Sep 15, 2008 - communication (IEC) campaigns and post-distribution âhang-up campaignsâ need to ..... visits simply sleep in a hammock between two trees.
EVALUATION OF BEHAVIOUR CHANGE COMMUNICATION INTERVENTIONS AMONG MOBILE AND MIGRANT POPULATIONS IN WESTERN CAMBODIA
SARA ELENA CANAVATI DE LA TORRE
A THESIS FOR THE DEGREE OF THE DOCTOR OF PHILOSOPHY (TROPICAL MEDICINE)
FACULTY OF TROPICAL MEDICINE MAHIDOL UNIVERSITY BANGKOK, THAILAND
2013
0
Table of Contents
ACKNOWLEDGEMENTS ........................................................................................................... iii ABSTRACT .................................................................................................................................. iv LIST OF TABLES .......................................................................................................................... 3 LIST OF FIGURES ........................................................................................................................ 4 LIST OF ABBREVIATIONS ......................................................................................................... 5 CHAPTER I
INTRODUCTION ................................................................................................. 9
CHAPTER II
LITERATURE REVIEW .................................................................................. 15
CHAPTER III
METHODS ....................................................................................................... 33
Chapter IV
Assessment of the effectiveness of current BCC interventions ............................. 53
Village Malaria Workers (VMWs) ..........................................................................................................54 Mobile Malaria Workers (MMWs) .........................................................................................................56 Health Education thought VMWs and MWs ...........................................................................................59 Lending Scheme at Farms .......................................................................................................................61 Taxi Driver Scheme .................................................................................................................................67 Health Education .....................................................................................................................................68 Clients Feedback ......................................................................................................................................72
CHAPTER V
Targeting Behaviours not Individuals................................................................ 76
Net use .....................................................................................................................................................76 Net Preferences ........................................................................................................................................79 Treatment Seeking Behaviors ..................................................................................................................81 Correct Treatment ....................................................................................................................................82 Treatment Adherence ...............................................................................................................................83 Direct Observed Treatment ......................................................................................................................84 Side effects ...............................................................................................................................................87 Traditional Beliefs ...................................................................................................................................87
CHAPTER VI
System Issues.................................................................................................... 89
MMW Motivation ....................................................................................................................................89 Incentives .................................................................................................................................................90 Stock outs.................................................................................................................................................90 VMWs/MMWs lack of basic supplies .....................................................................................................92 1
Training of Workers in low endemic settings ..........................................................................................92 The Private Sector ....................................................................................................................................93 Client Satisfaction ....................................................................................................................................95 Perceived lack of need of MMWs due to low endemicity .......................................................................96 Perceived lack of malaria-related tasks by MMWs due to low endemicity ............................................96 Misinformation ........................................................................................................................................99 Limited exposure to intervention: MMW ..............................................................................................102 Lack of outreach by MMWs ..................................................................................................................103 Long distances and lack of transport .....................................................................................................105
Chapter VII
The way forward: MMP Framework .................................................................. 107
CHAPTER VIII CHAPTER IX
DISCUSSION .............................................................................................. 124 CONCLUSION .............................................................................................. 143
REFERENCES ........................................................................................................................... 149 BIOGRAPHY ............................................................................................................................. 157
2
LIST OF TABLES Table Page 2.1 Definitions of artemisinin resistance containment zones 2.2 Health promotion approaches 2.3 Categorisation of human population movements 2.4 Categories of mobile population groups and profiles 3.1 Study Variables linked to quantitative survey: Objectives 1&3 3.2 Study Variables linked to quantitative survey: Objective 2 3.3 Study Operational definitions 3.4 Study Operational Definitions 3.5 Study variables and their corresponding conceptual definitions and methods 3.6 Responsibilities of the two moderators 3.7 MMP strategy definitions 4.1 Indicators pertaining to VMWs/MMWs compared with assessment results 4.2 Health education messages provided by V/MMWs to MMPs 4.3 Summary of observations and insights (from all stakeholders) from Pailin informal survey (Mar 2012) 5.1 Reported Reasons for not using ITNs 5.2 Reported Reasons not to take ITN to forest 5.3 Migrants who slept under an ITN the night before the survey 5.4 Reported places for seeking malaria treatment in migrant population. 7.1 Malaria vulnerability index 7.2 Population Movement Framework 7.3 MMP profiles definition 7.4 Forest/malaria exposure index 7.5 MMP Access/outreach index 7.6 MMP Malaria Risk Index 7.7 Summary intervention packages by MMP profiles 7.8 Interventions and delivery channels 9.1 Proposed BCC Indicators for National Surveys in Cambodia
3
LIST OF FIGURES 5.1
MMP Malaria Ecosystem
6.1
Causes of confusions in messages (VMWs and Community)
4
LIST OF ABBREVIATIONS ACD
Active Case Detection
ACT
Artemisinin-based Combination Therapy
AS
Artesunate
ASEAN
Association of Southeast Asian Nations
BCC
Behaviour Change Communication
BMGF
Bill and Melinda Gates Foundation
CMS
Cambodia Malaria Survey
CNM
National Centre for Parasitology, Entomology and Malaria Control Cambodia
CPA
Comprehensive Package of Activities
DOT
Directly-observed Treatment
EDPT
Early diagnosis and prompt treatment
FHI360
Family Health International 360
GF
Global Fund
GFATM
Global Fund to Fight AIDS, Tuberculosis, and Malaria
GFR9
Global Fund Round Nine
GMAP
The Global Malaria Action Plan
HC
Health Centre
HCP
Health Coverage Plan
HS
Health System
HSS
Health Systems Strengthening
IEC
Information, Education, and Communication
IPTp
Intermittent Preventive Treatment in Pregnancy
IRD
Institut de Recherche pour le development – Institute of Research for Development
IRS
Indoor-residual sprays
ITN
Insecticide Treated Net
ITNs
Insecticide-treated nets
KAP
Knowledge, Attitudes, and Practices
LLIHN
Long Lasting Insecticidal Hammock Nets 5
LLIN
Long-Lasting Insecticidal Net
LLNs
Long Lasting Nets
M&E
Monitoring and Evaluation
MC
Malaria Consortium
MDG
Millennium Development Goal
MMP
Mobile and Migrant Population
MMW
Mobile Malaria Worker
MOH
Ministry of Health
MPA
Minimum Package of Activities
MSH
Management Sciences for Health
NGO
Non-Governmental Organization
NHSR
National Health Statistics Report
NMCP
National Malaria Control Programme
NMS
National Malaria Service
OD
Operational District
PD
Positive Deviance
Pf
Plasmodium falciparum
PFD
Partners for Development
PHC
Primary Health Care
PMC
Prevention, management and control
PSI
Population Services International
Pv
Plasmodium vivax
RDS
Respondent-Driven Sampling
RDT
Rapid Diagnostic Test
RH
Referral hospital
SMS
Short Message Service
SP
sulfadoxine-pyrimethamine
URC
University Research Co. LLD
USAID
United States Agency for International Development
VCNs
Volunteer Collaborator Networks
VCs
Volunteer Collaborators 6
VHSG
Village Health Support Group
VM
Volunteer Medicators
VMW
Village Malaria Worker
WHO
World Health Organisation
7
Abstract Background Village and Mobile Malaria Workers (VMWs/MMWs) are the backbone of all malaria related activities in Cambodia such as surveillance, diagnosis and treatment of malarious cases within their respective villages or migrant worker populations based at farms. Mobile and migrant populations (MMPs) are usually poorly connected to routine public health interventions and surveillance systems and therefore represent a vulnerable group with regards to the malaria control. The MMW scheme was implemented in 2009 in order to deal with MMPs in an attempt to contain artemisinin resistance in Pailin Province. Methods A mixed methods study was conducted in six provinces of Western Cambodia in the artemisinin resistance containment zones 1&2. A total of 498 semi structured interviews, 60 FGDs (416 participants) and 65 IDIs were conducted through October 2011 till January 2012, including VMWs/MMWs, public health facility staff, village chiefs, villagers, migrants and malaria patients. Results This study found that although MMWs have a strong potential to perform all malaria related tasks especially that of health education, however, they are not doing active outreach to the migrant populations, The preferred means of communication among migrants is interpersonal communication, yet MMWs were not the preferred source of information. The levels of satisfaction were very high among migrants who knew the MMW. MMWs face several challenges when proving malaria related services to MMPs in Pailin such as difficulties observing treatment, distances, transport Treatment seeking behaviors of migrants (prefer the private sector) Net preference (prefer traditional nets), net usage (lack of nets), treatment adherence, and taking the correct treatment (due to many traditional beliefs). There are also system issues that need to be solved such as stock outs, training and supervision of MMWs, the unregulated private sector, lack of perceived need of MMWs by the villagers and of malaria tasks by the MMWs due to such a low endemicity as well as miscommunication of malaria messages by MMWs to the migrant populations (such as boiling water as a method for malaria prevention). Two other interventions targeting migrants were assessed: the net lending scheme and the taxi driver scheme. The net lending scheme is working for those who have access to it. Yet the taxi scheme does not seem to be working for the migrant populations in Palin due to lack of supervision and lack of motivation and knowledge of the taxi drivers—malaria sensitisers. Based on the above evidence, a framework consisting on several semi quantitative indexes based on scoring of knowledge of malaria transmission-prevention-diagnosis and treatment; ownership and use of prevention measures (i.e. LLIN/ITN); housing type and conditions; and immune status, has been developed. Intervention packages per MMP profile were also developed.
