Thank you for your interest in this work. I have an easy answer for your second question: for anything that is a “select multiple” question (see the “type” column of the “survey” sheet in the xls form), you can encode multiple selections as a space-separated list, i.e “1 2” for cigarettes and pipe.
Here is where to look to find out if a question is a “select multiple” in PHMRC_Shortened_Instrument_8_20_2015.xls for when this comes up with other questions:
(This already came up with a different group mapping injuries, see here if you want to know more: https://healthyalgorithms.com/2016/01/28/injuries-query-on-mapping-data-for-smartva-analyze-1-1/ )
In order to answer your first question, I think I need more information. Can you give me an example of what you are trying to do? It also be helpful if you can try it, and let me know what goes wrong if it does not work.
In [previous post] …
From: Abraham D. Flaxman
Sent: Wednesday, December 30, 2015 5:16 PM
Subject: RE: VA Data Update
Cool, it looks like you are making progress. I do again encourage to do this work incrementally. So make a mapping that just gets the age and sex into the right columns and run that through, and then add in a few questions at a time to make sure things keep changing in a way that makes sense (i.e. when you add the column on chest pain to your mapping, the number of heart attack deaths should increase…).
You had a question about injury coding in your spreadsheet. Here is the coding:
1. Road traffic crash/ injury
5. Bite or sting by venomous animal
7. Violence (suicide, homicide, abuse)
11. Other injury, specify (__________)
8. Refused to answer
9. Don’t know
If multiple injury causes were endorsed, you may record them as a space-separated list, i.e “2 3” for a fall that resulted in drowning.
My new paper, which I believe is the first published application of DisMod-PDE: Projected growth of the adult congenital heart disease population in the United States to 2050: an integrative systems modeling approach
I have a new verbal autopsy paper out, which goes deep into the weeds of how to measure the accuracy of a method for identifying the underlying cause of death at the population level from survey data. http://www.pophealthmetrics.com/content/13/1/28
A fun online appendix includes probabilistic combinatorial calculations of the sort that I was actually trained in: http://www.pophealthmetrics.com/content/supplementary/s12963-015-0061-1-s2.zip
Finally got one!
I have a real printed copy of the DisMod book now, and if I have understood correctly, it is really available for purchase starting today!
I meant to have a picture of it in my hands to go here, but I didn’t get it in time. Here is a multiscale rainbow smudge instead: