Skip to main content

Advertisement

Table 3 Factors influencing the adherence to each PRISMA item adjusted for type of review

From: Quality of reporting of systematic reviews and meta-analyses in emergency medicine based on the PRISMA statement

  Total adherence n = 112 (100%) PRISMA mention vs. No PRISMA mention in article p – valuea Journal requiring PRISMA adherence vs. no journal requirement p - valuea
1. Title 93 (83) 0.053 0.607
2. Structured summary 104 (93) 0.887 0.192
3. Rationale 112 (100) NA NA
4. Objectives 100 (89) 0.796 0.118
5. Protocol and registration 21 (19) 0.128 0.954
6. Eligibility criteria 104 (93) 0.622b 0.542
7. Information sources 96 (86) 0.301 0.571
8. Search 59 (53) 0.270 0.091
9. Study selection 90 (80) 0.205 0.0202
10. Data collection process 64 (57) 0.431 0.130
11. Data items 75 (67) 0.701b 0.909
12. Risk of bias in individual studies 87 (78) 0.352 0.027b.2
13. Summary measures 63 (56) 0.851 0.960
14. Synthesis of results 79 (71) 0.908 0.116
15. Risk of bias across studies 29 (26) 0.554 0.450
16. Additional analyses 39 (64) 0.900 0.473
17. Study selection 29 (26) 0.002b.1 0.297
18. Study characteristics 99 (88) 0.936 0.0392
19. Risk of bias within studies 63 (57) 0.989 0.006b.2
20. Results of individual studies 96 (86) 0.357 0.0152
21. Synthesis of results 110 (98) 0.657 0.997b
22. Risk of bias across studies^ 29 (27) 0.791 0.797
23. Additional analysis˟ 32 (55) 0.666b 0.706b
24. Summary of evidence 110 (98) 0.872b 0.997
25. Limitations 96 (86) 0.540 0.991
26. Conclusions 84 (75) 0.406 0.611
27. Funding 87 (78) 0.137 < 0.0012
  1. NA not applicable, PRISMA preferred reporting items for systematic reviews and meta-analysis
  2. 51 missing cases
  3. 1 missing case
  4. ^4 missing cases
  5. ˟54 missing cases
  6. 1Articles with no PRISMA mention had better reporting
  7. 2Articles published in journal requiring PRISMA adherence had better reporting
  8. aCochran-Mantel-Haenzal test used with controlling for type of study (systematic review or meta-analysis)
  9. bLogistic regression used with controlling for type of study if Breslow-Day had p-value > 0.05