Discussion MMWs must to be the backbone of all malaria interventions among MMPs. MMWs need to do more active outreach especially because MMPs in Pailin are too mobile and are isolated and have little or no access to mass communication. Lack of knowledge must be address among all other system issues in order for the MMW programme to benefit the migrant populations. In order to target MMPs, a package of interventions needs to be implemented. The lending and the taxi driver schemes must be reinforced and continue to be implemented and expanded. However a comprehensive evaluation of the taxi scheme must take place. Due to the low prevalence of symptomatic cases, village active surveillance as an intervention must be implemented in Pailin through the MMWs in order to detect the last parasite among asymptomatic carriers. Targeting migrants at pre-departure stage is also highly recommended due to the human population movement form malaria free-provinces to endemic provinces in Cambodia. 8
CHAPTER I
INTRODUCTION
It is believed that behavioural factors might have contributed to the confirmed emergence and spread of drug resistance including artemisinin resistance along the Thailand-Cambodia border[1] [2] [3] [4]. In an attempt to change behaviour of people in order to contain the spread of drug resistance, a number of BCC key messages have been delivered to the general population, private sector and mobile and migrant populations in zones 1 and 2 of the containment project. This has included using Information Education Communication (IEC) materials, mass media campaigns and a major focus on village volunteers[5]. Eliminating artemisinin-resistant strains of malaria in Cambodia is of global epidemiological significance [6]. Under the Bill and Melinda Gates Foundation-funded (2007-2009) “A Strategy for Containment of Artemisinin-resistant Plasmodium falciparum parasites in South-East Asia” (also known as the Containment Project), the National Malaria Control Programme (NMCP) and other partners were striving to retain currently effective antimalarial drugs, namely artemisininbased therapies by containing and/or eliminating artemisinin-resistant parasites in Cambodia and Thailand [7]. Additionally, the project aimed to undertake research to characterise the resistance, define the extent of the spread, and test new strategies for future containment of drug resistance. The overall goal of the containment project was to contain the spread of artemisinin-resistant Plasmodium falciparum parasites by removing selection pressure, and reducing and ultimately eliminating falciparum malaria [8]. Village Malaria Workers (VMWs) have been very successful in extending malaria diagnosis and treatment in malaria-endemic areas [9, 10]. However, mobile/migrant populations remain highly at risk, and are not always reached by MMWs and other village-based services[2]. As a result, they may also spread drug-resistant strains elsewhere in Cambodia and the region[6]. The National Malaria Control Programme has recognised the importance of accurate information on migrants to guide containment strategy. Internal mobile/migrants are an important component of this strategy within the containment zones[11] [7] . However, it is difficult to sample migrants as it is difficult to generate an accurate sampling frame[12]. Due to their frequent movement makes it difficult to obtain accurate sampling and population numbers[13] [14] [15] [16] [17].
9
Population movement has been implicated in the spread of resistant malaria[13] [18]. For this reason, internal migrants are a key focus of activities for the containment project[19]. In addition, very little statistically reliable information is available on movement, care seeking, and malaria prevention behaviours of internal migrants in Cambodia. Mobile/migrant workers are a key component to controlling and eliminating malaria in Cambodia [20] [21]. The first study in the field of malaria on migrants is the malaria respondent driven sampling which is a new method in malaria. It has been used for HIV research for quite some time for surveying hidden populations 1. The main behaviour change theories used to explore and influence people’s health behaviours are the Health Belief Moderl (HBM), the stages of change, and the theory of reasoned action[22]. One of the earliest theoretical models developed for understanding health behaviours was the health belief model[23]. The model was developed in the 1950s to explain why people did not engage in behaviours to prevent or detect disease early. It is based on the notion that the frequency of a behaviour is determined by its consequences. Perceived susceptibility is the perception of personal risk of developing a particular condition, and it involves a subjective evaluation of risk rather than a rigorously derived level of risk. An important contribution of the model is the recognition that prevention requires people to take action in the absence of illness[24]. The limitations of the HBM [25]are related to lack of predictive value for some of its central tenets. For example, the perceived severity of a risk does not reliably predict protective health behaviours[24]. For the purpose of this study the HBM has been used to guide the design of the study tools as this model looks at perceived threat of disease and net benefit of changing behaviour to determine if and why a person will adopt a behaviour change. BCC is now an integral component of the national elimination strategy (2011-2025) with the aim to increase awareness and improve health seeking behaviours of communities to fight artemisinin resistant parasites [26]. It has been well documented that the population residing in Western Cambodia have a high level of knowledge of malaria and prevention key messages (>90%) but this does not necessarily translate into preventive malaria practices[7]. After intense interventions (including free mass LLINs distribution) and large amounts of money spent in the Cambodian
1
Study not yet published
10
national malaria programme, only 40-60% of the population in zones 1 and 2 of Western Cambodia was sleeping under an LLIN [7]. This gap between high levels of knowledge but low/lower levels of practice is often called the knowledge, attitudes and practices (KAP) gap. Information alone is rarely enough to bring about a sustained change in behaviour and the underlying determinants of behaviour must be addressed by creating more supportive and enabling environments. Gender relations, socio-economic power, support and self-esteem are key components of these environments[27]. To examine why a KAP-GAP exists needs more in-depth examinations that goes beyond KAP information gathered through surveys [28]. In-depth quality data was an integral part in this study for obtaining information concerning sensitive topics, e.g. use of traditional methods, issues related to migrants. The strength of this study is that we have chosen the appropriate methods in relation to the study objectives. This research was able to study health-seeking knowledge, attitudes, and practices in context, including focus group discussions, in-depth interviews, and participant observation. The use of mixed methods, has strengthened the validity and reliability of the data[29]; hence, triangulation has provided the ability to validate the integrity of this study results[30]. This study was developed as part of the package to provide more concise and stronger data for analysis than would be possible through the use of a single method as well as helping to assure the trustworthiness of the data. Therefore, the use of mix methods is most appropriate in this study to understand behavioural factors for the prevention of malaria in a context of containment of drug resistance and malaria elimination. Objectives Overall objective The main objective of this study was to assess the nature and quality of the VMWs’ services by interviewing VMWs and to identify the similarities and differences of the quality of the VMWs’ services from the perspective of community members in order to inform sound policy making. Specific objectives This study specifically aims to: 11
1. assess the overall effectiveness of BCC/Information Education Communication interventions on attitudes, behaviours and practices of the internal migrants, indicated by changes in knowledge, attitude and practices toward malaria prevention, diagnosis and treatment after 1-year of implementation of the BCC/IEC program; 2. assess the Village/Mobile Malaria Workers performance focusing on the delivery of the key messages in regards to malaria control, prevention and treatment; 3. inform sound policy making based on the results of the above research as well as the compilation of grey literature in Cambodia. Study Questions What is the overall effectiveness of mobile malaria workers, in terms of changes in target communities’s knowledge, attitudes, and practices toward malaria prevention? What are the factors that contribute towards the successful implementation of the BCC program among internal migrants after 1 year of implementation of the program, (including performance of Village/Mobile Malaria Workers in zones 1 and 2 of the containment project), in a malaria elimination setting in Western Cambodia? Study Hypotheses 1. Mobile malaria workers are effective in changing individual’s knowledge, attitudes, and practices toward malaria prevention and control among internal migrants. 2. Village/Mobile Malaria Workers are effective key personnel in the delivery of the key messages in regards to malaria control, prevention and treatment. 3. Behavioural factors, including net use, treatment compliance and treatment completion are important to the success malaria prevention in a drug resistance setting. Study Rationale This is the first study that evaluates the nature and quality of the VMWs’ services by interviewing VMWs and identifies the similarities and differences of the quality of the VMWs’ services from the perspective of community members. Mobile Malaria Workers (MMWs) were assigned to work with MMPs in selected farms in Pailin province on malaria preventive behaviours and drug behaviours, following the containment project, under Global Fund Round 9. This number has further increased significantly during the last four years. In other parts of Western Cambodia, the VMW/MMW programme has been 12
running for the last 10 years and no formal evaluation on VMWs performance had taken place at the time of the study. In the World Health Organisation’s Global Plan for Artemisinin Resistance Containment (GPARC) operational research into MMPs is highlighted as a vital part of containing and preventing resistance.[31] According to the strategy document, building scalable models to reach MMPs should be the highest priority for research. Significance of the Study The current evaluation of these interventions has contributed to the efforts towards the understanding of the underpinning issues associated to the control of drug resistance in Western Cambodia, especially those related to risky behaviours. The results of this study have and will continue to inform the strategies of the National Malaria Control Programme and other groups working with migrants in Cambodia (for instance with malaria volunteers from their own groups) to ensure better access to health services; providing tailor-made prevention tools in the form of long-lasting insecticide-treated hammock nets (LLIHNs); developing specific behaviour change and communication (BCC) strategies; and attempting to incorporate them in the routine surveillance systems by •
Developing an assessment instrument and methodology that could be used throughout the SSF sites in order to develop innovative tools to be used in the GFR9 VMW/MMW monitoring and assessment.
•
Producing results that will help CNM improve the VMW/MMW programme throughout the country by identifying potential challenges and weaknesses in the current system.
In order to more effectively undertake these activities, detailed information has been obtained on the MMPs patterns of migration, work patterns, access to and use of prevention measures, health seeking behaviour and barriers to services, and their knowledge of malaria and malaria control tools. The results have not only helped to improve the health of the migrant population, but in doing so, could potentially also contribute to decreasing the risk of malaria among the Khmer population and reducing the spread of resistant malaria parasites. This thesis addresses both the recipients (migrants) and the deliverers (Village Malaria Workerss/MMWs) of the behaviour change interventions. 13
The lack of statistically reliable information available on internal migration in Cambodia is worrying as MMPs represent a potential threat to controlling and eliminating malaria efforts in Cambodia. This study has aimed at addressing a set of indicators and recommendations for the national malaria programme (CNM) and partner NGOs aiming at more effective BCC interventions on malaria prevention and risk reduction; identify possible factors determining the uptake of interventions; identify possible reasons for not-using LLINs which may need to be addressed in next rounds/pulses of BCC; identify sources and channels of IEC/BCC activities supporting LLIN distribution and enhancing appropriate use of all interventions and to improve the existing knowledge and information about migrant populations. In addition, the VMW assessment has developed a comprehensive toolkit with assessment instruments and methodologies that could be used throughout the SSF sites in order to develop innovative tools to be used in the VMW/MMW monitoring and assessment.
