Evidence-based medicine - What does it mean to the average practitioner?

i) Defining the research question

This step is the key to facilitate the retrieval of the best research evidence. While defining the research question, one has to describe a population and the intervention intended to be undertaken. It needs to be compared with another common option of management with well-defined outcome measures and a defined follow-up. Hence, PICO (Population (P), intervention (I), comparison (O), and outcome O) is the key to frame a research question. For example, in the case of a displaced fracture neck or femur in the elderly, the two treatment options available are internal fixation of fracture and replacement arthroplasty (hemiarthroplasty or total hip replacement). The reoperation rate at one year is taken as an outcome measure. So research question could be “effect of arthroplasty and internal fixation on one year reoperation rate at one year follow-up.”

ii) Retrieval of research evidence

The research evidence is to be collected for the research question framed. We have required to retrieve the best research evidence on a given research question. It must be ensured that the retrieved literature is comprehensive and not selective, where few articles have been retrieved from general search engines, such as Google. The proper keywords (MESH terms) should be used, and literature should be searched on PubMed, Cochrane Library, and other well-established search engines.

iii) Appraise the evidence

The retrieved evidence is to be appraised for robustness of the studies and quality of evidence. Randomized control trials (RCTs) or Meta-analysis of RCTs are labelled as level 1 evidence, while Prospective Cohort Studies are level 2, Case Control Studies are level 3, and Retrospective Case Series are level 4. Individual opinion is described as the lowest level, 5. Level-1 studies have less bias while Level-5 evidence has the most bias. We also need to practice how to analyze these studies for their robustness. Bhandari and his team have published very simple checklists for randomized trials,⁴ metanalysis,⁵ prospective cohort study,⁶ case series,⁷ and diagnostic tests.⁸

Poorly conducted RCTs (Level-1) are inferior to well-conducted Cohort studies or level 4 retrospective studies. Robust RCT should have valid results. The interventionist should be well-trained in performing both types of interventions. It is important to know how subjects were randomized. Some questions to be asked are: Was the randomization concealed? Were they aware of allocation? Were the outcome assessors aware of allocation? Were the subjects analyzed? Were they initially randomized into enrolment? Was the follow-up completed? and How were the trial results measured? Similar criteria have been described for other types of studies. The clinician has to be confident about the quality of evidence to provide the expected outcome, only then it can be applied to a given patient.

iv) Patient value (applicability of evidence on a given patient)

On retrieval of literature, we may get variable evidence. While making a decision on a patient, we need to assess the patient’s expectations, available infrastructure, and the expertise of a surgeon/clinician to achieve the best treatment outcome. This is the clinical expertise of the clinician to use evidence for the best clinical outcome.

Above denotes that the evidence cycle starts with a patient where a question is asked about clinical problem, literature is acquired, and appraised. After assessing the patient’s requirement and the feasibility of its applicability in a given circumstance, the treatment option is based on the best available evidence.

Barriers to practice EBM

The most common barrier is the availability of too many articles on the digital platform. It is to be understood that data available on the internet may not be authentic and instead may be industry-driven. We have to learn how to identify the best quality evidence. We have to practice how not only to retrieve but also to appraise the best available evidence.

Another common barrier to implementing EBM is a lack of adequate outcome effectiveness studies, particularly in developing/resource-crunch countries. In India, we face a vast variety of clinical challenges and see complex clinical conditions to be treated in variable infrastructures and with variable expertise. For many clinical problems, we do not have research-based evidence generated from the West, and not too many Indian researchers are publishing research. Unless we create evidence on clinical conditions unique to our country, we will not have the best available evidence. We have to publish to generate evidence of clinical conditions unique to us. It is important that we publish our data. In certain clinical situations where we do not have robust data, we need to use guidelines developed by government agencies or WHO.

We should continue to be updated about literature and be prepared to change the practice as evidence changes. One such platform is OrthoEvidence, for example, bringing best evidence summaries in surgery to help surgeons and clinicians stay up to date. There are similar platforms in medicine.⁹

To conclude, evidence-based medicine allows us an opportunity to rationalize the best treatment option to a patient by integrating the best available research evidence.