Chapter 8- Cohort studies Flashcards
Cohort study
An observational study where the investigator selects a group of exposed individuals and a group of unexposed individuals, and follows both groups over time to compare the incidence of disease or rate of death from disease in both groups. These studies have to wait for the outcome to occur, which may require a long period of follow up
Positive association
If a positive association exists between the exposure and the disease, we would expect that the proportion of the exposed group where the disease develops would be greater than the proportion of the unexposed group where the disease develops. If we are identifying new (incident) cases of disease as they occur, we can determine whether a temporal relationship exists between the exposure and the disease- where exposure preceded the onset of disease
Limitations/considerations of cohort studies (4)
- How to select a study population?
- Types of cohort studies
- Selection biases
- Information biases
Selection of study populations
You can identify participants based on known exposure or can recruit a defined population without regard to exposure
Types of cohort studies (2)
- Prospective (longitudinal, concurrent)
- Retrospective (historical, nonconcurrent)
Prospective cohort studies
The investigator identifies the original population at the beginning of the study and follows the subjects concurrently through a period of time, until the disease either develops or does not develop. Exposure and unexposure are determined as they occur during the study. This type of study can take decades and can be expensive
Retrospective cohort studies
The study design does not differ from the prospective study. A retrospective study uses historical data in order to shorten the time period of the study. Exposure is ascertained from past records and the outcome is determined when the study begins
Framingham Heart Study
A well known cohort study in Massachusetts. Residents of Framingham could participate if they were 30-62 so they would be more likely to develop cardiovascular disease during the study period. The study population was examined every 2 years for incidence of new coronary events. The study was designed to test whether the incidence of CHD increases with age and whether it increases due to various exposures. This study was advantageous because it allowed investigators to study multiple exposures (hypertension, smoking, obesity, cholesterol levels) and complex interactions between the exposures.
Potential biases in cohort studies (2)
- Selection bias
- Information bias
Selection bias
Occurs when selection or participation is related to exposure AND outcome. Participants who refuse to join a study may skew the results if the participants who refuse to join are also more likely to have a specific exposure. Also, participants who are lost to follow up may impact the results.
Healthy worker effect
A type of selection bias comparing occupationally-exposed cohorts with unexposed persons from general population
Information bias
Occurs when measurement is study is related to exposure AND outcome. If the quality and extent of information obtained is different for exposed persons than for the unexposed persons, bias will be introduced. This is more likely in retrospective studies. If the person who decides whether the disease has developed in each subject knows whether the subject has been exposed and is also aware of the hypothesis being tested, it could bias their judgement of whether the illness has developed. This person should be masked. The epidemiologists analyzing the data may introduce bias if they have strong preconceptions
Benefits of cohort studies (2)
- Direct comparison of disease risk between exposed & unexposed (goal in analytic epi)
- More practical, when relationship between exposure & outcome is short
Cohort Studies
If exposure is associated with disease, we would expect to find
A greater incidence of disease in the exposed group than in the unexposed group
When is a cohort study indicated?
When good evidence suggests an association of a disease with a certain exposure
Case control studies based within a defined cohort
Combines case control and cohort studies- a population is identified and followed over time. When the population is identified, baseline data is collected. When a small proportion of the participants develop the disease over time, a case control study is then carried out using people who developed the disease as cases and using a sample of people who did not develop the disease as controls. There are 2 types of these studies- nested case control studies and case-cohort studies
2 types of Case-Control Studies within a Defined Cohort
- Nested case-control design
- Case-cohort design
Nested case-control studies
The controls are a sample of individuals who are at risk for the disease at the time each case of the disease develops. At the time each case develops, the same number of controls is selected. A control who is selected early in the study could develop the disease and later become a case in the study
Case-cohort studies
Cases develop at the same times that were seen in the nested case-control design, but the controls are randomly chosen from the defined cohort with which the study began. The subset of the full cohort is called the sub cohort.
Advantage of case-cohort studies
Because controls are not individually matched to each case, it’s possible to study different diseases (different sets of cases) in the same case-cohort study using the same cohort for controls
Benefits of Case-Control Studies within a Defined Cohort (4)
Combines advantages of cohort and case-control studies
1. Better establishes temporality
2. Limits information bias [differential recall of exposure]
3. Less expensive than cohort study
4. Greater comparability between cases & controls because BOTH are ascertained from same original cohort