14
CHAPTER II
LITERATURE REVIEW
The Kingdom of Cambodia has a population of approximately 14.2 million 2 [32]. Around 77% of people live in rural areas [33]. Overall 28% live below the national poverty line but poverty is disproportionately high amongst the rural population 3 [34]. Malaria transmission is very focal and generally restricted to relatively sparsely populated forested areas. An estimated 7.5 million people are considered to be at risk of infection and 6.2 million live in areas where reported annual incidence exceeds 1 case per 1,000 population [3]. Cambodia sits within the Greater Mekong Subregion (GMS), a region of six countries/regions sharing the Mekong river basin (Cambodia, Yunnan Province in The People’s Republic of China, Lao People’s Democratic Republic, Republic of the Union of Myanmar, Kingdom of Thailand, and Socialist Republic of Vietnam). Although reductions in prevalence and incidence have been seen in recent years, malaria remains an important public health risk in the GMS. The epidemiology of malaria varies widely both between and within countries, though the majority of cases throughout the region are Pf and Pv, and countries share many of the same challenges for control: lack of strong health systems, counterfeit and sub-standard drugs, outdoor transmission, drug resistance and highly mobile populations[35]. Outdoor transmission is a product of the major mosquito species in the region. Prominent in Cambodia are Anopheles minimus and An. dirus complexes[36, 37]. Both are associated with forests forest-fringe regions[38, 39]. Members of the An. dirus complex are highly efficient vectors, exhibiting substantial exophilic and exophagic behaviour, and biting early evening before people are under nets[38, 40]. Control measures have included a ban on artemisinin monotherapy since 2010, use of Rapid Diagnostic Tests (RDTs) at the community level and free distribution of treated nets. Village and Mobile Malaria Workers (VMWs/MMWs) are responsible for surveillance, diagnosis and treatment of malarious cases within their respective villages or migrant worker populations based at farms[41]. A scheme is underway to have VMWs/MMWs alert Public Health Departments (PHDs) to cases via use of an SMS system, which can update the HIS in real time (the HIS does not currently include VMW/MMW data). Details are also taken on an individual’s migrant status and recent travel history[42]. 2 3
Based on a 2008 population of 13.4 million and a population growth rate of 1.54 per annum. 48% of the rural population is in the lower 2 wealth quintiles compared to 4.1% of the urban population.
15
While a positive picture overall, the result of control efforts is a concentration of malaria infection within certain areas and specific high-risk groups. The 2010 malaria survey (CMS2010) found that mobile populations had a positivity rate of 1.5%, compared to 0.9% in the general population[43]. Most at risk of infection were forest-goers (individuals that slept overnight in the forest at least once during the previous 6 months) with 2.5% prevalence. Male 15-49 year-olds had significantly greater prevalence than any other age-group and gender category (1.5%). It is this group that makes up the majority of travellers and forest-goers, moving around to find employment or work in activities within the forest, such as logging, where the source of transmission is found[44]. There has also been a shift in the ratio of Pf:Pv infection, with the proportion of Pf falling and Pv rising[41], because Pv is harder to control and eliminate than Pf[45]. The latest Figure of mixed infections is not reported in the CMS 2010 [43] Overall, there has been recent improvement around the use of preventative measures. The CMS 2010 estimates that the proportion of respondents who slept under an Insecticide Treated Nets/ Long Lasting Insecticide Treated Nets the previous night has increased to over 50%, compared to a 29% in 2004 and 25% in 2007 [43]. Despite this improvement, there is a particular need to increase usage amongst forest-goers. Half of the respondents who reported going to the forest previously spent more than 4 weeks there and only one-third of these reported using a mosquito net[26] Increasing and maintaining the use of ITN and LLINs remains key, as identified in the current National Malaria Strategic Plan. Each household living in a malaria endemic area is eligible to receive one ITN/LLIN and hammock net4. Between September and December 2011 approximately 2.7 million ITNs/LLINs were due be distributed across 4057 villages and 45 operational districts of 20 provinces of Cambodia. The use of ITN/LLINs remains a key approach to reducing malaria transmission, especially in endemic areas. 5 With the start of the Global Fund Round 9 (GFR9) grant and following the National Strategy Plan of Royal Cambodian Government from 2011- 2025 implemented by the national malaria control programme, parasitology and entomology (CNM), the following vector control activities were approved for 2011-2015: (i) Distribution of LLINs with a ratio of 1 net per person, (ii) distribution
4 5
Usage of hamock nets was not measured in the CMS 2010. CNM (2012) Press release: http://www.cnm.gov.kh/, retrieved on February 23, 2012
16
of Long Lasting Insecticide Hammock Nets (LLIHNs) with a ratio of 1 net per family in villages at risk, and (iii) re-treatment of existing conventional nets with long-lasting insecticide. For those villages that already received a LLIN every 2 people under the previous campaign, an extra LLIN every 2 people was distributed to ensure the targeted ratio of one LLIN per person [46] (Figure 4). Under the containment project several innovative interventions were piloted. An LLIN lending scheme has been implemented to provide high coverage of LLIN amongst mobile and migrant workers and family members. Both URC and FHI have been involved in the employer ‘Lending Schemes’ begun in Pailin and now being extended to throughout the country through the national programme. In order to increase access to accurate diagnosis and treatment in remote forested areas, the CNM launched the Village Malaria Worker (VMW) project in 2001. The CNM identified malaria-prone villages, where two VMWs (a male and a female) per village were selected through community consensus[10]. The number of villages with trained village malaria workers has increased to approximately 1390 and there are now 103 trained mobile malaria workers[5]. The CMS 2010 suggests that knowledge around the correct treatment has been steadily increasing since 2004[47]. However, although most people are aware of treatments to cure malaria, some are still not aware of the need to take the entire course of treatment in order for it to be effective[48]. Pailin has been the focus of control and elimination efforts ever since drug resistance to artemisinin was reported as originating from there. Labelled as ‘Zone 1’ the focused efforts have meant a huge reduction in malaria transmission[49, 50]. The artemisinin resistance containment project on the Cambodia-Thailand border in 2008 established three containment zones as defined in Table X. Most of the current epidemiological data relates to these Zones. 2.1
Definitions of Artemisinin Resistance Containment Zones [19]*
Zone
Zone 1
Definition
Population / Provinces 0.27 million people in 254 The areas of the Thai-Cambodian border where villages covered by 23 health there was evidence of artemisinin tolerant centres in 8 administrative Plasmodium falciparum. districts in 5 operational districts in 4 provinces
17
Areas where there was no evidence of 4.02 million people in 336 artemisinin resistance, but the risk was high villages in 9 provinces, excluding Zone 2 because they are near to Zone 1 (plus one urban areas district in Kampong Speu – Oral - where slow parasite clearance time had been reported). Excluding non-target zones such Zone 3 The rest of Cambodia as Phnom Penh, Kandal, Prey Veng and Svay Rieng provinces, *Based on definitions in the original Project Proposal document -– 2008 The recent new initiative by WHO called “Emergency response to artemisinin resistance in the Greater Mekong subregion: regional framework for action” highlights calls key action areas in which progress is urgently needed in the coming years if we are to contain resistance and move towards elimination of malaria in GMS. It recalls the overarching containment goal of protecting ACTs as an effective treatment for P. falciparum malaria. The framework seeks to do this by rallying stakeholders to urgently scale-up and increase the effectiveness of interventions to address artemisinin resistance[51]. BCC interventions (Table X) have been defined according to the stratification guidelines of regional framework stratetgy[51]. Drug Resistance Drug resistance in malaria has historically first developed along the Thai-Cambodia border. Resistance developed to chloroquine in the late-1950s (virtually simultaneously with emergence in Colombia, South America), sulfadoxine-pyrimethamine and quinine in the mid-1960s, mefloquine in the late-1980s, and more recently to artemisinin [52, 53]. The Global Plan for Artemisinin Resistance Containment (GPARC) was launched in 2011 to contain artemisinin resistance along the Thai-Cambodia border[50]. However, there has since been evidence of increased treatment failure across the GMS[54] [55]. Primary Health Care Improving access to treatment and providing appropriate health education and support to people at risk of or using treatment for malaria are essential components of successful malaria control. One way of improving access is through community participation. International efforts to improve access to primary healthcare (PHC) have begun since the Alma Ata Declaration in 1978 highlighting the need of PHC improvements to achieve a minimum standard quality of life [56].
