Yackobovitch-Gavan et al. [1] executed a three-arm parallel randomized controlled trial (RCT). Participants were randomly assigned to one of three groups: a control group or one of two behavioral interventions, towards either both parents and children, or parents only. The “primary outcome measures were the change in BMI-SDS [BMI-standard deviation score] from baseline at completion of the intervention (3 months) and at the end of follow-up (24 months) in each of the groups.” In parallel RCTs, the correct analyses for inferences about treatment effects compare outcomes among randomized groups [2–5]. Yackobovitch-Gavan et al. did these analyses (“comparison of the change in BMI-SDS after 3 months among the groups yielded no significant difference (P = .440)”) with similarly non-significant among-group findings at 24 months (P = 0.208). Therefore, the correct conclusion to draw for this study is that the data did not show that either of the treatment interventions was effective with respect to the primary outcomes.
Instead, the abstract concludes that “[a]n intervention programme that focuses on both parents and children was found to have positive short-term and long-term effects on BMI-SDS”. The analytic strategy used to reach this conclusion, a DINS (difference in nominal significance) error, can result in an inflated Type-I error rate higher than 50% with more than two groups [3–5]; its invalidity has been noted repeatedly in the literature [3–7].
Because the primary conclusions communicated by Yackobovitch-Gavan and colleagues are not substantiated by the data, we believe that the paper should be corrected.
Acknowledgements:
Supported in part by NIH grants R25DK099080 and R25HL124208. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health or any other organization.
Footnotes
Competing interests.
None pertinent.
References
- 1.Yackobovitch-Gavan M, Wolf Linhard D, Nagelberg N et al. Intervention for childhood obesity based on parents only or parents and child compared with follow-up alone. Ped Obes 2018; 10.1111/ijpo.12263 [DOI] [PubMed] [Google Scholar]
- 2.David Moher, Sally Hopewell, F Schulz Kenneth et al. CONSORT 2010 Explanation and Elaboration: updated guidelines for reporting parallel group randomised trials BMJ 2010; 340 :c869. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Bland JM, & Altman DG (2011). Comparisons against baseline within randomised groups are often used and can be highly misleading. Trials, 12(1), 264 10.1186/1745-6215-12-264 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Bland JM, Altman DG. Best (but oft forgotten) practices: testing for treatment effects in randomized trials by separate analyses of changes from baseline in each group is a misleading approach. Am J Clin Nutr 102: 991–994, 2015. doi: 10.3945/ajcn.115.119768. [DOI] [PubMed] [Google Scholar]
- 5.George BJ, Beasley TM, Brown AW et al. (2016), Common scientific and statistical errors in obesity research. Obesity, 24: 781–790. doi: 10.1002/oby.21449 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Allison DB, Brown AW, George BJ et al. Reproducibility: A tragedy of errors. Nature(2016). Available at: http://www.nature.com/news/reproducibility-a-tragedy-of-errors-1.19264. (Accessed: 11th August 2017) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Gelman A & Stern H (2006). The Difference Between “Significant” and “Not Significant” is not Itself Statistically Significant. The American Statistician, 60(4), 328–331. 10.1198/000313006X152649 [DOI] [Google Scholar]