Predicting models of depression or complicated grief among bereaved family members of patients with cancer.

Abstract

To derive and validate a risk scoring system for predicting major depressive disorder (MDD) and complicated grief (CG) among bereaved family members of patients with cancer that is feasible for clinical use. We conducted a secondary analysis of two cross-sectional nationwide bereavement surveys in Japan. From a total of 17,312 bereaved family members of patients with cancer, 8618 and 8619 were randomly assigned to a derivation and a validation group. The Patient Health Questionnaire 9 (PHQ-9) and the Brief Grief Questionnaire (BGQ) were used to assess MDD (PHQ-9 score≥10) and CG (BGQ score≥8), respectively. We compared five models with potential predictive variables that could be easily obtained in daily practice and were included in the bereavement survey (i.e., sociodemographic data). The model which included variables such as the families' physical/mental health status and preparedness toward bereavement, in addition to their sociodemographic data, was considered modest for predicting the risk of both MDD and CG. The areas around the curve for MDD and CG were 0.74 (95% CI: 0.73-0.76) and 0.74 (95% CI: 0.72-0.75) and 0.78 (95% CI: 0.76-0.79) and 0.77 (95% CI: 0.76-0.79) in the derivation and validation groups, respectively. We developed a clinical risk score for predicting MDD and CG among bereaved family members of patients with cancer. However, further research is needed for external validation and assessment regarding its implementation in actual practice.