AI for Global Health

Munai, a health technology company in Curitiba, Brazil, works on a specific version of a common problem: a patient's condition changes through several small signals held in different systems, and no single view brings them together. The case study describes INTI, the company's clinical intelligence platform, connecting existing hospital data sources and delivering alerts, dashboards, and decision support inside the electronic health record rather than in a separate tool.

The case study reports more than 30 hospital partners, roughly 18 million medical encounters processed, and a multicenter validation of its clinical deterioration prediction across six hospitals. It also describes a gap that matters for implementation: the people who use the alerts and the people who buy the platform are not the same. Clinicians act on a signal; administrators decide whether the platform is worth paying for, and the case study describes the financial case being anchored in the reported cost of an avoidable deterioration event.

Bive works in rural Colombia, connecting underserved communities to preventive services. In this project it added an AI-supported eligibility step to an existing outreach model: provider data goes in, women eligible for cervical cancer screening are identified, automated voice calls reach them where connectivity and smartphones are limited, and community health workers follow up in person so that the existing provider delivers the care.

An evaluation by Universidad de Caldas reported large shifts in the participating community: attitudes toward HPV vaccine safety at 95.9%, up from 63.5%; HPV vaccination among girls aged 9 to 13 at 87.6%, up from 28.5%; and Pap smears at 97.9%, up from 67.2%. The more consequential finding was about ownership. Participating providers handed over their population databases and relied on Bive to run identification, contact, and follow-up rather than integrating the tool into their own routines. The project worked; the capability did not necessarily stay behind.

Jacaranda Health is a nonprofit working on maternal and newborn health with public health systems. Its PROMPTS program sends guidance and support to mothers by SMS, while a companion program strengthens frontline providers' emergency skills. The case study reports that by November 2021, at the start of its peer-reviewed evaluation period, PROMPTS had enrolled more than 750,000 women through over 900 health facilities in Kenya. A cluster-randomized trial across 40 facilities in eight counties found modest but statistically significant improvements in five of six evaluated domains.

The implementation story is less about whether the intervention works than about what scaling requires. Expanding from Kenya to Ghana surfaced the need for clearly defined institutional roles, coordination with mobile network operators, and explicit decisions about who funds message volume and support over time. The case study describes work with government and telecom partners, and an interest in bringing government in earlier so those arrangements can be built rather than retrofitted.