Annie George BHSc1, Ydo V. Kleinlugtenbelt MD PhD2 and Kim Madden PhD1,3 1 Department of Health Research Methods, Evidence, and Impact, McMaster University, Hamilton, ON, Canada 2 Department of Orthopaedic Surgery, Deventer Hospital, Deventer, Netherlands 3 Division of Orthopaedic Surgery, Department of Surgery, McMaster University, Hamilton, ON, Canada The number of clinical studies in the field of orthopedics is overwhelmingly large and continually growing. It stands to reason that some of this research is of higher quality than others. How do we know which studies we can trust? Proponents of evidence‐based medicine have developed a hierarchy of evidence which divides studies into groups according to study design to highlight the foundational role that appropriate study design plays in study quality. It should be noted that study design is not the only factor involved in study quality, but it does form the basis for making a decision about the trustworthiness or credibility of the evidence. It is important to have a thorough understanding of the theoretical underpinnings of each major study design and how these designs and methodological decisions affect study quality and credibility of the evidence. This principle of a hierarchy of evidence became prominent in the early 1990s as evidence‐based medicine was formalized and physicians started to appraise and apply evidence to their practice,1,2 and this developed into a pyramid structure, with the best evidence placed at the top of the pyramid and called level I evidence, and the lowest‐quality evidence at the bottom, called level V evidence (Figure 2.1). Figure 2.1 Levels of evidence pyramid for therapy studies. There are several major types of research questions that researchers can answer. These are typically classified into therapy, prognosis, harm, diagnosis, and economic questions.3 In this chapter we will focus on those studies addressing therapy, as this is generally the most common type of study in the orthopedic surgical literature. In the case of therapeutic trials, RCTs and meta‐analyses of RCTs are considered the “best evidence.” This is because randomization allows investigators to control for both known and unknown prognostic factors, which cannot be done with other study designs. High‐quality observational studies can control very well for known prognostic factors, but only well‐done randomization with allocation concealment can control for all prognostic variables. Allocation concealment ensures that the investigators and the participants cannot manipulate the treatment group that they are randomly assigned to. The second level of evidence is prospective cohort studies. This study design involves two (or more) groups of participants who are exposed to a factor of interest, for example direct anterior versus lateral surgical approach in total hip arthroplasty. The difference between a prospective cohort study and an RCT is that the participants are not randomized to groups: they are assigned by choosing who is in each group or by some other nonrandom means. Modern statistical methods are excellent at controlling for prognostic variables in large high‐quality prospective cohort studies, but we can never be sure that all unknown prognostic variables are accounted for, as with RCTs. This is why this design is on the second level of the hierarchy of evidence. It should be noted that some groups put retrospective cohort studies on the second level of evidence, but some put them on the third level due to their retrospective nature. For example, the Oxford Centre for Evidence‐Based Medicine’s well‐used 2009 hierarchy of evidence table puts all cohort studies on level II regardless of whether they are retrospective or prospective.3 However, the Journal of Bone and Joint Surgery puts retrospective cohort studies in level III with other retrospective studies.4 The third level of evidence is case‐control studies. This study design is unique in that it starts with an outcome and looks backward to determine if there was a particular exposure of interest. Case‐control studies are useful for rare outcomes or cases where there are long latency periods between an exposure and an outcome. However, they are vulnerable to recall bias and are not randomized. The fourth level on the hierarchy is case reports and case series. These studies are often retrospective in nature (but can be prospective) and they are characterized by having no comparison group. This means that we cannot compare the treated/exposed group to untreated/unexposed controls, and it is therefore very difficult to determine whether the participants would have improved (or not) had they not received the treatment/exposure. Finally, the lowest level on the hierarchy is expert opinion. Experts have worked hard to develop their expertise usually over a period of many years. However, there are myriad biases that an expert can encounter that can color their opinions. Systematic research, such as the levels of evidence and study designs noted above, are theoretically more sound than expert opinion alone. It should be noted that the study design is not the only consideration when assigning a level of evidence to a study. Studies can be downgraded for poor quality. For example, if an RCT has a major methodological flaw that leads to bias, it can be downgraded to level II evidence. RCTs, if conducted rigorously, are the gold standard for ascertaining the effectiveness and safety of a treatment.2
2 Hierarchy of Evidence and Common Study Designs
Introduction
Top five questions
Question 1: What is the hierarchy of evidence for therapy studies?
Question 2: What are randomized controlled trials (RCTs)?
Related posts:
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
