Generating Mortality Rate by Year in a Panel Data Set

[Rushambwat]

Greetings Stata Users, I trust you are all keeping well.

Guys am working on a project in which I have a very large Panel Dataset with approximately 11 million observations and am trying to develop a panel data model for estimating the relationship between Mortality and Education and a list of other predicting variables. I want to use mortality as the outcome variable.

And there lies my challenge, I have a few time related variables, Year (just the Year), Date of Birth (DoB), Date of Death (DoD), number of people dead (Died), Gender (Sex) and have uploaded a subset of the data here.

All my code options have not worked and hence I have not uploaded them here.


I need your assistance with possible strategies I can employ to generate the Mortality Rate variable per 1000 people

Your assistance would be greatly appreciated.

[Carlo Lazzaro]

Rushambwat:
please do not upload spreadsheets. Due to the risk of active contents nobody I know on this list will ever download them (see the FAQ about -dataex- instead). Thanks.
See -xtpoisson-.

[Rushambwat]

Hi Carlo, my apologies I didn't know that about Excel files. I have tried to edit the post to remove the Excel file and replace with a Stata Data file which i have upload here. Please do let me know how i can remove the excel file. But i also still need help with the predicament i'm facing above .

[Carlo Lazzaro]

Rushambwat:
you probably need to consider survival analysis:

. use "C:\Users\carlo\Downloads\Subset Data Stata (2).dta" 

. encode Died, g(death)
. encode Sex, g(gender)

. stset Year, failure(death==1)

Survival-time data settings

         Failure event: death==1
Observed time interval: (0, Year]
     Exit on or before: failure

--------------------------------------------------------------------------
      1,000  total observations
          0  exclusions
--------------------------------------------------------------------------
      1,000  observations remaining, representing
        998  failures in single-record/single-failure data
  1,991,883  total analysis time at risk and under observation
                                                At risk from t =         0
                                     Earliest observed entry t =         0
                                          Last observed exit t =     1,992

. strate gender , per(1000) 

        Failure _d: death==1
  Analysis time _t: Year

Estimated failure rates
Number of records = 1000

  +-------------------------------------------------------+
  | gender     D          Y      Rate     Lower     Upper |
  |-------------------------------------------------------|
  | Female   522    1.0e+03   0.50205   0.46077   0.54701 |
  |   Male   476   952.1360   0.49993   0.45698   0.54692 |
  +-------------------------------------------------------+
   Notes: Rate = D/Y = failures/person-time (per 1000).
          Lower and Upper are bounds of 95% confidence intervals



.

For the future, please use -daatex- to share ab excerpt of your dataset (as per FAQ). Thanks.

[Rushambwat]

Dear Carlo L, I can't thank you enough.

This is insightful, after I saw the reply, I went and studied a number a books including Hamilton Data Analysis with Stata 12 and Handbook of Statistical Analyses with Stata/R and did quite some reading and truly, this was the solution to my predicament.

I am implementing it and learning further


Thank you so much.