Evidence Analysis Projects and Toolkits Esther Myers, PhD, RD, FADA Academy of Nutrition and Dietetics (AND) Formerly called the American Dietetic Association emyers@eatright.org Research and Strategic Business Development Team 12/9/2012 Overview • Introduce Academy of Nutrition and Dietetics EAL concept • Content in EAL ® • Users of the EAL • • • • Evidence Analyses process Guideline & Toolkit Development process Usability/feasibility testing EBNPG &Toolkit Other products from EAL ® 2 Academy of Nutrition & Dietetics (Formerly called American Dietetic Association) •Largest food and nutrition professional organization in the world •+71,800 members • • • • +51,000 Registered Dietitians +1500 Dietetic Technician Registered +17,000 students +450 International members •28 Dietetic Practice Groups (DPG) • • Diabetes Care and Education = 6187 members American Overseas Dietitians = 888 members 3 Academy’s Commitment The Academy believes that research is the foundation of the profession providing the basis for practice, education and policy. Castellanos, Myers, Shanklin, Research Priorities Contribute to Bright Future, JADA, 2004 4 Academy’s Evidence Analysis Library: The best available research on nutrition topics in a online practitioner-friendly website and includes Evidence-Based Nutrition Practice Guidelines. The EAL was launched in September 2004. It contains a significant amount of analysis completed on over 42 different disease/conditions, food, nutrients and other topic areas. The Academy is proud of it’s rigorous process of evidence analysis which was recognized by JCAHO as exemplary of bringing the best research to practice and served as the template for the Nutrition Evidence Library used by the US Federal Government for synthesizing the science to develop the 2010 US Dietary Guidelines 5 About the Evidence Analysis Library 6 19 EvidenceBased Nutrition Practice Guidelines Year Guideline 2005 • Disorders of Lipid Metabolism 2006 • Adult Weight Management • Critical Illness 2007 • Pediatric Weight Management • Oncology 2008 • Diabetes Mellitus Type 1 and 2 • Hypertension • Heart Failure • Chronic Obstructive Pulmonary Disease • Gestational Diabetes 2009 • Celiac Disease • Spinal Cord Injury • Unintended Weight Loss 2010 • Chronic Kidney Disease • HIV/AIDS 2011 • Vegetarian Nutrition •Disorders of Lipid Metabolism Update 2012 • Critical Illness Update • Food and Nutrition for Older Adults (Recommendations) How much content is on EAL®? Abstracted Articles/Worksheets Number of Published Worksheets on EAL 6000 5176 5000 4450 4000 3351 3000 1008 2005 2006 2009 1356 Sep-10 Sep-11 115 Sep-12 0 2004 2004 2007 2274 2000 1000 4677 2005 2006 2007 2009 Sep-10 Sep-11 Sep-12 8 EAL ® Usage Worldwide Users from 212 different countries As of October 1, 2012 – 15,162,099 Page Views Evaluating the Science 10 Universal Principles and Process 11 Application of Research to Practice Guidelines Two Basic Principles of Evidence Based Medicine (EBM) are: • A hierarchy of strength of evidence behind recommendations • The judgment of the clinician when weighing alternative management strategies, considering patient values and preferences, and societal values Hierarchy of Strength of Evidence Type and Description of Evidence Primary Studies • Selecting studies that are both: – Highly relevant – Study designs minimize bias • High strength of inference Guyatt et al, JAMA.2000;284(10):1290-1296 Which research design? Type of primary study preferred varies by use: Diagnosis • CrossSectional Studies Treatment • Randomized Controlled Trials Prognosis Etiology/Harm • Cohort Studies (over time with outcomes) • Cohort/Case Control (2 groups over time) Steps in Evidence Based Practice • Identify a specific problem/area of uncertainty • Formulate problem as question/analytic framework • Search for and find evidence • Evaluate the reports and grading the evidence • Form recommendation/make decisions • Summarize and disseminate findings What is the Gold Standard for Evidence Based Reviews?? 16 Cochrane Reviews Considered as gold standard by many Protocol developed for review • RCT (occasionally some quasi RCT) Review includes • Summary • Tables of evidence Central Database of RCTs and CCTs • 350,000 citations searchable by keywords TYPICALLY NO EPIDEMIOLOGICAL STUDIES http://www.cochrane.org Institute of Clinical Systems Improvement (ICSI) Methodology Classes of Research Reports Report Quality Categories Conclusion Grades I,II, III and Grade Not Assignable Greer et al, Jr Comm J Qual Improv. 2000;26(12):700-712 ICSI Evidence Grading System. Available at: http://www.icsi.org/evidence_grading_system_6/evidence_grading_system__pdf_.html. Accessed August 2007 The Academy’s Process 19 Steps in the Evidence Analysis Process Topic and Workgroup Selection Formulate Question Gather Research Appraise Articles Summarize Develop Conclusion Statement and Grade Guideline Development: Develop algorithms based on Nutrition Care Process Draft guideline recommendations Internal/external review and revise Publish guideline on EAL® Evidence Analysis Teams Roles and Responsibilities Evidence Based Practice Committee Expert Work Group Members Project Managers Lead Analysts Analysts Academy Staff 21 PICO Format Well-built clinical questions usually contain the following four elements: P opulation with a specific problem I ntervention, procedure, or approach Comparison intervention Outcome of interest 22 Sample Questions 12/9/2012 23 