OLATHE, KAN—The evaluation for allergy typically begins with a careful history and physical examination (HPE). But while this initial, somewhat subjective assessment is of great value, a new study shows that it may overly influence the final diagnosis—sometimes even overriding more conclusive evidence to the contrary.[1] The study also demonstrates that allergen sensitivity as predicted by the HPE often does not coincide with the results of skin testing (ST) and serologic immunoglobulin E (IgE) measurements.

P. Brock Williams, PhD, who led the investigation, said that because the HPE is the standard by which other allergy tests are interpreted, physicians’ impressions of how well these tests perform are not likely to be accurate. “Most allergists have been trained to believe [that] the HPE is the key to a correct diagnosis,” but this is not necessarily the case, he says, because of the development of heuristic biases—cognitive errors that affect clinical decision making.

The study included 152 children ages 6 to 18 and took place at two sites: a university-based allergy training center and a private practice. When each child presented initially, an HPE was performed first to establish the participant’s sensitivity to seven common household allergens; ST (which was not standardized) and serum IgE analysis followed. All individual test results were classified as positive, negative, or indeterminate.

Similar categories were used to compare the individual test results. For example, a diagnosis of allergy based on the HPE was deemed to be a false-positive if the ST and IgE results were both negative; an HPE-derived diagnosis of no allergy was considered a false-negative if both laboratory tests produced positive results.

Rates of agreement between the HPE and laboratory diagnoses rarely exceeded 50% and differed with both the allergen and the allergist. At site 1, where 597 diagnoses corresponding to specific allergens were made by HPE, only 262 of them (44%) agreed with both the ST and IgE results. The lowest rate of agreement was for Alternaria allergen (34%); the highest was for cat allergen (55%). The rate of false-positive HPE diagnoses at this site was 16%; the false-negative rate was 6%.

At site 2, where 441 specific diagnoses were made by HPE, overall agreement was only 22%. The rate of agreement was lowest for dog allergen (5%) and highest for ragweed allergen (32%). No false-negative HPE diagnoses were made at this site, but 22% of the HPE diagnoses were false-positives.

Dr. Williams explained that the higher rates of agreement for cat and ragweed allergen probably relate to the rather straightforward association between exposure and symptoms. He attributed the low agreement for Alternaria to a bias called anchoring, which is a strong attachment to one’s first impression. “As it happens, the allergist at site 1 performed basic research on Alternaria and thus had a tendency to overdiagnose this problem,” he said.

In contrast, the allergist at site 2 used aqueous extracts that are known to be unstable for skin testing; previous experience with a poor dog allergen extract may have led this physician to form the opinion that few people are sensitized to dogs—a bias that Dr. Williams termed availability, in which perceived patterns influence estimates of disease likelihood.

Dr. Williams further explained that although ST results are not as reliable as serum IgE findings, ST results can nevertheless lead to patterning when experiences with different patients are integrated into the physician’s thinking, building the impression that sensitivities to certain allergens are more or less common than they actually are.

The strong influence of the HPE was also evident by the independent decision of the allergist at site 2 to employ select intracutaneous skin tests. Eighty-seven such procedures were performed when the ST results were negative but the HPE was judged positive. Although in only six of the 87 cases did the patients have serum IgE to the allergen in question, a positive result with the intracutaneous test was considered proof that original HPE diagnosis was correct. The authors speculate that use of such a test, which has a significant rate of false-positivity, reinforces the HPE diagnosis “as a self-fulfilling prophecy.”

—Verna L. Schwartz, MS