The caregiver placebo effect in veterinary medicine
After a long break from blogging on professional topics (although not from blogging in general – if you would like access to the blog from our fantastic trip around the western part of Australia please leave a comment to that effect) I have returned to work and am reading Ben Goldacre’s Bad Science, an eye-opening look at the sometimes questionable studies on which current treatments and accepted wisdom may be based. I had not previously heard of Goldacre, a UK-based doctor, epidemiologist, broadcaster and author who has an incisive mind and a lovely turn of phrase, my current favourite of which is ‘The plural of anecdotes is not data.’ As a bonus in discovery of interesting thinkers, researching this post also lead me to discover The SkeptVet blog, written by a blogger who ‘takes a sceptical and science-based look at veterinary medicine.’
One of the many topics covered in Bad Science is the placebo effect. As a student my somewhat primitive understanding of this concept was that placebos were substances which had no effect in treating a condition and were most often used in clinical trials to provide a control with which to compare the medication being investigated. I assumed that the placebo effect, where beneficial effects of the substance were thought to be due to the patient’s belief in the treatment rather than the treatment itself, was not a significant issue in veterinary practice because placebos didn’t have the same psychological effect in animals as in people. Once graduated, I encountered some vets who occasionally gave sterile water or vitamin injections to patients to satisfy owner demands. This was my first experience with the caregiver placebo effect, also known as the placebo effect by proxy, of which more later.
Surprising researching findings
Goldacre reports on several studies which investigated the placebo effect, some of which had some very revealing results. Placebo tablets have been found to have a dose-response curve in the way you would expect for other drugs (two tablets were perceived to be more effective than one), to be influenced by the colour of tablets (pink tablets helped maintain concentration better than blue tablets) and to be being more effective when injected than when taken in tablet form. In a more extreme form of placebo, patients with pacemakers implanted but not switched on reported feeling better than before having the device implanted (although they did better still after the device was activated!)
Fascinatingly, the doctor’s faith in a particular treatment can also have a significant effect on the outcome for a patient. In a creative study in the mid-1980s, two groups of doctors who were blinded to which medication they were injecting, gave one of three possible treatments to patients. Of the three, only one was effective for the patients’ condition. One group was told that the treatments they were giving were ineffective and the other group were told the truth – that there was a chance their treatment could benefit the patient. Doctors were forbidden to tell patients the chance of their treatment being effective. No prizes for guessing that the patients of the second group did better. Apparently the manner of the doctor was enough to influence the outcome for the patient.
The caregiver placebo effect has significant influence
In veterinary medicine, the vet obviously communicates about the likely success of treatment with the caregiver rather than directly with the patient. The vet’s and owner’s opinion on whether the patient has improved are significant in the vet’s assessment of the success or otherwise of treatment both in day-to-day patient management and in clinical trials. The SkeptVet blog led me to a study1 which attempted to assess the impact of the caregiver placebo effect on outcomes for dogs with lameness due to arthritis. Using animals assigned to the placebo arm of a clinical trial, owner and veterinarian assessments of dog’s response to treatment (placebo) were compared with objective measurements of weight bearing in affected limbs. A placebo effect was noted in over 55% of owner assessments and over 40% of veterinarian assessments.
In an ingenious attempt to elicit the impact of the caregiver placebo effect, a separate study2 on the effects of a non-steroidal anti-inflammatory treatment (meloxicam) on cats with degenerative joint disease (DJD) divided patients into two groups. One group received meloxicam, well-established as a treatment DJD in cats, for three weeks and then placebo for a further three weeks whilst the second group received only placebo throughout the trial. Owners did not know which group their cat was in. Cats were assessed by owner surveys prior to the study, after three weeks i.e. at the time of cessation of meloxicam for those receiving it and finally after six weeks. At the midpoint of the study, cats from both groups were perceived to have improved to an extent not significantly different between the two groups. At the end of the trial, cats in the meloxicam treatment group were perceived to have deteriorated compared with their midpoint assessment while those in the placebo group remained unchanged. The authors propose that the positive effect of the meloxicam in the first three weeks of the trial was masked by the caregiver placebo effect but once the meloxicam was replaced by the placebo, the return of clinical signs after withdrawal of active medication negated this effect.
Such studies illustrate the importance of the caregiver placebo effect in veterinary medicine, both for their veterinarian assessment of the patient and those of the patient’s owners. As veterinary educators, we should ensure that our students are forewarned of the phenomenon and remind mindful of it while working in clinical practice.
1 Conzemius M, Evans R. Caregiver placebo effect for dogs with lameness from osterarthritis. J Am Vet Med Assoc 2012;241:1314–1319
2 Gruen ME, Griffith E, Thomson A et al. Detection of Clinically Relevant Pain Relief in Cats withDegenerative Joint Disease Associated Pain. J Vet Intern Med 2014;28:346–350