Skip to main content
. 2020 Jul 2;10:10939. doi: 10.1038/s41598-020-67666-3

Table 1.

For random design RS with sample size of 100, we show the sample size needed to achieve a similar accuracy using an adaptive design AS.

Country Ai Num. obsv. Accuracy (%) PPV (%) Sensitivity (%) MSE (×10-4)
RS AS RS AS RS AS RS AS RS AS
Côte d’Ivoire 1 100 27 85.2 85.3 64.9 78.6 64.8 65.1 17.3 20.5
10 100 30 85.0 85.3 78.6 78.7 64.7 65.2 16.5 20.4
50 100 50 85.1 85.5 79.2 78.6 64.3 65.5 17.7 20.0
Malawi 1 100 36 81.8 81.9 80.8 80.6 59.9 59.4 14.0 14.9
10 100 40 81.6 82.0 79.5 80.5 60.8 59.7 14.1 15.0
50 100 50 82.1 82.1 80.3 80.3 61.9 60.4 13.8 14.5
Haiti 1 100 31 82.4 82.6 70.3 75.7 43.0 35.0 1.0 1.4
10 100 30 81.5 81.6 71.7 71.0 38.9 36.4 1.0 1.5
50 100 50 81.5 82.3 70.8 70.5 38.1 39.8 0.9 1.5
Philippines 1 100 7 95.2 95.2 93.7 94.4 69.7 67.7 2.5 4.3
10 100 10 95.1 95.2 94.1 93.6 68.9 68.6 2.3 5.1
50 100 50 95.2 95.6 94.6 85.0 68.9 79.8 2.7 5.5

Additional validation statistics: PPV, sensitivity and MSE are also shown. Along the rows, results are shown per country and batch size Ai.