Workshop 4Reginald Finger, MD, MPH Jiajoyce Conway, DNP, CRN.docx

Workshop 4
Reginald Finger, MD, MPH
Jiajoyce Conway, DNP, CRNP
Avoiding Epidemiologic Traps
1
Ecologic Fallacy
The ecologic fallacy, simply stated, is the error made when one
makes incorrect inferences about an individual or small group’s
probability of having a certain characteristic, based on the
probability of that characteristic in the population from which
that individual or small group comes.
Let’s look at an example:
Ecologic Fallacy
Let’s say that we know that in a University with lower and
upper campuses, that because upper campus houses graduate and
honors programs, that the average GPA of students on upper

campus is higher.
Does it follow that in a particular introductory psychology class
offered to students from across the campus, if the professor
tabulates the GPA of students enrolling in that class, that the
ones from upper campus will have a higher GPA than those
from lower campus?
3
Ecologic Fallacy
Not necessarily!
The mistake the psychology professor would make with such an
assumption would be to substitute risk at the population level
for risk at the individual level. The population-level difference
in GPA, in this case, is attributable to the influence of students
who probably are not going to take introductory psychology.
Confounding
Consider the following: 7 of 110 women, and 17 of 100 men,
are positive for a new viral antibody “N” on when screened with
a blood test. The odds ratio is 3.01, easily significant with a
chi-square test.
However … look what happens when the data are analyzed by

who has or has not ever lived outside the U.S.:
5
Confounding
Among those having lived outside the U.S., 15/50 men and 3/10
women (both 30%) have the N antibody. Among those never
having lived outside the U.S., 2/50 men and 4/100 women have
N antibody (both 4%).
The odds ratio in each group is 1.00 – there is actually no
relationship at all between gender and having the N antibody. It
only appeared so because of a statistical fluke.
Confounding
This is an example of what is known as confounding. The
relationship between gender and the N antibody is confounded
by whether the person has lived outside the U.S. Confounding
occurs when the relationship between “A” and “C” is distorted
(in either direction) by “B”, which is associated with “A”, and,
independent of its relationship with A, is associated with “C”.

7
Bias
Confounding is an illustration of what is known as “bias” – that
is, a situation where an unrelated factor obscures the true
outcome.
Bias can come from a number of sources: four of the most
common are recall bias, nonresponse bias, selection bias, and
publication bias.
8
Selection Bias
If, for instance, hospital workers are surveyed for their opinions
about the medical care system and their responses considered
representative of the population at large, the result will be
inaccurate as the result of selection bias. Hospital workers are
likely to have opinions about medical care for reasons that
would not be applicable to most people.
Recall Bias

When, for instance, persons with a specific illness are surveyed
about their prior exposures to possible risk factors, and their
responses compared to those who are not ill, investigators must
consider that the experience of facing illness may make a
person much more likely to have thought about and to recall
prior experiences. This is an example of recall bias. Recall
bias is a particular problem in case-control studies.
Nonresponse Bias
When investigators are trying to assess the opinions of a group
of people, and a significant fraction of the group chooses not to
respond (call back, fill out the survey, log in, etc.), it is quite
likely that those who do respond are doing so because they have
an interest in the topic which influences their opinions in a way
not typical of the others. This is an example of nonresponse
bias.
Publication Bias
Here is one that people sometimes do not think about: it is
simply that studies with positive or remarkable findings are
more likely to be published in journals, while studies that find
no association are less likely to see the light of day. Over time,
this can give the impression that certain associations are
stronger than they really are.
Be careful with this one – not every rejection of an article is
due to bias! Sometimes the editors discover another form of

bias.
Effect Modification
Consider the following: 100 of 200 (50%) men, and 140 of 200
(70%) women are found to have antibody “R” on blood testing.
Clearly, the relationship is significant.
However … look what happens when the data are analyzed by
whether the men or women take multivitamins regularly:
Effect Modification
In the group that does not take vitamins, 50/100 men and 50/100
women have the R antibody (both 50%). In the group that does
take vitamins, 50/100 men and 90/100 women have the
antibody! Clearly, the relationship between gender and the R
antibody is very different depending on whether the person
takes vitamins.
Effect Modification
This phenomenon is known as “effect modification”. We say
that vitamins modify the effect of gender on R antibody status.

Women (and not men) are susceptible to “R” in some way that
requires the presence of a vitamin to develop the antibody.
The relationship between gender and R is nonetheless real, but
it cannot be understood properly unless one considers the effect
of the vitamin.
15
MY FUTURE AS A 2
4.5 Career Paper Rubric – Highlighted section indicates points
earned
Criteria
5 points
4-3 points
2-0 points
1. Career and likely epidemiological responsibilities.
Provided clear and complete description of career role and
relevancy of epidemiology.
Described career role but with some ambiguities or
inconsistencies.
Career goal not present or had to be inferred from other
comments.
2. Outbreaks that could realistically happen.
Accurate and detailed description of a known type of outbreak
that could likely occur in the setting and uses statistics to
describe its history and current trend of incidence and
prevalence.

