Cohort studies

Question 1

person years

Answer 1

1 person is followed up for 10 years = 1x10 = 10 person-years
3 people are followed up for 2 years = 2x3 = 6 person-years
by adding all multiplications = number of person years

allows for a dynamic population to be studied

Question 2

describe the logical basis of cohort studies

Answer 2

observe a study population over time and compare the development of disease outcomes across categories of determinants.
record any non-identical features and adjust for them
may be a dynamic population
the outcomes are measured without relationship to the determinant - otherwised = biased
allows absolute and relative measures of disease risk to be measured

Question 3

Give characterisitics of cohort

Answer 3

recruit people who are outcome free
over time follow up and observe how many develop disease
compare risk according to exposure/personal characteristics

comparative - compare exposed group with unexposed
calculate risk of each group and compare using relative risk (IRR or SMR)

Question 4

define cohort

Answer 4

a defined group of people with known exposure or personal characteristics

Question 5

give the different types of cohort study

Answer 5

prospective/ concurrent

retrospective

Question 6

describe how you would make internal comparison and its problems

Answer 6

collect data of outcomes and calculate the incidence rate of both exposed and unexposed groups - could be influenced by RANDOM variation so….

calculate error factor
calculate CI - (observed/e.f.) and (observed x e.f.)

sub-cohorts may be of different sizes
internal comparisons may need large studies
sub cohorts may not be comparable due to other factors - confounding

Question 7

describe how you would do an external comparison and its problems

Answer 7

use SMR
o/e x 100 - may be subject ot random variation

calculate error factor and the confidence intervals

for most serious diseases large cohorts need to be obsreved over a long period of time - during the time cohort membrers age and age - specific morbidity and mortality rates in ref population will change and cacluating the expected cases is more complicated

Question 8

give the relationship between sample size and error factor

Answer 8

the greater the number of events, the smaller the error factor

Question 9

give the difference between internal and external comparison e.f.

Answer 9

internal comparison e.f. are larger than external

as observed casesa are being subdivided whereas with external comparisons e.g. just uses observations

Question 10

Give the postives and negatives of external and internal cohort studies

Answer 10

external have less random variation than internal
external may have less information on the reference pop - confounders
there may be selection bias between study and referenc epopulations - healthy worker effect
may only be able to use binary comparisons with external comparison
but you have more control over the data with internal comparisons and more detailed information can be collected allowing for studies such as dose-response effects

Question 11

give general positives and negatives of cohort studies

Answer 11

cohort studies measure absolute risks and relative risks
good for internal or external comparisons

need large number of participants 
extended follow up periods 
expensive 
need many years to follow up 
not good for rare diseases 

and retrospective cohort studies may be quicker and easier but historical data may be incomplete or not available

suitable for rare exposures but not for rare disease

can look at many different harmful effects form a single exposure