Estimating and correcting 'missing' people with disabilities in Indonesian statistics: Improving understanding of health care use at a regional level

There is a lack of information about people with disability (PWD) in Indonesia. Due to various reasons,
including fear of stigma and unclear definitions of disability in many surveys, there is a large
underreporting of the numbers of PWD across the country. This creates different issues: a lack of focus in
government policies, poor health services and a lack of visibility of PWD. The aim of this research is to
improve estimates of the level and characteristics of PWD in Indonesia. Further, to assess, test and apply
different methods of survey and population-based data collection and statistical analyses for estimating
and correcting 'missing' and 'misclassified' PWD. This project will use a mixed-methods approach which
will include qualitative semi-structured interviews with health professionals alongside statistical analysis
of available data. First, bivariate analysis and regression techniques will be used to explore the PWD who
are 'missing' from the statistics. The project will then use new and innovative quantitative methods for
this specific substantive area of disability. Missing data will be investigated using advanced methods for
missing data analysis such as multiple imputation for investigation of values missing not at random
(MNAR). Monte Carlo simulations will also be used to create records for missing PWD in Indonesia.
Regression analysis will be complemented by Latent Class Analysis (LCA) which will help to perform
classifications of those "missing" and those "misclassified" after they have been imputed and this new
summary variable will allow for further investigation of characteristics of people in these important
categories.