• study where one or more groups of people that are free of disease and that differe according to extent of exposure (ex: exposed and unexposed) are compared with respect to disease incidence

Unlike case control, we start with exposure (case control starts with disease)

• we ID exposed and nonexposed comparable groups and follow forward in time for the development of disease (NOT randomly assigned to exposed and unexposed group- it happens naturally)
• may study morbidity and mortality
• design can be retrospective, prospective, or ambidirectional

• Advantages
  • Usually cheaper and faster to complete
  • Efficient with diseases with long latent period
• disadvantage- Exposure data may be inadequate

• Disadvantages
  • More expensive and time consuming
  • Not efficient for diseases with long latent periods
• Advantages
  • Better exposure and confounder data (easier to classify exposure)
  • Less vulnerable to  selection bias

1.   Identify cohort based on exposure or a common experience

2.   Follow entire population forward in time

3.   Choose unexposed population for comparison or members of cohort with differing degrees of exposure

4.   Compare rates of disease or mortality in exposed group to rates in unexposed/less exposed group

• Must enumerate (identify) entire population
• Methods to identify population include:
  • birth records
  • union records
  • Occupational or work histories
  • personnel records
  • census or other residence records
  • household survey (complete)
  • other records, e.g. school, military
• Characterize exposure variation within cohort  (may be homogeneous or heterogeneous)

• determine vital status for each cohort memeber
• determine cause of death for deceased indivuals

• pre-existing records- inexpensive data recorded before disease occurence
  • disadvantages- detal can be inadequate, records can be missing, doesn't have confounders
• questionarree interviews- good for info, but not routinely recorded but could have recall and interview bias
• direct physical exams, tests, environmental monitoring may be needed to ascertain certain exposures

• in any cohort study, ascertainment of outcome data involves tracing or following all subjects from exposure into future
• many resources used
• time consuming, but high losses to follow up raise doubt about validity of study

• general population rates aka external comparison group
• use comparable group from outside you cohort (type of external comparison group, and is sometimes called a double cohort study) (It doubles size and cost)
• divide cohort into different exposure groups and compare exposed to unexposed individuals aka internal comparison group

relative risk = incidence rate among exposed/    incidence rate among unexposed

RR = ad/bc

• advantages
  • gives incidence rates
  • selection bias usually not a problem
  • can study rare exposures
  • can look at multiple outcomes
• limitations
  • expensive, time consuming, long follow up, cant study disease with long latent period (PROSPECTIVE)
  • require large study population
  • if record based, there are often limited exposure data and usually no data on smoking, other confounders
  • not effective for studying rare diseases

• healty worker effect
  • you expect to see lower SMR because in order to work, you need to be relatively healthy
• exposure misclassification
• outcome misclassification
• ascertainment bias

• standardized mortality rates (SMR) used with general population comparison groups
• standardized incidence ratio (SIR) and standardized rate ratio (SRR) generally used with other external comparison groups