Systematic Reviews their contribution to knowledge Morag Heirs

Systematic Reviews: their contribution to knowledge Morag Heirs

Research Fellow Centre for Reviews and Dissemination University of York t: +44 (0)1904 321070 f: +44 (0)1904 321041 e: mkc 500@york. ac. uk www. york. ac. uk/inst/crd Ph. D student (NIHR funded) Health Sciences University of York

What is a systematic review �Systematic Review of a clearly formulated question that uses explicit methods to minimise bias in the location, selection, critical evaluation and synthesis of research evidence. (may or may not involve quantitative synthesis) �Meta-analysis Statistical techniques used to combine the results of two or more studies and obtain a pooled (combined) estimate of effect. (informative meta-analysis will usually also be a systematic review)

Traditional reviews �‘Unscientific’ rarely pre-specify or make methods explicit �Usually subjective, opinions of individual �Often incomplete, filing cabinet or MEDLINE review �Difficult to make sense of conflicting or equivocal trials on qualitative reading alone

Why we need systematic reviews Synthesis �Health care providers, researchers and policy makers are inundated with unmanageable amounts of information �Need systematic reviews to summarise existing information and provide data for rational decision making �Enable practitioners to keep up to date with evidence accumulating in field and to practice evidence-based medicine

Why we need systematic reviews Totality of evidence �Evaluations and recommendations should be based on results of all trials �not just published / well known trials that are likely to be biased towards positive (publication bias) �Results of any one trial should be interpreted in the context of all relevant evidence �consistency / inconsistency �generalisability

Why we need systematic reviews Power and precision �Often the benefits that can be expected of a new intervention are moderate �These moderate benefits can be important clinically and in terms of public health �Often trials recruit too few patients to detect such differences with reliability

Definitions and dilemmas �Systematic reviews �Are not restricted to including randomised controlled trials (RCTs) alone �Do not always include meta-analysis (but should always provide a synthesis) �Are a flexible and powerful methodology for answering a variety of questions

Appraising systematic reviews �Adequate search �Defined inclusion criteria – appropriate choices to answer the question �Study selection/quality assessment/data extraction �Avoidance of bias and error �Synthesis (narrative, statistical, qualitative) taking into account quality of the primary studies

Conclusions �Systematic reviews ≠ meta-analyses �Systematic reviews ≠ only looking at RCTs �Important to assess primary studies for risk of bias �Quality assess the systematic review itself �Systematic reviewing is a well established, adaptable methodology suitable for most topics and questions