Table 2 The comparison of pooled performance measures in subgroups

Machine learning(n)

 LR (8)

0.77 (0.73–0.81)

0.54 (0.37–0.70)

0.82 (0.72–0.89)

3.06 (2.19–4.28)

0.56 (0.41–0.76)

5.47 (3.35–8.95)

 Boosting (7)

0.84 (0.81–0.87)

0.68 (0.57–0.78)

0.85 (0.77–0.90)

4.44 (3.15–6.27)

0.37 (0.28–0.50)

11.86 (7.80–18.01)

 RF (6)

0.75 (0.71–0.78)

0.67 (0.59–0.73)

0.74 (0.65–0.81)

2.54 (1.82–3.54)

0.45 (0.36–0.57)

5.59 (3.29–9.49)

Sample size (n)

  < 10000 (7)

0.68 (0.64–0.72)

0.51 (0.36–0.65)

0.77 (0.63–0.87)

2.24 (1.61–3.11)

0.64 (0.53–0.77)

3.52 (2.44–5.08)

  > 10000 (16)

0.82 (0.78–0.85)

0.64 (0.56–0.72)

0.83 (0.78–0.87)

3.71 (3.00–4.58)

0.43 (0.35–0.53)

8.62 (6.25–11.89)

Age group(n)

 Children (4)

0.72 (0.67–0.75)

0.59 (0.40–0.76)

0.75 (0.55–0.88)

2.33 (1.54–3.51)

0.55 (0.41–0.74)

4.23 (2.72–6.57)

 Children_adults(6)

0.88 (0.84–0.90)

0.53 (0.37–0.68)

0.89 (0.86–0.92)

5.02 (4.05–6.22)

0.53 (0.38–0.73)

9.49 (5.83–15.44)

 Adults (13)

0.79 (0.75–0.82)

0.65 (0.55–0.74)

0.78 (0.72–0.83)

2.95 (2.32–3.75)

0.45 (0.35–0.57)

6.56 (4.37–9.86)

Outcome(n)

 ED/HP (15)

0.81 (0.77–0.84)

0.60 (0.49–0.70)

0.84 (0.78–0.88)

3.65 (2.86–4.65)

0.48 (0.38–0.61)

7.58 (5.35–10.74)

 AE (8)

0.78 (0.74–0.81)

0.64 (0.52–0.74)

0.77 (0.70–0.84)

2.82 (2.06–3.86)

0.47 (0.35–0.63)

6.01 (3.47–10.41)