What is the right answer? Framing and medical decision making

In the Sunday NYT, Denise Grady reports the results of a study out of the University of Florida where they have determined that open breast biopsies appear to be vastly over utilized. They reviewed a Florida database and found that the open biopsy rate was approximately 30% while there is a consensus among experts that the rates should be closer to 10% (http://www.nytimes.com/2011/02/19/health/19cancer.html?ref=health).

What I though was the most intriguing part of this report is the fact that we are now setting up quality standards which are measured not by the fact that we get our patients to do any one  thing, but that we now appear to be targeting some degree of variability (or diversity) in decisions that are the result of our actions. As noted in the article:

She said that when she asked surgeons in the study why they were doing open biopsies, many said patients wanted them. “My comeback was, ‘Do you think you had an inherent bias in the way you explained it?’ ” In the past seven years, she said she had only one patient choose an open biopsy over a needle biopsy.

In the book "Nudge" Cass Sunstein uses (?coins) the term choice architect. We are choice architects having what could be characterized as an unhealthy influence over our patient's decisions. I would venture to suggest that there are very few situations where our we should consider it desirable if our patients followed our recommendations all of the time, particularly when we are dealing with some issue in an asymptomatic person who by all other measures is without disease. 

What this article talks about is the possible driver of the discrepancy between the ideal and real rates of open biopsy, that being the reticence of surgeons to refer to radiologists to do needle biopsies because of the loss of business. Before radiologists rejoice in their new found income stream, it might behoove us to look and an earlier choice also shaped by other choice architects, that being the choice to do mammography, on whom, starting when, and how often. These same issues arose when the US Preventative Services Task Force looked at breast cancer and found that screening mammography impact  was marginal at best in a number of subgroups and its impact so modest that it seemed reasonable that at least some women (and perhaps many women) should be forgoing screening if the data were presented in a balanced way.

While the current article presents the figures as being very black and white, the reality is almost certainly more nuanced. Needle biopsies are very small and although I am not a pathologist, I suspect that the risk of sampling error is much greater than an excisional biopsy. Not all cases are optimal for needle biopsy and depending upon the individual and their preferences, some circumstances warrant the more aggressive diagnostic approach. The question becomes if there is no one right answer and the "ideal" is for the doctor patient discussion to yield a variety of decisions, how you define where the target should be?

Furthermore, how will this translate into other metrics? Should the targets for blood pressure control or  HA1c be different for different patients? Of course they should but how do quality metrics account for this? We also must ask the question as to whether we should be held accountable to individual patient vs. an aggregate goal. If a given patient, after being explained the pluses and minuses of needle or open biopsy, wants an open biopsy because of concerns about sampling error, should we accommodate them or tell them the open biopsy is simply not a choice for them?

Now that the overwhelming majority of encounters in health care occur outside of an urgent setting, where decisions revolve around acting now in some way to prevent some low frequency event which may happen in the future (often way in the future), this is what physicians and patients will face. Physicians need to understand that what they have come to believe is true and base their recommendations to patients will change over time. Patients also need to understand that if they are not engaged in the decision making process, the end result is not likely to reflect their desires. Finally, all parties need to understand that we have wandered into a realm where the results of our decisions will almost never be clearly linked to a given outcome in a given patient. This may be the hardest pill to swallow since both patients and physicians desire the control associated with targeted actions which purportedly save lives. The truth is not so clear.