Happy Nice Mall

Please consider to support MIT Technology Review journalism subscribe to.

FIND monitors whether the tool’s recommendations are working and are being implemented. “This should be something that countries can have themselves,” Albert says. “One important thing is to put the power of data in the hands of people. [who] He has to make decisions.”

Catching up with smartphones

BroadReach uses a simulation tool called Vantage, created in partnership with Microsoft, to identify understaffed clinics and send healthcare workers where they’re needed most. In 2020, during the first few weeks of the pandemic, BroadReach worked with FIND to assess how prepared the two provinces in South Africa were for covid, citing a shortage of protective equipment and staff at more than 300 clinics in just three days.

Before going to medical school, Sargent says she learned firsthand about African health systems while working in refugee camps. He later founded BroadReach with his doctor friend, Ernest Darkoh, who grew up in Tanzania and Kenya. “You can go to rural clinics in places like Zambia and you’ll find patient records are on paper,” Sargent says. There’s technology, though: “Nurses use smartphones and Facebook suggests posts they might like.”

In addition to monitoring deficiencies, BroadReach continues to follow up on this by following individual patients in more than a thousand clinics in various African countries to see if they are getting the treatment they need. Clinics already do this, but Vantage also uses machine learning trained with hundreds of thousands of anonymized clinical records and social data to identify people who are likely to discontinue treatment and recommend that healthcare professionals proactively engage with them.

The Nigerian Institute of Virology used Vantage in 2021 to estimate which of the 30,000 people treated for HIV at three sites in Nigeria are at risk of going off their medication. As a result, the tool found that 91% of those called or visited by a healthcare professional had up-to-date information about their medication, compared to 55% of those who were not contacted.

Health professionals at a number of HIV clinics said the tool helps them build closer relationships with their patients by helping them focus on those who need intervention most, according to BroadReach.

The so-called developed world

BroadReach now wants to make its software available in the US. “At the time of Covid’s hit, we kind of woke up and realized that many of the unquoted advanced world health systems weren’t all that good and that a large part of the population was lagging behind,” Sargent says.

BroadReach is involved in four pilot projects with US healthcare providers and insurers. In one, he struggled with low vaccination rates in parts of Colorado, using machine learning to predict where vaccination sites will be set up and which communities should be screened. Local health officials had assumed that resources should be focused on urban areas to vaccinate as many people as possible. But Vantage found that focusing on low-income rural minority communities would have a greater impact.

BroadReach also works with an insurance company in California, which sees significant disparities in the way people in various groups adhere to a treatment regimen of statins, drugs used to treat high cholesterol. BroadReach wants to look at the data to identify possible explanations for what the insurance company has found. Some communities have poor transport links to clinics, which can prevent people from visiting their doctors to update their prescriptions. Sargent says others have long lacked confidence in the healthcare system.

Finally, he wants to see Vantage predict risk factors for individual patients. For example, for a Spanish speaker who doesn’t live near a clinic, the software recommends that the insurer provide a Spanish-speaking social worker and a coupon for the taxi, he says.

However, getting the data needed to train AI to make such predictions accurately is difficult. In the US, healthcare data is generally not shared between providers. Sargent says BroadReach solves this problem by combining medical records with people’s socioeconomic data, such as zip codes and credit histories. “We have partnerships with consumer data companies because you can tell a lot about a patient if you know their behavior patterns and the conditions in which they live,” he says. “We’re putting all this together to get a view of the general population and each patient.”

How people feel about this type of oversight will depend on what benefits it actually brings them. A number of organs, including loan companies, recruitment agencies, policeand more—already use the socioeconomic data BroadReach leverages to predict the likely future behavior of individuals. The prejudices in these systems have rightly led to strong rejection from civil rights groups.

Nicholson Price, who studies legal and ethical questions about the use of personal data at the University of Michigan, says that the government’s proposals to share medical data in many countries, including the US, UK and Australia, have sparked backlash. But that hasn’t stopped companies from combining medical and consumer data. “Companies have been doing this for a few years only at a lower profile,” Price says.

“There’s a sense of surrender that it’s happening and we don’t seem to be able to stop it,” he says. “However, maybe there is a silver lining that good things will come out of it, rather than just being advertised and manipulated.”