2009 Systematic Review Idc
This document was uploaded by user and they confirmed that they have the permission to share it. If you are author or own the copyright of this book, please report to us by using this DMCA report form. Report DMCA
Overview
Download & View 2009 Systematic Review Idc as PDF for free.
More details
- Words: 1,020
- Pages: 38
Systematic Reviews
Felix Eduardo R. Punzalan, MD Dept. of Medicine, Section of Cardiology Dept. of Clinical Epidemiology
Scenario • To resolve conflicting evidence • To answer questions where the answer is uncertain • To explain variations in practice
Scenario • Have there been other good studies of the same question? • Is this study the only one to show an effect? Are there other studies?
Reviews - done in many different ways - different strengths and weaknesses
TYPES OF INTEGRATIVE STUDIES
REVIEW = any 2 or more articles SYSTEMATIC REVIEW= comprehensive, systematic and objective search META-ANALYSIS=results of trials are combined statistically
Traditional Reviews - summary of evidence and recommendations - usually, broad-gauged questions “management of the diabetic patients” - experts know the literature have the actual practice - But lack structure articles are cited but personal experience and conventional wisdom strength of original researches not known
Shorthand Indicators of Quality 1. 2. 3. 4.
prestige of journal author currency/latest number of articles for and against a given point
5. design
# of RCTs
SYSTEMATIC REVIEWS AND META-ANALYSES COMPREHENSIVE SYSTEMATIC OBJECTIVE
Systematic Reviews - rigorous reviews of specific questions - useful in addressing a single, focused question Ex: Effect Aspirin in prevention of cardiovascular events Effect Statin treatment in prevention of cardiovascular events
Systematic Reviews - rigorous reviews of specific questions - useful in addressing a single, focused question Ex: ACE inhibitors of CHF skin adhesives vs. sutures for superficial lacerations
Elements of Systematic Reviews 1. Define the clinical question. 2. Identify all completed studies of the question, published and unpublished. 3. Select the studies that meet high standards for scientific validity. 4. Look for evidence of bias in the studies selected. 5. Describe the scientific quality of the studies.
Elements of Systematic Reviews 6. Ask whether the quality is systematically related to results of the studies. 7. Describe the studies with a figure (forest plot). 8. Decide whether the studies are similar enough to justify combining them. 9. If they are similar enough to combine, calculate a summary measure of effect and confidence interval.
Finding all Relevant Studies • • • •
Search MEDLINE. Read recent reviews and textbooks. Seek the advice of experts in the content area. Consider articles cited in the articles already found by other approaches. • Consult databases of articles such as the Cochrane Database of Systematic Reviews. • Review registries of clinical trials (to detect publication bias, if present).
Limiting to Scientifically Strong Studies 1. CONSORT Criteria for RCTs Moher D, et al Annal Intern Med 2001:134:657-662
2. STARD criteria diagnostic test performance Bossuyt PM, et al Annal Intern Med 2003:138:40-44
Publication Bias • Positive vs. negative studies • editors’ bias - registration of studies • industry - sponsors requirements • language
Funnel Plots Way of detecting bias in the selection of studies for systematic reviews plot main results vs. sample size effects size of small trials vs large trials
Fig. 12.1
How Good are the Best Studies? A Simple Scale for Measuring Quality of Randomized Controlled Trials Questions: 1. Was the study described as randomized? 2. Was the study described as double-blind? 3. Was there a description of withdrawals and dropouts? Scoring: Give a score of 1 point for each “yes” and 0 points for each “no”. There are no in-between marks.
Is Scientific Quality Related to Research Results? – critically examine each papers in a systematic manner
How Good are the Best Studies? A Simple Scale for Measuring Quality of Randomized Controlled Trials Questions: 1. Was the study described as randomized? 2. Was the study described as blinded? 3. Were the groups comparable? 4. Was follow up adequate?
