Article
Are the effects of homeopathy attributable to a statistical artefact?
Search Medline for
Authors
Published: | September 2, 2009 |
---|
Outline
Text
Background: Several outcome studies, mostly designed as uncontrolled cohort studies, have found that an individualised homeopathic treatment has marked beneficial health effects, be them specific or unspecific. These results however are likely to be biased by regression to the mean (RTM), a statistical artefact originating in non-random sampling from the extremes of a distribution.
Objective: To evaluate whether the observed effects of homeopathy are attributable to RTM and to estimate RTM adjusted treatment effects (specific plus placebo effects) of homeopathy.
Methods: We reanalysed quality-of-life data from a German outcome study on nearly 3.000 chronic diseased and homeopathically treated adults [1] and applied Mee and Chua´s modified t-test [2], [3] to physical and mental summary scores of the MOS-SF-36 questionnaire, hereby making the conservative assumption that the true population mean is identical to the mean of the (healthy) Germany population [4]. Estimates are presented as standardized mean differences.
Results: Overall RTM adjusted estimates of homeopathic treatment effects were 0.10 (95%CI: 0.04 to 0.15, p<0.001) in the physical and 0.26 (CI: 0.22 to 0.29, p<0.001) in the mental summary score. Treatment effects could be confirmed for most single diagnosis in the mental summary scores (migraine: p<0.001; tension-type headache: p<0.001; eczema: p<0.001; anxiety disorders: p<0.001; low back pain: p=0.007; atopic eczema: p=0.001; asthma: p<0.001; allergic rhinitis: p<0.001), but not in the physical summary scores (all p>0.05, except for allergic rhinitis: p=0.002).
Conclusions: Effects of homeopathy on quality-of-life are probably small but cannot be explained by RTM alone. Due to the uncontrolled study design they may however completely be due to placebo effects. Due to the restrictive and conservative assumption of Mee and Chua´s modified t-test, true treatment effects can be expected to be larger than presented.
References
- 1.
- Witt CM, Lüdtke R, Baur R, Willich SN. Homeopathic medical practice: Long-term results of a cohort study with 3981 patients. BMC Public Health. 2005;5:115.
- 2.
- Mee R, Chua T. Regression Toward the Mean and the Paired Sample t Test. Am Statistician. 1991;45:39-42.
- 3.
- Ostermann T, Willich SN, Lüdtke R. Regression toward the mean – a detection method for unknown population mean based on Mee and Chua's algorithm. BMC Med Res Methodol. 2008;8:52.
- 4.
- Bullinger M, Kirchberger I. SF-36 Fragebogen zum Gesundheitszustand. Göttingen: Hogrefe; 1998