We’ve heard it before: Artificial intelligence (AI), with its giant data sets and rapidly evolving technology, is transforming health care into a science capable of preventing illness and prolonging life.
We believe this will be true over the long term, but the pace of this change needs to be incremental in order to get things right.
In the meantime, humans will continue to age as we always have. As we grow older, we’re more susceptible to serious illnesses such as cardiovascular disease, cancer, lung disease, and Alzheimer’s, which are leaving their mark on countless lives. At Amgen, we’re pioneering a 21st-century approach to developing and bringing life-saving therapies to patients living with these diseases.
Improving diagnosis and treatment
When managing serious illness, a correct diagnosis must be present for successful treatment. We’re applying machine learning to help predict the risk of specific serious illnesses before they manifest, partnering with medical research and data science startups to develop predictive algorithms.
Let’s play out an example of an absent or incorrect diagnosis.
Familial hypercholesterolemia (FH) is a genetic condition that causes high levels of low-density lipoprotein cholesterol beginning at birth, potentially leading to events associated with early aging.
Though the prevalence of FH is thought to be at least one in 250, 90% of people with the condition have not been properly diagnosed and are therefore undertreated.
When people with FH are not treated, they are at 20-times greater risk of having an early heart attack. Eighty-five percent of men with FH will have a heart attack by age 60, some as early as in their 40s. Women with FH have a similarly increased risk, with heart attacks striking as early as in their 50s.
Recognizing this high unmet diagnostic need for patients and families, Amgen developed FH Score, an online clinical diagnostic support tool for health-care providers that leverages the Dutch Lipid Clinic Network criteria, a clinical diagnostic standard for FH. Currently used in 11 countries, FH Score helps identify patients at risk of FH before a cardiovascular event such as a heart attack takes place.
Supporting persistence to therapy
Another key variable for treatment success is simply staying the course. For patients with chronic illness, the Centers for Disease Control and Prevention (CDC) identified medication adherence as critical to improving outcomes. According to a recent meta-analysis, non-adherence to chronic disease medications accounts for approximately $290 billion of health spending in the United States.
One reason for nonadherence? Forgetfulness.
So, we’re developing approaches to serve patients where they are when taking medications—at home—with a simple, user-friendly reminder system.
Patients prescribed self-injectable biologics—medicines developed from living cells—must store their medications in the refrigerator. With this in mind, we are creating a refrigerator-mounted device that provides smart reminders based on proximity sensors, designed specifically for these biologics. Preliminary results show that most patients find the device easy to use, noting that it fits into their daily lives—promising signs for incorporating it into their routines.
On the horizon
With everything we do at Amgen, we’re working to deliver on our mission to serve patients. We’re inspired by the progress we’ve already made in realizing a 21st-century approach to treating serious illness and energized by the work still to be done in bringing AI applications to modern health care.
Learn more about Amgen’s approach to predicting and preventing illness at www.amgen.com.
A horrifying new AI app swaps women into porn videos with a click
Deepfake researchers have long feared the day this would arrive.
How AI is reinventing what computers are
Three key ways artificial intelligence is changing what it means to compute.
How AI digital twins help weather the world’s supply chain nightmare
Just-in-time shipping is dead. Long live supply chains stress-tested with AI digital twins.
DeepMind’s AI predicts almost exactly when and where it’s going to rain
The firm worked with UK weather forecasters to create a model that was better at making short term predictions than existing systems.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.