Case-control study |
Less expensive, less time-consuming |
Vulnerable to bias (recall bias, selection bias, confounding bias) |
Good for the study of rare disease |
Can assess multiple risk factors at once |
Cohort study |
Effective to establish cause and effect |
Possibility of selection bias, information bias |
Useful to identify the timelines over which certain exposures can contribution to outcome |
More expensive, more time-consuming (prospective cohort study) |
Can collect a wide variety of data |
Risk bias in sampling the cohort (retrospective cohort study) |
Nested case-control study |
Can reduce the cost to perform the study |
Require the selection of a new set of controls for each distinct disease |
Confounders can be matched in matching process |
Case-cohort study |
The ability to study several diseases using the same subcohort |
Require a more complicated statistical analysis |
Cross-sectional study |
Useful to assess the prevalence of disease |
Cannot infer causality |
Can suggest a natural progression with less cost |
Cannot estimate incidence rate |
Not good for studying rare disease |
Susceptible to nonresponse bias and recall bias |