To evaluate the accuracy of patient self-reports of health care insurance coverage relative to actual health care benefits and to explore patient predictors of accuracy among patients with depression.
Data were analyzed from 767 patients with current depressive symptoms and disorder who completed a self-report mail survey that included questions about health insurance and for whom we also obtained actual health care benefits records. Percentage agreement and κ statistics were calculated for different types of benefits (medical visit, pharmacy and mental health visit copays, and mental health visit limits), and we used multivariate analysis to examine predictors of accuracy.
Accuracy of self-reports relative to administrative data were good for measures of medical visits and prescription copays (κ = 0.79 and 0.64), but less accurate for measures of mental health benefits (κ = 0.19 for copays and 0.40 for visit coverage). Recent use of health care services (past 6 months) increased the accuracy of medical benefits but not of mental health benefits. Inaccuracy of mental health benefits was in the direction of patients perceiving more generous coverage than was actually available to them. Men and those with greater wealth were more knowledgeable about their medical benefits. Those from a nonminority ethnic group, who were less satisfied, and were less sick had better knowledge about their mental health benefits.
Agreement between self-reports and actual benefits was stronger for general medical services than for mental health care. Background, experience, and sickness lead to more accurate recall of insurance information.