A Model of household choice among medical treatment alternatives in rural Kenya

Abstract

The purpose of this dissertation is to study patients' demand for health services in a given illness period. Particular attention is devoted to a model of patients' behavior throughout an illness episode and to a statistical estimation of its parameters. An observed visit to a health facility is assumed to have been

generated by a sequence of two decisions. The first decision is about whether or not to seek treatment for a given illness. If the outcome of this decision is to seek treatment, the next decision concerns the source of medical care. At any given moment, each of these decisions is discrete - in the sense that a patient chooses to seek treatment or not;

to seek it from one provider and not from any of the others. Further, over the entire illness period, these choice decisions display a sequential pattern of responses as patients visit different providers or return to the previous one(s) for follow-up care. To explain the above sorts of decisions, a demand model in which patients learn from their experience with health care providers is

developed. The factors whose effects on patients' behavior are studied with the aid of the model are the time and monetary costs of treatment, proximity, the facility-specific attributes, the socio-economic characteristics of the patients, patients' experience in treatment, patterns of income distribution and seasonality. The data used to estimate the effects of these variables were collected through a household survey in a rural area in Kenya. The estimated results indicate that households' choice of health facilities is highly sensitive to changes in time costs and money prices. It is also found that an increase in incomes shifts demand from low quality to high quality clinics but rather slowly. These and other results are used in a micro-economic simulation

exercise in the last chapter of the dissertation, to predict demands for health services' at different facilities, to evaluate the effects on households' welfare of imposing user fees in government clinics and to estimate the amount of revenue the government can hope to collect from its clinics when certain levels of user charges are in effect.