A Conversation with Henry DePhillips, MD: The Doctor in the Teladoc House


Interview by Peter Wehrwein


Being a chief medical officer is nothing new to Henry DePhillips, MD. He has held that title at several companies, including Audax Health, PDR Network, and Medem. But none of those jobs were as prominent as his current one as chief medical officer of Teladoc, the country’s largest telemedicine company by several measures. DePhillips, a board-certified family and community medicine physician, was interviewed by phone in March.

Telehealth or telemedicine—which one do you use and is there a meaningful difference?

First, there is not a standard definition for either term, and as a result of that, there is a variety of definitions in legislation, regulation, different federal agencies, and even in the private sector. We tend to view telemedicine as a form of medicine, rendering patient care using technology to connect to parties that are not in the same proximity. Telehealth includes health-related education, health monitoring of ongoing chronic conditions, and so forth.

So, Teladoc, you think of yourselves as a telemedicine company?

I think based on the definitions that I just shared with you, as of today, yes, I think you would characterize us as a telemedicine company. But if you follow our company, you’ll see we’re expanding our offerings exponentially at this point, and it’s only a matter of time before we implement telemedicine programs around things that are more broadly defined as telehealth.

There’s been a lot of investment in primary care and the patient-centered medical home in American health care. Doesn’t telemedicine undercut those efforts that are trying to make American health care more holistic?

I certainly understand how you might come to that observation, but at Teladoc specifically, we work incredibly hard to not only not impede the existing physician–patient relationship, but actually to support it. So, I think the answer is it really depends on the telemedicine company’s business model.

In our primary care program, which is our largest program to date, our marketing materials are very specific about saying we provide access to U.S. board-certified physicians in the scenario when you cannot get in to see your own primary care physician in a timely way.

On top of that, every patient must establish a full electronic health record prior to their first visit with us, and we ask every single patient who has a visit with us permission—required under HIPAA—to share that electronic health record with their primary care physician, if they have one.

We actually ask the questions: Do you have a medical home? Do you have an ongoing care relationship with a primary care physician? And if the answer is no, then several of our health plan clients would like to know that, so that subsequent to a visit with us, they can connect that patient with a primary care physician in the health plan’s network.

We also have a partnership with Compass, which is specifically designed to help identify a medical need, and then, identify an in-person provider who is best suited to provide that need.

Have you done an analysis of how many of your encounters result in a follow-up visit to a primary care physician in person?

We have a 92% complete resolution rate for the reason for the initial call over a 30-day follow-up period. In other words, 92% of Teladoc users do not need any subsequent care for the reason they called over a period of 30 days.

The 8% breaks down roughly half and half. About 4% are contacting us with issues that are totally inappropriate for remote care—drug seekers and people with medical issues that just are clearly not appropriate. And the other 4% are folks who, despite appropriate care need additional care. This is the human body we’re talking about here. Sometimes people will get worse.

What are the three or four most common problems people call Teladoc doctors with?

We now have multiple lines of business. Our episodic primary care program is our busiest program. The top three diagnoses there are sinusitis, bronchitis, and urinary tract infections, rounded out by influenza, which, by the way, is an example of a diagnosis that is ideally suited for telemedicine.

In our dermatology program, chronic recalcitrant acne is among the top. Suspicious lesions, which you would think would be popular, are our number three diagnosis. And then rashes are number two.

In our behavioral health program, it tends to be disorders that you would expect—depression, anxiety. Our number three diagnosis is PTSD. That tends to be very amenable to telemedicine care because some folks are not comfortable venturing out in the community, being seen with others, going to in-person care.

And then, we have a sexual health program that anonymously allows folks to get comprehensive testing and treatment for the entire range of sexually transmitted diseases. The anonymity and the remote nature make it more comfortable for many people.

Why is flu ideal?

I think influenza is a prototypical diagnosis for telemedicine. Influenza, when it is epidemic, you can make the diagnosis very, very accurately based on history. You know that if one third of the teachers in the school are out and a quarter of the students are out and all of them have influenza, and you develop prototypical symptoms, chances are really good you have influenza. And unless you have some kind of complications or complicated health history, the CDC actually recommends that you stay out of physician offices so as not to expose other patients, as well as the medical staff.

Also Tamiflu is a prescription medication—non-DEA controlled—which when prescribed very early in the course of influenza significantly reduces the duration and severity of symptoms.

Shifting focus, what did you think of the Rand study in the March issue of Health Affairs that found that 88% of Teladoc use by CalPERS beneficiaries was new utilization and that it added $45 of cost per acute respiratory patient? The gist—and I think this has been found with urgent care clinics—is that telemedicine is feeding new utilization and is going to cost more.

