. Author manuscript; available in PMC: 2015 Oct 27.

Published in final edited form as: Stat Med. 2009 Jul 20;28(16):2095–2115. doi: 10.1002/sim.3599

Table V.

Mockup accuracy for the CVD model.

		Using incident counts for the UKPDS		Using the UKPDS risk equation
	Mockup probability values	Absolute Error	Relative Error	Absolute Error	Relative Error
$\hat{p}$ ₀₁	0.005	−8.8E–12	−1.8E–09	−1.2E–11	−2.5E–09
$\hat{p}$ ₀₂	0.03	1.85E–11	6.16E–10	−3.8E–11	−1.3E–09
$\hat{p}$ ₁₂	0.05	4.29E–10	8.58E–09	2.3E–10	4.6E–09
$\hat{p}$ ₁₄	0.05	3.48E–09	6.96E–08	1.08E–09	2.16E–08
$\hat{p}$ ₂₃	0.8	8.32E–09	1.04E–08	−3.3E–09	−4.2E–09
$\hat{p}$ ₃₂	0.1	−3.1E–10	−3.1E–09	−2.2E–09	−2.2E–08
$\hat{p}$ ₃₄	0.08	2.3E–09	2.88E–08	−1.1E–10	−1.4E–09

The first two columns represent the model input probabilities which are also the expected estimation results. The following columns show the accuracy of the estimation engine from these expected results for both models.