Health_Equity_and_financial_protection_report.pdf

TIMOR LESTE

HEALTH EQUITY AND FINANCIAL PROTECTION REPORT

TIMOR LESTE

HEALTH EQUITY AND FINANCIAL PROTECTION REPORT

P ub lic D is cl os ur e A ut ho riz ed

P ub lic D is cl os ur e A ut ho riz ed

P ub lic D is cl os ur e A ut ho riz ed

P ub lic D is cl os ur e A ut ho riz ed

WB456286

Typewritten Text 89850

ABOUT THE HEALTH EQUITY AND FINANCIAL PROTECTION REPORTS

The Health Equity and Financial Protection reports are short country-specific volumes that provide a picture of equity and financial protection in the health sectors of low- and middle-income countries. Topics covered include: inequalities in health outcomes, health behavior and health care utilization; benefit incidence analysis;

financial protection; and the progressivity of health care financing. Data are drawn from the Demographic and Health Surveys, World Health Surveys, Multiple Indicator Cluster Surveys, Living Standards and Measurement Surveys, as well as other household surveys, and use a common set of health indicators for all countries in the series. All analyses are conducted using the health modules of the ADePT software. Also available are Health Equity and Financial Protection datasheets that summarize key measures of equity and financial protection.

The most recent versions of the Health Equity and Financial Protection reports and datasheets can be downloaded at www.worldbank.org/povertyandhealth.

Photo credit: Alex Baluyut

ABOUT THE HEALTH EQUITY AND FINANCIAL PROTECTION REPORTS

The Health Equity and Financial Protection reports are short country-specific volumes that provide a picture of equity and financial protection in the health sectors of low- and middle-income countries. Topics covered include: inequalities in health outcomes, health behavior and health care utilization; benefit incidence analysis;

financial protection; and the progressivity of health care financing. Data are drawn from the Demographic and Health Surveys, World Health Surveys, Multiple Indicator Cluster Surveys, Living Standards and Measurement Surveys, as well as other household surveys, and use a common set of health indicators for all countries in the series. All analyses are conducted using the health modules of the ADePT software. Also available are Health Equity and Financial Protection datasheets that summarize key measures of equity and financial protection.

The most recent versions of the Health Equity and Financial Protection reports and datasheets can be downloaded at www.worldbank.org/povertyandhealth.

Photo credit: Alex Baluyut

ABOUT THE HEALTH EQUITY AND FINANCIAL PROTECTION REPORTS

The Health Equity and Financial Protection reports are short country-specific volumes that provide a picture of equity and financial protection in the health sectors of low- and middle-income countries. Topics covered include: inequalities in health outcomes, health behavior and health care utilization; benefit incidence analysis;

financial protection; and the progressivity of health care financing. Data are drawn from the Demographic and Health Surveys, World Health Surveys, Multiple Indicator Cluster Surveys, Living Standards and Measurement Surveys, as well as other household surveys, and use a common set of health indicators for all countries in the series. All analyses are conducted using the health modules of the ADePT software. Also available are Health Equity and Financial Protection datasheets that summarize key measures of equity and financial protection.

The most recent versions of the Health Equity and Financial Protection reports and datasheets can be downloaded at www.worldbank.org/povertyandhealth.

Photo credit: Alex Baluyut

About the Health Equity and Financial Protection Reports

The Health Equity and Financial Protection reports are short country-specific volumes that provide a picture of equity and financial protection in the health sectors of low- and middle-income countries. Topics covered include: inequalities in health outcomes, health behavior and health care utilization; benefit incidence analysis; financial protection; and the progressivity of health care financing. Data are drawn from the Demographic and Health Surveys, World Health Surveys, Multiple Indicator Cluster Surveys, Living Standards and Measurement Surveys, as well as other household surveys, and use a common set of health indicators for all countries in the series. All analyses are conducted using the health modules of the ADePT software. Also available are Health Equity and Financial Protection datasheets that summarize key measures of equity and financial protection.

The most recent versions of the Health Equity and Financial Protection reports and datasheets can be downloaded at www.worldbank.org/povertyandhealth.

Full citation: World Bank. 2014. Health Equity and Financial Protection Report – Timor-Leste. Washington, D.C.: World Bank.

Photo credit: Alex Baluyut

1Timor-Leste

Health Equity and Financial Protection Report

This report was produced and written by a task team consisting of Caroline Ly (Consultant), Yi-Kyoung Lee (Senior Health Specialist) and Xiaohui Hou (Senior Economist) under the task team leaders Yi-Kyoung Lee and Xiaohui Hou. The authors thank Leander Buisman (Consultant), Caryn Bredenkamp (Senior Economist) and Lourenco Camnahas (Consultant) for their additional analysis, technical advice and/or written contributions. Toomas Palu (Sector Manager, Health, Nutrition and Population of the East Asia and Pacific Region) provided technical comments and overall supervision on this report. The team would also like to thank Franz Drees-Gross (Country Director for Timor-Leste, Papua New Guinea & Pacific Islands, East Asia and Pacific Region) and Luis Constantino (Country Manager for Timor-Leste) for their guidance and support.

The report benefited from comments from Caryn Bredenkamp and Ajay Tandon (Senior Economist) on the earlier draft and peer reviews from Owen Smith (Senior Economist) and Tania Dmytraczenko (Senior Economist) and Augustine Asante (Research Fellow, Health Economics & Financing, UNSW).

