Plant-based bioassays have a long tradition in homeopathic basic research. This research approach relies on organisms that are more complex and differentiated than single cells, but do not raise any ethical concerns (as in animal or clinical research). Plant-based bioassays can be grouped into four main areas: assays with unimpaired plants, assays with abiotically stressed plants, phytopathological models, and agricultural field trials.
The research field of plant-based bioassays has been reviewed systematically in 2009/2011 (Betti et al., Majewski et al., Jäger et al.), and a thorough update was published by Ücker et al. in 2018. In this review, a manuscript information score (MIS) was used to assess the reporting quality of the papers. Further quality criteria included the use of statistical models to assess the results, and the use of adequate controls to reliably identify specific effects of homeopathic preparations.
In total, 192 publications were found describing 202 experimental studies. 119 studies included inferential or descriptive statistics, and 74 were rated with MIS > 5 (out of 10 points), thus allowing proper and detailed interpretation. Forty-two out of these 74 studies used succussed or potentized controls to identify specific effects of homeopathic preparations. Forty out of 42 studies (95%) reported significant effects against controls, including dilution levels beyond the inverse Avogadro’s number. Many single studies with diverse methods, and only a few replication trials were identified. 18 studies reported on the use of systematic negative control (SNC) experiments to assess the stability of the experimental set-up.
The quality of the studies increased in recent years. In the update by Ücker et al., covering the time period of 2009/2011-2017, all studies (100%) used some kind of statistical analysis, compared to 50% of the studies from the period 1920-2008/2010. The percentage of publications reaching at least 5 points in the MIS increased from about 60% (1920–2008/2010) to about 75% (2009/2011–2017), and the percentage of studies conducting systematic negative control experiments nearly doubled to 23% compared to the first reviews of 2009/2011.
The most consistent and intriguing result of these reviews is the observation of alternating biologically effective and ineffective potency levels within a series of potency levels of a given substance. All studies that tested series of consecutive potency levels reported such a non-linear and discontinuous relationship between effect and potency level. The specific form of this dose-response-relationship was different for each combination of experimental bioassay and substance potentized. The corresponding “pattern” of effective and ineffective potency levels within a given experimental system proved to be reproducible over weeks to months but seemed to be subject to change after a longer period of time (years). Since such non-linear effects may be considered as surprising and unexpected, they highlight the need for highly controlled experimental set-ups to exclude false-positive results. The best way to control and document the stability of the experimental set-up are SNC experiments, which were implemented in 18 out of 42 studies in these reviews, and yielded evidence for stable experimental conditions.
It is remarkable that 40 out of the 42 studies which used succussed or potentized controls, reported significant effects against controls. One may suspect a publication bias, e.g. the withholding of trials without significant effects. Due to the heterogeneity of the experimental designs and outcomes, a quantitative meta-analysis could not be performed; correspondingly, there is no available estimation of the number of possibly unpublished trials to achieve a zero result in meta-analysis. Based on our experience in the field, we assume it highly unlikely that there should exist a sufficient and larger number of unpublished studies reporting non-significant effects to outweigh 40 published studies reporting significant effects.
Summarizing, in 95% of all published investigations of homeopathic preparations with whole-plant bioassays, significant effects of homeopathic preparations were observed against adequate controls. We thus conclude that there is strong evidence for specific effects of homeopathic preparations over placebo. The empirical observation base thus justifies and calls for further investigations to identify the mode of action of homeopathic preparations.
