Table 5.

The pooled results of subgroup analysis of the exercise-based cardiac rehabilitation effect on all-cause hospital admissions for patients after PCI or CABG

Subgroups	Sample size	Pooled results			Heterogeneity
		RR [95% CI]	P-value		I2	Analytical effect model
Operation
CABG	242	0.48 [0.20, 1.16]	0.10		0%	Fixed-effect model
PCI	869	0.71 [0.36, 1.37]	0.30		11%	Fixed-effect model
PCI + CABG	415	0.52 [0.18, 1.48]	0.22		7%	Fixed-effect model
Follow-up time
6 to 12 months	1022	0.50 [0.31, 0.82]	0.006		50%	Random-effect model
12 to 36 months	504	0.98 [0.77, 1.23]	0.83		0%	Fixed-effect model
Total duration
≤6 months	357	0.52 [0.36, 0.75]	0.0004		0%	Fixed-effect model
6 to 12 months	442	0.52 [0.16, 1.74]	0.29		44%	Random-effect model
Male
>80%	984	0.63 [0.40, 1.00]	0.05		61%	Random-effect model
<80%	319	0.86 [0.54, 1.35]	0.58		0%	Fixed-effect model
CR components
Exercise	500	0.41 [0.13, 1.27]	0.12		45%	Random-effect model
Exercise + education	599	0.65 [0.41, 1.02]	0.06		50%	Random-effect model
Exercise + education + psychological support	223	0.63 [0.18, 2.16]	0.46		0%	Fixed-effect model

CABG, coronary artery bypass graft; PCI, percutaneous coronary interventions; CR: cardiac rehabilitation; RR, relative risk; CI, confidence interval.