The authors would also like to thank Hans Beck (Senior Economist) and Martin Cumpa Castro (Consultant) for contributions and input on the data analysis, Sarah Harrison (Consultant) for contributions on the comparative analysis as well as Emiliana Gunawan (Program Assistant) and Daniela Hoshino (Program Assistant) for administrative support, Devon Rohr (Consultant) for graphic design and Melody Molinoff (Consultant) for proofreading.

The team would like to thank the Ministry of Health in Timor- Leste for their support and the development partners for their feedbacks on the earlier drafts.

Financial support for this work was received from Rapid Social Response Multi-Donor Trust Fund.

ACKNOWLEDGEMENTS

ARI Acute respiratory infection

BIA Benefit-incidence analysis

CHC Community health centers

CI Concentration index

CPI Consumer price index

DHS Demographic and Health Survey

EAP East Asia and Pacific

GDP Gross domestic product

GHE Government health expenditures

GNI Gross national income

HIES Household Income and Expenditures Survey

HNGV Hospital Nacional Guido Valadares

MCH Maternal and child health

MICS Multiple Indicator Cluster Survey

MoH Ministry of Health

NHA National health accounts

OOPs Out-of-Pocket Spending

PPP Purchasing power parity

SAMES Serviço Autónomo de Medicamentos e Equipamentos de Saúde

SISCa Sistema Integradu Saude Communitaria

THE Total health expenditure

TLSLS Timor Leste Survey of Living Standard

WDI World Development Indicators

WHO World Health Organization

LIST OF ABBREVIATIONS AND ACRONYMS

3Timor-Leste

Executive Summary

1 . Timor-Leste’s health system

1.1. Equity and financial protection as policy goals

1.2. Health financing system

1.3. Health care delivery system

2 . Inequalities in health

2.1. Data availability

2.2. Inequalities in health

3 . Inequalities in health care utilization

3.1. Data availability

3.2. Inequalities in health care utilization

4 . Benefit incidence in government spending

4.1. Data availability

4.2. Inequalities in benefit incidence

5 . Financial protection in health

5.1. Data availability

5.2. Catastrophic out-of-pocket payments

5.3. Impoverishing out-of-pocket payments

6 . Progressivity of health financing

6.1. Data availability

6.2. Progressivity of health care financing

7 . Comparison with other countries on key equity measures

8 . Summary and policy implications

8.1. Summary

8.2. Policy Implications

9 . References

10 . Annexes

10.1. Measurement of indicators

10.2. Methodological notes

TABLE OF CONTENTS

5Timor-Leste

Is ill health more concentrated among the poor?

Yes, with exceptions. In general, ill health is concentrated among the poor in Timor-Leste. This includes some selected indicators of child health such as infant mortality, under-5 mortality, stunting, and underweight. But wealthier households are more likely to report child illnesses such as diarrhea and fever, potentially because of their ability to more easily recognize these symptoms. With respect to available measures of adult health, the better-off are more likely to experience sexually transmitted diseases among men, and obesity among non-pregnant women. Some risky health behaviors such as smoking are more prevalent among the poor.

Do the poor use health services less than the rich?

Yes. Health care utilization in Timor-Leste is concentrated among the better-off. Immunizations, treatment of acute respiratory infections, and use of bed nets are higher among children from wealthier households. Skilled antenatal care and delivery, contraceptive prevalence and use of bed nets are also higher among women from wealthier households.

Is the distribution of government spending on health pro-rich or pro-poor?

Pro-rich. Government spending on hospital care is found to be pro-rich, regardless of the assumptions made in the benefit-incidence analysis. When it comes to lower level health care services such as community health centers, health posts and mobile clinics, government spending is pro-poor using two of three methodological assumptions. Taken together, total subsidies for health are found to be pro-rich. Generally, government expenditure on health favors the better-off.

What is the effect of out-of-pocket payments on household financial well-being?

Minor. Out-of-pocket spending as a share of total health expenditures is only about 4 percent in Timor-Leste. Only

0.9 percent of households spend 10 percent of their household income or more on out-of-pocket health payments and 9.6 percent of households spend 10 percent or more of nonfood consumption on out-of-pocket payments. Health spending has a small effect on increasing the poverty rate. Less than 1 percent of households fell under the $1.25/day poverty line because of health spending. This is similar to results from 2001/02 and 2007/08.

Is health financing progressive or regressive?

Progressive. Taxes and out-of-pocket payments are also progressive relative to consumption. The wealthiest 20 percent cover 74 percent of total health care payments, compared to 39 percent of consumption. Health spending among the poorest two quintiles is only 5 percent, while their share of consumption is 22 percent.

What are the policy implications?

1. Continue improving the availability and quality of services at the frontline. The poor use community health centers and mobile clinics more while the wealthy tend to seek care from hospitals. The availability and quality of services offered at the community health centers, health posts, and SISCA sites needs to be strengthened so patients, particularly the poor, can access quality primary care. Recently, a large number of Cuban trained Timorese doctors have returned and present a unique opportunity to significantly strengthen the primary health care system, especially in rural areas. This increase in the number of qualified health workers should be accompanied by better management of human resources and the creation of conducive environments for health care workers. The instruments include, but are not limited to: incentivizing good performance to reduce inefficiency such as total absenteeism and reduced working hours; providing basic infrastructure (e.g., water and electricity) and essential commodities (e.g., medicines and medical supplies);

and taking measures to improve quality (e.g., step-wise quality accreditation system). Engaging community and empowering citizens to effectively monitor service delivery will also be imperative to ensure effective utilization of the services by citizens, particularly the poor.

