Year: 2015
Type: letter
Abstract: Potential conflict of interest: Nothing to report. We thank Wei and colleagues for their interest in our evidence synthesis.1 In general, most of our analyses were associated with low heterogeneity; therefore, we believe that there is no significant impact of heterogeneity on our certainty in the estimates or conclusions. We do not believe this affected the applicability or generalizability of our results. We agree with Wei and colleagues that further knowledge about subgroup interactions and effect modifiers like duration of treatment, dose, and associated morbidities will be helpful. However, such analyses require individual patient data. Because there is only a small number of available trials, performing such exploration using meta‐regression techniques is not feasible and can potentially lead to misleading and spurious findings if performed using aggregate estimates.2 We also certainly would not recommend pursuing post hoc sensitivity analyses without a clear rationale. Unfortunately, we do not believe that there is a particular “model” or statistical technique that can remedy this issue.
Source: Hepatology
DOI: 10.1002/hep.28412
Language: English
Subfield: Statistics, Probability and Uncertainty
Field: Decision Sciences
Domain: Social Sciences
Open Access status: bronze
OpenAlex ID: W2278579758