18
Health promotion is the process of enabling people to increase control over, and to improve, their health. It moves beyond a focus on individual behaviour towards a wide range of social and environmental interventions [57]. Theories are useful in behaviour research to inform the design and to guide the implementation of interventions [58]. The main behaviour or change theories about reasons people change their behaviours are the health belief model, social cognitive theory, cognitive-behavioural theory, the stages of change and the theory of reasoned action[22]. One of the first models created with the main objective to explain health behaviour was the HBM. The basis of the theory is that knowledge of perceived susceptibility, severity, benefits and barriers jointly predict health behaviours [58]. The range of behaviours that have been examined using health belief models have been categorised into three broad areas: preventive health behaviours, sick role behaviours and clinic use [59]. In this type of model, individual beliefs offer the link between socialisation and behaviour. Hochbaum’s study (1958) offers the earliest examples of the uptake of screening for TB, where he identified that a belief that sufferers could be asymptomatic was linked to screening uptake. Hochnaum’s discovery is the highly relevant for the purpose of this study as in a low endemic area, such as Pailin, asymptomatic parasitemia carriers are the focus of most interventions. The Health-Belief Model focuses on the person's perception of the threat of a health problem and the appraisal of recommended behaviour(s) for preventing or managing the problem. In this model “for a behaviour change to take place, individuals must have an incentive to change, feel threatened by their current behaviour, feel a change would be beneficial in some way and have few adverse consequences; and must feel competent to carry out the change” (p. 223) [60]. The key concepts of the model are perceived susceptibility, perceived severity, perceived benefits of action, cues to action and self-efficacy. Health belief models focus on two elements: ‘threat perception’ and ‘behavioural evaluation’[59]. Threat perception depends upon perceived susceptibility to illness and anticipated severity; behavioural evaluation consists of beliefs concerning the benefits of a particular behaviour and the barriers to it. ‘Cues to action’ and general ‘health motivation’ have also been included[25] (see Figure 1).
19
The health belief model has been criticised by some for portraying individuals as asocial economic decision makers, and its application to major contemporary health issues, such as sexual behaviour, have failed to offer any insights [59] Social-cognitive theory as developed by Bandura[61] assumes that healthy functioning is determined by the interactions among behavioural, physiological, and cognitive factors and the environment [61]. According to Sallis [62], this theory has been useful in predicting maintenance for physical activity as well as other health behaviours. Agency is defined “as the embodiment of the endowments, belief systems, self-regulatory capabilities and distributed structures and functions through which personal influence is exercised, rather than residing as a discrete entity in a particular place” [63]. Within the social cognitive theory, to be an agent is to intentionally make things happen by one's actions[61]. The core features of agency enable people to play a part in their self-development, adaptation, and self-renewal with changing times. It is argued that for the most part, behavioural theories grant humans little, if any, agentic capabilities[61]. The theory of reasoned action is concerned with the role of anticipated material and social consequences in people’s decisions and intentions to engage in health-related behaviours. In short, this theory states that intentions are the most immediate influence on behaviour. It focuses on intentions to perform a behaviour based on attitudes toward the behaviour and social acceptance of a behaviour and perceived behavioural control[64, 65]. The transtheoretical model is based on the observation that people tend to go through similar stages of change no matter what treatment is applied [66]. This model is basically composed of two ideas which are that different intervention approaches are needed for people at different stages of behaviour change and that different processes of change may be occurring respectively. This model helps program planners to tailor interventions based on a person's stage of readiness to change: Precontemplation, contemplation, preparation, action, or maintenance. People’s needs differ in their different stages of change and therefore need to receive different interventions[67] Due to the fact that public health problems are usually addressed at different levels of society, the social action theory was designed to specifically integrate social-cognitive models into a public health framework[68] (Table X). 2.2 Approach
Health promotion approaches [60] Aim
Methods 20
Criticisms
Medical
Behaviour Change
Educational
Reduce morbidity and Medical intervention, premature mortality risk education
Seeks to increase medical interventions and compliance; ignores social and environmental factors Encourage individuals to Prompting, cueing, “Victim‐blaming”; ignores adopt “healthy” vicarious learning social and economic behaviours factors; difficulties in addressing multi‐factorial causes of behaviour Provide information and Information‐giving Assume the relation develop skills so that through small groups or knowledge‐attitude‐ people can make massmedia; group behaviour change informed decisions discussions; role‐plays
Empowerment
Identify individual or community health needs and gain the knowledge, skills and attitudes to act upon them through a program of action
Counselling, problem solving, community development, advocacy, public participation
Is difficult to quantify results; assume that rational choices are healthy choices
Social change
To bring about changes Lobbying, advocacy, in physical, social, and policies, fiscal measures economic environment, which enables people to enjoy better health.
It may require major structural changes. Vulnerable to official oppositions.
Behaviour Change is a health promotion approach. Behaviour Change Communication (BCC) is an interactive process of working with individuals and communities to 1) develop communication strategies to promote positive behaviours as well as 2) create a supportive environment to enable them to adopt and sustain positive behaviours [24]. Behaviour change is a primary objective in public health programmes[69], both to modify lifestyles which risk individual well-being, and to achieve health-improvements and environmental change[70]. In efforts to induce voluntary behaviour change, without economic or legal intervention, there are two basic choices: interpersonal and mass communication [71, 72], The most commonly accepted method is interpersonal communication through face-to-face education, either in individual or small group sessions, group teaching, and other techniques designed to influence the behaviour of participants. Behaviour change interventions targeted at MMPs In Western Cambodia, there have been several interventions in BCC conducted by a range of partners. The components of these campaigns have included: information, education, 21
communication (IEC) campaigns and post-distribution “hang-up campaigns” need to explain which are strongly recommended for malaria elimination [73]. •
Taxi Driver Scheme
Taxi drivers – malaria sensitizers are paid an incentive of $10-$15 per month for participation in the program. The aim of the strategy is to enable taxi drivers bringing new migrants from the neighbouring city of Battambang to Pailin and back, to sensitize the newcomers to the area about malaria prevention, diagnosis and treatment messages. Cars are outfitted with ‘Ask me About Malaria’ external car stickers; sun visors and music tapes with the ‘Ayay’ malaria song and program (of malaria messages) to play in the car. •
Net Loaning Scheme
LLIN Lending Scheme is implemented through private farm employers who supply LLINs to seasonal migrant labourers in order to subsidise the full cost of the LLINs, which are supplied free to farms. Monitoring and administration provided with NGO support (FHI/URC). BCC provided by VMW and MMWs with support from FHI/URC •
Media products for broadcast on radio and television
BBC Media Action and Women Media Centre produced media products for broadcast on radio and television and for community use targeting segmented populations, about malaria resistance, treatment and prevention. BBC’s media products were broadcasted nation-wide. •
Mobile Broadcasting Units (MBUs)/Listener and Viewer Clubs (LVCs)
In a recent evaluation of mass media practices in Cambodia conducted by the BBC Media Action of BCC interventions among migrant populations in Western Cambodia it was shown that more respondents had radio than TV access[74]. Key listening times were 6 to 7am, 11am to 1pm and 5 to 7pm. Respondents listened to the radio less in rainy (March/April) and harvest (July) seasons than at other times of the year. TV access and ownership was low amongst respondents, especially the new-comers to malaria endemic regions. For households with TVs, children’s viewing habits were influential in determining what households watched. The most popular viewing times of the national TV channels were 5 to 6pm and 9 to10pm. The new-comers to malaria endemic areas 22
reported particular difficulties receiving mass media and therefore had watched or listened to the spots less than the other groups[74]. Women Media Centre’s (WMC) media products are being broadcasted in seven target provinces as well as through Mobile Broadcasting Units (MBUs) and Listener and Viewer Clubs (LVCs) in 30 target villages in three provinces (Pailin, Battambang and Koh Kong). •
Village Health Volunteers (VHV)
Asia Medical Doctors Association (AMDA) is working with Village Health Volunteers in Kampong Speu province to encourage preventive and positive health seeking behaviour in communities, build the capacity of heath centre staff and improve the utilisation of the public health system for malaria diagnosis and treatment. A recent assessment showed that there is no difference in BCC practices among villages with VHVs and no VHVs[75]. •
Village Malaria Workers (VMWs)/ Mobile Malaria Workers (MMWs)
CNM, FHI and URC and are working with Village and Mobile Malaria Workers (VMWs and MMWs) with a focus on continuing containment strategies as well as encouraging preventive and positive health seeking behaviour in communities. Village/Mobile Malaria Workers Since the Alma Ata Declaration, Primary Health Care system development shifted toward the deployment of various types of community health workers (CHW).
