It is a terrific idea to compare effectiveness of treatments for the same illness—if “effectiveness” and “same illness” mean the same to the doctors, the patients, and the payers and patients who will benefit or lose according to the rules that follow. It is a terrific idea if the comparison meets the needs of each participant and if effectiveness measured in the short term also predicts what happens after many years.
In chronic illness these ifs are mostly wrong. For practical reasons, treatment studies mostly look at outcome after one or two years and at only one measure of success, such as, did a blood test improve? For chronically ill persons a short-term result many not predict the distant future. Some treatment choices require quick decisions: how best to repair a fracture after a fall. The immediate outcome may say the treatment worked (the fracture healed), but a later measure may give a different answer (walking is worse because the wrong repair was made). An illness relapses; the result—did permanent damage occur because the wrong treatment was used?—may take decades to know. A treatment effective in the short term (relief from pain) may be harmful later on (the chosen treatment weakens bones). The most important measure of effectiveness may be that disability does not occur, which we may not know for decades.
Treatments have side effects that some patients tolerate and others do not. Some patients choose one treatment that prevents crippling despite the fact that it increases risk of heart attacks, while another chooses a treatment that spares the heart but carries a high risk that the patient will end up in a wheel chair. Which is more effective? A third person cannot know the patient’s choice.
Effectiveness trials do not study complicated patients. Although a patient with rheumatoid arthritis who has had breast cancer cannot participate in an arthritis trial, this patient will still need treatment at some time. Patients with very mild disease and patients with severe disease are not part of trials but need to be treated as well. Exceptional patients—perhaps the majority of those with a given diagnosis—may not be able to take the “effective” drug, but, knowing the patient only by his diagnosis, the third party suggesting the most effective treatment may make the wrong call.
In a treatment trial, patients and researchers agree on the goals, relief of pain, for instance, or a joint less swollen. Outside the trial patients may have different goals. Methotrexate, the most effective first drug for rheumatoid arthritis patients, can deform or kill the baby of pregnant woman. Some women will not take it because they hope to or may accidentally become pregnant. Others will not take it for religious reasons. But even today insurers challenge the prescription of a less preferred drug while offering the physician no opportunity to explain deviation from their version of the effectiveness norm.
Personal goals can also trump the most-effective-treatment choice. To care for an ailing parent or child, to stay employed, to save a marriage, many patients will select short term gain over long-term risk. A few years ago, when Vioxx and Bextra were taken from the market, several of my patients ranked the relief provided by those drugs more important than risk of heart attack and hoarded as much of these drugs as they could find. Many patients so hate the way a drug makes them feel that they refuse a second course of treatment no matter how effective the drug is deemed. “Effectiveness” seldom acknowledges to patients’ personal concerns.
Nonetheless, it is possible and right to compare effectiveness of medical treatments. In doing so, we should not use a one diagnosis, one time period, one outcome trial to decide the best treatment for complicated patients with chronic disease. We should instead compare treatments under real world conditions that include patients with second illnesses, with different severities of illness or different disease courses, and whose personal needs are respected when the treatment is chosen. The term “effective” should consider whether the effect is true for both the short- and for the long-term. Above all, labeling a treatment effective should weigh the patient’s opinion whether the treatment has or has not enhanced his life.