Reference no: EM132845399
Section A: Short questions
Question 1. Define or explain the following terms with the aid of an example involving brief calculations
i. Conditional probability
ii. Confounding
iii. Standardising to the Z - normal distributions
iv. Bayes theorem's use on diagnostic tests on populations
v. Distinguish Fisher's exact test and McNemar's test
vi. Confidence interval
vii. Continuity correction
viii. Poisson distribution
ix. Binomial distribution
x. Non parametric tests
Question 2. Suppose a hospital instrument goes an average of 5000 minutes between recharges, with a standard deviation of 150 minutes (i.e., µ = 5000 and σ = 150) .
a. What is the probability that the instrument will go for less than 4775 minutes without recharging?
b. What is the least duration over which the instrument will go more than 10.75% after recharging?
c. What is the interquartile range for instrument i.e lower quartile (q1) - upper quartile (q3)?
Question 3. Hb was measured among patients who presented with fever in a malaria endemic region. The researcher anticipated 2 units reduction among patients who were found to have malaria compared to those who did not have malaria. Using a two-tailed test at 5% level of significance.
i. How many patients should be enrolled in the study to ensure that the power of the test is 93.51% to detect this difference assuming the mean Hb of 14.4 units for those who did not have malaria and a standard deviation of 3.8?
ii. What would be the power of the study when 45 patients are enrolled in each arm and 3 units reduction are anticipated? Assume everything else remains the same.
iii. Explain why you got a greater power than 80% when you had a smaller sample of 45 persons in ii)?
Section B: Topical Questions
Question 4. Given the following table of a radionuclide ventriculography test for coronary heart disease, answer the questions that follow
|
Disease
(coronary artery disease)
|
Test Result
(radionuclide ventriculography)
|
Present
|
Absent
|
|
Positive
|
32
|
8
|
|
Negative
|
18
|
22
|
i. Work out the probability of P(Test Positive T+ or Disease positive D+)?
ii. What is specificity in terms of conditional probability terminology? Work out the specificity of this ventriculography test?
iii. Work out the predictive value of a negative test?
iv. For a population in which the probability of having coronary artery disease is 0.10, estimate the probability that an individual has the disease given that (s)he tests positive using radionuclide ventriculography.
Question 5. You are provided with STATA output Appendix A1, A2 and A3 based on a study carried out in Uganda. Use these outputs to answer the following questions.
a. What two assumptions were tested for, based on outputs A1 and A2?
b. For each of these two tests state the hypotheses and conclusions that can be deduced from the Stata outputs A1 and A2.
c. State the statistical test used, also write the null and alternate hypotheses based on the output A3?
d. Work out the student's t-statistic based on the A3 output (t-statistic)?
e. What conclusion do you draw from the results based on A3 output?
f. Perform the testing of significance again using the critical value and draw a conclusion. Compare this with the decision you made on e) above
Question 6. A student performs a oneway analysis of variance test on his Biostatistics 1 assignment, using the outcome total number of symptoms (totsym) and the exposure variable WHO stage at Baseline (whostbas). Use the given output in Appendix B1, B2 and B3.
a. Work out the values for p, q, r and s from B2 oneway anova Stata output? Utilise the summary statistics provided on B1.
b. State only the hypotheses for the oneway anova and two crucial assumptions for this test? What conclusions do you draw for each of these hypothesis based on the given Stata output?
c. What is the main purpose of the Bonferroni's test and what results did this yield using the given output on B3?
d. What is the non-parametric alternative to the oneway anova?
Question 7. An upper lung cancer is investigated in a case control study. Given that 200 mothers of cases and 400 controls, one potential risk factor is whether the father smokes or not. Given the Stata output in Appendix C1 and C2, answer the questions that follow.
a. Based on the output C1 the Pearson's Chi-Square test was used, state the null and alternate hypotheses.
b. Work out the Pearson's Chi-Square test statistic (test statistic) on output C1?
The Stata output C2 investigates confounding, using these output results;
c. State clearly the exposure, confounding and outcome variables
d. Distinguish between the crude and adjusted odds ratios
e. What are the null hypotheses and results corresponding to these two outputs from the results.What conclusion can you draw from these analysis?
i. Test of homogeneity (M-H) chi2(1) = 0.00 Pr>chi2 = 1.0000
ii. Test that combined OR = 1: Mantel-Haenszel chi2(1) = 0.00 Pr>chi2 = 1.0000
f. Was there confounding, effect modification or neither, support your answer?
Attachment:- ventriculography.rar