Generalizing from clinical trial data: A case study. The risk of suicidality among pediatric antidepressant users

Abstract

For the results of randomized controlled clinical trials (RCTs) and related meta-analyses to be useful in practice, they must be relevant to a definable group of patients in a particular clinical setting. To the extent this is so, we say that the trial is generalizable or externally valid. Although concern about the generalizability of the results of RCTs is often discussed, there are few examples of methods for assessing the generalizability of clinical trial data. In this paper, we describe and illustrate an approach for making what we call generalizability judgments and illustrate the approach in the context of a case study of the risk of suicidality among pediatric antidepressant users.

Is the toxicity of adjuvant aromatase inhibitor therapy underestimated? Complementary information from patient-reported outcomes (PROs)

“…This might explain the difference between the reported symptom frequencies in the naturalistic after-care setting of our study and the Superscripts mark statistically significant differences in comparison with the results of the following refs: a Cella et al [23] b ATAC [3] c BIG 1-98 [4] significantly lower rates registered by Cella et al [23] who investigated patient-reported side-effects of adjuvant anastrazole therapy in a subprotocol of the ATAC trial. Although randomized-controlled clinical trials (RCTs) provide higher internal validity [27], the generalizability of results from RCTs is limited due to a lack of representativity of the study population and clinical setting [28]. Fraser et al [29] have highlighted in their communitybased population study the problems of applying clinical trial data to a generalized, non-trial population and revealed a distinct difference to the PACS 01 trial results regarding toxicity of a chemotherapeutic regime.…”

From Sample Average Treatment Effect to Population Average Treatment Effect on the Treated: Combining Experimental with Observational Studies to Estimate Population Treatment Effects

“…Previous research has proposed using non-randomized studies (NRSs) to assess whether RCT-based estimates apply to a target population (Cole and Stuart, 2010;Greenhouse et al, 2008;Kline and Tamer, 2011;Imai et al, 2008;Shadish et al, 2002;Stuart et al, 2011). A common concern is that there may be many baseline covariates, including continuous measures, which differ between the RCT and target population, and modify the treatment effect.…”

Do antidepressants make children and adolescents suicidal?

“…48 Depressed adolescents enrolled in an RCT have been estimated to have half the rate of suicidal thoughts and behaviours of depressed youth in the community. 49 There are methodological issues in pooling RCTs for a metaanalysis. The RCTs are predominantly studies in which suicidal thoughts and behaviours were not assessed prospectively.…”