Worldwide, malaria
community-based interventions by CHWs have improved health outcomes and decreased malaria mortality in resource poor settings[9] [76] [77]. Cambodia is one of the countries suffering from a health workforce shortage. The country has the greatest sub-national inequities in the distribution of medical doctors among ten countries of the Association of Southeast Asian Nations (ASEAN) [78]. In Cambodia, most public health facilities are under-staffed or do not have enough health providers with sufficient clinical skills to deliver the services in rural and remote areas[79] [80]. CHWs can play a crucial role in addressing the shortage of health workforce, especially in rural and remote areas of the country [81]. 23
Village-based volunteer workers have played an important role in malaria diagnosis and treatment in many different settings for more than 35 years. Two of these programmes stand out in terms of their size and longevity: the Volunteer Collaborator Network of Latin America[82] and the Village Voluntary Malaria Collaborator Programme of Thailand [83]. The success of these programmes was based on a tradition of active community participation and sustained commitment and support from national malaria control programmes. In Cambodia, village malaria workers (VMWs) were first introduced in June 2001 as part of an insecticide treated bednet (ITN) trial conducted by the national malaria control program (CNM) in Ratanakiri Province[84]. Between 2004 and 2005, the VMW scheme was rolled-out to cover 300 villages. Then, in 2006 Pharmaciens Sans Frontieres (PSF) in collaboration with CNM, the Ministry of Health’s (MoH) department for Communicable Disease Control-IMCI and the WHO implemented a pilot project in 52 remote villages in Stung Treng Province to assess the feasibility of adding case management of acute respiratory tract infections (ARIs) and diarrhea in under-fives to the scope of work of VMWs[85]. VMWs were trained to treat simple coughs, colds and diarrhea as well as non-complicated cases of pneumonia and to refer severe cases to the nearest health facility. Three different approaches were tested by randomly selecting VMWs to manage: malaria and ARIs; malaria and diarrhea; or, malaria, ARIs and diarrhea. Although a detailed comparison of feasibility and efficiency within the three intervention groups was never fully assessed, findings indicated that it was feasible for these ‘expanded VMWs’ (eVMWs) to treat all three illnesses given proper training. Small, established rural communities in Cambodia also have a strong tradition of community participation[86] and the CNM has an international reputation for commitment to the implementation of innovative health interventions. This likely contributed to the reported success of the VMW project in Cambodia. The VMW project was launched in 2004 as a vertical CNM led scheme that allowed essential supplies to bypass the delays and other problems inherent in the health system (Figure 7). Migration and health Population movement, migration and mobility in relation to health outcomes and potential health threats (emerging or reemerging diseases) is a global concern, fueled by globalization and demographic and socio-economic disparities.. In 2010, it was estimated that migrants represented
24
almost one billion persons, comprising about 214 million international migrants (40% moving between neighbouring countries) and 740 million internal migrants[87]. There is a high volume of cross-border movement across the Greater Mekong Sub-Region (GMS), particularly of migrant workers. Official Figures reported just over 1 million migrants across the GMS in 2008[11]. However, porous borders mean migrants can cross at numerous unofficial border points, undetected and unregistered. Accounting for this, a study estimated the actual number of migrants as closer to 4 million in 2008, climbing to almost 4.5 million in 2013[11]6. Cambodia is no exception to this trend, with movement common between it and its neighbours: Thailand, Laos and Vietnam. Within the GMS, Cambodia is the third ranking country in terms of out-migration, and the second in terms of destination choice, housing over one million migrants[11]. The importance of migration is only expected to rise because the Cambodian population is so young. Over 45% are under 20 years old[88], so many will be entering the work force in coming years and needing to find work. A study of Cambodian migrant workers in Thailand highlighted the frequency of this movement. A large proportion were residing in Thailand for a short-term period, many had returned to Cambodia within the previous 3 months, and many were planning to return in the future to various provinces across the country[89, 90]. Malaria and Human Population Movement (HPM) The relationship between malaria and population movement is not recent, having been identified at least 50 years ago, at the time of the Global Malaria Eradication Program (GMEP) [13, 14]. The importance of the distinction between migration and circulation and the need to apply various temporal and spatial dimensions to distinguish different categories of human mobility were developed in subsequent years [14] and applied to mobility patterns in northern Thailand in relation to malaria [91]. More recent work, based on the typology of HPM developed [91] aiming at quantifying HPM at various spatial scales, further integrated into models, to inform global, regional and national strategies both in the context of spread of anti-malarial drug resistance, and in the context of the feasibility of malaria elimination. [92, 93] [94], [95-97] In those studies, population movement and population activities are represented along space and time axis. They address the effects of
6
It is believed that accounting for unofficial points it would be 4 times higher.
25
population movement on the circulation, introduction, reintroduction and spread (including resistance) of the parasite at various spatial scales. In South East Asia, the malaria ecosystem and related transmission are closely related to forest areas (10). Therefore, the intensity, duration and frequency of the interaction and exposure with forest condition determine the malaria incidence in a population. Based on this assumption, the National Malaria Strategy developed an approach of stratification of villages “at risk” of malaria which has been used to target malaria interventions. Since the late 1990’s, this approach included impregnated bed net distribution and Village Malaria Workers (VMWs), has been based on the distance of villages from the forest. During this time, Cambodia has gone through major political, economic and environmental changes. The political stability and peace has allowed for development of road infrastructure, improving access to those areas, leading to land development projects, mainly in the agriculture sector. This has led to increased interaction between local and migrant population with the forests. This trend has increased since the mid 2000’s and has led to changes in malaria epidemiology as documented in the Cambodia Malaria Survey 2007 and 2010[47, 48], where forest goers were shown to have significantly higher risk than non-forest-goers. In has been documented that the movement of populations in malaria-endemic areas poses a major threat to the spread of artemisinin resistance along the GMS, more specifically on the Cambodia–Thailand border [89]. It is not only the drive for equity of health care that makes mobile and migrant people a priority population, but also their recognised role in the spread of Artemisinin resistance. Operational research into MMPs is highlighted as a vital part of containing and preventing resistanceas well as building scalable models to reach MMPs should be the highest priority for research[31]. In recent years, large-scale cross-border collaboration has been used to target MMPs in the Greater Mekong Subregion, including Cambodia, through a two year Bill and Melinda Gates Foundation funded containment project. Border populations remain a key area of focus for malaria programs, however, research suggests that the majority of movement amongst MMPs in Cambodia is internal [15, 16]. Therefore, there is a need to strengthen efforts on a domestic level to reach those moving about the country, as well as those moving internationally. According to research findings [98], those mobile populations come to the new place with a variety of purposes which include farming work, investment, trade, visiting relatives, and eventually a prospect to find a new settlement. The following “push and pull” factors have been identified in Cambodia [98]: 26
•
Land use and land resources are one of the main driver of population movement as landlessness in southern provinces due to increased demographic and related land pressure (with high population density) has pushed people to move to border and forested low density areas which contained high natural resources wealth and where land development has been ongoing in the last decade for agriculture and/or exploitation of natural and mineral resources.
•
Poverty is closely related to migration and circulation and affect families both in the place of origin as a push factor and at the place of destination where migrants and mobile population get caught further in a poverty cycle. Health and malaria issues are closely intertwined with migration and poverty, both as cause of migration but as well as a consequence, as non-immune individuals are exposed to malaria while moving to the border areas.
One of the biggest challenges to malaria control in the GMS is the movement of populations within and between countries. When populations move from areas of high to low transmission they hinder control and elimination of malaria by importing infections and acting as a source of local transmission, as well as facilitating spread of drug resistance[13, 17]. Population movements can be categorized according to spatial and temporal characteristics: spatially, migration can occur within a country (rural/urban, rural/rural & urban/urban) or between countries (contiguous and non-contiguous international movement)[99]; while temporally, distinctions are made between migration (permanent/very long term change of residence) and circulation (shorter-term and cyclical movements, no change of residence)[14]. Circulation occurs for different reasons and at different levels of frequency (Table X). It is circulation of populations that results in importation of new infections and spread of drug resistance, while contiguous or cross-border movement has been attributed to maintenance of ‘hot spots’ of high transmission along international borders and potential spread of drug resistance [100]. 2.3
Categorisation of human population movements[14, 17] Circulatory period:
Frequent
Periodic
27
Seasonal
Long-term
Period of
Daily
return trip Reason
Few
Few months
days/months
Yearly/several years
Commute,
Trade, tourism,
Agriculture,
Labour,
trade,
mining
industrial
colonization,
demand
official transfer
cultivation
People who live or work in forested areas or on private farms near the forest are the single greatest sources of malaria cases in Cambodia [7]. Although the forested areas have drastically reduced over the past few years, forest-related activities are still important sources of income for a significant proportion of Cambodians. Often young men from cities go to the forest looking for work. The housing conditions are often primitive; sometimes there are no houses so temporary visits simply sleep in a hammock between two trees. The workers may not have mosquito nets or any sort of on-site treatment. Having come from the city where they are not exposed, they generally have little knowledge of malaria or how to protect themselves and given their low immunity - they often get sick. When sick, many return home for treatment thereby providing a source of transmission so that villages which might have been designated malaria free are at risk of developing cases due to local transmission. 2.4
Categories of Mobile Population Groups and Profiles Mobile Group
Stable, affiliated with local village
Group Profile Includes ethnic minority groups. Visits forest for days/weeks for farming (chamkar), hunting, fishing, and/or collecting valuable products
Semi-stable, affiliated with company
Works for a company (e.g., rubber plantation) in a malaria risk area; may be from another province
Seasonal aggregation of agricultural Works for landowner to plant or harvest agriculture labour, affiliated with land owner products (2-4 months); often from other provinces; (malaria risk depends on terrain and often non-immune environment) 28
Security personnel,
affiliated
government
with Military, police, and border guards who patrol in forests or other malaria risk areas; additionally wives/family
members
often
follow
husbands/relatives and live in nearby risk areas
Population groups have been identified on this basis specifically within Cambodia; three groups were identified7: 1) Local population – permanent resident in the area >1 year 2) Mobile population – Resident in the area for 6 months and 37.2oC), of which 4 tested positive for Pv, giving a positivity rate in the symptomatic sample of 8.2%. However, it is reasonable to suspect that many febrile individuals would not usually make the long, hot journey across the border, and thus that the current protocol is failing to identify many symptomatic and, most significantly, all asymptomatic infections as only those found with fever are tested for malaria infection. Other, there have been no studies on the malaria risk in unofficial cross-border migration along this border. A pilot study currently being carried out by a collaboration among MC, FHI360, IPC and the LSHTM in Ou Chhue Krom and Tun Trom Dey unofficial border points in Pailin using a structured questionnaire and blood collection techniques RDT, microscopy and DBS has not shown any positive cases by RDT. Two official border crossing points (Phsar Prom (Thai) and Trapaing Kreal (Laos) have been selected for comparison. Participants are recruited as they crossed the border and tested for fever, malaria RDT and dry-blood spot (DBS) samples for PCR analysis. Participants are interviewed for information on a priori risk factors for infection. Prevalence and ORs are calculated for each exposure on the outcome of malaria infection. For Trapaing, there was increased odds with fever, security personnel, forest-goers and previous malaria. Only one case was found in Phsar Prom. Multivariate analysis of main effects reduced ORs but associations remained with fever, occupation and forest-goers. Cross-border transfer of malaria appears to be a problem in Trapaing, but not Phsar Prom. Border surveillance activities should be implemented along the Laos border to identify infected border-crossers and prevent import of infection8.
