Misdiagnosis by intuition: Seven diagnostic axioms to enhance patient care

Most would agree that a correct diagnosis is a key prerequisite to providingsafe and effective treatment for various illnesses. However, our diagnosesare often a matter of opinion rather than matter of fact.

In fact, isn't it true that most diagnoses often require us to make decisionsin the absence of certainty? Our diagnoses are based on probability. Thisbeing the case, we must use caution not to fall into the trap of makingdiagnoses on the basis of faulty logic or insufficient information. It isone thing to make a diagnosis and another to be able to substantiate it.Though we name the things we know, we do not necessarily know them becausewe name them.

Audit the diagnosis

Shortcuts in diagnostic reasoning tend to become increasingly prevalentwhen veterinarians are subjected to the pressures of a high case load ina busy hospital. In this context, short cuts are often defended on the basisof "practicality". Although practicality is a virtue, we mustuse caution not to use the concept of practicality as an excuse for ignorance.

A misdiagnosis may be more detrimental to the patient than the illness.A wise sage once penned this thought: "Heaven defend me from a busydoctor."

How do we know when our diagnoses are in error? If we do not have a systemdesigned to periodically audit our diagnoses for accuracy, we are unlikelyto recognize and correct our errors. If the accuracy of our diagnoses isnever questioned, we may become overconfident in our judgments with a tendencyto rely less and less on clinical data and more and more on our intuition.

What is the inevitable result? Experience has revealed that diagnosisby intuition is often a rapid method of reaching the wrong conclusion.

Based on the premise that a well-defined problem is half solved, theprimary objective of this Diagnote is to review some clinical axioms thatfoster the diagnostic process. An axiom is a statement universally acceptedas true.

Seven diagnostic axioms

Axiom 1: There is a difference between knowledge and wisdom.

Knowledge is facts; it consists of familiarity with information gainedby study and observation (that is, empirical experience or investigation).Unfortunately, most of us have been taught to over-emphasize the accumulationof new knowledge to a point where we neglect the development of acquiringwisdom.

Whereas knowledge consists of our familiarity with relevant information(facts), wisdom consists of the ability to properly apply knowledge. Itimplies sufficient breadth of knowledge and depth of understanding to providesound judgment. Although essential, facts (knowledge) by themselves arerarely of useful value. Facts are not science, just as the dictionary isnot literature.

In context of diagnosis of diseases, facts become useful only to theextent that they can be wisely used to define, solve and prevent problems.If we have knowledge but have not learned how to make practical applicationof it, we lack wisdom.

Axiom 2: There is a difference between problem definition andproblem solution.

We use the term diagnosis in context of defining the cause(s) of clinicalsigns. The ability to define a patient's medical problems without overstatingthem is a crucial first step in the diagnostic process, since one must beable to define problems before they can be solved.

No veterinarian has or ever will be trained to single-handedly solveall types of medical problems. No one can recall enough knowledge and beproficient with enough techniques to guarantee that (s)he alone can providethe best care of every patient. Veterinarians can be trained to accuratelyidentify problems, however. They can and should be master "problemdefiners."

Accurate definition of a patient's clinical problems will permit us inour role as diagnosticians to more efficiently use available resources,such as journals, books, the Internet, consultations and referrals, to helpresolve diagnostic problems. A problem well defined is half solved.

Axiom 3: There is a difference between observations and interpretations.

Discernment of the difference between observations (facts) and interpretationsof observations (inferences or assessments) is a critical component of thediagnostic process. In the process of defining problems, we must use carenot to consider the meanings of observations and interpretations as equal.

Likewise, we should avoid mixing observations and interpretations randomly.Why? Because observations and interpretations represent distinctly separatefacets of diagnosis. Consider this example. As veterinarians, we frequentlyinterview clients who confuse observations and interpretations when describingthe illness of their animals to us. A classic example is to misinterpretthe observation of tenesmus as constipation in a male cat with urethralobstruction.

This type of error in reasoning is not limited to clients. It affectsus all at one time or another. For example, when asking for specific laboratorydata such as the hematocrit value (an observation), we may be told thatit is normal (an interpretation).

But a hematocrit value of 37 percent (an observation), which is interpretedas normal may actually be abnormal in a severely dehydrated patient. Althougheither observations or interpretations may be erroneous, in our experiencemisinterpretation of a correct observation is the most common pattern oferror.

