Timor-Leste_Health_Worker_Survey_-_Final_Report_14June15.pdf

June 2015

Health Worker Survey in Timor-Leste Final Report

This project was carried out by World Bank and Oxford Policy Management (OPM). Dili Institute of Technology (DIT) was the local partner. The Ministry of Health of the Government of Timor-Leste provided support at every stage of this project.

If you have any queries please contact Xiaohui Hou, Senior Economist, World Bank (xhou@worldbank.org) or Rashid Zaman, Consultant, Health Portfolio, Oxford Policy Management (rashid.zaman@opml.co.uk).

Full citation: World Bank and OPM (2015) “Health Worker Survey in Timor-Leste”, Dili: World Bank and Oxford: Oxford Policy Management (OPM).

Oxford Policy Management Limited 6 St Aldates Courtyard Tel +44 (0) 1865 207300

38 St Aldates Fax +44 (0) 1865 207301

Oxford OX1 1BN Email admin@opml.co.uk

Registered in England: 3122495 United Kingdom Website www.opml.co.uk

© World Bank and Oxford Policy Management i

Preface

This report presents the results of the Health Worker Survey in Timor-Leste. The survey collected information on labour market dynamics, health worker preferences and concerns, and the skills, competence and motivation of doctors working in the public sector. The overall goal is to provide evidence for policy guidance on the health workforce in Timor-Leste.

If you have any queries or need any clarifications please feel free to contact us using the information at the bottom of this page.

mailto:xhou@worldbank.org mailto:rashid.zaman@opml.co.uk

© World Bank and Oxford Policy Management ii

Acknowledgements

This report was produced and written by a task team from Oxford Policy Management (OPM) consisting of Rashid Zaman, Kay Engelhardt, Marc-Francois Smitz, Firdaus Hafidz, Sophie Witter and Tomas Lievens under the World Bank task team leader Xiaohui Hou (Senior Economist).

World Bank staff including Christophe Lemiere (Senior Health Specialist), Eileen Sullivan (Senior

Operations Officer for Health), Sarah Harrison (Health Consultant), Sara Maria Pereira

(Consultant), Jennifer Shin (Consultant) and Cornelio Quintao de Carvalho (Team Assistant) provided technical inputs and logistical support.

Toomas Palu (Practice Manager, Health, Nutrition and Population Global Practice) 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 Bolormaa Amgaabazar (Country Manager for Timor-Leste) for their guidance and support.

The authors would like to thank H. E. Maria do Céu Sarmento Pina da Costa (Minister of Health, Timor-Leste and H.E. Ana Isabel Soares (Vice-Minister of Health, Timor-Leste) for their overall supervision and guidance.

The authors thank the staff members and external consultants of OPM including Juan Munoz, Katie McIntosh, Sourovi De, Mahmud Hassan and Wirya Wardaya for their support in sampling, analysis, survey management, data-entry programming and data management. The team would also like to thank Prima Ratrikaningtyas, Sutarman and Lukman Chandra, the medical researchers from the University of Gadjah Mada who joined the field teams. Dili Institute of Technology (DIT) was the local partner for this survey and the team included Estanislau Saldanha, Fernanda Julia, Afonso Almeida, Alvaro Amaral, Calisto Godinho Quiteria Fernandes, Joao Silva, Tiago Soares, Valenti Martins, Diamantino Soares, Francisco Godinho, Palmira Lemos and Jeferinu Ulan. The members of the data management team were Octovianus Pereira, Roberto Silva, Haryanto

Santoso, Raul Amaral, Natalino Costa, Maria Amaral and Ticiana Caldas.

The team would like to thank the Ministry of Health (MoH) officials who provided important support at various stages of this project, including Jose dos Reis Magno (Director General), Duarte

Ximenes (National Director, Human Resources and Administration), Marcelo Amaral (National

Director, Finance and Procurement), Amelia Barreto (Chief, Department of Quality Control), Lourenco Camnahas (Project Manager, National Health Sector Strategic Plan), Celeste Cham

(Head, Department of Pharmacy), Odete Belo (President of Ad Hoc Commission, SAMES) and their colleagues.

The report benefited from comments from Augustine Asante (Research Fellow, University of New

South Wales), Joao Martins (Dean, Faculty of Medicine, University of Timor-Leste), Andre Cote

(HR Specialist, National Health Sector Strategic Plan), Ian Morris (Consultant, World Bank), Sanjoy

Nayak (Human Resources for Health Adviser, National Health Sector Strategic Plan) and peer reviewers Akiko Maeda (lead health specialist, World Bank), Christopher Herbst (Health Specialist, World Bank) and Armandina Gusmao-Amaral (Senior Health Officer, the Australian Department of

Foreign Affairs and Trade). The team would also like to thank development partners and non-governmental organisations including the Department of Foreign Affairs and Trade (DFAT) of the

Australian Government, the European Commission (EC), the World Health Organization (WHO), UNFPA, USAID, Marie Stopes and Health Improvement Project for their feedback during the stakeholder consultation workshop and individual consultations. Financial support for this work came from the Multi-Donor Trust Fund, which received funding from the Australian Government and the European Commission.

