According to the World Health Organization, an estimated 80 percent of the South African population has latent Tuberculosis (TB), with about 450,000 active TB cases in 2013 alone. Last year, TB deaths overtook those of HIV/AIDS, making it the world’s deadliest infectious disease.
A significant number of infectious TB cases (about 37.5% globally) go undetected and therefore untreated for years – an issue referred to as the detection or treatment gap. These cases continue to spread infection creating a vicious cycle of infection propagation within populations. Identifying these cases will advance health officials’ ability to treat the infections – which is why IBM is supporting the World Health Organization’s End TB effort with a data-driven, Internet of Things to identify, predict, and curb TB’s spread.
One of the biggest challenges in tracking the infection is the reporting system administered by various health organizations; surveys, which need to be manually filled in, are often forgotten about, dismissed, or incorrectly filled out. Additionally, particularly in South Africa, having TB can be considered a social taboo, so many people aren’t willing to volunteer such information.
IBM researchers in the newly opened Johannesburg lab are tackling the challenge with data-driven approaches, with the intent to design more-effective TB prevention and control strategies. Using inexpensive radio frequency (RF) tags, the team is piloting the deployment of two dozen devices in a specific area in Johannesburg to anonymously trace transmission of the disease.
Researchers Toby Kurien and Darlington Shingirirai are working together in the lab’s Braamfontein headquarters to create the RF tags to apply toward business use cases for hospitals, clinics and other organizations, and eventually deploy them into the community. Kurien, well known in the local maker community, built the trackers. Darlington, a PhD in bioinformatics joined IBM Research – Africa eight months ago, and brings experience in healthcare research.
A better way to tag TB
Typically, RFID trackers are confined to one particular area, and can only be tracked within a specific location, like a shopping mall. The readers have limited range so it’s impossible to determine who was in contact with whom, just who was in an area at a specific time – to be alerted about a discount or some other special, in the case of a shopping mall example.
Kurien and Shingirirai’s tags solve this range and communication problem by communicating with each other.
“Our tags are designed to talk to each other, so that when one comes in contact with another, the interaction is recorded,” Kurien says. For example, a person wearing a tag may come in contact with 10 people at work, eight on the bus, and three at home. If everyone is wearing one of these tags, each interaction is recorded.
The data from the tags is generated for analysis on a 3D, motion-detecting dashboard display that can be used by scientists and clinical researchers to pinpoint heavy concentrations of the disease, or trace the spread from a particular location to others. Because of the high cost of vaccinations, health officials need data to prioritize immunization targets.
“As the tags collect data, we create visualizations of their interactions to identify large clusters or communities,” Kurien says. “When we notice a large cluster, it means that one person has come in contact with more people than most, and we call them ‘super connectors,’ or people with high centrality. The data suggests that if we select a super connector to immunize, we disintegrate the network and prevent the spread of the disease, further.”
“Using these devices, we will be able to understand the dynamics of infection transmission patterns in diverse population settings, which is critical for predicting the spread of disease among different individuals,” Shingirirai says. “This knowledge is essential to identify specific mechanisms and routes that favor disease transmission between and within groups in different social settings, and allows us to devise more effective intervention strategies, such as targeted immunization (Isoniazid Preventative Therapy), intensified case finding, infection control and other measures aimed at preventing the spread of disease.”
Aside from misrepresented survey data, TB is also difficult to track because traditional tools like GPS sensors are not only expensive, but people aren’t willing to wear them. Tracking devices for medical data are typically perceived as a violation of privacy, and there’s also sensitivity about displaying a physical object that others would recognize as a disease tracker.
With this in mind, the team has built their third rendition of the device, which is the smallest yet. “Each board contains a small sensor, storage device and battery,” Kurien says. “We can get this small enough to wear as a band or bracelet so that it eliminates the stigma and is more fashionable for people to wear.” Shingirirai suggests that in order to anonymize the devices even more, they could be designed as headwear for women or other types of wristbands for men.
The team came up with the concept and worked with a local tech start-up, SiGNL, that manufactured the boards. They intend to pilot the trackers first in a controlled environment, like a hospital, then deploy the trackers in a larger sampling of the general population. The solution can also be applied in tracking infectious diseases other than TB.
“If we can learn more, then we can begin to understand more,” Shingirirai says. “When we understand more, we can take positive action – in this case create a bigger impact in the fight against TB.”
In support of the WHO’s End TB strategy, IBM’s infectious disease tracking service provides an innovative and accessible tool designed to contribute valuable data for improving global health security, while removing social and cost-related barriers to existing tracking techniques.