• Open Access

General health in Timor-Leste: self-assessed health in a large household survey

Authors


Correspondence to:
Dr Jaya Earnest, Associate Professor and Postgraduate Research Co-ordinator, Centre for International Health, Curtin University of technology, GPO Box U1987, Perth WA, 6845. Fax: (08) 9266 2608; e-mail: J.Earnest@curtin.edu.au

Abstract

Objective: Timor-Leste is one of the world's newest nations and became a democracy in 2002. Ranked 150 out of 177 in the 2007 UNDP Human Development Index, the country has the worst health indicators in the Asia-Pacific region. The objective of this study was to collect and analyse data on subjectively assessed general health, health service use, migration and mobility patterns.

Methods: The data collection involved recording self-reported status of general health using a structured questionnaire. The survey was administered to 1,213 Timorese households in six districts using a multi-stage random cluster sampling procedure. Basic descriptive statistical analyses were performed on all variables with SPSS version 13.

Results: More than a quarter (27%) of respondents reported a health problem at the time of the survey. Only approximately half of respondents assessed their health to be good (53%) or average (38%). Barriers reported in the uptake of healthcare services were no felt needed; difficulty in accessing services and unavailability of service.

Conclusions: Results reveal that Timor-Leste needs a more decentralised provision of healthcare through primary healthcare centres or integrated health services. Trained traditional healers, who are familiar with the difficult terrain and understand cultural contexts and barriers, can be used to improve uptake of public health services. An adult literacy and community health education program is needed to further improve the extremely poor health indicators in the country.

Implications: Key lessons that emerged were the importance of understanding cultural mechanisms in areas of protracted conflict and the need for integrated health services in communities.

The Democratic Republic of Timor-Leste (East Timor) is the world's newest nation. Prior to Indonesian occupation in 1975, the country was a Portuguese colony for more than 450 years.1 It is widely recognised that the decades of Indonesian occupation were marked by violence, human rights abuses and an estimated 200,000 deaths.2 Following the vote for independence from Indonesia in August 1999, anti-independence and militia groups trained by the Indonesian army led a violent assault on the nation's people and infrastructure, in which thousands of East Timorese were killed and injured and public buildings, private homes and businesses were destroyed.1

During the crisis of 1999, more than 75% of the estimated population of 850,000 was displaced and there was widespread damage to infrastructure. Dili and other towns had been razed; basic infrastructure had been decimated; there was no water, sanitation, electricity or phone system and over 90% of health and educational infrastructure had been destroyed.3 In addition, the emigration of core health and education professionals who were mainly Indonesian caused the ‘total collapse’ of the health and education system.2,3 While much ‘nation-building’ work has since been undertaken by the United Nations, international organisations, Non Government Organisations (NGO), donor nations such as Australia, and the new East Timorese Government itself, health remains a core and vital priority for the nation's development.4,5

Australia has been a key partner and supporter of Timor-Leste and has a shared interest in the country in terms of economic development and security. It has provided more than $641 million in Development Assistance to Timor-Leste between 1999 and 2007. After the conflict in 2006/07, it significantly lifted aid to provide emergency support in the area of policing and security. The new Australia-Timor-Leste Development Assistance Strategy for 2007-11 will provide long term support in the areas of finance and budgets, water and sanitation, policing and newer areas of partnerships that will assist in supporting and strengthening health service delivery, vocational education, the court and justice system and infrastructure projects. Australia is determined to support Timor-Leste, its closest neighbour, in building governance and stronger health and education systems.6

As in many other developing countries, infectious and parasitic diseases such as TB, malaria, dengue and diarrhoeal diseases, are endemic in Timor-Leste. Communicable diseases account for 60% of all deaths. The main causes of death are malaria, dengue fever, acute respiratory infections including tuberculosis and diarrhoeal diseases. Immunisation rates are extremely low; with only 5% of children fully immunised and national vaccination programs being carried out by UNICEF to address this problem.7,8

The poor health infrastructure, with only one tertiary referral centre (Dili National Hospital), and less than 75 physicians (including general and specialist) in the whole country,3 in combination with water and sanitation problems, food insecurity, malnutrition and poverty, is an enormous challenge for the country and its people. The Cuban government is currently supporting Timor-Leste's Ministry of Health by providing nearly 286 doctors.9

Although 70% of the population have access to some form of health services (with an average walking time of 70 minutes), a lack of trained health workers and effective health systems means the overall population's health remains extremely poor.7 Currently, institutions such as the World Health Organization (WHO), United Nations Children's Emergency Fund (UNICEF), United Nations Population Fund (UNFPA), United Nations Development Programme (UNDP) and United States Agency for International Development (USAID) are assisting the East Timorese Government by providing funding, and technical and program assistance to develop local skills for emergency and disaster planning; surveillance of communicable diseases; health staff training and an extended program of immunisation.3,8 It is also hoped that when international organisations leave the country, an effective health system and structure will remain in place to guarantee the health needs of the Timorese population.

