Virtual care has its gaps - it’s time we talked about it
We teamed up with Omada Health to quantify how much of patient care can be delivered virtually vs. in person
For this piece, we wanted to get a deeper industry perspective. So we teamed up with Omada Health CEO Sean Duffy. If you haven’t heard of Omada, it’s one of the OG digital health companies that started out in preventing diabetes and now works across diabetes, hypertension and more. So this analysis represents both of our views.
As an industry, we all talk a lot about virtual-first care. But let’s get real. Telemedicine and remote care have limits. There are times when it works extremely well. And there are times when the patient needs to be seen in person.
We all know that to be true on a deeper level. And yet, we wondered if there would be a way to actually quantify that. How much of our care services could be administered in a virtual setting and what type of treatment needs to stay brick and mortar? And how will that ratio change in the future - and why? As we move from hype to balance, we think that the real work lies ahead. Both ambulatory care managers and new digital entrants alike need to think through the right balance between virtual and in-person.
As a next step, we asked the data and clinical teams at Omada to analyze when, how, and why office-based encounters are required for safe and effective care. That involved parsing through the catalog of coded, categorized medical services and procedures via the American Medical Association’s CPT (Current Procedural Terminology) code library.
Omada’s researchers began with analyzing the full repository of 9,523 Category I CPT codes against a rubric of “clinical feasibility” for virtual care. How Omada’s team defined that:
If a particular service could be reasonably accomplished without an office visit;
And this could be done safely and effectively without undue patient burden.
This is not meant to be perfect, as within some codes, one can find cases where you could argue both ways, or it may depend on some sub-specifics of a procedure, yet is meant to be directional to offer a grounded starting point.
Via this analysis, we hypothesized that 1.6% of Category I CPT codes in existence could be administered virtually, with significant differences amongst primary coding categories. The vast majority of virtually-feasible codes are E&M codes, referring to an evaluation of a patient by a doctor. As you can see, there are many CPT coding categories where no care can be delivered virtually, including radiology, surgery and so on.
Yet, this clearly tells a partial story. This simply describes feasibility, not utilization. Thus, a more important dimension of analysis is evaluating the utilization (quantity) of care relative to common needs and opportunities. In other words, a significantly larger portion of care is for the roughly 1.6% of codes that can be administered virtually.
To provide a better estimate, we asked Komodo Health for help. Komodo has developed a Healthcare Map, which is an administrative claims database of about 330 million de-identified patient journeys in the United States. Omada’s team then analyzed over 165 million patients from payer-complete sources who received medical care in 2019 (pre-pandemic), to study utilization against the full library of CPT codes.
We learned that for the average individual, 28% of the services they receive in a given year are feasible via virtual first care. On an encounter level, 24% of encounters with health care practitioners included no clinical services that required in-office interaction.
This analysis gives us a starting point for discussion in a way that begins with metrics versus pure observations. We’ll note here that we don’t believe virtual and office-based care should be viewed as competing models. They need to be integrated in the pursuit of optimal care for a patient. Patients will value and optimize for convenience, yet biopsies need to be performed, moles need to be removed, and wounds need to be stitched up.
Perhaps we should orient toward an acceptance that roughly 30% of in-year patient care can be delivered virtually? If this is true, it speaks to an important input in optimal service design. By definition, a virtual-only approach does not work to support holistic care if 70% of care rendered needed in person interaction.
Yet the positive side is that 30% of care can be virtualized and delivered in a more convenient, accessible, and equitable manner. For this segment of care, innovations are possible that even the very-best in-person provider could not do.
This potential is what digital health companies are constantly thinking about. Many will self-select into areas where more virtual care makes sense, including in musculoskeletal conditions, diabetes management and behavioral health. And others are consolidating to offer a more complete solution so patients can choose between virtual or brick & mortar offerings. One example from this past week is OMERS Ventures’ portfolio company PeerWell getting acquired by Bardavon to provide hybrid Physical Therapy.
