Assignment: Evaluating Cost-effectiveness using decision trees: The HPV case
The goal of this assignment is to provide students with applied practical experience using decision trees to evaluate the cost-effectiveness of alternative decisions, in this case providing HPV vaccination within the case study context versus not providing universal vaccination.
The decision trees included in this homework assignment are not intended to provide a realistic evaluation of the likely Cost-effectiveness of each alternative given the limited nature of decision trees to evaluate public health promotion interventions in open populations. This becomes more complicated in the case of HPV vaccination, which also has a dynamic nature due to the potential herd immunity against HPV infection.
However, the exercises below provide a good entry point to understanding the calculations used to develop more complicated Markov cohort a microsimulation models and are considered a foundational aspect of introductory economic evaluation courses.
Students will be provided with a list of variables needed to build an effective decision tree analysis aligned with the content covered in Unit 6.12. We will not be applying discounting in these examples in order to simplify the exercise.
However, the exercises below provide a good entry point to understanding the calculations used to develop more complicated Markov cohort a microsimulation models and are considered a foundational aspect of introductory economic evaluation courses.
Students will be provided with a list of variables needed to build an effective decision tree analysis aligned with the content covered in Unit 6.12. This unit should explain everything that you need to know to complete this homework assignment. We will also be reviewing the homework in class before the final version is due. We will not be applying discounting in these examples in order to simplify the exercise.
Remember that we can measure average incremental cost-effectiveness ratio (ICER) of an intervention from the perspective of an individual or for a cohort, the results should be the same. The only difference between running a population through the tree and running the "average" person is that we can't estimate the magnitude of the population effects with a single average person.
Background:
|
Model Parameters
|
Estimate
|
|
Lifetime Incidence of cervical cancer
|
36 cases per 100000 women, or 0.00036 per person
|
|
% of cervical cancer associated with HPV type 16/18
|
70%
|
|
% local cancer
|
25%
|
|
% regional/distant cancer
|
75%
|
|
Probability of receiving 1 dose (incomplete)
|
80%
|
|
Probability of receiving two dose fall series
|
75%
|
|
Probability of individual reduction with full series
|
100%
|
|
Probability of individual protection with incomplete series
|
94%
|
|
Duration of protection
|
100 years
|
|
Life expectancy 80 years
|
80 years
|
|
Age of cervical cancer onset
|
50 years
|
|
Local cancer disability weight (percent healthy time loss)
|
0.08
|
|
Regional cancer disability weight (percent healthy time loss)
|
0.75
|
|
Average five year survival (local cancer)
|
81%
|
|
Average five year survival (regional cancer)
|
23%
|
|
Average cost per woman per local cancer treated
|
US$ 6321
|
|
Average cost per woman per regional cancer case treated
|
US$ 8709
|
|
Cost per fully vaccinated girl
|
US$ 20
|
|
Cost per partially vaccinated girl
|
US$ 16
|
Assume that every girl not vaccinated at age 11 is at 100% risk of developing persistent HPV infection. Ignore any herd immunity effects. Assume that any woman with cervical cancer who has passed the five year survival mark is no longer at risk of early death but retains their disability weight for the rest of their lives. For those developing cervical cancer at age 50, their disability begins immediately. Assume that regardless of vaccination status all other sources of disability and life-expectancy are equal across the alternative arms.
Deliverables
1. Completed decision tree with calculated cost-effectiveness ratio in terms of cost per DALY averted. A table with expected values in terms of costs and DALYS averted and probability for each branch of the decision tree should be submitted separately.
. All probabilities and expected values (in terms of both costs and DALYs) should be included in the submitted decision tree as partial credit will be given for completion of part of the tree.
. You may choose to complete this in a specialized software program (such as TreeAge), in a word processing or powerpoint file, or hand-drawn and scanned as long as it is legible
2. 250 word commentary on the limitations of using the decision tree in this case as well as the broader limitations of decisions trees.
Download:-Decision tree.rar