Outbreak described but with inconsistencies, or as unlikely to
occur in the setting. Uses statistics to describe history or trend
of incidence and/or prevalence.
Outbreak not described or description is entirely off base
factually. Stats not used to describe the outbreak.
3. Function in the event of an outbreak.
Describes workable role for this professional in this outbreak,
in significant detail, protection of self, staff, other patients, etc.
Lacks detail, inappropriate role, or mismatched to type of
outbreak. Protection described in a limited fashion.
Role not described or has to be inferred from other comments.
Protection not addressed.
4. Working with the media and public.
Describes how to educate the public, gain media cooperation,
and gives a full narrative about what the public should do to
protect self and others, and how to treat if affected, using
language that avoids panic.
Description of public or media response lacks detail, or is less
than workable. Narrative to public about protecting self and
others and steps to treat is sketchy.
Reader cannot tell what the student intends to do about public
or media response. No narrative regarding how to protect self
and others.
5. Epidemiologic traps.
Correctly identifies one of the epidemiologic traps presented in
the course, gives a correct example of it, and a workable
strategy to avoid.
Description of the epi trap has inconsistencies, or the plan to
avoid it has flaws.
Example of epi trap, if given, is not correct or unlikely to occur
in the setting.
10 points
8-9 points
7-5 points
4-0 points

6. Apply the 10 steps of outbreak investigation.
Strongly applies all of the 10 steps model, to this outbreak and
incorporates disease specific information into each step.
Applies most of the 10 steps model, to this outbreak and
incorporates disease specific information into each step.
Description of outbreak response model is inconsistent or
incomplete; information is lacking.
Reader cannot reasonably discern that an outbreak response
model is applied.
5
4
3-2
1-0
7. APA format: margins, font, etc.
Less than one correction per page.
One to two corrections per page, deduction depends on
seriousness of errors.
Three to five corrections per page, deduction depends on
seriousness of errors
Frequent corrections made; not adhering to graduate level
expectations.
8. Grammar, tense, punctuation, noun/verb agreement sentence
structure.
Less than one correction per page.
One to two corrections per page, depending on seriousness of
errors.
Three to five corrections per page, deduction depends on
seriousness of errors.
Frequent corrections made; not adhering to graduate level
expectations.
Total Points
/50

Outbreak Investigations:
The 10-Step Approach
Zack Moore, MD, MPH
Medical Epidemiologist
North Carolina Division of Public Health
Learning Objectives
1. List three reasons why outbreak
investigations are important to public health
2. Know the steps of an outbreak investigation
3. Give an example of a single overriding
communication objective (SOCO)
Reasons to Investigate an Outbreak
• Identify the source (and eliminate it)
• Develop strategies to prevent future
outbreaks
• Evaluate existing prevention strategies
• Describe new diseases and learn more
about known diseases
• Address public concern

• It’s your job!
When to Investigate
Consider the following factors:
• Severity of illness
• Transmissibility
• Unanswered questions
• Ongoing illness/exposure
• Public concern
Environmental Investigation
• Vital part of investigation
• Should be done with (not instead of)
epidemiologic investigation
Collecting and Testing
Environmental Samples
• Ideally, epidemiologic results guide sample
collection
– Often collected at the same time
• Can support epidemiologic findings
– Positive or negative results can be misleading

Principles of Outbreak Investigations
• Be systematic!
– Follow the same steps for every type of outbreak
– Write down case definitions
– Ask the same questions of everybody
• Stop often to re-assess what you know
– Line list and epi curve provide valuable
information; many investigations never go past
this point
• Coordinate with partners (e.g., environmental and
epidemiology)
10 Steps of an Outbreak
Investigation
1. Identify investigation team and resources
2. Establish existence of an outbreak
3. Verify the diagnosis
4. Construct case definition
5. Find cases systematically and develop line
listing
6. Perform descriptive epidemiology/develop
hypotheses
7. Evaluate hypotheses/perform additional
studies as necessary
8. Implement control measures

9. Communicate findings
10. Maintain surveillance
Investigation
Investigation Resources
• Local
– Epi teams
• State
– CD Branch epidemiologists / subject matter experts
– Nurse Consultants
– PHRST teams
– Disease Investigation Specialists (DIS)
• Other
– Team Epi-Aid (UNC)
– CDC
Investigation

What is an Outbreak?
Increase in cases above what is expected in
that population in that area
care center in January?
eating at Restaurant A?
Investigation
Verify the Diagnosis
• Obtain medical records and lab reports
– Contact Public Health Epidemiologist in
Hospital & Infection Preventionists
• Conduct clinical testing if needed
– Consult with CD Branch, State Lab

Investigation
Components of Case Definition
• Person...... Type of illness
(e.g., “a person with...”)
• Place......... Location of suspected
exposure
• Time.......... Based on incubation
(if known)
Sample Outbreak Case Definition
Hepatitis A outbreak:
• Person: An acute illness involving
jaundice or elevated liver function tests
• Place: Occurring after visiting or residing
on Property A
• Time: During May–August 2006

Investigation
5. Find cases systematically and develop line listing
What to Put on a Line List
1. Clinical information
• Symptoms (type, duration)
• Onset dates and/or times
2. Demographic information
3. Exposure information
Use line list to summarize information
Investigation

hypotheses
Descriptive Epidemiology
• Person, place and time
• Line lists and epi curves useful in
developing hypotheses
Can suggest type of exposure
– Point-source
– Person-to-Person
Epi Curves
Epi Curve A
0
20
40
60
80
100