Inclusion/Exclusion Criteria Decisions about • • • • • Language Duplicates Abstracts Peer-reviewed Unpublished—Gray Literature Within the research report • • • • • • • • Age Disease status Setting Interventions Specific outcome measures Sample size Dropout rate Study duration 24 Search Plan & Results for Each Question •Inclusion Criteria •Exclusion Criteria •Databases Searched •Search Terms •Included Articles •Excluded Articles Search Plan & Results List reason for exclusion for each article not included in the analysis; e.g. Sample size too small 26 Critically Appraise the Research • Critically review the evidence- review each study • Worksheet developed • Quality Criteria Checklist completed -quality rating assigned to each study (positive, negative or neutral) • Based on the answers to several relevance and validity questions Example Worksheet •Citation / PubMed ID •Date •Study Design •Class •Rating (+/0/-) •Research Purpose •Inclusion Criteria •Exclusion Criteria •Description of Study Protocol •Data Collection Summary •Description of Actual Data Sample •Summary of Results •Author Conclusion Reviewer Comments Classes of Research Report PRIMARY A RCT B Cohort study C Nonrandomized trial concurrent/ historical controls Case control study Population based descriptive study D Cross sectional, Case series, Case report SYTHESIZED M Meta-analysis, systematic review, decision analysis, c/b or c/e study R Narrative review, Consensus statement/report X Medical opinion Example Quality Criteria Checklist Questions related to relevance and validity Determines if article Quality is rated as: • Positive • Negative • Neutral Summary of the findings Synthesize the studies into a narrative evidence summary and an overview table. Evidence Summary - Narrative Studies showing little or no effect of glycemic index on metabolic outcomes Gilbertson et al, 2001 found in a positive-quality 12-month randomized controlled trial of 104 children with type 1 diabetes, that children given flexible dietary advice with an emphasis on choosing low-GI foods consumed a diet with a mean 49% of the energy from carbohydrate and a mean 21 g of dietary fiber, resulting in significantly lower HbA1c levels and lower rates of hypoglycemic episodes. However, the dietary GI values between the two groups did not differ. Gilbertson et al, 2003, in a new analysis of the previous study, concluded that the children in the study did not have lower dietary quality or more limited food choices than children given traditional carbohydrate exchange advice. In a positive-quality 8-week randomized controlled trial of subjects with type 2 diabetes, Heilbronn et al, 2002 ……. Organized by topic….e.g. by direction of findings Summarizes Studies showing an effect of glycemic index on metabolic outcomes details In a positive-quality randomized controlled trial by Frost et al, 1994,key 51 subjects with type 2 diabetes were given either standard dietary advice based on theabout British Diabetic study Association recommendations or standard dietary advice with an emphasis on low in advice narrative glycemic index foods for 12 weeks. Subjects in the low glycemic index group had a lower calculated mean diet glycemic index intake (GI 77 vs 82), format and a significant correlation was detected between the glycemic index of the diet and the fall in fructosamine (r = 0.54, p < 0.01) and fasting blood glucose (r = 0.41, p < 0.05)……… …….. Summary of Findings . A total of fifteen studies examining the effect of the GI on glycemic response: • nine randomized controlled trials • seven positive-quality • two neutral-quality • two cohort studies • one positive-quality • one neutral-quality • two positive-quality population-based descriptive studies • one neutral-quality meta-analysis • one neutral-quality narrative review. Evidence Summary -Overview Table Columns customized to question Summarizes key details about study in table format Conclusion Statement Bottom Line – Answer to question based on the science Determined after research analyzed Graded based on quality of supporting evidence Grades I, II, III, IV, and V for Good, Fair, Limited, Expert Opinion only, and Not Assignable. Conclusion Grades Grade I Studies of strong design to answer questions Clinically important and consistent Adequate statistical power Grade II Studies of strong design to answer questions Some uncertainty in conclusion/inconsistent results, doubt about statistical or clinical significance or generalizability Separate studies of weaker design, all consistent Conclusion Grades (Continued) Grade III Strong design, substantial uncertainty/ inconsistent results/generalizability unclear Limited studies of weaker design or small effect size or intermediate or surrogate outcome Grade IV Expert opinion or consensus statement Grade Not Assignable Unsubstantiated by research studies Informed medical commentators based on clinical Conclusion statement What is the relationship between glycemic index and metabolic outcomes in persons with type 1 and type 2 diabetes? Conclusion Fifteen short-term studies ranging from two to 12 weeks and one longer term (one year) studies report mixed effects on A1C levels. These studies are complicated by differing definitions of "high GI" and "low GI" diets or quartiles, as well as possible confounding dietary factors. Grade II Question, Conclusion Statement and Grade Link to Evidence Summaries Link to Search Plan Evidence Based Nutrition Practice