META-ANALYSIS: AN OVERVIEW WITH APPLICATION TO PHYSICAL ACTIVITY by George A. Kelley, DA, FACSM, Professor, Dept. of Co

Presentation Transcript

META-ANALYSIS: AN OVERVIEW WITH APPLICATION TO PHYSICAL ACTIVITY by George A. Kelley, DA, FACSM, Professor, Dept. of Community Medicine/Director, Meta-Analytic Research Group, West Virginia University, Morgantown, WV E-mail: gkelley@hsc.wvu.edu

Acknowledgments Kristi Kelley, M.Ed. – West Virginia University Russell L. Moore – University of Colorado Zung Vu Tran, Ph.D. – Colorado Prevention Center James Hagberg, Ph.D. – University of Maryland @ College Park Doug Seals, Ph.D. – University of Colorado @ Boulder Christine Snow, Ph.D. – Oregon State University Charlotte Sanborn, Ph.D. – Texas Woman's University William Haskell, Ph.D. – Stanford University

Objectives List and depict the 9 levels of the confirmation based pyramid List and portray the 2 sorts of precise audits List and portray the 3 fundamental things to consider while assessing a deliberate survey List and portray the 4 methodological strides for directing a meta-investigation

Topics Evidence-Based Medicine Overview of Systematic Reviews Application of Systematic Review

I. Prove Based Pyramid Source: SUNY Downstate Medical Center . Therapeutic Research Library of Brooklyn. Confirm Based Medicine Course. A Guide to Research Methods: The Evidence Pyramid: http://servers.medlib.hscbklyn.edu/ebmdos/2100.htm

A. Preferences II. Outline of Systematic Reviews Study address particular & limit Data accumulation far reaching & particular Study choice in view of consistently connected criteria

B. Sorts of Systematic Reviews Notes: Quantitative is meta-examination; SM, Summary means, IPD, singular patient information; *, IPD meta-examination best however frequently hard to get IPD from agents (Sources: Kelley GA, Kelley KS, Tran ZV. Recovery of individual patient information for a practice meta-investigation. Am J Med Sports 2002;4:350-4; Duval S, Vazquez G, Baker WL, Jacobs DR. The Collaborative Study of Obesity and Diabetes in Adults (CODA) extend meta-examination outline and depiction of taking an interest considers. Weight Reviews 2007;8:263-76).

Number of Citations from a PubMed Search utilizing Keyword "meta-investigation" (Accessed 09/01/07)

C. Significance Need to utilize a methodical approach for orchestrating data Proliferation of data in today's general public (Evidence-based prescription note pad. On the requirement for confirmation based prescription. Confirm Based Medicine 1995;1:5-6)

D. Evaluation of Systematic Reviews Validity of trial technique Magnitude and accuracy of treatment impacts Application of the outcomes to your patient or populace Source: Akobeng AK. Understanding deliberate surveys and meta-examination. Curve Dis Child 90(8); 845-848, 2005.

III. Use of Systematic Review (Summary Means Meta-Analysis) Kelley GA, Kelley KS, Tran ZV. Strolling and non-HDL-C in grown-ups: A meta-examination of randomized controlled trials. Prev Cardiol 8(2);102-107, 2005. Financed by the National Institutes of Health, National Heart, Lung, and Blood Institute (NIH-R01-HL 069802), G.A. Kelley, Principal Investigator

Rationale for study Non-high thickness lipoprotein cholesterol best indicator of CVD (Pischon T, Girman CJ, Sacks FM, Rifai N, Stampfer MJ, Rimm EB. Non-High-Density Lipoprotein Cholesterol and Apolipoprotein B in the Prediction of Coronary Heart Disease in Men. Course 2005;112(22)33:75-83) Physical movement suggested for enhancing lipids/lipoproteins (National Cholesterol Education Program; National Heart Lung and Blood Institute; National Institutes of Health. Third Report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III) Final Report. NIH Publication No. 02-5215 , September 2002)

Walking most basic type of physical movement in US (Eyler AA, Brownson RC, Bacak SJ, Housemann RA. The study of disease transmission of strolling for physical action in the United States. Med Sci Sports Exerc 2003;35(9)1529-36) RCTs have not revealed the impacts of strolling on non-HDL-C

