UNIVERSITY OF NAIROBI

SPECIAL/ SUPPLEMENTARY EXAMINATIONS 2023/2024

FOURTH YEAR EXAMINATIONS FOR THE DEGREE OF BACHELOR OF SCIENCE

SAC 402/STA 437: SURVIVAL ANALYSIS

TIME: 11.30 am-1.30pm

DATE: AUGUST 23rd , 2024

INSTRUCTIONS

Answer Question ONE and any other TWO questions. You are strongly advised NOT to answer more
than three questions.

Question 1 (30 Marks)

a)In an experiment individuals had to solve a difficult task. Some of the individuals did not want to complete
the task and stopped early. Others did not complete the task by the end of the experiment.
The experiment lasted for two hours and the time( in minutes ) until finishing were recorded

50 51 66∗ 82 92 120∗ 120∗ 120∗

i) Compute the survival probability S(t) and sketch

ii) Compute the cumulative hazard function and sketch

iii) Estimate S(t) for t = 60minutes

(8 Marks)

b) The population of elderly people in prison is observed during the period 01/01/1994 to 31/12/1996. The
duration of residence (measured to the nearest number of months) is recorded for those who die during the period,
for those released from the prison during the period and for those who are still in residence on 31/12/1996. The
data recorded in months are

6∗ , 6, 6, 6, 7, 9∗ , 10∗ , 10, 11∗ , 13, 16, 17∗ , 20, 23∗

i) State the type(s) of censoring present in this data

ii) Calculate the Kaplan- Meier(Product limit) estimate of survival function S(t)

(7 Marks)
 c) For a mortality study with right censored data, you are given

Time No. of death Number at risk

5 2 15

7 1 12

10 1 10

12 2 6

i Obtain the estimate of the cumulative hazard function

ii Calculate S(t) at t = 9 using the Nelson Aalen estimate

(8 Marks)
d)Suppose at a University over a period of several years, 44 faculty members (20 females and 24 males ) were
recruited below the rank of full professor, and the accompanying table gives their waiting period (in years ) for
promotion to the full professor rank.
Test the hypothesis that there is no sex difference in the waiting time for promotion to full professor rank.
Females Males
6 6 6 6+ 8 5 5 5 5 5 5

9+ 10+ 11 12+ 13 5 5 5+ 5+ 6 6

16 17 18+ 19 20+ 8 8 8 8 11 11

22 23 25+ 32+ 32+ 12 12+ 15 17 22+ 23+

(7 Marks)

Question 2(20 Marks)

a) The table below gives the data for a small sample of employees in a factory. It shows the time in months
until the first absence from work.
Observations marked + show the time of leaving for those employees who left employment without being
absent from work. Use Cox proportional hazard model

Male employees: 6+ 11 13+ 15 16+ 19+ 20

Female employees: 1 4+ 7 8+ 10+ 12+ 17 21+

i) to determine the partial log likelihood

ii) to calculate the maximum partial likelihood estimate of the regression parameter

(7 Marks)

b) Let
X1 = Age − 50
X2 = Sex(male = 0, female = 1)
X3 = Smoking status(Non-smoker = 0, Smoker = 1)
X4 = Sex and Smoking status = X2 X3

2
 The CPHM is given by

h(t) = h0 (t)exp[β1 X1 + β2 X2 + β3 X3 + β4 X2 X3 ]
h(t) = h0 (t)exp[0.01X1 − 0.4X2 + 0.5X3 − 0.25X2 X3 ]

Find the relative risks holding X1 constant

hF S(t) hF N S(t) hM S(t)
i) ii) iii)
hM N S(t) hM N S(t) hM N S(t)
where

FS=Female smoker FNS=Female non- smoker

MS=Male smoker MNS=Male non- smoker

(8 Marks)

c) What is the survival function and hazard function of an exponential random variable

(5 Marks)

Question 3(20 Marks)

a) Suppose 10 students were admitted in a given year to a particular post-graduate program of a university.
This program allows the students to elect one of the two possible tracks (specialization). If time (in years)
taken by students to complete is as follows Use the log -rank test to compare the two tracts. Is there a

Track A: 3.5, 5.0+, 6.5, 12.25,

Track B: 4.25+, 6.5, 8.5+, 10+, 10, 12

difference between the two tracks with respect to the time required for the completion of the program?

(6 Marks)

b) The following data sets refer to the survival times, in years, of parents in a certain clinical trial. An asterisk
indicates that the observation was censored. It is assumed that censoring was random and that patients
were randomly assigned to the treatment or control groups.

Control Group : 3, 6∗ , 7, 8∗ , 8∗ , 8∗ , 9, 11, 13

Treatment Group : 4∗ , 5, 8∗ , 8∗ , 8∗ , 8∗ , 10, 12∗ , 14∗
Assuming a proportional hazard model,
I) Find the estimate of the log hazard ratio β using CPHM
ii) Determine β̂

(8 Marks)

c) In an experiment individual has to solve a difficult task. Some of the individuals did not want to complete
the task and stopped early. Others did not complete the task by the end of the experiment. The experiment
lasted for two hours and the time( in minutes ) until finishing were recorded

50 51 66∗ 82 92 120∗ 120∗ 120∗

Compute the survival probability S(t) at time t = 92

(6 Marks)

3
Question 4(20 Marks)
a) Individuals may enter a study at different times. Consider a hypothetical study of HIV positive sur-
vivals.Subjects were enrolled from 01/01/2000 to 01/01/2001. The study ended in 01/01/2002. After a con-
firmed diagnosis of HIV, subjects were followed untilAIDS related deaths. Of interest is the time from diagnosis
to death. The data is given in the following table

Subject Entry End Survival(months) Comment

1 01/01/2000 01/03/2001 14 died

2 01/02/2000 01/02/2001 12∗ lost

3 01/06/2000 01/01/2002 19∗ alive

4 01/09/2000 01/04/2001 7 dead

i) Calculate the survival probabilities and sketch the survival curve

ii) Compute the cumulative hazard function and sketch the curve

iii) Estimate S(t) at 17 months

(12 Marks)

b)A photocopier firm experiments with two types of bulbs to use in its machines. The data below on time in
months was collected from a pilot study the firm conducted. Use the Wilcoxon statistics to test if there is a

Group A: 1 2 3+ 4 4+ 7+ 8 9 11 15
Group B: 1+ 2 2 3 3 4 4 5+ 6 7+

difference in the survival time of bulbs in the two groups

(8 Marks)

4
 Question 5(20 Marks)
a) A group of individuals with sexually transmitted disease (STD), all initially HIV free was monitored for 12
years until HIV infection diagnosis. The number of those either diagnosed with HIV or censored are presented
in the table below.

Interval in years HIV infected Number censored

0-2 4 4

2-4 5 7

4-6 6 7

6-8 5 10

8–10 4 8

10-12 6 14

i) How many individuals were involved in the study

ii) Estimate S(t) and plot

iii) Estimate a 10 year survival rate from the data

(10 Marks)

b) Consider failure times in hours from two populations π1 and π2 as given below: Compare the survival curves

From π1 : 2, 3, 5, 5, 5∗ , 7, 8∗ , 8∗ , 10

From π2 : 2∗ , 4, 4, 7, 7∗ , 8 9 9 10∗ 11

from the two populations using appropriate test statistic.

(10 Marks)