Confounding Bias: Stratification

Similar documents
Main objective of Epidemiology. Statistical Inference. Statistical Inference: Example. Statistical Inference: Example

Confounding and Interaction

RANDOMIZED CLINICAL TRIALS. Study Designs. Randomized Clinical Trials (RCT) ASSIGN EXPOSURE Follow up Check for OUTCOME. Experimental studies

Study Designs. Randomized Clinical Trials (RCT) RCT: Example 1. RCT: Two by Two Table. Outcome. Exposure. Yes a b No c d

115 remained abstinent. 140 remained abstinent. Relapsed Remained abstinent Total

Confounding, Effect modification, and Stratification

Stratified Tables. Example: Effect of seat belt use on accident fatality

INTERNAL VALIDITY, BIAS AND CONFOUNDING

Case-control studies. Hans Wolff. Service d épidémiologie clinique Département de médecine communautaire. WHO- Postgraduate course 2007 CC studies

Cohort Study: Two by two table. Study Designs. Cohort Study: Example 1. Cohort Study: Direction of inquiry. Outcome. Exposure.

Bias and confounding. Mads Kamper-Jørgensen, associate professor, Section of Social Medicine

Confounding and Effect Modification. John McGready Johns Hopkins University

Biases in clinical research. Seungho Ryu, MD, PhD Kanguk Samsung Hospital, Sungkyunkwan University

Person-years; number of study participants (number of cases) HR (95% CI) P for trend

Confounding and Bias

Strategies for data analysis: case-control studies

INTERPRETATION OF STUDY FINDINGS: PART I. Julie E. Buring, ScD Harvard School of Public Health Boston, MA

Biases in clinical research. Seungho Ryu, MD, PhD Kanguk Samsung Hospital, Sungkyunkwan University

EPIDEMIOLOGY-BIOSTATISTICS EXAM Midterm 2004 PRINT YOUR LEGAL NAME:

3. Factors such as race, age, sex, and a person s physiological state are all considered determinants of disease. a. True

Is There An Association?

Epidemiological study design. Paul Pharoah Department of Public Health and Primary Care

Confounding. Confounding and effect modification. Example (after Rothman, 1998) Beer and Rectal Ca. Confounding (after Rothman, 1998)

Biostatistics for Med Students. Lecture 1

Incorporating Clinical Information into the Label

Strategies for Data Analysis: Cohort and Case-control Studies

Outline. Case control studies. Study Designs. Case Control Study. Start with OUTCOME Go backwards Check for EXPOSURE. Experimental studies

1 Case Control Studies

Biostatistics and Epidemiology Step 1 Sample Questions Set 1

Tool to Assess Risk of Bias in Cohort Studies

Controlling Bias & Confounding

In the 1700s patients in charity hospitals sometimes slept two or more to a bed, regardless of diagnosis.

University of Wollongong. Research Online. Australian Health Services Research Institute

Study Design STUDY DESIGN CASE SERIES AND CROSS-SECTIONAL STUDY DESIGN

5 Bias and confounding

Health Studies 315. Clinical Epidemiology: Evidence of Risk and Harm

Epidemiologic Study Designs. (RCTs)

Epidemiology: Overview of Key Concepts and Study Design. Polly Marchbanks

Challenges of Observational and Retrospective Studies

Challenges in design and analysis of large register-based epidemiological studies

Consideration of Anthropometric Measures in Cancer. S. Lani Park April 24, 2009

first three years of life

Cohort studies. Training Course in Sexual and Reproductive Health Research Geneva Nguyen Thi My Huong, MD PhD WHO/RHR/SIS

Understanding Statistics for Research Staff!

Penelitian IKM/Epidemiologi. Contact: Blog: suyatno.blog.undip.ac.id Hp/Telp: /

Online Supplementary Material

Bayesian methods for combining multiple Individual and Aggregate data Sources in observational studies

Module 2: Fundamentals of Epidemiology Issues of Interpretation. Slide 1: Introduction. Slide 2: Acknowledgements. Slide 3: Presentation Objectives

Two-sample Categorical data: Measuring association

Observational Study Designs. Review. Today. Measures of disease occurrence. Cohort Studies

Cohort studies. Training course in research methodology and research protocol development Geneva Nguyen Thi My Huong, MD PhD WHO/RHR/SIS

Bias and confounding special issues. Outline for evaluation of bias

INTRODUCTION TO EPIDEMIOLOGICAL STUDY DESIGNS PHUNLERD PIYARAJ, MD., MHS., PHD.

PhD Course in Biostatistics

Using Epidemiology to Identify the Cause of Disease: Cohort study

ARTICLE REVIEW Article Review on Prenatal Fluoride Exposure and Cognitive Outcomes in Children at 4 and 6 12 Years of Age in Mexico

Study design. Chapter 64. Research Methodology S.B. MARTINS, A. ZIN, W. ZIN

Trial Designs. Professor Peter Cameron

Epidemiology EPIB-695 Final exam Monday, 10 April 2006

Observational Medical Studies. HRP 261 1/13/ am

Improved control for confounding using propensity scores and instrumental variables?

Purpose. Study Designs. Objectives. Observational Studies. Analytic Studies

Systematic Review & Course outline. Lecture (20%) Class discussion & tutorial (30%)

ADENIYI MOFOLUWAKE MPH APPLIED EPIDEMIOLOGY WEEK 5 CASE STUDY ASSIGNMENT APRIL

Understanding Confounding in Research Kantahyanee W. Murray and Anne Duggan. DOI: /pir

How do we know that smoking causes lung cancer?

Case-Control Studies

Low birthweight and respiratory disease in adulthood: A population-based casecontrol

Bias. A systematic error (caused by the investigator or the subjects) that causes an incorrect (overor under-) estimate of an association.

Food Diversity in the First Year of Life and the Development of Allergic Disease in High-Risk Children. By Cheryl Hirst. Supervisor: Dr.

Lecture Slides. Elementary Statistics Eleventh Edition. by Mario F. Triola. and the Triola Statistics Series 1.1-1

TORCH: Salmeterol and Fluticasone Propionate and Survival in COPD

Section 6.1 Sampling. Population each element (or person) from the set of observations that can be made (entire group)

GATE: Graphic Appraisal Tool for Epidemiology picture, 2 formulas & 3 acronyms

Intermediate Methods in Epidemiology Exercise No. 4 - Passive smoking and atherosclerosis

DECLARATION OF CONFLICT OF INTEREST

Søren Friis. Center for Kræftforskning, Kræftens Bekæmpelse. Epidemiologi/Bias & confounding/phd/sep 2012/SF

Biases in clinical research. Seungho Ryu, MD, PhD Kanguk Samsung Hospital, Sungkyunkwan University

Lecture 2. Key Concepts in Clinical Research

Assessment Schedule 2015 Mathematics and Statistics (Statistics): Evaluate statistically based reports (91584)

Gathering. Useful Data. Chapter 3. Copyright 2004 Brooks/Cole, a division of Thomson Learning, Inc.

ARTICLE REVIEW Article Review on Prenatal Fluoride Exposure and Cognitive Outcomes in Children at 4 and 6 12 Years of Age in Mexico

Types of Biomedical Research

AP Statistics Exam Review: Strand 2: Sampling and Experimentation Date:

COPD and environmental risk factors other than smoking. 14. Summary

STUDY DESIGNS WHICH ONE IS BEST?

12/26/2013. Types of Biomedical Research. Clinical Research. 7Steps to do research INTRODUCTION & MEASUREMENT IN CLINICAL RESEARCH S T A T I S T I C

Lecture 4. Confounding

Confounding and Effect Modification

Since 1980, obesity has more than doubled worldwide, and in 2008 over 1.5 billion adults aged 20 years were overweight.

Evidence-Based Medicine Journal Club. A Primer in Statistics, Study Design, and Epidemiology. August, 2013

Søren Friis. Center for Kræftforskning, Kræftens Bekæmpelse. Epidemiology/Bias and confounding/phd/sep 2012/SF

Pre-Calculus Multiple Choice Questions - Chapter S4

Rapid appraisal of the literature: Identifying study biases

Observational study II

Chapter 3. Producing Data

HARM. Definition modified from the IHI definition of Harm by the QUEST Harm Workgroup

You can t fix by analysis what you bungled by design. Fancy analysis can t fix a poorly designed study.

10/21/2014. Considerations in the Selection of Research Participants. Reasons to think about participant selection

Epidemiologic Methods and Counting Infections: The Basics of Surveillance

Transcription:

OUTLINE: Confounding- cont. Generalizability Reproducibility Effect modification Confounding Bias: Stratification Example 1: Association between place of residence & Chronic bronchitis Residence Chronic bronchitis Yes No Urban 194 2219 RR= 1.48 Rural 69 1208 Confounding Bias: Stratification Example 1: Stratified by Smoking: Smokers Non-smokers Bronchitis Bronchitis Residence Yes No Yes No Urban 167 1094 27 1125 Rural 53 417 16 791 RR=1.18 RR=1.17

Confounding Bias: Stratification Example 1: But 1.18 1.17 = 1.48 (crude) Smoking is a confounding variable! Confounding Bias: Stratification Example 2: Study the association between Alcohol consumption & Myocardial Infarction Alcohol MI Yes No Yes 71 52 OR=2.2 No 29 48 Confounding Bias: Stratification Example 2: Stratified by smoking: Smokers Non-smokers Alcohol MI No MI MI No MI Yes 8 16 63 36 No 22 44 7 4 OR=1 OR=1

Confounding Bias: Stratification Example 2: But 1 = 1 = 2.2 (crude) Smoking is a confounding variable! Confounding Bias: Stratification Example 3: Study the association between Treatment 1 & Treatment 2 with survival Alive Dead Total T1 40 60 100 T2 60 40 100 Total 100 100 200 RR (Crude) = 0.67 Confounding Bias: Stratification Example 3: Stratified by gender: Females Males Alive Dead Alive Dead T1 24 3 16 57 T2 58 30 2 10 RR=1.35 RR=1.32

