Table 6. Balancing achieved by different propensity score-based matching strategies for macrolide use.

Covariate	Macrolide	No macrolide	Before matching		Matching within studyes, without replacement*		Matching within studies, with replacement*		Matching across studies without replacement*		Matching across studies with replacement*
	Mean (sd)	Mean (sd)	Std dif # (%)	p-value+	Std dif (%)	p-value+	Std dif (%)	p-value	Std dif (%)	p-value+	Std dif (%)	p-value+
Age	38.75 (13.11)	39.13 (13.45)	1.02	0.6192	-3.57	0.628	-8.11	<0.001	-7.28	0.31	-7.68	<0.001
Male gender	0.64 (0.48)	0.68 (0.47)	-8.7	0.127	0.00	1.0000	8.64	<.0001	-1.56	0.8273	4.72	0.0021
Extensive disease	0.79 (0.41)	0.72 (0.45)	15.8	0.008	6.64	0.4054	-3.93	0.0174	-18.03	0.0116	-0.42	0.7973
Prior TB	0.57 (0.5)	0.7 (0.46)	-26	0	7.40	0.3458	-11.72	<.0001	7.72	0.2752	16.47	<.0001
Prior MDR-TB	0.26 (0.44)	0.08 (0.27)	47.8	0	-11.00	0.1573	-2.44	0.1159	-7.87	0.2850	-9.11	<.0001
Known HIV co-infection	0.04 (0.19)	0.17 (0.38)	-44.7	0	-4.73	0.3173	0.24	0.7389	18.08	0.0196	6.29	0.0002
No. drugs (intensive)	6.44 (1.69)	5.12 (1.33)	-30.55	<0.0001	-12.91	0.059	-31.10	<0.001	-10.03	0.028	-5.89	<0.001
Total therapy (months)	24.38 (16.32)	17.55 (8.28)	-18.68	<0.0001	-3.49	0.337	-7.70	0.337	-14.69	0.04	-5.53	<0.001
Logit of predicted propensity score	3.95 (1.6)	2.86 (1.28)	-26.45	<0.0001	12.48	0.06	9.01	<0.001	-6.37	0.118	-0.03	0.542

*1:1 matching.

^# Std dif = Standardized difference for unexposed relative to exposed. The differences in characteristics between exposed and unexposed individuals receiving were tested for significance using a paired t-test for continuous variables and McNemar’s test for dichotomous variables [34].

MDR-TB = multi-drug resistant tuberculosis. TB = tuberculosis. HIV = human immunodeficiency virus.