Assignment 3
Assignment 3
QUESTION 1 (1 point)
a) Statistical combination of the results of two or more studies
QUESTION 2 (1 point)
Figure 2 of the Jordan et al., meta-analysis depicts a forest plot illustrating the estimates of the association between hysterectomy and ovarian cancer. How many of the individual studies (in figure 2) reported statistically significant reductions in the risk of ovarian cancer in women with hysterectomy, compared to women without hysterectomy?
Ans) There were 20 individual studies that reported statistically significant reductions in the risk of ovarian cancer in women with hysterectomy, compared to women without hysterectomy.
QUESTION 3 (1 point)
How would you interpret the overall summary estimate (relative risk) provided in Figure 2? (1-2 sentences)
Ans) The overall relative risk was estimated to be 0.81, with a 95% confidence level of 0.72-0.92.
QUESTION 4 (2 points)
Is it appropriate to use the overall summary estimate provided in figure 2 to understand the true association between hysterectomy and ovarian cancer? Why/ why not? (2-3 sentences).
Ans) No, it is inappropriate to use the overall summary estimate as true association between hysterectomy and ovarian cancer because the study was conducted only in women with menopause. Women before that age were not included in the study. Moreover, the study was conducted only on the available number of articles, and hence can be biased.
QUESTION 5 (1 point)
What significant sources of heterogeneity did the authors identify in Table 2 (Select all answers that are correct. Note there may be more than one answer).
b) Year of diagnosis
c) Control response rate
d) Country of study
e) Hysterectomy prevalence
QUESTION 6 (1 point)
In the forest plots in figures 2 and 3, you can see that each of the studies report confidence intervals. The narrower the confidence interval, the higher the weight assigned to the study (as depicted by the size of the shaded square/rectangle symbols around each study estimate). What is generally the reason some studies have narrower (smaller) confidence intervals compared to others? (Select 1 answer only)
b) The studies have larger sample sizes
QUESTION 7 (1 point)
If an infectious disease has a reproductive number of 4, what proportion of the population would we need to vaccinate to achieve the herd immunity threshold? (Select 1 answer only)
c) 75%
QUESTION 8 (1 point)
The defining characteristic of an active surveillance programme is that: (Select 1 answer only)
e) It can be used to track risk factors as well as infectious and chronic diseases
QUESTION 9 (1 point)
Communicable disease control measures are generally not aimed at which of the following? (Select 1 answer only)
c) Eliminating the host
QUESTION 10 (1 point)
A food poisoning outbreak from some contaminated seafood at a restaurant is an example of what type of epidemic? (Select 1 answer only)
b) Point source
QUESTION 11 (1 point)
When a screening test simply increases the period of time between detection (through screening) and usual clinical diagnosis, without improving survival time, what type of bias does this represent? (Select 1 answer only)
b) Length-time bias
QUESTION 12 (1 point)
You wish to evaluate a new screening test that has become available for an infectious disease that results in chronic health problems if left undiagnosed. Results from your evaluation found that of the 1448 people with a positive test result, 797 were subsequently confirmed to have the infectious disease. Among the 10,892 people with a negative test result 129 were later diagnosed with the infectious disease.
What is the sensitivity of this test? (Please round your answer to 1 decimal place)
Positive Test Group:
Total number of cases = 1448
Let the number of cases confirmed with disease in the positive test group be ‘a’, and the number of cases tested negative for the disease be ‘b’.
Negative Test Group:
Total number of cases = 10,892
Let the number of cases confirmed with disease in the negative test group be ‘c’, and the number of cases tested negative for the disease be‘d’.
Thus, we have:
a = 797, b = 651, c = 129, d = 10,763
Sensitivity = a/ (a + b) = 797/ (797 + 651) = 797/1448 = 0.6
QUESTION 13 (1 point)
What is the specificity of the test? (Please round your answer to 1 decimal place)
Specificity = d/(c + d) = 10,763/ (129 + 10,763) = 10,763/10,892 = 1.
QUESTION 14 (3 points)
Five years after the successful implementation of the screening test mentioned in Question 13, early intervention had reduced the spread of infection and the prevalence of the disease has reduced to 1.2%. Assuming the sensitivity and specificity of the test remains the same, what will happen to the positive predictive value of the test now (compared to five years ago)? (2-3 sentences – please justify your answer with the PPV calculations from both 5 years ago (in question 13) and now to receive full marks)
Positive predictive value = a/a + c = 797/797 + 129 = 797/926 = 0.9
The prevalence of the disease has reduced to 1.2% after 5 years.
Thus, Positive predictive value after 5 years = sensitivity*prevalence/(sensitivity*prevalence) + (1-specifity)*(1-prevalence) = (0.6 * 0.1) / (0.6 * 0.1) + (1-1) * (1-0.1) = 0.06 / 0.06 = 1.0
Positive predictive value before 5 years = 0.9
Positive predictive value after 5 years = 1.0
QUESTION 15 (1 point)
Based on your calculations of sensitivity, specificity and positive predictive value, would you recommend the implementation of the new screening program when the prevalence is 1.2%? Why/ why not? (2-3 sentence)
Ans) It can be seen that the positive predicted value had increased slightly by 0.1, indicating that the probability of the disease present after 5 years increased slightly thus proving that the screening program had not made much of a difference in reducing the spread of the infectious disease. Hence, I would not recommend its implementation.
