Table 3

Odds ratios with 95% credible intervals of infection control strategies based on the random effect MTC model

	No systemic antibiotics	Plain cement	Conventional ventilation	Systemic antibiotics	Antibiotic-impregnated cement	Laminar airflow	Body exhaust suit	OR	95% Credible interval
T1	✓	✓	✓	×	×	×	×	Referent
T6	×	×	✓	✓	✓	×	×	0.13	0.03–0.35
T2	×	✓	✓	✓	×	×	×	0.31	0.12–0.65
T3	✓	✓	×	×	×	✓	×	0.26	0.03–0.95
T4	×	✓	×	✓	×	✓	×	0.25	0.06–0.66
T7	×	×	×	✓	✓	✓	×	0.27	0.03–0.93
T5	✓	×	✓	×	✓	×	×	0.38	0.09–1.12
T8	×	×	✓	✓	✓	×	✓	0.52	0.03–2.12
T9	×	×	×	✓	✓	✓	✓	0.74	0.05–2.69
T2 vs T5	–	–	–	–	–	–	–	1.28	0.38–3.38
T6 vs T7	–	–	–	–	–	–	–	1.96	0.52–5.37
Model fit statistic (posterior mean residual deviance) 34.3*			Model fit statistic (DIC) 180.6			Heterogeneity (between-study SD 0.63

*Compared with 32 data points (model fit is considered to be adequate if the posterior mean residual deviance is approximately equal to the number of total data points; see online supplementary appendix 7).
Note: T1–T9: nine infection control strategies. Refer to table 1 for details.
✓: The strategy contains the infection control measure as indicated by the column heading.
×: The strategy does not contain the infection control measure as indicated by the column heading.
DIC, deviance information criterion; MTC, mixed treatment comparison.