A practical solution to estimate the sample size required for clinical prediction models generated from observational research on data

European Radiology Experimental

Table 3 Results of sample size calculations

Model	Riley’s sample size	10 EPP	5 EPP	Riley’s EPP
NB 5-year mortality risk	1111	978	490	11.37
NB relapse/progression risk	1168	1166	584	10.03
NB time to death	1397	978	490	7.00
HR-NB 5-year mortality risk	1043	600	300	17.38
HR-NB 5-year relapse/progression risk	1060	736	369	14.42
HR-NB time to relapse/progression	1060	536	268	11.23
DIPG 1-year mortality risk	1043	668	334	15.65
DIPG 2-year mortality risk	1345	1776	889	7.58
DIPG 1-year progression risk	1208	1278	639	9.46
DIPG time to death	1273	1776	889	7.87
DIPG time to progression	1130	1278	639	9.67

Sample size estimated by Riley’s methodology, the 10 EPP “rule of thumb” and 5 EPP. The number of events per predictor (EPP) variable derived from the sample size obtained with Riley’s methodology was also calculated. EPP event per predictor parameter