I thought Morgan Levine's group had a good idea. They did a study to take test results available from the NHANS data sets, https://wwwn.cdc.gov...eBeginYear=1999, and do a proportional hazard model to explain all cause mortality. I have just started to digest the paper. I am having problems reproducing their numbers, or making sense of their particular model.
Since cardiovascular disease is the leading cause of death in the US, I expected HDL, LDL and triglycerides to be in their final equation. None of these variables were there. Perhaps there was a high correlation with blood glucose and C reactive protein, or other variables?
I attached a png of their final equation.
I took my most recent lab results, and plugged them into their equation, and I calculated that I have the phenotypic age of 149 years.
I attached the python code, and their original paper.
I went to the NHANS site, and tried to download their data. It is in a strange format. It has an XPT file extension. I don't know how to read this data. I don't see any mortality data on their page.
This is the title of their paper: "A new aging measure captures morbidity and mortality risk across diverse subpopulations from NHANES IV: A cohort study"
Does anyone have any thoughts on how to use the formula, or why my calculation is so off?
Does anyone see how they did a proportional risk model from the NHANES data?
Attached Thumbnails
Attached Files
-
pheno_age.txt 2.75KB
229 downloads
