Leicester Warwick Medical School Health and Disease in Populations Case-Control Studies Paul Burton.

Slides:



Advertisements
Similar presentations
Bias Lecture notes Sam Bracebridge.
Advertisements

Observational Studies and RCT Libby Brewin. What are the 3 types of observational studies? Cross-sectional studies Case-control Cohort.
KRUSKAL-WALIS ANOVA BY RANK (Nonparametric test)
Study Designs in Epidemiologic
Cohort Studies.
CONCEPTS UNDERLYING STUDY DESIGN
Observational Designs Oncology Journal Club April 26, 2002.
Case-Control Studies (Retrospective Studies). What is a cohort?
1 Case-Control Study Design Two groups are selected, one of people with the disease (cases), and the other of people with the same general characteristics.
Categorical Data. To identify any association between two categorical data. Example: 1,073 subjects of both genders were recruited for a study where the.
Extension Article by Dr Tim Kenny
Biostatistics ~ Types of Studies. Research classifications Observational vs. Experimental Observational – researcher collects info on attributes or measurements.
We’re ready to TEST our Research Questions! In science, how do we usually test a hypothesis?
Bios 101 Lecture 2 September 27, Hierarchy of Designs Expert opinion, usual practice Case series and case reports Ecological studies/Correlational.
Measures of Disease Association Measuring occurrence of new outcome events can be an aim by itself, but usually we want to look at the relationship between.
Measures of association
Comunicación y Gerencia 1Case control studies15/12/2010.
Cohort Studies Hanna E. Bloomfield, MD, MPH Professor of Medicine Associate Chief of Staff, Research Minneapolis VA Medical Center.
COHORT AND CASE-CONTROL DESIGNS Dr. N. Birkett, Department of Epidemiology & Community Medicine, University of Ottawa SUMMER COURSE: INTRODUCTION TO EPIDEMIOLOGY.
COHORT STUDY DR. A.A.TRIVEDI (M.D., D.I.H.) ASSISTANT PROFESSOR
Are exposures associated with disease?
BC Jung A Brief Introduction to Epidemiology - X (Epidemiologic Research Designs: Cohort Studies) Betty C. Jung, RN, MPH, CHES.
EPIDEMIOLOGY Why is it so damn confusing?. Disease or Outcome Exposure ab cd n.
Case Control Study Manish Chaudhary BPH, MPH
Analytic Epidemiology
Multiple Choice Questions for discussion
Dr. Abdulaziz BinSaeed & Dr. Hayfaa A. Wahabi Department of Family & Community medicine  Case-Control Studies.
Copyright © 2012 Wolters Kluwer Health | Lippincott Williams & Wilkins Chapter 7: Gathering Evidence for Practice.
Epidemiologic Study Designs Nancy D. Barker, MS. Epidemiologic Study Design The plan of an empirical investigation to assess an E – D relationship. Exposure.
CHP400: Community Health Program- lI Research Methodology STUDY DESIGNS Observational / Analytical Studies Case Control Studies Present: Disease Past:
Health and Disease in Populations 2001 Sources of variation (2) Jane Hutton (Paul Burton)
Retrospective Cohort Study. Review- Retrospective Cohort Study Retrospective cohort study: Investigator has access to exposure data on a group of people.
Types of study designs Arash Najimi
Lecture 6 Objective 16. Describe the elements of design of observational studies: (current) cohort studies (longitudinal studies). Discuss the advantages.
Mother and Child Health: Research Methods G.J.Ebrahim Editor Journal of Tropical Pediatrics, Oxford University Press.
A short introduction to epidemiology Chapter 2b: Conducting a case- control study Neil Pearce Centre for Public Health Research Massey University Wellington,
Epidemiologic design from a sampling perspective Epidemiology II Lecture April 14, 2005 David Jacobs.
S. Mazloomzadeh MD, PhD COHORT STUDIES Learning Objectives To develop an understanding of: - What is a cohort study? - What types of cohort studies are.
MBP1010 – Lecture 8: March 1, Odds Ratio/Relative Risk Logistic Regression Survival Analysis Reading: papers on OR and survival analysis (Resources)
A short introduction to epidemiology Chapter 10: Interpretation Neil Pearce Centre for Public Health Research Massey University, Wellington, New Zealand.
Types of study designs.
Basic concept of clinical study
Overview of Study Designs. Study Designs Experimental Randomized Controlled Trial Group Randomized Trial Observational Descriptive Analytical Cross-sectional.
Study designs. Kate O’Donnell General Practice & Primary Care.
Organization of statistical research. The role of Biostatisticians Biostatisticians play essential roles in designing studies, analyzing data and.
Case-Control Studies Abdualziz BinSaeed. Case-Control Studies Type of analytic study Unit of observation and analysis: Individual (not group)
Leicester Warwick Medical School Health and Disease in Populations Cohort Studies Paul Burton.
COHORT STUDY COHORT A group of people who share a common characteristic or experience within a defined period of time. e.g. age, occupation, exposure.
CHP400: Community Health Program - lI Research Methodology STUDY DESIGNS Observational / Analytical Studies Cohort Study Present: Disease Past: Exposure.
Case Control Studies Dr Amna Rehana Siddiqui Department of Family and Community Medicine October 17, 2010.
BIOSTATISTICS Lecture 2. The role of Biostatisticians Biostatisticians play essential roles in designing studies, analyzing data and creating methods.
A short introduction to epidemiology Chapter 2: Incidence studies Neil Pearce Centre for Public Health Research Massey University Wellington, New Zealand.
Analytical Studies Case – Control Studies By Dr. Sameh Zaytoun (MBBch, DPH, DM, FRCP(Manch), DTM&H(UK),Dr.PH) University of Alexandria - Egypt Consultant.
Types of Studies. Aim of epidemiological studies To determine distribution of disease To examine determinants of a disease To judge whether a given exposure.
Headlines Introduction General concepts
Odds Ratio& Bias in case-control studies
Leicester Warwick Medical School Health and Disease in Populations Revision Paul Burton.
1 Study Design Imre Janszky Faculty of Medicine, ISM NTNU.
Case control & cohort studies
Case Control study. An investigation that compares a group of people with a disease to a group of people without the disease. Used to identify and assess.
Epidemiological Study Designs And Measures Of Risks (1)
Chapter 9: Case Control Studies Objectives: -List advantages and disadvantages of case-control studies -Identify how selection and information bias can.
Present: Disease Past: Exposure
Epidemiologic Measures of Association
CASE-CONTROL STUDIES Ass.Prof. Dr Faris Al-Lami MB,ChB MSc PhD FFPH
Study design IV: Cohort Studies
Mpundu MKC MSc Epidemiology and Biostatistics, BSc Nursing, RM, RN
Study design IV: Cohort Studies
Cohort Study.
Presentation transcript:

Leicester Warwick Medical School Health and Disease in Populations Case-Control Studies Paul Burton

Lecture Objectives You should be able to: 1.Describe the principles underlying case-control studies 2.Describe the differences and similarities between case-control studies and other epidemiological designs 3.Outline the factors which suggest that a case- control design might be suitable for a particular epidemiological question

Lecture Objectives 4.Describe the limitations and assumptions inherent to case-control designs 5.Estimate the strength of an association from the result of a simple case-control study, and calculate and interpret the error factor and 95% confidence interval for this estimate Recommended reading from Prescribed book: Farmer and Miller, Ch 6, pp56-67

A hierarchy of study designs

Cohort Studies Exposed Unexposed Time Count events and pyrs

Cohort and case-control studies Cohort study (bladder cancer and cigarette smoking) Exposed: 100 cases in 100,000 people followed over 10 years = 1,000,000 pyrs Unexposed: 10 cases in 200,000 people followed over 10 years = 2,000,000 pyrs IRR = (100/1,000,000)  (10/2,000,000) = 20

Cohort and case-control studies Now ask question the other way around What are the odds of having been a smoker if you are a case? 100:10 = 100/10 = 10 What are the odds of having been a smoker if you are not a case? 99,900:199,990 = 99,900/199,990 = What is the ratio of the odds (OR=odds ratio)? 10/ = NB (100/99,900)  (10/199,900) = (100/10)  (99,900/199,900) = Very similar result to cohort analysis

Cohort and case-control studies What if we only have a 10% sample of the non-cases? What are the odds of being a smoker if you are a case? 100:10 = 100/10 = 10 What are the odds of being a smoker if you are not a case? 9,990:19,999 = 9,990/19,999 = What is the ratio of the odds (OR=odds ratio)? OR=10/ = Exactly the same result!!

Cohort and case-control studies What if we only have a 50% sample of the cases and a 20% sample of the non-cases?: What are the odds of being a smoker if you are a case? 50:5 = 50/5 = 10 What are the odds of being a smoker if you are not a case? 19,980:39,998 = 19,980/39,998 = What is the ratio of the odds (OR=odds ratio)? OR=10/ = Exactly the same result again!!

