Twenty-three studies were included (n=22,849 patients); 61% were randomised controlled trials (RCTs). Sample size ranged from less than 100 to more than 5,000. Nine studies were classified as high quality and 14 were classified as low quality.
Overall interventions were associated with a statistically significant increase in mammography screening rates compared to control (effect size using random-effects model 0.078, 95% CI 0.043 to 0.113). Significant heterogeneity was found (p<0.001).
The intervention type with the largest effect size was access-enhancing strategies (effect size 0.155, 95% CI 0.087 to 0.223; six studies) followed by individually directed strategies (effect size 0.099, 95% CI 0.073 to 0.110; 19 studies).
Tailored therapy-based interventions were associated with a larger effect size (0.101, p<0.001; four studies) than non-tailored interventions (0.076, p<0.001; 19 studies).
Ethnically matched interventions had an effect size of 0.067 (95% CI 0.015 to 0.120; 14 studies) and culturally matched interventions had an effect size of 0.051 (95% CI 0.009 to 0.092; 15 studies).
High-quality studies had an effect size of 0.099 (95% CI 0.076 to 0.122; nine studies) and low-quality studies an effect size of 0.061 (95% CI 0.008 to 0.114, 14 studies).
The funnel plot was slightly asymmetrical, but the fail-safe N value of 411 suggested that review findings were robust.