Understanding real research 2. Case control studies.
What can studies do? Describe the situation: Descriptive. Explain the situation: Analytical. Compare approaches: Experimental.
Study designs. Descriptive Cross-sectional, longitudinal. Analytic Case-control studies. Cohort studies. Quasi-experimental Natural experiments, policy interventions. Experimental Randomised controlled trial.
Study designs
Prevalence Cross-sectional Cause/ Aetiology Cross-sectional; Case-control; Cohort. Prognosis Harm Effectiveness Randomised controlled trial.
Past Present Future Cross-sectional Case-control Cohort From Altman. Practical Statistics for Medical Research, 1991.
Case control studies: Key features. Retrospective, i.e. backward looking. Relate an effect or outcome to a probable cause. Individuals with disease (cases) compared to individuals without disease (controls).
Case control study. From: http://library.downstate.edu/EBM2/2500.htm
Advantages & disadvantages Quicker than cohort studies No measure of incidence Relatively small study population required Can only investigate one disease at a time & no temporal association Good for rare diseases – can select all known cases Retrospective, so data may be biased, e.g. poor records; memory recall Case selection bias – must define cases very explicitly Control selection bias Matching controls to cases can be difficult Can only determine the odds ratio, not relative risk
Appraising case control studies Methodological approach: Are cases and controls similar, except for exposure to the putative cause? Is collection of retrospective data objective? Is there evidence of “causation”? Statistical reporting: Type of data – influences statistical analysis. Reporting of risk.