User-fee-removal improves equity of children's health care utilization and reduces families' financial burden: evidence from Jamaica

Abstract

Background: The impact of user-fee policies on the equity of health care utilization and households' financial burdens has remained largely unexplored in Latin American and the Caribbean, as well as in upper-middle-income countries. This paper assesses the short- and long-term impacts of Jamaica's user-fee-removal for children in 2007.

Methods: This study utilizes 14 rounds of data from the Jamaica Survey of Living Conditions (JSLC) for the periods 1996 to 2012. JSLC is a national household survey, which collects data on health care utilization and among other purposes for planning. Interrupted time series (ITS) analysis was used to examine the immediate impact of the user-fee-removal policy on children's health care utilization and households' financial burdens, as well as the impact in the medium- to long-term.

Results: Immediately following the implementation of user-fee-removal, the odds of seeking for health care if the children fell ill in the past 4 weeks increased by 97% (odds ratio 2.0, 95% confidence interval (CI) 1.1 to 3.5, P = 0.018). In the short-term (2007-2008), health care utilization increased at a faster rate among children not in poverty than children in poverty; while this gap narrowed after 2008. There was minimal difference in health care utilization across wealth groups in the medium- to long-term. The household's financial burden (health expenditure as a share of household's non-food expenditures) reduced by 6 percentage points (95% CI: -11 to -1, P = 0.020) right after the policy was implemented and kept at a low level. The difference in financial burden between children in poverty and children not in poverty shrunk rapidly after 2007 and remained small in subsequent years.

Conclusions: User-fee-removal had a positive impact on promoting health care utilization among children and reducing their household health expenditures in Jamaica. The short-term and the medium- to long-term results have different indications: In the short-term, the policy deteriorated the equity of access to health care for children, while the equity status improved fast in the medium- to long-term.

Figure 1

Healthcare utilization among under-18 children fell ill in the past 4 weeks. 1. To generate this figure, we split the 2007 sample into two parts–the sample interviewed before the implementation of user-fee-removal policy and the sample interviewed four weeks after it. 2. The observations numbers in the JSLC surveys vary by year (Most years have observation numbers between five thousand and eight thousand. For several years, the observation number is above fifteen thousand, such as 2008, and 2012). To increase the observation numbers involved in the generation of each data point in the figure above, we combined data from 1996 and 1997, 1998 and 1999, 2000 and 2001, 2001 and 2002, 2009 and 2010. 3. Sample weight is applied to all available years.

Figure 2

The difference in health care utilization between children in poverty and children not in poverty, among under-18 children fell ill in the past 4 weeks. 1. The observations numbers in the JSLC surveys vary by year (most years have observation numbers between five thousand and eight thousand. For several years, the observation number is above fifteen thousand, such as 2008, and 2012). To increase the observation numbers involved in the generation of each data point in the figure above, we combined data from 1996 and 1997, 1998 and 1999, 2000 and 2001, 2001 and 2002, 2009 and 2010. 2. Subjects under 18 years old in 2007 interviewed before 28 May 2007 are combined to year 2006 to prevent losing observations. 3. Sample weight is applied to all available years.

Figure 3

The proportion of households with under-18 children suffered catastrophic health expenditure in the 4 weeks preceding the survey if the children fell ill in the past 4 weeks. 1. To generate this figure, we split the 2007 sample into two parts–the sample interviewed before the implementation of user-fee-removal policy and the sample interviewed four weeks after it. 2. The observations numbers in the JSLC surveys vary by year (most years have observation numbers between five thousand and eight thousand. For several years, the observation number is above fifteen thousand, such as 2008, and 2012). To increase the observation numbers involved in the generation of each data point in the figure above, we combined data from 1996 and 1997, 1998 and 1999, 2000 and 2001, 2001 and 2002, 2009 and 2010. 3. Sample weight is applied to all available years.

Figure 4

Difference in probability of experiencing catastrophic health expenditures between households in poverty and households not in poverty with sick children. 1. The observations numbers in the JSLC surveys vary by year (Most years have observation numbers between five thousand and eight thousand. For several years, the observation number is above fifteen thousand, such as 2008, and 2012). To increase the observation numbers involved in the generation of each data point in the figure above, we combined data from 1996 and 1997, 1998 and 1999, 2000 and 2001, 2001 and 2002, 2009 and 2010. 2. Subjects under 18 years old in 2007 interviewed before 28 May 2007 are combined to year 2006 to prevent losing observations. 3. Sample weight is applied to all available years.