Table 3. Overall sensitivity and specificity of the observation-based classifier computed against the ADOS-G and ADOS-2 algorithm scores and best estimate clinical diagnosis.
ADOS-G |
ADOS-2 |
Clinical diagnosis |
||||
---|---|---|---|---|---|---|
Spectrum | Non-spectrum | Spectrum | Non-spectrum | Spectrum | Non-spectrum | |
OBC | ||||||
Spectrum | 2280 | 64 | 2304 | 40 | 766 | 3 |
Non-spectrum | 53 | 219 | 69 | 203 | 23 | 15 |
Sensitivity | 0.977 | 0.971 | 0.971 | |||
Specificity | 0.774 | 0.835 | 0.833 | |||
Accuracy | 0.955 | 0.958 | 0.968 | |||
PPV | 0.973 | 0.983 | 0.996 | |||
NPV | 0.805 | 0.746 | 0.395a |
Abbreviations: ADOS-G, Autism Diagnostic Observation Schedule-Generic; NPV, negative predictive value; OBC, observation-based classifier; PPV, positive predictive value. Sensitivity was comparable across the three baselines for comparison; specificity was highest (>83%) when compared with ADOS-2 and best estimate clinical diagnosis.
Based on limited number of non-spectrum individuals.