UWS Epidemiology (401076) Assignment 2: CHD Study Analysis, 2018

This assignment provides a comprehensive analysis of epidemiological studies related to coronary heart disease (CHD). It begins by examining a cohort study on the long-term effects of physical activity on CHD incidence, addressing aspects such as study design, data sources, and the impact of confounders like hyperlipidemia, hypertension, and diabetes. The assignment further delves into a case-control study investigating the effect of dietary fat intake on melanoma, calculating relative risks and attributable risks to assess the association between low-fat diets and skin cancer. Additionally, it explores the concept of bias in cohort and case-control studies and discusses methods for addressing confounding factors. The analysis covers key epidemiological concepts such as risk ratios, rate ratios, and the interpretation of statistical results, providing a detailed understanding of the factors influencing CHD and melanoma risks. Desklib offers a wealth of similar solved assignments and past papers to aid students in their studies.

The University of Western Sydney
School of Science and Health
Introduction to Epidemiology (401076)
ASSIGNMENT 2
Spring, 2018

Based on the following study
Sundquist K., Qvist J. Johansson SE., Sundquist J. The long-term effect of physical activity
on incidence of coronary heart disease: a 12-year follow-up study Preventive Medicine
2005;41:219-225
Question 1
(a) The study design employed in the study [1 mark]
- This is a cohort study utilized prospective study. The outcomes were assessed as
the study progressed and outcomes measured at the end of the study.
(b) Data sources for coronary heart diseases in the study [2 marks]
- Swedish national discharge register data source was used to identify the events
coronary heart disease from the respondents.
(c) Why the authors excluded 1894 participants who rated their general health as “poor”
[2 marks]
- The authors excluded these participants as it would have affected the study
because of its confounding effects on the issue of interest in the study which was
coronary heart disease emergence. Having respondents with underlying or poor
health conditions could lead to coronary heart disease being linked to other
causative factors part from the issue of interest which is the respondent’s physical
activity.
(d) The crude overall CHD incidence in the study [2 marks]
- The overall crude incidence rate was 59 cases per 10,000 persons over physical
activity levels.
(e) Comparison of physical activity being undertaken at least twice a week for low
income and all other income categories [4 marks]
- The risks obtained at 95% confidence level reveals that physical activity of twice
daily had RR of 0.72 (CI 0.52-1.01) compared to low-income earners having RR
of 1.20 (0.95-1.52) and those of all other income earners at RR equals to 1.
(f) The interpretation of the relative risk above [3 marks]
- The relative risks above illustrate that exposure to undertaking physical activity
twice a week decreased the exposure of developing coronary heart disease.
- Low income was further associated with the risk of developing coronary heart
disease, thus being a low-income earner had a greater risk of developing coronary
heart disease, while all other income earners, in general, had no effect on
developing coronary heart disease.
(g) The relative risk in this study is better described as risk ratio in that [2 marks]
- Risk ratio refers to the probability of an outcome in an exposed group to the
probable outcome of the exposed group, while rate ratio refers to comparing rates
of events taking place at any given time. This study is best explained when risks
ratio is assessed on the risks of developing coronary heart disease to factors such
as physical activity, income, smoking and BMI assessments for the respondents.

In this case the strength of the association between the risks factors and outcome
of developing coronary heart disease.
(h) The major confounders in this study which were not included include the following
[4 marks]
- Some of the major confounders which were not assessed in this study include
hyperlipidemia, which has been considered as a powerful indicator of coronary
heart disease, with the positive association being linked to cholesterol levels.
Further hypertension could be a major confounding fact in the development of
coronary heart disease. Hypertension has been attributed as a strong and
independent risk factor for the development of coronary heart disease causing
morbidity and mortality.
- Further diabetes mellitus could a major confounding factor. Both type 1 and 2,
have been associated with elevated risks for developing heart disease. Thus it
could act as a confounding factor in the development of heart disease.
(i) These confounders were not adjusted due to the following reasons [4 marks]
- The confounding factors were not added into the study due to the longitudinal
nature of the study, which followed the group cohorts over 10 years, in this case,
effects of variables under study was sought.
- The study was designed in such a way that those respondents who attributed their
health state being poor or bad were excluded in the participation. Hence those
having poor self-rating of their disease condition were not included in the study,
and the study only focused on those who had good health status.
Question 2:
Tasmanian researchers conducted a case-control study to investigate the effect of dietary fat
intake on melanoma (Skin Cancer). They hypothesized that people whose dietary fat intake is
low will be more susceptible to skin cancer. The study compared 500 cases of melanoma
with 500 controls. The controls were randomly selected from the state’s electoral roll. The
researchers categorized the dietary fat intake into three categories High, Moderate and Low.
They found that among skin cancer patients 150 were classified at the low and 80 at the high
dietary fat intake whereas among control 130 were at the low and 100 were at the high.
a) Summarised table for estimation of the level of association of the diseases and fat
intake [4 points]
Having disease The absence of the
disease
Total
High fat intake 80 100 180
Moderate fat
intake
270 270 540
Low fat intake 150 130 280
Total 500 500 1000