We know the author well. She is a very knowledgeable researcher. But that study, I think, illustrates the fact that study design will have a significant impact on study outcomes.

For example, that study used telemedicine visit data that was four- to five-years old—somewhere in there. The design also had a very short follow-up time period. So, there were a number of factors that caused that study to be very different from some of the much more extensive, current, and exhaustive studies that we have done.

When we look at much more recent data and have a 30-day follow-up period, our studies actually show results that are the exact opposite of the study in Health Affairs. That is, about 88% of utilization would have happened anyway in some other setting, whether someone would have gone to the ER, or whatever, and about 12% to 13% of utilization is what we call “new” utilization.aa

The financial ROI in many of the very large, multi-state, longitudinal studies that we’ve done demonstrates substantial cost savings in the neighborhood of $472 or so per visit saved over the weighted average cost of that same visit in another setting—a blend of other settings. Also, we have found, give or take, about a $20 per-member, per-month reduction in medical costs across the entire population served by telemedicine.

Have you published those results in a peer-reviewed journal?

We have not. It’s a very fair question. Having a study peer reviewed adds a significant amount of time to the time of publication, and because we are a publicly held company and we have specific business objectives, we prioritized the importance of getting those studies out to the market sooner, rather than having them peer reviewed. So, we did the next best thing, and that was have very knowledgeable industry, well-known Harvard Medical School researchers actually do that work. We didn’t do the work; they did the work. And they went right to our clients for the data. And we chose to publish it in that format, mostly because of speed to market.bb

I have questions about quality-of-care issues. I’m interested in the JAMA Internal Medicine study that showed that broad-spectrum antibiotics were more likely to be prescribed in a Teladoc encounter than an in-person physician encounter. Do you see that as a credible finding and, if so, have you taken steps to address the problem?

Obviously as chief medical officer of Teladoc, my number one, two, and three responsibilities are quality of patient care. Specific to your question, we do think that the results were credible, and we do think they were correct, and we did take that as constructive feedback to our program. So, not only have we looked at the basis for those results, but we have taken extensive actions to improve our performance in that area. That study shined a nice light on something we needed to go to work on, and we did.

A couple things. During the study period—which was a few years back—there were multiple states that had a legislative three-day prescription limitation for any telemedicine-related encounter. I think that was a regulatory or legislative effort to try to get patients to circle back to their own primary care physician after a telemedicine encounter.

Unfortunately, it had some unintended consequences. There’s a broad-spectrum antibiotic available where you can give a three-day course that results in a full course of treatment. So, there was about a 40% increase in those types of prescriptions in the states that had that legislation or regulation in place. Since then, all states except one have done away with that limitation, which has been helpful in dealing with the broad-spectrum issue.

Everything through the Teladoc program occurs on the Teladoc platform, so we have the ability to look at every single aspect of every single encounter. So, the quality oversight committees—of which we now have five—looked specifically at broad-spectrum prescribing rates and appropriate antibiotic utilization.

And as a result of very specific, very targeted, very focused quality interventions, very often physician-to-physician and peer comparison interventions, our broad-spectrum prescribing rate has come down substantially since that study was published. Our overall prescribing rate and our broad spectrum prescribing rate are now below what we believe the national average is in the U.S. today.cc

Is there anything that resembles the broad-spectrum antibiotic problem that you said you’ve dealt with—when you have found something that was, systematically, maybe a little off-kilter?

The other issue that has come up that we’re now looking at is the use of steroid prescriptions for respiratory illnesses. And it looks like in the community there’s a fair amount of that, and we’ve seen more than we’re comfortable with, issuing a prescription for steroids in the setting of acute respiratory infections. So, we now have a very analogous—very similar to the broad-spectrum issue—quality intervention going on where we have updated our clinical practice guidelines and they call out the appropriateness or lack of appropriateness for using steroids. And we have a very similar crackdown-type intervention on folks that tend to be a little more generous with steroids.

Once the docs know that you’re looking, and once you identify that it is not the standard care for certain diagnoses, and once you compare them to their peers, which seems to be very effective, the rates tend to come down very nicely. And we are seeing that in our program.

A different subject. Has Medicare Advantage been a growth area for you? My understanding is there are restrictions on telemedicine in traditional Medicare, but a Medicare Advantage plan has a fair amount of discretion on how it uses telehealth.

Most of Teladoc’s customers are commercial insurers and employers. Medicare? Not so much. Medicaid makes “perfect sense.”

