Using SMS alerts to stop preventable disease

PATH is scaling text-based data reporting to improve health workers reports and contain malaria outbreaks.

Preventing community spread through SMS

In Zambia’s Southern Province, the fight to end malaria begins with a knock on the door. An army of community health workers, dedicated volunteers chosen by their communities, are trained and equipped to test and treat their neighbors for malaria.

A positive malaria cases triggers a targeted response: the community health worker visits the home of the patient and tests everyone there. If anyone tests positive for malaria they are given treatment.Then she expands the search for the malaria parasite out to 140 meters—about the length of a football field.

Zambia is using this community surveillance approach because malaria infection tends to cluster. It’s the latest step in the country’s long fight to end malaria for good.

Delivering support efficiently

Ten years ago, targeted attention to individual malaria cases was not happening. Back then, Zambia was demonstrating that by scaling up nationwide the proven tools to control malaria—bed nets, insecticides, diagnostics and drugs—a country could put a huge dent in the disease burden. Because of these efforts, malaria in children was cut in half.

The gains were encouraging, but Zambian authorities weren’t satisfied. To refine the attack on malaria, the Ministry of Health sought to better understand the situation on the ground and established a weekly malaria reporting system. Data from more than a 1/3 of the country’s health facilities—malaria cases and the number of test kits and treatment remaining on the shelf—are now fed into a national database.

Meanwhile, massive household campaigns revealed that a lot of malaria was being missed. In some communities, more than 80 percent of malaria cases lacked symptoms. While an individual may not look or feel sick, a mosquito can still suck up parasites during a blood meal and transmit them when she bites again later.

To combat this latent transmission, Zambia introduced mass drug administration: test everyone in an area, regardless of symptoms. It worked. Malaria fell dramatically in the study area: an 87 percent reduction including a 97 percent reduction in facility-reported deaths.

Now Zambia is taking another big step toward malaria elimination by using data analysis and community-based outreach to identify and attack the disease. Health workers testing in their communities meet every month at the local health facility to share data. After checking for quality, the data is submitted into the national system. Thanks to these efforts, Zambia has rich, near real-time data at all levels and can use that information to guide strategy. 

And as the country pushes toward malaria elimination, success has been marked by unwavering political commitment and steadfast collaboration from a range of partners at local to global levels. PATH, a global health non-profit based in Seattle, has been working in Zambia since 2005. Their primary focus is supporting the National Malaria Elimination Center, from data collection and management to research studies to engaging communities during routine and campaign malaria activities.

PATH is dedicated to innovation in public health, and that extends to initiatives such as Visualize No Malaria, a partnership with the Tableau Foundation. Along with a coalition of technology companies, including Twilio, PATH and the Tableau Foundation seek to make Zambia’s rich and growing base of data rapidly accessible and high quality for decision-makers and those on the frontline.

In collaboration with the Ministry of Health, a team at PATH identified data quality and analysis as key areas for action. The team implemented a Tableau Server to optimize reported data that was updated both weekly and monthly. Using native Tableau features such as data joins and blends, they were able to offer new insights on malaria case occurrences and trends. Through Tableau dashboards, a senior official at the Ministry of Health in the capital Lusaka could see what is taking place in districts, health facilities, and even the area where a community health worker operates. These visualizations offered a fresh perspective on the disease, but data gaps remained.

Designing the right solution

To improve data completeness, the team developed a program called Viz Alerts, and integrated Twilio communications. Based on input from district-level user groups in Zambia—engaged to ensure the solution would meet their needs—it was clear that any technological fix must blend existing systems and user behavior. Ultimately, PATH chose SMS for communication because of its prevalence among health workers and the simplicity of the format.

Viz Alerts is an open source Python script designed to identify data gaps, alert the local health workers of those gaps, and to integrate missing data when provided by health workers. When the Viz Alerts algorithm detects an incomplete dataset, the script prompts an SMS message via Twilio to the appropriate health worker to submit the missing report; the health worker can quickly reply by submitting the missing data over SMS. Viz Alerts incorporates the missing data automatically, and users can access the updated dataset through Tableau visualizations.

Automated messaging and data collection saves the Ministry of Health significant time reviewing reports for incomplete data, providing health officials more time to analyze results and respond to incidents. Most importantly, those working on the ground quickly saw the value of the platform. Marianne Soko is the malaria focal point person in Itezhi-Tezhi district, located some four hours west of the capital Lusaka. “This is really great because the alert allows me to follow up on the missing reports,” said Marianne, “and they can be updated both at facility and community level.”

Jeff Berson agreed. “Seeing the drop in missing reports was a sign that we had tapped into something potentially transformative,” said Jeff.

PATH is now scaling the program across Zambia’s Southern Province and aiming to expand it to Western Province and beyond. With this scale, the team is working to further improve the data collection process. Tracking data submissions over time will allow them to identify reporting patterns and improve message targeting.

“Much in the same way businesses motivate their employees through messaging,” says Jeff, “we see Twilio as a powerful channel for the Zambian Ministry of Health to do the same with frontline health providers.”

Jeff Bernson is PATH’s Senior Director of Results Management, Measurement, and Learning. He’s been in this business a while and said the approach that resulted in Viz Alerts was critical. “Involving and listening to users early is key,” explained Jeff. “The technology is powerful, but understanding how to frame the message, the content and the frequency are all keys to success.”

The team’s initial research with the user group helped the SMS message content, frequency, and target audience.

Viz Alerts was introduced this year across 200 health facilities and the results are strong: missing report volume dropped more than 21 percent.