Intended for healthcare professionals

Rapid response to:

Views & Reviews From the Frontline

Good medicine: homeopathy

BMJ 2012; 345 doi: (Published 14 September 2012) Cite this as: BMJ 2012;345:e6184

Rapid Response:

Re: Good medicine: homeopathy

In commenting upon the Cuban Leptospirosis paper, Alan Henness appears to have either misread the paper, or misunderstood it, or both.

Leptospirosis is a zoonotic disease which is a particular problem to humans mainly in tropical and developing countries. Symptoms listed in the paper include meningitis, pneumonitis, hepatitis, nephritis, mastitis, myocarditis, haemorrhagic crisis and multi organ failure. Infection occurs through mucosa and open skin lesions from contact with water contaminated with urine from domestic and wild animals. Farmers and people working with livestock are therefore at the highest risk of exposure. In areas where drinking water supplies are open and sewage disposal is primitive, whole populations are at risk. These risks are increased when natural disasters such as flooding or earthquake disrupt water supplies. Leptospirosis is endemic throughout Cuba, but the three eastern regions chosen for this study are particularly prone to epidemics throughout the rainy and hurricane season, beginning in October through to the end of December. This study commenced at the end of October 2007 and completed at the end of 2008.

From the second half of 2007, it was apparent that conditions in the Eastern regions were causing a relatively high incidence of the disease, based upon the weekly reports of confirmed cases. As a result of two further major rainfall events in October and November it was clear that a major epidemic was likely. A programme of prophylaxis was therefore instituted by means of a homeopathic preparation of the four strains of Leptospirosis endemic in that part of Cuba and administered to as much of the population in that area as possible. For the purposes of the study, this was designated the Intervention Region (IR) whilst the rest of Cuba was designated the Rest of Country (RC). The homeoprophylaxis was administered throughout the IR during week 45 onwards, i.e. the first week in November, achieving coverage of 90% of the IR population who received the intervention by week 50. Because the programme only started in week 45, about one third the way through the high risk period, confirmed cases appear to have already reached about 38 per week. However, from that point onwards, the numbers of cases plummeted to 2 or 3 by week 49 and remained at that level for the rest of the high risk period. Throughout 2008, cases bumped along at between 2 and zero cases per week, when the programme was repeated in September 2008 achieving 96% coverage. The 2008 high risk period rarely rose above 2 cases per week.

Despite this considerable success, Alan Henness comments as follows:
‘No control’ – This is not a Randomised Controlled Trial. It is a prospective observational cohort study. The intervention arm comprises some 2.1 million people (IR) and the comparison arm (RC) of 8.8 million people.
‘No randomisation’ – There is no need, the study effectively enrols the entire population of Cuba.
‘The intervention region was very significantly different at base line from other affected areas of Cuba.’ – That is perfectly true. The IR was historically by far the most badly affected region. That is why an emergency programme was instituted there.
‘.....the paper gives scant detail about the models and how the validation of the models was done....’ – The details of the model derivation are described in the section headed ‘Statistical analysis’. The validation of the model actually used was compared to the cases for the RC, where the authors state that the actual cases do not differ significantly from the model predictions.
‘Many possible non-specific effects were not discussed....’ – Such as?
‘Without proper randomisation, these effects could not have been taken into account.’ – See ‘No randomisation’ above.
‘Many high-risk individuals........ were also vaccinated’ - in 2007, only 15,000 high risk individuals were vaccinated in the IR, amounting to 0.6% of the total IR population. Given the considerable flooding in this region at this time, this tiny percentage is unlikely to have affected the results or conclusions of the study.
‘There were other more plausible reasons.....’ – Again, such as?
‘I suspect these severe limitations are the reason why the study wasn’t accepted for publication in a respectable scientific journal, but in a homeopathy trade journal.’ – Homeopathy is a peer reviewed scientific journal published by Elsevier.

In his comments, it is clear that Alan Henness has not understood the distinctions between an observational study of a population; a randomised controlled trial using a sample of a population; and the need and reasons to randomise the sample in the latter, but not the former. This distinction is covered in most books on basic medical statistics, to which I direct his kind attention.

This study, and the emergency which gave rise to it, demonstrates that the Cuban health authorities acted boldly and swiftly in responding to what was clearly an emerging epidemic of a very nasty disease. Not only were several hundred people prevented from contracting Leptospirosis in 2007 and 2008, but many lives were most likely saved. Furthermore, by getting on top of the outbreak as early as they did, they managed to ensure that resources needed to cope with other emergencies were better available. Despite this, Alan Henness asserts confidently that: ‘ would be ethically wrong to base a health intervention on such an unsound trial.’ The people of Eastern Cuba may count themselves fortunate that Alan Henness is not their Chief Medical Officer.

Competing interests: User of homeopathy on sheep and cattle.

25 October 2012
David Eyles
Livestock farmer