Using DALY in modelling

Asked: 31 Jul 2018 | 1711

Hello there, I am Amanuel from Ethiopia. I am conducting a thesis paper in cost effectiveness of HIV testing modalities and assisting others in economic evaluation. In my work, we conduct our research using decision tree modelling and Markov modelling. How should DALY be used for decision tree modelling? How do we calculate total DALY from Markov modelling? Thanks a lot

Modelling Utility measure

Expert Replies:

Mark Jit

Professor | United Kingdom | Replied: 22 Aug 2018 at 01:02

Dear Amanuel,
Apologies; I only noticed that you had asked a follow up question today.
For premature death, you would simply calculate the number of years of life lost by the person, which would be the different between his/her age and the standard life expectancy at that age.
You should use a disability weight of 1 for each year of life lost when calculating DALYs (however if you are calculating QALYs then you should ideally weigh the years of life lost according to the population average quality of life).
Mark

Mark Jit

Professor | United Kingdom | Replied: 03 Aug 2018 at 22:33

Dear Amanuel,
In a decision tree, the end node and possibly some of the chance nodes represent particular health states. Each of them may be associated with incurring certain DALYs, so you would multiply the probability of reaching that node with the DALY associated with it. Eg. there may be a node representing hospitalisation with pneumonia, which you would multiply with the DALY associated with that health state.
In a Markov model, the states represent health states rather than health conditions, so you would multiply the time spent in a certain state with the disability weight associated with that state. For instance, there may be a state representing having chronic hepatitis B; you would multiply the disability weight for that state with the length of time spent there (i.e. the sum of the probability at each time step of being in that state over the time horizon of the model).
Mark

Amanuel Yigezu

Health Economist | Ethiopia | Replied: 11 Aug 2018 at 13:56

Dear Mark,
Thank you for your reply. what about if that person dies before expected life expectancy. for this I am not going to use the disability weight. how am I going to use the YLL in Markov modelling while I am using Treeage software? should I use disability weight of 1 for death at that health state? Thanks a lot.