Time
#
C
as
es
Point Source
0
20
40
60
80
Time
#
C
as
es
Propagated
(Person-to-Person)
Epi Curve B

• Suggest type of exposure
– point-source, person-to-person
• Suggest time of exposure
– if agent known
• Suggest possible agents
– if time of exposure known
Epi Curves
0
10
20
30
40
50
60
70
80
Time
#
C
as
es
Average incubation
Exposure

Known Time of
Exposure
0
10
20
30
40
50
60
70
80
Time
#
C
as
es
Maximum incubation
Minimum incubation
Est. exposure period
Known
Agent

Investigation
hypotheses
7. Evaluate hypotheses/perform additional studies as
necessary
Additional Studies
• Types
Cohort
Case-control
• Designed to assess exposures equally
among ill and non-ill
Cohort Studies
• Include EVERYONE who could have been
exposed
– Only use if a complete list is available
– Meeting attendees, students, LTCF residents, etc.
• Measure of association = Relative Risk

Relative Risk (RR)
• RR = 1.0
Risk same among exposed and unexposed
• RR > 1.0
Risk is HIGHER among exposed
• RR < 1.0
Risk is LOWER among exposed
Case-Control Studies
• Compare exposures among ill persons
(case-patients) and non-ill persons (controls)
• Used when a complete list is not available or
too large
– Restaurant outbreaks, national outbreaks, etc.
• Measure of association = Odds Ratio
Interpretation of Odds Ratio
• OR = 1.0
Same odds of exposure among ill and non-ill
• OR > 1.0
HIGHER odds of exposure among ill

• OR < 1.0
LOWER odds of exposure among ill
Investigation
hypotheses
necessary
Control Measures
• Can occur at any point during outbreak
• Isolation, cohorting, product recall
• Balance between preventing further
disease and protecting credibility and
reputation of institution
• Should be guided by epidemiologic results
in conjunction with environmental
investigation

Investigation
hypotheses
necessary
Inform Public and Media
• Public & press are not aware of most
outbreak investigations
• Media attention desirable if public action
needed
• Response to media attention important to
address public concerns about outbreak
– Single overriding communication objective
(SOCO)

• Results of investigations public information
Investigation
6. Perform descriptive epidemiology/develop hypotheses
7. Evaluate hypotheses/perform additional studies as necessary
10. Maintain surveillance
Maintain Surveillance
control measures
Conclusions
• Epidemiologic investigations are essential
to determine source of outbreaks
• Be systematic
• Follow the steps!

Outbreak Investigations: �The 10-Step ApproachLearning
ObjectivesReasons to Investigate an OutbreakWhen to
InvestigateEnvironmental InvestigationCollecting and Testing
�Environmental SamplesPrinciples of Outbreak
Investigations10 Steps of an Outbreak Investigation10 Steps of
an Outbreak InvestigationInvestigation Resources10 Steps of an
Outbreak InvestigationWhat is an Outbreak?10 Steps of an
Outbreak InvestigationVerify the Diagnosis10 Steps of an
Outbreak InvestigationComponents of Case DefinitionSample
Outbreak Case Definition10 Steps of an Outbreak
InvestigationWhat to Put on a Line List10 Steps of an Outbreak
InvestigationDescriptive EpidemiologyEpi CurvesEpi Curve
AEpi Curve BEpi CurvesSlide Number 26Slide Number 2710
Steps of an Outbreak InvestigationAdditional StudiesCohort
StudiesRelative Risk (RR)Case-Control StudiesInterpretation of
Odds Ratio10 Steps of an Outbreak InvestigationControl
Measures10 Steps of an Outbreak InvestigationInform Public
and Media10 Steps of an Outbreak InvestigationMaintain
SurveillanceConclusions
10/22/2018 Biases and Confounding | Health Knowledge
https://www.healthknowledge.org.uk/public-health-
textbook/research-methods/1a-epidemiology/biases 1/17
Sign in / Register
Search
1a - Epidemiology
PLEASE NOTE:

We are currently in the process of updating this chapter and we
appreciate your patience whilst this is being completed.
Bias in Epidemiological Studies
While the results of an epidemiological study may reflect the
true effect of an exposure(s) on the development of the outcome
under investigation, it
should always be considered that the findings may in fact be
due to an alternative explanation1.
Such alternative explanations may be due to the effects of
chance (random error), bias or confounding which may produce
spurious results, leading
us to conclude the existence of a valid statistical association
when one does not exist or alternatively the absence of an
association when one is truly
present1.
Biases and Confounding
HOME ABOUT PUBLIC HEALTH TEXTBOOK TEXT
COURSES VIDEO COURSES TRAINING
https://www.healthknowledge.org.uk/
textbook/research-methods/1a-epidemiology
https://www.healthknowledge.org.uk/
https://www.healthknowledge.org.uk/about-us
https://www.healthknowledge.org.uk/public-health-textbook
https://www.healthknowledge.org.uk/e-learning
https://www.healthknowledge.org.uk/interactive-learning
https://www.healthknowledge.org.uk/teaching

Observational studies are particularly susceptible to the effects
of chance, bias and confounding and these factors need to be
considered at both the
design and analysis stage of an epidemiological study so that
their effects can be minimised.
Bias
Bias may be defined as any systematic error in an
epidemiological study that results in an incorrect estimate of the
true effect of an exposure on the
outcome of interest.1
Bias results from systematic errors in the research methodology.
The effect of bias will be an estimate either above or below the
true value, depending on the direction of the systematic error.
The magnitude of bias is generally difficult to quantify, and
limited scope exists for the adjustment of most forms of bias at
the analysis stage. As a
result, careful consideration and control of the ways in which
bias may be introduced during the design and conduct of the
study is essential in
order to limit the effects on the validity of the study results.
Common types of bias in epidemiological studies
More than 50 types of bias have been identified in
epidemiological studies, but for simplicity they can be broadly
grouped into two categories:
information bias and selection bias.