© World Bank and Oxford Policy Management iii

Executive Summary

Introduction

At the time of independence in 2002, Timor-Leste had a seriously weak health system with only a handful of doctors in the country. In this context, the governments of Timor-Leste signed an agreement and the Cuban Medical Brigade started to train medical students and deploy them in the country, particularly in rural areas. While the initial massive shortage has now been minimised, there are concerns over more complex issues including facility functionality, rural retention, motivation, preferences and competence of health workers. The objectives of this survey were to understand the labour market dynamics among health workers, to learn more about the preferences and concerns of health workers, and to assess the skills, competence and motivation of doctors.

Methods

This survey included four components: a health facility survey, a health worker survey, a discrete choice experiment (DCE) and direct clinical observations (DCO). A qualitative study was carried out prior to the main study. The health facilities and health workers were sampled using probability proportionate to size (PPS). Three field teams collected the quantitative survey data between July and August 2014. Data were double entered using a customised data-entry program. Associations between various predictors and outcome variables are presented through descriptive statistics and statistical tests. Further work could be done in future to examine the relationships between the three main datasets.

The study has several limitations. One of them is the lack of demand-side information; the sample only included the health facilities and health workers and did not include any patients. To mitigate this limitation the survey methodology and instruments were carefully designed so that the results would not be biased based on health workers’ self-reporting on functionality or performance.

Private health providers are not included in this survey, which is another important limitation.

Private health care providers provide about one-quarter of the health services in Timor-Leste;

including them may have increased the comprehensiveness of the survey. However, the primary goal of this survey was to inform policy-makers on government health workers.

Findings

Labour Market Dynamics: Findings from the Health Worker Survey

The general health worker survey was administered to 443 health workers, including 175 doctors, 150 nurses and 118 midwives (about 20% of the health workers in the country). The majority (56%) of health care workers were women, including 58% of the doctors and all of the midwives. The only exception was nurses, where the majority (64%) were men. Close to 53% of all doctors interviewed were posted in rural facilities, while nurses were more concentrated in urban areas (70%). Twenty-one percent of doctors were working in Health Posts (HP), compared to 8% of nurses and 16% of midwives. Since there were only a few doctors in country in Timor-Leste at the time of independence, almost the entire cohort of doctors was newly trained; most of the doctors (96%) interviewed in the survey had less than five years of experience in the sector. The survey shows that health workers got opportunities to advance to higher types of facility with more experience.

Motivation to join and long-term preferences: The vast majority of staff selected medicine as a career in order to help people and, generally health workers indicated high levels of satisfaction:

© World Bank and Oxford Policy Management iv only 4% of the respondents indicated that they were ‘unsatisfied’ or ‘very unsatisfied’ with their work. Intrinsic motivation is also shown by the fact that eight in ten medical staff said that they would stay in the facility until the last patient was treated – even if they did not receive additional money for this.

In the long run the majority of respondents would like to continue to work in the government/public sector, including 99% of the doctors. This was also the case for the broad majority of nurses and midwives, although 6% of nurses would prefer to move to the private sector. HPs were the least attractive stations for medical staff, with only 6% of doctors stating a HP as their long-term preferred facility, although 32% were content to work in Community Health Centres (CHC). Only

22% saw themselves working in rural sucos long term. However, 88% acknowledged that the MoH would determine where their next assignment would be. The majority of respondents (97%) were not looking for another job in the short term.

Salary and financial benefits: Wage differentials within each cadre were relatively small, and income did not vary much by the years of experience, particularly for the doctors. On average, a doctor with more than ten years of experience earned just US$ 50 more than a newly joined doctor. More than half of the health staff interviewed believed their salary to be too low. However, they did not see other labour market opportunities that they would prefer. All medical staff receive their money through a direct deposit to their bank account and only very few (2%) have experienced any delays in receiving their money. Very few report private practice (but this may be under-reported). More non-financial benefits (such as housing and motorbikes) were reported by doctors and at lower-level facilities. However, only half of them reported that they get either sufficient fuel or funds to buy fuel for the motorbike, and more than half of all respondents also stated that these benefits were delayed.

Workload: Although most respondents worked five days a week in HPs, CHCs and district/regional hospitals (63%, 66% and 54%, respectively), 22% to 36% worked six and sometimes seven days a week. The number of patients seen by the sampled doctors who participated in the DCOs was counted; the mean number of patients per day was 10.2, with the standard deviation (SD) of 7.5. In urban facilities the mean (SD) patient load was 11.5 (6.9) compared to 9.6 (7.8) in rural facilities. The number of patients also varied by level of facility, with higher tier facilities having more patients. Forty-five percent of respondents agreed with the statement that they have ‘too much work to do’. However, we found no clear correlation between workload and feelings of overload.

Training: Almost all respondents believed that their training adequately prepared them to diagnose and treat clinical cases in Timor-Leste. When asked how many short-term training sessions (below

30 days) they had attended in the last three years, roughly half of the nurses and midwives reported doing three or more, while one-quarter did not attend any such training in this timeframe.

At the same time, roughly one-third (35%) of the doctors attended three or more training sessions and around the same proportion (37%) attended none. Despite all the training opportunities, roughly half (52%) of respondents stated that ‘there is not enough opportunity to learn current medical knowledge’.

