How Will We Know When the “Right” Number of People Choose Active Surveillance?:

In this month's issue of Thyroid, Hwangbo et al. (1) report on the effects that physician specialty, a decision aid, and other factors have on the patient's choice of treatment for their papillary thyroid microcarcinoma. Although perhaps not the intent of the study, the results show the challenges we face in helping our patients make “the right” decision.

In the MAeSTro study, patients with papillary thyroid microcarcinomas were offered a choice of surgery or active surveillance, and they were followed for both cancer and patient-reported outcomes. More than half of the patients in the MAeSTro study reported they had heard of active surveillance before they were diagnosed with their cancer, a testament to the effective media strategy that has been undertaken by the nation's physicians and lay media to make this option known. Overall, ∼64% of people enrolled in the study chose active surveillance, with the other 36% choosing surgery.

Some patients were recruited to the study by surgeons, others by endocrinologists. Many of the patients who participated in the study were counseled using an active surveillance conversation aid. Conversation aids are a newer style of decision aid designed to support physician–patient discussions (2).

Who the patient saw for consultation had a profound effect on choice of treatment. Patients who happened to be recruited by an endocrinologist chose active surveillance 87% of the time, while those recruited by surgeons chose active surveillance only 11% of the time. These differences strongly suggest that the primary driver in treatment choice was the physician. It might be tempting to think the physicians in this study were biased in their beliefs, but which was the biased party—the surgeons or the endocrinologists? Or both? It might also be tempting to write off these results as reflective of physician-centric decision-making styles of Korea that might not be seen in other locations, but decades of data suggest otherwise.

The strong role of the physician in determining treatment was first described in the United Kingdom in 1938, by showing the tremendous variation in rates of tonsillectomy by physician availability and practice style, with Wennberg leading a generation of researchers in this area starting in the 1970s (3). This kind of practice variation has been termed “unwarranted variation,” it is when rates of a specific medical practice vary greatly across regions and the variation cannot be explained by illness, medical need, or the dictates of evidence-based medicine.

By the mid 2010s, demonstrations of unwarranted variation in virtually all categories of preference or supply-sensitive health care had been shown in most of the Organization for Economic Cooperation and Development (OECD) countries, including disparate cultures and health care systems such as Estonia, Ireland, Japan, Norway, the United States, and Canada (4). It seems none of us is immune to having influence on our patients' decisions, and most work now is focused on ways to decrease the unwarranted variation.

One approach to overcoming unwarranted variations in health care is to try to decrease the influence of the physician. Decision aids were invented to be sources of unbiased information about the evidence supporting different treatment options, to help patients make a decision that is right for them. Exposure to the conversation aid style decision aid used in this study influenced choice of treatment. Those exposed to the decision aid had 2.46 higher odds of choosing active surveillance than those who were not. In the post-treatment decision survey of study participants, ∼76% of those who chose active surveillance had been exposed to the aid, whereas only ∼47% of those who chose surgery had been exposed to it.

How exposure to the aid varied by physician specialty would be interesting to know, but was not separately reported. Nevertheless, the distribution of treatment choice by decision aid exposure raises the question—how many people in the MAeSTro study exposed to the conversation aid “should have” chosen active surveillance? All of them? Only a small proportion? What is the right number of people choosing active surveillance overall? This study cannot tell us that.

To be sure, observational data from studies such as MAeSTro will create important evidence that can be used to help patients and physicians make better decisions. But as we look around the corner to the next steps in advancing the evidence base for early stage and low-risk thyroid cancer treatment options—including emerging options such as radiofrequency ablation, we will need to take our thinking about decision making to the next level, too.

Our observational studies of treatments for early-stage thyroid cancer and evaluations of decision support tools, when used, should include 1) assessments of concordance between patients expressed values and priorities and their ultimate choice of treatment, and 2) tests of association between these outcomes, patient-reported outcomes, and levels of decisional satisfaction and regret.

It is time also to closely examine and improve the processes by which we work with our patients to come to these decisions. Tools such as the rating scales of Elwyn allow patients to rate the level of collaboration they experienced in their decision making, and observers to assess the level of shared deliberation seen in the interaction (5). Through the inclusion of these sorts of tools and approaches, we can simultaneously work to gather evidence on the treatment options for early-stage thyroid cancer as we also improve how we involve our patients in decisions for their care, a win–win for all.