One scientist’s personal mission to tackle comorbidity
“Even as a child, I liked solving difficult problems,” says Marzieh Nabi. “I like the challenge of studying a puzzle, and the satisfying feeling of solving it. My father was a math teacher. That’s his influence.”
Marzieh’s interest in complex math problems initially drove her to study aeronautical systems design. But soon after her studies began, a shock event propelled her career in a very different direction. Her father suffered a seizure, caused by a brain tumor.
The revelation that her father had cancer — at the age of just 40 — hit the young Marzieh hard. As she was growing up, he had been her inspiration; he encouraged her to study hard and apply to one of Iran’s top universities.
“Rural Iranian society tends to discourage professional ambition in women,” she says, “but I was raised differently. My father always said: ‘it doesn’t matter if you’re a girl or a boy; you have to grow to your full potential.”
A stellar academic career
And Marzieh did just that, ranking 128th out of 500,000 applicants to secure a place on a degree course in aeronautics and astronautics at Tehran’s prestigious Sharif University. There, she applied her analytical mind to designing complex systems; deconstructing enormous challenges into many smaller problems and developing techniques for tackling them in concert.
But her dad’s illness weighed on her. Marzieh wondered if she should change paths: abandon the engineering degree and apply instead to medical school.
“I thought that if I was doing medicine I’d be more able to help him,” she says. “But then I decided it would be better to stick to what I was doing and try to acquire some relevant knowledge along the way.”
Marzieh’s journey eventually led her to leave Iran for a PhD in aeronautics in Seattle. “Moving abroad was a really tough decision. It meant leaving my mother and siblings to do the looking after of my father. As the oldest child I felt it should be my responsibility to care for him. But my father was all for me getting my PhD. He encouraged me to go.”
An emotional departure from Iran
Leaving Iran was made worse by the fact that, after an initially positive prognosis, Marzieh’s dad had begun to experience a series of setbacks. Part of it was the side-effects of the cancer treatment; “this poisoning medication,” as Marzieh calls it. But other complications were caused by factors initially unrelated to the cancer.
One of those factors was a head injury Goodarz (Marzieh’s father) sustained in a car accident. It was relatively minor, but it damaged a vulnerable patch of scalp where he had had radiotherapy. That necessitated more surgery, placing greater stress on her dad’s heart. Soon he was also being treated for heart disease, in addition to cancer. Finally, he was also diagnosed with chronic lymphocytic leukemia.
Comorbidity: The healthcare challenge of the 21st century
There is a medical term for this: Comorbidity. And learning about it was to change Marzieh’s life. “As I talked to my father’s doctors, I discovered that our understanding of comorbidity, and of how to treat patients with multiple chronic conditions, is extremely poor,” she says.
It’s also not an isolated phenomenon. A quarter of all adults in the United States suffer from multiple chronic conditions; that number expands to two-thirds for people over the age of 65. An estimated 70 percent plus of the nation’s $2 trillion in annual health spending is focused on patients with comorbidity.
And as the population ages, the problem is becoming more acute. “The White House calls it the healthcare challenge of the 21st century,” Marzieh says.
Applying systems thinking to multiple chronic conditions
Though at the time Marzieh was focused on aeronautical science, she saw co-morbidity — and what was unfolding with her father — in terms of another complex problem to solve.
“I had been looking at airplanes,” she says, “but essentially my training is all about systems. About observing them and figuring out how they work, about modelling them mathematically, about designing new systems and optimizing existing ones.
“The human body, the way we deliver and fund healthcare — these are also complex systems where the same kind of thinking can be applied.”
A chance discovery and a career turning point
When a doctor friend told Marzieh about the existence of a database of electronic medical records of individuals with co-morbidity, she realized she’d stumbled onto something profound. Marzieh’s friend suggested she analyze the dataset to look for insights into how multiple conditions occur, and how they affect each other.
Gaining access to this type of information was the turning point in Marzieh’s scientific career. Finally, she could use her analytical training to help not just her father, but everyone with multiple chronic conditions.
Bringing the problem to Xerox
At first, she worked on her new “project” in her spare time — combing the data for patterns and correlations between patients’ diagnoses, their age, the operations they’d had, their gender, their lifestyle, and many other variables. But once she’d finished her PhD, she found an opportunity to work on the problem full-time, as a research scientist at PARC.
“When I applied to work at Xerox it wasn’t specifically for this,” she says. “I just wanted to work somewhere I could apply my scientific knowledge to projects that would have a tangible impact, rather than sit in an ivory tower.” It was there, she said, where she “successfully made a case for turning my side-project into a Xerox research project, and it became my day job.”
