Second in an occasional series that will examine how new health information should or should not affect personal decisions about health and prevention.
Mrs. P, in her early 50s and worried about her strong family history of osteoporosis, is taking postmenopausal hormone therapy.
Neither breast cancer nor heart disease is particularly prominent in her family, and she isn’t concerned about altering her risk of those conditions.
Mrs. A, in her mid 50s wonders about taking hormone therapy to reduce her risk of the heart disease and Alzheimer’s disease that her mother had.
Women take hormone therapy to alleviate menopausal symptoms, and also with the hope of decreasing the risk of various aging-related conditions. Each situation involves a balance of fears and the possibility of benefits and risks.
On May 31, 2002, a randomized controlled trial of combined hormone replacement therapy versus placebo, conducted as part of the Women’s Health Initiative, was stopped early. The media reported that the trial was halted because of evidence, in women receiving combined hormone treatment, for an increase in the risk of invasive breast cancer, and possibly other harms that outweighed the benefits. The message was that, instead of being beneficial, postmenopausal hormone replacement therapy did more harm than good.
What has happened when a medical research trial is abruptly halted because of adverse effects? What are women to think now?
Early termination is unique to randomized controlled trials. In these, participants agree to be randomly assigned to one of several treatment groups (or other interventions like dietary modification or x-ray screening), defined in advance by the investigators. The investigators generally suspect that there are benefits to one or more of the treatments, but there isn’t strong enough evidence to prove it. A randomized trial is conducted only when the question of whether there is a difference among treatments remains unanswered.
Participants are informed about this uncertainty when granting their consent. If, in fact, during the trial the evidence becomes convincing that there really is a difference (either beneficial or harmful) among treatments, then it would be unethical to continue the trial.
This generally applies only to randomized trials. For most other studies, instead of being randomly assigned to a treatment group, participants make all of their own choices about treatments, tests, diets, lifestyles or exposures to potentially beneficial or harmful agents. Investigators observe them to learn about possible associations between their choices and health outcomes. Stopping is not an issue.
Increasingly, randomized controlled trials are carefully monitored for safety by an independent board of experts. This is considered particularly appropriate for a large, long-term prevention trial, such as the Women’s Health Initiative trial, involving essentially healthy participants and having the potential for both beneficial and harmful effects involving a number of diseases or conditions. The board monitors the conduct of the trial and periodically reviews the information coming out of it. The board watches for unexpected results, and particularly for potential harmful effects on the participants.
Considering both the results and the rules established when the study was designed, the board may recommend continuing, modifying or stopping all or part of the trial. Such recommendations could be based on evidence for benefit or for harm, or both.
While the first responsibility of the monitoring board and those sponsoring the trial is to the safety of the participants, in considering early stoppage they may take into account broader interests in public health as well.
In the Women’s Health Initiative trial, guidelines for possible early termination because of harms or benefit were part of the original design. The monitoring board reviewed results semiannually. When it appeared that for 16,000 women followed over the first five years of the trial the consequences of taking a particular preparation of estrogen plus progestin were indeed different from the consequences of taking a placebo, the trial was halted approximately three years early. An increased occurrence of breast cancer in women receiving combined hormone treatment met the guidelines, indicating a real difference. The board “concluded that the evidence for breast cancer harm, along with evidence for some increase in [coronary heart disease], stroke, and [pulmonary embolism], outweighed the evidence of benefit for fractures and possible benefit for colon cancer over the average 5.2-year follow-up period.”
While the meaning of the differences that were observed in the trial may vary from person to person, the monitoring board made a decision to terminate, based on its responsibility to protect the safety of the participants as a group.
The early termination does not mean that no one should consider hormone therapy. The trial provided new information about the consequences of a particular treatment given to a particular group of women observed over about five years. This does not encompass all women, all hormone treatment regimens, all time periods or all consequences. This one study adds an important piece to a much larger and still incomplete picture painted by all of the available information.
Individual women, whether members of the study or not, can now consult with their healthcare providers to evaluate their own particular tradeoffs of benefit and harm in light of all of the available evidence.
Mark Zweig, MD, previously a staff physician at NIH, has more than 35 years of experience in medical research and 25 years in patient care. He has particular interest in medical decision-making and preventive medicine.