Original ArticleA meta-regression analysis shows no impact of design characteristics on outcome in trials on tension-type headaches
Section snippets
Objective
Systematic reviews and meta-analysis are designed to help the clinician base their clinical decisions on the best available evidence. In the conduct of a systematic review or meta-analysis, many possible sources of bias exist, such as publication bias [1], [2], language bias [3], [4], and bias caused by design characteristics [5], but also clinical heterogeneity. All these sources of variation possibly have an influence on the reported outcome.
One can assess design components such as concealed
Study selection
For this study, we selected randomized clinical trials out of a large systematic review, which fulfilled the following criteria: (1) including a control group with placebo treatment/no treatment or waiting list controls; (2) presenting sufficient data (means and measures of variability or number of patients successful on treatment outcome). For the crossover trials, we summed the outcomes over all active treatments periods and over all control periods, as none of these trials provided separate
Study selection
We divided the original data set of 146 trials into four main categories according to the intervention: acute pain medication (n = 41), preventive medication (n = 36), physiotherapy interventions (n = 8), and behavioral interventions (n = 43). One category of trials (n = 11) concerned children with TTH. Five trials were included in two categories (preventive medication and physiotherapy or behavioral interventions) and in total seven trials did not fit in either one of the categories and were excluded,
Discussion
Meta-regression evaluates whether certain factors, in this case design characteristics, explain heterogeneity of treatment effect between trials. We did not find any significant association between design characteristics as measured with the Delphi list and the effect estimate.
Conclusion
In this study, sample design characteristics do not show to have an impact on treatment effect estimates, but the way the treatment effect is measured has a significant impact. Whether assessment of design characteristics should or should not be a part of the design of systematic reviews and meta-analysis is still undecided. Bias by design characteristics cannot be ruled out, but the direction is unclear.
Acknowledgment
Two authors (A.P.V., T.S.) had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
References (27)
- et al.
Language bias in randomised controlled trials published in English and German
Lancet
(1997) - et al.
The art of quality assessment of RCTs for systematic reviews
J Clin Epidemiol
(2001) - et al.
Does quality of reports of randomised trials affect estimates of intervention efficacy reported in meta-analyses?
Lancet
(1998) - et al.
Meta-regression detected associations between heterogeneous treatment effects and study-level, but not patient level, factors
J Clin Epidemiol
(2004) - et al.
The effectiveness of physiotherapy and manipulations in patients with tension-type headache: a systematic review
Pain
(2004) - et al.
The Delphi list: a criteria list for quality assessment of randomised clinical trials for conducting systematic reviews developed by Delphi consensus
J Clin Epidemiol
(1998) - et al.
Publication bias in the medical literature: a review by a Canadian Research Ethics Board
Can J Anaesth
(2007) Assessing the implications of publication bias for two popular estimates of between-study variance in meta-analysis
Biometrics
(2007)- et al.
Direction and impact of language bias in meta-analyses of controlled trials: empirical study
Int J Epidemiol
(2002) - et al.
Empirical evidence of bias; dimensions of methodological quality associated with estimates of treatment effects in controlled trials
JAMA
(1995)