This paper reports results on the perceptions of self-assessed general health drawn from the results of a larger study that investigated migration patterns, HIV/AIDS awareness, knowledge and practices, and general health after independence.

Methods

The Migration Patterns Survey

The Migration Patterns Survey (MPS) was developed using a modified version of the survey tool published in ‘Situational Assessment on Migration and HIV/AIDS’10 together with surveys that had been developed and tested by IOM in the Caribbean.11 A 106 item questionnaire was developed by the two authors; pilot tested for understanding, and the revised version which was translated into Tetun (the widely spoken local dialect), Portuguese and Bahasa Indonesia. This household survey formed part of a larger project on migration, mobility and HIV knowledge, attitudes and practices. The survey was administered to heads of households and questions regarding overall general health, health seeking behaviour, and barriers to health service utilisation were asked of participants. The household head was defined as an adult person (≥18 yrs) who was present at the time of the survey and self identified as the head of the household.

The survey was administered between October and December 2005, in six of the 13 districts of Timor-Leste. The six districts accounted for 55% of total households in the country. The districts identified for the study (with assistance from the International Organisation for Migration) were Baucau, Bobonaro, Liquica, Cova Lima and Oecussi, and suburbs of Dili, the capital of Timor-Leste. These sites were chosen because they were border districts of Timor-Leste, and/or districts with reportedly higher percentage of migrant and greater mobility. The questionnaire also consisted of questions concerning demographic and socio-economic status, mobility and displacement, HIV/AIDS knowledge, attitudes and practices and general health. The questionnaire was administered in Tetum, the local language, or Bahasa Indonesian.

The multi-stage, random cluster sampling procedure

Drawing on the Tanner, Magnani & Shuaib's improved cluster sampling method for resource poor settings, and Demographic and health survey (DHS) administered in Timor-Leste in 2003, a multi-stage random cluster sampling strategy was designed.12,13 The term ‘cluster’ in the standard sampling of this study means a natural grouping of households within a village (suco) and sub-village (aldeia). Clusters are an aggregation of sampling units from which a smaller sub-sample can be selected. In the field, clusters usually had clear boundaries, were located close to one and another, and the size of the cluster is available prior to sample selection.12

Using a probability proportional to population measure drawn from the DHS (2003),13 the number of households that needed to be contacted for the survey in the six districts was estimated at 1,150 at 5% significance and 80% power.

  • 1In the first stage data from the Timor-Leste Census 2004, was used to obtain the total number of households in the six districts. According to census data, 55.1% of households are found in these six districts. A probability proportion to population sample percentage calculation with in each of the six districts aided in obtaining the total number of surveys to be administered in each district.
  • 2In the second and third stages, data obtained from the 2001 suco survey was used to identify sucos (sub-districts) based on population characteristics per 1,000 persons. The most populous sub-districts and sucos were selected. In the third stage, aldeias were randomly selected based on number of households in the aldeia.
  • 3In the final stage, maps of the aldeias were obtained from the Statistics Bureau of the Ministry of Planning to aid in sampling the cluster of households in the aldeias. The maps provided an aerial image of the aldeias and the location of households.
  • 4Using a convenience sampling approach every second household in the aldeia was sampled. In households where there was no head of household available at the time, were replaced by other households, where a head of household was available at the time of the interview. In order to reduce cluster bias, clusters never exceeded 40 households per aldeia.

