Personal opinion: Can antifungal sensitivity tests predict clinical treatment outcomes?

Laboratory tests done with clinical specimens can only ever measure directly the quantity or condition of one or more parameters of the specimen. Responsibility for deciding the meaning of the test result in the clinic lies with the physician attending the patient concerned. The laboratory must, of course, ensure that each test is done with the most stringent possible quality controls to minimize the probability of technical artefacts leading to results that might be misinterpreted in the clinic. As with all medical laboratory tests, the results of measurement of sensitivity of a fungus to one or more antifungal agents in vitro lie on a spectrum between two extremes. At one end are tests (measurements of serum electrolytes are examples) where results that differ from "normal" cannot on their own possibly indicate a precise clinical diagnosis or prognosis. At the other end are tests (many serodiagnostic tests for virus infections are examples) where a positive or negative result has a very high probability of indicating a specific clinical situation. Indeed, it is no accident that serological tests are usually evaluated for their efficacy in terms of statistics that include their "predictive value". My personal opinion of all antimicrobial sensitivity testing, whether with antibacterials or antifungal agents, is that the results usually have only a low predicti-ve value for the clinician, but that they are far too often interpreted as if their predictive value is high. I believe that microbiologists should work much harder to educate physicians to understand the limitations of sensitivity tests. Such tests provide useful advisory information in determining treatment, but they do not and cannot predict definitively when a particular antimicrobial agent will and will not successfully treat an infected patient. Over the years, antibacterial testing has developed a mystique close to a dogmatic, almost religious belief. If a bacterium is reported as "sensitive" to a particular agent, then that agent is automatically regarded as a potentially useful treatment for eradication of the bacterium. If a bacterium is reported as "resistant" to an agent, then that agent is usually excluded for use in the infected patient. In reality, this black-and-white interpretation of a sensitivity test result is a massive oversimplification. A sensitivity test predicts only how to treat an infected test tube [l]. There are many reasons why a "sensitive" bacterium may not respond to treatment with an agent that works in vitro: they include achievement of inadequate …