Frequently Asked Questions

In this self-referred adult sample, the online ADHD assessment had an overall accuracy of 78.0% (95% CI, 73.2%–82.2%) when compared with a virtual clinical interview used as the reference standard. Its positive predictive value was 94.9% (95% CI, 92.8%–96.3%), sensitivity was 80.6% (95% CI, 75.9%–84.8%), specificity was 44.0% (95% CI, 24.4%–65.1%), and negative predictive value was 15.1% (95% CI, 9.8%–22.6%).

These results indicate that in adults who self-referred for online ADHD evaluation, a positive online result was usually correct, while a negative result did not reliably rule out ADHD.

The online assessment was more conservative than the clinical interview in assigning ADHD-positive diagnoses. ADHD-positive results were returned in 78.8% of online assessments versus 92.8% of clinical interviews.

More than 80% of discordant cases (62 of 76) were cases in which the online assessment was ADHD-negative but the clinical interview was ADHD-positive. The authors interpreted this pattern as evidence that the online assessment did not increase false-positive diagnoses and was less likely to overcall ADHD in this help-seeking population.

No. In this study, a negative online result did not reliably rule out ADHD. The negative predictive value was 15.1% (95% CI, 9.8%–22.6%), and most discordant cases were participants who were ADHD-positive on clinical interview but ADHD-negative on the online assessment (62 of 76 cases).

The study states that when the online assessment did not diagnose ADHD, further clinical evaluation was always recommended.

When discordant cases were reviewed with full available data, the online assessment and clinical interview showed similar overall agreement with the adjudicated diagnosis, but each had different strengths. Agreement with adjudicated diagnoses was κ = 0.49 (SE = 0.06, 95% CI, 0.37–0.61, p < .001) for the online assessment and κ = 0.46 (SE = 0.05, 95% CI, 0.30–0.63, p < .001) for the clinical interview.

• The online assessment had higher PPV: 98.5% (95% CI, 96.4%–99.4%) versus 94.1% (95% CI, 92.2%–95.5%; p < .05).
• The online assessment had higher specificity: 88.2% (95% CI, 72.6%–96.7%) versus 44.1% (95% CI, 27.2%–62.1%; p < .05).
• The clinical interview had higher sensitivity: 96.8% (95% CI, 94.2%–98.5%) versus 86.2% (95% CI, 81.8%–89.8%; p < .05).

This means the online tool performed better for avoiding false positives, while the interview was better for minimizing missed ADHD cases.

In discordant cases reviewed against adjudicated diagnoses, the false-positive rate for the online assessment was 11.7% (4 of 34), which was significantly lower than the clinical interview’s 55.9% (19 of 34). The authors cited this as evidence that the online assessment was more conservative in assigning ADHD-positive diagnoses.

No ADHD presentation type was consistently misclassified or overrepresented among discordant cases. The most common presentations on both assessments were combined and inattentive, but follow-up analysis did not find evidence that disagreement was driven by systematic misclassification of a particular presentation type.

The authors concluded that discordance between methods appeared to reflect diagnostic threshold differences rather than failure to detect a specific ADHD subtype.

This was a cross-sectional study comparing a proprietary online ADHD assessment with virtual clinical interviews conducted by licensed doctorate-level clinicians applying DSM-5 criteria. Adults 19 years or older from the United States were recruited if they were seeking online resources for ADHD diagnosis, treatment, or management and had not previously been diagnosed with or referred for ADHD assessment.

Participants completed the clinical interview first, then the online assessment 1–4 weeks later. The online assessment included the Adult ADHD Self-Report Scale and other validated measures and DSM-5-aligned questions, and its results were reviewed offline by a clinician blinded to the interview results. If the two assessments disagreed, a third clinician reviewed both and assigned an adjudicated full-data diagnosis.

The findings apply primarily to adults who self-referred for online ADHD evaluation, not to the general adult population. The final sample included 345 community-dwelling US adults who completed both assessments. The sample was predominantly female (80.9%), had a mean age of about 35 years, and had high ADHD symptom burden; 92.8% were ADHD-positive on clinical interview.

Because this was a help-seeking sample with a very high prevalence of ADHD, the reported sensitivity, specificity, and predictive values should be interpreted in that context. The authors note that replication in a more general population with lower ADHD prevalence could produce different performance characteristics.

Yes. The Adult ADHD Self-Report Scale showed acceptable internal reliability in this sample, with Cronbach α = .74. The study also found significantly higher ASRS scores in the clinical interview ADHD-positive group than in the ADHD-negative group.

The main limitations are that the sample was self-referred, mostly female, and had a very high rate of ADHD, which limits generalizability. The authors state that participants were drawn from adults attracted to online ADHD advertisements and likely experiencing symptoms or impairment, so selection and response bias are possible.

Additional limitations included the fixed assessment order, with the clinical interview always preceding the online assessment; the use of virtual rather than in-person interviews, which may have missed some behavioral observations; unmeasured error in the clinical interview reference assessment; and adjudication being performed only for discordant cases rather than the full sample.