's validity is established.
• It shows satisfactory accuracy and strong correlation with known groups, outperforming alternative methods.
• The algorithm serves as a crucial clinical tool for identifying patients needing timely, targeted symptom management.