B. Strategies 1. Information Sources Computer seeks (Medline, Embase, Sport Discus, Current Contents, Dissertation Abstracts) Cross-referencing from past survey articles/unique reviews Expert to audit reference list (W. Haskell, individual correspondence)

2. Think about Selection Randomized controlled trials Walking program for > two months Adult people > 18 years old English-dialect just Published & unpublished ( 1955 to 2003) TC and HDL-C surveyed

3. Information Abstraction C oding sheet (more than 200 things/consider) Major factors coded: think about, subject, & preparing program attributes, essential & auxiliary results Dual-coding, free of each other Every thing assessed for precision

4. Measurable Analysis Overall outcomes (essential/auxiliary results) Changes in pre to post contrasts in strolling and control amasses Each result weighted by the reverse of the difference Multilevel model (two-arrange) with an arbitrary impacts part & 95% CIs for pooled results (Dersimonian R, Laird N. Meta-examination in clinical trials. Control Clin Trials 1986;7:177-88) dj = γ 0 + uj + ej

Heterogeneity utilizing the Q measurement and p < 0.10 (Cochran WG. The blend of evaluations from various trials. Biometrics 1954;10:101-29) Inconsistency as I 2 = 100% x (Q - df)/Q (Higgins JPT, Thompson SG, Deeks JJ, Altman DG. Measuring irregularity in meta-investigations. Br Med J 2003;327:557-60) Study quality – (0 to 5 scale) (Jadad AR, Moore RA, Carroll D, Jenkinson C, Reynolds DJM, Gavaghan DJ, McQuay HJ. Evaluating the nature of reports of randomized clinical trials Is blinding important? Control Clin Trials 1996;17:1-12) Publication inclination – Trim & Fill (Duval S, Tweedie R. Trim and fill a straightforward pipe plot-based strategy for testing and altering for distribution inclination in meta-examination. Biometrics 2000;5:64:55-63)

Categorical Analysis – Multilevel (two-organize) arbitrary impacts ANOVA-like models for meta-examination at p < 0.01 Multilevel model (two-arrange) with covariate & irregular impacts part (p < 0.01) dj = γ0 + γ1 W 1 j + u j + ej

C. Comes about Table 1. Ponder qualities

Table 2. Starting attributes of subjects. N, number of gatherings detailing information; M + SD, mean + standard deviation; BMI, body mass file; VO 2max , most extreme oxygen utilization; TC, add up to cholesterol; HDL-C, high-thickness lipoprotein cholesterol; Non-HDL-C, non-high thickness lipoprotein cholesterol, figured as aggregate cholesterol less high-thickness lipoprotein cholesterol .

Table 3. Strolling program attributes Notes: N, number of gatherings; M + SD, mean + standard deviation; Compliance, rate of practice sessions went to.

Table 4. Essential and optional results. N, number of results; BMI, body mass record; VO 2max , most extreme oxygen utilization; TC, add up to cholesterol; HDL-C, high-thickness lipoprotein cholesterol; Non-HDL-C, non-high thickness lipoprotein cholesterol (add up to cholesterol short high-thickness lipoprotein cholesterol); CI, Confidence interim; *, altogether not the same as zero (0).

Table 5. Relative Changes in Outcomes

Categorical and Regression Analysis Changes in BMI and changes in Non-HDL-C (r = 0.46, p = 0.004)

D. CONCLUSIONS 1. Strolling is related with clinically vital diminishments in non-HDL-C in grown-up people. Pedersen TR, Olsson AG, Faergeman O, Kjekshus J, Wedel H, Berg K, Wilhelmsen L, Haghfelt T, Thorgeirsson G, Pyorala K, et al. Lipoprotein changes and decrease in the occurrence of real coronary illness occasions in the Scandinavian Simvastatin Survival Study (4S). Course 1998;97(15):1453-60.

E. Constraints Availability of information Because of numerous tests, noteworthy relapse results could be because of the play of chance Efficacious yet not known whether compelling

THANK YOU FOR LISTENING Crooked River, Idaho