While it may seem a contradiction in terms, digital automation tools may make possible a new level of personalization in medical care.
Over the past several years, there has been a growing movement toward customizing medical care for the individual. This is not just a "feel good" social movement; it is a scientific realization that individualized care can be more effective than care designed for the masses.
From N=many to N=1
Before I explain about the automation, let me share a little history. The research of disease processes has, in the past, focused on studying large populations of people with common symptoms in an effort to find common denominators that could explain the underlying disease process. For example, 19th-century physician John Snow studied a population of cholera epidemic victims and discovered that contaminated drinking water was the source of contagion. Similarly the "gold standard" in medical research has been the "randomized clinical trial." In this study, a large number of patients are split into two equal groups, each of which receives a different therapy. The group that has the best overall outcome or response defines which therapy then becomes accepted.
This N=many approach (where N represents the size of the population in the study) has worked well, but it has its limitations. Human physiology is complex, with an enormous number of variations which can affect the disease process and response to treatment in an individual. Not all members of a group of patients with a given disease are alike.
And, too, a patient's environment -- including social network, education, personality and access to money and other resources -- can affect his or her ability to follow through with recommended treatments.
As our knowledge of human physiology has grown, we've realized that we need an approach that accounts for these variations and designs treatments at the individual level. An N=1 approach, if you will.
After 23 years of searching, N=1 finds a cure
Eric Dishman of Intel is the perfect example of how the variation in human physiology can confound medical science. Dishman was diagnosed in college with what was thought to be a rare form of kidney cancer. Over the next 23 years he received a wide variety of treatments that had proved effective in treating kidney cancer (an N=many approach). While the treatments kept him alive, they were devastating to his system and didn't cure his cancer.
But Dishman got lucky. As part of his job as Intel's general manager for health and life sciences, he visited a number of enterprises involved in next generation genome sequencing. On one of those visits, on a whim, he had his entire genome sequenced and sent the results to his physicians. The genomic data revealed that the cancer in his kidneys was genetically much closer to a type of cancer commonly found in the pancreas. His treatment was changed to address the specific mutations of his cancer cells (an N=1 approach) and, 18 months later, he was free of cancer and healthy again.
Dishman, who has been an innovator in the field of patient-centered care and an advocate of the personalized approach to medicine, turned out to be the poster-child for why this matters. At this week's HIMSS14 conference, he is speaking on "N=1: Customizing Care for My Life and My DNA." I am privileged to join him during that talk to share an update on an "N=1" childhood neuroblastoma project that uses high-performance computing to speed up the process of whole genome sequencing. The genomic data pinpoints the best treatment for each young patient based on the specific mutations in that child's tumor.
As the title of Dishman's presentation suggests, our genomic data are clearly important, but make up only one piece of the N=1 approach. Creating a treatment plan that takes into account other physiologic data (proteins for example) as well as individual differences in lifestyle, environment and resources matters just as much as genomics.
Automation makes individualization possible
That's where automation can help clinicians provide personalized care. On a very simple level, clinicians can use customized text messages, sent automatically on a pre-set schedule, to remind patients to take medications on time, check their blood pressure or do other important health-related tasks.
Often, a patient's ability to comply with a treatment plan can be bolstered by having the right information at the right time. For example an automated text message could say, "7 a.m. Time to take your blood pressure medicine and your statin. That's one of the round blue pills and one of the pink oval pills. Take those pills with a full glass of water."
For older patients with multiple medications, having a reminder offered at the moment it's needed, complete with personalized instructions, could mean the difference between health and hospitalization. Without automation, caregivers couldn't offer that level of specificity and support to patients.
In the area of health coaching, which turns out to be pretty effective in helping prevent chronic disease complications, automated feedback messages can be very useful, offering multiple opportunities to guide patients to appropriate diet and exercise choices, timed to arrive at the moment when the advice is most needed. They can also offer positive reinforcement, which can be a powerful motivator.
Better care, lowered costs, will earn support for N=1
While these may seem like simple tools, there is growing data that they can be very effective at keeping patients out of the hospital, saving money and improving the quality of life. And there are thousands of developers working on many more digital tools that will be customized to individual needs. Medicare and private health plans are starting to take note of the effectiveness of digital tools and have begun reimbursing for some of them.
As the N=1 approach proves its effectiveness in reducing costs by reducing hospitalizations and expensive complications, payers will not only support it, they will insist on it. And automated digital tools will quickly become ubiquitous in healthcare.