1. Information bias
Information bias results from systematic differences in the way
data on exposure or outcome are obtained from the various
study groups.1 This may
mean that individuals are assigned to the wrong outcome
category, leading to an incorrect estimate of the association
between exposure and outcome.
Errors in measurement are also known as misclassifications, and
the magnitude of the effect of bias depends on the type of
misclassification that has
occurred. There are two types of misclassification – differential
and non-differential – and these are dealt with elsewhere (see
“Sources of variation,
its measurement and control”).
Observer bias may be a result of the investigator’s prior
knowledge of the hypothesis under investigation or knowledge
of an individual's exposure
or disease status. Such information may influence the way
information is collected, measured or interpretation by the
investigator for each of the
study groups.
For example, in a trial of a new medication to treat
hypertension, if the investigator is aware which treatment arm
participants were allocated to, this
may influence their reading of blood pressure measurements.
Observers may underestimate the blood pressure in those who
have been treated, and
textbook/research-methods/1a-epidemiology/sources-variation-
measurement-control

overestimate it in those in the control group.
Interviewer bias occurs where an interviewer asks leading
questions that may systematically influence the responses given
by interviewees.
Minimising observer / interviewer bias:
Where possible, observers should be blinded to the exposure
and disease status of the individual
Blind observers to the hypothesis under investigation.
In a randomised controlled trial blind investigators and
participants to treatment and control group (double-blinding).
Development of a protocol for the collection, measurement and
interpretation of information.
Use of standardised questionnaires or calibrated instruments,
such as sphygmomanometers.
Training of interviewers.
Recall (or response) bias - In a case-control study data on
exposure is collected retrospectively. The quality of the data is
therefore determined to a
large extent on the patient's ability to accurately recall past
exposures. Recall bias may occur when the information
provided on exposure differs
between the cases and controls. For example an individual with
the outcome under investigation (case) may report their
exposure experience
differently than an individual without the outcome (control)
under investigation.

Recall bias may result in either an underestimate or
overestimate of the association between exposure and outcome.
Methods to minimise recall bias include:
Collecting exposure data from work or medical records.
Blinding participants to the study hypothesis.
Social desirability bias occurs where respondents to surveys
tend to answer in a manner they feel will be seen as favourable
by others, for example
by over-reporting positive behaviours or under-reporting
undesirable ones. In reporting bias, individuals may selectively
suppress or reveal
information, for similar reasons (for example, around smoking
history). Reporting bias can also refer to selective outcome
reporting by study authors.
Performance bias refers to when study personnel or participants
modify their behaviour / responses where they are aware of
group allocations.
Detection bias occurs where the way in which outcome
information is collected differs between groups. Instrument bias
refers to where an
inadequately calibrated measuring instrument systematically
over/underestimates measurement. Blinding of outcome
assessors and the use of
standardised, calibrated instruments may reduce the risk of this.

2. Selection bias
Selection bias occurs when there is a systematic difference
between either:
Those who participate in the study and those who do not
(affecting generalisability) or
Those in the treatment arm of a study and those in the control
group (affecting comparability between groups).
That is, there are differences in the characteristics between
study groups, and those characteristics are related to either the
exposure or outcome under
investigation. Selection bias can occur for a number of reasons.
Sampling bias describes the scenario in which some individuals
within a target population are more likely to be selected for
inclusion than others.
For example, if participants are asked to volunteer for a study,
it is likely that those who volunteer will not be representative of
the general
population, threatening the generalisability of the study results.
Volunteers tend to be more health conscious than the general
population.
Allocation bias occurs in controlled trials when there is a
systematic difference between participants in study groups
(other than the intervention
being studied). This can be avoided by randomisation.
Loss to follow-up is a particular problem associated with cohort
studies. Bias may be introduced if the individuals lost to

follow-up differ with
respect to the exposure and outcome from those persons who
remain in the study. The differential loss of participants from
groups of a randomised
control trial is known as attrition bias.
• Selection bias in case-control studies
Selection bias is a particular problem inherent in case-control
studies, where it gives rise to non-comparability between cases
and controls. In case-
control studies, controls should be drawn from the same
population as the cases, so they are representative of the
population which produced the
cases. Controls are used to provide an estimate of the exposure
rate in the population. Therefore, selection bias may occur when
those individuals
selected as controls are unrepresentative of the population that
produced the cases.
The potential for selection bias in case-control studies is a
particular problem when cases and controls are recruited
exclusively from hospital or
clinics. Such controls may be preferable for logistic reasons.
However, hospital patients tend to have different characteristics
to the wider population,
for example they may have higher levels of alcohol
consumption or cigarette smoking. Their admission to hospital
may even be related to their
exposure status, so measurements of the exposure among
controls may be different from that in the reference population.
This may result in a biased
estimate of the association between exposure and disease.

