"Youthful epidemic" or diagnostic bias? Differential item functioning of DSM-IV cannabis use criteria in an Australian general population survey

Abstract

Background: Prevalence rates of cannabis abuse and dependence are higher for younger adults when compared with older adults. It is necessary to examine the performance of criteria across age groups to establish whether these differences in prevalence rates are a result of diagnostic bias. The current study aims to investigate whether there is any age-related differential item functioning in the DSM-IV diagnostic criteria for cannabis use disorders using an item response theory approach.

Method: The sample consisted of 10,641 participants in a population-based survey of Australian adults aged 18 years and older. DSM-IV cannabis use disorders were assessed in all respondents who had used cannabis more than five times in the previous twelve months (n=722). Age-based differential item functioning was assessed in each of the DSM-IV criteria for cannabis use disorders.

Results: Age-based differential item functioning was only detected in the Hazard criterion of the DSM-IV cannabis use disorders. The Hazard criterion was found to be more discriminating for those aged 18-24 when compared with those aged 25 and over.

Conclusions: The DSM-IV criteria for cannabis use disorders appear to function similarly across age groups. Differential item functioning was only detected in the alpha parameter for the Hazard criterion. These results are discussed with regard to implications for future editions of the DSM system.