Measurement invariance and differential item functioning in cancer quality-of-life scales across cultural contexts in ghana

Abstract

Health-related quality of life (HRQoL) is a critical outcome in oncology care. While numerous QoL instruments exist, their psychometric properties and cross-cultural validity have not been fully examined in Ghanaian populations. The study aimed to evaluate the reliability, factorial validity, measurement invariance, and differential item functioning (DIF) of a culturally adapted 24-item cancer QoL scale among adult oncology patients in Ghana. A cross-sectional, multi-center survey was conducted with 887 adult cancer patients across major oncology centers in Ghana. The adapted QoL scale assessed Physical, Emotional, Social, and Functional Well-Being. Analyses included descriptive statistics, internal consistency (Cronbach’s α, McDonald’s ω), confirmatory factor analysis (CFA), multi-group CFA for measurement invariance across Akan, Ewe, Northern language speakers, and other minority groups, DIF using Mantel-Haenszel, logistic regression, and item response theory. The scale demonstrated excellent reliability (α = 0.87–0.94, ω = 0.88–0.95) and acceptable factorial validity (CFI = 0.958, TLI = 0.951, RMSEA = 0.038). Configural, metric, scalar, and strict invariance were supported across cultural groups. Minor DIF was detected for Emotional Well-Being, Fatigue, and Family Support items. Latent mean comparisons indicated small variations across groups, with most domains largely comparable. The adapted 24-item cancer QoL scale is reliable, valid, and largely cross-culturally equivalent for assessing HRQoL among diverse Ghanaian oncology populations. It provides a culturally sensitive tool for clinical assessment, research, and cross-cultural comparisons.