Although a large number of doctors were in rural areas, it is interesting to see that only 19% underwent training on community health. Around 75% of doctors indicated that they require training on Integrated Management of Childhood Illness (IMCI) and 64% on Emergency Obstetrics and

Newborn Care (EmONC). The doctors located in urban areas were found to benefit more: roughly half (51%) of them had attended three or more training sessions in the last year, as opposed to

26% in rural areas.

© World Bank and Oxford Policy Management v

Older or highly satisfied doctors tended to be less interested in any kind of training, while females were keener than males to have the opportunities for visits from specialists. Workers in HPs and

CHCs were more interested in specialisation. Highly satisfied doctors had a lower utility for all kinds of training. Older nurses and midwives, as well as those with more medical experience, tended to be more interested in any kind of training. Rural facility health workers were significantly more willing to finish their bachelor’s degrees.

Supervision: Eighty-five percent of respondents indicated that they have a supervisor who is responsible for providing feedback on their performance. The majority of staff had supervisory meetings at least every three months, with urban-based staff having more meetings. Supervisors’ activities were heavily biased toward activities that support technical staff development and quality control: supervisors observed the consultation, provided health instructions and asked knowledge assessment questions. Checking records, an administrative task, topped the list of supervisory tasks. There were no important differences between rural and urban facilities. 75% of all respondents stated that they felt the need to discuss difficulties with their supervisor within the last year; most of them (95%) actually discussed the issue with their supervisor and 66% noticed improvements afterwards.

Challenges: Low salaries were identified as a problem (with 63% agreeing or strongly agreeing), followed by inadequate opportunities to learn (52%), lack of transport to see patients (50%), inadequate housing (48%), too much work (47%), security problems (39%), lack of supervision

(30%), lack of feedback on performance (23%), and lack of motivation (20%).

Absence from work: Eight percent of respondents reported being absent from work in the thirty days before the interviews due to sickness. The percentage was the same across all types of medical staff. Five percent of staff was absent in that timeframe because of personal reasons. In the event of an absence, roughly 13% of respondents stated that they were called by the facility head. In 5% of the cases, the supervisor discussed the absence with staff and 2% indicated that money was deducted from their salary as a reaction.

Working Environment: Findings from the Health Facility Survey

Sixty-nine health facilities in all thirteen districts of Timor-Leste, including six hospitals, thirty-three community health centres (CHCs) and thirty health posts (HPs) were visited. The data suggest that the national and referral hospitals were generally well equipped and that CHCs were moderately equipped. However, most of the HPs were poorly equipped and lacking basic facilities and services. The facilities located in rural areas were also considerably under-resourced compared to facilities in urban areas. The storage conditions of medicines were found to be of concern, especially in rural areas.

Service availability was typically high in most of the areas except for treatment of HIV/AIDS, caesarean section and blood transfusion. The HPs offered fewer services than CHCs and hospitals.

Health Worker Preferences: Results from the DCE

In the DCE the preferences of health workers were explored by giving them the option to choose hypothetical jobs with different attributes. The findings revealed that the ‘probability of specialisation’ was the top-ranked most important factor for doctors. Visits from specialists, the availability of equipment, good housing, working in higher-level facilities and urban locations were the other significant attributes (by order of importance). Newly graduated doctors were totally neutral toward wages. For nurses and midwives, in-service training was valued most highly, followed by transportation, equipment, remote location and housing.

© World Bank and Oxford Policy Management vi

Older doctors valued wages positively and the same was true for males, while females’ coefficient was negative and strong (-0.003 per dollar). Nurses and midwives valued a higher wage and factors associated positively with it were age, experience, being male, being a nurse (compared to a midwife), being married and owning a house, being in the same district as one’s family, working in a CHC in a rural area, not being provided with housing and having a relatively lower wage and working in a health facility with a better supply of drugs and equipment.

Nurses and midwives would be ready to give up US$ 194 per month to access workshops, and

US$ 113 to have a motorbike. Having medium scale of equipment availability ranked third place with a value of US$ 102.

Competence of Doctors: Results from DCOs and Vignettes

In total, 635 cases were observed, mostly in outpatient departments, with 40% in urban and 60% in rural areas. The vignettes focused on patients presenting with cough, fever and diarrhea.

Knowledge and practice scores: The overall clinical performance, as measured by DCO, of general practitioners was very good in terms of attitudes (91% score) and moderate at history taking (57%), health education (50%) and treatment accuracy (69%). However, the average physical examination performance score was 28%. Lack of examination tools was suspected to be the one of the reasons they could not perform physical examinations as well as expected; some healthcare facilities did not even have any examination bed, blood pressure monitor

(sphygmomanometer) or thermometer. The overall average performance score was 62%.

When comparing observations with vignettes (to assess knowledge), there were significant differences in history taking, appropriateness of treatment and health education. The average scores for history taking and health education were higher in the vignette group, suggesting a

‘know–do’ gap. However, scores on the appropriateness of treatment were much better for DCOs than vignettes, which probably means that vignette cases were harder than daily practice cases.

One worrying finding for both observations and vignettes was the low rate of hand-washing – slightly above 20% in observed cases and below 10% for simulated vignettes. Absence of facilities cannot explain the low rate in the vignettes.

Determinants of clinical performance: HPs had a negative correlation with clinical performance when the clinical performance was distinguished by the type of health facility. As HPs were found to be comparatively poorly equipped, doctors might face difficulties in performing to ideal clinical practice. The location of the practice also had a significant effect on the clinical performance of doctors; doctors working in a rural area were less skilled than their counterparts in urban areas, which again may link to the absence of an enabling work environment.

