REVIEW OF: Harrison et al, J Atten Disord 2023;27(12):1343–1359

STUDY TYPE: Systematic review

When a patient presents with a positive score on an ADHD screener, how much weight should we give the test? Overreliance on self-report may lead to overdiagnosis of adult ADHD. This study aimed to describe the psychometric properties of commonly used ADHD screening tools to help clinicians interpret results. 

The authors conducted a systematic review, including 20 studies of several self-report measures, such as the Adult ADHD Self-Report Scale. They looked at sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) to see how well the tests separated ADHD from both healthy adults and symptomatic non-ADHD adults. False positives are expected with screening tools, which are designed to be more sensitive (that is, catching those who need further evaluation) than specific (that is, identifying those who don’t). 

Surprisingly, none of the studies used a comprehensive “gold standard” evaluation to verify the diagnosis. Many instead relied on other self-reports, often with high false positive rates. PPV scores were poor, particularly in clinical settings, where most tests had scores below 30%. In other words, a positive result for most tests had less than a coin flip’s chance of correctly identifying ADHD. Conversely, nearly every measure exhibited an NPV score of 95% or more, meaning a negative result rarely missed true ADHD, even in the presence of comorbid psychiatric conditions. This held true for most tests, except for two that focused on patients undergoing substance use treatment. Because the DSM criteria for ADHD include domains not captured by most self-reports (such as childhood onset and functional impairment), it is perhaps unsurprising that diagnosing by self-report alone is not reliable.

CARLAT TAKE 

While a negative result on an ADHD screener helps rule out the diagnosis, a positive result doesn’t tell us that the patient has ADHD. To diagnose adult ADHD, first rule out other causes. Seek ­third-party input and look for symptoms that date back to childhood (before age 12), are stable through time, and cause specific problems in multiple areas of life.