“If you build it, will they come?” Facility-Level Characteristics that Determine Demand for Health Care Services in Rural Uganda

Background: Improvement in population health outcomes is an intrinsic goal of the health system that is facilitated when there is increased access to quality health care. Access to care enables individuals to avail themselves of health care resources that preserve or ameliorate their health outcomes. Good health as an outcome is desired by both individuals and national governments. Individuals desire good health for the obvious benefit of a better quality of life that it brings. National governments benefit when their citizens are in good health because they are able to contribute to economic growth. Policy makers in low-income countries, such as Uganda, are therefore challenged consistently to come up with innovative ways to improve access.In Uganda, promoting access to health service – physically and financially – has been an explicit goal of the national government.

Objective: This study examines the determinants of modern health facility use in rural Uganda.  It also assesses the relative importance of facility characteristics compared to individual characteristics in determining the use of health care services when ill. 

Data/Methods: Pooled data on health seeking behavior from two rounds of the Uganda National Panel Surveys is used. Both household and community module data are used in the analysis. The main outcome of interest is use of modern curative care. It is a binary variable indicating whether or not a respondent who was ill sought care from a modern health care provider, defined as government or private/NGO provider. Probit and multilevel models are used to accommodate the hierarchical nature of data. 

Results: The results are mixed. Although the results show the hypothesized positive association with facility attributes such as drug availability and service availability, these relationships are not significant and individual attributes seem to drive health facility use. Probit models also seem to fit the data better than the multilevel specification.

Conclusions: An understanding of the drivers of health service use is especially important in resource scarce environments where the need to be efficient with the use of the limited available resources is more urgent. The findings from the study will also be relevant to similar low-income countries interested in improving access to services. Further work will investigate the use of full information maximum likelihood estimators and heckman-type estimators to corroborate these findings.