Table 3. Operating Characteristics of the DSM-IV ASRS Screening Scale Optimally Scored With RiskSLIM to Predict DSM-5/ACDS ADHD.
| Score Threshold | Predicted Prevalence, % | Sensitivity, %a | Specificity, %b | AUCc | PPV, %d |
|---|---|---|---|---|---|
| Pooled NCS-R and managed care development samplese | |||||
| ≥9 vs 0-8 | 32.9 | 94.1 | 72.6 | .83 | 23.5 |
| ≥10 vs 0-9 | 24.9 | 89.7 | 80.9 | .85 | 29.5 |
| ≥11 vs 0-10 | 16.6 | 84.2 | 89.5 | .87 | 41.7 |
| ≥12 vs 0-11 | 6.0 | 57.9 | 98.7 | .78 | 79.6 |
| ≥13 vs 0-12 | 4.3 | 44.4 | 99.3 | .72 | 84.7 |
| NYU Langone validation samplef | |||||
| ≥9 vs 0-8 | 63.3 | 90.2 | 72.2 | .81 | 82.1 |
| ≥10 vs 0-9 | 56.3 | 85.0 | 82.7 | .84 | 87.0 |
| ≥11 vs 0-10 | 50.3 | 79.8 | 89.8 | .85 | 91.4 |
| ≥12 vs 0-11 | 45.3 | 74.6 | 94.5 | .85 | 94.8 |
| ≥13 vs 0-12 | 34.0 | 58.4 | 99.2 | .79 | 99.0 |
Abbreviations: ADHD, attention-deficit/hyperactivity disorder; ASRS, Adult ADHD Self-Report Scale; AUC, area under the curve; NCS-R, National Comorbidity Survey Replication; NYU Langone, New York University Langone Medical Center Adult ADHD Program; PPV, positive predictive value.
The proportion of DSM-5 and Adult Clinical ADHD Diagnostic Scale (ACDS) cases classified as cases by the screener at the threshold.
The proportion of DSM-5/ACDS noncases classified as noncases by the screener at the threshold.
The probability that a randomly selected DSM-5/ACDS case would have a higher screening scale score than a randomly selected noncase at the threshold, with ties on the screening scale assigned a predicted probability of 0.500.
The proportion of respondents with screening scale scores above the threshold that meet DSM-5/ACDS criteria for the disorder.
Sample of 337 participants; weighted prevalence, 8.2%.
Sample of 300; participants; prevalence, 57.7%.