A breakthrough moment: the discovery of a huge new dataset
Marzieh’s work took on greater urgency when she learned that Xerox subsidiary Midas+, which develops healthcare management software, had access to an anonymized dataset of nine million EMRs.
“Combined with other information sources, this much larger dataset has the potential to transform our understanding of comorbidity, how it occurs, and how it should — and should not — be treated,” Marzieh says.
“In five or ten years’ time, we should have a much better understanding of how patients with multiple chronic conditions are likely to progress, and we’ll be able to optimize their medication accordingly. This is information my father didn’t have. The doctors didn’t know how to fine-tune the dosage to his specific conditions, to minimize the side-effects, or anything like that.”
Medicine, surgeon Atul Gawande has written, is an “imperfect science.” Marzieh believes that increased computing power and ever more powerful data analysis techniques will combine to drive those imperfections out of the system.
“In my father’s case, we personally observed how trial and error was the basis of many procedures and prescribed medications,” she says. “I think in not that far future we will be able to combine the experiences of all patients and doctors across the globe to make sense of all the data and assist us to make optimal treatment decisions.” In contrast, she says, today “we struggle with much simpler question of how to predict the next stage of a disease — especially in cases of comorbidities with many different variables involved.”
Another problem is the slowness of medical discovery. Marzieh cites an example: Metabolic syndrome, which increases the risk of heart disease, stroke and diabetes and found in a third of Americans. It took health care more than 90 years to recognize the syndrome is caused by a cluster of conditions — increased blood pressure, high blood sugar, excess body fat around the waist and abnormal cholesterol levels. With that knowledge, treatments have improved.
“Ninety years is a long time,” Marzieh said. “The process of medical discovery has historically been very slow. It typically starts with a small set of observations and many pre-clinical and clinical trials on different patient population cohorts. Heterogeneous environments, uncertainties in original hypotheses, the passage of time and accumulating costs make it a very complex process.
“But the promise of big healthcare data is set to significantly pick up the pace, kicking off a new age of intelligent medicine where information from different medical resources will become integrated.’’
Even nine million records don’t hold all the answers
Marzieh’s project could help millions of people manage and survive the onset of multiple chronic conditions. But she warns there is a long way to go, and that even a dataset of nine million records has severe limitations. The database only covers the United States, for example, so Marzieh must take care in making assumptions.
“The geographical limitation is just the start. We always have to be aware of what information is missing. These are just codifications of individual doctors’ notes,” she says.
There are many sources of bias and noise, Marzieh says, ranging from different coding practices at different hospitals to different habits in patients with and without insurance and even different physician quality levels.
“But we can still learn a lot about what’s happening for patients, and how different conditions and treatments interact in their body. And from that we can take steps towards understanding how best to tailor treatments to patients based on their individual circumstances.”
One promising application of artificial intelligence, machine learning and systems science in medicine is the use of so-called “inferential statistics” that leads to improved guidelines for improved clinical decision making.
“Guidelines play a big role in current medicine,” she said.
A new challenge: convincing doctors to trust algorithms
One of the challenges in this emerging discipline is convincing medical professionals of the potential of data — to trust algorithms as much as their own eyes.
“Doctors have their own way of thinking about the world, and mathematicians and system scientists have another,” Marzieh admits. “Creating a bridge between the two communities is definitely a challenge to be addressed.”
A data-driven future for healthcare
In the longer term, Marzieh hopes that data-driven analysis will become an integral part of medical training and drug development. “We are just scratching the surface of big data in healthcare. The datasets will only get bigger — and that will give us progressively better knowledge of the causal relationship between diseases. In 15 to 20 years’ time, it won’t take us nearly as long to gain new knowledge in medicine, or to develop new treatments.”
Sadly, for Marzieh’s father, those breakthroughs will come too late. Goodarz died in 2014 of complications from brain cancer. But his endless encouragement and support still spur his daughter forward today. “It’s very tough to do this work sometimes, because it reminds me constantly of my father,” she says. “But ultimately, that’s what motivates me to keep going.”
We’ve all changed the world. Every one of us. With every breath we take, our presence endlessly ripples outwards.
But few of us have the opportunity to change many lives for the better. And even fewer are challenged to do so every day. That’s the gauntlet thrown daily at Xerox research scientists — to try and effect change.
In return, we give them time and space to dream. And then the resources to turn dreams into reality — whether they’re inventing new materials with incredible functions, or using augmented reality to bolster the memory of Alzheimer’s patients.
We’re proud of our Agents of Change in Xerox research centers across the world.
Here are some of their stories.