Primary objective: Several published meta-analyses indicate that mild traumatic brain injury (MTBI) is associated with a favourable course of recovery over a period of days-to-weeks, with no indication of permanent impairment on neuropsychological testing by 3 months post-injury in group studies. These meta-analyses provide important but not definitive information relating to outcome from MTBI in individual patients. The purpose of this paper was to illustrate that a sub-group of patients with residual cognitive deficits could exist, yet be obscured using group inferential statistics.
Main outcome and results: A sample of 30 concussed amateur athletes and a hypothetical sample of 30 adults who had sustained MTBIs were used to illustrate these statistical issues. In both groups, a minority of subjects with residual cognitive deficits were not identified using group statistics.
Conclusions: It is important to appreciate that MTBI meta-analyses represent an aggregation of effect sizes derived from multiple groups across multiple studies. Therefore, this methodology could, theoretically, obscure small sub-group or individual effects. Implications for interpreting meta-analyses are discussed.