Article type
Year
Abstract
Background: More so than efficacy events, safety events can be quite rare. Asymptotic approximations are known to be imprecise or even invalid in the presence of rare events. Nevertheless, the great majority of systematic reviews continue to use normal approximations regardless of the distribution of the data.
Objectives: Establish and describe a sample of Cochrane and non-Cochrane meta-analyses of safety data. Determine which factors (i.e. event rates, number of studies etc.) are most important in addressing when exact methods are preferable to asymptotic methods.
Methods: A large sample of Cochrane (n=500) and non-Cochrane (n=200) systematic reviews was randomly selected and a database of safety meta-analyses established. Point estimates and confidence intervals for each meta-analysis were recalculated using exact methods and compared to the results of the asymptotic methods used in the systematic review.
Results: Cochrane reviews were nearly four times as likely as non-Cochrane reviews to contain meta-analyses of safety data (35% compared to 9%). More than 50% of safety meta-analyses contained an outcome with a rare event rate (< 5%). For rare event metaanalyses, exact point estimates differed substantially from asymptotic estimates 38% of the time, compared to 17% for those without rare events. Exact confidence intervals differed substantially from asymptotic ones 67% of the time compared to only 19% for those without rare events. The magnitude of differences was also correlated with the number of studies and the summary statistic used to combine the data.
Conclusions: Asymptotic methods will not always be a good approximation for exact methods in safety meta-analyses. Event rates and the number of studies should be closely examined when choosing the statistical method for combining rare event data. This project was funded by a grant from the Canadian Co-ordinating Office for Health Technology Assessments (CCOHTA) now called Canadian Agency for Drugs and Technologies in Health (CADTH).
Objectives: Establish and describe a sample of Cochrane and non-Cochrane meta-analyses of safety data. Determine which factors (i.e. event rates, number of studies etc.) are most important in addressing when exact methods are preferable to asymptotic methods.
Methods: A large sample of Cochrane (n=500) and non-Cochrane (n=200) systematic reviews was randomly selected and a database of safety meta-analyses established. Point estimates and confidence intervals for each meta-analysis were recalculated using exact methods and compared to the results of the asymptotic methods used in the systematic review.
Results: Cochrane reviews were nearly four times as likely as non-Cochrane reviews to contain meta-analyses of safety data (35% compared to 9%). More than 50% of safety meta-analyses contained an outcome with a rare event rate (< 5%). For rare event metaanalyses, exact point estimates differed substantially from asymptotic estimates 38% of the time, compared to 17% for those without rare events. Exact confidence intervals differed substantially from asymptotic ones 67% of the time compared to only 19% for those without rare events. The magnitude of differences was also correlated with the number of studies and the summary statistic used to combine the data.
Conclusions: Asymptotic methods will not always be a good approximation for exact methods in safety meta-analyses. Event rates and the number of studies should be closely examined when choosing the statistical method for combining rare event data. This project was funded by a grant from the Canadian Co-ordinating Office for Health Technology Assessments (CCOHTA) now called Canadian Agency for Drugs and Technologies in Health (CADTH).