Table 3.
Empirical example statistics for illustration: Meta-analyses for systolic blood pressure outcomes across three reviews.
|
Glynn et al., 2010
|
Bray et al., 2010
|
Fletcher et al., 2015
|
|||
|---|---|---|---|---|---|
| Citation | WMD (95% CI) | Citation | WMD (95% CI) | Citation | WMD (95% CI) |
| Artinian (2001) | −26.00 (−40.38, −11.62) | Artinian (2001)bc | −25.60 (−41.78, −9.42) | Missing | |
| Bailey (1998)* | 5.00 (−4.72, 14.72) | Bailey (1999) bc | 5.00 (−6.07, 16.07) | Missing | |
| NI a | Baque (2005) | −0.14 (−2.05, 1.77) | Missing | ||
| EXCL - no usual care | Binstock (1988) | −18.00 (−27.13, −8.87) | Missing | ||
| NA | NA | Bove (2013) | 30 (−9.58, 0.98) | ||
| EXCL - study design issues | Broege (2001) | −2.00 (−16.33, 12.33) | Missing | ||
| Carnaham (1975) | −7.50 (−14.18, −0.82) | Carnaham (1975) bc | −7.50 (−14.28, −0.72) | Missing | |
| NA | NA | de Souza (2012) | −8.40 (−30.60, 13.80) | ||
| NA | NA | Fikri-benbrahim (2013) | −4.70 (−8.17, − 1.23) | ||
| Freidman (1996) | −0.40 (−3.88, 3.08) | Freidman (1996) bc | −0.90 (−4.98, 3.18) | Missing | |
| NI | Green a (2008) | −3.40 (−5.91, −0.89) | Missing | ||
| NI | Green b (2008) | −9.30 (−11.80, −6.80) | Missing | ||
| NA | NA | Green (2014) | −2.50 (−8.14, 3.14) | ||
| Halme (2005) | −3.20 (−7.25, 0.85) | Halme (2005) | −3.10 (−7.93, 1.73) | Missing | |
| NA | NA | Hosseininasab (2014) | 1.00 (−0.84, 2.84) | ||
| NA | NA | Magid (2011) | −6.00 (−10.35, −1.65) | ||
| NI | Marquez-Contreras (2006) | −4.60 (−9.01, −0.19) | Marquez-Contreras (2006) | −4.50 (−8.91, −0.09) | |
| McManus (2005) | −2.30 (−5.39, 0.79) | McManus (2005) c | −2.30 (−5.47, 0.87) | Missing | |
| Mehos (2000) | −10.10 (−19.82, −0.38) | Mehos (2000) | −10.10 (−20.61, 0.41) | Missing | |
| Midanik (1991) | −2.00 (−7.38, 3.38) | Midanik (1991) | −2.60 (−7.26, 2.06) | Missing | |
| NI | Mulhauser (1993) | −5.00 (−10.45, 0.45) | Missing | ||
| NI | Parati (2009) | −0.20 (−3.84, 3.44) | Missing | ||
| Rogers (2001) | −4.80 (−9.80, 0.20) | Missing | Missing | ||
| Rudd (2004) | −8.50 (−14.29, −2.71) | Rudd (2004) bc | −8.50 (−14.16, −2.84) | Rudd (2004) | −8.50 (−14.42, −2.58) |
| Soghikan (1992) | −3.30 (−6.39, −0.21) | Soghikan (1992) | −3.30 (−6.77, 0.17) | Missing | |
| NA | NA | Stewart (2014) | −5.40 (−10.65, −0.15) | ||
| NI | Verberk (2007) | 0.50 (−3.65, 4.65) | Missing | ||
| Vetter (2000) | −0.50 (−2.77, 1.77) | Vetter (2000) bc | −0.50 (−3.07, 2.07) | Missing | |
| NI | Zillich (2005) | −4.40 (−10.52, 1.72) | Missing | ||
| Overall Pooled Effect | −2.53 (−3.73, −1.34) | Overall Pooled Effect | −3.82 (−5.61, −2.03) | Overall Pooled Effect | −4.07 (−6.71, −1.43) |
Note. Rows are organized alphabetically by primary study included in each analysis. If a primary study is missing from a particular review, then EXCL = excluded; NA = not applicable because study published after review published; NI = not identified, only available when review authors reported excluded studies at full text level; otherwise, this information was “Missing.” Effect sizes and/or 95% CIs for the same outcome/trial that differ between meta-analyses appear in boldface.
WMD = Weighted mean difference effect size.
Appeared in “eligible for review” section but the only instance was in "Table 1. Quality of included randomized trials" and it did not appear in any meta-analyses.
missing r – calculated.
missing change SD – calculated