“What is not surrounded by uncertainty cannot be the truth.”
–Richard Feynman
“Indecision and delays are the parents of failure.”
–Attributed to George Canning
In medicine, every clinical decision we make involves uncertainty. This lack of certainty can come from the personal biases of both clinician and patient, but it more frequently stems from ambiguity as a result of either deficient scientific data, abstract criteria, or the multiple possible presentations of disease and illness. When making decisions about people’s health, it often feels as though we face simultaneously overwhelming and insufficient amounts of information, inadequate to accurately understand or describe a clinical situation.
While our clinical notes reflect our uncertainty in 71% of cases, we tend to avoid sharing this with each other and with our patients. Uncertainty is frightening, as it threatens to reveal to patients, peers, staff, and ourselves that we are not quite who we thought we were — that we do not have the “special knowledge” typically attributed to practitioners.
Worse than the potential blows to our egos is the hasty or overcorrective outcome that can ensue when we are uncertain. In our desperate attempts to simplify situations, we may choose unfocused labs, imaging, invasive tests, and even treatments, a form of medical error that raises costs, risks, and morbidity for patents. Rather than forming actual hypotheses to test, we may overuse existing resources, gathering poorly targeted additional general information that ultimately adds to the opacity of a dilemma. Unorganized steps commonly lead to as much delay in helpful treatment as inaction will. Not only is this uncomfortable and discouraging for patients, it may also worsen their ultimate prognosis.
Uncertainty can also lead to indecision. Indecision is not just described by taking or not taking action. It results when we don’t utilize a conscious cognitive strategy to work through uncertainty once we have faced it. This state of indecision leaves clinicians frozen and not cognizant of well reasoned steps that could help to resolve the clinical situations brought to us. Whereas uncertainty is inherent to medicine, indecision is fatal and must be avoided.
Fortunately, evolving models of clinical reasoning have identified the best practices to directly address and manage our routine uncertainty before it becomes dysfunctional indecision. These begin with self-awareness of our cognition, monitoring our use of various modes of problem solving, and skillfully pivoting when stymied by impasses.
The first steps are anticipation, awareness, and acceptance of our uncertainty, in addition to estimating the degree of it we experience in any situation. We physicians often discount the unspecified probabilities which flow from ambiguity, and some personalities are more comfortable tolerating uncertainty than others. When we accept that all situations are initially ambiguous, that our most common medical problems are the most likely to display atypical presentations, we can comfortably look directly at, rather than ignore, incomplete data. We can then follow the deficiencies we find to generate new clinical questions to answer. Our uncertainty is our cue that data is missing. Identifying these information gaps gives essential direction to our problem solving.
Competitive hypothesis revision, an element of abductive reasoning, is an approach that lessens diagnostic and therapeutic error. With this method we form initial hypotheses, test them, then return to amend the original guesses with the new information we have gathered (referred to as Bayesian inference). New hypotheses are then formed which better fit the additional data and are subsequently tested.
With this procedure, only the best hypotheses are carried forward, and only after being updated by our new findings. The search for new data cannot be broad and indiscriminate; it must be informed by the results of previous hypothesis testing. Then, since acknowledgment of ambiguous situations has helped us gather pertinent missing data, we are able to revise or form new hypotheses. The result is a progressive reduction in uncertainty, as we are guided by new and relevant data toward our next decision. This is an iterative process, as we continually reassess our information and refine questions and hypotheses. If steps are not being taken to choose new hypotheses and revise our working models of clinical problems, we have lost our way in clinical reasoning and are left truly indecisive.
As Box and Draper famously observed, “Essentially, all models are wrong, but some are useful … The practical question is how wrong do they have to be to not be useful.” Impasses are expected throughout assessment and treatment, and quick resolutions of problems are as fortunate as they are rare. We err, though, when we consider a clinical impasse merely as a failed or suboptimal treatment attempt, and fail to grasp the opportunity it affords: the chance to correct our original conceptualization of a problem and possible solutions to it. Instead, we can discover new features of a dilemma, repair or complete our modeling of it, broaden our mindset, and form new associations to consider and test.
The complexity of our practice has increased in recent decades as more treatments become available and our patients live longer lives. As a result, there are a greater number of acute and chronic comorbidities and treatments for them that we must contend with. Complexity can be viewed as a gift instead of a problem, however, as it provides rich granularity containing the individual answers we and our patients seek: solutions amid clinical uncertainty. We must learn to become comfortable with complexity, as well as ambiguity.
Communicating our uncertainty to patients, rather than displaying our indecision, must be thoughtful and applicable to the specific situation. How patients react to these disclosures is primarily determined by our style of communication with them. While variable, patients’ responses are usually more positive when we share our information with them in full, speak positively, and explain the steps we will take to resolve uncertainty. Both clinician and patient should begin their work together with the expectation that impasses are likely, alongside a shared understanding of how these will be converted into helpful answers.
Overall, the inherent everyday uncertainty in our work is best managed with awareness of our method of clinical reasoning; shared decision-making; thorough, methodical, and consistent history taking and examination; steps to exclude concerning relevant conditions in our differential diagnoses; and the trust we share with our patients. Admitting in advance that we are unlikely to know everything we need to after an initial examination, we can come prepared with a cognitive strategy to keep moving forward in a positive and productive direction for our patients and our own professional satisfaction.
Dr. Putman, in Austin, TX, is the author of “Rational Psychopharmacology: A Book of Clinical Skills,” “Encountering Treatment Resistance: Solutions Through Reconceptualization,” and “Thinking Again: Reducing Cognitive Error in Psychiatric Practice.” He blogs at drpaulputman.com. Dr. Putman was a 2024–2025 Doximity Op-Med Fellow, and continues as a 2025–2026 Doximity Op-Med Fellow.
Image by Denis Novikov / Getty
