Scribd
Upload
173 views5 pages

AMA 4424 Survival Analysis B

This document outlines the examination structure for the course AMA 4424: Survival Analysis at the Technical University of Mombasa, including instructions for candidates and a series of questions covering key concepts in survival analysis. The exam consists of five questions, with the first question being compulsory, and it tests knowledge on survival functions, hazard functions, Kaplan-Meier estimators, and Cox proportional hazard models, among other topics. Students are required to demonstrate their understanding of theoretical concepts and practical applications related to survival analysis.

Uploaded by

ellyotieno856
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
173 views5 pages

AMA 4424 Survival Analysis B

This document outlines the examination structure for the course AMA 4424: Survival Analysis at the Technical University of Mombasa, including instructions for candidates and a series of questions covering key concepts in survival analysis. The exam consists of five questions, with the first question being compulsory, and it tests knowledge on survival functions, hazard functions, Kaplan-Meier estimators, and Cox proportional hazard models, among other topics. Students are required to demonstrate their understanding of theoretical concepts and practical applications related to survival analysis.

Uploaded by

ellyotieno856
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 5

TECHNICAL UNIVERSITY OF MOMBASA

SCHOOL OF APPLIED AND HEALTH SCIENCES


DEPARTMENT OF MATHEMATICS AND PHYSICS
UNIVERSITY EXAMINATIONS FOR DEGREE OF B.Sc IN B.Sc
IN STATISTICS AND COMPUTER SCIENCE
AMA 4424: Survival Analysis
END OF SEMESTER EXAMINATION
SERIES: NOVEMBER 2024
TIME: 2 HOURS DATE: NOVEMBER 2024

Instructions to Candidates:
You should have the following for this examination: Answer Book-
let, examination pass and student ID. This paper consists of FIVE
questions Answer question ONE (COMPULSORY) and Any other
TWO Questions. The maximum marks for each question is shown.
Do not write on the question paper. Mathematical tables and calcu-
lators may be used

QUESTION ONE (30 MARKS)


a) Define the following terms in the context of survival analysis
i) Survival Function (1 mark)
ii) Hazard Function (1 mark)
iii) Censored Data (1 mark)
iv) Kaplan-Meier Estimator (1 mark)
v) Cox Proportional Hazard Model (1 mark)
b) Explain why Cox model is called a semiparametric model (2 marks)
c) Let T be a continuous random variable (survival time). Prove that
i) S(t) = e−H(t) (3 marks)
ii) f (t) = h(t)e−H(t) (1 mark)
where f (t), h(t), S(t) and H(t) are density function, hazard function, sur-
vival function and cumulative hazard function, respectively.
d) A lifetime distribution has hazard function

h(t) = ρθ /(ρt + 1)

where ρ > 0 and θ > 0. Find


i) the probability density function. (4 marks)
ii) the median. (2 marks)
e) Consider the following set of survival times (in months) for a group of pa-
tients: 1, 3+, 4, 4+, 6, 7, 8, 10+, 11, 12+, 14+
where (+) denotes censored data. Calculate the Kaplan-Meier estimator of
the survivorship function and sketch a graph of the estimated function. (5
marks)
f) The following survival data were collected from 5 patients, where Z is a
binary random variable. Each column represents a patient in the table
below.

Subject 1 2 3 4 5
Ti 13 11 18 3 8
δi 0 1 0 1 1
Zi 1 0 1 0 1

Suppose that you decide to use a Cox PH model for data analysis. Write
down the partial likelihood L(β) (5 marks)
g) Suppose the hazard function for individual i is
ρi
hi (t) =
1 + ρi t
Show that this constitutes an accelerated Failure Time model (3 marks)

QUESTION TWO (20 MARKS)

a) Suppose T is a discrete lifetime that comes from a geometric distribution;


f (t) = θ(1 − θ)t , t = 0, 1, 2, . . . Obtain;
i) The survival function (3 marks)
ii) The hazard function (2 marks)