Is Scientific Quality Related to Research Results? – critically examine each papers in a systematic manner
Summarizing Results Forest Plots - points estimates and confidence interval of each studies Six Kinds of Useful Information in a Forest Plot 1. Number of studies 2. References to publications of component studies 3. The pattern of effect sizes 4. The number of studies that are statistically significant 5. Comparison between large vs small studies 6. Chronological listing show how results may change in time and and when size increases
Meta-Analysis • Practice of combining results of individual studies (or patients in these studies) • Provided the studies are similar enough to be combined
Summarizing Observational and Diagnostic Test Studies
Summarizing Diagnostic Test
Are the studies similar enough • • • •
patient intervention follow-ups outcomes
2 General approaches in pooling study results 1. Statistical test for homogeneity - The extent to which trials results are similar
2. Make an informed judgment on the similarities of the judgment
What is Combined-Studies or Patients? • usual study results are combined • more powerful - individual patient data of component studies Ex. Anti-thrombotic Trialists’ Collaboration - effect of antiplatelet therapy on vascular events in patient at high risk for CVD 195 clinical trials, 200 patients/trial Overall – decrease vascular events by ¼ Patient-level data – decreased among patients who had MI, stable angina and fibrillation, and cardiac valve surgery
How are the Results Summarized 2 Kinds of Mathematical Model 3. Fixed-effect Model All studies are asking similar questions
4. Random-effects Model Studies are asking somewhat different but related questions
Figure 12.5
Summary Effect
Cumulative Meta-Analysis - another way of looking at same information - new survey effect age and confidence calculated each time results of the study become available
Fig. 12.7
Advantages and Risk of Combining Studies • Advantage: - can do subgroup analysis - detecting rare events • Disadvantage: - may give misleading impression
Advantages Combining Studies • Advantage: - can do subgroup analysis - detecting rare events
• Ask – Focused Clinical Question • Acquire – Systematic, comprehensive • Appraise • Assess Results – combined result • Apply – summarize conclusion and apply to practice
• Among patients at risk of CV events, what is the effect of statins in preventing CV outcomes? • Among patients at risk of CV events, what is the effect of aspirin in preventing CV outcomes?
Summary • • • • •
Systematic Reviews Traditional Review Elements of a Systematic Review Summarizing Results Meta-analysis
Felix Eduardo R. Punzalan, MD Dept. of Medicine, Section of Cardiology Dept. of Clinical Epidemiology
Scenario • To resolve conflicting evidence • To answer questions where the answer is uncertain • To explain variations in practice
Scenario • Have there been other good studies of the same question? • Is this study the only one to show an effect? Are there other studies?
Reviews - done in many different ways - different strengths and weaknesses
TYPES OF INTEGRATIVE STUDIES
REVIEW = any 2 or more articles SYSTEMATIC REVIEW= comprehensive, systematic and objective search META-ANALYSIS=results of trials are combined statistically
Traditional Reviews - summary of evidence and recommendations - usually, broad-gauged questions “management of the diabetic patients” - experts know the literature have the actual practice - But lack structure articles are cited but personal experience and conventional wisdom strength of original researches not known
Shorthand Indicators of Quality 1. 2. 3. 4.
prestige of journal author currency/latest number of articles for and against a given point
5. design
# of RCTs
SYSTEMATIC REVIEWS AND META-ANALYSES COMPREHENSIVE SYSTEMATIC OBJECTIVE
Systematic Reviews - rigorous reviews of specific questions - useful in addressing a single, focused question Ex: Effect Aspirin in prevention of cardiovascular events Effect Statin treatment in prevention of cardiovascular events
Systematic Reviews - rigorous reviews of specific questions - useful in addressing a single, focused question Ex: ACE inhibitors of CHF skin adhesives vs. sutures for superficial lacerations
Elements of Systematic Reviews 1. Define the clinical question. 2. Identify all completed studies of the question, published and unpublished. 3. Select the studies that meet high standards for scientific validity. 4. Look for evidence of bias in the studies selected. 5. Describe the scientific quality of the studies.
Elements of Systematic Reviews 6. Ask whether the quality is systematically related to results of the studies. 7. Describe the studies with a figure (forest plot). 8. Decide whether the studies are similar enough to justify combining them. 9. If they are similar enough to combine, calculate a summary measure of effect and confidence interval.