2. Improve accessibility of secondary care to the poor when needed. The utilization of secondary care, which is largely publicly-subsidized, is pro-rich. In order for the poor to have better access to secondary care services, the up and down referral system between primary and secondary care levels needs to be strengthened. The major barriers for the poor to access secondary care might not be point of service user fees, but rather physical access to district and national hospitals. The government needs to explore options which can improve the poor’s access to secondary care when clinically required by providing, for example, travel vouchers, or (partial) reimbursement of travel cost.

Anecdotal evidence has shown that oversea referrals are pro-rich benefitting those who are economically and socially better off. Increased transparency on oversea referral programs is required to ensure the equitable allocation and utilization of resources.

3. Monitor the impact of tightening fiscal space on public health spending on poor. In the environment where fiscal space is tightening, the poor are usually more affected. The utilization of some essential health services that often incur out-of-pocket expenditure such as skilled delivery is much lower among the poor. This highlights the need for the government to closely monitor the impact of the tightening fiscal space on service utilization especially among the poor as well as the catastrophic out-of-pocket expenditures. The household surveys conducted by the MoH (e.g., DHS) and the MoF (e.g., TLSLS) can provide information to do so.

EXECUTIVE SUMMARY

This report analyses equity and financial protection in the health sector of Timor-Leste. In particular, it examines inequalities in health outcomes, health behavior and health care utilization; benefit incidence analysis; financial protection; and the progressivity of health care financing. Data are drawn from the 2009/2010 Demographic and Health Survey, the 2001-2002 and 2007-2008 Living Standards and Measurement Surveys as well as 2011-2012 Household Income Expenditure Survey, and the Ministry of Finance. All analyses are conducted using original data and performed using the health modules of the ADePT software.

1 .1 . Equity and financial protection as policy goals

The Government of Timor-Leste is strongly committed to improving financial protection in health and improving equity.

The following quote illustrates this commitment1:

“The Mission of the Ministry of Health is to strive to ensure the availability, accessibility and affordability of health services to all East Timorese people, to regulate the health sector and to promote community and stakeholders participation...”

Ministry of Health, Dili, Timor Leste

1 .2 . Health financing system

Health expenditure2

Timor-Leste spent US$ 82.1 (PPP) per capita on health or 5.1 percent of GDP in 20113. Government spending on health as a share of government expenditures has been declining in recent years as a result of shifting government priorities. Health spending as a share of total government spending declined from 12 percent in 2005 to an estimated 4 percent in 2012 4,5,6. In 2011, this amounted to about US$ 58.1 (2011 PPP) or US$ 33.1(2011 current) per capita. External financing for the health sector, which may flow through government and private channels, increased from 38.2 percent of total health expenditures (THE) in 2007 to 50.8 percent in 20117.

Table 1.1: Health expenditure data, 2011

Indicator

Health expenditure as share of GDP 5.1% Government expenditure on health, per capita US$33.1 (Current) US$58.7 (PPP adjusted) Government expenditure on health as share of total health expenditures 71.5%7 Out-of-pocket (OOP) expenditure on health as share of total health expenditures 4.0%

Source: WHO National Health Accounts database (2013)

1 Mission statement available at http://www.moh.gov.tl/?q=node/2, accessed 11/5/12 2 Data are from 2011 and available from the WHO Global Health Expenditure database, accessed 6/21/13 3 Ibid.

4 Timor-Leste Ministry of Finance Budget Portal, accessed 11/12/12.

5 Goods and Services as a share of total Goods and Services have dropped from about 13% to 4% between 2005 and 2012.

6 Without the human capital fund and infrastructure fund, the share of health is 6.4% and 6.2% in 2011 and 2012, respectively.

7 Note that external financing may be channelled through both the government and private sector. Thus, external financing and government financing may exceed 100%. Data are for 2011 and available from the WHO National Health Accounts database, accessed 6/21/13.

This section provides a brief overview of Timor-Leste’s health system, focusing on features that are likely to be especially salient for equity and financial protection.

1 . TIMOR-LESTE’S HEALTH SYSTEM

7Timor-Leste

Centralization and revenue-raising/sources of funds

Health financing in Timor-Leste is highly centralized.

Government spending accounted for 71.5 percent of total health spending as shown in Figure 1.1. In 2011, a large share or 80.8 percent of general government revenues came from the Petroleum Fund. Only 5.0 percent of revenues came from domestic taxes. Such heavy reliance on petroleum revenues makes the health sector vulnerable to petroleum price fluctuations and supply. External sources make up about a half of THE8. Generally, publicly provided health care is free at the point of service and thus out-of-pocket payments (OOPs) are low in comparison to other East-Asia and the Pacific countries9. OOPs in Timor-Leste account for 4.0 percent of

THE, compared to 11.7 percent, 7.0 percent and 0.1 percent in PNG, Samoa and Tuvalu10. This includes OOPs on private sector health care. According to the 2007-2008 Timor-Leste Survey of Living Standard (TLSLS), most of those who incurred OOPs did so from visiting private sector. Most households (89 percent) that sought health care reported visiting a public health care provider and only three percent of visits to the public provider incurred a payment to the provider11. However, households that visited private providers were more likely to pay OOPs for their services. More than half of patients who went to private providers made a payment to a provider12.

