research techniques made simple: the role of systematic reviews and meta-analysis in dermatology...

10
Research Techniques Made Simple: The Role of Systematic Reviews and Meta-Analysis in Dermatology Katrina Abuabara, MD, MA 1 Esther E Freeman MD, PhD 2 ; Robert Dellavalle, MD, PhD, MSPH 3 Affiliations: 1 Department of Dermatology, University of Pennsylvania; 2 Department of Dermatology, Harvard Combined Dermatology Residency Program, Massachusetts General Hospital; 3 Dermatology Service, Denver VA Medical Center

Upload: sarah-dixon

Post on 26-Mar-2015

220 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Research Techniques Made Simple: The Role of Systematic Reviews and Meta-Analysis in Dermatology Katrina Abuabara, MD, MA 1 Esther E Freeman MD, PhD 2

Research Techniques Made Simple: The Role of Systematic Reviews and

Meta-Analysis in Dermatology 

Katrina Abuabara, MD, MA1 Esther E Freeman MD, PhD2; Robert Dellavalle, MD, PhD, MSPH3

Affiliations: 1Department of Dermatology, University of Pennsylvania; 2Department of Dermatology, Harvard 

Combined Dermatology Residency Program, Massachusetts General Hospital; 3Dermatology Service, Denver VA Medical 

Center 

Page 2: Research Techniques Made Simple: The Role of Systematic Reviews and Meta-Analysis in Dermatology Katrina Abuabara, MD, MA 1 Esther E Freeman MD, PhD 2

Summary Points

• A systematic review is a comprehensive summary of available data pertaining to a specific question, organized through a rigorous design. Systematic reviews often contain a meta-analysis, which is a statistical method for synthesizing data from multiple studies. 

• These techniques are used to answer specific research questions and may minimize bias, improve precision of intervention estimates, and increase the statistical power of identifying a real effect. They can also help to settle controversy when individual studies show conflicting results, and they may identify research gaps. 

Page 3: Research Techniques Made Simple: The Role of Systematic Reviews and Meta-Analysis in Dermatology Katrina Abuabara, MD, MA 1 Esther E Freeman MD, PhD 2

Study design methodology• The PRISMA (Preferred Reporting Items for 

Systematic reviews and Meta-Analyses) Statement describes criteria for developing, carrying out, and evaluating these types of studies.

• Formulating an appropriate and specific research question is critical. 

• Eligibility criteria should be prespecified and a search strategy for included studies should be comprehensive and clearly described; these are often presented in a flow diagram.

Page 4: Research Techniques Made Simple: The Role of Systematic Reviews and Meta-Analysis in Dermatology Katrina Abuabara, MD, MA 1 Esther E Freeman MD, PhD 2

Example of a Flow Diagram 

*Originally published with the PRISMA Statement (Liberati A, Altman DG, Tetzlaff J, et al. (2009) PLoS Medicine 6:e1000100)

Page 5: Research Techniques Made Simple: The Role of Systematic Reviews and Meta-Analysis in Dermatology Katrina Abuabara, MD, MA 1 Esther E Freeman MD, PhD 2

Assessing Risk of Bias in Individual Studies

• A descriptive approach of assessed methodological components such as selection bias, performance bias, detection bias, attrition bias, and reporting bias is recommended. 

Page 6: Research Techniques Made Simple: The Role of Systematic Reviews and Meta-Analysis in Dermatology Katrina Abuabara, MD, MA 1 Esther E Freeman MD, PhD 2

Statistical Meta-analysis

• Meta-analysis allows data from multiple studies to be statistically combined. 

• First, a summary statistic is calculated for each study. 

• Next, an overall intervention effect estimate is calculated. 

• Finally, to ensure that the results are valid and robust, researchers typically test for heterogeneity and publication bias and perform sensitivity analyses. 

• Results are often presented in a forest plot.

Page 7: Research Techniques Made Simple: The Role of Systematic Reviews and Meta-Analysis in Dermatology Katrina Abuabara, MD, MA 1 Esther E Freeman MD, PhD 2

Forest Plot

• Used to present the results of a meta-analysis.• The summary statistic for each study is shown as a 

square with a horizontal line indicating the confidence interval. 

• At the bottom of the graphic the overall intervention effect estimate is represented by a diamond, with the center showing the point estimate and the horizontal tips illustrating the confidence interval. 

• The significance of each study and the overall estimate are highlighted by whether they cross a vertical line of no effect. 

Page 8: Research Techniques Made Simple: The Role of Systematic Reviews and Meta-Analysis in Dermatology Katrina Abuabara, MD, MA 1 Esther E Freeman MD, PhD 2

Example of a Forest Plot

*Originally published with the PRISMA Statement. (Liberati A, Altman DG, Tetzlaff J, et al. (2009) PLoS Medicine 6:e1000100)

SB
COMP: In the second to last and last lines of Figure 2, please change "hetrogeneity" to "heterogeneity" and "overal" to "overall"
Page 9: Research Techniques Made Simple: The Role of Systematic Reviews and Meta-Analysis in Dermatology Katrina Abuabara, MD, MA 1 Esther E Freeman MD, PhD 2

Limitations of Systematic Reviewsand Meta-analyses 

• Limitations include the risk of misleading results if individual studies are biased or their reporting is not standardized. 

• Although often useful for summarizing an intervention effect from randomized controlled trials, systematic reviews and meta-analyses are less effective for capturing adverse effects or summarizing observational studies. 

Page 10: Research Techniques Made Simple: The Role of Systematic Reviews and Meta-Analysis in Dermatology Katrina Abuabara, MD, MA 1 Esther E Freeman MD, PhD 2

Rating the Strength of Recommendations 

• GRADE– Used by the World Health Organization in its guideline 

development process– Classifies strength of recommendations into strong or weak. A 

strong recommendation means that based on the available evidence, clinicians are very certain that that benefits either do or do not outweigh the risks of an intervention.

– Website: http://www.gradeworkinggroup.org/intro.htm• SORT

– Used by the American Academy of Dermatology– Grades strength of recommendation into A, B, and C, with A-

level evidence based on consistent, good-quality, patient-oriented evidence and C based on consensus, usual practice, or opinion. 

– Website:  http://www.aad.org/education-and-quality-care/clinical-guidelines/guideline-development-process

SB