Methodological quality of randomised controlled trials comparing short-term results of laparoscopic and conventional colorectal resection

Abstract

Randomised, controlled trials (RCT) and systematic reviews of RCT with meta-analysis are considered to be of highest methodological quality and therefore are given the highest level of evidence (Ia/b). Although, "low-quality" RCT may be downgraded to level of evidence IIb, the methodological quality of each individual RCT is not respected in detail in this classification of the level of evidence. Within a systematic Cochrane Review of RCT on short-term benefits of laparoscopic or conventional colorectal resections, the methodological quality of all included RCT was evaluated. All RCT were assessed by the Evans and Pollock questionnaire (E and P increasing quality from 0-100) and the Jadad score (increasing quality from 0-5). Publications from 28 RCT printed from 1996 to 2005 were included in the analysis. Methodological quality of RCT was only moderate [E & P 55 (32-84); Jadad 2 (1-5)]. There was a significant correlation between the E & P and the Jadad score (r = 0.788; p < 0.001). Methodological quality of RCT slightly increased with increasing number of patients included (r = 0.494; p = 0.009) and year of publication (r = 0.427; p = 0.03). Meta-analysis of all RCT yielded clinically relevant differences for overall and local morbidity when compared to meta-analysis of "high-quality" (E & P > 70) RCT only. The methodological quality of reports of RCT comparing laparoscopic and open colorectal resection varies considerably. In a systematic review, methodological quality of RCT should be assessed because meta-analysis of "high-quality" RCT may yield different results than meta-analysis of all RCT.