Training of Survey Research Assistants

The research assistants (RAs) were recruited from shortlisted Timorese university students and those who had applied for the position in Dili. The authors were assisted in the shortlisting process by two local NGOs in Dili. The training of the RAs consisted of a two-day workshop in introducing the project, concepts of administering surveys and standardised interviews and data collection. Several RAs had experience working for NGOs in Timor-Leste and some had collected survey data as part of that work previously. During survey administration, RAs were instructed to proceed to the next house if the head of household was absent at the time of the visit. The survey was administered by research assistants under supervision of a qualified Timorese research supervisor. The authors were also in the field at the time of data collection, with the second author co-ordinating the whole in-field data collection process.

Ethics approval

Prior to the field research, ethical approval was obtained from the Human Research Ethics Committee at Curtin University of Technology, and from the Ministry of Health in Timor-Leste. Permission was also obtained from the International Organisation of Migration who facilitated the study and the UN Theme group on HIV/AIDS. The questionnaire as well as the study proposal and protocol were discussed with involved NGOs, UN agencies and the Ministry of Health of Timor-Leste and piloted before use. All participants gave written or oral consent to participate in the study.

Data Analysis

Data were entered and analysed in Western Australia using SPSS Version 13. Basic descriptive statistical analyses were performed on all variables. For some variables, univariate and multivariate analyses were performed using ordinary and binary logistic regression to test for associations between outcomes of interest. The main outcomes of interest presented in this paper relate to migrant status and general health. Statistically significant relations found in univariate analysis were examined in multivariate analysis while controlling for gender and age variables. The final multivariate models did not include any terms that were non-significant. Different models were used for the univariate and multivariate analysis; these included binary, logistic and ordinal models. The binary logistic model giving odds ratio was used to determine independent predictors of migrant status. Differences observed were tested using the Chi-square and student t-test, and further analysis was conducted testing for correlations. Statistically significant differences and correlations have been reported in the text, and/or in Table 1 and 2.

Table 1.  Determinants of general health.
 Health Status (n=1176)
 Very Good n=85 (7.23%)Good n=626 (53.23%)Average n=441 (37.50%)Poor n=23 (1.96%)Very Poor n=1 (0.09%)p-value for associationaOdds Ratiob
  1. Notes:

  2. a) χ2 test of association.

  3. b) (95% CI) and p-value for linear trend.

  4. c) The number of missing values represents participants who did not provide a response to the corresponding question (of the 1176 participants who stated their self-rated health status).

  5. d) Missing values include non-response and those participants who did not go to school.

  6. e) Missing values include non-response and those participants who did not earn a regular income or answered “unsure/don't know” (30%, n = 352).

  7. Participants who did not answer the question on self-rated health status were excluded, thus the total analysis population is 1,176.

Age group [n = 10 missing (0.85%)]c
  18-24 years201365110<0.01N/A
  25-39 years241636801  
  40-54 years191396840  
  55-69 years1211010470  
  70-85 years871149110  
Went to school? [n = 20 missing (1.70%)]c0.51
  No16166202110<0.01(0.42-0.61)
  Yes67452229121 <0.01
Education level [n = 402 missing (34.18%)]d
  Some Primary School13736030<0.01N/A
  Completed Primary School5805630  
  Some Secondary School8622721  
  Completed Secondary School291997840  
  Some Tertiary Education732500  
  Completed Tertiary Education617400  
SES [n = 405 missing (34.4%)]e
  leqslant R: less-than-or-eq, slant10 USD1235194   
  <60 USD1656140 <0.01N/A
  <120 USD1191290   
  leqslant R: less-than-or-eq, slant120 USD1680351   
Sex (n = 0 missing)c0.69
  Male47289274151<0.01(0.58-0.83)
  Female3833716780 <0.01
Migrant (n = 0 missing)c0.54
  No33412346131<0.01(0.45-0.65)
  Yes5221495100 <0.01
Left home in 1999 [n = 14 missing (1.19%)]c1.19
  No22856041<0.01(0.93-1.53)
  Yes63528380190 0.17
Family member (went missing in 1999) [n = 18 missing (1.53%)]c0.79
  No785654152110.49(0.55-1.13)
  Yes6482220 0.20
Died during conflict [n = 16 missing (1.36%)]c0.96
  No765494061710.03(0.7-1.30)
  Yes7663260 0.80
Table 2.  Multivariate analysis of determinants of general health.
Health status RRRStd. Err.ZP> |z|95% Conf. Interval
      LowerUpper
  1. Notes:

  2. 1. Nominal logistic regression was used instead of ordinal logistic regression Results are presented as Relative Risk Ratios (RRR)

  3. 2. Health status ‘very good’ (Category No. 1) is the base outcome i.e. the RRR's for each category are the RR's compared to category 1.

