It's been said that the placebo effect is the only effect that's common to all drugs, yet is the sole effect of some. Something so common yet elusive can be difficult to control for. To understand a drug's efficacy, however, clinicians should take into account the placebo effect. Here are important facts about the effect that you need to know.

Coping with the Placebo Effect
A placebo is defined as a medical treatment or preparation with no specific pharmacological activity with effects that are only psychological. In a study, comparing a drug to a placebo can help determine efficacy and side effects.

In a perfect world, only the active drug in a study would exert an effect, and the placebo would allow the natural course of the condition to remain unchanged. Many studies show a placebo effect, however. If you look at the results of these studies and actual clinical experience, you'll find that an agent that doesn't have any active ingredients can have both subjective and, even more surprisingly, objective changes related to it. 

The true placebo effect becomes a difficult concept to deal with when you recognize that, in order to control for it, you have to mask patients against any knowledge as to whether they're receiving an active agent or not.

Be careful when wording an informed consent document. The chances for a placebo response can be influenced by details such as stating that the patient "may" receive placebo during the course of the study versus saying that he "will." The influence of such details and of the intrusive nature of a study upon patients' daily routines and lifestyles can make placebo effects or other, non-placebo effects difficult to eliminate.

Other Effects
It helps to be aware of other, similar non-placebo effects that can be disruptive, as well.

 • The Hawthorne effect. This is the impact on patient behavior that can result from being observed. Change in routine mandated by authority figures for a clinical trial (such as the influence of a doctor or study staff giving specific instructions) is perceived by participants as concern for their well-being. This perception, coupled with the feeling of being studied, increases their motivation to carry out instructions and fulfill the study criteria. This can result in all study patients appearing to improve, when it's actually the result of the increased compliance with their drug regimen.

Enrollment exuberance must be controlled. The desire of patients to meet the enrollment by exaggerating symptoms and the desire of the physician to stretch the entry criteria to admit certain patients can enhance this effect as patients will seem to have improved later on. 

 • Regression towards the mean. This may be likely to occur in studies of conditions with fluctuating severity. Patient inducement to seek treatment or enter a study occurs when symptoms are at a peak. This will cause the appearance of improvement with time. For example, in a glaucoma study, patients may be brought in multiple times until they achieve a pressure required for inclusion, such as 25 mmHg. Then, on their next visit, whether they received drug or placebo, their pressures will likely represent an average pressure for them (usually lower), rather than the near-peak level. This creates the impression of a placebo effect.

One can control for regression towards the mean by using multiple defined-in-advance visits at time points prior to entry while also limiting the number of measurements. 

 • Treatment by vehicle. The difference between placebo and vehicle is that the vehicle is equivalent to active drug, minus the active component. Thus, vehicle contains only relatively inert substances while a placebo is intended to have no pharmacological activity. As these are defined in such similar ways, vehicle and placebo for U.S. Food and Drug Administration studies are very similar.
When studying topical drops, in many cases a vehicle that's used as a placebo and that resembles many tear substitutes will bring about some temporary improvement in the severity of a patient's symptoms. If taken frequently enough, it may even exhibit marginal alleviation of some clinical signs based on its lubricating function.

In the case of infection, the preservative benzalkonium chloride is often an ingredient in a placebo or a tear substitute that actually has bacteriostatic and bactericidal properties.

In allergy, the eye wash diluent and barrier function provided by a tear substitute has created placebo effects across dozens of allergy studies, with rates as high as 70 percent.1,2

The placebo effect makes it more difficult for drugs to become approved, but, once approved, improves the likelihood that patients will perceive genuine treatment and satisfaction with any eye drop or other medication. In some ways, this works to the advantage of the pharmaceutical industry. It's clear however that, whenever possible, the FDA does require placebo-controlled studies that are double-masked and randomized. Managed-care formularies follow suit, and, in addition to the negative controls of placebo, require positive controls. These controls often take the form of a separate active agent to which the test agent is intended to show equivalent or superior efficacy.

Studying Side Effects
The placebo group also provides data on the course of side effects.
If one asks enough questions, uses a well-constructed diary and does a long enough safety study, a placebo group will reflect the entire spectrum of ailments, from dog bites to headaches.

It's against this backdrop that the drug must be evaluated for any increase in frequency of side effects or the rare occurrence that could not be expected to be found in a similar group of placebo patients. That's why in new chemical entities the International Committee on Harmonization of Regulatory Standards requires 500-patient safety studies with at least 300 participants on a drug for a six-week minimum for short-term agents, and at least 12 months for chronic-use agents.3 The side effects observed in a study must be compared to the general population. Adverse events that occur in fewer than 1:500 people, however, may not be detected in a clinical trial. Over the long term, these have been shown to be extremely rare for ophthalmic studies because of the general safety of ophthalmic agents and low level of systemic absorption. This situation is fortunate and isn't shared by those who evaluate systemic agents, where rare side effects may occur in 1:10,000 or 1:100,000 people. In these studies, it's often unclear whether these side effects are random or idiosyncratic.

The development of aplastic anemia in a couple of patients using chloramphenicol eye drops that resulted in their being banned from the United States shows how naturally occurring side effects can skew our view of an agent's side effects. The instance of aplastic anemia being approximately 1.5:200,000 suggests that it would conceivably have occurred in the normal population, since millions of people had used the agent. The fact that this agent is widely used and effective in Europe and Asia shows that there's both an upside and a downside to interpreting such data.

It's good to know that all drugs currently approved work better than placebo. However, always bear in mind that even a placebo will work for some people. 

Dr. Abelson, an associate clinical professor of ophthalmology at Harvard Medical School and senior clinical scientist at Schepens Eye Research Institute, consults in ophthalmic pharmaceuticals. Ms. Fink is a research associate at Ophthalmic Research Associates in North Andover.

1. Leino M et al. Double-blind group comparative study of 2% nedocromil sodium eye drops with 2% sodium cromoglycate and placebo eye drops in the treatment of seasonal allergic conjunctivitis. Clin Exp Allergy 1992; 22:10:929-32.
2. Tinkelman DG et al. Double-masked, paired-comparison clinical study of ketorolac tromethamine 0.5% ophthalmic solution compared with placebo eyedrops in the treatment of seasonal allergic conjunctivitis. Surv Ophthalmol 1993;38:(S)133-40.
3. ICH Harmonised Tripartite Guideline: The extent of population exposure to assess clinical safety. Oct. 27,1994.