Development and validation of a patient reported outcome measure for health-related quality of life for locally recurrent rectal cancer: a multicentre, three-phase, mixed-methods, cohort study
Deena Harji, Cherry Koh, Niamh McKigney, Michael J. Solomon, Ben Griffiths, Martyn Evans, Alexander G. Heriot, P. M. Sagar, Galina Velikova, Julia Brown
Abstract
Background: Locally recurrent rectal cancer (LRRC) occurs in 5-10% of patients following previous treatment of rectal cancer. It has a significant impact on patients' overall health-related quality of life (HrQoL). Major advances in surgical treatments have led to improved survival outcomes. However, due to the lack of disease-specific, validated patient-reported outcome measure (PROM), HrQoL, is variably assessed. The aim of this study is to develop a disease-specific, psychometrically robust, and validated PROM for use in LRRC. Methods: A multicentre, three phase, mixed-methods, observational study was performed across five centres in the UK and Australia. Adult patients (>18 years old) with an existing or previously treated LRRC within the last 2 years were eligible to participate. Patients completed the proposed LRRC-QoL, EORTC QLQ-CR29, and FACT-C questionnaires. Scale structure was analysed using multi-trait scaling analysis and exploratory factor analysis, reliability was assessed using Cronbach's and the intra-class coefficient, convergent validity was assessed using Pearson's correlation, and known-groups comparison was assessed using the student t-test or ANOVA. Findings: Between 01/03/2015 and 31/12/2019, 117 patients with a diagnosis of LRRC were recruited. The final scale structure of the LRRC-QoL consisted of nine multi-item scales (healthcare services, psychological impact, pain, urostomy-related symptoms, lower limb symptoms, stoma, sexual function, sexual interest, and urinary symptoms) and three single items. Cronbach's Alpha and Intraclass correlation values of >0.7 across the majority of scales supported overall reliability. Convergent validity was demonstrated between LRRC-QoL Pain Scale and FACT-C Physical Well Being scale (r = 0.528, p < 0.001), LRRC-QoL Psychological Impact scale with EORTC QLQ CR29 Body Image (r = 0.680, p < 0.001) and the FACT-C Emotional Well Being scale (r = 0.326, p < 0.001), and LRRC-QoL Urinary Symptoms scale with EORTC QLQ-CR29 Urinary Frequency scale (r = 0.310, p < 0.001). Known-groups validity was demonstrated for gender, disease location, treatment intent, and re-recurrent disease. Interpretation: The LRRC-QoL has demonstrated robust psychometric properties and can be used in clinical and academic practice. Funding: None.