Family and professionals underestimate quality of life across diverse cultures and health conditions: Systematic review

Objective To examine how accurately proxies evaluate quality of life (QoL) in people they know, using cross-cultural data from the multidimensional, multilingual World Health Organization Quality of Life assessment short-form – The WHOQOL-BREF, and whether accuracy varies by health condition or proxy type (e.g. family/professional). Study design and setting Systematic review with meta-analysis: We searched five databases for reports of proxy-completed WHOQOL-BREF scores and aggregated results using a random-effects model. Minimal clinically important difference values were calculated. Results Analyses included nine studies (1980 dyads) of physical (n=762) or mental (n=604) health conditions, or intellectual disability (n=614), in 10 countries. Mean person-proxy correlations ranged from 0.28 (social QoL) to 0.44 (physical QoL). Proxy measures were underestimates (i.e. significantly lower than persons reported for themselves) for social (mean difference (MD) = 4.7, 95% confidence interval (CI) 1.8 to 7.6), psychological (MD=3.7, 95% CI 0.6 to 6.8), and physical (MD=3.1, 95% CI 0.6 to 5.6) QoL. Underestimates varied significantly between health conditions for social (p