Table 2: Assessment of attrition bias risk.

First author Journal Year Vol. Worst-case scenario
First page DS Test group Amalgam group LTF adjusted effect estimate Original effect estimate Bias risk
LTF N = BSL teeth n+LTF LTF N+LT n
RR 95% CI p RR 95% CI p
Sachdeo [31] EurJProsthodontRestDent 2004 12 15 CA01 no LTF apparent during trial No
Soncini [30] JADA 2007 138 763 CA02 nil Yes
Bernado [2] JADA 2007 138 775 CA04 no LTF apparent during trial No
Li [27] PractClinMed 2005 6 105 GA02 nil Yes
Frencken [28] JDR 2006 85 622 GA07 18 52 21 10 33 10 1.33 0.72 - 2.46 0.36 0.20 0.06 - 0.66 0.008* Yes
GA10 127 288 135 105 218 5 20.44 8.52 - 49.03 <0.00001** 1.12 0.38 - 3.34 0.83 Yes
Estupiñán-Day [29] PAHO-report/Panama 2006 GA17 144 769 166 118 677 8 18.27 9.03 - 36.86 <0.00001** 2.46 1.10 - 5.48 0.03** No
First author Journal Year Vol Best-case scenario
First page DS Test group Amalgam group LTF adjusted effect estimate Original effect estimate Bias risk
LTF N+LTF teeth n LTF N = BSL n+LTF
RR 95% CI p RR 95% CI p
Sachdeo [31] EurJProsthodontRestDent 2004 12 15 CA01 no LTF apparent during trial No
Soncini [30] JADA 2007 138 763 CA02 nil Yes
Bernado [2] JADA 2007 138 775 CA04 no LTF apparent during trial No
Li [27] PractClinMed 2005 6 105 GA02 nil Yes
Frencken [28] JDR 2006 85 622 GA07 18 52 3 10 33 20 0.10 0.03 - 0.30 <0.00001* 0.20 0.06 - 0.66 0.008* No
GA10 127 288 8 105 218 110 0.06 0.03 - 0.11 <0.00001* 1.12 0.38 - 3.34 0.83 Yes
Estupiñán-Day [29] PAHO-report/Panama 2006 GA17 144 769 22 118 677 126 0.15 0.10 - 0.24 <0.00001* 2.46 1.10 - 5.48 0.03** Yes

LTF = Number of restorations lost to follow-up; Vol. = Journal volume; DS = Dataset number; N = Number of restorations evaluated; BSL = Number of restorations at baseline; n = Number of failed restorations; RR = Risk ratio; CI = Confidence interval; EurJProsthodontRestDent = European Journal of Prosthodontic Dentistry; JADA = Journal of the American Dental Association; PractClinMed = Practice Clinical Medicine; JDR = Journal of Dental Research; PAHO = Pan-American Health Organisation; *Difference statistically significant in favour of test group; ** Difference statistically significant in favour of control group.