Thinking into the future, here are a few recommendations we came up with for health care providers for optimal virtual care design:
Personalize care experiences in a way you could never do so in an in-person setting. Your local clinic doesn’t repaint the walls based on your preferences when you arrive. Done right, virtual care can truly feel like it was shaped for the patient.
Create efficient, consistent, and simple coordination between different types of care professionals. The best in-person providers try to put several specialties in the same building. With no buildings, this could become even more powerful (and as a side note here, we’re seeing some exciting startups pop up in this space to help coordinate care between different telemedicine providers).
Enable social and emotional support from peers who’ve lived in the patient’s shoes or have the same struggles. The current health “system” leaves patients feeling very alone (shout out here to Marigold Health, a company that’s leveraging peers and helping them be as effective as possible).
Iterate and adjust care plans with more speed, data, and consistency. A patient shouldn’t have to wait weeks for a visit to tune their treatment plan.
Deliver more value per medical credential. By combining people and technology, a medical professional’s credential can have a further reach and deliver more value at lower unit cost.
And a few other things that could change how care is administered in the future - plus an idea or two for future posts:
What might a similar analysis look like for at-home care? Clearly we should not be setting up portable operating rooms in people’s kitchens, but what medical services can be done safely and efficiently in the home setting that cannot be done virtually? This thought experiment sets up a vision for somewhat of a “convenience triage:” virtual → home-based → clinic-based.
Just because care can be delivered either at home or virtually, it doesn’t mean it should. Patient preference at the end of the day needs to be the most important voice. Today’s consumer expectations are shifting rapidly.
What are the looming and most promising technologies – diagnostics, devices, etc. – that might enable positive changes in what can be delivered virtually? Innovations in even simple things like specimen collection might enable more to be completed from afar.
But we recognize we’re still in the early innings and there’s still work to do.
"The reality is that brick and mortar primary care is overflowing with patients wanting and needing primary care, with demand exceeding the supply,” Aledade’s chief medical officer Dr. Emily Maxson told us. We talked to Aledade because of its network of primary care physician groups across the U.S.. The company told us it is seeing roughly 5 and 10 percent of visits taking place through telehealth.
We’re at an important juncture.
The last two years seemed reactive for than anything else. Covid-19 prompted a lot of providers to embrace telemedicine, but many of them didn’t prepare for it or expect it. Faced with few other options, providers did their best but had to grapple with imperfect processes and poorly optimized technologies. There wasn’t a moment to breathe to do this right.
Armed with data, it’s time for us as an industry to think more deeply about what tomorrow’s care can and should look like.
This piece is a joy to read - smart, beautifully written, tight … super absorbing. It’s not overly hopeful in all its possibility, but the notion of positive external forces helping (!) clinicians/teams to work, at least potentially, at the high ends of their credentials, was valuable to ponder.
Are patients and communities similarly able to work at the high end of their games, and if not, why not?
Some patients now “get” and use MyChart, report and even upload data, and do all sorts of things related to health virtually. Some, who even two or three years ago, didn’t yet do any of this, do it now. Many others can’t or don’t. What is most feasible to change?
What inning are we in in terms of reaching great, consistent, broadband access nationally? Is that a solvable problem, a moonshot, or even eminently addressabl? How much is it moving now? How much do we want change here?
Relatedly, is it kosher to say that “patient preference needs to be the most important voice,” when many can’t start or persist due to somewhat or altogether weak broadband access?
Since some broadband access needs fairly radical change in some venues, I’d love to know who are among the most impressive leaders or stakeholders in terms of propelling multi-stakeholders to establish broadband as a starting point for all?
This addresses doc. centricity. If the Q is optimal care shift to consumer centricity.
How do optimally eat,hydrate,move, sleep?Omada will never reverse lifestyle disease in a clinic. It happens in the 99.9% of the time we aren’t. Habit stacking using a care plan and infotaining feedback and wearables with a constant data exhaust are the answer to VPC value add for consumers,docs,costs,outcomes,employers to bend the unsustainable trajectory.