Surveillance technology in Thailand, Cambodia, and Tanzania: Case Study from Ending Pandemics

Introduction

San Francisco-based nonprofit Ending Pandemics works with partners in lower-income countries using innovative methods to prevent, detect, and contain disease outbreaks. Its collaborations take a “One Health” approach, which starts from the evidence that because three in four emerging infectious diseases arise in animals before spreading to humans, detecting disease in animals faster can prevent outbreaks in humans. This approach also enables earlier detection of and rapid response to outbreaks that spill over into human populations.

Working with local organizations, Ending Pandemics supports the development of technology-enabled surveillance systems that rely on community input. These systems include mobile applications in Tanzania and Thailand and a digital hotline in Cambodia. In both cases, Ending Pandemics and its partners hosted in-country EpiHacks^TM (epidemiology hack-a-thons) to engage local software developers in the challenge of building efficient, robust surveillance systems. These sessions also generated buy-in from local health professionals, communities, the private, and other key stakeholders. As implementation partners continued fine-tuning the technology, local partners began recruiting community volunteers and raising citizen awareness to empower people to use the systems to the greatest effect.

Thailand’s Participatory One Health Disease Detection (PODD) project launched in 2015 and Cambodia’s 115 Hotline launched in 2016. Both systems have effectively reduced detection and response time and limited the spread of animal disease, clearly establishing that user-friendly technology alongside community mobilization can facilitate successful “bottom-up” disease surveillance.

The two systems have since been adapted to serve other public health purposes, including early warning of weather shocks and, more recently, COVID-19 surveillance. With continued support from Ending Pandemics and buy-in from other local governments, they are being extended and replicated in other countries. In Tanzania, a project built on the PODD experience and led by the SACIDS Foundation at Sokoine University of Agriculture is demonstrating promising early results.

Context

In 2004, Thailand was at the center of an avian flu outbreak that killed 12 people.¹  Dr. Lertrak Srikitjakarn, former dean of Faculty of Veterinary Medicine, Chiang Mai University, believed that more such outbreaks would be forthcoming, due to the number of families that relied on backyard animals for survival. Dr. Lertrak and the president of Ending Pandemics, Dr. Mark Smolinski, began working together on PODD in 2014, bringing on a Bangkok-based technology social enterprise Opendream as a third partner.

In 2015, Ending Pandemics began working with the SACIDS Foundation, a longtime partner, at Sokoine University of Agriculture in Tanzaniato replicate the PODD approach there. Unlike PODD in Thailand, which focused on backyard farms, Tanzania’s initiative, called AfyaData, centered on Maasai herding communities. Like PODD, AfyaData was initially developed through an EpiHack that brought together local stakeholders from the health and technology sectors. Subsequently, it was championed at SACIDS by Dr. Mark Rweyemamu, Dr. Esron Karimuribo, and Mr. Eric Beda.

In Cambodia, where contact between animals and humans resembles the patterns in Thailand, the One Health approach made just as much sense. Ending Pandemics, the Cambodian CDC in the Ministry of Health , and a local technology partner InSTEDD iLab Southeast Asia worked together to develop a phone-based technology solution, the 115 Hotline, linked to local communities and the formal health system.

Detection

Thailand

Dr. Lertrak and his team worked with public health and livestock officials in Chiang Mai Province to pilot PODD, and in the past five years it has scaled up quickly. PODD is currently in use by 410 local governments across 30 provinces (out of a total of 76 in the country)^{2  3}, with significant coverage in Chiang Mai, Chiang Rai, and Khon Kaen provinces.

Local residents are educated about the initiative so they can report all potentially relevant human and animal health events to volunteers, who then submit that information electronically through the PODD smartphone app. To aid in community education and mobilization, the acronym for the project was carefully selected: “poh-dee-dee” in Thai means “look closely and you will see.”

The volunteers, who are selected by local government leaders, are trained to use the app. Some volunteers already belong to Thailand’s Village Health Volunteer network, but this is not a requirement. The partners took great care to design an easy-to-use app, with a recent survey showing that 89 percent of users can navigate it effectively after basic training even though half had never owned a mobile phone. To date, 19,400 volunteers have received training.²

The reports filed electronically by the volunteers are triaged by analysts, usually Chiang Mai University staff and graduate students in epidemiology. They verify the reports, either by contacting the volunteers to ask follow-up questions or, increasingly, by using newly developed automated processes. Then they route them to the appropriate local public health and livestock officials who implement containment measures as appropriate, including diagnostic testing, vaccination and treatment, and community education. When the outbreak is detected earlier, less resource-intensive containment measures are more likely to be effective.

