Attention-deficit hyperactivity disorder in people with intellectual disability: Statistical approach to developing a bespoke screening tool

Background 

Attention-deficit hyperactivity disorder (ADHD) is common among people with intellectual disability. Diagnosing ADHD in this clinically and cognitively complex and diverse group is difficult, given the overlapping psychiatric and behavioural presentations. Underdiagnoses and misdiagnoses leading to irrational polypharmacy and worse health and social outcomes are common. Diagnostic interviews exist, but are cumbersome and not in regular clinical use. Aims We aimed to develop a screening tool to help identify people with intellectual disability and ADHD. 

Method 

A prospective cross-sectional study, using STROBE guidance, invited all carers of people with intellectual disability aged 18-50 years open to the review of the psychiatric team in a single UK intellectual disability service (catchment population: 150 000). A ten-item questionnaire based on the DSM-V ADHD criteria was circulated. All respondents' baseline clinical characteristics were recorded, and the DIVA-5-ID was administered blinded to the individual questionnaire result. Fisher exact and multiple logistic regressions were conducted to identify relevant questionnaire items and the combinations that afforded best sensitivity and specificity for predicting ADHD. 

Results 

Of 78 people invited, 39 responded (26 men, 13 women), of whom 30 had moderate-to-profound intellectual disability and 38 had associated comorbidities and on were medication, including 22 on psychotropics. Thirty-six screened positive for ADHD, and 24 were diagnosed (16 men, eight women). Analysis showed two positive responses on three specific questions to have 88% sensitivity and 87% specificity, and be the best predictor of ADHD. 

Conclusions 

The three-question screening is an important development for identifying ADHD in people with intellectual disability. It needs larger-scale replication to generate generalisable results.