Page 2
iii) The cumulative hazard function (2 marks)
b) Suppose survival times (in weeks) are:

10 13+ 18+ 19 23+ 30 36 38+ 54+


56+ 59 75 93 97 104+ 107 107+ 107+

where + denote censored times.


i) Construct the Kaplan-Meier estimates for the hazard function and the
survival function (11 marks)
ii) Make a plot of the survival function based on the estimates obtained in
part (i) above (2 marks)

QUESTION THREE (20 MARKS)

a) The following are remission times (in month) of a group of cancer patients
who received one of three treatments: Surgery (S), Chemotherapy (C) and
Radiotherapy (R).

Patient Survival time Outcome Treatment


number (months)
1 2 Cured R
2 4 Cured R
3 5 Unknown R
4 6 Unknown C
5 9 Unknown R
6 10 Cured S
7 12 Cured R
8 12 Cured C
9 15 Unknown R
10 18 Unknown C
11 22 Cured R
12 26 Cured S
13 30 Cured C
14 35 Unknown S
15 37 Cured S
16 55 Cured S
17 72 Cured C
18 80 Censored S

Page 3
Test whether all treatment are equally effective at 5% level of significance
Take (χ22,0.05 = 5.99 ) (14 marks)
b) In an experiment to evaluate the effect of diets on mammary carcinogenesis
risk of rates. A sample of rates were treated with the drug DMBA. Starting
6 weeks after DMBA administration, each rat was examined once weekly for
14 weeks. The survival time of concern is the time until a tumor is detected
after the DMBA administration.

i) What is the failure event for this study? (1 mark)


ii) What is the appropriate time scale for this study? (1 mark)

Describe in detail the types of censoring that are represented by the


following rats, and provide the survival time or censoring time data in
days:
iii) A rat that had a tumor at the first examination at 6 weeks. (1 mark)
iv) A rat that survived the study without having any tumors.(1 mark)
v) A rat that did not have tumor at week 12 but had tumor at week 13.(1
mark)
vi) A rat that died (without tumor present and death was unrelated to the
occurance of cancer) at day 37. (1 mark)

QUESTION FOUR (20 MARKS)

a) The survival time T follows the accelerated failure time (AFT) model

ln(T ) = β0 + β1 x + ϵ

The survival function of T is S(t) = exp −t/ eβ0 +β1 x

i) Find the density function of T (3 marks)


ii) Find the hazard function of T (2 marks)
iii) What is the median survival time (2 marks)
iv) Derive the time ratio at the median survival time for x = 1 and x = 0
(2 marks)

b) Consider the following sequence of survival times for a set of individuals


measured from birth: 10, 7, 32+, 23+ 11, 12+, 6+, 27, 30, 16
where numbers followed by a “+” indicate right-censored values.

Page 4
i) Calculate the Nelson-Aalen estimate of the cumulative hazard function
for these data, presenting the results in a table. (7 marks)
ii) Calculate a linear 95% confidence interval for H(15). (4 marks)

QUESTION FIVE (20 MARKS)


a) Briefly describe buckley-James method (3 marks)
b) Explain how the Cox-Snell residuals are used to assess the appropriateness
of the Cox proportional hazards model for these data. (4 marks)
c) To compare two treatments for solid tumors, two groups of mice were inoc-
ulated with a highly malignant neoplasm, and after randomization into the
treatments, survival (days) was observed as follows:

Drug A: 17, 21+, 23, 24+, 25, 32, 37+, 39


Drug B: 9, 10, 12+, 13, 15, 19+, 21, 28

where “ + ” indicates censoring


i) Compute the Kaplan-Meier product-limit estimate of each group and
plot the two survival curves in the same figure. K-H estimator is
k  
Y di
Ŝ(t) = 1− f or tk ≤ t ≤ tk+1 (8 marks)
i=1
ni

ii) Find the median survival time for each group. (2 marks)
iii) Compute the cumulative hazard at 20 days for each group. Which drug
is more effective? Why? (3 marks)

Page 5

You might also like