For example, in a case-control study exploring the effects of
smoking on lung cancer, the strength of the association would
be underestimated if the
controls were patients with other conditions on the respiratory
ward, because admission to hospital for other lung diseases may
also be related to
smoking status. More subtly, the effect of alcohol on liver
disease could potentially be underestimated if controls are taken
from other wards: higher
than average alcohol consumption may result in admission for a
variety of other conditions, such as trauma.
As the potential for selection bias is likely to be less of a
problem in population-based case-control studies,
neighbourhood controls may be a
preferable choice when using cases from a hospital or clinic
setting. Alternatively, the potential for selection bias may be
minimised by selecting
controls from more than one source. For example, the use of
both hospital and neighbourhood controls.
• Selection bias in cohort studies
Selection bias can be less of problem in cohort studies
compared with case-control studies, because exposed and
unexposed individuals are enrolled
before they develop the outcome of interest.
However, selection bias may be introduced when the
completeness of follow-up or case ascertainment differs
between exposure categories. For

example, it may be easier to follow up exposed individuals who
all work in the same factory, than unexposed controls selected
from the community
(loss to follow-up bias). This can be minimised by ensuring that
a high level of follow-up is maintained among all study groups.
The healthy worker effect is a potential form of selection bias
specific to occupational cohort studies. For example, an
occupational cohort study
might seek to compare disease rates amongst individuals from a
particular occupational group with individuals in an external
standard population.
There is a risk of bias here because individuals who are
employed generally have to be healthy in order to work. In
contrast, the general population
will also include those who are unfit to work. Therefore,
mortality or morbidity rates in the occupation group cohort may
be lower than in the
population as a whole.
In order to minimise the potential for this form of bias, a
comparison group should be selected from a group of workers
with different jobs performed
at different locations within a single facility1; for example, a
group of non-exposed office workers. Alternatively, the
comparison group may be
selected from an external population of employed individuals.
• Selection bias in randomised trials
Randomised trials are theoretically less likely to be affected by
selection bias, because individuals are randomly allocated to the
groups being
compared, and steps should be taken to minimise the ability of
investigators or participants to influence this allocation process.
However, refusals to

participate in a study, or subsequent withdrawals, may affect the
results if the reasons are related to both exposure and outcome.
Confounding
Confounding, interaction and effect modification
Confounding provides an alternative explanation for an
association between an exposure (X) and an outcome. It occurs
when an observed association
is in fact distorted because the exposure is also correlated with
another risk factor (Y). This risk factor Y is also associated
with the outcome, but
independently of the exposure under investigation, X. As a
consequence, the estimated association is not that same as the
true effect of exposure X on
the outcome.
An unequal distribution of the additional risk factor, Y, between
the study groups will result in confounding. The observed
association may be due
totally, or in part, to the effects of differences between the
study groups rather than the exposure under investigation.1
A potential confounder is any factor that might have an effect
on the risk of disease under study. This may include factors
with a direct causal link to
the disease, as well as factors that are proxy measures for other

unknown causes, such as age and socioeconomic status.2
In order for a variable to be considered as a confounder:
1. The variable must be independently associated with the
outcome (i.e. be a risk factor).
2. The variable must also be associated with the exposure under
study in the source population.
3. The variable should not lie on the causal pathway between
exposure and disease.
Examples of confounding
A study found alcohol consumption to be associated with the
risk of coronary heart disease (CHD). However, smoking may
have confounded the
association between alcohol and CHD.
Smoking is a risk factor in its own right for CHD, so is
independently associated with the outcome, and smoking is also
associated with alcohol
consumption because smokers tend to drink more than non-
smokers.
Controlling for the potential confounding effect of smoking may
in fact show no association between alcohol consumption and
CHD.

Effects of confounding
Confounding factors, if not controlled for, cause bias in the
estimate of the impact of the exposure being studied. The
effects of confounding may
result in:
An observed association when no real association exists.
No observed association when a true association does exist.
An underestimate of the association (negative confounding).
An overestimate of the association (positive confounding).
Controlling for confounding
Confounding can be addressed either at the study design stage,
or adjusted for at the analysis stage providing sufficient
relevant data have been
collected. A number of methods can be applied to control for
potential confounding factors and the aim of all of them is to
make the groups as similar
as possible with respect to the confounder(s).
Controlling for confounding at the design stage
Potential confounding factors may be identified at the design
stage based on previous studies or because a link between the

factor and outcome may
be considered as biologically plausible. Methods to limit
confounding at the design stage include randomisation,
restriction and matching.
• Randomisation
This is the ideal method of controlling for confounding because
all potential confounding variables, both known and unknown,
should be equally
distributed between the study groups. It involves the random
allocation (e.g. using a table of random numbers) of individuals
to study groups.
However, this method can only be used in experimental clinical
trials.
• Restriction
Restriction limits participation in the study to individuals who
are similar in relation to the confounder. For example, if
participation in a study is
restricted to non-smokers only, any potential confounding effect
of smoking will be eliminated. However, a disadvantage of
restriction is that it may
be difficult to generalise the results of the study to the wider
population if the study group is homogenous.1
• Matching
Matching involves selecting controls so that the distribution of
potential confounders (e.g. age or smoking status) is as similar
as possible to that
amongst the cases. In practice this is only utilised in case-
control studies, but it can be done in two ways:
1. Pair matching - selecting for each case one or more controls

