ECON 317 SPRING 2020:
DUE MARCH 17th, 2019 BY 11:59 PM
Last 3 digits of SID
Honor Code: I guarantee that all the answers in this assignment, except those for
a question specifically marked as a group discussion question, are entirely my own
work. I have cited any outside sources that I used to create these answers in APA
Question Mark Out of
1 a 12
IF YOU HAVE ALREADY COMPLETED FOUR ASSIGNMENTS, YOU
DON’T HAVE TO WRITE THIS ONE! (BUT YOU MAY, IF YOU
WISH). ONLY YOUR HIGHEST FOUR ASSIGNMENT MARKS
1. [Critique] Read the following article:
Mortaz, S., Wessman, C., Duncan, R., Gray, R., & Badawi, A. (2012). Impact of screening and
early detection of impaired fasting glucose tolerance and type 2 diabetes in Canada: a Markov
model simulation. ClinicoEconomics and outcomes research : CEOR, 4, 91–97. Retrieved from
(The link given should work off-campus.)
While reading this paper, your focus should be on understanding how decision trees and Markov
models can help in economic evaluation of health care treatments.
a. (12 marks) Write a 3-2-1 report in the usual fashion.
2. [Markov Models] Consider the following situation:
There are three possible states: healthy, sick and dead.
Healthy people can get sick, and sick people can recover and become healthy.
Healthy people can die, and sick people can die. Dead people stay dead.
Each ‘cycle’ is one day long.
If you are healthy today, there is an 80% chance that you will be healthy tomorrow, and
a 5% chance that you die. If you do not stay healthy or die, you become sick.
If you are sick today, there is a 60% chance that you will be sick tomorrow, and a 10%
chance that you will die. If you do not stay sick or die, you recover (become healthy).
In ‘Cycle 0’, there are 1,000 healthy people, 800 sick people and 0 dead people.
a. (6 marks) Draw a Markov diagram for the situation above (hint: it’s the diagram with the
circles and arrows).
b. (6 marks) Calculate the number of healthy people, sick people and dead people in Cycle 2. If
necessary, round your answers to the nearest whole number. Show your work. (Hint: Cycle 2 is
2 cycles after Cycle 0. If Cycle 0 is today, then Cycle 1 is tomorrow, and Cycle 2 is the day after
Healthy People in Cycle 2: ____________
Sick People in Cycle 2: ___________
Dead People in Cycle 2: __________
3. [Decision Trees] Consider the following situation:
The QALY weight for being dead is 0. The QALY weight for being perfectly healthy is 1.
A hospital has to decide whether to treat a patient for a difficult-to-diagnose disease.
There is a 70% chance that the patient is perfectly healthy, and a 30% chance that the
patient has a fatal illness that will kill them within the next hour.
A treatment exists for the illness, but the treatment leaves any patient who takes it with
permanent digestive problems. Living with these digestive problems has a QALY weight
If the patient is sick, there is an 80% chance that the treatment works, leaving the patient
alive but with the permanent digestive problems mentioned above, and a 20% chance
that the treatment fails, and the patient dies anyway.
If a healthy patient takes the treatment, they live, but suffer from the permanent
digestive problems mentioned above.
Simplifying assumptions: If the patient survives (is not killed in) the next hour, they will
live for exactly 20 more years and then die (whether they are perfectly healthy or have
digestive problems). There is no time discounting of QALY.
a. (6 marks) Draw a decision tree for the situation above using standard node symbols (squares,
circles, triangles). Label all relevant probabilities and outcomes on your Decision Tree:
b. (6 marks) Use your decision tree to calculate the expected QALY from administering the
treatment to the patient, and from NOT administering the treatment to the patient. Show your
work. Should the hospital give the treatment to the patient? (You may assume the treatment is
free or already paid for, so you only need to consider expected QALY.)
Expected QALY if the patient is treated: ___________
Expected QALY if the patient is NOT treated: ___________
Should the hospital treat the patient? Yes / No
4. [Rates and Probabilities] Suppose the three-year probability of getting the flu, if you are
unvaccinated, is 65.7%, and that the incidence rate (1-year rate) is constant over those three
a. (3 marks) Calculate the 1-year incidence rate using the appropriate formula from Lecture 21.
(That formula is r = -ln(1 – p)/t ). Show your work. Have your answer correct to four decimal
places (e.g. 0.1234). This is to make sure your answer to part b. is accurate.
b. (3 marks) Calculate the 1-year probability of getting the flu, using the rate you calculated in
part a. and the appropriate formula from Lecture 21. (That formula is p = 1 – e-rt , which is just
the formula mentioned in part a, rearranged to have p on the left.) Show your work. Have your
probability correct to two decimal places (e.g. 33.33%).
Hint: You can double-check your work using standard probability math. If the one-year probability
of getting sick is S, then the probability of getting sick within a three year window is the probability
of getting sick in Year 1, S, plus the probability you didn’t get sick in year 1, but did get sick in year
2, (1 – S) x S, plus the probability you didn’t get sick in years 1 or 2, but did get sick in Year 3, (1 –
S) x (1 – S) x S.
Probability: ______________ %
5. [Research Methods] In order for a journal article to be a cost-effectiveness study about a
drug that is used in a specific treatment, the following conditions must be satisfied:
It must be a study presenting original results, not a study of studies (‘book report’, such
as a systematic review, review of the literature, meta-analysis, etc.).
It must refer to that specific drug and treatment (some drugs can be used in various
treatments – for example, aspirin can be used as a painkiller or a blood thinner).
It must present, as one of its main results, data on the cost of the treatment.
It must present, as one of its main results, data on the effectiveness of the treatment.
Consider the following three articles, all submitted as part of a Group Assignment 21:
Levy, P., Lechat, P., Leizorovicz, A. & Levy, E. (1998). A Cost-Minimization of Heart Failure
Therapy with Bisoprolol in the French Setting: An Analysis from CIBIS Trial Data. Cardiovascular
Drugs and Therapy, 12, 301-305. Retrieved from https://doi-
Gilbert, E. M. (2002). Cost-Effectiveness of Beta-Blocker Treatment in Heart Failure. Reviews in
Cardiovascular Medicine, 3(S3), 42-47. Retrieved from
Ekman, M., Zethraeus, N. & Jönsson, B. (2012). Cost Effectiveness of Bisoprolol in the
Treatment of Chronic Congestive Heart Failure in Sweden. Pharmacoeconomics, 19(90), 901-
916. Retrieved from https://doi-org.ezproxy.library.uvic.ca/10.2165/00019053-200119090-
a. (2 marks) Which of the three articles is a cost-effectiveness study of the drug Bisoprolol as a
treatment for Heart Failure? (There’s only one.)
b. (2 marks) In what units is cost measured in the main result of the cost-effectiveness study?
(Hint: You only need to check the Abstract.)
1 This is not intended to single anyone out. The group in question actually ended up doing reasonably well on the
assignment, since this was only part of their submission. This question attempts to address some very common
issues I came across while marking the second group assignment.
c. (2 marks) In what units is effectiveness of Treatment measured in the main result of the cost-
effectiveness study? (Hint: You only need to check the Abstract.)
d. (6 marks) Why are the two other articles NOT cost-effectiveness studies? What points in the
bullet-pointed checklist of what must be present in a cost-effectiveness study do they not
Other Article 1:
Other Article 2:
- Name or Signature for Honor Code: ______________________________________________
The table below is for TA use only.