Prevalence of substance use disorder

Transcription

1 Prevalence of substance use disorder Institut für Therapieforschung München A question of definition? Ludwig Kraus 1,2), Elena Gomes de Matos 1) & Daniela Piontek 1) 1) IFT Institut für Therapieforschung, Munich, Germany 2) Centre for Social Research on Alcohol and Drugs (SoRAD), Stockholm University, Stockholm, Sweden First European Conference on Addictive Behaviours and Dependencies September 2015, Lisbon, Portugal

2 Introduction Definitions of substance use disorders (SUD) have substantially changed over the past 40 years Recent change from DSM-IV to DSM-5 - Merging abuse and dependence into a one-dimensional scale AUD measured with DSM-IV increased by 11.8 % when measured with the corresponding DSM-5 criteria (US: Agrawal et al., 2011); AUD diagnosed under DSM-5 increased prevalence by 61.7 % (Australia: Mewton et al., 2011) Tobacco disorder based on DSM-5 much higher than measured by DSM-IV (Chung et al., 2011) Small differences between DSM-5 and DSM-IV with regard to illegal drugs (Peer et al., 2013; Proctor et al., 2012) No comparative studies on disorders related to pharmaceuticals 2

3 Introduction Consumption measures not part of the classical symptomoriented medical diagnostic systems (DSM, ICD), despite - Results from IRT analyses: consumption measures similar informative as current criteria (for alcohol Li et al., 2007; Saha et al., 2010; for cannabis Piontek et al., 2011) - Evidence that greater levels of use generally associated with higher odds of being affected by SUD (Moss, Chen & Yi, 2012; Bohn et al., 1995; Knight et al., 2002) Proposal of an alternative definition of addiction based on heavy use over time (Rehm et al., 2013, 2014) - All consequences listed as SUD criteria linked to heavy use 3

4 Introduction Almost all of what is currently conceptualized under the heading of addiction or use disorders is a consequence of heavy use over time (Rehm et al., 2013; 2014a,b) A redefinition in terms of heavy use over time better fits the epidemiological data, would reduce stigmatization and help initiate early lifestyle changes and interventions 4

5 Introduction Different definitions lead to - Differences in prevalence rates - Differences in individuals classified for SUD Associations between diagnostic criteria and consumption - Correlation, regression analysis or risk functions - Using cut-off values and assess concordance Aims - Investigating quantitative and qualitative differences in diagnoses between the two DSM versions for various substances, - Analysing the link between the current criteria definition (DSM-5) and various indicators of consumption, and - Using definitions for at risk consumers in order to assess the extent of concordance and discordance with DSM-5 diagnoses 5

6 Methods Data Epidemiological Survey of Substance Abuse (ESA) (Kraus et al., 2013) Substances - Alcohol, tobacco, cannabis, cocaine, analgesics Measures (last 12 months) - SUD symptoms according to DSM-IV and DSM-5 - Consumption measures: frequency, quantity (alcohol, tobacco), volume/day, volume/consumption day 6

7 Methods Statistical analyses - 12-month prevalence estimates for disorders according to DSM-IV and DSM-5 - Sensitivity (the proportion of DSM-5 positives to all DSM-IV positives) and specificity (the proportion of DSM-5 negatives to all DSM-IV negatives) - Bivariate and multivariate associations between substance use and number of DSM-5 criteria - Various cut-offs for consumption (dichotomous measure) Assessment of concordance between consumption groups and DSM-5 substance use diagnoses Sensitivity, specificity, PPV, NPV, correctly classified individuals 7

8 Results: DSM-IV vs. DSM-5 DSM-IV vs. DSM-5: prevalence of SUDs among users 8

9 Results: DSM-IV vs. DSM-5 Institut für Therapieforschung München Sensitivity Specificity Alcohol Tobacco Cannabis Cocaine Analgesics

10 Results: DSM-5 vs. consumption Institut für Therapieforschung München Association between number of symptoms and consumption level r =.23 r =.33 Tobacco: r =.33 to.49 Cannabis: r =.57 Cocaine: r =.39 Analgesics r =.40 10

11 Results Cut-off Prevalence above threshold a) Prevalence DSM-5 a) Correctly classified Sensitivity Specificity PPV NPV Alcohol No. of days, drinking Frequency of EHD days/week monthly Daily mean volume US) 42/ Daily mean volume Can) 27/ Daily mean volume UK) 24/ Daily mean volume Germany) 12/ Mean volume per drinking day Mean volume per drinking day 20/

12 Discussion 1) DSM-5 leads to substantial changes in estimates of SUD across substance - No clear pattern of direction of changes - Low sensitivity (alcohol, cannabis, analgesics), low specificity (tobacco) 2) Substantial associations between substance use disorder and various consumption indicators 3) Dichotomized continuous consumption measures largely fail in the identification of individuals qualifying for SU diagnoses - DSM-5 criteria identify a disorder in individuals at low levels of consumption and vice versa fail to identify individuals who consume at high levels 12

13 Discussion Empirical and conceptual considerations Identification problem - Heavy users that under DSM-5 would not qualify as in need for treatment or intervention - Individuals with a profile of low frequent consumption/low quantities that qualify for SUD Misinterpretation of alcohol-attributable consequences - Differential responses by age (Pabst et al., 2011) - Confounding of responses: overestimation of AUD in patients with major depression; depressive individuals rated symptoms according to their negative views of themselves (Baggio et al., 2014) 13

14 Discussion Epidemiological perspective - It is suggested to base risk measures of negative consequences on objective measures rather than on diagnoses which are subject to various biases Clinical perspective - A combination of consumption and diagnostic measures may be optimal - Using established screenings tests such as the Alcohol Use Disorder Identification Test (AUDIT, Babor et al., 2001) or the Fagerstöm Test for Nicotine Dependence (FTND, Heatherton et al., 1991) - Capture both groups that otherwise would be missed by either one of the approaches (false positives/ negatives) 14

15 Discussion Perspective of early intervention - A numerical approach has been proposed advantageous (Nutt & Rehm, 2014) - The proposed approach does not limit interventions to diagnoses of a disorder or illness - A continuous monitoring of frequency and amount consumed works in the same way as the monitoring of calories, cholesterol or blood pressure - The question of cut-off points may be considered as recommendations rather than as a fixed line distinguishing between ill and healthy 15

16 16

17 Thank you for your attention! Institut für Therapieforschung München

18 Introduction Kettil Bruun Dependence is a rather useless term..the term is often used in such a way that one assumes, on the basis of consequences, that dependence is at hand, which means that we generally have no indications on dependence which by definition are separate from the consequences. Therefore I will from here on principally disregard the concept of dependence (Bruun, 1973)