Table 5.

Meta‐analysis results for overall calibration (O/E) when using model (16), as estimated in a frequentist or Bayesian framework.¹

Example	Predicted PPV and NPV obtained using…	Statistical framework	Summary O/E (95% CI)	$\widehat{\tau }$ (95% CI)	95% prediction interval for O/E in a new population	Probability 0.9 < O/E < 1.1 in a new population
Temperature data	Option A	Bayesian	1.02 (0.93 to 1.11)	0.10 (0.01 to 0.23)	0.78 to 1.31	0.67
	Option A	Frequentist	1.02 (0.95 to 1.10)	0.08	0.85 to 1.24	—
PTH data 1–2 h	Option B	Bayesian	1.02 (0.73 to 1.38)	0.14 (0.01 to 0.45)	0.56 to 1.74	0.40
	Option B	Frequentist	1.01 (0.79 to 1.29)	0	0.68 to 1.50	—

¹All frequentist analyses used method of moments to estimate the model.

All Bayesian analyses used a prior N(0, 1 000 000) for the mean ln(O/E) and a prior uniform(0, 0.25) for τ, with a 10 000 burn‐in followed by 100 000 samples for posterior inferences.

O/E, observed/expected; PPV, positive predictive value; NPV, negative predictive value.