Spanish AEMPS Homeopathy Report Misrepresents the Scientific Evidence

A recent claim by the Spanish Agency for Medicines and Health Products (AEMPS) that there is “no scientific evidence” of efficacy of homeopathy for any condition, is not backed up by their own scientific data. This statement was made in a press release¹ announcing release of a report² commissioned as part of the government’s #ConPrueba programme³, explicitly designed to “dismantle” pseudo-therapies. Production of the report was also overseen by a Minister who publicly described homeopathy as “a scam”⁴ 18 months before the report was published. This context raises strong concerns about distortion of the scientific process, as we see a low quality report, designed with critical flaws incorporated from the outset, appearing to be used to fit a pre-determined political agenda.

The Spanish report is highly misleading due to the following factors:

1. The AEMPS has misreported their own inconclusive result for homeopathy as a negative result

The report by the AEMPS states in its opening Summary of findings that “the evidence is insufficient to make a treatment recommendation with homeopathic products in any pathology”.

This carefully worded conclusion accurately reflects the AEMPS’s assessment of 64 studies (systematic reviews) presented in the report: 5 reviews looked at the evidence of efficacy of homeopathy across multiple health conditions and all 5 are listed as being inconclusive. In addition, 58 systematic reviews assessing homeopathy for treatment of specific medical conditions were summarised by the conclusion, “No condition in this group has an adequate evidence base to support a treatment recommendation.”

The authors of the AEMPS report have misreported these results from their own data assessment, altering the actual result of insufficient evidence to recommend homeopathy, to there is “no evidence” at all that homeopathy works.

2. The AEMPS makes inaccurate, scientifically unsubstantiated negative claims about the evidence

Reporting the result that there is ‘some evidence, but not enough’ as a complete lack of evidence, is scientifically unacceptable, yet the AEMPS goes even further with two inaccurate and unsubstantiated negative claims.

Firstly, the authors inexplicably draw the definitive conclusion that, “It has been established in published works that the efficacy observed for homeopathic products is comparable to placebo.” No reference is given to support this statement. Not only is it completely unsubstantiated, but the authors fail to mention any of the reviews which found some evidence of beneficial effects of homeopathy beyond placebo (e.g. Mathie et al. 2014⁵; Perry et al. 2024⁶) and therefore directly contradict this statement.

Secondly, the report claims that it is a “well-known empirical fact” that the efficacy of homeopathic preparations decreases as study quality increases, and that when only high-quality studies are analysed, the effect disappears entirely. However, the study the AEMPS cite to back up this supposedly definitive point, not only fails to support the statement, it directly contradicts it.

The authors of the paper (Mathie et al.)⁷ identified this trend in the studies, but further analysis established that this correlation seen between effect size and study quality, was not statistically significant, meaning that this observation may simply be due to chance (p=0.417 for risk-of-bias sub-group comparison; p=0.617 for meta-regression). For AEMPS to state that this statistically unconfirmed trend is a “known empirical fact” is unjustifiable.

3. The review is scientifically low quality, failing to use usual appropriate methods

The report’s credibility is further undermined by its failure to use the standard, accepted methodology for a study of this type (a review of reviews or ‘umbrella review’)⁸. Incomprehensibly, the AEMPS report:

• Lacks PROSPERO pre-registration – publishing precise details of how the study will be conducted before it begins is the usual, established safeguard against bias, minimising the ability to influence the results during the process
• Uses the wrong tool to assess study quality – the validated quality assessment instrument for a review of reviews is AMSTAR-2, yet the report uses an unspecified NIH tool not designed or validated for this purpose⁹
• Omits the Cochrane Library from its database search – a critical omission for a review of reviews
• Restricts the scope of the review to RCTs published from 2009 onwards
• Conducts no overlap analysis to account for the same clinical trials (RCTs) being counted more than once across multiple included reviews
• Applies no GRADE certainty-of-evidence ratings to its conclusions
• Provides no sensitivity analyses.

None of these weaknesses or omissions are explained, or even acknowledged as limitations in the report’s conclusions, yet they result in a report of such low scientific quality that any conclusions drawn cannot be considered valid.

