October 1, 2016 (Vol. 36, No. 17)

Vicki Glaser Writer GEN

Smartphone-Connected Patients Are Asking, “Can You Treat Me Now?”

Mobile. Smart. Disruptive. These words dominate discussions of technology-driven trends in e-commerce, social networking, and the sharing economy. And soon these words will apply to healthcare, which represents the next barricade to be stormed by smartphone-wielding revolutionaries. Already, the popular press is celebrating the “doctor in your pocket.”

In the vanguard of the new medicine are “think different” types such as Eric Topol, M.D. As if his duties as the director of the Scripps Translational Science Institute weren’t enough to keep him fully occupied, Dr. Topol has found the time to write a popular book about the medical impact of smart, handheld, networked devices.

The book is titled “The Patient Will See You Now: The Future of Medicine Is in Your Hands” (Basic Books, 2015). It describes the democratization and individualization of medicine. As Dr. Topol explains in this interview, these developments will ultimately be enabled by the digitalization of essentially all forms of omics and biomedical imaging and analysis..


Parents are now able to get a child’s ear infection diagnosed using a smartphone and algorithms. [Martin Barraud/Getty Images]

GEN: If your technological vision is realized, will medicine cease to be as much an art as a science?

Dr. Topol: Traditional medicine sees people from 30,000 feet and misses granular detail. It has been much more art than science. We lacked the ability to digitize a human being. Now we can develop a kind of Google medical map with multiple layers that include the phenome, anatome, physiome, and all of the different biological “omes.&rdquo

We now have a mosaic composite of the individual. We can capture data about a person through wearable biosensors, obtain a DNA sequence, profile the gut microbiome—and many of these examples represent real-time, streaming data.

Medicine is becoming a much more quantitative science. It is changing to accommodate a new level of data, big data, for each individual.

When you refer to the patient-doctor relationship, I would emphasize that the individual will soon be a big driver of this relationship going forward, because a lot of the data will be individual-generated through sensors or by doing self-lab tests or self-imaging.

The changing dynamics of the patient-doctor relationship will mean more individualized care and greater patient safety. To give you an idea of how bad things are now, there are 12 million serious medical diagnostic errors each year, and these errors have been estimated to cost physician practices more than $15 billion per year. I think that a big part of the problem is the imprecision that exists due to the lack of understanding of each person at the level that is needed.


GEN: Please explain what the phrase “the democratization of healthcare” implies.

Dr. Topol: Democratization of healthcare means that everything is available to all in the medical sphere. Many things that are happening today reinforce that democratization is taking hold. One of the first things is that people are generating data through their smartphones.

People are performing their own electrocardiograms or taking images of skin lesions or rashes, and they can receive diagnoses immediately via algorithms. Parents can get a child’s ear infection diagnosed using a smartphone and algorithms. Sleep apnea can be assessed in the comfort of a patient’s own bedroom, and the results would more accurate than those from a hospital sleep lab, while saving about $3,500.

For the first time, over the past year, one of the major central laboratories, LabCorp, has enabled people to order their own lab tests without needing a doctor’s order. That’s a big step forward, especially in light of the falling costs of genetic testing.

Soon, patients may no longer be totally dependent on doctors, hospitals, and health systems. Consumer surveys indicate that at least 80% of people want to be able to take much greater charge and direction of their medical care. That said, I do not believe that democratization of healthcare will be truly achieved until all individuals have ownership rights to their data, protected by new laws and legislation, with technology that protects the privacy and security of that data.


GEN: In your book, you present a vast array of smartphone attachments, apps, and uses available or in development, including various kinds of biosensors. You describe “some of the remarkably diverse lab-on-a-chip assays that have been or soon will be integrated with a smartphone.” You even suggest a future of smartphone-based genotyping capabilities. Is all of this possible with today’s smartphone technology?

Dr. Topol: The digital transformation of medicine is already happening. You can use Uber to bring a doctor to your house. On-demand medicine is a big part of the democratization of medicine.

But the big shift going forward will not be the capability to have a video chat with the doctor, or summon the doctor to your house, but rather the ability to generate data and share it with the doctor. The patient-doctor exchange will go beyond chatting. It will involve the review of data, getting oversight of the data, and getting guidance and an injection of wisdom, experience, and compassion.

Smartphone labs are part of the story. For example, it is possible to make a diagnosis of infection through a smartphone. That has been demonstrated in as remote a place as Rwanda, where 99% accuracy was achieved for detecting HIV and syphilis, at a cost of 50 cents, with results available in minutes.

Additionally, we now have the ability to consider portable sequencing in the field, to provide an agnostic approach to identifying pathogens. This is still some years away, but it’s becoming technically increasingly likely.

Many companies are working on processing the breath using a smartphone, for example, to be able to measure aldehydes and other organic compounds that would increase the suspicion for cancer—not just lung cancer, but any type of cancer. We will also be able to monitor the environment in general, for radiation exposure or for air quality and pollutants.


