Errors in assessing DSM-IV substance use disorders
Archives of General Psychiatry (Impact Factor: 14.48). 04/2007; 64(3):379-80; author reply 381-2. DOI: 10.1001/archpsyc.64.3.379
[Show abstract] [Hide abstract]
- "The combination of abuse and dependence into a single diagnosis reflected substantial evidence from factor and item response theory (IRT) analyses supporting a single latent construct underlying all DSM-IV SUDs, with abuse and dependence criteria always intermixed along the severity spectrum (e.g., Compton et al., 2009; Mewton et al., 2011; Saha et al., 2006, 2007, 2010, 2012; Shmulewitz et al., 2010). Additional considerations included the well-documented reliability (Hasin et al., 2006a) and validity (Compton et al., 2007; Dawson et al., 2000; Grant et al., 2004; Hasin et al., 2007) of dependence but less desirable psychometric properties of abuse. Moreover, whereas a syndrome by definition requires more than one symptom, almost half of all abuse cases received a diagnosis by meeting only one criterion, typically hazardous use (Hasin et al., 1999; Hasin & Paykin, 1999). "
ABSTRACT: The purpose of this study was to examine prevalences and concordances between Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV), and Fifth Edition (DSM-5) substance use disorders (SUDs) in a newly completed U.S. epidemiologic survey. The National Epidemiologic Survey on Alcohol and Related Conditions-III surveyed 36,309 civilian, noninstitutionalized adults. SUDs were assessed using the Alcohol Use Disorder and Associated Disabilities Interview Schedule-5. Concordances between DSM-IV and DSM-5 disorders were assessed using kappa statistics. Prevalences of past-year substance-specific DSM-5 disorders (2+ criteria) were modestly higher than those of DSM-IV dependence and abuse combined for alcohol, sedatives/tranquilizers, opioids, and heroin, but lower for cannabis, cocaine, and stimulants. Lifetime prevalences were lower under DSM-5. Prevalences were similar between moderate to severe (4+ criteria) DSM-5 disorders and dependence, whereas prevalences of DSM-5 disorders at 3+ criteria (DSM-5 [3+]) were higher, particularly for cannabis. Past-year concordances were excellent for DSM-IV dependence and abuse combined versus any DSM-5 and DSM-IV dependence versus DSM-5 moderate to severe disorders; lifetime concordances were fair to excellent. Past-year concordances between DSM-IV and DSM-5 (3+) were generally similar to or modestly higher than those with any DSM-5 disorder; lifetime concordances were mostly lower. Findings are consistent with those informing the development of DSM-5. Future research should examine differences in patterns between past-year and lifetime disorders, particularly for cannabis. Other questions warranting investigation include whether different combinations of the same numbers of criteria carry different clinical or nosologic implications, whether changes innosology yield changes in treatment demand, and whether changes in characteristics of individuals with DSM-5 SUDs dictate modifications to screening and intervention.Journal of studies on alcohol and drugs 05/2015; 76(3):378-388. DOI:10.15288/jsad.2015.76.378 · 2.76 Impact Factor
[Show abstract] [Hide abstract]
- "Part of the problem is that there is no such object as a " gold standard " measure of mental health. For example , while Kessler et al. (2005) provide strong evidence that the National Comorbidity Study (NCS) provide results similar to the earlier Epidemiologic Catchment Area (ECA) Study and to clinical results, Kessler et al. (2004) acknowledge that there is no " gold standard " to compare to the NCS, and Grant et al. (2007) argue that the NCS underestimates substance dependency because of the nature of the screening questions. "
ABSTRACT: I present an improved methodology for estimating local prevalence rates using classical econometric methods. I provide information on the variation within national mental health surveys associated with ICD versus DSM coding. Conditional on the validity of national survey responses, I estimate precise and statistically significant models associated with binary measures of mental health diagnoses. I also present estimates from polychotomous discrete choice allowing for covariance in errors. Focusing on binary discrete measures, empirical results from NCS-R and NSADMHP are qualitatively similar though very different from NHIS. I speculate that, to a significant degree, this occurs because both NCS-R and NSADMHP rely on popular screening tools to mechanically diagnosis sample participants, while NHIS relies on self-diagnosis. I also discuss the effects on local prevalence estimates caused by unobserved community-specific effects. Finally, I use the results to make policy statements about the provision of public mental health services in central Virginia; the results document a severe shortage of services for people who are unlikely to be able to afford services in the private market.Health Services and Outcomes Research Methodology 09/2014; 14(3). DOI:10.1007/s10742-014-0120-2
[Show abstract] [Hide abstract]
- "Consider the following example: the two most recent surveys for alcohol use disorders revealed a 12.5% prevalence of alcohol dependence in Latvia (Snikere et al., 2011) and <0.5% in Italy (de Girolamo et al., 2006, using standard survey methodology—Composite International Diagnostic Interview CIDI; see Rehm et al., 2005, 2012, for background on methodology and a European overview of prevalence rates). Even if we allow for methodological problems with the Italian estimate, as it was based on the World Mental Health survey (Grant et al., 2007), which at the time asked dependence questions only if the respondent had previously answered affirmatively to the abuse criteria, other evidence from Italy would suggest prevalence of alcohol dependence at most 1/5 to 1/10 that of Latvia (Scafato et al., 2005). If we look at the per capita consumption of the two countries for 2010, we find a recorded consumption of 6.1 for Italy and 9.8 for Latvia, with unrecorded consumption being estimated at an additional 1.4 for Italy and 1.8 for Latvia (http ://www.who.int/substance_abuse/activities/gisah/en/). "
ABSTRACT: Aims: The aim of the study was to explore whether the concept of heavy substance use over time can be used as definition of substance use disorder. Methods: Narrative review. Results: Heavy use over time clearly underlies the neurobiological changes associated with current thinking of substance use disorders. In addition, there is evidence that heavy use over time can explain the majority of social problems and of burden of disease (morbidity and mortality). A definition of substance use disorders via heavy use over time would avoid some of the problems of current conceptualizations, for instance the cultural specificity of concepts such as loss of control. Finally, stressing the continuum of use may avoid the high level of stigmatization currently associated with substance use disorders. Conclusion: 'Heavy substance use over time' seems to be a definition of substance use disorders in line with results of basic research and epidemiology. Additionally, it reduces stigmatization. This approach should thus be further explored.Alcohol and Alcoholism 08/2013; 48(6):633–640. DOI:10.1093/alcalc/agt127 · 2.89 Impact Factor
Data provided are for informational purposes only. Although carefully collected, accuracy cannot be guaranteed. The impact factor represents a rough estimation of the journal's impact factor and does not reflect the actual current impact factor. Publisher conditions are provided by RoMEO. Differing provisions from the publisher's actual policy or licence agreement may be applicable.