The clinical performance was also better when consultation periods were longer. The average consultation time per case was nine minutes. Doctors spent less time in DCOs (eight minutes) compared to vignettes (ten minutes), which was also statistically significant (p= <0.01). Conducting the vignettes after work may have permitted this longer focus on each case.

Conclusion

The study highlights some encouraging findings, including the gender balance of health staff

(overall, if not within specific cadres), the concentration of doctors in rural areas, the high overall reported satisfaction of staff with their work and high intrinsic motivation, the positive intention to stay in the public sector, the feeling of being well prepared by training for work, the relatively

© World Bank and Oxford Policy Management vii frequent and satisfactory supervisions reported, and the good attitudes to patients identified in observations and vignettes. However, some areas require more investigation and investment.

The following recommendations are based on the evidence gathered from this survey:

Ensure Policy Compliance at All Levels

1. Service availability should match the Basic Service Package (BSP): The survey found that the health facilities, particularly the rural ones often lacked services and facilities that they are supposed to have according to the BSP. Ensuring policy compliance to BSP will not only maintain the consistency of services across the sector, but will also improve the functionality of the rural health facilities which will eventually encourage health workers to stay motivated to work in rural areas.

2. Availability of medicines should comply with the Essential Drug List (EDL): The drug supplies were frequently inconsistent with the EDL recommendation. Complying with the EDL may improve service delivery by reducing the chance of stock-outs and misuse of drugs. However, it is possible that there is a functioning decentralised mechanism in place, which is effective, but not fully compliant to the central polity. This should be investigated before implementing this recommendation.

3. Clinical guidelines should be enforced: Enforcing clinical guidelines may increase the performance of the doctors. However, there are few related issues. First, changing practice usually requires a lot of effort. Second, since some health providers may lack elements of basic knowledge, the clinical guidelines, which are mostly practice oriented, may not be fully effective. And finally, the clinical guidelines in all relevant topics might not be available or updated.

Strengthen the Health Human Resource Policy

4. Ensure salary progression for doctors: The data revealed that the salary difference between the junior and senior doctors was very narrow. This may demotivate experienced health workers and cause them to leave the public sector or engage in dual practice in future. A clear salary progression policy will keep health workers motivated to serve in the public sector and will also encourage better performance.

5. Provide necessary training for health workers: The overall low score on clinical competence and the narrow know-do gap indicates doctor’s lack of knowledge in Timor-

Leste, especially in rural areas. Regular in-service training and visits from specialists are highly recommended to increase health worker’s competence. This strategy will also improve the confidence of young health workers and motivate them to stay in rural areas.

6. Improve the effectiveness of supervision: While the supervision visits were frequent, in most cases these visits were mostly focussed on checking official records. More efficient supervision and monitoring tools and exercises are recommended so that these visits are more meaningful and action oriented.

7. Adjust the expectations of health workers on their workload: Nearly half of the respondents reported to have a high workload. However, the average number of patients consulted, as observed during the survey, did not reflect that statement. It might be useful to better understand the reason behind this statement and adjust the workload expectation by establishing clear terms of reference for the work.

© World Bank and Oxford Policy Management viii

Improve the Functionality of Health Facilities

8. Improve the pharmaceutical chain of procurement and distribution: The survey results indicate a loosely functioning pharmaceutical chain with an incoherent distribution mechanism. A strong, demand-driven pharmaceutical procurement and disbursement system in which demand will be raised in a bottom-up process along with a sequential distribution mechanism is recommended.

9. Improve availability of medical supplies: The rural health facilities were frequently under-equipped and many health workers expressed this as one of the important attributes in deciding workplace. Improving availability of medical supplies in health facilities will improve the service delivery and at the same time will enhance the rural retention of the health workers.

10. Improve infrastructure and the access to water, electricity and communication:

Many rural facilities lacked basic infrastructure including access to water, electricity and communication. Improving the infrastructure will optimise service delivery and rural health worker retention. An inter-sectoral initiative, involving the other concerned ministries, is recommended to identify and prioritize potential development areas, and improve the infrastructure of health facilities, including the accommodation of health workers.

Further studies on assessing the training gaps, migration and exit to private sector, absenteeism and dual practice, pharmaceutical distribution, and location preference (urban vs. rural) of health workers will broaden the understanding on some of the survey findings. A follow-up survey, using similar methodology to monitor the trend over time is recommended.

This survey highlights some important policy areas that can be addressed to better manage the health workforce in Timor-Leste and ensure better health service to the population, particularly in rural areas. The Ministry of Health can consider prioritising the above mentioned recommendations and group them into short, medium and long term initiatives.

