How digital tools are improving access to healthcare (13 min read)
Through interdisciplinary partnerships, digital technologies such as artificial intelligence affect every aspect of how people access medical services and treatment – and they’ve already begun helping patients worldwide.
Jun 07, 2018
Listen to audio version (12:44)
Heat kills. In one of the deadliest weather events in Europe in the past century, a heat wave in August 2003 killed nearly 15,000 residents of Paris – in less than two weeks. Thousands of patients with heat-related conditions crowded emergency rooms and strained hospital admission capacities. This also happened during a summer vacation month, when fewer staff members were available. The physicians, nurses and other healthcare workers on duty were pushed to their limits. Despite emergency mobilization of healthcare resources, 42 % of the city’s heat victims died in hospitals. Better hot-weather preparedness by hospitals might have saved lives.
To better anticipate emergency peaks, as well as to more effectively manage daily hospital operations, the question arises: Is it possible to forecast the number of patients coming to a clinic on a given day – or even in a specific hour – the way meteorologists forecast the weather? This is what four clinics in the Assistance Publique-Hôpitaux de Paris, Europe’s largest university hospital, are testing right now. The clinics wanted to improve access to medical care for patients by reducing waiting times and making optimal use of their existing staff. So technology company Intel helped them develop an analysis system that predicts emergency visits and hospital admissions for the coming 15 days each month. The project is based on a massive, anonymized data base of 470,000 patients that has been collected for more than a decade. Then data scientists combined this information with external factors like flu rates, seasonality and temperatures in Paris. Should all of the tests be successful, a computer will help keep emergency room staff optimally ready and patients better cared for.
Potential : Increasing every aspect of patient access
Interdisciplinary partnerships with tech giants such as IBM, Apple and Google make big sets of health data accessible in real time. What might sound like sci-fi scenarios hold huge potential for all aspects of medicine. Credit: Shutterstock/Elnur
When talking about the digital future of healthcare, some patients think of sci-fi scenarios such as robotic nurses. But as with computer models that predict emergency admissions, the biggest changes are already taking place. “We are on the verge of a digital revolution across every aspect of the healthcare sector, from the lab bench to the patient’s bedside,” says Vas Narasimhan, CEO of Novartis, of which Sandoz is a division. He sees opportunities to provide patients with new, improved and more holistic solutions that lead to better outcomes and also help reduce the burden of illness. “Digital technologies and big data are changing every aspect of how companies operate, across myriad industries,” adds Narasimhan.
A big data revolution is under way in healthcare.
Physicians, researchers and a plethora of business leaders echo this sentiment. “A big data revolution is under way in health care” is the key sentence in a report by global consulting firm McKinsey. Bringing big data to healthcare is a combined effort that includes interdisciplinary partnerships with tech giants such as IBM, Apple and Google. Such companies work with healthcare organizations so that patients benefit directly from computing power. Put simply, when big data – huge amounts of data – is processed using artificial intelligence and software that “learns” how to apply the data, medical care takes a quantum leap forward. Reducing waiting time in the hospital is just the beginning of how patients experience these systems. Perhaps paradoxically, the more data systems collect from people, the better they can tailor individual diagnosis, drug development and choice of treatment. And the applications reach into patients‘ everyday lives when they use digital devices to monitor their own conditions: They can actively contribute to their care or help prevent diseases in the first place.
Diagnosis: Intelligent systems save lives
Digital technologies and big data offer holistic solutions for patients that lead the way to personalized medicine. Credit: Novartis
When a patient displays a combination of symptoms, physicians rely on their training, experience and usually also on medical equipment to diagnose the condition. For diagnosing heart diseases, for example, cardiologists have relied primarily on echocardiograms. As the diagnosis is based on a limited range of factors, they are only correct in 80 % of cases, according to researchers at John Radcliffe Hospital in Oxford, UK. This means a diagnosis may miss imminent heart attacks, or it may lead to an unnecessary operation.
The Oxford researchers aim to reduce some of the uncertainty of diagnosing heart disease. Their system, Ultromics, whose solution is based on artificial intelligence, analyzes 80,000 data points from every image and increases the diagnosis accuracy to 90 %. These machine learning models are fed with information from one of the largest heart imaging databases in the world, which is stored at Oxford University. The system, which is planned to start in 2018, could save the UK’s National Health System GBP 300 million a year by making the diagnosis of heart diseases more reliable.
When making a diagnosis using digital data, artificial intelligence systems might have access to millions of patient cases. This can save an individual patient’s life, as in the case of a 60-year-old woman in Japan. At the University of Tokyo in Japan, IBM’s computer system Watson helped physicians to compare the gene sequence of this woman with 20 million clinical oncology studies from a database. As a result, the doctors diagnosed a rare form of leukemia in the woman, and they were able to provide her with a successful therapy. Before this, the woman hadn’t responded to treatments based on conventional diagnosis. Reportedly, Watson needed just 10 minutes to analyze the woman’s form of leukemia. Success stories like this raise hope in researchers and patients all over the world that digital healthcare will help to find a cause – and cure – for their conditions.
Treatment: Cognitive computing improves care
For some patients, for example, following a cancer diagnosis, a long, hard treatment road lies ahead of them. Here, the right combination of therapy approaches can go a long way. This is why Novartis started an initiative to optimize cancer care. The company has joined with IBM Watson Health to explore development of a cognitive solution that uses real-world data and machine learning (see interview, page 26). The common aim is to provide better insights on the expected outcomes of breast cancer treatment options.
