Proving a negative? Methodological statistical and psychometric flaws in Ullmann et al. (2017) PTSD study

摘要

Ullmann et al. recently published a pilot study in Translational Psychiatry in which they report failing to find a statistically significant reduction in either hair cortisol or hair cortisone levels in circumcised men as compared with genitally intact (noncircumcised) men. Based on such null findings, the authors purport to have "refuted the psycho-pathological long-term effects of circumcision" and that the lack of significant results, "add to the growing body of evidence in the literature that male circumcision is not likely psychologically traumatizing across the life-span." In addition, they claim that they have proven a "healthy functionality of the LHPA axis" in men subjected to circumcision during infancy or childhood. However, it is not possible to draw any such conclusions on the basis of a null finding, especially one derived from an underpowered study in which the trend in the data suggest, if anything, that an adequately powered study may have shown the opposite of what the authors claim.Relevance for patients:When combined with other weaknesses in study design, measurement, and interpretation, it becomes apparent that the authors' conclusions are not supported by their data.