What is the point? A misinterpreted problem is the worst of all problems.Why? Because if misinterpretations are unknowingly accepted as facts, misdiagnosisfollowed by misprognosis and formulation of ineffective or contraindicatedtherapy may result. This is indeed ironic since the patient may then bein a worse condition as a result of having visited us in our roles as doctors.What can we do to minimize this problem? One thing is to put the axiomsin this column into practice. An observation or an interpretation is unlikelyto mislead us if we learn how to avoid being misled.

Axiom 4: There is a difference between possibilities and probabilities.

The need to discern the difference between diagnostic possibilities anddiagnostic probabilities is another key diagnostic axiom.

In general, collection and interpretation of relevant clinical data abouta patient's illness allows us to reduce numerous diagnostic possibilitiesto a few or one diagnostic probability.

However, even after collection of a large quantity of relevant data,many diagnostic probabilities still represent a matter of educated opinionrather than a matter of fact.

Recall that absence of clinical evidence of suspected diseases is notalways synonymous with evidence of absence of these diseases. As a corollary,detection of evidence that is consistent with a certain, specific type ofdisease is not always pathognomonic for a specific disease.

It follows that we as veterinarians should convey to our clients thatour diagnoses, prognoses and treatment recommendations are based on probability.

Axiom 5: There is a difference between disease and failure.

Discernment of the conceptual difference between organ disease and organfailure is also fundamental to proper diagnostic refinement. Organ functionthat is "adequate" to sustain homeostasis is often not synonymouswith "total" organ function.

For example, patients with only one kidney have adequate renal functionto live a "normal" life without manifestations of renal dysfunction.Even when slowly progressive irreversible lesions occur, signs of organdysfunction do not develop if adequate quantities of functional parenchyma(i.e. nephrons, hepatic lobules, etc.) remain to sustain homeostasis. Thisconcept is the basis for distinguishing organ disease (such as cardiac valvularinsufficiency) from organ failure (such as altered circulation associatedwith abnormal cardiac rate and rhythm which ultimately occur as a resultof irreversible progressive cardiac valvular insufficiency).

Won't you agree that the approach to management of a patient with cardiacvalvular insufficiency and adequate cardiac function is very different frommanagement designed for a patient with cardiac valvular insufficiency andcongestive heart failure?

Axiom 6: There is a difference between clinical signs inducedby diseases and the body's compensatory response to disease-induced signs.

Clinical manifestations of disease can be subdivided into the followingtwo classes: 1) signs directly induced by the disease (such as impairedurine concentrating capacity and obligatory polyuria associated with damageto the countercurrent system in patients with bilateral bacterial pyleonephritis),and 2) the body's compensatory response to these signs (such as compensatorypolydipsia needed to maintain fluid balance because of obligatory polyuria).

Other examples of this relationship include compensatory inflammationin response to damaged tissue, fever in response to systemic infectiousagents, polychromasia and reticulocytosis in response to anemia, and hyperparathomonemiain response to hypocalcemia. It follows that making a diagnosis of urinarytract infection solely on the basis of pyuria would be an overdiagnosisbecause pyuria may be a compensatory response to both infectious and noninfectiousdiseases.

Axiom 7: There is a difference between events that occur consecutivelyand cause and effect relationships.

The ability to recognize true cause and effect relationships is not aninnate characteristic - it must be learned. The important point to be madehere is that just because two or more events occur in consecutive orderdoes not prove a cause and effect relationship. Why? Because, unrelatedcoincidences commonly occur in the lives of all of our patients. Considerthis example. In the late 1970's and early 1980's, vesicourachal diverticulawere cited as playing an etiologic role in some cats with lower urinarytract disease (LUTD). Treatment by surgical extirpation was recommendedin most veterinary textbooks at that time. The observation that clinicalsigns subsided coincidentally with diverticulectomy, and lack of studiesof the biologic behavior of macrosopic diverticula without surgery, reinforcedthe interpretation that this anatomic abnormality was a cause of LUTD.

However, subsequent studies revealed that vesicourachal diverticula werea sequela, rather than a cause, of LUTDs. Most of them spontaneously resolvedwith appropriate medical therapy of the underlying problem. Surgery wasunnecessary. This example highlights the fact that favorable outcomes associatedwith our treatments do not prove that our diagnoses were correct, or thatour treatments were.