Going Digital, Integrating Data Science, and Enlisting a Citizen Network
By Matej Novak, Head of Programs, Premise Data
It’s been one year since Premise partnered with Cali, Colombia on a digital overhaul of the city’s vector surveillance and control operations. In this blog we will discuss our accomplishments, the results of a data science experiment, and the city’s plan to enhance its capacity with a citizen network of data contributors.
The fight against mosquito-borne diseases, such as Zika, malaria, or chikungunya, requires field workers to cover vast terrain performing time and labor intensive activities such as vector surveillance (setting-up and harvesting mosquito traps, monitoring mosquito densities) and vector control (fumigation, larviciding, and spraying campaigns). The city of Cali is painfully aware of the investment required to monitor and abate mosquito (aedes) breeding sites. Colombia had the second largest Zika outbreak in the world in 2015–2016. Cali, a city with 2.3 million people, accounted for 23% of Colombia’s Zika cases in 2016 and is hyper-endemic for dengue. Vector-borne diseases have therefore been at the center of attention of the city’s Secretaría de Salud Pública (Public Health Department) for years. Despite the city’s focus, the aedes mosquito density index remains stubbornly close to 5%.
Digitization of Vector Control Workflows
A year ago, Cali started an ambitious undertaking utilizing Premise’s technology to modernize their vector surveillance and control efforts with mobile technology, data science, and analytics. The project, funded by a USAID Grand Challenge grant targeting Zika, first sought to digitize Cali’s vector surveillance program that monitors and larvicides the city’s sanitary sewer system. The sewer system has over 54,000 sewage openings (our blog about the initial stage is here), and has been a major contributor to mosquito density rates, especially during the city’s rainy season.
Since March 2017, over 615,000 sewage opening inspections have been performed with the Premise platform, painting a dynamic picture of aedes mosquito density across communas (city districts), with data granularity down to the level of street intersections and sewer openings. Currently, the Cali Public Health Department uses the data for directing larvicide interventions and planning field workforce routes and coverage. Access to highly granular data has dramatically improved the efficiency of Cali’s identification of vector control “hotspots,” which allows them to rapidly deploy fumigation and larvicide treatments to targeted locations.
More surveillance and control workflows are now being digitized by Premise. The Public Health Department’s next focus with Premise’s technology will be fumigation campaigns, house visit monitoring campaigns, and public institution monitoring.
Investment in Integrated Mobile Technology
Cali has invested substantial time and resources to maximize the potential of the USAID funded project with Premise. One area where Cali Public Health has recently made upgrades is with investments in their vector surveillance program mobile devices. The integration of Premise’s technology started on older mobile devices that were donated several years ago, but at the end of 2017, Cali invested in 100 new integrated mobile devices specifically for field data collection with the Premise app. A central operating console managed by IT staff tracks device location, battery life, and manages software updates on all devices.
Data Science Experiment
After several months of data collection in 2017, the Premise data science team started surfacing trends for Cali to utilize in their vector control operations. For example, it became clear that with the city’s blanket approach to larviciding, some low-risk places rarely got infected, while other areas were continuously driving risk. While those trends are less visible at the sector level, the trends become significant when reviewed at the more granular level of specific street intersections or sewer openings.
This insight led to a joint Premise-Cali larvicide experiment. The purpose of the experiment was to validate whether or not treatment frequency should be varied at the hyperlocal street intersection/sewer opening level as a function of the historical risk. In the experiment, several sectors with either historically high or historically low aedes prevalence (but bordering each other) were identified. Public Health Dept. field staff were instructed to inspect all sewer openings in each selected sector twice a week, which is more than a traditional workload, and to treat with larvicide as needed. Each time, the treatment frequency and aedes index was assessed.
The experiment confirmed that risk indeed is hyperlocal and the proportion of high risk places does not vary significantly at the sector level. The critical data comes from these sub-sector levels. Areas (intersections) at high risk rapidly declined in risk over the 4 week experiment. 90% of intersections that were treated biweekly were negative for aedes at each following reported visit. At the end of the experiment, the average aedes index of the highest-risk intersections reduced by 27 percentage points on average from 30.6% to 3.7%; while the aedes index average for low risk areas remained below 5% (3.6%).
Premise is hopeful that such data science driven analysis can be integrated into Cali’s vector surveillance and control work and lead to better outcomes for the community. More and longer term data will continue to improve the quality and quantity of insights such analysis can produce.
Enlisting Citizens in Cali’s Vector Surveillance
Even with a dedicated professional staff equipped with advanced technology, achieving sufficient coverage and regularity with a city-wide vector surveillance program is difficult due to the inherent time and resource constraints. Thanks to their new digitally enabled workforce, Cali’s Department of Public Health is working more efficiently than ever to monitor and combat Zika and other vector borne diseases. Looking to add further innovation to their program, Cali and Premise will now work to engage Cali’s citizens to help monitor and predict vector sites across the city.
Starting in early 2018, Premise will recruit and engage a citizen network in Cali to work side-by-side with Cali’s Public Health Department to monitor mosquito activity. Utilizing the public to monitor vector activity holds several benefits. It will drive down costs of surveillance, allow for rapid operational onset with hyperlocal precision, and generate complementary data for Cali’s Department of Public Health staff to incorporate into their analysis and workflows. The vision is to enhance the city’s overall vector surveillance program with an innovative government staff + citizen data collection strategy.
Premise’s platform will utilize its micro-payment capability to reward citizens for their submissions. After a submission has been processed and validated by a quality control protocol (both automated and manual, depending on context), a user can “cash-out” their rewards.
Premise has partnered with community leaders in two of Cali’s local districts (Communa 7 & 20) to organize focus-group discussions and user experience testing with local volunteers that have been instrumental in the city’s dengue-related operations in the past. This human-centered design stage resulted in an early prototype of a civic surveillance model, which is based on core groups of local activists at the community level. Those individuals will maintain mosquito traps in their backyard, provide regular data on mosquito densities from the traps, regularly walk routes suggested by the Premise app to identify breeding sites, and carry out community awareness house visits. Premise’s user-friendly app interface directs action and provides abundant hints and examples to maintain quality of submissions.
Cali’s Public Health Department has rapidly become a leader in technology enabled vector surveillance and control. Cali has digitized critical workflows, equipping their staff with innovative tools as they fight Zika in their community. Cali is also breaking new ground for hyperlocal vector control targeting by integrating data science experiments directly into their programs. And finally, Cali and Premise are busy setting the stage for a truly innovative government-citizen data collection partnership that will enhance the city’s capabilities against vector-borne diseases. Stay tuned for future updates.