The explosion of clinical research has led to a conundrum in practice: Never before has so much evidence been generated to guide practice, but the sheer volume generated makes it difficult for practitioners to keep pace with the knowledge, and new knowledge rapidly eclipses existing practice. In 2009, it is estimated that more than 120 randomized clinical trials in neonatology were published. This dilemma has made it imperative that every physician become skilled at evidence-based medicine (EBM), which, at its core as defined by Sackett in 1997 is “…a process of life-long, self-directed learning in which caring for our patients creates the need for clinically important information about diagnosis, prognosis, therapy, and other clinical and health issues…”. This chapter will review the components of EBM and the contribution of neonatal research networks to the generation of high-quality evidence.
The Evolution of Evidence-Based Medicine
When first conceptualized in 1992 by Guyatt, the fundamental principle of EBM was real time application of the best available clinical evidence at the bedside. The chief barriers to such application in neonatology were the absence of high quality evidence and the tedious search for, and synthesis of, available evidence. The development of large research collaboratives has led to the generation of high-quality evidence. Advances in computer technology and information management have made evidence available on the desktop of every clinician. The Cochrane Collaboration in 1990 developed standard approaches to literature review and analyses that have placed the practice of EBM within the reach of most practitioners. Neonatologists are indeed fortunate that the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) has funded online publication of the Neonatal Cochrane reviews for more than a decade. This has contributed to the rapid uptake of EBM among neonatal practitioners. The next innovation in EBM will incorporate rigorous assessments of quality improvement methods to aid us in determining which methods most rapidly lead to the incorporation of evidence-based treatments into practice. Many authors have documented that on average it takes more than 7 years for a new practice that has strong evidence of efficacy to achieve high penetration at the bedside. Methods are needed to enhance the dissemination and uptake of these innovations. Physicians who are skilled in EBM are more likely to recognize and incorporate these advances.
A Prescription for Evidence-Based Medicine Focused Practice
Sackett and colleagues synthesized the steps needed to ask and answer a relevant question using EBM ( Box 1-1 ). To these steps we have added a first step using the phrase by Horbar, “developing the habit for using evidence and implementing change,” which has been disseminated among neonatologists by the Vermont Oxford Collaborative.
- 1.
Develop the habit for the use of evidence.
- 2.
Frame the question in a manner that can be answered.
- 3.
Search for evidence with maximum efficiency from the most reliable sources.
- 4.
Critically appraise the evidence for its validity (closeness to the truth) and usefulness (clinical application).
- 5.
Apply the results of this appraisal in practice.
- 6.
Evaluate the performance of the treatment.
Developing the Habit for Evidence Use
Medical students and residents who are educated in a culture that values, teaches, and models the use of EBM are more likely to apply the method themselves in later practice. Nevertheless, all physicians can learn and practice the steps needed. Research has shown that physicians who use EBM are more likely to be current in practice 15 years out of training than those who are not practicing EBM. Today, the American Board of Medical Specialties has mandated continuous maintenance of certification, rather than permanent or intermittent recertification, as the best practice for documenting physician competency. EBM will facilitate self-directed lifelong learning and support maintenance of certification.
Framing the Question
To be easily answered, the exact question must be carefully framed. Strauss and colleagues have summarized the four elements of a good question as “PICO”: Patient population, Intervention, Comparison, Outcome.
Patient Population
Describe precisely the patient population under consideration; for example, “infants born at <28 weeks’ gestation,” OR “inborn infants <28 weeks’ gestation,” OR “very low-birth-weight (VLBW) neonates who remain intubated and mechanically ventilated at 14 days of age.” The more precisely the population is defined, the more targeted the search for evidence will be.
Intervention
Describe the main intervention in which you are interested. For example: “Is clindamycin superior to ampicillin in the treatment of necrotizing enterocolitis?” Other questions that may be explored may relate to prognostic factors or to risk factors.
Comparison
What is the main alternative to compare with the intervention (e.g., when compared with supportive therapy alone).
Outcome
State the outcome of interest in as specific terms as possible including a time horizon. For example: “Will adding clindamycin to ampicillin in a VLBW infant with stage 2 necrotizing enterocolitis reduce mortality prior to hospital discharge?”
A busy clinician will generate more questions than they have time to address. To avoid frustration, the questions may be prioritized by how critical the patient is, or which question is of most interest to the clinician. Other questions can be added to a list, which can be used when off-service time can be directed to self-education. Through this process the clinician will be actively practicing lifelong learning.
