Seema Verma, Administrator
Centers for Medicare & Medicaid Services
Department of Health and Human Services
200 Independence Avenue SW.,
Washington, DC 20201
Dear Administrator Verma:
On behalf of Conversa Health, which supports health providers with virtual check-in technology, I write in response to the proposed rule titled, “Medicare Program; Revisions to Payment Policies under the Physician Fee Schedule and Other Revisions to Part B for CY 2019; Medicare Shared Savings Program Requirements; Quality Payment Program; and Medicaid Promoting Interoperability Program” as published by the Centers for Medicare & Medicaid Services (CMS) in the July 27, 2018 Federal Register.
We applaud CMS for recognizing that advances in communication technology have changed patients’ and practitioners’ expectations regarding the quantity and quality of information that can be conveyed via communication technology, and that virtual check-ins using evolving technology may become more effective at addressing patient concerns and needs. Your proposal acknowledges that practitioners are beginning to leverage technology for virtual check-ins, obviating the need for unnecessary office visits. This type of proactive engagement with patients will reduce costs and improve care across the healthcare system. And as a result of reducing office visits through more efficient and timely virtual care, reimbursement for virtual check-ins will increase clinician availability for patients that will benefit from in-person visits and will expand health system capacity.
Your proposal also heralds an important change to our health care system. Virtual check-ins and remote patient monitoring (RPM) transform our health care system into one that is proactive rather than reactive, and one that is more continuous and collaborative rather than episodic. We should not underestimate the importance of that change. CMS’s proposal supports this positive change by rewarding providers for their expertise and time, while rewarding patients with greater convenience and lower cost.
CMS has asked for comment on the types of communication technology that are utilized by physicians or other qualified health care professionals in furnishing virtual check-in services. We are pleased to share with you our experience in deploying automated conversational technologies for this purpose over nearly five years.
Conversa Health’s conversational artificial intelligence (AI) technology allows healthcare organizations to deliver automated doctor-patient conversations that lead to more informed and meaningful patient relationships, effective population health management and, ultimately, better clinical outcomes for patients at lower cost. Through the use of automated technologies that are implemented by providers for virtual check-ins, patient-generated health data (PGHD) can be systematically and reliably gathered and analyzed in order to support the provider’s decision-making regarding the need for an in-person visit. At the same time, using automated technologies for virtual check-ins can deliver clinically-endorsed, real-time automated feedback to the patient. You can experience an example of our virtual check-in patient experience by texting “CHF” to the short code telephone number 83973.
Through our on-the-ground experience implementing and integrating automated virtual check-ins into the physician workflow, we’ve concluded the following:
—The majority of virtual check-ins (approximately 75 percent) are initiated by the provider, not the patient. The provider initiates them based upon their knowledge of the patient’s health condition. When patients do initiate virtual check-ins, about half of the time they are seeking additional health information on their health condition, while the other half of the time they are seeking support from the provider for a problem or concern that the automated check-in process did not address.
—The content and frequency of virtual check-ins are guided by provider-approved logic governed by institutional practice guidelines or common, publicly available clinical practice standards.
—Virtual check-ins are most effective when they collect structured and clinically-coded PGHD because such data can be aggregated, analyzed, and presented for efficient decision-making by the provider.
—Virtual check-ins are also most effective when they are designed to deliver clinically endorsed, real-time, automated feedback to the patient, thereby providing the guidance and support necessary for the patient to achieve a healthier state.
—The process of initiating virtual check-ins and utilizing the resulting structured PGHD can and should be integrated seamlessly into the clinical workflow, prioritizing clinician preference and minimizing administrative burden.
—Most often, physicians will designate other team members such as a care coordinator or care management nurse to review the PGHD from the check-ins, reserving physician time to those results which require additional expertise and judgment. These providers are not always billing practitioners under Medicare.
