Article Text
PDF
Statistics from Altmetric.com
‘More research is needed’ is a most unhelpful conclusion for the clinician reader of a systematic review. Clinicians look to researchers for the ‘research evidence’ part of Professor Sackett’s ‘three circles’ of evidence-based practice.1 The clinician can not ignore the patient in front of him or her just because there is no evidence from systematic reviews or randomised controlled trials (RCTs): ‘Kindly forget about your shoulder problem, until we have better research to tell me what to do’? This will not do! In this editorial, we share three practical tips for authors synthesising lower level evidence in systematic reviews, when high-quality RCTs are absent. Our goal is to help authors help clinicians help patients.
Consider the patient who plays tennis and presents with shoulder pain. Right now, the clinician who is seeking RCT evidence on how to progress from impairment-focused tasks in neutral shoulder positions to high-level, sport-specific tasks will be disappointed. Being able to function in shoulder elevation positions above 90° is critical for overhead athletes. But research evidence is lacking. This is a major gap because sport-specific rehabilitation is an important progression towards returning to sport.
When we have few (or no) high-quality studies, it is reasonable to consider lower quality evidence. But this must be done judiciously and with a clear data synthesis plan. Lower quality research should not be assigned the same weight in a synthesis as higher quality research, which means there are three key things to consider: (1) which evidence to include, (2) the quality of the evidence and (3) how to combine all available evidence in a meaningful synthesis.
Expert opinion can be valid
If RCTs are not available to answer a clinical question, as in our shoulder example, we need to consider other study designs. If no (randomised) controlled trials or observational studies are available, consider expert opinion that comes from clinical experience—often, research is playing catch up to clinical practice.2 The advantage of ‘practice-based evidence’3 is that, if carefully selected, it can fill in important gaps in the research-based evidence—illustrating how a clinical question is addressed in current clinical practice, when research has not yet addressed that question.
Clinical practice guidelines/position statements
One step higher than expert opinion is the opinion from a group of topic experts, in clinical practice guidelines or consensus statements, especially if these have been endorsed by reputable professional organisations. Quality examples are the practice guidelines for anterior cruciate ligament rehabilitation4 endorsed by the Royal Dutch Society for Physical Therapy; or the Australasian College of Sports Physicians position statement on the use of stem cell therapy in sport and exercise medicine.5 Checking for intellectual and financial conflict of interest is important when using consensus statements and clinical practice guidelines. Endorsement from a professional organisation is one hallmark of a reputable consensus statement or clinical practice guideline. But beware! Some ‘respected’ organisations have major, declared, corporate partners.6 That does not invalidate a statement but it warrants documenting and taking into careful consideration.
Combining the evidence in a meaningful and robust synthesis
Quality systematic reviews are underpinned by robust syntheses. When authors synthesise evidence from across the spectrum of ‘levels of evidence’, they must carefully consider how to weigh evidence. Higher quality evidence (eg, RCT) should contribute more to a decision than lower quality evidence (eg, expert opinion). However, in the shoulder example, where RCTs do not address the specific needs of the overhead athlete, expert opinion can still provide helpful ‘evidence’. A best evidence synthesis approach, where evidence is weighted according to quality, and combined for overall statements, can provide meaningful information without pooling data statistically. It is up to authors to decide when it is appropriate to include lower levels of evidence, and when to prioritise lower levels of evidence. Authors should be explicit in every case and provide strong justification if prioritising a lower level of evidence over a higher one.7
Supporting quality clinical practice
Professor John Ioannidis concluded that ‘many systematic reviews are false, misleading and unnecessary’.8 9 Therefore, there is a need for high-quality systematic reviews that address clinically relevant questions. RCTs will continue to be valued as the highest levels of evidence in systematic reviews. However, other levels of evidence can be used for clinically meaningful best evidence synthesis when RCTs are not available or do not address the clinical question at hand, provided the susceptibility to bias is carefully accounted for.
BJSM prioritises quality systematic reviews that address a clear, clinically relevant question and provide value for the clinician. We look forward to receiving and publishing (and awarding! See the BJSM blog for past winners of the systematic review competition) such clinically relevant systematic reviews, with considered findings.
Footnotes
Contributors CLA and MW contributed equally to the production of this editorial. CLA is a BJSM Deputy Editor responsible for systematic reviews. MW is a BJSM Senior Associate Editor responsible for systematic reviews.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.