with similar characteristics (e.g. same age and smoking habits)
2. Frequency matching - ensuring that as a group the cases have
similar characteristics to the controls
Detecting and controlling for confounding at the analysis stage
The presence or magnitude of confounding in epidemiological
studies is evaluated by observing the degree of discrepancy
between the crude
estimate (without controlling for confounding) and the adjusted
estimate after accounting for the potential confounder(s). If the
estimate has changed
and there is little variation between the stratum specific ratios
(see below), then there is evidence of confounding.
It is inappropriate to use statistical tests to assess the presence
of confounding, but the following methods may be used to
minimise its effect.
• Stratification
Stratification allows the association between exposure and
outcome to be examined within different strata of the
confounding variable, for example
by age or sex. The strength of the association is initially
measured separately within each stratum of the confounding
variable. Assuming the stratum

specific rates are relatively uniform, they may then be pooled to
give a summary estimate as adjusted or controlled for the
potential confounder. An
example is the Mantel-Haenszel method. One drawback of this
method is that the more the original sample is stratified, the
smaller each stratum will
become, and the power to detect associations is reduced.
• Multivariable analysis
Statistical modelling (e.g. multivariable regression analysis) is
used to control for more than one confounder at the same time,
and allows for the
interpretation of the effect of each confounder individually. It is
the most commonly used method for dealing with confounding
at the analysis stage.
• Standardisation
Standardisation accounts for confounders (generally age and
sex) by using a standard reference population to negate the
effect of differences in the
distribution of confounding factors between study populations.
See “Numerators, denominators and populations at risk” for
more details.
Residual confounding
It is only possible to control for confounders at the analysis
stage if data on confounders were accurately collected. Residual
confounding occurs
when all confounders have not been adequately adjusted for,
either because they have been inaccurately measured, or
because they have not been
measured (for example, unknown confounders). An example

would be socioeconomic status, because it influences multiple
health outcomes but is
difficult to measure accurately.3
Interaction (effect modification)
Interaction occurs when the direction or magnitude of an
association between two variables varies according to the level
of a third variable (the effect
modifier). For example, aspirin can be used to manage the
symptoms of viral illnesses, such as influenza. However, whilst
it may be effective in
adults, aspirin use in children with viral illnesses is associated
with liver dysfunction and brain damage (Reye’s syndrome).4 In
this case, the effect of
aspirin on managing viral illnesses is modified by age.
Where interaction exists, calculating an overall estimate of an
association may be misleading. Unlike confounding, interaction
is a biological
phenomenon and should not be statistically adjusted for. A
common method of dealing with interaction is to analyse and
present the associations for
each level of the third variable. In the example above, the odds
of developing Reye’s syndrome following aspirin use in viral
illnesses would be far
greater in children compared to adults, and this would highlight
the role of age as an effect modifier. Interaction can be
confirmed statistically, for
textbook/research-methods/1a-epidemiology/numerators-
denominators-populations

‹ Association and Causation up Interactions, methods for
assessment of effect modification ›
example using a chi-squared test to assess for heterogeneity in
the stratum-specific estimates. However, such tests are known
to have a low power for
detecting interaction5 and a visual inspection of stratum-
specific estimates is also recommended.
References
1. Hennekens CH, Buring JE. Epidemiology in Medicine,
Lippincott Williams & Wilkins, 1987.
2. Carneiro I, Howard N. Introduction to Epidemiology. Open
University Press, 2011.
3. http://www.edmundjessop.org.uk/fulltext.doc - Accessed
20/02/16
4. McGovern MC. Reye’s syndrome and aspirin: lest we forget.
BMJ 2001;322:1591.
5. Marshall SW. Power for tests of interaction: effect of raising
the type 1 error rate. Epidemiological perspectives and
innovations 2007;4:4.
© Helen Barratt, Maria Kirwan 2009, Saran Shantikumar 2018

textbook/research-methods/1a-epidemiology/association-
causation
textbook/research-methods/1a-epidemiology/confounding-
interactions-methods
http://www.edmundjessop.org.uk/fulltext.doc%20-
%20Accessed%2020/02/
Navigation
Use of routine vital and health statistics to describe the
distribution of disease in time and place and by person
Numerators, denominators and populations at risk
Methods for Summarising Data
Incidence and prevalence including direct and indirect
standardisation
Years of Life Lost
textbook/research-methods/1a-epidemiology/use-of-routine-
vital-and-health-statistics
textbook/research-methods/1a-epidemiology/numerators-
denominators-populations

textbook/research-methods/1a-epidemiology/methods-
summarising-data
textbook/research-methods/1a-epidemiology/incidence-
prevalence
textbook/research-methods/1a-epidemiology/years-lost-life
Measures of disease burden (event-based and time-based) and
population attributable risks including identification of
comparison groups
appropriate for public health
Sources of variation, its measurement and control
Common errors in epidemiological measurements, their effects
on numerator and denominator data and their avoidance
Concepts and measures of risk - The odds ratio, the rate ratio
and risk ratio (relative risk)
Association and Causation
Biases and Confounding
Interactions, methods for assessment of effect modification
Strategies to allow/adjust for confounding in design and
analysis

