Beyond using composite measures to analyze the effect of unmet supportive care needs on caregivers’ anxiety and depression

Abstract

Objective: Caregiver research has relied on composite measures (e.g., count) of unmet supportive care needs to determine relationships with anxiety and depression. Such composite measures assume that all unmet needs have a similar impact on outcomes. The purpose of this study is to identify individual unmet needs most associated with caregivers’ anxiety and depression. Methods: 219 Caregivers completed the 44-item Supportive Care Needs Survey and the Hospital Anxiety and Depression scale [minimal clinically important difference (MCID)=1.5] at 6-8 months, 1, 2, 3.5, and 5 years following the patients' cancer diagnosis. The list of needs was reduced using Partial Least Square regression and those with a Variance Importance in Projection > 1 were analyzed using Bayesian Model Averaging. Results: Across time, eight items remained in the top 10 based on prevalence and were labelled “core”. Three additional ones were labelled “frequent”, as they remained in the top 10 from 1- year onwards. Bayesian Model Averaging identified a maximum of four significant unmet needs per time point – all leading to a difference greater than the MCID. For depression, none of the core unmet needs were significant, rather significance was noted for frequent needs and needs that were not prevalent. For anxiety, 3/8 core and 3/3 frequent unmet needs were significant. Conclusions: Prevalent Those unmet needs that are most prevalent are not necessarily the most significant ones, and findings provide an evidence-based framework to guide the development of caregiver interventions. A broader contribution is proposing a different approach to identify significant unmet needs.