Uncertainty is the essence of medicine


The practice of medicine may be based on science, but it is filled with uncertainty: idiopathic conditions, culture-negative infections, equivocal test results, presumptive diagnoses, empiric therapy. In the clinical care of patients, physicians have been trained to manage uncertainty with skill and confidence. In fact, this is the essence of medicine. We begin with differential diagnoses based on signs and symptoms, then proceed to eliminate options (i.e., uncertainty) until a diagnosis is confirmed.

Sometimes conditions and diagnoses cannot be confirmed with certainty, however, and must be treated anyway. Physicians often are uncomfortable documenting these situations. Perhaps it's the desire to reach that one final diagnosis in the differential, to get the right answer. Maybe it's a fear of missing something or of being wrong.

Photo from Thinkstock
Photo from Thinkstock.

Yet even Medicare's ICD-9-CM Official Guidelines for Coding and Reporting assure us there is an element of clinical judgment and uncertainty in dealing with patients and establishing diagnoses: “If the diagnosis documented at the time of discharge is qualified as ‘probable,’ ‘suspected,’ ‘likely,’ ‘questionable,’ ‘possible,’ or ‘still to be ruled out,’ or other similar terms indicating uncertainty, code the condition as if it existed or was established.”

Other acceptable terminology indicating uncertainty includes: “consistent with,” “compatible with,” “comparable with,” “indicative of,” “suggestive of,” and “appears to be.” “Evidence of” does not qualify, nor does comparative/contrasting terminology like “versus” and “either/or.” Physicians should verify the probable or suspected diagnosis in the discharge summary (or final progress note) to confirm it did not change.

Physicians must exercise reasonable medical judgment when documenting a probable or suspected diagnosis, and the diagnosis must be consistent with and supported by the rest of the medical record—particularly when there is comprehensive treatment of the condition. Authoritative evidence-based criteria and guidelines can inform a reasonable judgment of what is probable or likely. One wouldn't diagnose “suspected” pneumonia and only give 3 days of antibiotics, for example. Likewise, antiplatelet or anticoagulant therapy would support a diagnosis of a “probable” cerebral vascular accident.

Uncertainty is more common with some diagnoses than others. Health care-associated pneumonia (HCAP) is a good example, where the causative organism is rarely identified but is likely, suspected, or probable based on the clinical circumstances and scientific medical data. The predominant cause of HCAP when culture results are negative is gram-negative and/or Staphylococcus, and most such patients receive a full course of broad-spectrum antibiotics targeted at one or both. In such circumstances, a diagnosis of “HCAP probably due to gram-negative and/or staph” or a similar statement is warranted to ensure correct coding that accurately reflects the clinical reality of the patient's condition and management.

Contrary to what many physicians are told about inpatient documentation and coding, CMS permits uncertain diagnoses for documentation and coding of patient visits on a daily basis so long as they remain uncertain. Once a diagnosis is confirmed or excluded, it must be documented as such. According to the December 2010 Documentation Guidelines for Evaluation and Management (E/M) Services from the CMS Medical Learning Network: “For a presenting problem without an established diagnosis, the assessment or clinical impression may be stated in the form of a differential diagnosis as ‘possible,’ ‘probable,’ or ‘rule out’ (R/O) diagnosis.”

Commercial payers may disagree with the guidelines, but they represent an industry standard of practice for physician documentation and coding.

Physicians should embrace uncertainty. Knowledge, experience, and wise clinical judgment tell us the most likely or probable condition when faced with limited data. Anyone can diagnose and treat pneumonia with obvious X-ray findings and positive culture results. Were it not for our professional judgment and clinical intuition in uncertain situations, a computer could practice medicine better than we.