Source: WHO National Health Accounts database (2013)

Figure 1.1: Health care financing mix, 2007-2011

8 This indicates that external resources are channeled through both the public and private revenue sources.

9 Out-of-Pocket payments are typically direct spending after deducting third party payments such as insurance (WHO 2006) available at http://www.who.

int/nha/methods/estimating_OOPs_ravi_final.pdf. The TLSMS includes payments for health care services and medicines as part of health spending.

10 WHO National Health Accounts estimates (2013) 11 According to Lewis (2010), the frequency of informal payments to public health care workers among users of health services vary widely from 3 percent in Peru to 96 percent in Pakistan. Informal payment in this study is defined as “payments to individual and institutional providers, in kind or in cash that are made outside official payment channels or are purchases meant to be covered by the health care system including value of medical supplies purchased by patients and drugs obtained from private pharmacies but intended to be part of government-financed health care services,”

12 Nearly half of patients who went to private providers reported to pay nothing. This might be due to the fact that patients paid in-kind (e.g., chicken), and that the survey didn’t capture that.

Source: World Development Indicator (WDI, 2014)

Source: WDI (2014)

Figure 1.2: Government expenditure on health as a share of THE, 2011

Figure 1.3: Government health expenditure per capita, 2011

Comparison with other countries

Compared with other countries in East Asia, the share of public financing for the overall health expenditure is relatively high in Timor-Leste: government expenditure as a share of THE accounts for 71.7 percent in Timor, similar to Thailand (75.5 percent), but higher than Indonesia (34.1percent), Vietnam (40.4 percent) and Laos (49.3 percent).

Government spending accounts for a large share of total health expenditure in Timor, but in absolute value, the amount is quite small compared to other countries in East Asia. The government needs to increase public health spending especially as the economy grows. Per capita government health expenditure in

Timor (US$ 46) is only slightly better than Myanmar (US$

23) and Laos (US$ 37) in parity purchasing power (PPP) term, but still much lower than Thailand (US$ 202) which achieved universal health coverage (Figure 1.3).

The public expenditure as a share of total government expenditure is also low compared to other countries, only 2.9 percent in Timor, slightly better than Myanmar (1.3 percent), far less than Thailand (14.5 percent). As noted earlier, this figure might be under-estimated given the share of health was 6.4 percent and 6.2 percent in 2011 and 2012 without human capital funding and infrastructure funding (footnote 6).

However, this is still significantly lower than in 2005, at which time public health spending accounted for 12percent of total government spending in Timor. Yet, the argument to increase fiscal allocation on health is still thin given the projected limited government revenue growth in next few years, the competing demand from other sectors and the need for the health sector to show the effectiveness of resources spending in translation to improved health outcomes.

9Timor-Leste

Source: WDI (2014)

Figure 1.4: Public health expenditure as a share of government expenditure, 2011

Figure 1.5A: Percent allocation of health budget items, 2008 - 2012

Government spending on health

Spending on the health sector mirrors national priorities on infrastructure development (Figure 1.5A and Figure1.5B).

Recent and forecast health spending is concentrated on expansion and rehabilitation of hospital capital works projects.

Minor and major capital development accounted for 33 percent of expenditures in 2010 but has recently decreased to about 16 percent in 2012. Salaries increased from 20 percent in 2008 to 35 percent in 2012; the absolute amount has more than doubled from US$ 6 million in 2008 to about US$ 16 million in 2012. During this period of time, the share of goods and services has been decreasing from about 59 percent in 2008 to 30 percent in 2012; with the absolute amount dropping from

US$ 18 million in 2008 to US$ 14 million in 2012. There is no clear explanation as to why goods and services have gone down. The demand for health services is anticipated to increase.

Therefore, goods and services, which account for a large part of the recurrent expenditures, should have increased. The observed decrease raises concerns regarding the effectiveness of service delivery at the frontline of services. In 2012, transfers accounted for 20 percent in the total budget, roughly US$ 10 million per year. It raises serious equity concerns given a large part of the transfers are for the oversea treatment, which the better-off are most likely to benefit from.

Figure 1.5B: Amount allocation of health budget items, 2008 – 2012 (US$, million)

Provider organization

The health sector is divided into four organizational levels:

district health services, hospitals, Central services, and personalized services (i.e., (semi) autonomous agencies).

District health services are primarily responsible for delivering the basic health service package. They operate 66 community health centers (CHCs), 192 health posts, mobile clinics as well as the Sistema Integradu Saude Communitaria (SISCa). SISCa, introduced in late 2007, is a community based program that delivers preventative and curative health services, namely family registration, mother and child care, nutrition monitoring including food supplements, curative care and integrated vector control and health promoting activities. The distribution of health posts around the country is designed so that every citizen has a health facility within walking distance13. Sparsely populated areas are served by mobile clinics and SISCas.

The district health services work with the Central services in implementing a wide range of programs such as child immunization, malaria and TB programs.

Hospitals deliver higher level health care through a chain of five referral hospitals14 in the districts and one national hospital (Hospital Nacional Guido Valadares: HNGV) in Dili. Cases which cannot be treated at the district level are sent to the hospitals or in rare cases, transported overseas for treatment.

Central services provide centralized administrative support as well as bulk purchasing services. Personalized services include the Institute of Health Sciences that provides research and in-service training, Serviço Autónomo de Medicamentos e Equipamentos de Saúde (SAMES; central medical store), and National Laboratory.

Payment mechanisms and provider autonomy

Health facilities are reliant on government financing allocated by the central ministry level. Facilities have no significant sources of internally generated funds (IGF). Only HNGV in Dili charges user fees for non-Timorese or VIP patients and was able to keep about 70 percent of its IGF (MoH 2009). The Ministry of Health (MoH) makes personnel decisions, limiting the level of provider autonomy.

