The placebo effect is a critically important, and much misunderstood phenomenon which can mislead us greatly when we are trying to decide if a medical therapy is or is not effective. While most people understand the placebo effect to mean a patient feels better because they think they are getting a real therapy, when actually they are getting a fake therapy, this is only one kind of error included under this label.
Belief and expectation can certainly cause people to imagine an improvement in their condition even when they aren’t getting a real treatment, and even when objectively their disease is unchanged or getting worse. One relatively recent example is a study of placebo treatments, including acupuncture, for asthma. In this study, fake therapies frequently made people feel their asthma symptoms were reduced, but objective measures of lung function didn’t change. This sort of result indicates pretty clearly that even if people report feeling better, it is not a good idea to rely on placebo effects alone because they don’t actually improve the condition being treated. There may be some value to improvements in subjective symptoms like pain even if they aren’t real, but we cannot ignore the fact that these improvements can occur even when the people feeling better are actually getting sicker.
There are many other sources of error in scientific studies of medical treatments that can lead to a false impression of improvement even if a therapy doesn’t work. Regression to the mean, the Hawthorne Effect, the natural course of disease, and many other similar phenomena can fool us into thinking a therapy is working when it isn’t, or lead us to believe the effect of a treatment is stronger than it really is. These are often called placebo effects because they can be controlled for in experimental studies by having a placebo treatment group, but they don’t involve belief or expectation on the part of the patient. Many of these effects occur in animals just as they occur in humans, even though animal patients likely do not have beliefs or expectations about their health. Even with a problem as apparently objective and straightforward as epileptic seizures, significant placebo effects have been seen in dogs.
One source of error seen in studies of medical therapies in humans, and which likely also operates in veterinary medicine, is the placebo effect by proxy (e.g. 1, 2). This is where the beliefs and expectations of someone caring for a patient (often a parent) influence the apparent effect of a medical treatment because the parent either reports on the effect of therapy or behaves differently in other ways related to the treatment being tested. Often, this effect makes it look like patients are improving by the subjective measure of caregiver evaluation even if objective measures show the patient actually is not getting better.
I have long argued that this form of placebo effect is a serious problem in veterinary medicine, where subjective measures of response based on owner or veterinarian opinion are often used to evaluate the effect of therapies for veterinary patients. Inconsistent and usually positive subjective evaluations of response to treatment show up in almost every veterinary clinical trial, and without adequate controls for placebo effects, these can lead to false impressions of a benefit. This is especially a concern with therapies shown in humans to operate mostly or entirely by placebo effects.
One area in which this problem appears to be significant is the treatment of arthritis in veterinary patients. Many therapies are used which have not yet been demonstrated to be effective through high-quality clinical trials (such as stem cell therapies) or which have been shown in humans to operate mostly as placebos (such as glucosamine, acupuncture, and others), are tested in dogs and cats using only subjective assessment by owners or vets to determine if they are effective. A new study specifically evaluating the impact of caregiver placebo effects on evaluation of arthritis treatments illustrates why studies that don’t use objective measures or control for placebo effects are very likely to be misleading.
Conzemius MG. Evans RB. Caregiver placebo effect for dogs with lameness from osteoarthritis. Journal of the American Veterinary Medical Association. 2012;241(10):1314-1319.
This study looked at the assessment of lameness in 58 dogs who were in the placebo arm of a clinical trial evaluating a non-steroidal anti-inflammatory medication for arthritis treatment. The authors compared owner and veterinarian assessments of patient lameness with the results of an objective measurement tool that evaluated the amount of weight the dogs actually put on their arthritic legs. There were several important results that illustrate the importance of proxy placebo effects:
- Objective measures were consistent throughout the study, while subjective evaluations were inconsistent and tended to show progressive improvement as the study progressed, even in dogs who were not actually improving.
- A placebo effect in the owner evaluations occurred 56.9% of the time.
- A placebo effect in veterinarian evaluations occurred between 40% and 45% of the time, depending on which specific method of evaluation was being used.
- These placebo effects were highly significant at every point at which the dogs were evaluated (P<0.001).
- Changes in lameness of </= 5% measured objectively were not unusual, but changes >/= 10% were rare, suggesting that small changes may occur naturally and can interfere with accurate evaluation of a therapy. Higher effect sizes should be seen if we are to have confidence that a therapy is truly working.
These placebo effects, seen in dogs not actually receiving any therapy for their arthritis, were highly significant and easily large enough to make an ineffective therapy appear effective without an objective measure of response. This illustrates quite clearly how important it is that we not accept only subjective assessments of arthritis treatments as proof that they work. We have to remember that the caregiver placebo effect means that the owner or veterinarian judge the patient to be improved, but actually the patient is likely in as much or more pain as they were without the treatment. Trusting in our uncontrolled observations to judge the value of treatments for arthritis pain very likely guarantees uncontrolled suffering for our pets and our patients.