ABSTRACT
A single, even epidemiologically impeccable study, does not always answer all of the important questions in a particular area of medical work. Does jogging do more harm than good? Is diethylstilbestrol a predominant cause of clear cell vaginal carcinoma? Do beta-blockers prevent death after myocardial infarction? Is the BCG vaccine equally effective around the world? Chapter 8 stresses consistency of findings as one of the criteria of causality. Systematic reviews of evidence are meant to uncover whether findings about a health problem are consistent across the body of available original studies and valid information in the literature. Readers of medical journals quickly realize how complicated and diversified information becomes. Original studies multiply fast and exponentially, especially in such crucial fields as cardiovascular disease and cancer, among others. These studies, as comprehensive as they may be, do not give homogeneous results across often enormous bodies of accumulated knowledge on a single subject. These studies are followed by ‘review articles’, ‘position papers’, ‘consensus on…’ with recommendations to practitioners as to what and how to diagnose, what treatment to choose, etc. To accept or refute such recommendations and translate them into office decisions or further research is crucial for patient well being, for institutional budgets,
and for good medicine and/or research. By analyzing and summarizing bodies of evidence, clinicians seek new and better premises and conclusions for logical arguments, supported by the best available evidence. Consider what can happen while treating a patient:
However, if the patient asks whether garlic will reduce the risk of myocardial infarction, blood cholesterol must be considered as a surrogate point in this discussion and further evidence focusing on garlic’s preventive role in myocardial infarction should be assessed. Concluding that‘gֹarlic reduces the risk of myocardial infarction' (Conclusion C) is justified only if the relation between the surrogate point and the outcome of primary clinical interest, i.e. not having a heart attack, is as clearly understood as possible in this situation. Additional evidence in this sense would be mandatory. Otherwise, such a resolution of a logical discourse should be refuted. In this fictitious discussion, ignoring evidence would lead the physician to make recommendations based solely on hearsay, faith and goodwill. This is what distinguishes health professionals from charlatans, wellintentioned lay persons and health gurus. Is there a more reliable clinical argumentation than the opinion of a respected medical professional, coming from an equally respectable institution and published in a reputable medical journal? This question was first tackled by authors in the fields of psychology and education. They were the first to develop a structured and systematic way of analyzing and synthesizing independent studies on the same topic. This method should allow the best possible conclusions to be reached by reflecting reality across the different studies under consideration.
