Table 3 Systematic reviews of randomized oncology trials with sensitivity analysis exploring the relationship between study quality scores and effect sizes for mortality
Systematic Review | Interventions | Outcome | Quality Scale | Definition of High Quality | Effect Size (95% confidence interval) | |
|---|---|---|---|---|---|---|
| Â | Â | Â | Â | Â | All Studies (# studies) | High Quality (# studies) |
McAlister et al, 1998 (26) | allogenic blood transfusion versus autologous or leucocyte-depleted allogenic blood during cancer surgery | relative risk of death* | Jadad (22) | score ≥3 out of 5 | RR, 0.94 (0.76 to 1.16) (n = 5) | RR, 0.84 (0.47 to 1.52) (n = 2) |
Caubet et al, 1997 (27) | nonsteroidal anti-androgens (plus LHRH or orchiectomy) versus LHRH or orchiectomy alone for advanced prostate cancer | relative risk of death* | Chalmers (31) | score ≥50 % of total possible score | RR, 0.81 (0.70 to 0.94) (n = 13) | RR, 0.78 (0.66 to 0.92) (n = 4) |
Dube et al, 1997 (28) | adjuvant chemotherapy versus control for colorectal cancer | odds ratio for death* | Chalmers (31) | score >50 % of total possible score | OR, 0.82 (0.77 to 0.89) (n = 29) | OR, 0.77 (0.71 to 0.85) (n = 14) |
Detsky et al, 1992 (29) | total parenteral nutrition versus control in cancer patients undergoing chemotherapy | odds ratio for survival** | Chalmers (31) | score >42 % of total possible score; quality score also used as a weighting factor in meta analysis | OR, 0.74 (0.42 to 1.3) (n = 8) | OR, 0.69 (0.38 to 1.3) (n = 2) weighted OR, 0.61 (0.23 to 1.6) |
Klein et al, 1986 (30) | total parenteral nutrition versus control in cancer patients undergoing surgery | odds ratio for operative death* | Developed specifically for the systematic review | quality score used as a weighting factor in meta analysis | OR, 0.44 (0.21 to 0.90, p = 0.02) (n = 10) | weighted OR not reported but p = 0.07 after weighting for study quality |