Decisional conflict, regret and the burden of rational decision-making

Faith and decision-making

I was first introduced to the idea of decisional conflict and regret at 12 or 13 years of age. The religious studies teacher at my school described a “moral dilemma” that he had heard on a radio phone-in. We were asked to imagine that we were in the jungle and had stumbled across a squad of soldiers about to execute some villagers. The officer in charge offers us a deal: if we kill one villager, he would let the rest go free. The teacher told us that he would know exactly what to do, he would refuse the offer because killing breaks one of the Ten Commandments. Indeed, he said that he would have no regrets afterwards, would never doubt his decision and would sleep well. He went on to explain that this moral dilemma illustrated one of the great joys of having faith, which is that you always knew what to do. I remember thinking that I was not at all sure of what the right choice was, and that I probably would always doubt my decision, and that sleep might well be hard to come by. So my first lesson in decision-making was this: low decisional conflict and regret was associated with the abnegation of rational decision-making; thinking through a decision and struggling with moral principles led to decisional conflict and regret.

Decisional conflict and regret in studies of medical decision-making

Every day we ask patients to make medical decisions that are perhaps no less agonizing than the choice of whether or not to kill one person to save many others. Those of us interested in medical decision making want to help patients make those decisions well. My problem is that how we often assess how we are doing in this task, is that we measure decisional conflict and regret, the very endpoints that proved so problematic in the moral dilemma. In a typical study testing the effectiveness of decision aids, for instance, patients are randomized to a decision aid versus standard counseling and then given a follow-up questionnaire to assess decisional conflict and regret, perhaps alongside other measures[1]. If conflict and regret are lower in the decision-aid group, this is seen to be a good thing, and a justification for clinical use of the decision aid. For instance, in a study recently published in Medical Decision Making, Politi et al. randomized 343 uninsured individuals to one of three strategies to support decision making about health insurance[2]. One of the endpoints was “confidence in choices”, a sub-domain of the Decisional Conflict Scale[3], in which participants were answered items such as “I feel sure about what to choose” and “This decision is easy for me to make”. Another recent example of a paper in Medical Decision Making that investigated decisional conflict is an observational study treatment decision making for localized prostate cancer[4]. Orom et al. measured decisional conflict expressly to answer the question “What is a good treatment decision?” With respect to regret, Becerra Perez and colleagues published a systematic review in Medical Decision Making that aimed to determine the extent and predictors of decision regret after medical decisions[5]. One express purpose of the review was to design interventions to reduce regret based on understanding of risk factors for regret. To be eligible for the review, decision regret had to be measured using the Decision Regret Scale, that includes questions such as “It was the right decision” and “The decision was a wise one”[5].

It makes sense that good decision-making would lower decisional conflict and regret. I recently watched a urologist at Memorial Sloan-Kettering Cancer Center work with a older man who had low-risk prostate cancer. He started by carefully and sensitively eliciting the patient’s preferences and understanding of treatment options, before presenting high-quality information in a systematic, empathetic fashion. He presented opportunities for the patient and his family to ask questions and answered these clearly, based on best evidence. At the end of the consultation, the patient decided to go forward with active surveillance, a program of careful monitoring that avoids aggressive treatment unless the cancer is found to progress. I have little doubt that the urologists’ expert handling of the consultation reduced decisional conflict, as the sense of relief in the room was palpable. I also believe that the patient will not regret his decision, as the course of treatment chosen avoided treatment-related side-effects but is unlikely to lead to a poor cancer outcome.

