In order to evaluate the homeopathic treatment system as a whole, meta-analyses can be carried out across different disease patterns.One problem with this approach is that both different clinical pictures and homeopathic remedies are mixed in the statistical evaluation, so that the result does not allow any generalisable conclusions, neither with regard to specific clinical pictures nor to individual homeopathic remedies or treatment strategies. It is evident that this methodological approach would never be applied in conventional treatments (e.g. summarizing and averaging trials in cardiology, oncology, psychiatry etc. all differing in researched treatments and approaches). Nevertheless, such meta-analyses have been performed and remain informative in assessing the homeopathic approach overall.
Meta-analyses across diagnoses and types of homeopathic approaches were performed for randomized controlled trials, i.e. to assess efficacy of homeopathic remedies (whether they exert specific remedy effects beyond placebo). In the following we discuss the most recent and most exhaustive meta-analyses.
Randomised controlled trials (RCTs) of individualised homeopathic treatment
Mathie et al. (2014) performed a systematic review and meta-analysis of randomised controlled trials (RCTs) of individualised homeopathic treatment. Thirty-two RCTs were eligible, covering a total of 24 medical conditions. Twenty-two trials had extractable data and were subjected to meta-analysis with a mean odds ratio of 1.53 (p<0.001). Sensitivity analysis yielded stronger treatment effects with decreasing risk of bias. Three trials with reliable evidence yielded an odds ratio of 1.98 (p=0.013). It can be concluded from this analysis that medicines prescribed in individualised homeopathy have specific treatment effects.
Randomised controlled trials (RCTs) of non-individualised homeopathic treatment
Mathie et al. (2017) performed a systematic review and meta-analysis of randomised controlled trials (RCTs) of non-individualised homeopathic treatment. Fifty-four trials had extractable data: pooled SMD was –0.33 (p<0.001), which was reduced to –0.16 (p<0.023) after adjustment for publication bias. Though three trials with reliable evidence yielded a non-significant pooled SMD: –0.18 (95% CI –0.46, 0.09), a risk of bias analysis yielded no significant difference between studies rated as ‘low’, ‘uncertain’ or ‘high’ risk of bias (p = 0.417). Sub-group analyses yielded no significant differences. Due to the conflicting results of the different evaluation approaches, the high number of included indications, and the low number of replication trials for each indication, the overall meta-analysis does not arrive at consistent and meaningful conclusions.
Randomised controlled trials (RCTs) of individualised homeopathic treatment with comparison groups other than placebo
Mathie et al. (2018) performed a systematic review and meta-analysis of randomised controlled trials of individualised homeopathic treatment with comparison groups other than placebo (e.g. another or no treatment). Eleven RCTs representing 11 different medical conditions were included in the study. Most trials were rated as having ‘high risk of bias’ overall. It is to be concluded that due to the small number and the heterogeneity of studies, no meaningful overall interpretation of the studies was possible.
Randomised controlled trials (RCTs) of non-individualised homeopathic treatment with comparison groups other than placebo
Mathie et al. (2019) performed a systematic review and meta-analysis of randomised controlled trials of non-individualised homeopathic treatment in which the comparison group received other treatment than placebo (e.g. other or no treatment). Ten trials provided extractable data for meta-analysis. However, the authors concluded that “significant heterogeneity undermined the planned meta-analyses or their meaningful interpretation”.
Systematic review of meta-analyses of randomised placebo-controlled homoeopathy trials for any indication
In this systematic review, Hamre et al. (2023) analysed meta-analyses of randomised controlled trials of homeopathic treatments to placebo with regard to results and methodological quality. Six meta-analyses were identified, covering individualised homoeopathy (Mathie et al. 2014, Linde et al. 1998), non-individualised homoeopathy (Mathie et al. 2017) and all homoeopathy types (Linde et al. 1997, Cucherat et al. 2000, Shang et al. 2005). The risk of bias (low/unclear/high) was rated low for three meta-analyses (Linde et al. 1997, Mathie et al. 2014, Mathie et al. 2017) and high for three meta-analyses (Linde et al. 1998, Cucherat et al. 2000, Shang et al. 2005). For the three meta-analyses with low risk of bias (Linde et al. 1997, Mathie et al. 2014, Mathie et al. 2017) , effect estimates for all trials in each meta-analysis showed a significant positive effect of homeopathy compared to placebo. In sensitivity analyses restricting to high-quality trials, the effect remained significant in two meta-analyses (assessing individualised homoeopathy, Mathie et al. 2014,and all homoeopathy types Linde et al. 1997) and was no longer significant in one meta-analysis (which assessed non-individualised homoeopathy, Mathie et al. 2017). The authors concluded that the quality of evidence (assessed as either high/moderate/low/very low) for positive effects of homoeopathy beyond placebo was high for individualised homoeopathy and moderate for all homoeopathy types and non-individualised homoeopathy: “There was no support for the alternative hypothesis of no outcome difference between homoeopathy and placebo.”
Conclusions
The attempts to analyse the specific efficacy of homeopathic preparations in overall meta-analyses of randomized controlled trials across all possible indications with comparison groups other than placebo did not arrive at meaningful conclusions due to the large heterogeneity of the available trials.
However, the meta-analyses with low risk of bias of randomised placebo-controlled trials (RCTs) of homeopathic treatments found that such treatments yielded an overall significant effect in favour of homeopathy. The quality of evidence was moderate for non-individualised homoeopathy and high for individualised homoeopathy, i.e. when provided by a trained homeopath.