b) The relative risk (RR) of high fat intake versus low fat intake and the RR of medium
to high fat intake in relation to the development of melanoma was found to be
[4 marks]
RR of high fat intake versus low fat intake
= a/(a+b)/ c/(c+d)
= a-80, b-100, c-150,d-130
= (80/180)/(150/180)
= 0.444/0.833
= 0.533
This result indicates that there is a 50% decrease in the development of the cancer
disease among those who are taking high fat intake compared to those who were
taking low fat intake.
RR of medium to high fat intake
= a/(a+b)/ c/(c+d)
= a-270, b-270, c-80,d-100
= (270/540)/(80/180)
= 0.5/0.44
= 1.1336
This relative risk indicates that RR medium to high fat intake show that melanoma
risk is increased with intake of high fat. Thus cancer risk is increased with the fat
exposure.
These results of the relative risks indicate an inverse association of risk that in that,
exposure portrays different results as expected. Hence the actual relative risk of
exposure is not obtained from the study calculations.
c) The attributable risk due to exposure to the low-dietary fat intake on Melanoma
[4 points]
Cases Controls Totals
Exposed 150 130 280
Unexposed 350 370 720
Total 500 500 1000
AR = IE-IU
= P(D/E)-P(D/U)
= a/(a+b)-c/(c+d)
= 150/280-350/720
= 0.53-0.48
AR% = 0.05x100

= 5%
This result indicates that there is a 5% difference in melanoma cancer in exposed and
unexposed individual in the study. Thus the attributable risk is lower in these findings.
d) Calculation of the population attributable risk of low-fat intake on melanoma reveal
that [4 points]
PAR = (IT-IU)/IT
= Pe(RR-1)/Pe(RR-1)+1)
= 5(1.1336-1)/ 5(1.1336-1)+1
= 0.668/1.668
= 0.40
Thus the excess of the disease in the population attributed to the exposure is 0.4. Thus
assuming that there is a causal association, there is low exposure of the disease in the
population among low fat intake eaters.
e) The conclusion of the PAR regarding exposure to a low-fat diet to reduce melanoma
shows that [4 points
The conclusions reveal that that low intake of intake of fat in the population is
attributed to low disease incidence. Hence the population has a low intake of fat have
a low occurrence of melanoma cancer.
Question 3:
In a small cohort study investigating the effect of a rare exposure ( E), the following results
were found:
Table 1 Disease
Exposure
A- The data above shows that
[2 points]
There is no association between disease and exposure in the above data; this is observed
through calculation of the relative risk which is
RR = a(a+b)/c(c+d)
= (120/480)/(120/380)
= 0.25/0.315
= 0.80
= thus no strength of association
Rate Ratio (Increased risk) Rate Ratio (Decreased risk) Strength of Association
1.0 – 1.2 0.9 – 1.0 None
1.2 – 1.5 0.7 – 0.9 Weak
>1.5 < 0.9 Moderate to Strong
*Adapted from Monson (1990)1
1 Monson R. Occupational Epidemiology, 2nd edition. Boca Raton, Florida: CRC Press Inc., 1990.
Yes No
Yes 120 360
No 120 360

B- A stratified analysis by Age-groups shows the following:
Younger adults Older adults
Disease Disease
Exposure
Yes No
Exposure Yes No
Yes 60 180 Yes 80 160
No 40 160 No 60 180
The relative risks for both older adults and younger adults is calculated as
[ 4 points]
RR = a(a+b)/c(c+d)
Among younger adults
RR = (60/240)/(40/200)
= 0.25/0.2
= 1.25
Among older adults
RR = (80/240)/(60/240)
= 0.33/0.25
= 1.32
c- The results [4 points]
From the above-calculated risks, the values obtained shows that there is a weak strength of
association between disease and exposure across the two population groups. Hence the
exposure might not be causing the disease; there is the presence of other factors which might
have an impact on disease state.
Question 4:
a) A classical example of a biasness in Cohort cases is [4 marks]
Subject selection biases. This is related to the exposure and outcome. The subjects are
enrolled in the prospective study before developing outcome of interest. This is easier to
observe how enrolment is related to exposure. This bias is common in a retrospective study,
where individuals have to provide informed consent before participation. Hence those with
the most interests in the disease have been exposed.
b) Typical biasness in a case-control study is [4 marks]
Selection bias. This occurs when the subjects chosen for the control group do not represent
the actual population. A typical example is observed when investigators while undertaking

cervical cancer, relating to socioeconomic status. The investigators identified control subjects
through a door to door recruitment in the community. The issue arose in that the cases were
selected using different method compared to controls hence displaying a typical selection bias
of cervical cancer patients based on their socioeconomic status.
Q 5 A cohort study was conducted to examine cigarette smoking and the risk of oral cancer.
The investigators selected exposed and unexposed subjects so that they had exactly the same
distribution of race. A typical method of addressing this probable confounding factor is
(2 points)
Stratification
An effective method to address this kind of error is through stratification by race. This
ensures that all the races in the study have an equal chance of participating in the study.
Question 6:
A study aims to determine the incidence of type 2 diabetes. A cohort of 200 people age 65
years or older who were initially disease-free participated in the study. One hundred and fifty
people were examined at the end of 3 years. Fifty other participants from the initial cohort
could not be examined, including 11 people who had died.
[4 points]
Participant loss in this study doesn't imply a source of biases. Studies undertaken have shown
that there is no difference on the impact of association between health states and variables
being assessed. The effect of participation doesn’t affect the exposure and outcome
association being investigated2. Various studies have indicated that this effect causes minimal
biases in assessing exposure and outcome associations.3
2 Banks J, Muriel A, Smith JP. Attrition and health in aging studies: evidence from ELSA and HRS.
Longit Life Course Stud. 2011;2(2):101-26
3 Pizzi C, De Stavola B, Merletti F, Bellocco R, dos Santos Silva I, Pearce N, Richiardi LJ Epidemiol
Community Health. 2011 May; 65(5):407-11.