Skip to main content
. 2015 Mar 20;34(13):2081–2103. doi: 10.1002/sim.6471

Table 4.

Results of the internal–external cross‐validation procedure for the temperature data.

Study Observed no. with fever (O) Observed prevalence Summary prevalence from meta‐analysis model (8), Summary sensitivity from meta‐analysis model (1), Summary specificity from meta‐analysis model (1), Option A Post‐test predictions obtained using summary sensitivity, summary specificity and summary prevalence (Equations (6) and (7)) Option B Post‐test predictions obtained using summary sensitivity, summary specificity and observed prevalence (Equations (9) and (10)) Option C Post‐test predictions obtained after bivariate meta‐regression of PPV and NPV on prevalence (Equations (12) and (13))
Predicted PPV Predicted NPV Expected no. with fever ^ (E) O/E Predicted PPV Predicted NPV Expected no. with fever ^ (E) O/E Predicted PPV Predicted NPV Expected no. with fever ^(E) O/E
Brennan 203 0.55 0.45 0.644 0.99 0.97 0.77 207.59 0.98 0.98 0.69 226.11 0.90 0.98 0.75 211.83 0.96
Davis 18 0.27 0.48 0.661 0.98 0.97 0.76 24.41 0.74 0.94 0.89 16.85 1.07 0.96 0.78 22.71 0.79
Green 9 0.43 0.46 0.631 0.98 0.97 0.76 10.92 0.82 0.97 0.78 10.58 0.85 0.96 0.76 10.80 0.83
Greenes 109 0.36 0.46 0.67 0.98 0.97 0.77 110.91 0.98 0.96 0.83 94.12 1.16 0.95 0.78 106.11 1.03
Hoffman 42 0.42 0.47 0.645 0.99 0.97 0.77 47.23 0.89 0.96 0.80 44.56 0.94 0.97 0.76 47.06 0.89
Hooker 15 0.38 0.46 0.65 0.98 0.97 0.76 16.63 0.90 0.96 0.82 14.97 1.00 0.96 0.78 16.07 0.93
Lanham 103 0.58 0.47 0.666 0.98 0.97 0.76 81.98 1.26 0.98 0.67 94.01 1.10 0.98 0.76 82.71 1.25
Muma 87 0.39 0.46 0.662 0.98 0.97 0.78 84.73 1.03 0.96 0.83 75.98 1.15 0.95 0.78 84.66 1.03
Nypaver 425 0.48 0.45 0.652 0.98 0.97 0.78 408.91 1.04 0.97 0.76 426.06 1.00 0.97 0.76 421.87 1.01
Rhoads 27 0.42 0.47 0.683 0.98 0.96 0.77 20.25 1.33 0.96 0.80 18.12 1.49 0.96 0.77 19.78 1.36
Stewart 59 0.75 0.43 0.65 0.98 0.97 0.78 59.88 0.99 0.99 0.49 67.79 0.87 0.98 0.56 65.61 0.90

Obtained from meta‐an the study in the row.

^

Expected number for the study in the row, based on using Equation (14) and the positive predictive value (PPV) and negative predictive value (NPV) in the prior two columns.