Resource availability and utilization

Taking into account the imminent inflow of new doctors, a density of qualified health workers in Timor-Leste is getting close to the benchmark recommended by WHO by the end of 2013 (i.e., 2.35/1000 population)15. While about two-thirds of the clinical health workers were nurses and midwives (n≅1500), the situation is changing very quickly with a major inflow of young medical graduates who have finished their training in Cuba and elsewhere (n≅790). The density of doctors per district (without incoming doctors) varies significantly from 0.04 to 0.35, while that of nurses (0.4-1.4) and midwives (0.2-0.9) is a little better. The addition of new health workers will not only increase the density of health professionals by 0.5/1,000 population, but it will also increase the costs of service delivery.

For example, it is expected that drug prescription costs will increase by $2/person/year based on the prescription pattern of skilled health workers. In addition, the MoH budgeted about $12m to cover additional salaries and allowances, equipment, housing, and training.

13 Walking distance is defined as a maximum of one hour. Due to the rough terrain in Timor, the bad conditions of many roads and the dispersed population, only between 40% and 50% of the population has a health facility within a one hour’s walk. The rest has to walk at least two or even three hours to reach a health post.

14 Baucau, Maliana, Suai, Oecussi and Maubesse 15 World Bank. Human Resources for Health in Timor-Leste: a Rapid Assessment, Nov 2012

1 .3 . Health care delivery system

Source: Ministry of Finance (2012)

11Timor-Leste

2 . INEQUALITIES IN HEALTH

A Demographic and Health Survey (DHS) was fielded in Timor- Leste from 2009 to 2010. The DHS has rich information for many health outcomes, particularly in relation to maternal and child health (MCH). The DHS lacks consumption and income measures, but one can construct an “asset index” using principal components analysis to rank households from poorest to richest (see Filmer and Pritchett 2001).

The tables in this section show how health outcomes vary across asset (wealth) quintiles. The tables show the mean values of the indicator for each quintile, as well as for the sample as a whole. Also shown are the concentration indices (CI), which capture the direction and degree of inequality. A negative value indicates that the indicator takes higher values among the poor, while a positive index indicates that the indicator takes higher values among the better-off. The larger the index in absolute size, the more inequality there is.

Table 2.1 shows that infant and under-five mortality, stunting, and underweight are worse among the poor. However, diarrhea, acute respiratory infections (ARI) and fever are worse among the better-off. The 2007/08 TLSLS confirms that diarrhea and fever/malaria were self-reported at higher rates among the better-off. This may be because the better-off were better able to identify the symptoms associated with these childhood diseases than the poor. Better educated households reported higher rates of diarrhea, fever and ARI than less educated households (data not shown).

Source: Authors’ estimates using ADePT and data from the 2009/10 Timor-Leste DHS

Note: Q1: Poorest Quintile, Q2: 2nd Poorest Quintile, Q3: Median Quintile, Q4: 2nd Wealthiest Quintile, Q5: Wealthiest

* CI is significant at 10%, **CI is significant at 5%, ***CI is significant at 1%.

Source: Authors’ estimates using ADePT and data from the 2009/10 Timor-Leste DHS

Note: Q1: Poorest Quintile, Q2: 2nd Poorest Quintile, Q3: Median Quintile, Q4: 2nd Wealthiest Quintile, Q5: Wealthiest; Obesity: Body Mass Index >30

* CI is significant at 10%, **CI is significant at 5%, ***CI is significant at 1%.

Table 2.2 shows certain risk factors such as smoking are more concentrated among the poor for both men and women: Smoking rates among men are high at 70 percent, with higher prevalence of smoking concentrated among the poor. Smoking is not very prevalent among Timorese women, but higher among poor women. Less than one percent of non-pregnant women are obese. The prevalence of sexually transmitted diseases (STDs) among men is worse among the wealthy.

In sum, the tables in this section indicate that children from poor households in Timor-Leste are disproportionately affected by ill health. Adverse adult risk behaviors such as smoking are concentrated among poorer households.

2 .1 . Data availability

2 .2 . Inequalities in health

Table 2.1: Inequalities in child health, 2009/10

Child (<5 years) health Q1 Q2 Q3 Q4 Q5 Total CI

Infant mortality rate (per 1,000 live births) 67.8 74.0 62.3 60.2 39.8 60.7 -0.094*** Under-five mortality rate (per 1,000 live births) 98.2 106.8 94.9 92.2 58.1 89.4 -0.092*** Stunting (%) 61.2 63.3 58.6 55.3 46.5 57.1 -0.054*** Underweight (%) 49.3 48.6 47.2 41.7 35.6 44.6 -0.065*** Diarrhea (%) 13.2 13.7 15.2 19.0 17.4 15.7 0.065*** Acute respiratory infection (%) 2.7 3.9 4.2 4.8 4.2 3.9 0.084*** Fever (%) 16.2 16.9 18.5 23.2 21.9 19.2 0.073***

Table 2.2: Inequalities in adult health and risk factors, 2009/10

Adult (15-49 years) health and risk factors Q1 Q2 Q3 Q4 Q5 Total CI

Obesity among non-pregnant women (%) 0.5 0.6 0.6 1.0 1.6 0.9 0.269*** Anemia among women (%) 40.6 36.7 41.8 39.7 33.8 38.5 0.098*** Smoking among women (%) 6.7 5.1 4.5 3.7 3.5 4.6 -0.141*** Smoking among men (%) 73.3 73.9 70.4 68.5 61.3 69.5 -0.027*** STDs among men (%) 0.0 0.8 2.1 1.4 1.3 1.1 0.414***

Most policymakers regard large inequalities in health outcomes between the poor and rich as undesirable. This section reports inequalities in child and adult health outcomes, as well as health behaviors.

