Over the last few decades, the nature of medical knowledge has changed significantly. Before the revolution in evidence-based medicine, clinical medicine was practiced as more of an art (in the “artisan” sense). Individuals were treated empirically with a strong knowledge of medical biology, and the guidance of “The Giants”, or particularly skilled and respected practitioners. While the opinions of skilled practitioners is still valued, EBM adds a new value—one of “show me the evidence”.
Evidence-based medicine refers to the entire practice of gathering and applying medical knowledge. This includes evaluating diagnostic tests (e.g. how well does an CT scan diagnose pulmonary embolism?) and evaluating treatments (e.g. which anticoagulant is most effective, which one is safer, how long should you treat, etc.) There will always be some questions that are untestable, and some for which no testing is needed, and practices for which evidence is sketchy.
In corresponding with a friend recently, I started thinking about how we look at the quality of medical evidence, and how we can communicate this to the lay public.
Let’s take, for example, cholesterol.
It has been found over the years that there is a strong association between elevated cholesterol and coronary artery disease. Through many studies, it was found that LDL cholesterol is a useful marker of cardiac risk due to elevated cholesterol. Finally, it was found that lowering LDL cholesterol, especially with statin drugs, dramatically reduces heart disease risk and mortality.
That’s the facts. But what are they based on? How strong is the evidence?
Evidence-based medicine has many ways of grading quality of evidence. The primary reference for cholesterol treatment in the U.S. is the National Cholesterol Education Program’s Adult Treatment Panel III (NCEP and ATP III).
Before the report gets to any recommendations, it presents a table explaining how evidence used in the report is graded. First is the type of evidence (from randomized controlled trials down to clinical experience) then is the quality (from “very strong evidence” to “strong trend”).
For example, the ATP gives the following recommendation (among many):
Evidence statements: Secondary prevention trials demonstrate that reduction of LDL-cholesterol levels significantly reduces risk for recurrent major coronary events in persons with established CHD (A1).
That “A1” at the end gives an idea what kind of evidence we’re working with. In this case the statement is based on RCTs and there very strong evidence to support it.
Evidence-based medicine is about evidence. Sometimes that evidence is quite good, sometimes it isn’t: the quality of evidence is something we take into consideration when treating patients.
This is in stark contrast to so-called alternative medicine. Alternative medicine never discusses quality of evidence. The quality is usually, however, quite poor, relying on patient and physician anecdotes, uncontrolled “trials”, and fantasy.
Making use of EBM doesn’t have to be difficult. You don’t have to be able to interpret every chi-squared analysis, you don’t have to calculate every number needed to treat. You just have to be able to read the basic literature in your field, look up recommendations, and know how strong they are.
Anyone practitioner who ignores evidence-based medicine is not practicing the best of modern medicine. Anyone who treats disease without understanding the difference between evidence-based practice and non-evidence-based practice should hand you a Quack Miranda Warning when you walk in the door.