GoodAge0.990.01<0.010.680.971.02
 Sex1.100.300.340.730.651.86
 Migrant0.300.08<0.010.000.170.51
 Education0.870.09<0.010.180.711.07
AverageAge1.020.011.120.260.991.04
 Sex0.730.22<0.010.280.411.30
 Migrant0.020.05<0.010.000.080.27
 Education0.780.78<0.010.000.560.87
PoorAge1.030.021.420.150.991.08
 Sex0.380.28<0.010.180.091.57
 Migrant1.220.800.300.770.344.43
 Education0.730.18<0.010.190.451.17

Results

The overall number of surveys administered in the six districts was 1,213. The number of surveys administered in the six districts was 365 in Dili, 240 in Baucau, 135 in Liquica, 188 in Bobonaro, 145 in Cova Lima, and 140 in Oecussi.

Demographic characteristics

The mean age of the respondents was 37.5 years (median, 35 years, mode 30 and range 15-85 years). Sixty-one per cent of the respondents were in the age group 20 years to 42 years revealing a significant young and middle aged population in the surveyed districts. The gender distribution of the respondents was 53.2% (n=638) males and 46.8% (n=561) females. Sixty-five per cent (n=785) of the respondents had some education. The largest proportion of the literate group had completed secondary school (40%, n=313), followed by those who had completed primary school (19%, n=147), or some primary education (19%, n=147).

Health concerns

When asked in general terms about their health concerns, 75% of respondents were concerned about malaria, followed by TB (28.5%), dengue (28.1%), respiratory infections (26.8%) and diarrhoea (17.5%). Only 14.5% of respondents were concerned about HIV/AIDS (Figure 1).

Figure 1.

Health concerns of respondents in Timor-Leste, 2005 (reporting valid per cent), multiple answers per participant allowed.

Overall, 27% (n=329) of respondents stated they had a health problem at the time the survey was administered. In 25% (n=292) of households, someone else within the same household had a health problem at the time of the survey. Asked to specify the health problem, most reported it to be respiratory problems or malaria. Only a small proportion of respondents perceived their health status to be either ‘good’ or ‘average’.

Displacement and migrant status

In order to determine displacement and impact due to the conflict in 1999, respondents were asked whether they had left their home and/or whether a family member had gone missing or was killed during the conflict. Nearly 83.9% of respondents had been displaced during the conflict in 1999. Of those displaced, 49% stated they had fled to another district or the hills in Timor-Leste and 40% had left the country. Of the displaced respondents, 28% returned within the first six months after the conflict; and equal proportions (9%) returned within six months to a year and after one year. Oecussi and Cova Lima had the largest proportions of displaced persons in this sample during the conflict. A migrant status was assigned if people had been displaced and not moved back to the village or had moved elsewhere since the conflict. Out of a total of 1,213 participants from six districts 363 were identified as migrants.

Determinants of general health

The study examined various determinants of general health to test for an association (see Table 1). The determinants of general health examined in the study were age, education, literacy, socio-economic factors, gender, migrant status and displacement impacts due to the violence in 1999.

Including all observed associations into a multivariate model, only two determinants: age and gender remained statistically significant. The relative risk of having poor health status compared to having very good health was 70% lower for migrants compared with non-migrants (p <0.001) after adjusting for age and sex. Similarly, after adjusting for age and gender, the relative risk of having average health status compared to having very good health status, was found to be 30% lower for each unit increase in education level (p=0.001) (see Table 2).

Place of treatment of respondents

About two thirds of the respondents (67%, n=774) had sought treatment for various illnesses in the past year. On average, the closest health facility was located just under one hour away when travelling on foot, or half an hour when travelling by vehicle. The most frequently used health providers were primary or community health centres (47.87%), followed by traditional healers (30.86%), governmental hospitals (8.42%), pharmacy/dispensary (7.76%) and NGO health services (7.67%) (see Table 3). Of those who had sought treatment, more than half (64%, n=745) were satisfied with the healthcare received, and felt that the service and care received had helped overcome their health problem (65%, n=755).