—The time between virtual check-ins initiated by the provider should be no greater than the time during which something is likely to arise that would become a bigger (and more costly) problem. This varies depending on the patient and her condition. 1.) For patients in a post-discharge situation, this might be every day to every few days. 2.) For patients preparing for surgery, the check-ins might occur 3 or 4 times, while post-surgically they might occur 5-10 times. 3.) For patients in a chronic or ongoing health management scenario, the check-ins might occur every week to every few months.
—For patients in a primary care scenario, we have found that in the course of a year that only about 25 percent of patients require an office visit to manage their conditions. These visits are generally driven by a need for better medication management, lifestyle modification, or more stringent monitoring.
—While only about 15 percent of responses to virtual check-ins are identified as concerning, about 90 percent of patients will have need for more support, i.e. an “escalation” at some point during their recovery from a hospitalization, or at some point during the year in managing a chronic condition.
—When an escalation is needed, a phone call is satisfactory for addressing the concern in about 90 percent of cases. Only about 10 percent of escalations warrant an in-person visit.
We’ve found that the most intensive of these virtual check-in frequencies is under a post-discharge scenario. The following chart shows how Medicare patients of a large health care provider fare on virtual check-in chats for the 30 days following discharge for any of the hospital Star Rating diagnoses. The chats may be completed by the patient or their designated caregiver. Note that patients are much more likely to be in the “red zone” and require an escalation in the days following their discharge, while patients are much more likely to get into the “green zone” by about a week to 10 days after discharge.
An escalation in this case is a telephone call with a nurse navigator or care manager, only rarely warranting an office visit or a return to the hospital.
The automated virtual check-ins are a convenient way to understand in real time how the patient is doing. From the patient perspective, it’s a timely, personalized, and convenient way to answer key questions about their health and receive real-time guidance.
The value to the provider cannot be overstated. When automated technology helps them to gather and analyze the patient’s PGHD on a systematic basis, recalibrating their frequency and content based upon the provider’s approved logic, providers are quickly alerted to the relevant information so they can act upon it. This drives medical discussion among the care team for the patients who need it most, and allows care team members to not waste time unnecessarily on patients who are doing fine. More importantly, automated virtual check-ins reduce practice variation as they help ensure that compliance with care plans and monitoring for complications is consistent across patient populations.
Comments on the Proposed Rule
1. Provider Role
The proposed rule states that the expectation is that a majority of check-ins would be patient-initiated. This may assume that inbound secure messages in an EHR’s patient portal are the most likely modality. However, many more innovative technologies exist in the market than EHR patient portals, and we encourage CMS to ensure that use of these more user-friendly technologies are also incentivized. Based upon our experience, we have found that automated, asynchronous technologies generated from the provider are the most efficient and effective form of check-in. Therefore, we recommend that CMS recognize the role of automated and asynchronous check-in technologies that are provider-initiatied and that gather and analyze PGHD to support providers’ decision-making under the GVCI1 Virtual Check-In code.
We recommend that CMS recognize that a virtual check-in may, or may not, result in an escalation to a health care provider from a nurse or care manager. Even when it does not result in an escalation, such as when automated technologies using pre-defined thresholds find that the patients’ responses are within normal limits, these technologies are still aiding the provider in determining if a visit is warranted.
We also recommend that CMS make it clear that clinical staff such as a care coordinator may review virtual check-in data and manage necessary escalations, incident to a physician or other Medicare billing practitioner’s activity.
2. Frequency and Reimbursement
The value of a virtual check-in can be realized whether or not it is in close proximity to a visit, but the seven days following a procedure or hospitalization is an especially important time for identifying and addressing complications and preventing readmissions. Therefore, while we applaud CMS’ decision to not apply any frequency limitation, we recommend that the limitation of not being able to bill for the virtual check-in in the 7 days after a visit be removed. For this same reason, we also recommend that virtual check-ins not be included in the global period for surgical episodes of care.
