Table 2 Results of prediction for the first and validation cohort.

From: Early identification of postpartum depression using demographic, clinical, and digital phenotyping

 

HC vs. PPD

HC vs. AD

PPD vs. AD

Cross-validation (1st cohort)

Out-of-sample validation

Chance (95% CI)

Cross-validation (1st cohort)

Out-of-sample validation

Chance (95% CI)

Cross-validation (1st cohort)

Out-of-sample validation

Chance (95% CI)

Balanced accuracy

0.87

0.93

0.50 (0.44; 0.59)

0.91

0.79

0.50 (0.45; 0.57)

0.76

0.73

0.50 (0.35; 0.67)

Sensitivity

0.85

0.88

0.11 (0.00; 0.27)

0.88

0.60

0.12 (0.03; 0.23)

0.76

0.56

0.28 (0.07; 0.50)

Specificity

0.89

0.99

0.90 (0.89; 0.92)

0.94

0.98

0.88 (0.86; 0.90)

0.76

0.90

0.73 (0.63; 0.83)

PPV

0.45

0.88

0.10 (0.00; 0.25)

0.66

0.86

0.18 (0.05; 0.33)

0.66

0.75

0.32 (0.08; 0.58)

NPV

0.98

0.99

0.91 (0.89; 0.92)

0.98

0.92

0.83 (0.81; 0.85)

0.79

0.79

0.68 (0.59; 0.78)

AUC

0.91

0.98

0.50 (0.34; 0.65)

0.97

0.92

0.50 (0.39; 0.62)

0.79

0.88

0.50 (0.31; 0.68)

  1. Balanced accuracy, sensitivity, specificity, positive and negative predictive value obtained for the validation cohort (out-of-sample validation). For comparison, results of the first cohort aside with chance-level classification for the validation cohort are also reported.
  2. AD adjustment disorder, AUC area under the curve, CI confidence interval, HC healthy controls, PPD postpartum depression, PPV positive predictive value, NPV negative predictive value.
Back to article page