Table 3.

Between-study heterogeneity

Study no	Study name	Baseline risk	Linear predictor (=predicted absolute risks)		Membership C
		Admission proportion	Mean	SD
1	ISCOPE	0.23	−1.27	−0.46	0.84
2	Opti-Med	0.16	−1.71	−0.28	0.69
4	PRIMUM	0.16	−1.72	−0.52	0.80
5	RIME	0.22	−1.35	−0.33	0.80

Heterogeneity between original studies is described in terms of baseline risk (proportion of participants with hospital admissions), casemix distribution with respect to predicted risks, and the discriminative ability of the membership model to identify membership of a specific study.