Table 2 Results of meta-analyses
Outcome | Number of studies | Summary statistic | Confidence interval | I 2 statistic (%) |
|---|---|---|---|---|
RR = relative risk | ||||
SMD = standard mean difference | ||||
Patient outcomes | ||||
 Dichotomous patient outcomes | 5 | RR = 1.00 | 1.00–1.00 | 0 |
 Dichotomous patient outcomes (sensitivity analysis with Shah 2014 removed) | 4 | RR = 1.09 | 0.91–1.29 | 0 |
Physician behaviour outcomes | ||||
 Continuous physician behaviour outcomes (N = physicians) | 4 | SMD = 0.35 | −0.06–0.76 | 67.88 |
 Continuous physician behaviour outcomes (N = patients) | 3 | SMD = 0.27 | −0.03–0.57 | 68.74 |
 Dichotomous physician behaviour outcomes (N = physicians) | 3 | RR = 1.01 | 0.96–1.07 | 0 |
 Dichotomous physician behaviour outcomes (N = patients) | 9 | RR = 0.99 | 0.97–1.01 | 0 |
 Dichotomous physician behaviour outcomes (PEM vs. workshop) | 2 | RR = 0.57 | 0.12–2.73 | 64.48 |
Physician cognition outcomes | ||||
 Continuous physician cognition outcomes | 3 | SMD = 0.65 | −0.21–1.51 | 89.75 |
 Dichotomous cognition outcomes | 3 | RR = 1.51 | 0.90–2.52 | 90.27 |