Adjusting confounders in ITSA (Single-group)

[Abinash Singh]

Dear all,
I am using Single-group ITSA (Linden, 2015), for measuring the population-level impact of a health intervention on child mortality. However, i am confused with the followings:

itsa depvar [indepvars] [if] [in] [weight], trperiod(numlist) [single treatid(#) contid(numlist) prais lag(#) figure posttrend replace prefix(string) model options]

1) In the above command specification by Ariel Linden, can I use confounding variables (GDP per capita, health expenditure, literacy rate, population) in place of [indepvars] to adjust the model for socioeconomic and demographic factors?
2) Since my dependent variable is a count variable (child mortality) and independent/ confounders which i want to adjust are continuous variables, in this case, do I need to check the stationarity of the variables and take the difference value if found non-stationary?
3) The ITSA package provides two models 'newey' and 'prais', which one should i prefer ideally (all data are annual, hence free from seasonality problem)?

I would be really grateful if I can get suggestions on the above issues.
Thanks
--
Abinash S.

[Abinash Singh]

Dear all,
I am using Single-group ITSA (Linden, 2015), for measuring the population-level impact of a health intervention on child mortality. However, i am confused with the followings:

itsa depvar [indepvars] [if] [in] [weight], trperiod(numlist) [single treatid(#) contid(numlist) prais lag(#) figure posttrend replace prefix(string) model options]

1) In the above command specification by Ariel Linden, can I use confounding variables (GDP per capita, health expenditure, literacy rate, population) in place of [indepvars] to adjust the model for socioeconomic and demographic factors?
2) Since my dependent variable is a count variable (child mortality) and independent/ confounders which i want to adjust are continuous variables, in this case, do I need to check the stationarity of the variables and take the difference value if found non-stationary?
3) The ITSA package provides two models 'newey' and 'prais', which one should i prefer ideally (all data are annual, hence free from seasonality problem)?

I would be really grateful if I can get suggestions on the above issues.
Thanks
--
Abinash S.

Prof Clyde Schechter, I will be glad if you can provide your valuable insights.
Thank you.

[Clyde Schechter]

Sorry, but I am not familiar with the -itsa- command and I have no advice to offer you on this.

[Jared Greathouse]

Abinash Singh Hey, you can just put your covariates in the [indepvars] part of the command.


Additionally, since your outcome is a count, you CAN use GLM modeling within the framework of ITS, BUT BE WARNED, you'll need to construct the appropriate intervention variable and interaction terms (not too big a problem since it's single group ITS). You'll also need to implement newey SEs correctly, by adjusting the variance inflation factor.

While we're talking, why can't you do a comparative ITS?