Lecture 6: Epidemiologic Studies - Case Control PDF

Types of Epidemiologic Studies (Typology2) OBSERVATIONAL –ANALYTICAL STUDIES Community Medicine & Epidemiology Dr. Sanabel Barakat Year 3 – semester 1 BDS., MSc., PhD., JDC. Week 5  IL...

Types of Epidemiologic Studies (Typology2) OBSERVATIONAL –ANALYTICAL STUDIES Community Medicine & Epidemiology Dr. Sanabel Barakat Year 3 – semester 1 BDS., MSc., PhD., JDC. Week 5  ILOs 1. Describe epidemiological methods 2. Define observational studies 3. Identify different types of observational analytical studies 4. Calculate measures of risk or association 2  Outline Case Control Study Cohort Study Prospective cohort Retrospective Cohort 3  Classification of Epidemiological studies Epidemiological studies classification According to objective : Descriptive (case report/series, cross-sectional (survey), ecological, time series studies) Analytical (case-control & cohort studies,) According to methodology: Observational Descriptive (case report/series, cross-sectional (survey), ecological, time series studies) Analytical (case-control & cohort studies) Experimental (RCTs, Field trial, community trial) 4  Case-Control Study Cases are people with disease, treatment outcomes or other criteria can be used to define a “case.” Controls are people without the disease or condition OR calculated 5 This Photo by Unknown Author is licensed under CC BY-NC  Observational, Analytical Case-Control Study Are common study designs used in oral health research. Case control studies ask the question: What are the determinants of this disease? Definition: Case-control studies compare cases and disease-free controls for their exposure status and compare the risk of exposure in cases and controls. These studies may be incorporated inside a cohort and are called NESTED CASE-CONTROL STUDIES. 6  Observational, Analytical Case-Control Study The investigator: selects a group of people with a disease/condition of interest (cases) selects a group of people who do not have the disease, (controls), compares the exposures of interest to find out which exposures are associated with the cases more than the controls. 7 8  Case-Control Study Selection of Cases: Clear basis of case definition needed: sign/symptoms; clinical examination; diagnostic tests; confirmatory tests. It is better to err on the side of restriction rather than inclusion in doubtful cases. Cases should have the disease. Use incident cases rather than prevalent cases Cases may be identified from clinic rosters, death certificates, disease/out come registries, surveys, administrative databases in some situations, adverse drug reactions (ADR) databases. 9  Case-Control Study Selection of controls: Defining the population and selection of appropriate controls are the two most important factors that determine the success of case- control studies in terms of arriving at correct inferences about strength, direction, and the importance of association of study factors and outcome 10  Case-Control Studies Selection of controls: Controls should represent the same population from which cases arise; that is, if the control group members previously had the disease, they would have become cases. Controls should have the same probability of getting disease as the cases. The controls should be similar to the cases in most of their characteristics so that the comparison of the differential effect of the exposure factors can be referred to their fundamental difference, which is being or being not affected by the disease or health condition investigated. 12  Case-Control Studies Selection of controls: Controls should provide exposure information about the population. Controls should be selected independent of their exposure status; that is, selection probability and sampling fraction of exposed and unexposed controls should be the same. The sampling of controls can be done randomly, 13  It is important to emphasize that the difference between the case and control groups lies in the outcome, not in the exposure measure because this terminology can cause some confusion in people unfamiliar with epidemiology. Participant will be asked about exposure history (risk factors) 14  Case-Control Studies Selection of controls: There are different strategies for the selection of controls, each with advantages and disadvantages, from the operational point of view and consequences for the inference of results. Sources of controls are selected from the general population (population-based studies), from neighbours, relatives, or companions to the health service in which the cases were attended and hospital controls recruited in the same health unit. 15  Case-Control Studies Types of Controls: Neighborhood Controls: Controls are selected from neighborhood residencies of the cases in a systematic way with or without matching. Random-Digit Dialing: Controls are selected randomly by calling numbers from a telephone book. 16  Case-Control Studies Types of Controls: Hospital/Clinic-Based Controls: Often used in hospital/clinic-based studies. - The source population may be people treated in the hospital, and controls may be selected form the same source population. - The catchment area of the hospital/clinic may be ill-defined, thus compromising defining the source population properly. - It may be possible to match cases and controls on disease criteria, but healthy controls may be difficult to obtain. - Controls cannot be selected randomly, and exposure–disease associations may bias the studies, e.g., Berkson’s bias. 17  Case-Control Studies Types of Controls: Dead Controls: Sometimes it may be possible to use dead people as controls if their exposure history prior to death can be ascertained. - Such controls might be a useful strategy if the cases are already dead. However, if the cases are living, then the dead controls do not exist in the source population. - Information for dead controls may be elicited from their medical records, vital registries, and/or proxy persons who knew the dead person well. 18  Case-Control Studies Types of Controls: Sibling Controls: In some studies, siblings of cases may be used as controls as they share similar family characteristics, neighborhood, socio-economic characteristics, etc. - Overmatching may be a problem. - All cases may not have siblings. 19  Case-Control Studies Types of Controls: Friend Controls: - Friends of cases may be used as controls. - Cases are asked to provide names of friends of same gender and age group for this purpose. - Overmatching may be a problem. - Having friends is a function of sociability which may vary between cases. - Cases may refer only certain friends based on some criteria that may introduce bias. 20  Case-Control Studies Matching Matching is the process of equating the groups being compared (e.g., cases and controls) on one or more factors so that whatever differences are noticed between the two groups would not be attributable to the factors on which they were matched. For example, if cases and controls were matched by age group and gender, then the differences between them would not be due to gender and age. Matching increases study efficiency in case-control studies by using fewer numbers of factors and variables, and may improve validity of cohort and experimental studies. 21  Case-Control Studies Matching May be done as individual matching, category matching, caliber matching, and frequency matching. Effects of factors used in matching cannot be examined in the matched study. May introduce selection bias in the study. Matching results in paired data that must be handled especially for analysis Over/under-matching can create various problems in the study 22  Odds Ratio In a case-control study, the risk of disease cannot be directly calculated because the population at risk is not known. Odds ratio is the measure used with case-control studies. The major method for analyzing results in case-control studies is the odds ratio (OR). The odds ratio is the odds of having a disease (or outcome) with the exposure versus the odds of having the disease without the exposure. The odds are the ratio of the probability that an outcome occurs to the probability that the outcome does not occur 24  OR= ad/bc 25  Example: occurrence of caries in school children, in relation to daily intake of candy bars Cases (have Control (caries Total caries) free) ODDs Ratio = a/b ÷c/d = (aXd) ÷(bXC) Exposure a) 65 b) 25 90 = (65X75) ÷ (25X35) (daily candy bar) = 5.57 No disease c) 35 d) 75 110 (no candy bar) Total 100 100 200 26  Case-Control Studies Advantages: Useful for studying uncommon diseases; less expensive (prospective studies short duration studies, logistically easy; yields a reasonable estimate of risk ratio (odds ratio)—if prevalence is low, then odds ratio approximates relative risk well. 29  Case-Control Studies Disadvantages: Temporal relationship between exposure and outcome can not be examined; subject to substantial selection, survivor, and recall bias; one outcome can be studied at a time; does not provide prevalence, incidence, and excess risk. 30  Case-Control Studies Assumptions: Disease prevalence is low. When not much information is available about the disease Cases and controls are representative of the population when population dynamics do not permit longer studies Relative risk cannot be directly calculated. When obtaining exposure data is difficult/expensive When disease has very long induction period/latency 31  Thank you 32