ADVANCED THEORY OF BIOSTATISTICS — DU B.Sc (hons) Statistics Previous Year Question Papers

Biostatistics (DSE) Please mention null and alternativc hypotheses wherever applicable. (a) Explain a prospectrve study with a suitable example. Write one merit and one demerit of this study design. (b) What is blinding? Explain single, double and triple blinding. (c) Explain Likelihood ratio measure of the test. Compute LR+ s1d LR- from the data given { __!{yqcqdial infarction Present Abs€nt (d) Discuss log rank test to compare two survival distributions. (e) Discuss Mantel - Haenszel tDst to cDnpLre two survival distributions. How this test is different from other tests. (f) How do you compare the two Weibull distributions based on the two sample test proposed by Thomas and Bain? (5x5=25) (a) Ten patients with cancer are randomized to receive two different treatments. The following remission times (in months) are recorded: Test whether the two treatments are equally effective using the Gehan's Generalized Wilcoxon test at 5% level of significance. Also explain the steps of the test clearly. (b) Discuss the Iog likelihood ratio test for comparing two survival distributions with p parameters when all the parameters of the distribution are unknown. In particular discuss the test when we compare two Weibull distributions with parameters (Ip vr) and (1"r, vr) respectivelY where I,, 't, and ?"r, v, are unknown. (a) Discuss the likelihood ratio test for comparing the two exponential distributions with parameters )'' and f,, respectively. (b) Ten patients were randomized to receive either treatment I or treatment 2. The remission times of the patients (in weeks) was obtained as follows: .It was assumed that the remission time follows ' exponential distributions with parameters 1', and ' ),,, respectively under the two treatments' Test whether the two treatments are significantly different using likelihood ratio test. (Use 12r,0.0, = 4 (a) Discuss the Cox-Mantel test for comparison of two survival dr stributions. (b) Ten TB patients were randomly allocated treatment I or treatment 2. The following times to relapse (in months) were recorded: Apply Cox-Mantel test to check whether the treatment I is more effective than treatment 2 at 5%o tevel of significance. \6,7) Define a simple stochastic epidemic model. Suppose there are n susceptible and one infective at time t = 0. Derive the expression for the probability that r individuals remain susceptible at time t. (13) (a) Define Sensitivity and Specificity of the test. Using the data provided below, calculate the sensitivity and specificity of the ECG test performed on 700 subjects suffering from prolonged acute chest pain. Interpret the results in terms of the diagnostic accuracy of the ECG test. F]CC Myocardial infarction Prcsent Absent (b) Using the data given below, calculate the Relative Risk and apply the Z-statistic and Chi-Square statistic to test whether the antecedent (elevated LDL levels) has a significant effect on the outcome (cardiovascular disease). (Use 12,,0.0, = 3.84)