Malaria and MMPs in Cambodia In Cambodia, the 2009 containment survey found the prevalence of malaria among mobile populations (which included travellers, visitors, and forest goers) to be substantially higher than the general population. Among mobile populations, forest goers had the highest malaria prevalence rates (4.1% by microscopy and 11.4% by PCR). Malaria control amongst Mobile and Migrant
Data only from 2 weeks of cross-border surveillance no microscopy or PCR results available yet. Preliminary findings analysed and written by Canavati, S. 8
30
Populations (MMPs) is a great challenge in Cambodia as members of these populations often travel between endemic and non-endemic areas [98]; mobility can increase risk-taking behaviours; poverty increases susceptibility; patterns are unpredictable due to changing land use; and sometimes illegal status leads to avoidance and reductions in care seeking behaviour. Many of these factors contribute to the high incidence of malaria amongst MMPs when compared with a population of similar socio-economic and demographic profile. In addition, MMPs are usually poorly connected to routine public health interventions and surveillance systems and therefore represent a vulnerable group with regards to the malaria control. Border populations remain a key area of focus for malaria programs, however, research suggests that the majority of movement amongst MMPs in Cambodia is internal [15, 16]. Therefore, there is a need to strengthen efforts on a domestic level to reach those moving about the country, as well as those moving internationally. According to research findings [98], those mobile populations come to the new place with a variety of purposes which include farming work, investment, trade, visiting relatives, and eventually a prospect to find a new settlement. From Research to Strategy MMP Strategy Background Increased movement of population from non-malaria endemic areas to malaria endemic areas in the west and the north-east of the country have created new challenges for the malaria control program, both in terms of strategy and implementation. Operational research plays a critical role in helping national programmes to identify and resolve implementation bottlenecks [104]. Particularly in light of national programmes now re-orientating their strategies from malaria control towards malaria pre-elimination and elimination, there is an urgent need to prioritize the operational research agenda in the Greater Mekong Subregion. In October 2010, Malaria Consortium, through funding from the United States Agency for International Development, organized a regional malaria operational research symposium to identify and prioritize key research areas and topics in vector control and prevention, case management, Plasmodium vivax and G6PD, vulnerable populations, M&E and surveillance, health systems and the private sector. These research areas were highlighted by national programmes in the Mekong region as critical to help countries move from control to elimination.
31
Other international efforts to identify research priorities for malaria elimination include the Asian Pacific Malaria Elimination Network (APMEN), Malaria Erradication Scientific Alliance (MESA)9 and the Malaria Elimination Group (MEG). These networks are critical for guiding and supporting the research agenda as well as information sharing amongst national programmes and partners. It should be through these networks or other fora that national malaria programmes drive the operational research agenda for their countries and regions and recently WHO’s Global Malaria Programme10. Current operational research conducted in Cambodia on focal screening and treatment (FSAT), use of G6PD RDTs and administration of primaquine, and SMS technology for improved malaria surveillance, are examples of how operational research can help to improve programme implementation towards the goal of malaria elimination.
9
Previously the Malaria Eradication Research Agenda (malERA)
10
About the WHO Global Malaria Programme http://www.who.int/malaria/about_us/en/
32
CHAPTER III
METHODS
A mixed methods study was conducted in six provinces of Western Cambodia in the artemisinin resistance containment zones 1&2. Two cross-sectional surveys were undertaken consisting of total of 498 semi structured interviews. Interviews consisted of 60 FGDs (416 participants) and 65 IDIs were conducted through October 2011 till January 2012, including VMWs/MMWs, public health facility staff, village chiefs, villagers, migrants and malaria patients. Ethical clearance for the broader study was obtained from the National Centre of Entomology, Parasitology and Malaria Control (CNM). Ethical clearance was also obtained from the Cambodian National Ethics Committee for Health Research in August 2011 (130NECHR) and from the Ethics Committee of the Faculty of Tropical Medicine of Mahidol University (MUTM 2012-021-01). Prior to the interview, the interviewer read carefully the consent form and written consent was sought from the study participants. This consent form contains information on the objectives of the survey, the risks, benefits and freedom of the participation, as well as information on confidentiality (Annex). Funding for this study was obtained from the BMGF through the WHO Containment Project as well as from GFR9. This study was conducted in six provinces within zones 1&2 of the containment project, Western Cambodia. Provinces included in zone 1were Pursat, Pailin and Kampot 11 Provinces. Zone 2 included Battambang 12 and Kampong Speu Provinces (Table X). Objectives Overall objective The main objective of this study was to assess the nature and quality of the VMWs’ services by interviewing VMWs and to identify the similarities and differences of the quality of the VMWs’ services from the perspective of community members in order to inform sound policy making. Specific objectives This study specifically aims to: 1. assess the overall effectiveness of BCC/Information Education Communication interventions on attitudes, behaviours and practices of the internal migrants, indicated by
11 12
Kampot province is split into two zones (1&2). Targeted OD was in Zone 1 Battambang province is split into three zones (1,2&3). Targeted OD was in Zone 2
33
changes in knowledge, attitude and practices toward malaria prevention, diagnosis and treatment after 1-year of implementation of the BCC/IEC program; 2. assess the Village/Mobile Malaria Workers performance focusing on the delivery of the key messages in regards to malaria control, prevention and treatment; 3. inform sound policy making based on the results of the above research as well as the compilation of grey literature in Cambodia. Expected results The results of this study were able to: •
Provide with information on the current knowledge, attitudes and practices towards antimalarial interventions of the VMWs and MMWs in order to: -
Ensure that behaviour change messages are appropriately targeted; Ensure that VMWs are complying with their TOR; Evaluate the outcomes of behaviour change communication activities within the containment project. Identify possible factors determining the uptake of BCC activities delivered by VMWs in all
•
zones 1 and 2 of the containment project. 3.1
Study Variables linked to quantitative survey: Objectives 1&3 Objectives 1&3
Independent variables Dependent Variables SocioMigrants Knowledge on: • age demographic Knowledge • malaria diagnosis • gender and economic • treatment • occupation characteristics • awareness of the key • ethnicity messages • marital status • education • income (per month) Migration characteristics
Other variables
• migrant-issues • consultationtrends • length-of-stay in the farms • province of origin of migrants • training • knowledge • treatment • motivation
Migrants’ attitudes
Migrants’ practices
34
• Perceptions • satisfaction • acceptability of VMW/MMWs job
• • • •
the
malaria diagnosis treatment prevention (bed net use) control activities
• • • • 3.2
acceptability referrals stocks records
Study Variables linked to quantitative survey: Objective 2
For this study operational definitions were developed: Objective 2 Independent variables Dependent Variables VMW’s Key messages about • Age SocioKnowledge • malaria prevention • Gender demographic • diagnosis • Occupation characteristics: • treatment (including • Ethnicity kind of treatment and • Marital status location of treatment) • Education Migration characteristics
• • • •
Other Variables
3.3
• • • • • • •
VMW’s Migrant-issues Attitudes Consultationtrends Length-of-stay in the farms Province of origin of migrants
• • •
Satisfaction Motivationa Acceptability
VMW’s Practices
• • • • • •
malaria diagnosis treatment stocks records referrals prevention and control activities delivery of key messages
Training Knowledge Treatment Motivation Acceptability Referrals Stocks
•
Study Operational Definitions
Age refers to the age of respondent in years at Gender refers to the sex of respondent. the time of the assessment.
35
Education refers to the level of respondent’s Marital status includes single (never education all obtainment/attainment. This can married), married/living with someone as be divided into six categories: Never attended married, widowed and divorced/separated. school, some primary, completed primary (grade 6), some secondary, completed secondary (grade 12), and more than secondary. Occupation includes farmer, laborer, fisherman, merchant/seller, housewife, soldier and other. Training refers to the respondent’s duration of training on what, frequency of training, and type of training. Acceptability refers to perceived acceptability and trust from patients, perceived acceptability and support from health centre staff. Recognition refers to the respondents’ recognition of severe malaria, knowledge of malaria transmission and malaria prevention.
Stocks refer to respondents’ stock levels of all antimalarial drugs and expiry dates on each type.13 Migrants’ Practices refers to bed net use, treatment compliance, Referrals refer to respondents’ practices on identifying malaria suspects in the community, number of referrals, barriers faced when referring patients, and referral records. Awareness of key messages refers to respondents’ awareness of specific drugresistant malaria messages in terms of specific messages on prevention, diagnosis and treatment and use and care of mosquito nets and sources of delivery of key messages.
13
Income per month is defined as the amount of money in Khmer-Riel the participant makes per month at the time of the assessment. Motivation refers to the respondent’s perceived benefits to work as a VMW/MMW, motivation to work as a VMW/MMW, problems faced in the job. Knowledge refers to respondents’ knowledge of signs and symptoms of malaria and differential diagnosis of malaria. Treatment refers to respondents’ understanding of malaria treatment (for both: P. falciparum and P. vivax) including, names of antimalarial drugs, duration of treatment, consequences of treatment non-compliance, and banned/fake/counterfake drugs. VMW/MMW’s Practices refer to performance according to TOR: referrals, DOT, bed net use, outreach to patients Records refer to respondents’ possession of an outpatient register book, whether cases are being recorded and reported to the health centre. Interpersonal communication refers to the respondents’ interpersonal communication with specific population on malaria related topics. Delivery of key messages refers to respondents’ delivery of specific malaria containment messages in terms of frequency, specific messages on prevention, diagnosis and treatment and use and care of mosquito nets and other sources of delivery of key messages.
