The Impact of Health Insurance in Low- And Middle-Income Countries

Chapter 9: Colombia's Big Bang Health Insurance Reform

155 Chapter 9 Colombia’s Big Bang Health Insurance Reform Ursula Giedion, Carmen Elisa Flórez, Beatriz Yadira Díaz, Eduardo Alfonso, Renata Pardo, and Manuela Villar Colombia is one of the few developing countries that have introduced government-­ subsidized universal health insurance by drastically changing social security schemes and breaking the public sector monopoly. The reforms began in 1993, when approximately 28% of the population was covered by insurance through the traditional Latin American approach of a social security system delivering services directly to the covered population. By 2005 health insurance (with choice of provider) reached more than 70% of the total population and close to 60% of the lowest two income quintiles. Recent estimates suggest that insurance coverage reached 86% of the population by the end of 2006, with another 2% covered by military and other programs. The population is covered through two regimes: the contributory regime for the employed and self-employed (covering 40% of the population in 2006) and the subsidized regime for the poor (covering 46% of the population in 2006; Clavijo 2009). By 2009 coverage had expanded to 89% (Tsai 2010). For this chapter the gradual implementation of the subsidized health insurance regime for the poor provides a unique opportunity to apply semiparametric methods (propensity score matching, double difference, and matched double difference) to identify differences in health-related outcomes between those with insurance and those without (see box 9.1 for details on the data and methodology of this study). The impact of the contributory regime on similar variables is analyzed using an 156 Chapter 9 instrumental variable approach. Those without insurance nevertheless remain eligible for services provided directly by the government through public facilities, so those without insurance in this analysis retain access to traditional public providers. Box 9.1 Data and methodology No single household survey in Colombia synthesizes data on access, use, health status, and financial protection for the population. Whereas recent Colombian Demographic and Health Surveys (DHS) (1995, 2000, and 2005) offer household data on access, use, and health status for small children and women of child-bearing age, the Living Standards Measurement Survey 2003 (LSMS) provides information on general use of health care and out-ofpocket spending as well as a wealth of socioeconomic data (including information on employment). We used DHS data to evaluate the subsidized regime on access, use, and health status. In contrast to the contributory regime, the gradual implementation and still incomplete coverage of the subsidized regime among the poor allowed us to apply semiparametric methods to identify differences in health outcomes between those with insurance and those without. We used LSMS data to evaluate the impact of the contributory regime on healthrelated outcome variables. LSMS data do not, however, include any health status variables that can be expected to change as a consequence of benefits provided under the contributory regime. The analysis of the contributory regime therefore did not try to look at the impact of health insurance on health status. Further, we had to resort to either propensity score matching or instrumental variables to evaluate the impact of the contributory regime, as we had only one cross-sectional data set (LSMS 2003). Given that almost all of those working as formal workers participate, matching affiliates (through propensity score matching) to similar nonaffiliates was impossible, so we used instrumental variables. The table below summarizes the data used in the analysis. Summary of data sources for evaluation Subsidized regime Contributory regime Access, utilization, and health status Financial protection Access and utilization Financial protection LSMS 2003 (cross-sectional data) ✓ ✓ ✓ DHS 1995, 2000, and 2005 (cross-sectional and repeated cross-sectional data) ✓ Administrative data at the municipal and state level ✓ ✓ ✓ ✓ Census data at the block level ✓ ✓ ✓ ✓ Source: Authors. Colombia’s Big Bang Health Insurance Reform157 The main goal of this chapter is to analyze existing household data to provide information on the impact of the Colombian health insurance scheme on key performance indicators of the health system. Specifically, it seeks answers to the following questions: Has insurance improved access to and use of health services for individuals in case of an adverse health event? Has insurance reduced the risk of having to confront an out-of-pocket health payment that destabilizes the financial welfare of the household or causes the household to fall below a poverty line? Has insurance improved health outcomes? Main features of the Colombian health sector reform Colombia is a...