Table 3.  Where treatment was sought (multiple responses allowed).
Place of TreatmentResponses
  1. Note: Participants were allowed multiple answers, thus the total number of places where treatment was sought is larger than the number of participants (n= 1199)

Primary health centre598 (49.87%)
Pharmacy/Dispensary93 (7.76%)
Private doctor28 (2.34%)
Traditional healer370 (30.86%)
Government hospital101 (8.42%)
Non-government health facility92 (7.67%)
Self-medicated11 (0.92%)
Shop/market2 (0.17%)

Barriers to health service utilisation

Out of the respondents who had not sought treatment (15% of the overall sample, n=176), the majority (n=138) identified various barriers to health service utilisation. The most frequently reported barrier was that respondents did not feel the need to use health services. Geographical distance and transportation difficulties were the other barriers to health service use. Service providers related barriers such as service unavailability; a lack of trust in the quality of the available service; a lack of awareness of available services; and frequent barriers of direct and indirect costs. A lack of family support was also identified as a barrier by 5% of respondents who had not sought treatment (see Figure 2).

Figure 2.

Reported barriers to the uptake of health services.

Discussion

These findings need to be interpreted against the background of a lack of objective verification of individual self-assessed health status. Other studies in Southeast Asia have found self-assessed health to correlate well with objective assessments by physicians.15,16 The results of this study strongly reflect the current health priorities of Timor-Leste's population. Malaria and dengue treatment, food security and access to water and sanitation are urgent needs which have been and remain high priority issues for the people of Timor-Leste. The study documented significant challenges in terms of resource allocation, planning and programming.14

Patterns of healthcare utilisation and reported barriers highlight the need for comprehensive primary healthcare services that are easily accessible and designed for Timor-Leste's very difficult mountainous terrain. The lack of trust in health services and service quality and the direct and indirect costs associated with healthcare utilisation need to be considered when planning further services. Previous studies undertaken in Timor-Leste have documented that there is a preference for traditional healers and traditional medicines5, thus providing training for traditional healers and integrating them into comprehensive primary healthcare services especially in remote rural areas could address shortages in healthcare provision through a culturally and contextually appropriate mechanism.

The important association between health and educational levels highlights the need of addressing the very low literacy and education levels in Timor-Leste. Whether this is undertaken through an adult literacy program or a community health literacy campaign needs to be determined against the background of feasibility, human capacity and resource availability. Research in other resource poor countries indicates that low levels of education and low socio-economic status are inter-dependant7 and impact over-all health and well-being. Migrants, as well as those with low levels of education need special attention when implementing health interventions and programs as they are significantly more vulnerable and often hard to reach. Further research is needed to establish whether the health problems encountered by migrants can be addressed by services and programs particularly tailored to the needs of a mobile population.

Conclusion

Timor-Leste has some of the worst health indicators in the Asia-Pacific region. As has been documented in various other developing and resource poor countries, poverty must be addressed in conjunction with implementing health programs, in order to create a sustained increase in health status. Unfortunately, the political situation deteriorated considerably since the study took place, with the conflict re-occurring in 2006/07 causing displacement and an assassination attempt on the President and Prime Minister in early 2008. This makes further research and implementation of programs especially in the rural and remote areas extremely difficult. The conflict has led to further displacement and increased vulnerability of the population.

In regions of protracted conflict, with limited human capacity it is often difficult to implement effective long-term policies. Nevertheless, the Government of Timor-Leste is committed to improving the health and literacy of its people. The Ministry of Health needs co-ordinated international support to overcome its health problems, of which high morbidity and mortality characterise some of the most pressing needs. To address the immense health needs in this tiny nation, the Ministry of Health will need to develop and implement a National Health Policy that will understand the complex cultural mechanisms in areas of protracted conflict and develop partnerships between the government, local communities and international donors that are sustainable. The international community will need to support Timor-Leste in the effective implementation, monitoring and evaluation of a nation-wide health policy specifically aimed at improving over-all public health, mortality and morbidity.

Acknowledgements

This research was part of a larger project funded through UNAIDS funds and administered by the International Organisation for Migration [IOM – Timor-Leste Mission]. The authors would like to acknowledge the invaluable contribution of the participants made by volunteering their time in responding to the survey. We acknowledge the local research team, the support of the IOM Timor-Leste, Chief of Mission, Mr Luiz Vieira and the IOM – Dili Mission staff. We thank Dr Richard Woodman for his assistance with statistical data analysis.

Ancillary