In the previous article, I defined cohort studies and explained the difference between prospective and retrospective cohort designs. In this article, I will discuss bias and the pros and cons of cohort studies. The main advantages and disadvantages of cohort studies are highlighted below.
Advantages, disadvantages, and bias in cohort studies
Cohort studies are unbiased in terms of disease development (periodontitis), since exposure (orthodontic therapy) is measured before the onset of the disease. Cohort studies can be used to explore an association between an outcome and rare exposures by appropriate selection of the exposure groups. For example, we could evaluate dental fluorosis by selecting participants residing in areas with high versus low concentrations of fluoride in the drinking water.
Multiple outcomes for a single exposure can be examined. We may examine periodontitis, temporomandibular joint pain, and root resorption over several years in orthodontically treated and untreated participants.
Cohort studies measure the incidence (new cases in a time period) of a disease in the exposed and unexposed groups and allow determining whether there is a temporal relationship between exposure and outcome.
Since exposure is confirmed before the outcome has occurred, the chances of selection bias based on the outcome are nonexistent compared with other designs. If we know that certain participant characteristics might be associated with the outcome, our patient selection could be biased. For example, if age is associated with the outcome, selecting younger patients in the orthodontic group compared with the control group could lead to selection bias.
Losses to follow-up (information bias) can be a problem in cohort studies because they require a long follow-up period; therefore, participants may be lost to follow-up (high attrition rates). Missing information could be a problem as in clinical trials, especially if the losses are large or unequal and are somehow associated with the exposure or outcome of interest. For example, the willingness of a subject to return for the follow-up visit might depend on whether he or she has experienced signs of the outcome of interest. If patients not experiencing the outcome are more likely to be lost to follow-up than patients experiencing the outcome and this is related to the exposure, then the results of the study may be biased. In other words, if the loss to follow-up differs between exposure groups, this can have an impact on the results.
Table I shows the types of biases associated with cohort studies.
|Orthodontic group||No treatment group|
|Selection bias||Select younger patients||Select older patients|
|Information bias||Round up outcome measurement if exposure status is known||Round down outcome measurement if exposure status is known|