CPT Accuracy for ADHD: What the Meta-Analyses Show

Continuous performance test (CPT) accuracy for ADHD, pooled across commercially available tests, comes out to roughly 0.75 sensitivity and 0.71 specificity for the overall score — below the 0.80/0.80 threshold researchers generally consider acceptable for a standalone diagnostic tool. That’s the headline finding from the largest recent meta-analysis on the topic, and it’s consistent with what earlier and more recent reviews have found. This is honest, useful information, not a reason to discard CPTs — it’s the evidence base for exactly how to use one: as one input into a fuller evaluation, not a substitute for it.

Photo by Vitaly Gariev on Unsplash

What do sensitivity and specificity actually mean here?

Sensitivity is the proportion of people with ADHD that a test correctly flags as positive. Specificity is the proportion of people without ADHD that a test correctly flags as negative. A test with 0.75 sensitivity and 0.71 specificity means that, in the pooled research samples, it missed about a quarter of true ADHD cases and false-flagged about three in ten people without ADHD. Both numbers matter for a clinical tool: low sensitivity means real cases get missed, and low specificity means non-cases get an unnecessary positive result. Neither number, on its own, tells you whether a specific patient in front of you has ADHD — that’s a probability estimate calibrated to the study population, not a certainty about an individual.

What the largest recent meta-analysis found

A 2024 meta-analysis in the Journal of the American Academy of Child & Adolescent Psychiatry (JAACAP), led by Arrondo and colleagues, searched four databases through January 2023 and identified 19 studies using commercially available CPTs, combining results from up to 835 controls and 819 cases in its sensitivity/specificity analyses (and up to 996 cases and 1,083 controls in a separate area-under-the-curve analysis) [1]. The pooled results:

The authors’ conclusion is direct: “None of the measures met minimal standards for acceptable sensitivity (0.8) and specificity (0.8).” At the clinical level, they write, CPTs “as a stand-alone tool have only a modest to moderate ability to differentiate ADHD from non-ADHD samples” [1].

Two things stand out in that table. First, the composite total/ADHD score outperforms either individual error type (omissions or commissions) on its own — a reminder that the summary score, not a single subscale, is the more informative number. Second, even the best-performing metric falls short of the 0.80/0.80 bar the authors used as their acceptability threshold. For how omission and commission errors map onto inattentive versus impulsive symptom presentations, see omission vs. commission errors in ADHD testing.

What product-specific reviews add

The JAACAP meta-analysis pools across products; a few narrower reviews look at specific tests or populations, and the pattern holds.

A 2024 systematic review focused specifically on the Conners CPT-3 screened 1,480 title/abstracts down to 13 included studies and organized findings into five themes. The authors report that five of those studies found the CPT-3 was “a weak or poor predictor” of an ADHD diagnosis on its own, while two found it “adequate,” and two others found it useful for distinguishing ADHD from comorbid anxiety or from obsessive-compulsive disorder rather than for diagnosis in isolation [2]. Their overall conclusion: as a standalone measure, the CPT-3 is a weak-to-poor predictor of ADHD.

A separate 2024 systematic review looked specifically at CPT utility in adult ADHD evaluations across multiple commercial products and concluded that CPTs “should not be used in isolation as a diagnostic test but may be beneficial when used as a component of a comprehensive assessment” [3]. Adult evaluations are a harder case for CPTs generally: base rates of comorbidity are higher, and the tests were originally normed and validated more heavily in pediatric samples.

Accuracy isn’t the same as clinical utility

It’s worth separating two different questions: “how accurately does this test classify ADHD on its own?” and “does giving this test change what happens in the visit?” The evidence above answers the first question modestly. A 2018 single-blind randomized controlled trial of QbTest (a different, hardware-based CPT) in 267 children and young people referred for ADHD assessment in England found that clinicians who received the QbTest report were more likely to reach a diagnostic decision at a given appointment — 76% had reached a decision at 6 months, versus 50% in the control arm — and appointment length fell by 15%, with “no difference in diagnostic accuracy” between the two arms [4]. That trial measured decision-making efficiency and consistency, not the CPT’s standalone sensitivity and specificity, and it studied a specific product in a specific pediatric NHS pathway. It’s a useful reminder that objective data can speed up and standardize a visit even when the underlying test, used alone, wouldn’t meet a diagnostic accuracy bar.