Through this collaboration with IBM Watson Health, Novartis will use real-world breast cancer data and cognitive computing to identify solutions that may help physicians better understand which therapy may be best for which patients, and what information is useful for establishing clinical practice guidelines. The goal is to improve patient outcomes and experiences. This collaboration might also uncover care efficiencies that can be applied beyond breast cancer.
R&D: Increasing access to new and better medicines
In research and development, scientists use virtual reality to find new and better treatment options for patients. Credit: Novartis
In many therapies, to relieve symptoms, manage conditions or fight the underlying causes of various diseases, patients rely on medications. And it is in pharmaceutical development where the potential of big data and artificial intelligence really unfolds. Digital technologies, says Novartis CEO Narasimhan, has tremendous undiscovered potential to transform the research and development of medicines. Because databases store extensive data on chemical substances, these can be analyzed to suggest novel therapies, such as for rare diseases. His vision: “Digital solutions can democratize the research process for new medicines by helping us reach previously underserved and understudied groups of people. Ultimately, we can bring new and better medicines to patients who need them.”
The key to developing the next-generation oncology medicine for patients with cancer, for example, may be found by gaining better understanding of the molecular interaction of peptides and proteins. In one project, researchers at Novartis are using different methods of machine learning to predict the docking of compounds with protein targets. And thanks to virtual reality, scientists can even explore the interplay between compound and target in vivid detail. In another project, a Novartis partnership with the University of Vienna, cognitive computing plays a role in trying to help patients with diabetes- or age-related diseases of the retina. This has led to the discovery of two novel biomarkers. When these are present, physicians will be able to predict retinal conditions one year before a noticeable onset of disease. This ability will allow healthcare professionals to intervene earlier with therapy or medication to reduce acute-care episodes.
Stringent clinical trials accompany the development of pharmaceuticals. Here, too, digital solutions help to democratize the research process. In 2015, hardware giant Apple demonstrated to the public that smartphones have become a serious measuring instrument for medical studies. The company revealed a new software platform called ResearchKit that helps researchers enroll an unprecedented number of participants in their studies. The platform collects medical data on conditions including rheumatoid arthritis and concussions. ResearchKit also allows scientists to program apps that take surveys, give patients tasks and use smartphone sensors that track users' wellbeing.
Prevention: On the way to personalized healthcare
Digital devices monitor health information in real time – and they might soon help prevent diseases. Credit: GettyImages/AFP/David McNew
Smartphones also host a range of apps that allow patients to collect and monitor their own health data. Yet collection is nothing without collaboration. “All the progress we see in multiple therapeutic areas is only possible with the data that is accessible,” says Dr. Spencer Jones, Head of Medical Affairs at Sandoz. “If we stop sharing health data, we stop innovation.” Jones is an ambassador for digitalization in healthcare because of the potential that technology and computing models offer. “AI solutions rely on huge data sets – imagine if we could monitor and collect all data relevant for our health,” Jones adds.
And this volume is enormous, indeed. Today, each person will generate enough health data in their lifetime to fill 300 million books. More data has been created in the past two years than in the entirety of human history – and by 2020, IBM expects medical data to double every 73 days. Now, for the first time, humans have the technical possibilities to process all of this data.
What are the possibilities for the future of medicine when researchers have access to more health data? This vision is currently being followed by a large-scale and first-of-its-kind study named Baseline. With the goal “to map human health,” Google and the health tech division of its parent company Alphabet, Verily, have teamed up with Duke and Stanford Universities to monitor around 10,000 volunteers (currently in the US) of different ages, backgrounds and medical histories for a four-year study that started last year. All participants wear a fitness tracker that transmits their heart rates, movements and other information to a central database. A sensor below people’s mattresses monitors their sleep patterns. In addition, Verily also collects genomic data, as well as information on participants’ feelings, health records, family histories and the results of periodic lab tests on urine, saliva and blood. And monitoring environmental clues, such as home temperatures during heat waves, could save lives, too.
Whether predicting hospital admissions peaks in Paris, making genome-based diagnoses in Tokyo or discovering biomarkers in Vienna, digital solutions have the potential to change every aspect of medical service. Jones goes one step further: “If we can predict diseases and severe conditions, we can prevent them before they happen.” In the history of medicine, he explains, the order always used to be disease, medicine and then patient. “Digital tools turn this order on its head.” Dr. Jones and other experts believe that the digital healthcare revolution will affect how people access – or even define – healthcare. From the lab bench to the bedside, the patient will come first and remain solidly at the center of medical care.
Artificial intelligence (AI)
In 1966, scientist Marvin Minsky defined AI as “the science of making machines do things that would require intelligence if done by men.” AI is applied colloquially when a machine mimics “cognitive” functions that humans associate with natural intelligence, such as learning and problem solving.
Cognitive computing is technology platforms that encompass machine learning, reasoning, natural language processing, speech and object recognition and human-computer interaction via dialogue. The goal of these platforms is to simulate human thought processes in a computerized model.
Big data is data sets that are so voluminous and complex that traditional data processing application software is inadequate to deal with them. This is where Machine Learning is applied. Data sets grow rapidly – in part because they are increasingly gathered by inexpensive and numerous information- sensing Internet of Things devices.
Personalized medicine (or precision medicine)
Personalized medicine is a medical procedure that tailors medical decisions, practices, interventions and treatment to individual patients based on their predicted response or risk of disease.