4. Selective reporting and factual errors throughout the AEMP report results in a distorted

     assessment of the homeopathy evidence base

Numerous instances of ‘selection bias’ (omitting relevant data which contradicts the information the authors chose to include), as well as information that is simply factually wrong, have created a highly biased report, rather than the useful, objective document one would expect from a Government-commissioned review.

Most strikingly, the report’s section on mechanism of action of homeopathy openly challenges the plausibility of homeopathic medicines being biologically active at high dilutions and states that existing theories of mechanism of action “have no empirical or scientific support^”. This false statement denies the existence of the entire field of basic research on homeopathy which demonstrates that homeopathic medicines have specific physicochemical characteristics and the ability to exert biological effects in multiple experimental models.

Systematic reviews have found that 72% of physicochemical studies reported specific structural, thermal or spectroscopic features of homeopathic preparations that differ from control solvents¹⁰; and 77% of biological experiments report measurable, reproducible effects of homeopathic medicines across various models¹¹ including laboratory studies on cells^{11, 12}, animals¹³ and plants¹⁴.

Most recently, evidence of physiological effects has extended to agriculture, with even highly-diluted homeopathic preparations (which one would not expect to be biologically active according to standard biochemistry) being shown to improve survival, growth and reproduction in molluscs¹⁵, fish¹⁶ and crustaceans¹⁷ under controlled farming conditions – results warranting further research.

5. Selective reporting gives a distorted picture of the global status of homeopathy

A section apparently presenting an international consensus against homeopathy by health authorities in several countries, is critically flawed. Four national reports are included: 1 positive (Switzerland) and 3 negative (UK, France and Australia). The degree of bias is clear as soon as one observes that the positive Swiss report description includes key limitations and criticisms of the report, whilst the 3 negative reports are presented with no mention of any of their limitations or the criticisms levelled against them – only active endorsement of their credibility.

The report also refers to decisions made by the UK and French authorities to reduce support for homeopathy, without clarifying that these were heavily-informed by the 2015 Australian NHMRC report, criticised for drawing its conclusions on the findings of only 5 trials out of 176 trials assessed in the study. Thus, these are not three independent, converging assessments of the evidence as they appear, but an echo-chamber of circular citation of a single document.

Furthermore, as no mention is made of India or Brazil, where homeopathy is formally integrated within the public healthcare systems, the AEMPS statement that “in recent years there has been an international consolidation of public health institutions towards increasingly critical positions on homeopathy” does not objectively reflect the varied status of homeopathy globally.

6. The AEMPS report fails to accurately report the strong safety profile of homeopathy

Six systematic reviews have evaluated the safety of homeopathy: 5 reviews concluded in favour of the relative safety of homeopathic treatment, describing adverse effects as mild to moderate and transient^18-22. The sixth review (Posadzki et al 2012)²³, is the only study to reach a negative conclusion, generally regarded to be due to flawed methodology^24,25. The report states that “There are few data on the safety of homeopathic products”, yet of these 5 studies, the AEMPS report chose to include only two:

• Posadzki et al. 2012²³ identified 4 deaths associated with homeopathy over a period of 30 years worldwide, but failed to make the usual, essential distinction used in safety studies i.e. that all 4 cases were examples of clinical negligence (due to misprescribing or failure to refer for conventional treatment) and not related to ordinary standards of care by qualified homeopathic practitioners.
• Stub et al. 2016²¹ provides useful insights from RCTs, finding that homeopathy has a similar safety profile to control groups. However, the most recent study by this research group on observational studies (Stub et al. 2022)²² is a surprising omission, as it also provides useful data, concluding that “…the proportion of patients experiencing adverse effects was significantly higher when receiving conventional medicine and herbs, compared to patients receiving homeopathy.”

Conclusion

It is clear that the AEMPS report presents a fundamentally distorted and incomplete picture of the scientific evidence on homeopathy, reaching conclusions that cannot be considered accurate or objective and are therefore invalid.

HRI does not dispute that the clinical evidence base for homeopathy from randomised controlled trials requires strengthening before the evidence of efficacy for specific pathologies can be considered sufficient to recommend homeopathic treatment; however, a finding presented by a Government body to the general public as settled scientific consensus, must be backed by accurate and definitive evidence. The AEMPS report falls far short on both counts.

ReferencesLess