Cohort Studies Flashcards
what do cohort studies do?
cohort studies define and measure the risk factor (exposure) and then evaluate the outcome in a suitable demographic group over time
what can the outcome be in a cohort study? (2)
outcome can be the presence or absence of a disease associated with a certain risk factor
can you evaluate multiple risk factors with a single cohort study?
no; would then have to be separate studies for each risk factor
what can you do with a cohort study that you can’t do with a case control study?
can pick dose or population
what is the major difference between cohort and case control studies?
in cohort studies the population being studied is well-defined in demographics and numbers
what are the 2 types of cohort studies and what do they depend on?
can be prospective or restrospective, depending on the availability of data
what is an advantage of retrospective cohort studies?
save time if data is available
in what direction do you work in a case control study?
from outcome backwards to identify exposure
in what direction do you work in a cohort study?
from exposure forward to see the outcome
what is the ideal scenario for the control group in a cohort study regarding exposure to the risk factor? what complicates this? how do you determine this?
ideally the control group should have no exposure to risk factor; is complicated to determine is the group is only known through published data; conduct personal interviews if you can
what is the key point regarding cohort studies?
the presence or absence of the risk factor is determined BEFORE the outcome occurs
what is the basic idea of cohort studies?
see if those WITH the risk factor develop MORE disease than those WITHOUT the risk factor
compare the odds ratio to the relative risk in cohort studies
odds ratio is often higher than relative risk due to smaller populations and a larger margin of error
what are person years?
adjusts for the amount of time that any individual member of the cohort’s data is available
why are person years helpful?
some individuals are in the study longer than others, so it is better to evaluate the number of person years rather than the number of people in the study
when are person years very important and helpful?
with long term regional studies as people are more likely to drop out, move, etc.
what is reverse causality?
confounding by disease
what is the problem with reverse causality, or confounding by disease?
associations between risk factors and some disease outcome sare confounded by pre-existing comorbidities (tricky)
what should you ask to determine if selection bias is present in a cohort study?
are participants in exposed and unexposed groups similar in all ways except for the exposure?
what should you ask to determine if selection bias is present in a case control study?
are cases and controls similaar in all important aspects except for the disease in question?
what should you ask to determine if information bias is present in a cohort study?
is information about the outcome obtained in the same way for those exposed and those exposed?
what should you ask to determine if information bias is present in a case control study?
is information about exposure gathered in the same way for cases and controls?
what should you ask to determine if confounding is present?
could the results be accounted for by the presence of a factor associated with both the exposure and the outcome but not directly invovled in the causal pathway>
what are 3 features to look for in a cohort study?
- what steps were taken to minimize information bias?
- how complete was the followup of both groups?
- were potential confounding factors sought and controlled for in the analysis?
compare and contrast cohort versus case control study (5)
- cohorts tend to be long term and require significant preliminary planning
- cohorts are less useful in the stuudy of diseases of low prevalence or exposure
- cohort deals better with confounding and bias
- in cohorts, the strength of association is typically stronger for relative risk than odds ratios
- with case-control, you can extrapolate the results beyond the study group
why can you extrapolate case control studies beyond the study group?
- the demographic latitudes inherent in control group formation
- the additional veriation built into the confidence interval by including factorials for the control groups
what are clinical trials?
cohort studies in which the exposure to the effector is controlled
what are the 4 types of clinical trials?
- placebo-controlled
- single blind
- doublle blind
- triple blind
what are the 4 methods by which a diagnosis can be made?
- quantitative
- qualitative
- subjective
- objective
what do sensitivity and specificity measure
how good a test is at detecting binary features of interest (disease/no disease)
how do you calculate senstivity?
(those who test positive) divided by (those who are positive for the disease)
how do you calculate specificity?
(those who test negative) divided by (those who are negative for the disease)
how is the true disease status usually determined?
by some gold standard method
describe the relationship between sensitivity and specificity for a specific test
for a specific test, sensitivity increases as specificity decreases and vice versa
what is sensitivity?
the capacity of a test to correctly idenitfy TRUE POSITIVES
what happens to the number of false negatives as sensitivity increases
the greater the sensitivity, the smaller the number of unidentified false positives
what is specificity?
the capacity of a test to correctly exclude individuals who are free of the disease (FLASE NEGATIVES)
what happens to false positives as specificity increases?
the greater the specificity, the fewer false positives will be included
what is positive predictive value?
the probability of the disease being present among those with positive diagnostic results
what is negative predictive value?
the probability of the disease being absent among those whose test results were negative
what is Bayes Theorem?
used to compare one lab test to a reference lab with the gold standard test
what is misclassification?
the placement of subjects into inappropriate groups or classifications
when does random misclassification occur?
when the means by which we measure risk factors or disease are not perfect
when does non-random misclassification occur?
when exposure or disease status influences group assignment
what do misclassifications usually insert into data?
typically insert unwarranted bias, but can be corrected for if detected
