Table 2.

Summary of the calculated minimum required sample size and the difference between observed and calculated estimates where the observed value is: the sample size used to develop the prediction model, or the total available sample size. Values are median [25th and 75th percentiles]

	Total required sample size	Total number of events	Events per predictor parameter
Calculated sample size: Criteria 1 - small overfitting defined by an expected shrinkage of predictor effects by 10% or less (n = 94)	965 [487 to 2268]	188 [101 to 343]	7.8 [4.2 to 14.7]
Calculated sample size: Criteria 2 - small absolute difference of 0.05 in the model’s apparent and adjusted Nagelkerke’s R2 value (n = 94)	673 [496 to 1095]	122 [83 to 213]	5.1 [3.9 to 7.6]
Calculated sample size: Criteria 3 - precise estimation (within +/- 0.05) of the average outcome risk (n = 94)	241 [150 to 331]	47 [16 to 108]	1.9 [0.6 to 5.2]
Minimum required sample size to meet all criteria (n = 94)	971 [543 to 2268]	211 [114 to 360]	9.0 [4.9 to 15.1]
Difference between minimum required sample size to meet all criteria and sample size that was used to develop the model	-387 [-1207 to 49], n = 94	-75 [-234 to 6], n = 94	-3.4 [-8.4 to 0.4], n = 89
Difference between minimum required sample size to meet all criteria and sample size that was available to develop the model	-360 [-1207 to 72], n = 94	-63 [-225 to 7], n = 93	-3.0 [-7.7 to 0.4], n = 88
Difference between minimum required sample size to meet criteria 3 and sample size that was used to develop the model	262 [-1 to 1165], n = 94	46 [0 to 165], n = 94	1.6 [0 to 6.8], n = 89
Difference between minimum required sample size to meet criteria 3 and sample size that was available to develop the model	373 [14 to 1369], n = 94	50 [2 to 210], n = 93	1.9 [0.1 to 9.1], n = 88