Guidelines 40 What are Evidence-Based Nutrition Practice Guidelines? Evidence Summaries & Conclusion Statements = what the evidence says Guideline = course of action for the practitioner based on the evidence Steps in the Evidence Analysis Process Analysis: Guideline Development: Formulate Question Develop algorithms based on Nutrition Care Process Gather Research Appraise Articles Draft guideline recommendations Summarize Internal/external review and revise Develop Conclusion Statement and Grade Approval By Oversight and Publish guideline on EAL® Features of Guidelines • Executive Summary of Recommendations: list only of recommendations, no supporting evidence • Introduction: scope, intent, methods, benefits/harms • Recommendations: a series of guiding statements that propose a course of action for practitioners • Algorithms: step-by-step flowchart for treatment of the specific disease/condition Executive Summary of Recommendations Free on EAL Published on National Guidelines Clearinghouse 12/9/2012 45 Executive Summary of Recommendations Recommendation Rating Label (conditional or imperative) 12/9/2012 46 Major Recommendations 12/9/2012 47 18 Major Recommendations for Type 1 and 2 DM Plus recommendations from other conditions as appropriate Weight Management Disorders of Lipid Metabolism Hypertension Recommendation Page Recommendation Page: -Recommendation -Risks/Harms -Conditions of Application -Potential Costs -Recommendation Narrative -Strength Rationale -Link to Supporting Evidence -References If the use of the glycemic index is proposed as a method of meal planning, the RD should advise on the conflicting evidence of effectiveness of this strategy. Studies comparing high versus low GI diets report mixed effects on AIC FAIR CONDITIONAL Ratings Strength •Strong – Clear benefits & strong evidence •Fair – Benefits but not as strong evidence •Weak – Benefits unclear •Consensus – Experts recommend •Insufficient Evidence Applicability •Imperative - applies to all target population •Conditional - only applicable in certain conditions…e.g. if xx…then y) 50 Recommendation Page 12/9/2012 51 12/9/2012 52 Guideline Algorithms Nutrition Assessment Algorithm Nutrition Intervention Algorithm Review and Revision Expert review Internal to Academy Membership External (MD, RN, Dentist) Revise if necessary Publish/disseminate Available Resources 55 Toolkit development 56 Objectives of Toolkits To assist RD in: • Implementing evidence-based practice • Implementing NCP/SL • Promoting consistency • Promoting quality care • Achievement of expected outcomes Toolkit Development Develop toolkits to apply guidelines Conduct 60-day usability test of toolkit and revise EBPC Review and Approval Make toolkits available for purchase Features of Evidence-Based Toolkits Set of companion documents for application of the practice guideline • Disease/condition specific • Include: • MNT protocol for treatment of disease/condition • Documentation forms (progress notes, summary notes) • Outcomes monitoring sheets • Client education resources • Case studies • Electronic downloadable purchase item Useability test 60 Nutrition Care Process & Standardized Language A means to apply the NCP in a standard way using common language. – – – – Nutrition Assessment Nutrition Diagnosis Nutrition Intervention Monitoring and Evaluation Evidence Analysis Library • Started as Academy “Member benefit” • Others wanted access • Group subscriptions for other dietetic associations (e.g. Australia, Israel) • Other health professional associations – American Association for Diabetes Educators, Association for Parenteral and Enteral Nutrition • Guidelines.gov = free access very top-line information from Guidelines • Other companion products for sale • Toolkits, powerpoint slides, consumer handouts, mobile phone app 62 EAL Recommendations at your fingertips! • NutriGuides App- NEW!! •Provides 300+ recommendations •Categorized by: –Disease/condition •Topic •Nutrition Care Process •Able to see the strength of the recommendation Available for iPhones and iPads from the iTunes store and for Android products from Google store. NutriGuides from the Evidence Analysis Library For smart iPads/iPhones and Androids Access to 300+ nutrition recommendations Easy searching $1.99 Are you using all of the EAL Resources? 65 Are you using all of the NCP resources? 66 Conclusion Value of using an evidence grading methodology Value of using all types of research in the body of evidence based on type of question being answered • Epidemiological research preferred for questions about harm, prognosis and diagnoses • RCT preferred for treatment questions • EAL contains ~ 39% observational types of research Clearly demonstrate which research studies were evaluated in tables of evidence EAL Topics are focused on a specific question Use existing research AND create new research Acknowledgements • Current and Past Evidence Based Practice Committees (~75) • Current and Past Research Committees (~150) • Evidence Workgroups (~325) • Evidence Analysts (~210) • Academy Staff • Dr Alison Steiber • Dr Paula Zeigler • Kari Kren, Mujahed Kahn, Lisa Maloney, Grace Whiteford • Toni Acosta 68 Questions? 69 Thank YOU! 70 Evidence Analysis Library •Diseases/Health Conditions •Nutrients •Foods •Life Cycle & Nutrition •Nutrition Care Process A – Z Index Click the + sign to expand a category 72 Search Engine Enter word or phrases in the search engine: View results by category 73 74