To account for the value of these automated check-ins, but ensure that each reimbursement is aligned with an amount of medical discussion time, we recommend that the 5-10 minutes of medical discussion be changed to “medical review and/or discussion”, and that the time may be allowed to accumulate per patient. By allowing the minutes to accumulate over time, you provide greater flexibility to the provider in how they realize the value of these modern virtual check-in modalities.
We also recommend that medical review and/or discussion should include time reviewing PGHD, discussion among care team members, or discussion with the patient or their caregiver.
3. Documentation of Medical Necessity
The documentation of medical necessity for a check-in is very important. In our experience, this documentation includes the PGHD that gets documented with each automated interaction, along with the logic that drove the outreach to the patient. It is important that providers review and approve that logic, i.e., that it be clinically-endorsed.
We recommend that the documentation should include the clinical logic that justified the frequency of automated digital check-ins as well as the PGHD that drove clinical decision-making.
4. Patient Consent and Financial Obligation
Because patients must consent to receiving electronic outreach from providers, virtual check-ins that are initiated by providers must by definition include the patient’s consent. We therefore recommend that CMS clarify that electronic consent is sufficient. We also request that CMS follow the example set by the Medicare Annual Wellness Visits (codes G0438 & G0439) and eliminate any patient co-pay obligations for such a service. Patients should not have to pay for increased efficiency and convenience.
As an estimate of the potential financial impact, we have calculated the impact of virtual check-ins in a primary care practice caring for 25,000 patients, equally distributed among 10 physicians. Of these patients, we estimated that 7,500 are Medicare beneficiaries, and that automated virtual check-ins could help providers assess the need for an office visit an average of 4 times each month for each of these patients. Some check-ins would be more frequent, and some less frequent. We also estimate that approximately 25% of patients would have six or more check-ins in a month, each requiring approximately 1 minute of “medical review and/or discussion,” which would cumulatively justify a virtual check-in reimbursement. Based upon our experience, we estimate that only 25% of patients will need to be seen in the office for the management of their existing conditions and that each patient would be seen once per year on average for new problems. Finally, we estimate that the average E&M billing code for in-person care would increase due to the acuity of patients being seen in the office. Putting all of this together, we estimate an overall savings to Medicare, but also a significantly reduced wait time for an office visit due to the number of office visits averted.
Note that the above financial impact does NOT include the savings from the hospitalizations and ER visits that would be avoided as a result of mitigating those clinical problems earlier than normal as a result of the check-in technology.
The following scenarios examine how virtual check-ins can be used to help manage patients under a couple of different circumstances:
Scenario 1: A primary care practice managing a chronic care patient
Within a given month, a Medicare patient with high blood pressure and type 2 diabetes has received four automated virtual check-ins from her provider. As part of the chats, her interim blood pressure and blood glucose data are imported with their trends analyzed and summarized. The patient was found to be doing reasonably well with the management of her blood pressure and diabetes, and the service automatically provided her guidance on diet and exercise that, along with care plan compliance, should help her to continue to live a healthy life. The date of her Annual Wellness Visit with her HbA1c check and in-office blood pressure check was confirmed to be in a couple months. The cumulative patient check-in session time for these check-ins was about twenty minutes.
She initiated an additional check-in herself for a new problem that arose.
The provider’s nurse care manager reviewed each of the four check-ins and determined that no visit was warranted at this time. The new problem check-in resulted in a secure email to the patient to say that the doctor recommended an over-the-counter medication and to let the office know if it got worse. No visit was recommended at the time.
Virtual Check-ins: 5
Patient virtual check-in session time: 27 minutes
Chronic conditions managed: 2
Acute problems resolved: 1
Types of PGHD gathered: Biometric, symptom, health history, care plan compliance, lifestyle, quality of life data
Savings: $80 (1 chronic care follow-up visit and 1 acute visit avoided at $75 for each)
Scenario 2: A hospital managing a post-discharge patient
In the month immediately after a discharge for congestive heart failure, a patient received 30 provider-initiated virtual check-ins to ensure that her weight, symptoms, medication management, and oxygen levels were doing well. From those 30 check-in notifications, the patient completed 25 check-ins for a total patient check-in session time of 100 minutes. The patient self-reported their weight. With each check-in, the patient received additional patient education on salt intake along with a variety of other topics.
