Each year, 136 million infections are caused by bacteria resistant to life-saving antibiotics [REF] and as many as 50 million people are estimated to die due to them [REF].

I saw firsthand how these invisible threats could devastate lives—personally, and then, during the COVID-19 pandemic, and across every healthcare setting from the world’s most advanced hospitals to resource-limited clinics.

In 2019, I had just begun a PhD in Mathematics at Imperial College London, where I was building predictive models to track the spread of antimicrobial resistance in hospitals. I had already worked in biodefence and infection detection—and just before COVID hit, I was training in exactly the skills that would soon become urgently needed.

During the first wave of COVID-19, I volunteered in hospitals and worked with Imperial College Healthcare NHS Trust to provide real-time forecasts on bed and ventilator usage [REF - the page]. But inside the hospital, it was clear we were missing something vital. Infections weren’t just arriving from outside—they were spreading within.

In fact, up to half of COVID-19 hospitalisations were acquired inside hospitals themselves [REF - BBC]. Outside the pandemic, up to 1 in every 15 hospitalised patients has an infection, acquired as a result of their medical treatment [REF]. Yet people seemed to lack the tools to track or anticipate these transmissions.

That’s when the idea struck: what if we could use data to predict the next outbreak? What if infection control could become proactive—not just reactive?

So we started building. Our early prototypes helped hospitals visualise where infections were—and where they were likely to go next. We developed predictive tools grounded in epidemiological evidence and validated our approach with NHS Trusts. This work culminated in our landmark study published in The Lancet Digital Health [REF], showing that we can anticipate—and potentially prevent—hospital-onset infections.