Skip to main content

Table 1 Sensitivity and specificity analysis (clinical study; n = 635 [Sierra Leone])—logistic model

From: Overcoming the challenges of iris scanning to identify minors (1–4 years) in the real-world setting

Age Estimate (95% CI) p value vs infant identification using parent/guardian performance accuracy p value vs adult performance accuracy
Sensitivity in matching population
 1 year 0.896 (0.834–0.936) 0.0028 < 0.0001
 2 years 0.932 (0.886–0.960) < 0.0001 < 0.0001
 3 years 1.0000 (0.961–1.000) < 0.0001 1.0000a
 4 years 0.988 (0.921–0.998) 0.0013 < 0.0001
Specificity in matching population
 1 year 0.993 (0.952–0.999) 0.0002 < 0.0001
 2 years 0.948 (0.905–0.972) < 0.0001 < 0.0001
 3 years 0.946 (0.876–0.977) 0.0007 <0.0001
 4 years 0.918 (0.837–0.960) 0.0050 < 0.0001
Sensitivity in enrolled participants (instances in which the matching procedure could not be performed included as failures)
 1 year 0.563 (0.498–0.626) 1.0000 < 0.0001
 2 years 0.802 (0.744–0.849) 0.4955 < 0.0001
 3 years 0.939 (0.870–0.972) 0.0008 < 0.0001
 4 years 0.977 (0.912–0.994) 0.0005 < 0.0001
Specificity in enrolled participants (instances in which the matching procedure could not be performed included as failures)
 1 year 0.624 (0.560–0.685) 1.0000 < 0.0001
 2 years 0.815 (0.759–0.861) 0.3030 < 0.0001
 3 years 0.888 (0.809–0.937) 0.0178 < 0.0001
 4 years 0.907 (0.825–0.953) 0.0088 < 0.0001
  1. Analysis based on logistic models for sensitivity and specificity with age as covariate
  2. aExact testing was performed for sensitivity in 3-year-olds