As stated previously, a DHS was fielded in Timor-Leste from 2009 to 2010. The DHS has data on utilization of maternal and child health interventions. The DHS lacks consumption and income measures, but one can construct an “asset index” using principal components analysis to rank households from poorest to richest (see Filmer and Pritchett 2001).

The tables in this section show how health utilization varies across consumption or asset quintiles. The tables show the mean values of the indicator for each quintile, as well as for the sample as a whole. Also shown are the concentration indices which capture the direction and degree of inequality. A negative value indicates that the indicator takes higher values among the poor, while a positive index indicates that the indicator takes high values among the better-off. The larger the index in absolute size, the more inequality there is.

Table 3.1 shows coverage of key MCH interventions, including the treatment of childhood illness is usually higher among the better-off. Around 53 percent of children are fully immunized, 70 percent of children with ARIs received medical treatment.

Use of mosquito nets by children (and pregnant women) is about 42 percent with children from the wealthiest quintile twice more likely to use the nets than those from the poorest quintile. Inequalities in medical treatment of diarrhea among children under five are not statistically significant. About 55 percent of expectant women receive at least 4 skilled antenatal care visits but only 30 percent of women deliver their baby assisted by a skilled attendant. Women in the wealthiest quintile are 6.5 times more likely to deliver with a skilled attendant than women in the poorest quintile. This is consistent with what the report will discuss later on utilization of secondary care by income quintiles. The poor usually do not seek care at the secondary level. Another reason is that the fertility rate is quite high in Timor (5.7 on average according to DHS data).

Use of a modern method of contraception is more prevalent for wealthy women.

Source: Authors’ estimates using ADePT and data from the 2009/10 Timor-Leste DHS

Note: Q1: Poorest Quintile, Q2: 2nd Poorest Quintile, Q3: Median Quintile, Q4: 2nd Wealthiest Quintile, Q5: Wealthiest

* CI is significant at 10%, **CI is significant at 5%, ***CI is significant at 1%.

3 .1 . Data availability

3 .2 . Inequalities in health care utilization

Table 3.1: Inequalities in maternal and child health interventions, 2009/10

Interventions Q1 Q2 Q3 Q4 Q5 Total CI

Full immunization (%) 43.4 52.9 56.5 64.9 45.5 52.6 0.030** Treatment of diarrhea (%) 80.8 76.1 84.0 75.7 76.0 78.4 -0.011 Medical treatment of ARI (%) 53.2 66.8 75.6 73.2 73.6 69.6 0.054** Mosquito net use by children (%) 23.9 34.0 42.7 56.9 55.9 42.3 0.167*** Skilled antenatal care (4+ visits) (%) 41.1 44.9 57.3 63.6 68.8 55.1 0.112*** Skilled birth attendance (%) 10.6 14.2 21.2 38.9 69.5 30.3 0.392*** Contraceptive prevalence (%) 9.2 9.5 9.9 14.6 17.8 12.4 0.153*** Mosquito net use by pregnant women (%) 27.1 32.3 39.8 59.7 54.6 42.2 0.155***

In many countries, for a variety of possible reasons, the pattern of health care utilization tends to be distributed unequally across income groups, even after taking into account differences in medical needs. This section reports on inequalities in utilization of health care in Timor-Leste for different types of care, and for different types of health care providers.

3 . INEQUALITIES IN HEALTH CARE UTILIZATION

13Timor-Leste

4 . BENEFIT INCIDENCE IN GOVERNMENT SPENDING

The two most recent Living Standards Measurement Surveys (LSMS) for 2007/08 and 2011/12 (i.e., 2007/08 TLSLS and 2011/12 HIES) provide data on utilization of health services.

Government expenditure data at the hospital and district levels from the MoH and Ministry of Finance (MoF) provide information on the amount of subsidies to public health care.

While the 2011/12 HIES provides more recent data, it has some limitations. The 2011/12 HIES has combined out-of-pocket spending data on public hospitals and clinics. The 2011/12 survey also does not collect data on the frequency of visits to health care providers nor does it separate out utilization data by whether or not a visit was made to a public or private facility. It instead asks whether or not an individual in a household visited a provider in the last month. So there is potential bias in the utilization data if an income group tended to visit health care providers more frequently. Data on health care expenditures would also not serve as an indicator of utilization since health care is supposed to be free. Approximately three-fourths of all households that reported a member sought health care also reported zero health care expenditure.

On the other hand, the 2007/08 TLSLS captured frequency of data and separated out the category of health expenditures by provider category. However, the 2007/08 TLSLS provided overlapping categories of public providers in their dataset:

doctor, nurse/paramedic, midwife in government health facility, mobile clinic, health posts, government community health center, and government hospital. For the 2007/08 analysis, providers were grouped into two categories: government hospitals and district visits. District-level visits are defined as all other public visits to a government facility other than hospital (i.e. doctor, nurse/paramedic, midwife in government health facility, mobile clinic, government sub-community health center, and government community health center).