Within the first six months of PODD’s launch in a subset of districts in Chiang Mai, more animal disease events were reported in just those districts than had been reported in the entire province the previous year. Another early success was a rapid response to foot-and-mouth disease in 2016, which saved an estimated US$4 million for the local economy. One study of PODD found that 73 percent of confirmed poultry outbreaks were contained within the villages of origin, with the remaining 27 percent contained within neighboring villages.⁴

As local Thai governments have seen the success of PODD, more and more have chosen to pay for it. Where PODD has been implemented, about half of local governments provide the funding.

Tanzania

Tanzania’s AfyaData tool is being piloted in two districts, and early results are promising. The Food and Agriculture Organization of the United Nations is already working toward integrating the tool with its mobile field reporting application. Like PODD, AfyaData is a set of digital tools that empowers community members to report human and animal health events to the public health and veterinary experts tasked with responding to them. The tool helps the experts receiving reports analyze the data quickly, visualize them, and present them in different formats, including graphs and maps. The dashboard provides a quick overview of intervention strategies. They also have access to the One Health Knowledge Repository, a compendium of expert-authored content that helps with early diagnosis and treatment of diseases based on reported symptoms. Finally, the tool is enhanced with an early warning system based on text messages (SMS or short message service) to provide key information to decision makers.

The SACIDS team has trained over 700 community volunteers and other health professionals to use the tool. From late 2015 to September 2019, more than 4,800 reports related to human health and nearly 8,000 reports related to veterinary health were submitted.⁵ The tool and the trainings have improved reporting timeliness at the community level from an average of ten days to an average of three days.

Cambodia

Through a partnership with the Ministry of Telecommunications and telecom companies, the 115 Hotline is free for anyone in Cambodia, even if their phone has no credits. The hotline serves three primary functions:

In the first four years of operation, the 115 Hotline received an average of approximately 500 calls per day, with 15 to 20 of the daily calls routed to surveillance officers for follow-up. Surveillance officers then decide which cases to send to the disease investigation and rapid response teams. The system has been used to verify hundreds of cases of both animal and human disease. Moreover, it has improved rates of “zero reporting” (when health centers formally report no cases of disease, rather than simply not reporting), which is important because it helps authorities pinpoint reductions in disease burdens and identify which health centers have not yet submitted reports. Before the rollout of the hotline, average on-time reporting across all provinces was below 80%. After the rollout, that average improved to well over 90%, with even greater improvement in provinces that had historically poor reporting.⁵ Vietnam is currently piloting a version of the hotline.

COVID-19

One of the advantages of these surveillance technology tools is their flexibility. They provide an easy-to-use link between communities and public health authorities no matter the issue at hand. For example, in 2017 Thailand successfully used PODD to convey a flood warning to communities downriver. When heavy rains resulted in flash floods, local volunteers used the PODD system to alert surrounding subdistricts to the threat, saving property and, potentially, lives.

As the COVID-19 pandemic emerged, PODD and the 115 Hotline proved immediately useful to detection efforts. AfyaData is now working with a multi-country regional disease surveillance network and Africa Centres for Disease Control and Prevention to scale up use of the tool across the continent.

The 115 Hotline has been scaled up considerably in response to COVID-19. It is the primary means of reporting suspected cases in Cambodia, with an increase from 500 calls per day to 10,000–18,000 calls per day. There are now 25 operators staffing the hotline for the Cambodian CDC.⁶  They document initial reports for contact-tracing teams that gather detailed information, enter data in the World Health Organization’s Go. Data platform, and refer clients to the appropriate facilities for care and sample collection. Samples are sent to the Pasteur Institute in Cambodia for testing, the result is sent back to the contract-tracing team, and in confirmed cases the hotline provides automated active monitoring services . The hotline has been the mechanism for detection of approximately 90 percent of suspected and confirmed COVID cases reported in Cambodia.

Conclusion

Community-driven surveillance platforms show promise in significantly accelerating the detection of disease in animals and humans. The results of this increased speed are:

The experiences described in this case study suggest that an important part of the enabling environment in countries that implement digital surveillance systems successfully is government capacity and commitment. This is especially important because governments ultimately have responsibility for disease surveillance, and these technology-based systems need to work alongside or be incorporated into existing government systems. Thailand, Tanzania, and Cambodia also have knowledgeable staff who triage incoming reports and functioning teams at district and province levels who can respond when they are verified. Countries have different levels of technology infrastructure, and the ability to choose between tools and approaches provides options depending on feasibility in the local context. Getting these solutions right requires a gradual process of piloting and scaling up. This process allows for fine-tuning the system and technology and demonstrating its impact to communities and governments.