Data for a difference

It begins with a cough that just won’t go away. For more than 27,000 people worldwide, today will be the day they’re diagnosed with tuberculosis (TB). For much of the world population, and especially for those living in poverty in Asia and Africa, a TB diagnosis holds the stark possibility of death; more than 1.5 million people will die from TB in the next 12 months.

For others, it begins with a fever that keeps coming back. More than a million people are infected each day with malaria, which will kill nearly 500,000 people this year—the vast majority of them living in Africa.

Malaria and TB are both infectious diseases. The former is caused by a parasite commonly spread via mosquito bites, while the latter is caused by a bacterium and spread person-to-person. And they both can be cured with low-cost drug treatments.

Why, then, do they each claim so many lives?

For Rada Savic, PhD, the answers are right in front of us, hidden in the reams of data coming from clinics on the front lines of the battle against these diseases. “We know that we shouldn’t have any deaths from TB or malaria,” says Savic, a faculty member in the Department of Bioengineering and Therapeutic Sciences, a joint department of the UCSF Schools of Pharmacy and Medicine. And she is determined to make that possibility a reality.

Health and big data

High on the fifth floor of Byers Hall, at the UCSF Mission Bay campus, two of Savic’s students—Emma Hughes, a PhD candidate in the Pharmaceutical Sciences and Pharmacogenomics program, and Kendra Radtke, PharmD ’18—are sifting through numbers on pediatric cases of TB and malaria, beamed from health clinics located halfway around the world. Buried within gigabytes of data are clues that could help save many thousands of children from these diseases each year.

Malaria, a scourge of humanity that has stalked us for millennia, is particularly deadly for the young, the old, and the poor. TB similarly preys on the impoverished.

Savic works with numbers, but she’s also an expert in pharmacology. She uses statistics and computer modeling to follow drugs and outcomes during disease treatment. Her team uses computer code created in the Savic Lab to sift through reams of clinical data, hoping to identify patients who may be at an increased risk of dying, despite diligently taking their medicines as prescribed.

Left to right: Emma Hughes, Kendra Radtke, PharmD ’18, Rada Savic, PhD

Savic thinks that changing the guidelines for these drug regimens could be the difference between life and death for tens of thousands of children. To develop a new drug regimen that will work for every patient, she needs ample data on who may or may not be benefitting from existing regimens.

One big problem with this approach, however, is that resources can be scarce and record keeping can be scattered. But the Savic Lab has a workaround. Instead of perfect data, Savic relies on lots of it.

Strength in numbers

Normally, when a drug therapy isn’t working, pharmacologists use blood samples to understand how the drug is moving into, through, and out of a patient’s body. This movement of drugs throughout the body is called pharmacokinetics, or PK for short, and such PK data can be used to dial in the correct drug dosages for patients.

Even in the best of circumstances, the data that Savic’s team receives can have gaps, inconsistencies, or missing numbers. Critical PK data, the gold standard for troubleshooting doses, must be collected by specially trained clinicians. And virtually no single study, on its own, comprises enough children, varying by age, gender, and nutrition status, to adequately track why treatments aren’t working.

But by harnessing data from dozens of studies, the Savic Lab’s algorithms can discern patterns in the noise of clinical records.

If PK data is only available for some of the children in a study, for instance, Savic’s team can infer the missing PK data for the other children—and then infer that the standard doses given to malnourished children aren’t cutting it. The results turn clinical data from a hodgepodge of sources into a giant clinical trial.

Key to all of this work is Savic’s own rapport with countless health care organizations, providers, and fellow scientists. Many of these providers are the first responders to cases of malaria and TB, and they often get funding from organizations like the World Health Organization (WHO). Because of her network, Savic is often the first collaborator to receive newly collected data.

“For tuberculosis, we have access to data from almost every single trial in the world,” Savic says. And her network of malaria specialists and trials is growing, too.

The relationship goes two ways: in return for collecting and forwarding data on their patients, these clinics not only are enabling the Savic Lab to develop better treatments for malaria and TB, but also may be able to beta test new treatments in clinical trials that Savic herself oversees.