During three of the check-ins in the first ten days after discharge, the patient was identified as being in the “red zone” for weight and water retention. The nurse care manager called the patient and adjusted her medications, but based upon a lack of other concerning symptoms she felt that the patient could continue to be monitored at home.
The nurse care manager took about 30 seconds per check-in to review the patient’s data from each session for a total of 12.5 minutes, and about five minutes for each phone call.
Virtual check-ins: 25
Phone calls: Three at 5 minutes each for total of 15 minutes
Patient virtual check-in time: 100 minutes
Types of PGHD involved: Biometric, symptom, health history, care plan compliance, lifestyle, quality of life data
Reimbursement: $84 (if a limit of 6 per calendar month is imposed)
Savings: $6 – A saved office visit ($90) minus the GVCI1 reimbursements
Scenario 3: A patient receiving opioids for lower back pain
A patient receives care for acute onset of back pain and is placed on a short course of an opiod and a long term plan that includes exercise, physical therapy, and enrollment in an virtual check-up program.
Over the next month the patient receives 6 scheduled virtual checkup sessions. These sessions determine the patient’s goals of treatment and assess how the patient is progressing towards these goals. The biological, psychological, and social aspects of the pain are assessed through patient reports on pain, activity and work impairment, and depression through validated measures. The sessions also include education on safety taking prescription opioids, assess common opioid side effects,, and non opioid based interventions for back pain with a focus on self care. Lastly, the sessions include reminders for the patient’s PT appointments and confirming she has transportation to the PT clinic. The patient also self initiates 4 check ins for additional education about pain management. The cumulative session time for this patient over the month was about 40 minutes.
Over the month a nurse manager reviews the patient reported outcomes for each of the 6 sessions and notes steady improvement in the patient’s report of pain and no report side effects from the discontinuation of opioids. The nurse notes the patient reports difficulty in attending her PT sessions due to transportation issues and contact the patient to arrange alternative transportation. No office visit is deemed necessary at that time.
Virtual Check-ins: 10
Patient virtual check-in session time: 40 minutes
Chronic plan compliance issues resolved: 1
Patient reported positive progression
Types of PGHD involved: Patient reported outcome measures for pain, social and psychological issues, medication side effects, medication compliance, care plan compliance, quality of life data
Reimbursement: $84 for GVCI1 for accumulated time in the month > 5 minutes of medical review and/or discussion.
Savings: $16 for higher acuity visit avoided
Remote Patient Monitoring
Comments on Proposed Rule
- Physiologic Data
We applaud CMS’ expansion of the Remote Patient Monitoring codes. Based on our experience, to be effective Remote Patient Monitoring must include a wide variety of PGHD. Specifically, we collect data regarding symptoms, treatment adherence, quality of life and PROs in addition to biometric readings. This information is critical for clinicians that are monitoring patients to collect in order to understand how their patients are doing on a new treatment regimen, whether their condition is getting better or worse. In order to be sure that PGHD is not just quantitative data but also the critically important qualitative and behavioral data that clinicians need to monitor patients, we recommend that CMS clearly articulate that “Physiologic Data” may include patient-reported data about their health as well as biometric data.
We hope that our experience in delivering automated virtual check-ins for providers helps to show the potential benefits of this technology. We have found that such technologies can truly transform health care to being more proactive, continuous, and collaborative. This technology also represents a tremendous opportunity to improve the convenience of health care for patients and providers alike.
We look forward to continue working with our provider partners and CMS to ensure that we balance the needs of providers and patients as we embark on this exciting era of virtual care technologies.
Philip Marshall MD, MPH
Co-Founder and Chief Product Officer