Epidemiology Assignment: Study Design, Sampling Methods, and Statistical Analysis

EPID6430 Epidemiology B - 2018
Assignment 1
Question 1 (Total = 4 marks)
Recommendations made.
Epidemiological studies are aimed at revealing any unbiased relationships between different
exposures which can include alcohol, stress, biological factors, mortality or morbidity.
Identifying this causal relationship is crucial in any study. Establishing the exposure and
outcome is a critical aspect 1. Thus, in this case, there is a determination of causality and
effect between tonsillectomy and mortality which is the eventual outcome. Thus in order to
undertake a successful study, a case-control study shall be undertaken. In this case,
individuals with disease positive are compared against disease negative. Case-control
reviews past action at exposure period which both groups have encountered and
association created based on outcomes 2.
Question 2 (Total = 6 marks)
Sampling methods to be used and their reasons
a) Prevalence of asthma in 4-year-old residents in Canberra? (2 marks)
- Purposive sampling
This method of sampling allows for subjective or selective of the subjects to
participate in the study. A researcher may choose a representative sample to suit
their needs based on their specific characteristics. In this study, only residents
who have lived for 4 years in Canberra are selected to participate in the study
while those who don’t meet these criteria are not selected to participate in the
study3.
b) Prevalence of diabetes mellitus in Australia. (2 marks)
- Simple random sampling
In this assessment study, the respondents will be chosen through chance and all
the population of Australia has an equal chance. This method is crucial as it gives
error sampling to be calculated and reduces selection biases. Thus all the
population has an equal chance of selection to assess diabetes prevalence 3.
1

c) Measles vaccine coverage in a district of Papua New Guinea. (2 marks)
- Clustered sampling
This technique allows for identification and selection of population subgroups
compared to undertaking individual selection. The population under this
consideration are the clusters which in this case Papua district being defined for
the coverage assessment to be undertaken 3.
Question 3 (Total = 10 marks)
Statistics table
Alcohol consumption group
Number of None Occasional Light Moderate Heavy
Subjects 466 1845 2544 2042 832
Deaths 41 142 143 116 62
i) Calculate the risk and the relative risk (4 marks)
Alcohol consumption Subjects Deaths
None 466 41
Occasional 1845 142
Light 2544 143
Moderate 2042 116
Heavy 832 62
Relative risk for occasional drinkers to die
RR = a(a+b)/c(c+d)
=466(507)/1845(1987)
= 0.919/0.928
= 0.99
= the risk of dying for occasional drinkers is less likely to occur.
Relative risk for light drinkers to die
RR = a(a+b)/c(c+d)
=466(507)/2544(2687)
= 0.919/0.946
= 0.97
2

= the risk of dying for light drinkers is less likely to occur.
Relative risk for moderate drinkers to die
RR = a(a+b)/c(c+d)
=466(507)/2042(2158)
= 0.919/0.946
= 0.971
= the risk of dying for occasional drinkers is less likely to occur.
Relative risk for heavy drinkers to die
RR = a(a+b)/c(c+d)
=466(507)/832(894)
= 0.919/0.93
= 0.98
= the risk of dying for heavy drinkers is less likely to occur.
ii) Conclusions and reasoning behind the rationale given. (2 + 4 = 6 marks)
The conclusion s on the above is misleading in that intervening factors were not
controlled during the study. This can be seen with the rate of none drinkers
having more deaths observed hence meaning that deaths being witnessed are
merely caused by other factors rather than the drinking of alcohol.
For an effective study, prospective study will be undertaken and followed among
the drinkers of alcohol. Data collection tools to be used include follow up
questionnaires, surveys, follow-up assessment. The variables of importance in
this study will be the disease status of the respondents, a number of alcohol
drinking per volume consumed, frequency of intake and type of alcohol
consumed 4.
Question 4 (Total = 10 marks)
Statistics table
The death rate by cause of death and by cigarette habit at last survey:
Cause of death Annual death rate per 100,000 men
Non-
smokers
Current smokers (cigarettes/day) All
smokers
1-14 15-24 >25
Lung Cancer 10 78 127 251 140
Ischaemic Heart 413 608 652 792 669
3