Typically, a National Health Accounts (NHA) exercise provides detailed health financing data on various levels of care such as sub-community health centers, health posts, outpatient hospital care and inpatient hospital care. An NHA exercise has not been completed for Timor-Leste. Instead, the MoF’s 2010 Review of the Health Sector and the MoH’s 2009 update to the 2007 Medium Term Expenditure Framework (MTEF) provides recurrent and capital expenditures for hospitals and districts.

For the purposes of this BIA, district recurrent expenditures are interpreted as total public outpatient subsidies and hospital recurrent expenditures as total public hospital subsidies16.

The tables in this section show the distribution across consumption quintiles of utilization for government facilities, fees paid to these facilities, and estimated subsidies to the health sector. The latter depend on the assumptions made to allocate subsidies to households; results are presented for three sets of assumptions. The tables show the shares of fees or shares of subsidies that go to each quintile. Also shown are the concentration indices which capture the direction and degree of inequality. A negative value indicates that the variable in question is higher among the poor, while a positive index indicates higher values among the better off. The larger the index in absolute size, there is more inequality in the indicator.

Table 4.1 shows the utilization of four types of public facilities separated in 2007/08. According to the 2007/08 TLSLS, the most frequently used type of service was the health post with 0.333 visits per year per person on average. On the other hand, hospitals, with an average of 0.236 visits per year per person, were less frequently used17. There is a striking contrast between the distribution of lower level health care services (sub- CHCs and mobile clinics) and use of general hospital services.

Utilization of the former steadily decreases with income. The decrease is most noticeable for sub-CHCs, which drop from an average number of 0.445 visits for the poorest quintile to

0.219 for the highest. In contrast, hospital visits increase with levels of wealth. These patterns are reflected in their respective concentration indices. The concentration index of sub-CHCs, mobile clinics and CHCs are noticeably negative, indicating that use of this type of service is higher among the poor. By contrast, the concentration index of hospital visits is positive, indicating that richer individuals use this service more than the poor. The concentration index for CHCs was not statistically

Policymakers typically take the view that government health expenditure (GHE) ought not to disproportionately benefit the better-off, and if anything ought to favor the poor more than the better-off. Benefit-incidence analysis (BIA) shows whether and how far GHE disproportionately benefits the poor. This section reports BIA results for Timor-Leste, using three methods for allocating GHE to households, namely (i) the constant unit cost assumption, (ii) the constant unit subsidy assumption, and (iii) the proportional unit cost assumption. The first is arguably the least plausible of the three, since it implies that higher fees are not a reflection of more costly care (see Table 4.2). However, it does have the attraction of not needing to be modified if part of (general) GHE goes on demand-side subsidies through, for example, a conditional cash transfer for health.

In Timor-Leste, recurrent spending on services is allocated to two major areas: district health services and hospitals.

Generally, spending on hospitals is pro-rich whereas spending on district health services is pro-poor. Overall total government spending on health favors the rich. The results presented below were obtained using the constant-unit-subsidy and proportional-unit-cost assumptions; it is assumed implicitly that supply- and demand-side subsidies have the same distributional impact.

4 .1 . Data availability

4 .2 . Inequalities in benefit incidence

16 Outpatient and inpatient subsidies are not distinguished in hospital subsidies. As a result, it is possible that we are underestimating outpatient subsidies because we do not account for the outpatient services delivered in the hospital setting.

17 Note: According to the MoF review of the health sector, district visits are 1.7 per capita and hospital visits are 0.2 per capita.

significant. Data on utilization in 2011/12 is incomplete but analysis (not shown) reveals that the concentration index for all public health facility visits was slightly negative, indicating that the poor utilized public facilities more often than the wealthy.

Table 4.2 shows the inequalities in visits to individual health care providers. The wealthy are more likely to see midwives as demonstrated by the positive concentration index. The concentration indices of visits to nurses/paramedics and doctors across indicators are not statistically significant.

In 2007/08, fees or OOPs paid to see public providers averaged about US$ 0.4. This low mean is driven by a high number of patients who paid nothing at the point of service. Among those who did see public providers, 97 percent paid nothing. But among the patients who went to a public provider and reported paying for services, they paid on average about US$ 11.7.

Table 4.3 shows the share of fees paid to district and hospital facilities in 2007/08 and all public facilities in 2011/12. In 2007/08, visits to health care providers in the TLSLS were categorized such that all visits to hospitals were categorized as public inpatient visits but could potentially include outpatient visits. All other visits including visits to doctors, midwives and nurses were categorized as outpatient visits but could potentially include visits to hospitals. In 2011/12, data on utilization was not clearly separated between public and private facilities but data on expenditures was disaggregated between public and private health facilities. Visits that were categorized as public hospital and clinic facility visits include all reported visits minus visits associated with positive payments to a private facility and visits to a traditional medical practitioner. It is possible that this utilization number still includes visits to a private provider in which no payments were reported. Based on this categorization, there appears to be a positive relationship between wealth and fees paid for public services.

Source: Authors’ estimates using ADePT and data from the 2007/08 TLSLS

Note: * CI is significant at 10%, **CI is significant at 5%, ***CI is significant at 1%.

Source: Authors’ estimates using ADePT and data from the 2007/08 TLSLS

Note: * CI is significant at 10%, **CI is significant at 5%, ***CI is significant at 1%.

Source: Authors’ estimates using ADePT and data from the 2007/08 TLSLS and 2011/12 HIES; District level facilities includes CHCs, sub-CHCs and mobile clinics

Note: * CI is significant at 10%, **CI is significant at 5%, ***CI is significant at 1%,† Standard error not computed; Data in 2007/08 reflect results after removing a household outlier.

