Year: 2024
Type: letter
Abstract: We appreciate Siddiqi and colleagues' comments, which enrich the discussion concerning cytisine [1]. Below we provide a point-by-point response. As detailed in our work, the exclusion of some randomized controlled trials (RCTs) was established within the framework of the sensitivity analysis [2]. A sensitivity analysis is a repetition of the primary analysis of the meta-analysis, substituting alternative decisions or ranges of values [3]. It is recommended, and desirable, that these decisions be previously specified in the protocol a priori [4]. This was the case for most of our sensitivity analyses. However, sometimes one cannot anticipate the many issues suitable for sensitivity analysis that are only identified during the review process, where the specific characteristics of the design and development of each study included are recognized [4, 5]. As recommended in the Cochrane Handbook (section 10.10.3), it is advisable to perform this analysis with and without some studies as part of the strategies for addressing heterogeneity [3]. The studies excluded in the sensitivity analysis are one of the valid strategies for the treatment of heterogeneity, and their selection was based on purely methodological criteria [5, 6]. As we detail in our study, the cited trial that was included in the primary analysis, with a low risk of bias, was the one that contributed the greatest weight to our results, and was mentioned in the Conclusions and Discussion sections [2, 7]. The initial judgement of low risk of bias refers to the individual study, while the subsequent work refers to the inconsistency introduced in the remaining set of studies in the context of the sensitivity analysis, given their peculiarities. Such a 'one-leave-out' approach can tell us whether a particular study heavily influences the summary effect and the heterogeneity [6, 7]. When exploring the causes of heterogeneity, the methodological characteristics of this pragmatic trial differed substantially from the rest of the trials included [7]. This particular approach is not performed with the intention of explaining the causes of the results of the excluded study, but instead evaluates the summary results of the remaining trials and their heterogeneity [5, 7]. Regarding adherence to cytisine treatment, the assertions in our study only reflect what was detailed by the authors of this particular trial, and are not inferences made explanatory by our research team [6]. Adherence to cytisine treatment was not assessed in the remaining trials because we were performing a sensitivity analysis and not a subgroup analysis, which was unnecessary under this approach [3-5]. Finally, we have reported the results of our exhaustive sensitivity analysis with multiple approaches not as alternative conclusions, but rather as an example of the degree of confidence that our primary analysis represents the true estimate. There are no funders to report. None to declare.
Source: Addiction
DOI: 10.1111/add.16451
Institution: Hospital Posadas
Language: English
Subfield: Statistics, Probability and Uncertainty
Field: Decision Sciences
Domain: Social Sciences
Open Access status: bronze
OpenAlex ID: W4391817484