![]() |
Epidemiology & Biostatistics Case-control Studies
Author:
Dr.Janet ForresterProfessor
Tufts University School of Medicine
USA
Access: |
1.1 Definitions of statistics, probability, and key terms Read Online
1.2 Data, sampling, and variation in data and sampling Read Online
1.3 Frequency, frequency tables, and levels of measurement Read Online
1.4 Experimental design and ethics Read Online
By the end of this chapter, the student should be able to:
You are probably asking yourself the question, "When and where will I use statistics?" If you read any newspaper, watch television, or use the Internet, you will see statistical information. There are statistics about crime, sports, education, politics, and real estate. Typically, when you read a newspaper article or watch a television news program, you are given sample information. With this information, you may make a decision about the correctness of a statement, claim, or "fact." Statistical methods can help you make the "best educated guess."
Since you will undoubtedly be given statistical information at some point in your life, you need to know some techniques for analyzing the information thoughtfully. Think about buying a house or managing a budget. Think about your chosen profession. The fields of economics, business, psychology, education, biology, law, computer science, police science, and early childhood development require at least one course in statistics.
Included in this chapter are the basic ideas and words of probability and statistics. You will soon understand that statistics and probability work together. You will also learn how data are gathered and what "good" data can be distinguished from "bad."
Lect 5: Epidemiology & Biostatistics Case-control Studies
We will teach you how to read and critique medical journal articles using examples from some of the most widely-read medical journals. To critique the medical literature you will need to understand the fundamentals of epidemiologic study design, the sources of bias, and the role of chance. Every discipline has its own jargon. we will cover the terminology used in clinical research, including the basic statistical jargon. The most important concepts are in the lectures and small groups provide you with an opportunity to apply what you have learned from the lecture material to actual medical journal articles.
As future physicians you have an obligation to remain current in your field of practice and to treat patients according to generally accepted standards of care.
Question: Consider the case-control study below to answer Questions 6 and 7 Maternal coffee and alcohol consumption during pregnancy, parental smoking and risk of childhood leukemia Florence Menegaux et al. Cancer Detection and Prevention 29 (2005) 487-493. Accessed from PubMed.gov We investigated the role of maternal alcohol and coffee drinking and parental smoking on the risk of childhood acute leukemia in a multicenter case-control study. Methods: The study included 280 incident cases and 288 hospitalized controls, frequency matched with the cases by age,
Choices:
The exposures of interest were maternal smoking and the consumption of coffee and alcohol during pregnancy. The outcome of interest was acute leukemia (ALL or ANLL) in the child This is an example of a case-control study that examined more than one exposure in relation to a single outcome.
Question: What challenges does measurement of the exposure in a case-control study present?
Choices:
Case-control studies, being retrospective require that the exposure is measured retrospectively. Sometimes it is difficult to get good quality data on the exposure.
Question: Consider the case-control study below to answer Questions 6 and 7 Maternal coffee and alcohol consumption during pregnancy, parental smoking and risk of childhood leukemia Florence Menegaux et al. Cancer Detection and Prevention 29 (2005) 487-493. Accessed from PubMed.gov We investigated the role of maternal alcohol and coffee drinking and parental smoking on the risk of childhood acute leukemia in a multicenter case-control study. Methods: The study included 280 incident cases and 288 hospitalized controls, frequency matched with the cases by age,
Choices:
Case-control studies are prone to recall bias. This is when a mother of a child with leukemia "recalls" coffee consumption during pregnancy differently from a mother whose child does not have leukemia. You would want to know that at the time of the face-to-face interviews that the parents responding were not aware of the study hypothesis because that could influence how mothers reported their coffee consumption during pregnancy. If they thought the investigators suspected that coffee consumption increase the risk of leukemia, the mothers whose children were diagnosed with leukemia might report less (or more) coffee consumption than actually occurred. This could be done by embedding the coffee questions in a general survey about eating and exercise habits during pregnancy.
Question: Both retrospective cohort and case-control study designs are classified as observational studies. What is the essential difference between the design of a retrospective cohort study and the design of a case-control study?
Choices:
In cohort studies, including retrospective cohort studies, the groups are assembled based of their exposure status. In a case-control study, the groups are assembled based on the disease status.
Question: You have learned that conducting case-control studies can be difficult as they are subject to a wide variety of biases. In what circumstances would an investigator choose to conduct a case-control study?
Choices:
If the outcome is a rare disease, such as lupus, a case-control study may be the only practical study design to identify risk factors for the disease.
Question: Why do we use the odds ratio as the measure of association in a case-control study?
Choices:
The investigator determines how many cases and controls will be in a study based on sample size requirements. Hence, we cannot use these data to estimate incidence as we can in cohort or RCT studies. Since we do not have incidence data, we cannot calculate a relative risk and are, therefore, confined to calculating an odds ratio. The odds ratio approximates the RR when the disease is rare. Since case-control studies are usually done when outcomes are rare, the OR is a good measure of association for this study design. (When the disease is not rare, the OR over-estimates the RR).
Question: If an investigator is conducting a hospital-based case-control study, what is the best source of controls for that study?
Choices:
Hospital based case-control studies should use hospital controls. Community controls may not represent the catchment population of the hospital (i.e. the population that uses that hospital).
Question: How would you interpret the odds ratio of 3.1 for the association of coffee consumption of > 8 cups a day with acute lymphocytic leukemia?
Choices:
The OR is interpreted like a relative risk. You would interpret this OR as meaning that a woman who consumes more than 8 cups of coffee per day during pregnancy is 3.1 times as likely to have a child who will be diagnosed with acute lymphocytic leukemia (ALL).