This would seem to be a vindication of the idea that conflict and regret are reduced by a good decision-making process. But it is a non sequitur to move from our intuition that “good decision-making tends to lower decisional conflict and regret” to “lower decisional conflict and regret distinguish good types of decision-making from bad ones”. Indeed, it is trivial to imagine how a bad decision-making process could lead to better conflict and regret outcomes than a good one. For instance, imagine if the urologist had informed the patient that he had a serious, life-threatening condition – it was cancer after all – that radiotherapy and active surveillance was really for older and sicker men, and that the patient needed to “man up” and do the right thing for his family: “if you want to live to see your grandchildren, I don’t see what choice you have”. Although such advice is grossly at odds with both the scientific literature and current treatment guidelines, it is highly plausible that it would lead to salutary effects on decisional conflict and regret: the patient might have been confused as to what to do, he now knows he needs surgery, and may not subsequently regret his decision (“I’d probably be dead had Dr. Jones not treated me”). Conversely, we might imagine that our older, low-risk patient was planning to have surgery, and had not heard of any other options, but received appropriate advice to avoid aggressive treatment. He does so, but this leads to an increase in decisional conflict from his pre-consultation level, which would have been zero, as he did not realize a decision needed to be made.

Regret and conflict are associated with a second problem, which is that they can be highly time-sensitive. Take two patients with leukemia, one of whom undergoes aggressive chemotherapy and the other of whom opts for a cure with diet and vitamins. A few months after diagnosis, the chemotherapy patient is feeling awful, regrets agreeing to treatment and wonders what the right decision was; the natural healing patient is feeling terrific, is certain of her decision and has no regrets at all. Levels of regret are likely to be very different when, a year or so later, the first patient has recovered from treatment and is cancer-free and the second patient is dying of advanced disease.

Regret is problematic in a second way because it relies on individual patients understanding the counterfactual of what life would have been like had a different decision been made. As a second example from prostate cancer, PSA testing is known to have led to widespread overtreatment of indolent tumors highly unlikely to affect a man’s survival or quality of life had he avoided screening. However, a man with an incomplete knowledge of the natural history of prostate cancer might well believe “I would be dead today if that PSA test hadn’t caught my cancer just in time”. Such a man would have no regret about his decision to get a PSA and be unnecessarily treated for low-risk disease. Interestingly, I have heard an anecdote that illustrates the same problem from the opposite direction. An unmarried man made a decision not to get a PSA test in his early 50’s. A little over 10 years later, having married and fathered a child, he was diagnosed with metastatic prostate cancer. Shortly before dying, he said that he had no regrets, because avoiding PSA had allowed him the joy of fatherhood. This ignores the point that there are medical techniques available to allow men undergoing prostate cancer treatment to become fathers. In both cases – the man who believed that treatment of an indolent cancer saved his life and the man who believes that a PSA test would have prevented him becoming a father - low decisional regret results from a false faith that is the very opposite of informed decision making.

I am not the first commentator to question the value of conflict as an endpoint. Nelson et al. argue that the use of decisional conflict as an endpoint in studies of decision aids involves the assumption that “decisional conflict and uncertainty represent an undesirable state that is detrimental to decision making”. They point out that “appropriate deliberation about alternative outcomes and personal goals, as well as ongoing engagement in the decision-making process” may increase conflict. As an example, they compare a patient who “becomes absorbed in obtaining information and considering the risks and benefits of breast cancer treatment options” with one who simply asks her doctor to make a decision, and predict it would be the former patient who would score higher on the Decisional Conflict Scale[6].

In sum, decisional regret and conflict may be sensitive but are certainly not specific to good decision-making processes. A model example of informed shared decision-making with an empathetic doctor may well lead to lower regret and conflict, but so might a strategy that either omits or misspecifies crucial information.

Is a life well-lived one without conflict and regret?

As the old adage goes, it is better to be Socrates unsatisfied, than a pig satisfied. Hard decisions – whether moral dilemmas about murder, or life-and-death decisions about cancer treatment -can be heart-wrenching, and we may always agonize over our choices. We can go to a book to tell us what to do rather than think through a moral problem, or numbly follow the advice of a blowhard surgeon rather than grapple with actual data, and we might well feel more tranquil for doing so: we might agree that we are “sure about what to choose”, that the decision was “easy to make” and later confirm that “the decision was a wise one”. But turning a hard decision into an easy one would not be consistent with how many of us choose to live our lives, and certainly would seem to negate what medical decision-makers have been struggling towards for years. If we measure that success of decision-making in terms of conflict and regret, we may favor easy decisions over rational ones.