PD has tremendous promise as an effective community mobilization and behaviour change tool in malaria control and elimination settings. Here, community feedback following a small-scale evaluation in selected villages of Cambodia has been presented; however, the approach has since been applied to sites in Myanmar and Thailand and similar feedback from community members has been received.
Instilling confidence in the applicability of this intervention across different settings. In all areas the PD activity was well-accepted into the community and created a high amount of motivation and empowerment among community members to perform the PD behaviours identified, which in turn was associated with positive outcomes of behaviour change and fewer malaria episodes. Of real promise, is the apparent sustainability of the intervention, with activities still performed in the target communities in Cambodia 1 year following the community handover seminar.
Suitability to malaria control and elimination
PD is particularly suited to targeting small, remote communities, which are a prominent feature of malaria risk in the GMS [
22], and to other areas reaching malaria pre-elimination and elimination status. Classic BCC interventions using media outlets, such as TV and radio, often cannot reach these communities due to poor electricity supply, network coverage, or the fact that those at risk are often out of doors for long periods. BCC messages are often not context-specific and communities can reach saturation point if constantly bombarded with them. Classic BCC approaches often follow a needs-based approach which means communities may be unable to find a way to meet and sustain the behaviour [
19]. PD overcomes these challenges by involving the community from the beginning and throughout the whole process, learning from them about their day-to-day environment and challenges faced, and identifying role model behaviours from within the community.
Maintenance of community participation and enthusiasm in times of diminishing disease becomes a great challenge, and practice of preventive behaviours can fall because there are very few or no cases and people forget how important/serious the disease can be; it is no longer considered a threat [
24]. The PD approach presented here was well accepted by the target communities, with high community engagement and motivation to participate throughout. The interactive nature of the activities and identification of role models from within each group in the community created a strong sense of empowerment among all community members and eagerness to take part in activities. It was easy to recruit volunteers to conduct the PD community sessions and individuals reported that they could more easily relate to the role models and behaviours they needed to follow. This sense of community cohesion should be a positive influence on uptake of behaviours even when number of cases is low. Thus, PD represents a strong community mobilisation tool that engages communities well even in elimination contexts.
In elimination settings, it is also highly important to find interventions that are sustainable over a prolonged period of time. Where countries have had successes in malaria control and reduced numbers of cases, there follows an inevitable refocus of donor interests to countries with higher burden of disease, resulting in less money and fewer resources available to sustain malaria control activities [
31]. Control and surveillance methods required in elimination stages are also often costly [
32]. Although resource intensive at the beginning, PD has the potential to be sustainable at very low cost (see below for limitations and challenges). High community engagement and empowerment, along with the symbolic handover of the project may well ensure sustainability of the intervention after activities by the project field staff have ended. Evidence of short-term sustainability was found in Cambodia when PD volunteers were still conducting activities 1 year after external support from our team had stopped. Investigation needs to be done to determine the sustainability over a longer time-frame and whether top-up sessions with the community or PD volunteer retraining would be of benefit, and at which point in time it would be required.
Furthermore, PD has the potential to foster other interventions such as participatory surveillance [
33], as demonstrated through the development of community malaria maps, which will become particularly critical in the last push to malaria elimination.
Limitations and challenges
A key limitation in advocating for the use of PD in malaria control is the lack of quantitative evidence that it has an impact on human behaviour, and a subsequent impact on malaria transmission and disease. Here, we report qualitative information from community feedback that there was behaviour change in the community, and respondents in turn associated this with a reduction in malaria cases; however, this was not tested and the drop in cases may be attributable to a variety of other factors, such as low rainfall, impact of other interventions, etc. Behaviour change is notoriously difficult to measure, but attempts can be made to quantify this through surveys and observational studies. However, behaviour change in itself is not the ultimate goal; behaviour change is only useful if it has an ultimate impact on malaria transmission. Thus, the use of epidemiological indicators to assess PD, such as malaria incidence and prevalence rates, needs to be explored. Challenges with the prospect of using behaviour change and epidemiological indicators are firstly, that the populations in the GMS are generally highly mobile, so may differ between baseline and end-line surveys; and secondly, that positive behaviours in these communities are already very common and malaria caseloads are already very low, so massive sample sizes are required in order to detect any statistical difference. An alternative may be the use of entomological indicators. Use of entomological indicators is already being explored for the assessment of PD in dengue control, through measurement of pupal and larval density indices, and this could also offer a potential quantitative method of assessing PD in malaria.
