Horizontal inequity in healthcare use among residents of a large brazilian metropolitan region

Introduction: This study assessed horizontal inequity in healthcare use and the main factors associated with healthcare use among adults residing in a large Brazilian metropolitan region. Methods: Data were from a regionally representative household survey conducted in 2010 (n=12,213). Three main healthcare outcomes were studied: doctor visits, specialist visits, and hospital visits during the past 12 months. Concentration Indexes (CI) and Concentration Curves (CC) were used to assess horizontal equity in healthcare use along the income distribution. In addition, hurdle models were employed for each of the three outcomes. Hurdle models included logistic regression to model the probability of healthcare use, and zero-truncated negative binomial regression to assess the frequency of healthcare among those who had any utilization. Two groups of independent variables were included: Need-related variables (e.g. age, sex, previous medical diagnosis of chronic conditions, poor self-rated health) were used to standardize the CI in order to adjust for healthcare needs. The second group (non-need variables) included private health insurance, schooling and a socioeconomic status (SES) index used as a proxy measure for living standards. Results: Demographic factors such as sex and age were important predictors of all three outcomes. Although there are inequalities in the use of services, the CI and CC for doctor visits indicated nearly equity, but with a slight pro-rich orientation when use was standardized by need variables. The hurdle model showed that the SES index was significant only for the first doctor visit. After the first visit, SES was no longer associated with increases in the number of visits. The CI and CC for specialist visits showed a higher degree of inequity (also with a pro-rich orientation) when compared with general doctor visits. Private health insurance and higher SES were important components in the CI, particularly in the case of specialist visits. Hospitalization was found to be pro-poor. However, the hurdle model showed that the SES index was neither associated with probability of hospitalization nor with increased number of hospitalizations. Conclusions: The horizontal equity principle calls for equal treatment of people in equal need regardless of socio-demographic factors such as income, education and place of residence. This study shows that health needs were the most important factors driving most healthcare use, suggesting presence of horizontal equity in the region. However, pro-rich inequities persist particularly in the use of specialized services.