The definition of stockout is out of stock for more than a week in the past three months. 36
For this study operational definitions were developed: 3.4
Study Operational Definitions
Age refers to the age of respondent in years at the Gender refers to the sex of respondent. time of the assessment. Education refers to the level of respondent’s Marital status includes single (never married), education all obtainment/attainment. This can be married/living with someone as married, widowed divided into six categories: Never attended school, and divorced/separated. some primary, completed primary (grade 6), some secondary, completed secondary (grade 12), and more than secondary. Occupation includes farmer, laborer, fisherman, merchant/seller, housewife, soldier and other. Training refers to the respondent’s duration of training on what, frequency of training, and type of training. Acceptability refers to perceived acceptability and trust from patients, perceived acceptability and support from health centre staff. Recognition refers to the respondents’ recognition of severe malaria, knowledge of malaria transmission and malaria prevention.
Stocks refer to respondents’ stock levels of all antimalarial drugs and expiry dates on each type. 14 Migrants’ Practices refers to bed net use, treatment compliance, Referrals refer to respondents’ practices on identifying malaria suspects in the community, number of referrals, barriers faced when referring patients, and referral records. Awareness of key messages refers to respondents’ awareness of specific drug-resistant malaria messages in terms of specific messages on prevention, diagnosis and treatment and use and care of mosquito nets and sources of delivery of key messages.
14
Income per month is defined as the amount of money in Khmer-Riel the participant makes per month at the time of the assessment. Motivation refers to the respondent’s perceived benefits to work as a VMW/MMW, motivation to work as a VMW/MMW, problems faced in the job. Knowledge refers to respondents’ knowledge of signs and symptoms of malaria and differential diagnosis of malaria. Treatment refers to respondents’ understanding of malaria treatment (for both: P. falciparum and P. vivax) including, names of antimalarial drugs, duration of treatment, consequences of treatment non-compliance, and banned/fake/counterfake drugs. VMW/MMW’s Practices refer to performance according to TOR: referrals, DOT, bed net use, outreach to patients Records refer to respondents’ possession of an outpatient register book, whether cases are being recorded and reported to the health centre. Interpersonal communication refers to the respondents’ interpersonal communication with specific population on malaria related topics. Delivery of key messages refers to respondents’ delivery of specific malaria containment messages in terms of frequency, specific messages on prevention, diagnosis and treatment and use and care of mosquito nets and other sources of delivery of key messages.
The definition of stockout is out of stock for more than a week in the past three months. 37
38
3.5
Study variables and their corresponding conceptual definitions and methods
Category
#
1
2
3 Sociodemographic Characteristics
4
5
6
7
Variable
Study population
Method
Migrants, VMW/ MMW
QUA N
Measurement for quantitative Tools methods
Operational definitions
Migrants, Age of respondent (in VMW/MM years and months) W Migrants, Gender Gender of respondent VMW/MM W Migrants, Main occupation of Occupation VMW/MM respondent W This variable reports the Migrants, Ethnicity ethnic identity of VMW/MM respondents. W Migrants, Marital Respondent's marital or VMW/MM Status civil union status. W Migrants, Level of Highest level of education VMW/MM education of the respondent. W Amount of money in Migrants, Income Khmer Riel the participant VMW/MM makes per month W Age
39
Nominal, ordinal Questionnaire, QUAL (categorical), IDI guide, FGD interval, and ratio guide Continuous
Questionnaire
Ordinal
Questionnaire
Ordinal
Questionnaire
Ordinal
Questionnaire
Ordinal
Questionnaire
Ordinal
Questionnaire
Ordinal
Questionnaire
Training/Capacity Building
8
9
10
Malaria Knowledge
12
Malaria prevention, control and 11 treatment -Practices
Respondent’s duration of training, frequency of Training training, and type of training. Respondents’ knowledge Knowledge- of signs and symptoms of Symptoms malaria and differential diagnosis of malaria. Refers to the respondents’ Knowledgerecognition of severe Transmissio malaria, knowledge of n and malaria transmission and prevention malaria prevention. Respondents’ knowledge of malaria treatment (for both: P. falciparum and P. vivax) including, names of Knowledge- antimalarial drugs, Treatment duration of treatment, consequences of treatment non-compliance, and banned/fake/counterfake drugs. VMW/MMW Practices refer to performance according to TOR: referrals, DOT, bed net use, outreach to patients Practices General population/migrants Practices refers to bed net use and treatment compliance. 40
VMW/MM W
Ordinal
Questionnaire/F GD/IDI
Migrants, VMW/MM W
Ordinal
Questionnaire/F GD/IDI
Migrants, VMW/MM W
Ordinal
Questionnaire/F GD/IDI
Migrants, VMW/MM W
Ordinal
Questionnaire/F GD/IDI
Migrants, VMW/MM W
Motivation
13
Motivation
Acceptability
14
Acceptabilit y
Referrals
15
Referral
Stock List
16
Stock
Refers to the respondent’s perceived benefits to work as a VMW/MMW, motivation to work as a VMW/MMW, problems faced in the job. Refers to acceptability and trust from the community, acceptability and support provided by health centre staff. Refers to respondents’ practices on identifying malaria suspects in the community, number of referrals, barriers faced when referring patients, and referral records. Refers to respondents’ stock levels of all antimalarial drugs and expiry dates on each type.
41
VMW/MM W
Likert scale, ordinal
Questionnaire/F GD/IDI
VMW/MM W
Ordinal
Questionnaire/F GD/IDI
VMW/MM W
Continous, Ordinal
Questionnaire/F GD/IDI
VMW/MM W
Continous, Ordinal
Questionnaire/F GD/IDI
Record of recent 17 cases
Record
Malaria-related BCC/IEC materials
18
BCC/IEC materials
19
Interperson al communicat ion
Interpersonal communication
Refers to respondents’ possession of an outpatient register book, whether cases are being recorded and reported to the health centre. Availability of malariarelated BCC/IEC materials, use of BCC materials, display of BCC materias
VMW/MM W
Ordinal
Questionnaire/F GD/IDI
Migrants, VMW/MM W
Ordinal
Questionnaire/F GD/IDI
Share malaria topics with Migrants, the community, which VMW/MM topics, with whom W
Ordinal
Questionnaire/F GD/IDI
42
QUALITATIVE METHODS A total of 60 FGDs (416 participants) and 65 IDIs were conducted through October 2011 till January 2012, including VMWs/MMWs, public health facility staff, village chiefs, villagers, migrants and malaria patients. Non-probability sampling/purposive sampling was used to select all FGDs and IDIs participants. The qualitative assessment was conducted in 8 HC within and 4 farms within 5 ODs. Qualitative data was gathered through focus group discussions, in-depth interviews and observational data. Thematic analysis was used to identify common themes among the participants. Sampling Frame and Sample Size Community members were selected in villages in the catchment HCs. All villages included in this study were VMW-villages. Two villages were selected per HC. In each village one male and one female FGD was conducted. A total of two male and two female FGDs were selected per HC catchment area. The selection process of the villages ensured gender and geographical balance in the sample selected. The sample size for community-FGDs was 16 for female and male participants respectively. A total of 36 community-FGDs were conducted. The selection of migrant workers was done by farm. The sample size for migrant-FGDs was of 8 for female and male participants respectively. A total of 16 FGDs were conducted. The selection of VMW and MMWs for the FGDs was done by OD as there were fewer than 5 VMWs/MMWs at a HC level to conduct the sampled number of FGDs 15. Two male and two female FGDs were conducted in each OD. Sampling for VMWs: 8 ODs x 2 (1male:1female) 16 FGDs x 8 members = 128 participants. A total of 65 IDIS were conducted among farm owners, MMWs, VMWs, malaria patients, HC staff, community members and village chiefs. Inclusion/Exclusion criteria Health Centre Staff There is one staff per health centre responsible for VMW/MMW performance and for coordinating VMW/MMW activities at the villages pertaining to the respective HC. This person was interviewed after voluntary consent to participate was obtained. A total of 8 HC staff IDIs were included in this study. Community members To be able to participate in the assessment, community members must:
Most HCs ranged between 3-10 VMWs and the number of MMWs was even fewer. It was not possible to separate MMWs by gender so most interviews were IDIs. 15
43
•
Be residents of that community for the last two years (in order to ensure exposure to the
VMW programme). •
Voluntarily agree to participate in the assessment
•
Age 25-45 years old
Farm owners For each selected farm, the farm owner was interviewed after voluntarily consenting to participate. Village Chief For each selected village, the village chief was interviewed after voluntarily consenting to participate. Migrant workers To be able to participate in the assessment, migrants workers had to be: ➢ of the selected farm ➢ voluntarily agree to participate in the assessment ➢ between 18-65 years of age MMWs For each selected farm, the MMW was interviewed after voluntarily consenting to participate. Tools Focus group discussions and in-depth interview guides were developed with a series of questions and prompts for the facilitator to use. These tools were pretested in a farm not selected for the study. The facilitator asked questions to the group and allow time for participants to respond to each other’s comments. The duration of each FGD and IDI was an average of 1.5 hours. These guides served as a checklist during the interview and aimed to ensure that fundamentally the same information was obtained from the different participants. The main sections of the discussion guides were: access to malaria treatment, mosquito nets, forest goers, health education, and conclusions.
The location of the FGDs and IDIs was crucial in order to have the desired participation from the members of the group 16. Focus group discussions were held in a location which provided privacy and where participants were comfortable. The group was set up in a circle so that all participants had eye contact with each other and can clearly hear each other speak. Focus group discussions were conducted by the moderator and the moderator’s assistant who were trained on qualitative techniques and use of the guides designed for the study. Language used was Khmer in all discussion groups which is a language I can fluently speak. All FGDs and IDIs started with a
16
http://assessment.aas.duke.edu/documents/How_to_Conduct_a_Focus_Group.pdf
44
general question and which was easy to discuss and then more complex topics were raised as the discussion progressed.