“We can reach out to clinicians on the ground via Skype or email,” says Hughes, whose specialty is malaria. “Oftentimes, if I have questions about data that we’ve just received, they try to clear up any confusions that we have.”

Adapting a cure for some to a cure for all

Early in their research, Savic Lab scientists noticed a common thread in data coming from clinics on the front lines of malaria and TB: malnutrition.

Radtke, who is investigating the TB drug regimen, says that some trials of standard treatments for TB were expected to demonstrate cure rates of more than 95 percent. Instead, “there was still 15 to 20 percent treatment failure or death,” she says. “And then we began to notice that many of the younger patients were malnourished.”

It may seem like a no-brainer that malnourished patients, children especially, might struggle to recover from disease even during treatment—after all, malnourishment can weaken the immune system, and the body more generally. Yet no clinical trials, to date, have conclusively linked malnourishment with poor outcomes for malaria and TB.

“Because little statistical evidence of this relationship has been shown [with malnourishment],” says Radtke, “we can’t expect anything to be done about it.”

If malnourished children do have a heightened risk of succumbing to malaria and TB during treatment, Radtke, Hughes, and Savic think that the solution might be surprisingly simple: give them doses that are appropriate for their age, rather than their weight, to ensure that every child receives enough drug to handily fight off their illness.

Currently, doctors choose drug doses based on patients’ weights: a 10 kg (22 lb.) two-year-old and a malnourished 10 kg four-year-old both receive the same dosage, based on existing guidelines that only account for weight.

© 2007 Bonnie Gillespie, Courtesy of Photoshare

A nurse in rural Tanzania explains to a child’s mother how to administer artemisinin-based combination therapy (ACT) to treat malaria.

“Malnourished children also typically have malabsorption of drugs into the bloodstream, so they’re getting an even lower drug exposure compared to a normal weight child,” Radtke says. “We really need to up the doses in those children in order to allow them to have the same chance of disease eradication.”

To confirm this hypothesis, Radtke and Hughes are continuing to collect, clean up, and analyze data from multiple clinics. Ultimately, the team wants to build models that can predict the doses that would be required to consistently cure every case of malaria or TB, based on an understanding of the pharmacokinetics of the most-used drugs.

Hughes is currently building a PK model of commonly used antimalarial drugs, based on the results of nine clinical trials that have already been carried out. She’s starting with just 370 children, but as more data is added to the sample size, her model will become more reliable.

And, on the TB front, one of Radtke’s preliminary analyses shows that, under her suggested regimen, malnourished children would beat tuberculosis.

When all is said and done, though, the team must bring their findings onto the world stage and ensure that their data actually makes a difference.

Making the case for a cure

The Savic Lab’s last challenge is to convince medical agencies that current dosing regimens for malaria and TB treatment must be altered to account for complications like malnourishment. But to make this argument, data from Radtke’s and Hughes’ models won’t be enough: they’ll need to demonstrate that these new dosages actually produce better outcomes in patients.

Curing TB and malaria could save millions of lives in the coming years—and could help some of the poorest parts of the world raise their standards of living.

Thanks to her decades of experience collaborating with clinics worldwide, Savic has the ears of investigators who are actively overseeing ongoing clinical trials—sometimes the same trials that helped provide data for her lab’s research.

“Once you communicate these findings in a useful way, a lot of these leadership entities really open up,” she says. “Recently, a collaborator at Rutgers, who is managing a big TB trial, told me, ‘I hope you feel good about yourself, because what you showed me just now, it's convinced me to totally redesign how we’re moving forward with our clinical trial.’”

Malaria, a scourge of humanity that has stalked us for millennia, is particularly deadly for the young, the old, and the poor. TB similarly preys on the impoverished. But the two diseases also contribute to poverty, putting enormous burdens on the families of the sick. Curing the two diseases could do more than just save millions of lives in the coming years—it could also help some of the poorest parts of the world raise their standards of living.

As exciting as the prospect of true cures for these two diseases is, it’s just as exciting to consider the wide range of disease treatments that could be improved using Savic’s methods.

“Any disease is a good target for this approach,” says Savic. “I’m choosing these two diseases because, to be frank, they are diseases for which you can really make a difference, today. These diseases are curable.”