Arriving at a diagnosis is the intellectual end point of defensive medicine. Defensive medicine is directed at finding the definitive test to establish the diagnosis or doing enough tests to rule out other possible diagnoses. This ruling out is considered an important legal protection should someone later determine the patient’s actual problem.

The ruling out of alternative diagnostic hypotheses is a valid strategy if the universe of possible alternatives is sufficiently large. But it is a dangerous strategy if the initial diagnostic assumptions constrain subsequent data collection and hypothesis generation. The physician must guard against ignoring the patient’s stated problem in favor of a diagnostically tidier problem. For example, if the patient has trouble walking because of a plantar wart, the physician should not attempt to repair the patient’s asymptomatic slipped disk as a substitute for removing the plantar wart.

This emphasis on diagnosis leads to laboratory diagnosis as the ideal of technological medicine. [Fortess EE, Kapp MB. Medical uncertainty, diagnostic testing, and legal liability. Law Med Health Care. 1985;13:213–218.] Clinical laboratory tests are seen as objective, while taking a history and physical is subjective. Laboratory tests are easy to replicate, may be discussed without reference to the context of the specific patient, and, at least superficially, are easy to interpret. Tests with numerical results lend an air of science to medical practice.

Before the advent of digital electronic calculators, calculations were worked out with slide rules. When digital calculators became cheap enough for general use in the classroom, many teachers opposed their use. There was a philosophical concern with the false rigor that electronic calculators created. The slide rule was accurate to only two or three digits; thus, using a slide rule constantly reminded the student of the approximate nature of the underlying data. Digital calculators and digital reading instruments display many digits—but most medical data are approximate, usually to only two digits of accuracy and sometimes less. Manipulations of these data that do not take into account their limited accuracy lead to spurious results, but results that appear to be scientific because they are numerical.

This problem of false rigor is exacerbated by the use of multiple test panels. These panels include large numbers of tests that measure mostly unrelated parameters. The test results are expressed numerically, but the evaluation of the results depends on the comparison of these numerical values with normal values. Normal values are determined by statistical techniques. They are usually set such that 5% or 10% of healthy people given the test will have values outside the normal range.

When a patient is given a panel with 20 independent tests, each with a normal value defined by the 95th percentile, then the probability is high that at least one test will be falsely normal or abnormal. The noise from the false test results makes it more difficult to evaluate the diagnostic content of the test panel and increases the chance that important information will be lost in the mass of data. The tests may also document important problems that the physician missed. When an injured patient seeks the advice of an attorney, the attorney has the luxury of working backward from the injury. If the physician did not order the proper tests, then all the other tests that were ordered become irrelevant. And if the proper test was ordered but the results were overlooked or ignored, the emphasis on ordering tests accentuates the error in not acting on the relevant test results.