A review pools trials of nurse-led diabetes education, but the studies differ greatly in participant populations, program content, and outcome timing. What should an appraiser examine before relying on the pooled estimate?