The general case ( full population data)

Sampling fractions: in cases, 0.02 in non-cases

Sampling fractions Regardless what proportion of all possible cases are collected (the sampling fraction in cases) and what proportion of all possible non-cases (the sampling fraction in non-cases) the two sampling fractions always cancel in calculating the odds ratio (OR) If we now call the non-cases “controls” this is a “case-control study”

Case-control studies We compare the odds of having been exposed in cases with the odds of having been exposed in the controls This gives us an odds ratio (OR) which is unaffected by the potentially different sampling fractions in cases and controls

Case-control Studies Case Non-Case (Control) Exposed? Time

Conducting a case-control study Identify a group of cases Identify a suitable group of non-cases Ascertain exposure status of everyone Compare level of exposure in cases and controls

Why use a case-control approach? Quick Fundamentally retrospective: no need to wait for a follow-up period

Why use a case-control approach? Cheap With a rare disease, most people in a cohort study will not develop disease and so most of the follow-up will be of people who contribute little information By using a low sampling fraction in controls in a case-control study you avoid having to collect information on a large number of non-cases

Expected yield of cohort studies:

The OR and the IRR Original example: IRR=20, OR=20.02 “The rare disease assumption” The approximation gets better and better as a disease gets rarer and rarer in the general population There is a special form of case-control study based on what is called “incidence density sampling” for which the approximation is always perfect – you don’t need to know about this for the HaDPop course Even when the IRR and OR are different ( e.g. IRR=5.1, OR=6.3) both are still measures of some sort of ‘risk ratio’ and both are therefore useful. It is not that one is ‘right’ and one is ‘wrong’: they express the same thing in a slightly different way.

Benefits of case-control studies Good for rare outcomes Possible to look at a lot of different exposures in detail Often no practicable alternative

Limitations of case-control studies No estimate of population incidence, only of relative risk The differing sampling fractions always cancel out in calculating ad/bc, but not in trying to calculating e.g. c/d (the odds of someone in the general population being a case if they are unexposed). Unless you know the sampling fractions More prone to bias: Information bias Selection bias It can be impossible to determine whether the disease causes the exposure or vice versa Not suitable for rare exposures

Information bias Does cigarette smoking cause ischaemic heart disease? Cases: average 5 cigarettes/day Controls: average 5 cigarettes/day Looks as if the exposure is not associated with the disease. But: True exposure in cases: 10 / day True exposure in controls: 5 / day Here, cases tend to understate their intake In addition Random errors push OR towards 1.0 (shrinkage)

Selection bias Case-control study of lung cancer and smoking Get cases of lung cancer from the respiratory medicine wards. Get controls as a random sample of patients from the same wards who do not have lung cancer But, smoking causes lots of other respiratory diseases as well as lung cancer so the patients on the ward are not a representative sample of the general population. Will underestimate OR.

Analysis 95%CI: OR  e.f., OR  e.f.

How many controls? Unlike an IRR, the precision of an OR is affected by the number of healthy people (x and z): So, it is worth increasing the number of controls - up to a point (typically up to 4-6 times as many controls as there are cases)

Creutzfeld Jacob Disease (CJD) and occupation Odds ratio = (9×104)/(3×13) = 24 95% CI: 24÷4.29, 24×4.29 = (5.59, 103.0)

Multiple levels of exposure

Retrospective v prospective? Confusing terminology: two different issues (1) Does the analysis look forwards or backwards? (2) Are the data collected as and when they occur ( i.e. prospectively) or from historical review - questionnaire, case-notes or other health records – ( i.e. retrospectively). Cohort analysis always looks forwards in time: Given exposure status at baseline, how many events occurred over time in how many person years and what is the incidence rate ratio? Simple case-control analysis is usually expressed as being backwards in time: Given case-control status now, what is the ratio of the odds of exposure at baseline?

Retrospective v prospective? Confusing terminology: two different issues (1) Does the analysis look forwards or backwards? (2) Are the data collected as and when they occur ( i.e. prospectively) or from historical review - questionnaire, case-notes or other health records – ( i.e. retrospectively). Conventional cohort study: prospective Historical cohort study: retrospective Conventional case-control study: retrospective

Comparison of cohort and case-control studies

Rooms for mid-module assessment th March 2002 Warwick Students Use their normal small group session rooms Leicester Students MSB room LT1 (candidate numbers 1-63) MSB room 206 (candidate numbers ) MSB room 320 (candidate numbers 115 onwards)