Requirements and Assessment of Laboratory Tests—Part 5 of a Series on Evaluation of Scientific Publications: A Good Example

Abstract

Borreliosis seems a good example to use in furthering this discussion. Professor Bautsch gives a probability of borreliosis of about 85% after a positive test result in cases where an indication for a blood test is given (for example, facial paresis). The range of symptoms that is associated with borreliosis by doctors, patients, and self help groups is extremely broad, however, and many symptoms are also reported for diseases with a psychosomatic origin. For this reason, the question is how these individual symptoms are weighted when an indication is defined. Spontaneously, “hard” indicators that spring to mind include particular dermatological changes (erythema migrans, benign lymphocytoma, acrodermatitis chronica atrophicans), arthritis, and cardiomyopathy, as well as individual neurological disorders (for example, meningo radiculitis and paresis of the facialis muscle, although the latter is rare in adults). If other indicators are selected then the assessment of the result may more closely resemble that of a “population screening,” because often no consensus exists about whether individual symptoms are considered to be associated with borreliosis. In everyday clinical practice, it is easy to lose sight of this. Epidemiological studies assume that 35% of the population will come into contact with borreliosis depending on the region, but mostly this population never develops apparent signs of illness. If it is true that no clear distinction is possible between active and cured borreliosis (p 405, right hand column) then that makes it even more difficult to reach reliable conclusions. Even if such conditions can be included in the statistical estimates, just which conclusions could possibly be drawn?