QUESTION 16 (1 point)
Australian law states that all cigarettes sold in Australia need to be packaged without brand logos and related art work on the outside of the package (this is called plain packaging). This is an example of a: (select 1 answer only)
d) Mass prevention strategy
QUESTION 17 (1 point)
Obesity is a risk factor for developing breast cancer in post-menopausal women. If we develop a weight loss intervention in a group of women over the age of 50 to reduce their risk of developing breast cancer, this is what level of breast cancer prevention? (Select 1 answer only)
a) Primary prevention
QUESTION 18 (1 point)
Which two of the following methods cannot be used in the design phase of a study to prevent confounding from occurring in a study? (Select 2 answers)
c) Stratification
e) Multivariable modeling
QUESTION 19 (4 points)
You are a policy advisor in the Department of Health in a country where road traffic accidents have been identified as a significant cause of death amongst the population. As funding is limited, you need to determine which risk factors should be given priority for prevention strategies to reduce the public health burden of road traffic accidents.
You identify a case-control study that has been previously conducted in your country, which examines the risk factors for road traffic accidents. Based on the data provided (see Table below), you need to prioritize the risk factors in order of which should be given funding priority to make the biggest impact on the public health burden of road traffic accidents (where 1 is the risk factor with the biggest impact, and 3 is the risk factor with the smallest impact). You will need to estimate the proportion of road accidents in the population attributable to each of the risk factors to justify your answer.
(NOTE: For full marks in your answer please provide: 1. The order of how you would prioritize risk factors for funding prevention interventions and 2. The calculations you used to come up with this order.)
It can be seen that the impact of not wearing a seatbelt has the greatest risk factor, followed by speeding and drink driving. I would give the greatest funding priority to promote wearing seatbelts, in order to reduce road traffic accidents.
QUESTION 14 (3 points)
Five years after the successful implementation of the screening test mentioned in Question 13, early intervention had reduced the spread of infection and the prevalence of the disease has reduced to 1.2%. Assuming the sensitivity and specificity of the test remains the same, what will happen to the positive predictive value of the test now (compared to five years ago)? (2-3 sentences – please justify your answer with the PPV calculations from both 5 years ago (in question 13) and now to receive full marks)
Positive predictive value = a/a + c = 797/797 + 129 = 797/926 = 0.9
The valence of the disease has reduced to 1.2% after 5 years.
Thus, Positive predictive value after 5 years = sensitivity*prevalence/(sensitivity*prevalence) + (1-specifity)*(1-prevalence) =
Positive predictive value before 5 years = 0.9
Positive predictive value after 5 years = 1.0
0.6*0.1/ (0.6*0.1) + (1-1)*(1-0.1) = 0.06/0.06 = 1
QUESTION 15 (1 point)
Based on your calculations of sensitivity, specificity and positive predictive value, would you recommend the implementation of the new screening program when the prevalence is 1.2%? Why/why not? (2-3 sentence)
It can be seen that the positive predicted value had increased slightly by 0.1, indicating that the probability of the disease present after 5 years increased slightly thus proving that the screening program had not made much of a difference in reducing the spread of the infectious disease. Hence, I would not recommend its implication.
QUESTION 16 (1 point)
Australian law states that all cigarettes sold in Australia need to be packaged without brand logos and related art work on the outside of the package (this is called plain packaging). This is an example of a: (select 1 answer only)
a) Mass prevention strategy
QUESTION 17 (1 point)
Obesity is a risk factor for developing breast cancer in post-menopausal women. If we develop a weight loss intervention in a group of women over the age of 50 to reduce their risk of developing breast cancer, this is what level of breast cancer prevention? (select 1 answer only)
Primary prevention
QUESTION 18 (1 point)
Which two of the following methods cannot be used in the design phase of a study to prevent confounding from occurring in a study? (select 2 answers)
QUESTION 19 (4 points)
You are a policy advisor in the Department of Health in a country where road traffic accidents have been identified as a significant cause of death amongst the population. As funding is limited, you need to determine which risk factors should be given priority for prevention strategies to reduce the public health burden of road traffic accidents.
You identify a case-control study that has been previously conducted in your country, which examines the risk factors for road traffic accidents. Based on the data provided (see Table below), you need to prioritise the risk factors in order of which should be given funding priority to make the biggest impact on the public health burden of road traffic accidents (where 1 is the risk factor with the biggest impact, and 3 is the risk factor with the smallest impact). You will need to estimate the proportion of road accidents in the population attributable to each of the risk factors to justify your answer.
(NOTE: For full marks in your answer please provide: 1. The order of how you would prioritise risk factors for funding prevention interventions and 2. The calculations you used to come up with this order.)
It can be seen that the impact of not wearing a seatbelt has the greatest risk factor, followed by speeding and drink driving. I would give the greatest funding priority to promote wearing seatbelts, in order to reduce road traffic accidents.
