Shalini wrote:
Dear all,
I have a question regarding statistical models. I have a cross-sectional
data
with obs on people's use of health services (type of treatment, type
of services (free or on payment basis) expenditure on various heads,
source of funding illness episode, selection of provider etc), and
out-of-pocket expenditure. If a person didn't get free public
hospital, there is an observable out of pocket expenditure; if not, it
is 0.
Now I want to study two things from this data
1) The components of out-of-pocket expenditure, but there are lot of
missing information, as people tend to report the total expenditure on
last illness episode instead of the decomposition into expenditure on
drugs, diagnostics, consultations etc.
Dear Shalini,
as far as managing missing data in health economic evaluations is concerned,
I would refer you to the following Briggs A, Clark T, Wolstenholme J and
Clarke P. Missing.... presumed at random: cost-analysis of incomplete data.
Health Econ. 12: 377-392 (2003).
Sorry I can't help you for the second (and probably more important) issue.
Kind Regards,
Carlo
2) The probability of a person (experiencing illness) incurring any
out-of-pocket expenditure i.e. greater than 0. For finding the
probability I want to use the independent variables like edu, place of
living, income, type of service use etc.
I am confused about what models to use. It looks like a standard probit
model, with expenditure at 0.
Or using OLS regression with fixed effects for the obs where out of
pocket expenditure is greater than 0.
Thanks for clarifying this for me!
shalini
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
