This study included 30 children with ADHD aged 7.0 to 11.0 and 30 matched controls. It also included 38 adults with ADHD aged 19 to 63 and 38 matched controls. Patients came from a university ADHD clinic in Germany.
Adults rated themselves and were rated by an observer. Parents rated the children. Everyone also did computer attention tasks, one of which tracked movement. A computer model then sorted each person into the ADHD or control group.
With ratings and tests combined, the model sorted 89.5% of adults correctly. For children, it sorted 86.7% correctly. The Conners rating scores were the most useful items in both age groups. With only the objective tests, accuracy fell to 79% in adults. In children it fell to 78%.
In adults, the most useful test results reflected inattention, such as missed targets and uneven reaction times. In children, they reflected hyperactivity, such as how much the child moved. Ratings and test results rarely lined up.
Each group was fairly small. Teacher ratings were left out because too much was missing. Children with ADHD had lower IQ scores than controls. People with ADHD were compared only with people without ADHD, not with people who have other conditions with similar symptoms.
The numbers Stats and what the terms mean
The statistics behind the plain-language claims above.
- Emser, 2018: how well ratings and tests combined picked out adults with and without ADHD, sensitivity = .90, specificity = .90.
- Emser, 2018: how well the objective tests alone picked out adults with and without ADHD, sensitivity = .82, specificity = .76.
- Emser, 2018: how well ratings and tests combined picked out children with and without ADHD, sensitivity = .83, specificity = .90.
- Emser, 2018: how well the objective tests alone picked out children with and without ADHD, sensitivity = .80, specificity = .77. What the terms mean:
- Sensitivity is the share of people with ADHD that the model correctly identified as having ADHD, where 1 would mean all of them.
- Specificity is the share of people without ADHD that the model correctly identified as not having ADHD, where 1 would mean all of them.
Source: Theresa S. Emser, Blair Johnston, J. Douglas Steele, et al.. Assessing ADHD symptoms in children and adults: evaluating the role of objective measures. Behavioral and Brain Functions (2018). doi:10.1186/s12993-018-0143-x