Confounding Bias: Stratification Example 3: But 1.35 1.32 = 0.67 (crude) Gender is a confounding variable! Mantel-Haenzel: Calculating an adjusted RR i th stratum: Outcome No outcome Total E ai bi No E ci di Total Ni RR MH = ai(ci + di) /Ni ci(ai + bi) /Ni Mantel-Haenzel: Example Example: Females Males Success Failure Success Failure T1 24 3 16 57 T2 58 30 2 10 RR= 1.35 RR= 1.32 RR Adjusted = RR MH = 24(58+30)/115 + 16 (2+10)/85 58(24+3)/115 + 2 (16+57)/85 = 1.34

Concerns in Epidemiologic studies I. Internal validity: Was the study carefully designed and analyzed? II. III. External validity (Generalizability): Are the results applicable to the external population? Reproducibility of the results (precision): Are the results reproducible? External validity Also Known as generalizability. The extent to which the analytic inference derived from study sample is correct for the external population. External population Target Population Study Sample External validity External validity is dependent on internal validity. Maximize external validity by selecting study subject from a target population as similar as possible to the external population.

Concerns in Epidemiologic studies I. Internal validity: Was the study carefully designed and analyzed? II. III. External validity (Generalizability): Are the results applicable to the external population? Reproducibility of the results (precision): Are the results reproducible? Reproducibility Subjects in a study are always a sample of a population. Repetitive sampling results in a range of estimates for different samples: Sampling variation. The smaller the sample, the less reproducible will be the sample estimate. To decrease sampling error, increase sample size. Reproducibility Target Population Sample 2 Sample 1 Sample 3 Study sample

Effect Modification (Interaction) Example 1: Association between severe injury & death. Death No Death Severe injury 45 55 No Severe injury 6 94 RR = 0.45/0.06 = 7.5 Effect Modification (Interaction) Example 1: Stratified by age: <65 years > 65years Death No Death No Death Death Severe injury 15 35 30 20 No Severe injury 2 28 4 66 RR= 4.48 RR= 10.5 Effect Modification (Interaction) Example 1: 4.48 = 10.5 = 7.5 (crude) Age is an effect modifier!

Effect Modification (Interaction) When the exposure- outcome relationship is different for the different levels of a third variable, we have interaction (effect modification). The crude RR is hiding important effects. Solution: Report the RR separately for each category of the variable (for the different levels of the effect modifier). Effect Modification (Interaction) Example 2: 1.69 = 3 = 1 (crude) Gender is an effect modifier! Effect Modification (Interaction) A Confounder or Effect modifiers? Males RR=2 Hip Fracture Young Old Males Hip Fracture Males Hip Fracture RR= 0.7 RR= 3

Effect Modification (Interaction) A Confounder or Effect modifiers? Obesity RR=1.6 Breast cancer Pre-menopausal Post-menopausal Obesity Breast cancer Obesity Breast cancer RR= 1.1 RR= 2 Effect Modification (Interaction) A Confounder or Effect modifiers? Smoking RR=3 Low birth weight (LBW) Young Old Smoking LBW Smoking LBW RR= 2 RR= 4 Effect Modification (Interaction) Example 2: A cohort study was conducted to evaluate the association between air pollution and a specific lung disease. From the information below would you conclude that gender is a confounder or an effect modifier??? D No D Total E 200 1800 2000 No E 400 3600 4000 Total 600 5400 6000 3500 individuals were females 500 females were exposed 490 females had the disease 110 females had disease & were exposed 1410 males were exposed & did not have the disease 20 males were not exposed & had the disease

Biases: Review T (True) or F (False): To ensure that the study is internally valid we need to check for the 4 main biases: Selection bias, information bias, confounding and effect modification. Generalizability is the extent to which the inference derived from the study sample is correct for the target population. Biases: Review External validity depends on If you increase the sample size you increase Biases: Review T (True) or F (False): Blinding the interviewer minimizes observation bias. Recall bias is common in case control studies while loss to follow up is common in randomized clinical trials.

Biases: Review A cohort study is planned to investigate the association between maternal alcohol consumption during pregnancy and fetal alcohol Syndrome (a disease that is difficult to diagnose) in newborn children. What are the possible biases? Biases: Review In a closed cohort study, most of those who were exposed and developed the disease died before the end of the study. What will happen to the RR? What are the characteristics of a confounder? Biases: Review T (True) or F (False): The crude relative risk for smoking and myocardial infarction was 2.8. When stratified by gender, the relative risk for males was 5.6 and 1.5 for females. The study investigators concluded that results are biased by gender which is an effect modifier and hence they adjusted for gender in the analysis.

Effect Modification (Interaction) A Confounder or Effect modifiers? Having Pets RR=2.3 Asthma Children Adults Having Pets Asthma Having Pets Asthma RR= 2.1 RR= 2.4