Choosing a reference group

1
Choosing a reference group
Doris Radun
15th EPIET Introductory Course, Lazareto,
Menorca 2009
2
Objectives of presentation

• Explore the meaning of representativeness
• Define source population
• Describe advantages and disadvantages of
  selecting different types of controls

3
Analytical epidemiology compare groups

• Questions to be answered by different methods
• Surveillance More cases than expected?
• Case-control study Exposure in cases vs.
  controls?
• Cohort study Incidence in exposed/unexposed?

Fundamental concept

4
Representativity and epidemiology
representativity regarding what?
in experiments? in surveys? in case-control
studies Cases? Controls?
5
Purpose of the comparison group (concept of
counterfactuality)
Cohort study The comparison group serves to
provide an estimate of the expected disease
incidence in the exposed group if the exposure
had been absent. Case control study The
comparison group serves to provide an estimate of
the exposure distribution in the source
population from which the cases
originate. Rothman KJ, 1986
6
Field epidemiology
A field epidemiologist often is involved after
the fact ? case control or retrospective cohort
study

• Constraints of time and resource
• Selection of reference group can be difficult
• Case-control studies prone to bias
• (why is that so?)

7
Outbreak Case Control Study

• May 2008
• 24 cases of Salmonella Typhimurium
• Cases age range 1656 years, 9 male, 6 female
• Cornwall (population 500,000)
• No recent travel abroad

8
Who are the right controls? Those ones?
9
Controls

• Controls
• should be representative
• of the population from which cases arose
• (i.e. the source population)

10
Control characteristics

• - Controls represent source population of
  cases -
• Shall have chance of exposure and illness
• Should have been identified as cases if they had
  disease under study
• Should have similar exclusion/restriction
  criteria as cases do

shall be susceptible
11
Source population
Exposed
Cases (n24)
Sample
Unexposed
Controls
- sample of the population - representative in
terms of exp.
Controls
12
Who is source population of cases?

• Start with your case definition!

13
Case definition

• Resident of Cornwall
• aged above 15 years
• with isolate of Salmonella Typhimurium in faecal
  sample
• during May 2008
• and
• who had not travelled abroad recently
• What is the source population?

14
Source population of cases was

• Residents of Cornwall aged above 15 years
• during May 2008
• who have not recently travelled abroad
• Controls should then be representative of this
  population

15
How to select controls?

• Aim for random sample of source population!
• not always feasible

16
Selecting controls (examples)

• Population based
• randomly from register/list/directory
• matched (age/sex/general practice)
• Friends
• Neighbourhood
• Hospital

17
a) Population controls

• Is there a list or register of source population?
  
• Such a list should
• be complete
• contain all cases
• be readily accessible
• identify specified characteristics e.g. age
• Take random sample!

18
Ad a) Population controls e.g., random
digit dialling

• using residential directories or mobile numbers
  (e.g. add 5 to cases number)
• quick and easy
• but may be bias in selection
• telephone ownership
• availability
• geographical area
• participation

- Not always applicable! (Germany) -
19
Ad a) Population controls or matching

• Matching
• e.g. same age, same sex, same doctor
• Matching useful if
• no full list of source population
• only specific group affected
• effect of matching variables is of no interest

20
b) Friend controls

• Advantages
• good matching for social factors
• can be quick and efficient
• Disadvantages
• Co-operation may be limited
• (concern about giving out names)
• If exposures same as in cases, you may not detect
  
• causal association overmatching

Hoopoe (Upups epops)
21
c) Neighbourhood controls

• Advantages
• no need for population register
• similar socio-economic status
• Disadvantages
• low co-operation
• may be time-consuming and expensive
• might be too similar to cases

22
d) Hospital controls

• Advantages
• useful if all cases identified from hospital
  register
• easily identified
• cost and time efficient
• Disadvantages
• different catchments for different diseases
• overmatching on exposures for other diseases

23
Sample size

• Often limited by number of cases available
• Usually inefficient to select more than 2-3
  controls per case
• (little extra statistical power beyond this
  number)

24
Controls may not be easy to find
25
Reminder Source population of cases was

• Residents of Cornwall aged above 15 years
• during May 2008
• who have not recently travelled abroad

26
Which reference group ?

• You are in charge of the case control study!
• What is your control definition?
• How would you select them?
• No population register or list is available
• ? Please discuss with a person next to you (2
  minutes!)

27
(No Transcript)
28
Some common questions regarding

• Non-cases as controls
• Asymptomatic cases
• Immune populations
• 100 exposure

?
29
What if Only non-cases as controls?
? You would artificially overestimate the Odds
Ratio!

• If attack rate is high,
• non-cases unlikely to represent exposure in
  source population
• If attack rate is low
• non-cases likely to represent exposures in source
  population
• can use them as controls
• More accurate not to exclude cases from control
  group
• Concept of case-cohort studies!

PS Compare with Tiramisu case study!
30
Non-cases as controls, by attack rate
end
start
High attack rate
Cases
Sourcepop.
If attack rate high, non-cases unlikely to
represent exposure in source population
Non- cases

• Low attack rate

If attack rate low, non-cases likely to represent
exposures in source population can use as
controls
31
What if Asymptomatic cases?

• Does it matter if we fail to identify mild cases?
  
• Analogous to non-response
• Example 40 cases, 40 controls

OR 20 x 30 / 20 x 10 3.0
32
Ad Asymptomatic cases

• If we only identify half the cases,
• and exposure stays the same

OR 10 x 30 / 10 x 10 3.0 (no bias)
33
Ad Asymptomatic cases

• If selection affects either case status
• as well as exposure status,

OR 10 x 30 / 20 x 10 1.5 the result will be
distorted
34
What if immune subjects?

• Not eligible as cases, so not in source
  population
• Difficult to identify immune status
• May have been cases in the past
• May have similar level of exposure to risk factor
  as current cases in study
• Bias in OR towards 1 (null value)

35
What if 100 exposure?

• What if close to 100 of population exposed?
• (e.g. food borne disease outbreaks where little
  choice in menu)
• ? Try to measure dose response
• (reference group lowest level of exposure)

36
What if 99 cases (attack rate very high)

• What if close to 100 of population are cases?
• ? Try to perform a cohort study!

37
Key points in choosing controls

• Define source population of cases
• Aim for sample representative for source
  population where cases arose from
• Review pros and cons of available options
• Plan to minimise bias, taking into account
  resources and urgency

38
Be prepared to defend your choice
and do the study!