© World Bank and Oxford Policy Management ix

Table of Contents

Preface i

Acknowledgements ii

Executive Summary iii

Introduction iii

Methods iii

Findings iii

Conclusion vi

List of Tables and Figures xi

List of Abbreviations xiii

1 Introduction 1

1.1 Background 1

1.2 Health Systems 1

1.3 Health Workforce 2

1.4 Policy Issues 4

1.5 Objectives 5

2 Methods 6

2.1 Components 6

2.2 Sampling 7

2.3 Survey Implementation 8

2.4 Data Analysis 10

2.5 Ethical Considerations 11

2.6 Methodological Limitations 11

3 Labour Market Dynamics: Findings from the Health Worker Survey 13

3.1 Characteristics of the Sampled Health Workers 13

3.2 Study Choice and Long-Term Career Preferences 17

3.3 Workload 21

3.4 Salary and Financial Benefits 23

3.5 Training 26

3.6 Supervision 29

3.7 Challenges 32

3.8 Absence from Work 34

4 Working Environment: Findings from the Health Facility Survey 37

4.1 Characteristics of Health Facilities 37

4.2 Service Availability 38

4.3 Infrastructure and Equipment 39

4.4 Drugs and Supplies 41

5 Health Worker Preferences: Findings from the Discrete Choice Experiment 44

5.1 Setting-up the Experiment 44

Concept 44

5.2 Descriptive Statistics for Doctors, Nurses and Midwives 45

5.2.1 A Bias Against Job A? 48

5.3 Multivariate Analysis for Doctors 49

© World Bank and Oxford Policy Management x

5.3.1 Basic Results 49

5.3.2 Multivariate Analysis 52

5.3.3 Modelling 54

5.4 Multivariate Analysis for Nurses and Midwives 55

5.4.1 Basic Results 55

5.4.2 Multivariate Analysis 57

6 Competence of Physicians: Findings from the Direct Clinical Observations and Vignettes 60

6.1 Characteristics of the DCO and Vignette Cases 60

6.2 Doctors’ Characteristics 61

6.3 Clinical Performance 61

6.4 Attitude 63

6.5 History Taking 64

6.6 Physical and Diagnostic Tests 67

6.7 Health Education 70

6.8 Determinant Factors of Clinical Performance 70

7 Conclusions 73

7.1 Policy Implications 73

7.2 Recommendations 76

7.3 Further Studies 77

7.4 Way Forward 78

References 79

Annex A Single Control Regressions for DCE 81

A.1 Single Control Regressions for Doctors 81 A.2 Single control regressions for nurses and midwives 84

Annex B Regression Models of DCO 87

B.1 Regression Model of Total Score Clinical Performance 87 B.2 DCO Regression Model of Total Score Clinical Performance 88

Annex C Additional Information on DCE Analysis 89

© World Bank and Oxford Policy Management xi

List of Tables and Figures

Table 1: Health Cadres in Timor-Leste Table 2: Proposed Sample Table 3: Distribution of Sampled Health Workers Table 4: Years of Experience in the Medical Sector by Occupational Group Table 5: Total Experience in the Medical Sector by Facility Table 6: Number of Facilities Worked in During Career by Occupational Group Table 7: Current Facility by Occupational Group Table 8: Long-Term Goals by Occupational Group Table 9: Number of Patients per Doctor per Day, by Types of Facility Table 10: Number of Benefits Received, by Occupational Group Table 11: Kinds of Benefit Received, by Occupational Group Table 12: Provision of Supplies for Motorbikes and their Timeliness Table 13: Respondent’s Share in Household Income Table 14: Money Available per Household Member After Expenses per Month Table 15: Number of Short-Term Training Sessions Undertaken in the Last Three Years Table 16: Number of Long-Term Training Sessions Undertaken in the Last Three Years Table 17: Training Sessions Attended in the Last Year, by Locality and Type of Facility Table 18: Supervisor by Occupational Group, Locality and Facility Type Table 19: Frequency of Meetings with Supervisors Table 20: Health Workers’ Perceived Problems – Average Scores Table 21: Number of Urban and Rural Facilities, by Facility Type Table 22: Mean and SD of Different Types of Health Worker, by Facility Type Table 23: Availability of General Services, by Facility Type (%) Table 24: Readiness for Disabled People and Gender Equity, by Facility Type Table 25: Water and Power Supply, by Facility Type Table 26: Infection Control Materials in Client Examination Rooms, by Facility Type Table 27: Availability of Antibiotics and Antimalarials, by Facility Type Table 28: Availability of Supply Items, by Locality Table 29: DCE Sample Table 30: Proportion of Health Workers Choosing Job A Over Job B, by Choice Set Table 31: Percentage of Job Selection per Level of Attribute: Medical Doctors Table 32: Percentage of Job Selection per Level of Attribute: Nurses and Midwives Table 33: Logit Fixed Effect Table 34: Marginal Rates of Substitution for Doctors Table 35: Urban MRS for Doctors Table 36 Control Variable Descriptive Statistics Table 37: Logit Fixed Effect (Nurses and Midwives) Table 38: MRS for Nurses and Midwives Table 39 Urban MRS for nurses and midwives Table 40: Non-Extremely Remote MRS for Nurses and Midwives Table 41: Nurses and Midwives: Control Variable Descriptive Statistics Table 42: Characteristics of the DCO and Vignette Cases Table 43: Patients' Primary Symptoms Table 44: Medical Doctors’ Characteristics Table 45: Different Scores Between DCOs and Vignettes Table 46: DCO Scores Table 47: Vignette Scores Table 48: Scores by Type of Facility Table 49: Statistical Association Between Non-Quality and Performance Attributes Table 50: Proportion of Job A Selection, by Type of Health Workers Table 51: Proportion of Choices That Were Rejected – by Type of Health Workers Table 52: Proportion of HW with Item Dominance Table 53: Regression Results Table 54: Regression Results for Nurses and Midwives