Traditional Medicare has severe restrictions on the use of telemedicine. Medicare Advantage plans are at liberty to use telemedicine as they see fit. So, the direct answer to your question is, we do have some Medicare Advantage business, and that’s increasing over time. It’s not a huge part of our portfolio. And, don’t forget, there’s legislative capability, and then there are resources. So, although Medicare Advantage organizations are able to use telemedicine, there’s no incremental funding for the addition of a telemedicine program within Medicare Advantage. It all comes under the umbrella of the payment that’s made today. Because there are not additional resources applied, you’re not seeing adoption grow quite as quickly as it would if there were additional resources.

Which payers have been your biggest growth area? Commercial payers?

The commercial health insurance companies are a huge growth area for us. When we began Teladoc back in 2002, we targeted employers. And I think we have 7,000 or so employers with plans at this point. About nine years ago, we started selling to health plans. We now have over 30, although it might be higher than that. We have more in the pipeline. Aetna is obviously one of our flagship clients. We have multiple Blues plans, we have some multistate Medicaid plans. But commercial insurance is huge as far as adopting telemedicine. The biggest benefits are seen initially with self-insured employers, for some reason. And then, fully insured employers are on the heels of that.

Because of the Medicare limitations, we’re not seeing quite as much deployment across Medicare. We are seeing an increase in deployment across Medicaid. If you think about it, it makes perfect sense. If you combine the improved access, the reduction in non-emergency care, emergency room utilization, and the medical cost savings, all Medicaid plans could benefit from those three things. So we are seeing significant adoption in Medicaid programs.

Do you have good data in the Medicaid population that telehealth reduces ER use?

We don’t have data that is as robust as we have in the commercial populations. The studies we will send you on cost savings are all done in a large, 50-state commercial employers—Fortune 1,000 companies, basically. We don’t have near as much data on the Medicaid populations, which is why we’re excited about the Medicaid clients that we have enrolled and are getting up and running.

The good news is that the outcome studies don’t take too long to run. One of the landmark studies that demonstrates cost savings was done with an 18-month pre- and post-implementation time for a 50-state home-improvement retailer, whose color is orange. The studies don’t take too long, and we hope to have some very, very robust Medicaid outcomes data in the not-too-distant future.

Home Depot—is that a self-insurance situation?

They are a self-insured organization, as are some of the other companies that we’ve studied, which you will receive information on. It was really an ideal company to study because they have a presence in all 50 states, which cuts across all care settings, models, ER costs, things like that, across all 50 states. The results blew me away. I expected some medical cost savings. I did not expect anywhere near the magnitude that study demonstrated.

Maybe it’s because all of those employees are used to doing do-it-yourself projects.

You can put that quote in your article if you’d like to. I’m fine with that.

Anything in the repeal and replacement of the ACA that Teladoc and the industry is looking for?dd

Telemedicine is a very nonpartisan issue. Regarding the ACA and whatever the replacement is going to be, we think that the benefits of telemedicine are going to occur whatever happens. There were benefits under the ACA; there will be benefits under whatever new program is implemented.

There’s no particular piece of what the Republicans are talking about—Medicaid block grants or larger HSAs—that you see as being particularly advantageous, either to Teladoc, or to the industry in general?

I would say that’s correct, with one comment. To the extent that any legislation—this or future legislation—continues to place responsibility for the cost of care on to the providers of care, which is sort of what the ACA was about, any time that becomes more important, telemedicine has a larger role.

If you extend your practice with telemedicine, you’re allowing patients, when you’re not in the office, to be able to access you or your colleagues for care, which prevents leakage—going outside of your system to maybe an unaffiliated ER, where the costs are going to be astronomical. So, to the extent that new legislation partners being responsible for the cost of care—the outcomes—together with the provision of care, telemedicine will actually have an increasing role in that scenario.

What are your membership goals?

Our goal for this year—and you’ll see it in our filings—we plan to end the 2017 calendar year with between 21.5 and 23 million members. Growing from about 17.5 million, which is where we ended 2016. One measurement of our growth is membership, but we think a far more important measure of our growth is utilization. The number of visits that our members are using telemedicine for is increasing quite a bit faster than the increase in membership.ee

What about revenues and losses? I see that your revenues are up, but you also have net losses. When does Teladoc expect to become profitable?

That’s also in the filings. We have published to the markets that we will achieve EBITDA breakeven in the fourth quarter of 2017. And we are aggressively pursuing that goal.

The transcript of this interview was edited for length and clarity.

Most of Teladoc’s customers are commercial insurers and employers. Medicare? Not so much. Medicaid makes “perfect sense.”