Since all discussions were taped, permission was sought from the
participants before taping. The facilitator introduced the note taker and explained her role. Table X was used to define the role of the moderator and the moderator’s assistant during the data collection phase. 3.6
Responsibilities of the two Moderators17
Moderator
Assistant Moderator
Sets up equipment, arranges Supports the Moderator in setting up equipment and refreshments and organizes the interview organizing the interview room. room. Welcomes the participants as they arrive and hands out honorariums. Oversees data gathering, negotiating with the Assistant Moderator the level of detail of note-taking (to supplement and not replace mechanically-recorded data). Facilitates the discussion.
During the interview monitors equipment, welcomes late-comers and resolves interruptions. Takes notes throughout the discussion for the purpose of debriefing (as negotiated with the Moderator). Does not take part in the discussion unless exceptionally requested
Thanks participants.
Looks through points/issues.
notes
and
summarises
key
Debriefs the session with the Assistant Contributes to debriefing immediately after the Moderator immediately after the interview. Supports ongoing data analysis process. interview. Transcribes and analyses interview data.
Data Management FDGs and IDIs were recorded, fully transcribed and translated into English by a hired team of trained translators specialised in anthropological and health research.
Data process and analyses FDGs and IDIs were recorded fully and transcribed in Khmer and translated into English by a team of trained translators. The translation was cross-checked by other members of the same team. Thematic analysis was used. Data analysis consisted of examining, categorizing, tabulating or recombining the data collected during the discussions to address the initial aim of the study. Responses were analysed reading the FGD/interview responses and arranging them in the general categories. After the responses are arranged, the different opinions were identified and 17
http://www.llas.ac.uk/resources/gpg/2399#toc_5
45
summarised. The themes or patterns that emerged were synthesised. Four main sources of information will be used in the analysis: the assistant moderator’s notes, memory of the focus group facilitator, note taker and technician, and the audio tape-recording of the session Transcripts and notes were assigned codes when relevant information was found. The codes aimed to identify text which represented a frequent idea or viewpoint in relation to one of the key questions of the study. Following coding, text was analysed using NVivo9. Finally, these themes were interpreted through the technique called analytic induction, which involves a summary statement which is true of each extract in the group. These statements are key themes which are communicated in this thesis. Informal survey of farms in Pailin An informal survey was conducted in Pailin in March 2012 to gain greater insight on the farm owners’ perspective of the lending scheme, to interview taxi drivers and their views on their job and to interview FHI360-PLN staff on the current IEC methods used in the farms. QUANTITATIVE METHODS Two cross-sectional surveys were undertaken consisting of total of 498 semi structured interviews. This included 196 VMWs/MMWs and 304 migrant workers. Assessment design and sample size Mobile/Village Malaria Workers For the VMWs/MMWs, a cross-sectional individual-based interview survey was undertaken. In zones 1 and 2 there are a total of 534 VMWs and 108 MMW villages attached to 43 HCs within the CNM containment area. Sampling was stratified by zone in order to detect any differences between the two zones in terms of satisfaction, malaria knowledge, performance, challenges and delivery of key BCC messages. The quantitative team visited six ODs in six different provinces located in zones one and two of the Containment Project as described in Table X. Two operational districts (ODs) were randomly selected from zone 1 and three for zone 2. Within these ODs, only HCs with 10 or more VMWs/MWMs were selected to optimize the number of VMWs interviewed per session (see Table X). The study was originally based on an expected 85% performance; that is, within a total of 534 CNM-VMWs and 108 CNM-MMWs and an expected outcome indicator of 85% of contact points (VMW/MMW) providing malaria diagnosis, treatment, prevention, and key messages to mobile/migrant populations; VMW/MMW are aware of key messages; and VMW/MMW who perform according to the VMW TOR, and Type I error of 0.05 and 10% of non-response rate. The sample size required was 165, yet during the implementation phase, the sample size calculation
46
was revisited and an expected 80% performance was considered to be more appropriate 18. The sample size was therefore increased to 196 participants, as described below. Within a total of 534 CNM-VMWs and 108 CNM-MMWs within the study area and an expected outcome indicator of 80% and Type I error of 0.05, a sample size of 178 was needed. In order to calculate for 10% non-response rate (18 participants) a total of 178+18 = 196 participants were estimated. Migrant Workers For the migrant workers a cross-sectional survey was conducted in 28 farms in Pailin province of the Kingdom of Cambodia (Figure 13). The study has covered specific target areas, with an average of 20-100 migrant workers in each farm. In the containment survey 2009, the main outcome indicator is the use of LLINs the previous night before the survey, which was 60%. Therefore, since estimates of net use are already available for the local population. Sample size calculations show that in order to demonstrate an improvement of 15% of the proportion of the main estimate/outcome indicator, which is LLIN use (60%) with a power of 80% and at a significance level of 5% a sample size of 304 migrant workers is needed in order to detect a 15% difference between before/after intervention. Towards the end of 2011, 40 MMWs were hired and trained to operate in 40 farms out of the 86 estimated farms in Pailin 19. However, at the time of the survey, there was a shortage of migrant supply20 and migrants were only found to be working in 28 farms. On these farms, a much smaller number of migrants than expected was found, and therefore, to reach the required sample size (n=304) 80% of the total number of migrants were randomly selected from a list of names provided by the farm manager. Each farm was considered as one cluster. Study Indicators Four main outcome indicators were assessed among village and mobile malaria workers: - Percentage of contact points (VMW/MMW) providing malaria diagnosis, treatment, prevention, and key messages to mobile/migrant populations; - Percentage of community level staff (VMW/MMW) who are fully trained; - Percentage of community level staff (VMW/MMW) who are aware of key messages; - Percentage of community level staff (VMW/MMW) who perform according to their TOR (see annex 2). For migrant workers, the below outcomes were measured: The CNM VMW programme manager suggested bringing down the expected performance to keep it consistent with the terminology used by donor agencies. Hence, after many discussions it was decided that 80% performance was accepTable. 18
19FHI360 monitoring
program data 2011 Two possible reasons: the situation on the Thai-Cambodia border improved significantly so possible migrants rather went to work on the Thai farms as it provides a better income for them. Other possible reason is that there were heavy floods throughout the country so migrants could not come to Pailin due to the flooded roads. 20
47
-
Percentage of migrant population in the targeted farms who slept under a long lasting insecticide-treated net (LLIN/conventional treated net/LLIHN) during the previous night of the survey; - Percentage of migrant workers who would seek diagnosis and treatment of malaria (from a public health facility); - Percentage of target population at risk of malaria who know the cause for malaria; the symptoms for malaria; and the treatment for malaria; - Percentage of target population at risk of malaria who know the preventive and control measures for malaria. - Percentage of target population who are informed about the key messages about malaria prevention and control provided by MMWs. Inclusion/exclusion criteria VMWs/MMWs • Be a CNM-VMW21/MMW22 within CNM containment zone working since 2009 • Agree to participate in the study Migrant Workers • Individuals 18 years and above • Migrants staying in one of the target farms: for a minimum of one week • Informed consent obtained • Agreement to participate in the study Field team composition and training Just prior to the field work, a three-day workshop took place. Training was given on the purpose and exact procedures of the interviews and note taking as well as conducting pilot interviews. In addition, a detailed guide with the standard operating procedures was prepared that will support the field team in the interviews. There were a total of three teams with two members per team respectively. The study tools have been developed in English and then translated into Khmer language. The questionnaire’s translation has been verified and cross-checked to ensure accuracy. It has been reported that the majority of the migrants speak Khmer standard language so cultural mediators were not needed. All study tools were pre-tested in a farm not selected for the survey, nearby the survey site. The reliability and validity of the survey questions were pre-tested. In order to test for reliability, answers provided by respondents in one pretest were compared with answers in another pretest. Subsequently, questions that needed further refinement or re-writing or did not add value were identified and noted for subsequent actions. All suggestions from respondents and ideas on the words or sentences which they find difficult to understand, once the questionnaires had been pretested, were reviewed and revised accordingly.
21 22
See annex 2 for VMW ToR, selection criteria and selection process See annex 3 for MMW ToR, selection criteria and selection process.
48
Study tools consisted of semi- structured questionnaires (see annex XX). The questionnaire lasted about forty minutes. Information was collected on household characteristics and VMWs/MMWs knowledge about malaria (causes, prevention), sociodemographic characteristics, training, malaria knowledge and recognition, malaria treatment, motivation and acceptability, referral practices, stock list, record of recent cases, malaria-related BCC/IEC materials, list of malaria-related BCC/IEC materials at the VMW/MMW house (Annex). Data collection The local authorities (village leader and commune chief) and community leaders were informed of the purpose and expected time of the study. Their approval and cooperation was sought in every aspect of data collection. In each interview, trained interviewers used the questionnaires in the form of a structured face-toface interview in a consistent format. Data for surveys was collected at an individual level for the migrant population. After the selection of individual respondents, the interview took place. Each interviewed individual received a unique identification number. Questionnaires were pre-coded with these ID numbers to avoid issue of double numbers. Data management and analysis Data was entered into a database format (EpiInfo database designed for this survey) by trained persons and cleaned. The data was then exported and analysed in STATA 11 for Windows where basic frequencies and simple proportions were calculated. Mantel Haenszel chi-square test or Fisher exact test for significance were calculated. Logistic regression to adjust for potential confounding factors was also used in those associations identified significant (P-value =