© World Bank and Oxford Policy Management xii

Figure 1: National Health Service Configuration of Timor-Leste Figure 2: Occupational Type by Gender Figure 3: Occupational Type by Age Group Figure 4: Types of Medical Staff by Locality Figure 5: Professional Experience of Respondents, by Locality Figure 6: Time Spent in Rural Areas by Occupational Group Figure 7: Satisfaction with Current Job Figure 8: Long-Term Preference of Doctors for Posting (by Current Work Location) Figure 9: Top Ten Statements on Motivation in Terms of Agreement Figure 10: Other Motivations Figure 11: Working Days per Facility Type Figure 12: Working Hours per Day per Locality Figure 13: Average Monthly Income by Occupation Type Figure 14: Average Monthly Income by Years of Experience Figure 15: Training Sessions Undertaken by Doctors Figure 16: Additional Training Required by Doctors Figure 17: Time of Last Meeting with Supervisor Figure 18: Supervisor Activities in Rural Areas Figure 19: Supervisor Activities by Locality Figure 20: Problems in the Current Job Figure 21: Personal Reasons for Absence from Work Figure 22: Availability of General Services, by Locality Figure 23: Communication Facilities, by Facility Type Figure 24: Privacy of Client Examination Rooms, by Facility Figure 25: Storage Condition, by Facility Type Figure 26: Average Clinical Performance Score by DCO and Vignette Figure 27: Doctors Attitude and Communication Figure 28: Fever-Related History Taking Figure 29: Cough-Related History Taking Figure 30: Diarrhea-Related History Taking Figure 31: Physical Examinations Figure 32: Fever-Related Physical Examinations and Diagnostic Tests Figure 33: Cough-Related Physical Examinations and Diagnostic Tests Figure 34: Diarrhoea-Related Physical Examinations and Diagnostic Tests Figure 35: Health Education Provided to DCO and Vignette Cases Figure 36: Non-Quality vs. Lack of Knowledge Figure 37: Non-Quality vs. Lack of Motivation Figure 38: Non-Quality vs. Lack of Equipment and Supplies

© World Bank and Oxford Policy Management xiii

List of Abbreviations

CHC Community Health Centre

DCE Discrete Choice Experiment

DCO Direct Clinical Observation

DFAT Department of Foreign Affairs and Trade

DHS District Health Service

DIT Dili Institute of Technology

EC European Commission

EmONC Emergency Obstetrics and Newborn Care

FGD Focus Group Discussion

HCW Health Care Worker

HF Health Facility

HP Health Post

ILO International Labour Organization

IMCI Integrated Management of Childhood Illness

ISCO International Standard Classification of Occupations

KII Key Informant Interview

MoH Ministry of Health

MRS Marginal Rate of Substitution

OPM Oxford Policy Management

PPS Probability Proportionate to Size

SD Standard Deviation

SISCa Servisu Integrado du Saude Comunidade

STI Sexually Transmitted Infection

TB Tuberculosis

ToR Terms of Reference

UNTL National University of Timor-Leste

WHO World Health Organization

© World Bank and Oxford Policy Management 1

1 Introduction

At the time of independence in 2002, Timor-Leste had a seriously weak health system and human health workforce with only a handful of doctors practicing in the country. In this context, the governments of Timor-Leste signed an agreement and the Cuban Medical

Brigade started to train medical students and deploy them in the country, particularly in rural areas. While the initial massive shortage has now been minimised, there are concerns over more complex issues including facility functionality, rural retention, motivation, preferences and competence of the health workers. The objectives of this survey were to understand the labour market dynamics among health workers, to learn more about the preferences and concerns of health workers, and to assess the skills, competence and motivation of doctors.

1.1 Background

Timor-Leste had a fragile health system when it became independent in 2002. During the war preceding independence more than 70% of the country’s health facilities were destroyed or seriously damaged and only approximately 20 doctors remained in the country (Cabral, et al., 2013). This radical shortage of health workers gradually started to improve as medical doctors trained in Cuba by the Cuban Medical Brigade began to return to Timor-Leste in 2004 (Asante, et al., 2012). However, there is still a shortage of health workers in all categories, including general doctors, specialists, nurses and midwives.

In 2012, the health worker density in Timor-Leste was 1.3 per 1,000 population (Asante, et al., 2014). This ratio was below the threshold of 2.2 which was indicated by WHO as the minimum to have desired level of coverage of skilled birth attended – a proxy indicator widely used to understand the density of health workforce(WHO, 2006). With the recent, large influx of doctors trained by the Cuban Medical Brigade the critical shortage of health workers has lessened, but more complicated issues that are related to human resource management, retention in rural facilities and the quality of health workers have emerged. In addition, there are concerns over the readiness-to-serve and functionality of health facilities.

1.2 Health Systems

The people of Timor-Leste have the constitutional right to health, medical care and a healthy environment. According to Article 57 of the Constitution, ‘All Timorese citizens are entitled to health care and the State has a duty to promote and protect this right free of charge, in accordance with its capabilities and in conformity with the law’ (MoH, 2012).

