Baseline differences in paper by Trichard, Chaufferin and Nicoloyannis

In the paper by Trichard, Chaufferin & Nicoloyannis, published in the January issue of Homeopathy (2005;94:3–9), there are baseline differences in important variables such as passive smoking, amount of preventive treatment, parents employment and childminding arrangements among others. Yet the authors state that the statistical advisor concluded that the baseline data were comparable. The only method that might have corrected for this bias is a regression analysis. I understand that this was recommended by one of the peer reviewers.

As an editor myself, I know that publishing sometimes means to strike a delicate balance between relevance, interest and rigour. Yet the results as reported are very misleading. It would be very helpful if the authors were to publish a regression analysis.