Efforts also need to be made to examine the cost-effectiveness of the PD intervention. The initial (but short) PD process is time and labour-intensive and requires individuals with some experience in qualitative research and data collection, and with community mobilization skills. There is also the necessity for high quality supportive supervision of volunteers during the initial implementation phase. However, should the intervention be successfully sustained by the community once external help has been removed, the long-term running of the intervention would essentially be cost-less. Short-term (1 year) sustainability was shown by the intervention in Cambodia. Longer-term sustainability needs to be investigated, along with a comprehensive cost-effectiveness analysis to confirm whether the intervention is indeed low-cost and high-impact compared to other community mobilization and behaviour change interventions. Efforts are now being made to assess this, as well as feasibility and costs of scale-up to a national level.
In areas that have had intensive malaria control activities, it may be difficult to identify positive deviant behaviours because the communities will already have high knowledge of malaria and use of preventive behaviours. In pre-elimination and elimination settings, universal coverage of nets would have been achieved and communities would have received multiple BCC initiatives to result in high practice of preventive behaviours. However, evidence from the GMS shows that even in areas where this has occurred, there remains some residual transmission [
34]. In this case, when other control interventions have been exhausted, it is important to find small deficits in prevention practices that can make any difference toward lowering residual malaria transmission. In the Cambodia context, community residents had a high amount of bed net use, however, by using nets to animate discussions, it became clear that there was a deficit in the correct usage of nets, e.g. to clear mosquitoes from the net, ensure it is tucked into the mattress/blanket, and to fix holes. Thus, although difficult, it was still possible to identify deviant behaviours within the communities that would not have been immediately apparent from use of standard surveys [
35].
Some aspects of knowledge and behaviour will still be a challenge to address, for example, some of the beliefs around the causes of malaria transmission and treatments for fever are based on cultural and spiritual beliefs and will be very hard to change; while others also rely on external factors, such as having enough money to travel to the health centre.
There are also challenges in targeting the most high-risk sub-populations within the target communities, usually males, forest workers, and MMPs [
22]. It was noted during our PD projects that males participated in the PD activities less frequently than females, because they were often at work or in the forest away from the village when PD activities are likely to take place. PD volunteers should encourage community members to cascade messages received at PD sessions through their families and social networks to ensure coverage of all community members.
Population movement in the GMS is highly complex, with a mix of short-term and long-term migrants, high internal mobility as well as international, and frequent border crossing; presenting a challenge for targeting them for PD [
36]. During the PD pilot in Cambodia, for example, the target villages had a consistent influx and efflux of migrant workers who usually only stay for up to 30 days, and a maximum of up to 3–4 months. New migrants may not have been exposed to the initial PD Process nor have been involved in any of the monthly PD activities held by the volunteers. For PD to be effective in target communities, regular PD activities need to be sustained and targeted to newly-arrived mobile and migrant individuals and families.
Future application of PD
Despite these challenges, the evidence provided here, the appropriateness of PD for malaria control and the recorded impact of PD on other health-related issues gives ample hope that the intervention can be successfully applied to malaria control and elimination programmes.
Following the initial proof of concept studies, better ways to assess the impact of PD are being explored, through quantifying behaviour change and measuring epidemiological and entomological indicators in target communities. This will give solid evidence of the impact that PD can have and will inform national malaria strategies of the effectiveness of PD as an elimination tool. A study into the use of PD for dengue control is underway in Myanmar, using entomological end points as a measure of impact, through measurement of pupal and larval density indices, which will gather evidence on the feasibility of these indicators. Furthermore, possible integration of PD for malaria, dengue, and other diseases and health issues should be considered, and it is essential to test the application of PD to target communities using points of access other than villages to maximize coverage of high-risk populations, such as at rubber plantations, farms, or other areas of migrant employment.
Next steps also include exploring the scaling-up and cost-effectiveness of PD at national level, as well as looking into the long-term sustainability of the intervention in target communities. In various national and regional meetings since this implementation, CNM and other NMCPs have showed their interest and commitment to use PD as an effective tool to reach out to high risk and hard-to-reach populations, particularly in elimination settings. As such, the scalability of PD and ability to integrate into national programmes is now being tested.