Commentary on Agrawal et al. (2011): Assessing the impact of changes to diagnostic criteria on the estimated prevalence of alcohol use disorders in general population surveys
Article first published online: 6 OCT 2011
© 2011 The Authors, Addiction © 2011 Society for the Study of Addiction
Volume 106, Issue 11, pages 1944–1945, November 2011
How to Cite
SLADE, T., TEESSON, M., CARRAGHER, N. and FARRELL, M. (2011), Commentary on Agrawal et al. (2011): Assessing the impact of changes to diagnostic criteria on the estimated prevalence of alcohol use disorders in general population surveys. Addiction, 106: 1944–1945. doi: 10.1111/j.1360-0443.2011.03597.x
- Issue published online: 6 OCT 2011
- Article first published online: 6 OCT 2011
- Alcohol use disorders;
There has been some substantial controversy and concern around the proposed changes to a range of diagnostic conditions in the proposed new DSM-5. Suggested modifications to the diagnostic criteria for the alcohol use disorders (AUDs) have been grounded in a rich body of empirical literature that has largely supported (i) the removal of the diagnostic distinction between alcohol abuse and dependence, (ii) the exclusion of a ‘recurrent alcohol-related legal problems’ criterion, (iii) the introduction of an ‘alcohol craving or strong desire to drink’ criterion and (iv) the recognition that AUDs should be diagnosed on a continuum of severity. In our view, dropping the abuse/dependence distinction and excluding the legal problems criterion are welcome changes that surely assist in greater diagnostic clarity. The addition of a severity dimension must be of potential assistance in assessing treatment need. The key question, then, is how do these changes work in practice?
Agrawal et al.  provide an elegant and thorough examination of the practical impact of the changes to DSM. This paper, and the other referred to in that paper , have capitalized on epidemiological data sets to assess the impact of these changes on estimated population prevalence of AUDs. The findings of Agrawal et al. indicate that the new definition results in a modest (11%) increase in the prevalence of AUD. The earlier paper suggested a greater (62%) increase in prevalence. Agrawal et al. demonstrate elegantly that this discrepancy can be explained almost entirely by how each study operationalized the ‘use of alcohol in hazardous situations’ criterion. They further observe that 3.3% (3.9% of alcohol users) of those who did not have a DSM-IV diagnosis were diagnosed with moderate DSM-5 AUDs, while 19.6% of those diagnosed with DSM-IV abuse/dependence no longer met criteria for DSM-5 AUDs. Neither the exclusion of the legal problems criterion nor the addition of a craving criterion made a major contribution to these diagnostic switches. Instead, the biggest contributor came again from the ‘use of alcohol in hazardous situations’ criterion. Nine out of every 10 individuals who switched from having a DSM-IV AUD to not having a DSM-5 AUD did so solely because of the ‘use in hazardous situations’ criterion. It seems imbalanced that a single criterion should carry such weight when estimating the prevalence of AUDs. These findings add to the growing body of evidence casting doubt on the utility of the ‘use in hazardous situations’ criterion [3–6].
On a broader level, the two studies by Agrawal et al. and Mewton et al. show that simple differences in how diagnostic interviews operationalize diagnostic criteria for a given disorder can have profound effects on the prevalence of that disorder. The success with which a given diagnostic interview question (or set of questions) can represent a criterion accurately depends to a large extent on the absence of ambiguity in the wording and intent of the criterion itself. Criteria sets with vague wording or ambiguously defined constructs open up the potential for diverse interpretations of the same putative concept, possibly resulting in differences between diagnostic interviews not only in content but also in intent of questions. Drawing upon the fields of cognitive psychology and survey methodology, Willis & Lessler  have pioneered a standardized methodology (known as the Questionnaire Appraisal System or QAS-99) to facilitate a direct critique of the clarity, understandability and potential assumptions that might exist in the wording and intent of interview questions. Applying the QAS-99 to the draft diagnostic criteria would be a novel application of a rigorous methodology that could identify and remedy problems with draft criteria before they are set for the next generation.
Overall, the two papers indicate that new criteria enable a better diagnostic framework for the diagnosis of AUDs. The milder end of the spectrum is more difficult to pin down, and the future criteria will benefit from further clarification regarding treatment need. It is assumed that the purpose of prevalence estimation is to develop needs assessment and intervention planning. The real test of new and different measures is how they match up and assist the process of assessing and intervening at different levels of problems. It is not entirely clear that our current classifications help assessment or have any leverage in predicting outcome; the real challenge is to work towards a future where criteria shape the longer-term treatment outcomes of those diagnosed with alcohol use disorders.
No potential or actual conflicts of interest. This manuscript was supported by an Australian National Health and Medical Research Council project grant (no. 630414).
Declarations of interest
- 71999., Question Appraisal System Manual. Raleigh, NC: Research Triangle Institute.