Table 4.1: Inequalities in use of public facilities (visits per capita per year), 2007/08

Quintile Sub-CHC Mobile Clinic CHC Hospital

Poorest 0.445 0.134 0.174 0.179 2nd poorest 0.366 0.119 0.208 0.200 Median/middle 0.351 0.119 0.183 0.234 2nd wealthiest 0.288 0.102 0.172 0.255 Wealthiest 0.219 0.096 0.186 0.313 Total 0.333 0.114 0.185 0.236 Concentration Index -0.139*** -0.074** -0.003 0.118***

Table 4.2: Inequalities in use of public providers (visits per capita per year), 2007/08

Quintile Midwives Nurses/Paramedic Doctors

Poorest 0.036 0.148 0.063 2nd poorest 0.039 0.168 0.060 Median/middle 0.054 0.131 0.065 2nd wealthiest 0.055 0.140 0.051 Wealthiest 0.068 0.125 0.063 Total 0.051 0.142 0.060 Concentration Index 0.137*** -0.040 -0.007

Table 4.3: Distribution of fees paid to public facility (%), 2007/08 & 2011/12 2007/08 2011/12 Quintile District level Hospital Hospital & Clinic

Poorest 2.2 1.4 0.1 2nd poorest 11.2 0.8 3.3 Median/middle 22.3 14.8 6.3 2nd wealthiest 22.8 8.1 7.9 Wealthiest 41.6 74.9 82.5 Concentration Index 0.4079*** 0.6954*** 0.6989†

15Timor-Leste

Table 4.4 shows the incidence of government health spending.

The first two lines of the table show how aggregate government spending on health varies across hospital and district level services18. The table contains three sets of estimates of subsidy distribution across consumption quintiles. The first assumption is the constant unit cost assumption in which each visit at a specific level of care cost the same amount irrespective of the different services provided (e.g. the hospital inpatient costs for an uncomplicated child delivery would be treated the same as the costs for a caesarean). This would be calculated by taking the total costs of service (i.e. subsidies plus user fees) divided by the number of units of utilization. This method however can result in negative imputed subsidies since the user fees paid for services may exceed the unit cost of service. To address this, any negative imputed subsidies are set to zero.

The second type of assumption is that the unit subsidy is constant, irrespective of the user fees. It is just calculated as the total subsidy divided by the units of utilization. The third assumption is the proportional cost assumption. It assumes that the size of subsidies is proportional to the size of user fees.

This implies that the size of user fees reflects the case complexity and therefore cost of services provided.

The first two lines indicate that 43 percent of recurrent government subsidies are spent on public hospitals and 57 percent is spent at the district level. The first set of results (based on the constant unit-cost assumption) for 2007/08 shows that the poorest quintile receives on average 15.3 percent of government health spending while the richest quintile receives 26.1 percent. The corresponding concentration index (0.1111) is slightly positive. Although district-level utilization is slightly pro-rich, the overall pro-rich result is mainly due to the effects of public spending in hospitals which are pro-rich.

When unit subsidies are assumed to be constant (the second set of results), the subsidies become slightly more pro-rich with a concentration index of 0.1233. Finally, when unit costs are assumed to be proportional to the amount spent OOP, the subsidies for all types of services become considerably pro-rich.

The resulting picture is a strongly pro-rich incidence of public spending that is driven by considerably pro-rich subsidies to hospital services.

For 2011/12 data on total public health spending, we see that concentration indices for both the constant unit cost and unit subsidy assumptions are slightly negative but insignificant, making payments somewhat proportional across quintiles.

However, the proportional cost assumption appears to be pro-rich. The top two quintiles capture about 90 percent of government subsidies to the health sector.

Taken together, these benefit incidence analyses find no absolute evidence that government spending on health favors the poor. On the contrary, if one assumes that higher fees also reflect higher subsidies, one would conclude that government spending is, in fact, very much pro-rich and worsening when compared to 2007/08. When using the other assumptions, one also finds that total subsidies favor the better-off in 2007/08 and are proportional in 2011/12, an improvement from 2007/08. Therefore, the assumptions are critical to interpreting whether or not the 2011/12 data reflect a worsening in the benefit incidence.

Source: Authors’ estimates using ADePT and data from the 2007/08 TLSLS and 2011/12 HIES. The MoF’s 2010 Review of the Health Sector and the MoH’s 2009 update to the 2007 Medium Term Expenditure Framework (MTEF) provides recurrent and capital expenditures for hospitals and districts.

Note: * CI is significant at 10%, **CI is significant at 5%, ***CI is significant at 1%, † Standard error not computed; Data in 2007/08 reflect results after removing a household outlier.

18 It would be preferable to categorize the level of services as inpatient and outpatient services. However, the financial data from the MoH is reported as hospital and district services. Districts are expected to be responsible for primary care services while hospitals are responsible for inpatient services.

However, it is possible that, without an effective gatekeeper system, care provided at hospitals may include outpatient services. Or, because of limitations in transport, districts may be providing inpatient services.

Table 4.4: Inequalities in the incidence of government health spending (%), 2007/08 & 2011/12 2007/08 2011/12 Districts Hospitals Total Total

Total Subsidies (US$, million) 7.8 5.9 13.7 19.5 Share of Total Subsidy 56.9 43.1 100.0 100.0 Constant Unit Cost Assumption Poorest 18.5 12.9 15.