The term "evidence-based medicine", at first glance, may seem counter-intuitive; isn't all medicine evidence based, and does that mean there is medicine not based on evidence, but perhaps on something more vague, like superstition? Such practices do exist, but the use of the term "evidence-based" does not serve to distinguish between modern medical practices that range from scientific to imaginary, but as a slightly redundant umbrella term for medicine which draws upon the resources made available through scientific inquiry.
The intention of evidence-based medicine is to give the health care provider information beyond the scope of their own experience with which to inform their decisions. EBM is meant to synthesize the practical, experience-based judgment of the physician with the empirical, mathematical data supplied by research. Too much of either factor, on its own, may be detrimental; for example, a research study may be too "clinical", and thus blind to real-world impacts on its conclusions. For example, a high BMI is not necessarily an indicator of obesity, and eating breakfast may not necessarily be linked to weight gain or loss; correlation is not causality. On the other hand, the doctor's own experience may not be sufficient to understand, diagnose or prescribe; it would generally be agreed that a well-informed doctor who can diagnose early warning signs of an unfamiliar condition would be a better health care provider than one who missed these signs and could only offer treatment when the condition was more visible.
The significance of this practice to the provision of health care is that a patient should, in theory, be able to receive the most scientifically informed treatment. The EBM process itself isn't terribly different from the scientific method; the provider acquires information from the patient, performs research, assesses the relevance and quality of the research, and then interprets it as appropriate to the patient. While this is by no means a process of guarantees, it provides both the patient and the doctor with information that will inform their decisions.
There is no single form of research which is most common to EBM, but a frequent form of study are clinical trials. These are usually intended to answer a single or specific series of questions, often with a single controlled variable. A further point of distinction lies between observational studies, in which the subjects are observed for perceived effects of treatment, and random controlled trials, wherein certain subjects are treated and others are part of a control group. The benefit of the random trials is that results unrelated to the treatment, such as a placebo effect, can be isolated. Thereafter, per the EBM process, the health care provider can attempt to create a mosaic of their patient, drawing from a variety of research which responds to specific qualities exhibited by that patient.