Bailar et al and Medical Research without Controls

In 1984, ethical considerations might not have made placebo boot studies impossible. Practical considerations, however, did make such studies, however, quite difficult. First, patients were generally referred to boot centers for treatment of serious foot problems that had defied "standard therapies". They had been under the care of various doctors who had prescribed appropriate antibiotics and had considered surgical revascularization procedures if they had not already been done. Further, their course had been observed over time to the point where they were commonly advised to have Circulator Boot therapy or amputation. Usually in pain and quite sick, they would not sign for a placebo study. They wanted therapy with some hope. Still many physicians were encouraging us to do a placebo study. How? What would be the placebo? Placing a leg in a boot that did not pump was not an alternative. Patients with ischemic legs are commonly restless and their movement may actually help improve what circulation they have. Placing the leg in our plastic bag likewise potentially might smother the skin surface. The sick patient was indeed difficult to study.

The remaining leg of the amputee had potential for a study; perhaps 50% will come to amputation within five years if the patient survives that long. We offered to a local rehabilitation hospital to perform boot therapy on the remaining leg of a population of diabetics that would be randomized to their routine rehabilitation program and their program plus boot therapy. The study was declined for political reasons: fear of loss of rehabilitaton referrals from physicians who might fear loss of the patients to our program. It was apparent that we at least were not going to do a prospective placebo-controlled randomized study. How should a physician evaluate therapies without such controls?

Bailar et al (N Engl J Med 311:156-162, 1984) noted that not all medical therapies lend themselves to rigid experimental investigation with internal controls. In designing studies in these areas, the authors made five recommendations:

• (1) The investigators should express before the study that the treatment will affect disease outcome (In Circulator Boot publications, both the patients and referring physicians were advised the treatments were likely to affect patient outcomes).
• (2) The data analysis should be planned before the data are generated ( In boot studies, both pertinent vascular data and photographs were obtained both before and after treatments).
• (3) The investigators should articulate a plausible hypothesis before the results are observed (The means whereby boot therapy benefits peripheral blood flow was described in our first publications).
• (4) The results should be of some interest even if "negative" or "opposite" findings are observed ( Negative results would have discouraged interest in the use of boots for vascular diseases while the positive results have led us to extend our use of the Circulator Boot to study its effects in patients with leg infections, neuropathy, venous insufficiency, Buerger's disease, congestive heart failure, angina, and arterial lesions of the arms).
• (5) There should be reasonable grounds to generalize the results of the study to a broader group of patients ( As just noted , we have generalized the treatment to other areas).