Scribd
Upload
50 views13 pages

Tuto - Bias, Confounding, Interaction - With Key

The document discusses different types of biases that can be introduced in epidemiological studies, including selection bias, recall bias, non-response bias, misclassification bias, and confounding. It provides examples of studies that may exhibit each type of bias and questions to help understand these concepts. It also covers interaction and how to assess for additive and multiplicative interaction using tables of study data.

Uploaded by

Kamran Ahmed
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
50 views13 pages

Tuto - Bias, Confounding, Interaction - With Key

The document discusses different types of biases that can be introduced in epidemiological studies, including selection bias, recall bias, non-response bias, misclassification bias, and confounding. It provides examples of studies that may exhibit each type of bias and questions to help understand these concepts. It also covers interaction and how to assess for additive and multiplicative interaction using tables of study data.

Uploaded by

Kamran Ahmed
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 13

BIAS, CONFOUNDING &

INTERACTION

Tutorial
23rd November 2023
1
• In a case control study, mothers of children with congenital disease were
more likely to remember the exposure to radiation as compared to mothers
whose children were perfectly healthy. This may introduce :
a. Selection bias
b. Interviewer bias
c. Misclassification bias
d. Non-response bias
e. Recall bias
E
2
• A survey was conducted through a mailed questionnaire to estimate
the prevalence of chronic bronchitis in an urban city. Response rate
was 85%. Those who did not respond to the questionnaire were
mostly smokers. What type of bias is introduced in the study? What
is its effect

a. Selection bias
b. Interviewer bias
c. Misclassification bias
d. Non-response bias
e. Recall bias
D
3
• In a survey on use of protective gloves while drawing blood samples,
most technicians actually did not practice wearing gloves but in order to
be in good books of hospital administration replied positively to the
interviewer. This will introduce:
a. Reporting bias
b. Attention bias
c. Exclusion bias
d. Compliance bias
e. Non response bias

A
4
• In a prospective study on relationship between using un-boiled water and
Hepatitis A, a few households which did not boil water in the beginning
started boiling water in the middle of the follow up period. This introduced:

a. Reporting
b. Attention bias
c. Exclusion bias
d. Compliance bias
e. Non response bias

B
5
• A study on relationship between HIV AIDS and drug abuse used a test with
low sensitivity to diagnose cases and low specificity to rule out controls.
a) Which bias may be introduced in the study?
b) What will be its effect?

a. Selection bias
b. Interviewer bias
c. Misclassification bias
d. Non-response bias
e. Recall bias
C
6
In a case-control study of the relationship of radiation exposure and thyroid cancer, 50 cases
admitted for thyroid cancer and 100 “controls” admitted during the same period for treatment of
hernias were studied. Only the cases were interviewed, and 20 of the cases were found to have
been exposed to x-ray therapy in the past, based on the interviews and medical records. The
controls were not interviewed, but a review of their hospital records when they were admitted for
hernia surgery revealed that only 2 controls had been exposed to x-ray therapy in the past.
Based on the description given above, what source of bias is least likely to be present in this
study?
a. Recall bias

b. Misclassification of cases and control

c. Exclusion bias

d. Compliance bias

e. Non response bias

D
7
• A case control study was conducted to see the relationship between intake of milk and Vitamin D Deficiency.
Comment on the confounding effect of Living in Flats?
a) From the following table calculate OR
Cases Controls
Milk+ 10 80
Not taking milk - 20 85

OR= ad/bc=0.53 (shows negative association)


b) When people living in flats were removed from the data, following results came. Calculate the
OR now? Interpret Cancer Cases Controls

Milk + 8 40
Living in Flats is a
Not taking milk 10 50
confounder between
-
relationship of Taking
milk and Vitamin D
Deficiency
OR =1
8
• A randomized control was conducted to see the effect of using SMS alerts on control of fasting
blood sugar?
a) From the following table calculate RR
Cases (uncontrolled FBS) Controls (Controlled FBS)
Received SMS alerts 50 150
Did not receive alerts 50 75

RR= 0.25/0.4=0.62 (shows negative association)


b) To adjust for effect of education, educated participants were removed from the data, following
results came. Calculate the RR now? Interpret
Cases (uncontrolled FBS) Controls (Controlled FBS)

Received SMS alerts 25 75


Did not receive alerts 38 62

RR =0.25/0.38=0.65
Education is not a confounder between relationship of
using SMS alerts on control of fasting blood sugar!
9
Factor A
- +
Factor - 4 8
B
+ 12

• Fill in the blank cell in the first table using the additive model of
interaction:
a) 16

• Fill in the blank cell in the first table using the multiplicative model of
interaction:
b) 24
10
Factor A
- +
Factor - 3 6
B
+ 9 21

• Is there interaction between factor A and B ?


Yes
• If Yes which model: Additive or multiplicative?
Multiplicative effect
11
Factor A
- +
Factor - 3 5
B
+ 8 8

• Is there interaction between factor A and B ?


No
• If Yes which model: Additive or multiplicative? If No: Why?
Effect is less than additive effect
ANY QUESTIONS?
Thank you!

You might also like