How can spine researchers improve their meta-analyses?
The meta-analysis can have a huge impact on how surgeons practice spine, but are all meta-analyses credible? How can they be improved? A new article in The Spine Journal examined just that. The researchers examined MEDLINE, EMBASE and The Cochrane Library where studies were assessed for credibility and completeness of reporting.
There were 132 eligible meta-analyses for spine surgical interventions. Here are five things to know:
1. The average credibility score was three out of seven, with the agreement for each item ranging from 0.54 to 0.83, "moderate" to "almost perfect."
2. For the clinical information:
• Questions were sensible: 95 percent
• Searches were exhaustive: 77 percent
• Bias risk assessments occurred: 69 percent
3. Only 5 percent of the meta-analyses addressed explanations for heterogeneity using a priori subgroup hypotheses.
4. There were 18 percent that presented results that were immediately clinically applicable.
5. The investigators conducted duplicate eligibility, risk of bias and data extraction assessments 35 percent of the time.
6. The overall confidence in the evidence was rated in 18 percent of the cases.
7. The factors significantly associated with higher credibility were:
• Later publication year
• Increasing journal impact factor
• Increasing number of databases
• Including randomized, controlled trials
• Including non-English studies
8. The average reporting score was 18 out of 27.
"The credibility of many current spine surgery meta-analyses is limited," concluded the study authors. Researchers can improve the meta-analyses by considering the success factors for higher credibility scores.