Purpose: A patient's physical function is a critical outcome variable for measuring and improving chronic care management. However, patient-reported outcome measures of physical function are not routinely assessed in cancer outpatients, in part due to limitations of tools available. This study presents the development and evaluation of the Cancer Ambulatory Patient Physical Function Longitudinal Evaluation Tool (CAPLET) as an adaptive response tool for routinely screening for physical dysfunction in oncology clinical practice.
Methods: In phase 1, 407 adult outpatients at Princess Margaret Cancer Centre completed the World Health Organization Disability Assessment Schedule (WHODAS) 2.0, Health Assessment Questionnaire Disability Index (HAQ-DI), EuroQuol-5D-3L ( EQ-5D-3L), and patient-reported outcome (PRO)-Eastern Cooperative Oncology Group (ECOG). CAPLET was developed based on a branching logic algorithm navigating patients to appropriate domains of HAQ-DI/WHOAS using their responses to the PRO-ECOG/EQ-5D-3L as screeners. Sensitivity/specificity of CAPLET screeners for HAQ-DI/WHODAS items were reported. In phase 2, CAPLET vs the WHODAS/HAQ-DI were alternatively administrated to 318 adult outpatients in a two-arm trial comparing time to completion and acceptability between the tools.
Results: Using a patient's ECOG status and the sum of the mobility, self-care, and usual activity dimensions of the EQ-5D-3L to dichotomize patients as with or without difficulty, CAPLET achieved a sensitivity > 90% against recommended WHODAS and HAQ-DI cutoffs for significant dysfunction. Sensitivity of screeners for capturing dysfunction in individual WHODAS/HAQ-DI items ranged from 85 to 100%. Compared to the HAQ-DI/WHODAS, CAPLET was associated with a 50% reduction in administration times and improved patient acceptability, while reducing question burden by 84% for half the sample population.
Conclusions: CAPLET improves the feasibility of capturing detailed assessments of patient-reported physical function in cancer outpatients.
Keywords: Cancer; Computer logic; Patient reported outcome; Physical function; Sensitivity; Specificity.