GEN: The increasing reliance on sensors and remote monitoring may make hospital stays obsolete—except for surgical procedures and emergency or intensive care. Eventually, it may even be commonplace to use blood-based nanosensors to enable continuous monitoring of the physiome. What advances are needed to realize these scenarios?

Dr. Topol: We need to overcome the “edifice complex,” our habitual reliance on big hospitals, clinics, and professional buildings. We should question the value of regular patient rooms in a hospital, which in the United States average more than $4,300 a night. By the same token, we should consider how many visits to doctors’ offices could be avoided if telemedicine were more common.

The only thing in the technology area that is missing for this transformation to happen is being able to take the data from remote monitoring and apply it to patient care. We need remote data-monitoring centers. The only one I know of is at Mercy Hospital in St. Louis. With sufficiently exquisite remote monitoring, we may, eventually, be able to show that the “hospital room of the future” should be the patient’s bedroom.

Another exciting technology is the liquid biopsy. Nearly 40 companies are now trying to harness the information from circulating DNA in the blood to make the earliest possible diagnosis of cancer, or to monitor people after treatment for cancer.

The use of circulating free DNA or RNA can be extended beyond cancer to study methylation and other epigenomic changes and to monitor various organs and systems. This relates to the idea of using nanosensors in the bloodstream, coupled with genomic signaling, with the results being reported to a mobile phone app.

We are also working on cardiac applications. We hope to pick up genomic signals that predict heart attacks. But the regulatory path to validate this type of test is long. You have to get the blood test approved first, then the sensor approved, before you can then get the combination of the sensor and the blood test approved. It takes multiple years and large clinical trials.


GEN You say that the extensive sharing of data from little devices presents bit opportunities for Massive Open Online Medicine (MOOM). Why is MOOM important?

Dr. Topol: MOOM is about achieving open medicine through massive medical data sharing. Today, doctors do not share patient data. For example, many oncologists in the United States do not routinely share patient data in a common knowledge resource.

But what if data were shared? Information could be collected about patients’ DNA sequences, the DNA sequences for the patients’ tumors, the treatment regimens, and the outcomes. All this information could be added to an expanding knowledge resource, such that when the next patient comes along, the doctor may perform a nearest neighbor analysis to help guide treatment decisions. The patients would benefit from “smart” therapies. We have yet to build a learning health system.

If we can have 1.6 billion people on Facebook, why can’t we get more than a billion people in a medical knowledge resource, where the data exist in a secure, encrypted peer-to-peer network? Again, 80–85% of people would want to put their data in this type of resource if it existed.

Such resources could be created, but first we need a whole new way of thinking. We need to transfer ownership of data from doctors to individual patients. Doctors are not into data sharing; in fact, they are more into data blocking.

We also need to overcome privacy and security are hurdles. Today, we have an epidemic of medical data hacking, and the way we currently handle the data supports this. In just the past year, more than 100 million people have had their medical data hacked, out of 300+ million people. We need to centralize the data and to rely on the personal cloud or electronic locker types of concepts, rather than having it stored in different hospital systems.

Genomic data is the most sensitive data, and it becomes increasingly sensitive as it becomes more informative. What happens when we will have a billion people sequenced by 2025, which is the projected number? If that data is not being shared, then it will not be nearly as informative and powerful as it could be.


GEN Why is genomic data not being used to the extent it could be? What main factors, in your view, could help accelerate and advance the application of pharmacogenomics?

Dr. Topol: Pharmacogenomics has been one of the biggest missed opportunities in medicine for many reasons. First, it is hardly used by doctors, in part because it is not available as a practical, inexpensive, rapid test. If I am going to start somebody on a medication for which there is unequivocal data that could tell me if it is the right medicine at the right dose and if there are any potentially serious side-effects for this patient, I want to get that information in minutes and at a very low cost.

That capability just does not exist today. We do not have point-of-care pharmacogenomics. We may gain it, however, after a billion people have had their genomes sequenced.

At last count, there were 125 FDA-approved drugs with a genomic label, but the sad truth is that it is a rare day when any patient in this country has a pharmacogenomic test. It uniformly requires sending out for the test, the results will not come back for a week or more, and it would cost a minimum of $150–200 for a single genotype, which is usually not reimbursed.

Drug companies should be doing sequencing for every drug being developed so they would know well into Phase II clinical testing if there is some type of sequence that correlates with an adverse event. Instead of stopping development of the drug, they could avoid administration to individuals with a certain genomic burden. But pharmacogenomics is seldom, if ever, used in drug development.

Many people have criticized genome-wide association studies (GWAS), which have often come up with low odds ratios, low penetrance, “non-important” findings. However, that could not be less true for pharmacogenomics and GWAS. Andrew Harper and I published an article in Nature Review Genetics [2015; 16: 689–701] that reviewed the field and showed that the genetic variants identified can serve as unequivocal signals.

Sometimes the odds ratios, instead of being 1.2 or 1.3 for common polygenic diseases, can be 50, 80, or 150. But we are just not using this precious information. Pharmacogenomic studies should be done for every commonly prescribed drug.

























This site uses Akismet to reduce spam. Learn how your comment data is processed.