The design, applications, strengths and weaknesses of
descriptive studies and ecological studies
Design, applications, strengths and weaknesses of cross-
sectional, analytical studies (including cohort, case-control and
nested case-control
studies), and intervention studies (including randomised
controlled trials)
Analysis of health and disease in small areas
Validity, reliability and generalisability
Intention to treat analysis
Clustered data - effects on sample size and approaches to
analysis
Numbers needed to treat (NNTs) - calculation, interpretation,
advantages and disadvantages
Time-trend analysis, time series designs
Nested case-control studies
Methods of sampling from a population
Methods of allocation in intervention studies
The design of documentation for recording survey data
Construction of valid questionnaires
Methods for validating observational techniques
Studies of disease prognosis

Appropriate use of statistical methods in the analysis and
interpretation of epidemiological studies, including life-table
analysis
textbook/research-methods/1a-epidemiology/measures-disease-
burden
textbook/research-methods/1a-epidemiology/sources-variation-
measurement-control
textbook/research-methods/1a-epidemiology/common-errors-
epidemiologoical-measurements
https://www.healthknowledge.org.uk/node/713
textbook/research-methods/1a-epidemiology/association-
causation
textbook/research-methods/1a-epidemiology/biases
textbook/research-methods/1a-epidemiology/confounding-
interactions-methods
textbook/research-methods/1a-epidemiology/descriptive-
studies-ecological-studies
textbook/research-methods/1a-epidemiology/cs-as-is
textbook/research-methods/1a-epidemiology/health-disease-
analysis
https://www.healthknowledge.org.uk/content/validity-
reliability-and-generalisability

textbook/research-methods/1a-epidemiology/clustered-data
textbook/research-methods/1a-epidemiology/nnts
textbook/research-methods/1a-epidemiology/time-trend-analysis
textbook/research-methods/1a-epidemiology/nested-case-
control-studies
textbook/research-methods/1a-epidemiology/methods-of-
sampling-population
allocation-intervention-studies
textbook/research-methods/1a-epidemiology/design-
documentation-recordingsurvey
textbook/research-methods/1a-epidemiology/construction-valid-
questionnaires
validating-observational-techniques
textbook/research-methods/1a-epidemiology/sudies-disease-
prognosis
textbook/research-methods/1a-epidemiology/statistical-
methods-analysis-interpretation

Epidemic theory (effective and basic reproduction numbers,
epidemic thresholds) and techniques for infectious disease data
(construction and
use of epidemic curves, generation numbers, exceptional
reporting and identification of significant clusters)
Systematic reviews, methods for combining data from several
studies, and meta-analysis
Electronic bibliographical databases and their limitations
Grey literature
Publication bias
Evidence based medicine and policy
The hierarchy of research evidence - from well conducted meta-
analysis down to small case series
The Cochrane collaboration
The ethics and etiquette of epidemiological research
Understanding of basic issues and terminology in the design,
conduct, analysis and interpretation of population-based genetic
association
studies, including twin studies, linkage and association studies
textbook/research-methods/1a-epidemiology/epidemic-theory
textbook/research-methods/1a-epidemiology/systematic-
reviews-methods-combining-data
textbook/research-methods/1a-epidemiology/electronic-

bibliographies
textbook/research-methods/1a-epidemiology/grey-literature
https://www.healthknowledge.org.uk/content/publication-bias
textbook/research-methods/1a-epidemiology/evidence-based-
medicine-policy
textbook/research-methods/1a-epidemiology/hierarchy-research-
evidence
textbook/research-methods/1a-epidemiology/cochrane-
collaboration
textbook/research-methods/1a-epidemiology/ethics-etiquette-
epidemiology-research
textbook/research-methods/1a-epidemiology/issues-
terminology-genetic-based-studies
https://googleads.g.doubleclick.net/aclk?sa=l&ai=CeafFqBDOW
4j6LtbpBdWogqgI95L_qVOX1Yes6QePx5PnlwwQASDxwbkjY
Mm-
soj0o8AQoAHeoo2YA8gBA6gDAcgDyQSqBOUBT9BIHlv9aEy
5gIQV0-
QwLb2Rcrtv1c0O97dySpUar6NXgbbD9xqIFXzsoM7wsjsH-
nZGEGdgiqganmnVPwUVtsBgnxH7Qw0DFWCP0o1iDKCD7hL
1TNXxHqzDUlgoOReS8wwLikA9Swz1odhQ58iQiUy7y5WPUc
afzdZ48-wfiHwgMQh8-
UbzudGmEjNFHxzRC5WJ2NBgxXgMuMexnQLmNMBquZDhK

