Skip to main content
. 2013 Jan 25;8(1):e54865. doi: 10.1371/journal.pone.0054865

Table 3. Validity of verbal autopsy based assignment of any cause* of neonatal deaths compared with the cause of death assigned by the treating neonatologist based on clinical and laboratory information.

Diagnosis Category Sensitivity Specificity Positive Predictive Value Negative Predictive Value Diagnostic Accuracy Area under ROC Kappa Agreement
N % N % % % % Proportion Agreement
(95% CI) (95% CI) (95% CI) (95% CI) (95% CI) (95% CI) (95% CI)
Congenital Anomaly 36 61.1 277 95.7 68.8 95.0 92.3 0.78 0.58
(43.5–76.4) (92.4–97.6) (49.9–83.3) (91.6–97.1) (88.8–95.0) (0.71–0.87) (0.44–0.73)
Preterm birth 87 81.6 226 69.0 50.4 90.7 72.5 0.75 0.43
(71.6–88.8) (62.5–74.9) (41.9–58.8) (85.1–94.4) (67.2–77.4) (0.70–0.80) (0.33–0.57)
Asphyxia 85 54.1 228 89.0 64.8 83.9 79.6 0.72 0.45
(43.0–64.9) (84.1–92.6) (52.5–75.5) (78.5–88.2) (74.6–83.9) (0.66–0.77) (0.34–0.57)
Severe infection (pneumonia, sepsis or meningitis) 91 78.0 222 66.2 48.6 88.0 69.6 0.72 0.37
(67.9–85.8) (59.5–72.3) (40.3–57.0) (81.9–92.4) (64.2–74-7) (0.67–0.77) (0.28–0.47)