Considering Heterogeneity of Treatment Effects

Cincinnati—Speakers at the AMCP meeting urged healthcare professionals initiating therapies or deciding what drugs to cover to consider that treatments may work differently for people with the same disease.

During a satellite symposium titled The Good, the Bad, and the Different: Deciphering Heterogeneity, the speakers defined heterogeneity of treatment effects as the idea that different patients respond differently to the same therapies.

For example, 2 patients with hypertension may have varying reactions to a medication. Thus, the speakers said it was important to examine risk factors to predict whether he or she will likely benefit.

Penny Mohr, MA, vice president of program development at the Center for Medical Technology Policy, noted that the United States spends approximately 18% of its GDP on healthcare, a much higher percentage than any other country and 30% higher than it should be. She warned that “the survival of Medicare is at stake” because of the costs, and added “we are far out of line of other developed countries.”

Advances in technology are playing a major role in making healthcare expensive, according to Ms. Mohr. She noted 81% of economists identified technology as the primary cost driver in healthcare in a study published recently.

Each year, >18,000 randomized, controlled trials and tens of thousands of other clinical studies are published, according to Ms. Mohr. However, she said despite being “overwhelmed” with information, evidence is “insufficient” for clinicians and patients to make informed decisions.

Medicare ruled it would cover implantable cardioverter-defibrillators only for patients with low ejection fraction and wide QRS. Ms. Mohr said clinicians criticized Medicare for its choice to restrict access.

Recently, there has been a transition toward patient-centered outcomes research, with the primary emphasis on tracking outcomes that help patients. In fact, the Patient Protection and Affordable Care Act created the Patient-Centered Outcomes Research Institute that has a mission of promoting evidence-based information “that comes from research guided by patients, caregivers, and the broader healthcare community,” according to Ms. Mohr.

Ms. Mohr also suggested understanding the heterogeneity of treatment effects could help defray costs and improve outcomes. When examining the effects, she said it is preferable to have confirmatory analysis to test hypotheses related to subgroup effects instead of exploratory analysis to generate hypotheses for future studies.

These options are important, according to Ms. Mohr, because clinical trials have numerous flaws, including research subjects are highly selected, research settings are often different than places where care is delivered, and most research is focused on physiologic or surrogate outcomes instead of measures such as functional status, long-term impacts, or quality of life that are important in the real world.

More on Heterogeneity

Dan Malone, RPh, PhD, professor at the University of Arizona College of Pharmacy, agreed with Ms. Mohr’s assessment that typical studies or analyses that focus on generalities for patients are not always helpful. He said that rather than prescribing a medication for a large group of people, clinicians and pharmacists must take into consideration factors such as age, sex, ethnicity, comorbidities, physiological status, level of activity, alcohol consumption, dietary intake, and treatment settings.

After discussing cases involving his patients, Dr. Malone described how risk factors for diseases can predict outcomes, but they are not independent of each other. For instance, bleeding risk depends on factors such as age, platelet function, presence of hypertension, use of anticoagulant therapy, and alcohol use.

When assessing bleeding risk, Dr. Malone believes that rather than individual risk factors, clinicians should use summary scores, in which factors are weighted based on assigned or derived weights and the ways the individual factors interact are taken into consideration.

Dr. Malone added that it is also sometimes difficult to compare different studies that evaluate the same drug. If healthcare professionals opt to perform a meta-analysis of studies, they should utilize a random effects model that allows for variations among trials and does not assume the trials estimate the same population.

Health Plan Perspective

J. Russell Teagarden, RPh, MA, vice president of scientific affairs in advanced clinical research and science at Express Scripts, said health plans are responsible for setting limits on healthcare coverage. For example, drugs are covered only for certain indications and in certain dosages. In general, coverage policies are focused on covering most needs for most people most of the time within the scope of the plan and usually are derived from population averages and the average population, according to Dr. Teagarden.

Before making the difficult decisions, Dr. Teagarden said plans consider numerous factors such as whether the drug will meet the needs of a general population or seriously ill patients, whether the drug has been proven to be effective and safe, and whether the drugs is cost effective.

There are other considerations, as well, including political (for local governments or federal programs such as Medicare and Medicaid), benefit payer objectives (such as helping to recruit employees and keep them happy and meet contractual obligations with unions), and regulatory (such as complying with mandated coverage and procedural requirements).

Dr. Teagarden admitted that plans sometimes do not account for heterogeneity of patient responses, which can leave people outside of the coverage limits even if that was not the intent of the plan. However, he added, they attempt to be flexible and accommodate heterogeneity of treatments for patients because it is a habit they have been doing for a long time and because it is an ethical imperative.

He described procedural principles that are used to deal with the issue, including making health plan policies and coverage decisions available for individuals to see and allowing people or groups to appeal policies or decisions.