Timor-Leste has a three-tiered government health system, which includes a National Hospital at the tertiary level, district hospitals at the secondary level, and CHCs, HPs and outreach centres known as Servisu Integrado du Saude Comunidade (SISCa) at the primary level (MoH, 2012).

According to the information provided by the MoH, there were five district hospitals, 66 CHCs, 205

HPs and 442 SISCas in country at the time of the survey. However, there are reports that not all of the SISCas were fully functioning (Martins & Trevena, 2014).

© World Bank and Oxford Policy Management 2

Figure 1: National Health Service Configuration of Timor-Leste

Source: Human Resource for Health, Timor-Leste (MoH, 2012)

The National Hospital in Dili is the only tertiary care facility in Timor-Leste. It has inpatient, outpatient and emergency care facilities. It also has specialised services and enhanced laboratory facilities. The district referral hospitals also have outpatient, inpatient and emergency services but these hospitals only provide specialist care in four clinical areas: medicine, surgery, paediatrics and obstetrics-gynaecology. There are thirteen districts in Timor-Leste and five of them have a referral hospital. The CHCs offer outpatient and inpatient care. Some CHCs are larger and have dentistal and laboratory facilities. The HPs are the first level of health facilities and only provide outpatient care. The SISCas are monthly satellite clinics run by health staff from the HPs (MoH, 2012). The MoH plans to deploy three doctors in each CHC and one doctor in each HP (Asante, et al., 2014).

The private health system is not very well developed in Timor-Leste. In 2011, there were twenty-six private health facilities in Timor-Leste, which covered about one-quarter of the country’s basic health service delivery, mostly in urban areas (Asante, et al., 2012). Government health workers are officially prohibited from private practice in Timor-Leste.

1.3 Health Workforce

There are thirty-six health cadres from eight categories in government health services in Timor-

Leste, as shown in Table 1.

Table 1: Health Cadres in Timor-Leste

Category Sub-category Cadre

Medical practitioners General practitioners General medical practitioner

Medical specialists Medical specialist

Dental practitioners

Dentists Dentist SI

Dental nurses and technicians

Dental Nurse DIII

Dental Nurse SPRG

Dental Lab Technician DI

Pharmacy practitioners Pharmacists Pharmacist DIII

© World Bank and Oxford Policy Management 3

Category Sub-category Cadre

Pharmaceutical technicians/ assistants

Junior Pharmacy Technician (D1)

Pharmacy Technician (SMF)

Nursing and Midwifery practitioners

Nursing professionals

Nurse SPK/DI

Nurse D III

Nurse SI

Midwifery professionals

Midwife DI

Midwife D III

Midwife D IV

Non-medical public health practitioners

Public health generalists Public Health Generalist

Public health specialists Public Health Specialist

Nutrition professionals Nutritionist D I

Nutritionist D III

Environmental health officers

Environment Health Officer DI

Environment Health Officer DIII

Environment Health Officer SI

Community health volunteers Promotores Saude Familia (PSF)

Medical technologists

Medical imaging technicians

Radiographer DI

Radiographer DIII

Radiographer SI

Electromedik DIII

Laboratory assistants

Laboratory Assistant SMAK/SPK

Laboratory Assistant DI

Laboratory Assistant DIII

Laboratory Assistant SI

Other health workers Optometrists Optometrist DI

Physiotherapists Physiotherapist DIII

Health management and support staff

Health service managers Health Service Manager

Medical records technicians Medical Records Technician

Support staff Support staff

Source: Human Resource for Health, Timor-Leste (MoH, 2012)

Although, the International Standard Classification of Occupations (ISCO) of the International

Labour Organization (ILO) categorises health professionals as ‘health professionals’ and ‘health associate professionals’ (ILO, 2008), there are practical challenges and standardisation issues in including all types of professionals under this definition. The health workforce is often defined as the aggregated total of physicians, nurses and midwives. The World Health Report 2006 used the aggregated total of these three professional groups to calculate the density of the health workforce and to compare across countries (WHO, 2006). Doctors, nurses and midwives were considered health workers in this survey.

The majority of doctors in Timor-Leste are trained, either in Cuba or Timor-Leste, by the Cuban

Medical Brigade (CMB). The CMB started enrolling students in 2004. Until 2007, medical students were trained simultaneously in Cuba and Timor-Leste. Students enrolled in or after 2008 are trained in Timor-Leste by CMB in collaboration with the University of Timor-Leste (UNTL). The program is a six-year Bachelor of Medicine course consisting of five years of medical schooling and one year of internship in the health facilities across the countries. The important characteristics of the Cuban medical training program are a commitment to community spirit and serving the people and the medical doctors sign an agreement with the government to work in public health sector for at least six years. In addition to the CMB-trained doctors there are some doctors who are

© World Bank and Oxford Policy Management 4 trained in other countries, mostly in Indonesia. Most of the nurses and midwives are trained in either UNTL or in the Institute of Health Science (Asante, et al., 2014).

1.4 Policy Issues

Lack and quality of infrastructure and limited resources at health facilities are longstanding problems in Timor-Leste. According to a recent report in The Lancet, there is no Magnetic

Resonance Imaging machine in Timor-Leste and there is only one Computed Tomography scan machine, which is quite old (McCall, 2014). Facility functionality is worst in rural areas. A recent article reported that only half of the SISCas were functioning (Martins & Trevena, 2014). A scoping review and analysis of published literature concluded that access to and use of health services is one of the top priority areas for Timor-Leste. The review summarised the findings of the survey reports that 12% of households in Timor-Leste did not seek health care when a household member was ill and 25% of households were more than two hours away from their usual health care provider (Deen, et al., 2013).

