Abstract
BACKGROUND: Epidemiological studies have shown that total scores in depression screening scales change with age, but the mechanism underlying these age-related changes remains unclear. Previous research has indicated that item responses in depression screening scales exhibit characteristic distributions in the general population. We analyzed Patient Health Questionnaire-9 (PHQ-9) data from a representative survey conducted in the USA, to determine how the response pattern for each item changed with age and whether the pattern of responses contributed to age-related changes in total scores. METHODS: We analyzed PHQ-9 data for 17,274 participants in the 2011-2016 National Health and Nutrition Examination Survey. The PHQ-9 allows respondents to self-rate the frequency of depressive symptoms using a four-point scale ranging from "not at all" to "nearly every day". RESULTS: The lines for all nine item responses followed the same characteristic pattern across all age groups, which was marked by intersection at a single point between "not at all" and "several days" and parallel patterns between "several days" and "nearly every day" on a logarithmic scale. The probability of "nearly every day" showed a reverse U-shaped pattern, in that it was low from 12-29 years, increased during 30-50 years, and then decreased at ≥60 years. The age-related change in the probability of a response of "nearly every day" coincided with the trajectory of the PHQ-9 total scores. CONCLUSIONS: This study demonstrated that item responses for the PHQ-9 followed a similar mathematical pattern across the adult lifespan. Moreover, our findings suggested that the probability of a response of "nearly every day" played an important role in age-related changes in PHQ-9 total scores across adulthood.