Examining utilization of specialized outpatient care in Norway : An analysis of the factors explaining total consumption, public provision, private provision and the possibility for substitution in a public-private mix provision of specialized outpatient care

Abstract

BACKGROUND: There is a scarcity of international and Norwegian studies investigating the relationships inherent in a public-private provision of specialist outpatient care. Information on these relationships is needed in order to better organize the provision of care at the specialist level. OBJECTIVE: To investigate factors that explain total consumption, public provision, private provision and the possibility for substitution in a public-private mix provision of specialist outpatient care. METHOD: A Weighted Least Squares regression method that incorporates a “fixed effects” model for all health enterprises has been used to run the analyses. A comprehensive cross sectional dataset covering observations on health status, socio-economic status and supply side factors has been used as a proxy for establishing the need for specialist care. The dataset contains observations from all the 430 municipalities disaggregated into municipal, gender and age-group units. RESULTS: The number of general practitioners has an insignificant effect on the utilization of specialist outpatient care. Personal income is important in explaining the use of private specialists but not the use of public specialists. Travel-time has a negative effect on the general use of specialist care. The share of the population represented by the elderly above 80 years has an insignificant effect on the utilization of both public and specialist care. Immigrants from countries outside Western Europe have a strong negative effect on the use of specialist outpatient care. The potential for substitution exists between public and private specialists in the treatment of a broad amount of need-groups represented by different variables. However, the substitution effect is very weak as evidenced by the extremely low estimate values ranging from 0.000 to 0.032.