In most developing countries, the retention of health workers in rural areas is a major challenge.

Timor-Leste is no exception and it also has some unique features. Various push and pull factors influence health workers’ choices to stay or not stay in rural health facilities. According to the strategic development plan of the Government of Timor-Leste, all HPs will have at least one doctor, two nurses and two midwives by 2020 (Government of Timor-Leste, 2011). Most of the

Cuban-trained medical students were enrolled from rural areas and therefore have some incentive to go back to their areas of origin. However, because of the poor functionality of the rural health facilities, doctors may be unwilling to work in those facilities. There are also concerns regarding a lack of supervision and unclear career development paths for those working in rural areas (Asante, et al., 2014).

While there is a strong political commitment to deploy health workers in rural areas, there are concerns over the utilisation of the services and the productivity of health workers posted in rural areas. A recent critical appraisal estimated that the patient load at the primary health care level in

Timor-Leste was on average less than three persons per day (Cabral, 2010). However, this finding is based on some assumptions and modelling and not backed up by rigorous primary data.

Since the literature suggests that the country’s rural facilities lack infrastructure and supplies, management function, supervision and opportunities to learn, there is a possibility that the health workers deployed in these facilities are not motivated. An article published on the tuberculosis (TB) services in Timor-Leste indicated that there was a lack of ownership by government health workers

(Martins, et al., 2006).

Finally, there are some concerns about the clinical competence and quality of care of the Cuban-trained doctors (Asante, et al., 2012). Although policy-makers believe that the competence of these doctors is adequate for the needs of the local health system, there have been no systematic independent reviews of the doctor’s performance in Timor-Leste (Asante, et al., 2014). A study of the use of medicine and adherence to the clinical guidelines by non-physician health workers (e.g.

nurses and midwives) in Timor-Leste was conducted and showed favourable results (Higuchi, et al., 2012). However, there is no evidence about doctor’s clinical performance that could support or assuage concerns regarding the competence of doctors in Timor-Leste.

All the above information indicates that there is a general lack of evidence in various aspects of the health labour market in Timor-Leste. This lack of information is mainly caused by the absence of an integrated information system and lack of properly designed research (Cabral, et al., 2013). A need for a properly designed sample survey on health workers in Timor-Leste was therefore

© World Bank and Oxford Policy Management 5 identified, not only to gather evidence on some of the complex issues like motivation, preference and rural retention but also to systematically assess the quality of the doctors working in the country.

1.5 Objectives

The primary objectives of this survey were to understand:

1. The labour market dynamics among health workers (especially the prevalence of dual job practice and the inflows/outflows between rural and urban areas);

2. The preferences and concerns of health workers, especially regarding their revenues and rural jobs; and

3. The skills, competence and motivation (i.e. the ‘know–do gap’) of doctors.

It is expected that this survey will provide important information and evidence to health policy-makers in Timor-Leste to facilitate better understanding of the incentives and constraints of health workers in the country.

© World Bank and Oxford Policy Management 6

2 Methods

This survey included four components: a health facility survey, a health worker survey, a

DCE, and DCOs. A qualitative study was carried out prior to the main study. The health facilities and health workers were sampled using PPS. Three field teams collected the quantitative survey data during July and August 2014. Data were double entered using a customised data-entry program. Associations between various predictors and outcome variables are presented through descriptive statistics and statistical tests Ethical clearance was obtained from both international and national institutional review boards. The important limitations of this study are the absence of data on demand side (i.e. user) perspective and not including the private health sector.

2.1 Components

The survey had four main components namely, general health facility survey, general health worker survey, discrete choice experiment (DCE) and direct clinical observations (DCO). A brief description of each of the four methods is outlined below.

General Health Facility Survey

This component aimed to gather basic information regarding the sampled health care facilities in order to understand the working environment. A structured questionnaire was administered to each of the sampled facilities.

The respondents of this survey were the head of the facility, the person in-charge of the facility, or the most senior health worker responsible for client services who was present at the facility at the time of the survey.

The questionnaire covered various dimensions of the health workers’ working environment, including inpatient service availability, processing of medical equipment, supplies and storage of medicines, user fees and sources of finance, sources of water and power supply, staff management and external supervision and disability and gender perspectives.

General Health Worker Survey

Data was gathered on health workers’ demographic characteristics and on various aspects on their profession. The field team members administered a structured questionnaire to the sampled health workers within the health facilities for this component.

The general health worker survey questionnaire included modules on job history, preferences and views on the profession, current job, training, supervision and absenteeism.

Discrete Choice Experiment (DCE)

A DCE was carried out to understand health workers’ preferences. Qualitative research was carried out to identify the job attributes health workers take into account when deciding on rural postings. DCEs draw on Lancaster’s theory of values, which states that each choice is composed of multiple attributes, also known as the choice characteristics of interest (Lancaster, 1966).

Assuming that the utility function of health workers follows the random utility theory, the observation of a large number of choices gives the relative preference of one attribute over the other across the sample of respondents, and is as such useful in terms of…