Searching for Evidence
Searching for evidence to answer clinically relevant questions is the most time consuming aspect of practicing evidence-based medicine. Strauss and others have suggested that this is the major barrier to effective implementation. Nordenstrom has recommended that clinicians search for evidence using online sources that contain critically reviewed data directed at clinical questions. By prioritizing sources, the clinicians’ time is used most efficiently. Nordenstrom recommends that the first source should be the Cochrane Collaboration, followed by meta search engines including Google Scholar. The next step is to search secondary sources focused on clinical questions such as the United Kingdom’s National Institute for Health and Clinical Excellence ( www.nice.org.uk ), the United States Agency for Healthcare Research and Quality Effective Health Care Program ( http://effectivehealthcare.ahrq.gov ) or Up To Date ( www.uptodate.com ), a commercial online source generated by content experts. Perhaps surprisingly, Nordenstrom recommends that PubMed be searched last, because 75% of the PubMed content deals with basic science research topics versus clinically relevant questions. Thus, for a busy clinician other sources are likely to yield a better answer faster.
Critically Appraise the Evidence for Validity, Applicability and Importance
In this discussion, we will focus on the appraisal of evidence regarding treatments. The highest hierarchy of evidence for these are results from a randomized controlled trial. The following critical questions to ask when assessing the validity of a trial are:
- 1.
Were patients randomly assigned to the treatment?
- 2.
Were all patients who were randomized accounted for in the analysis? Were they analyzed in the group to which they were assigned (intent-to-treat analysis)?
- 3.
Were patients, the clinicians caring for them, and those assessing the outcome kept masked to the treatment assignment?
- 4.
Were the groups similar at the beginning of the trial?
Randomized trials provide the most nonbiased assessment of the effect of a treatment. If the trial is not randomized, it may be best to stop reading and search for other sources. If the only evidence available is from a nonrandomized study, one must view the stated effects with some skepticism because the odds ratios from randomized trials are generally smaller than those from nonrandomized studies.
There are a number of different systems proposed for grading the quality of evidence. The proliferation of systems has made it difficult to adopt and understand any one method. Recently, a group of clinical epidemiologists have proposed a system that combines many of the elements of other systems and termed this the GRADE (Grading of Recommendations Assessment, Development, and Evaluation) system. The British Journal of Medicine has required a GRADE assessment of recommendations since 2006, and now more than 25 groups who generate systematic reviews, including the World Health Organization, the American College of Physicians, the American Thoracic Society, UpToDate ( www.uptodate.com ), and the Cochrane Collaboration have adopted the GRADE standard ( Table 1-1 ). The Grade system synthesizes the evidence into a recommendation based first on the quality of the evidence and second on the magnitude of effects, thereby yielding a recommendation which is either “strong” or “weak.” The GRADE system classifies quality of evidence into four levels: high, moderate, low, or very low. Evidence from randomized controlled trials (RCTs) begins as high quality, but may be rated down if trials demonstrate one of five categories of limitations. Observational studies begin as low-quality evidence, but may be rated up if associated with one of three categories of special strengths.
| Study Design | Quality of Evidence | Lower/Higher Level of Quality if: |
|---|---|---|
| Randomized trial |
|
|
| Observational trial |
|
The GRADE system suggests that when the desirable effects of a treatment clearly outweigh the undesirable effects, or the contrary, that guideline offers strong recommendations. When the data are less clear, such as when the quality of existing evidence is low or when undesirable effects outweigh desireable effects, the recommendations should be rated as weak, or equivocal. Such a standardized approach to rating the evidence would clearly benefit clinicians.
Applying the Evidence in Daily Practice
The Institute of Medicine (IOM) focuses on the promise of evidence-based medicine to improve the quality and effectiveness of health care, and has also highlighted barriers in the current system. The IOM cites “an irony of the information-rich environment is that information important to clinical decision making is often not available, or is provided in forms that are not relevant to the broad spectrum of patients—with differing levels of health, socioeconomic circumstances, and preferences—and the issues encountered in clinical practice.” In the IOM view, these limitations are driven by a paucity of clinical effectiveness research, poor dissemination of the evidence that is available, and too few incentives and decision supports for evidence-based care. Glenton and colleagues described several factors hindering the effective use of systematic reviews for clinical decision making. They found that reviews often lacked details about interventions and did not provide adequate information on the risks of adverse events, the availability of interventions, and the context in which the interventions may or may not work.
Evaluate the Performance of the Treatment
The final step in EBM is to assess the outcome of the treatment. Did the patient (or their parents) judge their condition to be improved? Was the treatment cost-effective? Did the treatment fit within the context of the unique circumstances and biology of the family? If a similar scenario was encountered again, what would the clinician do differently? This habit for critical self-appraisal and unremitting learning is at the heart of EBM. Only by widespread implementation of the principals of EBM is healthcare quality and value likely to improve.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
