Wk 3 - Calibrating cases

Question 1

CASE DEFINITION

i) what is it?
ii) what must not it include?
iii) give three reasons why it is important
iv) give two possible case definitions for asthma

Answer 1

i) set of criteria for the determination of whether someone has a disease
ii) must not include exposures of interest
iii) important because is affects prevalence estimates, affects identifcation of causal factors and outcomes, affects service planning and comissioning, affects treatment
iv) for asthma - taking medication, diagnosed by GO of wheeze in last 12 months in absence of a cold

Question 2

CASE IDENTIFICATION

i) what is it
ii) what is sensitivity?
iii) what is specificity?

Answer 2

i) operationalisation of the case definition - how we identify who is and isnt a ‘case’

ii) sensitivity is the proportion of people who have a condition and test positive for it
- eg breast cancer screening - 0.9 sensitivity means 9/10 women who test positive have it and 1/10 tests neg but does have it

iii) specificity is the proportion of people without the condition that correctly test negative
- eg 9/10 women who dont have BC test negative but 1 person who does have it also tests neg

Question 3

PREDICTIVE VALUES

i) what is a positive predictive value?
ii) what is a negative predictive value?
iii) name three things that predictive values depend on
iv) how do PPV and NPV change as prevalence increases? why?

Answer 3

i) PPV is the likelihood a patient with a positive test actually has the disease
ii) NPV is the likelihood a patient with a negative test does not have the disease
iii) dep on sensitivity and specificity of the test and the prevalence of the disease

iv) as prevalence increases - PPV increases and NPV decreases (and vice versa)
- the underlying frequency of the disease has increased in the population

Question 4

CASE CONTROL STUDY

i) what is it?
ii) what are the two groups of participants selected for to begin with?
iii) are past or future exposures measured?
iv) is this a prospective or retrospective type of study?
v) give two strengths and two weaknesess

Answer 4

i) groups who differ at the outset on disease status and compare their outcomes in relation to risk factors present or absent

ii) start with two groups - one with conditions (case) and one group without (control)
- need to match other variables as closely as possible

iii) measure past exposures in both groups - historical records
iv) retrospective

v) strength - can offer some evidence for cause and effect, can identify multiple exposures, good when disease is rare
weakness - cant calc prev, incidence or RR, less suitable for rare exposire, hard to ensure exposure occ before onset, retrospective data can be poor quality

Question 5

CASE CONTROL VS COHORT STUDY

i) what is the difference?
ii) what does a cohort study select participants on?
iii) what does case control select participants on?
iv) which is better for studying rare diseases? which is better for rare exposures?
v) which is better for diseases with long latency?

Answer 5

i) cohort selects people on basis of exposure status and case control selects people on basis of outcome

ii) cohort - selects on basis of exposure
- start with people who dont have the disease and compare exposure vs non exposure - compare incidence in new onset cases

iii) case control selects on basis of outcome
- start with people who have the disease and compare whether they had an exposure or not (retrospective only)

iv) rare diseases - case control
rare exposure - cohort

v) case control better if long latency

Question 6

ODDS RATIO

i) what does it test?
ii) how can it be expressed?
iii) what does an OR of <1, 1, >1 mean?
iv) what does OR of 3 mean?

Answer 6

i) test association between exposure and outcome of interest
ii) odds in exposed/ odds in non exposed
iii) OR <1 - odds of exposure for case in less than dds of exposure for controls (exposure may be protective)

OR = 1 - odds of exposure in cases is equal to controls (exposure not assoc with disease)

OR > 1 - odds of exposure for cases is greater than for controls
(exposure may be a TF for disease)

iv) OR 3 = odds of exposure for cases is 3x higher than for controls

Question 7

SUMMARY

i) what is case definition?
ii) what is sensitivity and specificity?
iii) what is predictive value? what three things do they depend on
iv) what are participants in a case control study selected on? give an adv and disadv
v) what is an odds ratio?

Answer 7

i) criteria for the determination of whether someone has a disease

ii) sensitivity - proportion of people with cond who test pos
specificity - prop of people wihtout cond who test negatuve

iii) likelohood an individual patient has a disease based on test result
- dep on sens, spec and prevalence of disease

iv) case control selects based on outcome status and historical exposures are measured
- adv - good in rare disease
- disadv- not useful for rare exposure

v) OR compares the odds of exposure (disease present) for cases vs controls (disease not present)