Try it with your next patient. AxonCPT gives you 2 free tests every month — no card required. Send a link, your patient finishes in ~16 minutes from home, and you get a norm-referenced report instantly. Create your free account →

What this means for how you actually use a CPT

None of this is a reason to skip objective testing — it’s a reason to use it correctly. A few practical implications follow directly from the evidence above:

• Don’t rule ADHD in or out on a CPT score alone. With a pooled specificity around 0.71–0.74 for individual metrics, roughly a quarter to a third of people without ADHD will still produce an abnormal result on any given metric. Weigh the CPT against the clinical interview, developmental history, and standardized rating scales, as covered in ADHD rating scales vs. objective testing.
• Treat a normal CPT result as reassuring, not disqualifying, and an abnormal result as a prompt for closer clinical review, not a diagnosis.
• Use repeat testing to track change, not to re-diagnose. The accuracy figures above describe single-administration diagnostic classification. Comparing a patient’s own scores across visits — before and after a dose change, for example — is a different and often more clinically useful application, since the patient serves as their own baseline. See using objective data in ADHD medication titration.
• Expect variability between CPT products and versions. The reviews above pool across different commercial tests with different norms and metrics; a specific product’s accuracy in your population may differ from the pooled average.

How this looks in practice

A psychologist building a diagnostic battery, or a psychiatrist reviewing a treatment response, typically applies this evidence the same way regardless of which CPT they use:

1. Order the CPT as one component of the evaluation, alongside a structured interview and validated rating scales, not as the sole basis for a decision.
2. Interpret an abnormal or normal result in context. A single elevated commission-error score, for instance, is weaker evidence on its own than a convergent pattern across the interview, history, and rating scales.
3. For monitoring, compare the same patient’s scores over time rather than treating each test as a fresh diagnostic classification. With objective ADHD testing with AxonCPT, each report breaks vigilance into quarters across the session and is built to be re-administered at follow-up visits, so a clinician can track a trend rather than a single snapshot.
4. Document how the result was weighed, not just what the number was — that’s what makes the chart defensible regardless of which way the score points.

You can see what the report looks like to judge how a CPT report fits into that kind of longitudinal review. For the broader question of what a CPT is and where it fits in an assessment, start with what is a continuous performance test? and for a step-by-step guide to reading the metrics on an actual report, see how to read a CPT report: a clinician walkthrough.

Frequently asked questions

Can a CPT diagnose ADHD by itself? No. Pooled sensitivity and specificity for commercially available CPTs fall short of the 0.80/0.80 threshold researchers use as a rough acceptability bar, and every major review cited here explicitly recommends against using a CPT in isolation [1][2][3].

Why do different studies report different accuracy numbers for the same test? Sample composition (age, comorbidity, referral source), the comparison group used, and which specific metric is analyzed (total score versus omissions versus commissions) all shift the reported sensitivity and specificity. That’s part of why pooled meta-analyses, which combine many studies, are more informative than any single study.

Are commission errors or omission errors more accurate for diagnosing ADHD? Neither performed as well as the composite total/ADHD score in the JAACAP pooled analysis. Omission errors had somewhat higher specificity (0.74) but lower sensitivity (0.63); commission errors were weaker on both dimensions [1]. See omission vs. commission errors in ADHD testing for what each error type reflects clinically.

If accuracy is modest, why give a CPT at all? Because an evaluation is a combination of several imperfect measures, not one perfect one. A CPT adds objective, real-time performance data that a rating scale or interview can’t provide on its own, and it’s especially useful for tracking change over time during treatment. See ADHD rating scales vs. objective testing for how the two data types complement each other.

Does test accuracy differ between children and adults? The literature is thinner and more mixed for adults. A 2024 systematic review focused on adult CPT use similarly concluded CPTs shouldn’t be used in isolation, and adult evaluations carry additional considerations — including the possibility of symptom exaggeration — covered in detecting feigned ADHD: symptom and performance validity.

Start testing your patients with AxonCPT

Objective testing works best as part of a full evaluation, not a replacement for one. AxonCPT gives you a standardized, norm-referenced measure to add to your interview and rating scales, with 2 free tests every month and no card required. Create your free AxonCPT account or learn more at axoncpt.com.

References

1. Arrondo G, Mulraney M, Iturmendi-Sabater I, et al. Systematic Review and Meta-Analysis: Clinical Utility of Continuous Performance Tests for the Identification of Attention-Deficit/Hyperactivity Disorder. J Am Acad Child Adolesc Psychiatry. 2024;63(2):154-171.
2. Callan PD, Swanberg S, Weber SK, Eidnes K, Pope TM, Shepler D. Diagnostic Utility of Conners Continuous Performance Test-3 for Attention-Deficit/Hyperactivity Disorder: A Systematic Review. J Atten Disord. 2024;28(6):992-1007.
3. Varela JL, Magnante AT, Miskey HM, et al. A systematic review of the utility of continuous performance tests among adults with ADHD. Clin Neuropsychol. 2024;38(7):1524-1585.
4. Hollis C, Hall CL, Guo B, et al. The impact of a computerised test of attention and activity (QbTest) on diagnostic decision-making in children and young people with suspected ADHD: single-blind randomised controlled trial. J Child Psychol Psychiatry. 2018;59(12):1298-1308.