Finding all Relevant Studies • • • •
Search MEDLINE. Read recent reviews and textbooks. Seek the advice of experts in the content area. Consider articles cited in the articles already found by other approaches. • Consult databases of articles such as the Cochrane Database of Systematic Reviews. • Review registries of clinical trials (to detect publication bias, if present).
Limiting to Scientifically Strong Studies 1. CONSORT Criteria for RCTs Moher D, et al Annal Intern Med 2001:134:657-662
2. STARD criteria diagnostic test performance Bossuyt PM, et al Annal Intern Med 2003:138:40-44
Publication Bias • Positive vs. negative studies • editors’ bias - registration of studies • industry - sponsors requirements • language
Funnel Plots Way of detecting bias in the selection of studies for systematic reviews plot main results vs. sample size effects size of small trials vs large trials
Fig. 12.1
How Good are the Best Studies? A Simple Scale for Measuring Quality of Randomized Controlled Trials Questions: 1. Was the study described as randomized? 2. Was the study described as double-blind? 3. Was there a description of withdrawals and dropouts? Scoring: Give a score of 1 point for each “yes” and 0 points for each “no”. There are no in-between marks.
Is Scientific Quality Related to Research Results? – critically examine each papers in a systematic manner
How Good are the Best Studies? A Simple Scale for Measuring Quality of Randomized Controlled Trials Questions: 1. Was the study described as randomized? 2. Was the study described as blinded? 3. Were the groups comparable? 4. Was follow up adequate?
Is Scientific Quality Related to Research Results? – critically examine each papers in a systematic manner
Summarizing Results Forest Plots - points estimates and confidence interval of each studies Six Kinds of Useful Information in a Forest Plot 1. Number of studies 2. References to publications of component studies 3. The pattern of effect sizes 4. The number of studies that are statistically significant 5. Comparison between large vs small studies 6. Chronological listing show how results may change in time and and when size increases
Meta-Analysis • Practice of combining results of individual studies (or patients in these studies) • Provided the studies are similar enough to be combined
Summarizing Observational and Diagnostic Test Studies
Summarizing Diagnostic Test
Are the studies similar enough • • • •
patient intervention follow-ups outcomes
2 General approaches in pooling study results 1. Statistical test for homogeneity - The extent to which trials results are similar
2. Make an informed judgment on the similarities of the judgment
What is Combined-Studies or Patients? • usual study results are combined • more powerful - individual patient data of component studies Ex. Anti-thrombotic Trialists’ Collaboration - effect of antiplatelet therapy on vascular events in patient at high risk for CVD 195 clinical trials, 200 patients/trial Overall – decrease vascular events by ¼ Patient-level data – decreased among patients who had MI, stable angina and fibrillation, and cardiac valve surgery
How are the Results Summarized 2 Kinds of Mathematical Model 3. Fixed-effect Model All studies are asking similar questions
4. Random-effects Model Studies are asking somewhat different but related questions
Figure 12.5
Summary Effect
Cumulative Meta-Analysis - another way of looking at same information - new survey effect age and confidence calculated each time results of the study become available
Fig. 12.7
Advantages and Risk of Combining Studies • Advantage: - can do subgroup analysis - detecting rare events • Disadvantage: - may give misleading impression
Advantages Combining Studies • Advantage: - can do subgroup analysis - detecting rare events
• Ask – Focused Clinical Question • Acquire – Systematic, comprehensive • Appraise • Assess Results – combined result • Apply – summarize conclusion and apply to practice
• Among patients at risk of CV events, what is the effect of statins in preventing CV outcomes? • Among patients at risk of CV events, what is the effect of aspirin in preventing CV outcomes?
Summary • • • • •
Systematic Reviews Traditional Review Elements of a Systematic Review Summarizing Results Meta-analysis
Related Documents
2009 Systematic Review Idc
May 2020 6
Idc Review
December 2019 6
Nus Systematic Review Assignment
May 2020 7
Systematic Review Appraisal Tools
June 2020 6
Idc
July 2020 6
Dil-idc
April 2020 16More Documents from ""