Skip to main content

Table 2 Missing data handling during development and external validation

From: Clinical prognostic models for sarcomas: a systematic review and critical appraisal of development and validation studies

 

Overall N = 241

Development only n = 182

External validation only n = 59

Missing data reporting

 None

178 (74.2%)

138 (75.8%)

40 (67.8%)

 Overall

40 (16.7%)

32 (17.6%)

8 (13.6%)

 Detaila

23 (9.6%)

12 (6.6%)

11 (18.6%)

Missing data method

 Complete case

161 (67.1%)

129 (70.9%)

32 (54.2%)

 Multiple imputation

6 (2.5%)

6 (3.3%)

0

 Unclear

73 (30.4%)

47 (25.8%)

27 (45.8%)

 Assumed mechanism of missingness reported

1 (0.4%)

1 (0.5%)

0

  1. aMissing data was provided for each variable