Substance use in individuals with mild to borderline intellectual disability: A comparison between self-report, collateral-report and biomarker analysis.

BACKGROUND AND AIMS: Individuals with mild or borderline intellectual disability (MBID) are at risk of substance use (SU). At present, it is unclear which strategy is the best for assessing SU in individuals with MBID. This study compares three strategies, namely self-report, collateral-report, and biomarker analysis. METHODS AND PROCEDURES: In a sample of 112 participants with MBID from six Dutch facilities providing care to individuals with intellectual disabilities, willingness to participate, SU rates, and agreement between the three strategies were explored. The Substance use and misuse in Intellectual Disability - Questionnaire (SumID-Q; self-report) assesses lifetime use, use in the previous month, and recent use of tobacco, alcohol, cannabis, and stimulants. The Substance use and misuse in Intellectual Disability - Collateral-report questionnaire (SumID-CR; collateral-report) assesses staff members' report of participants' SU over the same reference periods as the SumID-Q. Biomarkers for SU, such as cotinine (metabolite of nicotine), ethanol, tetrahydrocannabinol (THC), and its metabolite THCCOOH, benzoylecgonine (metabolite of cocaine), and amphetamines were assessed in urine, hair, and sweat patches. RESULTS: Willingness to provide biomarker samples was significantly lower compared to willingness to complete the SumID-Q (p<0.001). Most participants reported smoking, drinking alcohol, and using cannabis at least once in their lives, and about a fifth had ever used stimulants. Collateralreported lifetime use was significantly lower. However, self-reported past month and recent SU rates did not differ significantly from the rates from collateral-reports or biomarkers, with the exception of lower alcohol use rates found in biomarker analysis. The agreement between self-report and biomarker analysis was substantial (kappas 0.60-0.89), except for alcohol use (kappa 0.06). Disagreement between SumID-Q and biomarkers concerned mainly over-reporting of the SumID-Q. The agreement between SumID-CR and biomarker analysis was moderate to substantial (kappas 0.48 - 0.88), again with the exception of alcohol (kappa 0.02). CONCLUSIONS AND IMPLICATIONS: In this study, the three strategies that were used to assess SU in individuals with MBID differed significantly in participation rates, but not in SU rates. Several explanations for the better-than-expected performance of self- and collateral-reports are presented. We conclude that for individuals with MBID, self-report combined with collateralreport can be used to assess current SU, and this combination may contribute to collaborative, early intervention efforts to reduce SU and its related harms in this vulnerable group.

Capture recapture estimation of the prevalence of mild intellectual disability and substance use disorder.

Persons with mild to borderline intellectual disability (MID) have been identified as a group at risk for substance use disorder (SUD). However, prevalence estimates of co-occurring SUD and MID rely largely on single source studies performed in selected samples. To obtain more reliable population estimates of SUD and MID, this study combines data from an Intellectual Disability Facility (IDF), and an Addiction Treatment Centre (ATC) in a semi-rural area in the Netherlands. Capture-recapture analysis was used to estimate the hidden population (i.e., the population not identified in the original samples). Further analyses were performed for age and gender stratified data. Staff members reported on 88 patients with SUD and MID in the IDF (4.0% of the IDF sample) and 114 in the ATC (5.2% of the ATC sample), with 12 patients in both groups. Only strata for males over 30 years provided reliable population estimates. Based on 97 patients in these strata, the hidden population was estimated at 215. Hence the estimated total population of males over 30 years old with MID and SUD was 312 (95% CI 143-481), approximately 0.16% (0.05-0.25%) of the total population of this age and gender group. This illustrates that while patients with co-occurring SUD and MID often receive professional help from only one service provider, single source data underestimate its prevalence, and thus underestimate treatment and service needs. Therefore, population prevalence estimations of co-occurring SUD and MID should be based on combined multiple source data.