AGPRxgB5GV3TBOLpTsmqR6rkwXcbgM3Oq2BI-
Tpp8xfmrSmb0nDj5HpGdNqVyufIahRW9zwPlsfCF_AXgIX4qA
GA4AHlZLlR6gHjs4bqAfVyRuoB6gGqAfZyxuoB8_MG6gHpr4
bqAeaBtgHAdIIBwiAYRABGAKxCSqd6AijHxfsgAoB2BMM&
num=1&sig=AOD64_2e8xc9mYb92STG9B9_iTakbUTJ_A&clie
nt=ca-pub-
9562910130782731&adurl=https://www.liligal.com/Flash-Sale-
Top-vc-421-1.html
https://googleads.g.doubleclick.net/aclk?sa=l&ai=CeafFqBDOW
4j6LtbpBdWogqgI95L_qVOX1Yes6QePx5PnlwwQASDxwbkjY
Mm-
soj0o8AQoAHeoo2YA8gBA6gDAcgDyQSqBOUBT9BIHlv9aEy
5gIQV0-
QwLb2Rcrtv1c0O97dySpUar6NXgbbD9xqIFXzsoM7wsjsH-
nZGEGdgiqganmnVPwUVtsBgnxH7Qw0DFWCP0o1iDKCD7hL
1TNXxHqzDUlgoOReS8wwLikA9Swz1odhQ58iQiUy7y5WPUc
afzdZ48-wfiHwgMQh8-
UbzudGmEjNFHxzRC5WJ2NBgxXgMuMexnQLmNMBquZDhK
AGPRxgB5GV3TBOLpTsmqR6rkwXcbgM3Oq2BI-
Tpp8xfmrSmb0nDj5HpGdNqVyufIahRW9zwPlsfCF_AXgIX4qA
GA4AHlZLlR6gHjs4bqAfVyRuoB6gGqAfZyxuoB8_MG6gHpr4
bqAeaBtgHAdIIBwiAYRABGAKxCSqd6AijHxfsgAoB2BMM&
num=1&sig=AOD64_2e8xc9mYb92STG9B9_iTakbUTJ_A&clie
nt=ca-pub-
9562910130782731&adurl=https://www.liligal.com/Flash-Sale-
Top-vc-421-1.html

Our most popular content
Public Health Textbook
Identifying and managing internal and external stakeholder
interests
Management models and theories associated with motivation,
leadership and change management, and their application to
practical situations and
problems
Dietary Reference Values (DRVs), current dietary goals,
recommendations, guidelines and the evidence for them
Section 1: The theoretical perspectives and methods of enquiry
of the sciences concerned with human behaviour
Inequalities in health (e.g. by region, ethnicity, soci-economic
position or gender) and in access to health care, including their
causes
The impact of political, economic, socio-cultural, environmental
and other external influences
Introduction to study designs - intervention studies and
randomised controlled trials
2h - Principles and Practice of Health Promotion
Parametric and Non-parametric tests for comparing two or more
groups
Recently updated content
3b - Sickness and Health
5d - Understanding the Theory and Process of Strategy
Development
3a - Populations
2a - Epidemiological Paradigms
1a - Epidemiology
2d - Genetics
2c - Diagnosis and Screening

1d - The Principles of Qualitative Methods
1c - Approaches to the assessment of health care needs,
utilisation and outcomes, and the evaluation of health and
health care
5e Health and social service quality
https://www.healthknowledge.org.uk/public-health-textbook
textbook/organisation-management/5b-understanding-
ofs/managing-internal-external-stakeholders
textbook/organisation-management/5c-management-
change/basic-management-models
textbook/disease-causation-diagnostic/2e-health-social-
behaviour/drvs
textbook/medical-sociology-policy-economics/4a-concepts-
health-illness/section1
textbook/medical-sociology-policy-economics/4c-equality-
equity-policy/inequalities-distribution
textbook/organisation-management/5b-understanding-
ofs/assessing-impact-external-influences
https://www.healthknowledge.org.uk/e-
learning/epidemiology/practitioners/introduction-study-design-
is-rct
textbook/disease-causation-diagnostic/2h-principles-health-
promotion
textbook/research-methods/1b-statistical-methods/parametric-
nonparametric-tests
textbook/health-information/3b-sickness-health

textbook/organisation-management/5d-theory-process-strategy-
development
textbook/health-information/3a-populations
textbook/disease-causation-diagnostic/2a-epidemiological-
paradigms
textbook/disease-causation-diagnostic/2d-genetics
textbook/disease-causation-diagnostic/2c-diagnosis-screening
textbook/research-methods/1d-qualitative-methods
textbook/research-methods/1c-health-care-evaluation-health-
care-assessment
textbook/organisation-management/5e-health-and-social-
service-quality
Disclaimer | Copyright © Public Health Action Support Team
(PHAST) 2017 | Contact Us
Company Information - Public Health Action Support Team CIC
[registered in England and Wales under Company No.
06480440]
Registered Office - Sterling House, 20 Station Road, Gerrards

Cross, Bucks, SL9 8EL
http://www.phast.org.uk/
https://www.healthknowledge.org.uk/about-us/website-
disclaimer
https://www.healthknowledge.org.uk/about-us/terms-and-
conditions
https://www.healthknowledge.org.uk/contact
http://www.dh.gov.uk/en/index.htm

Workshop 4Reginald Finger, MD, MPH Jiajoyce Conway, DNP, CRN.docx

Recommended

Recommended

More Related Content

Similar to Workshop 4Reginald Finger, MD, MPH Jiajoyce Conway, DNP, CRN.docx

Similar to Workshop 4Reginald Finger, MD, MPH Jiajoyce Conway, DNP, CRN.docx (20)

More from dunnramage

More from dunnramage (20)

Recently uploaded

Recently uploaded (20)

Workshop 4Reginald Finger, MD, MPH Jiajoyce Conway, DNP, CRN.docx