A Deep Dive into the Details of Making Care Primary: What the CMS Request for Applications Tells Us about CMMI's Newest Primary Care Model
Post Written by Juliette Price, Chief Solutions Officer
With the just-released RFA, we can see how CMMI wants to reshape primary care–more integration with specialty care, renewed focus on clinical workflows and interventions, and some investments in practice delivery transformation. What’s not included? Downside risk and significant opportunities for shared savings produced outside of the primary care environment.
On August 14, CMS’s Center for Medicare and Medicaid Innovation (CMMI) released the Request for Applications (RFA) for their newest value based payment model, named “Making Care Primary.” The RFA contains all the program details that CMS will make public before the close of their application period. Applications are due to CMS by November 30, 2023.
If you’re just now tuning into the Making Care Primary model, take a first dive into this explanatory post that describes the overall model objectives and structure. In this post, we’ll dive into the granularities that were made available in the release of the Request for Applications on August 14, 2023.
Timeline
As usual, CMMI is giving potential participants a short window to decide on participation; the application period will open September 4, 2023 and close November 30, 2023. That may seem far off, but that’s only 12 weeks from the RFA release date.
Providers will be selected in Winter 2024 and onboard will occur between April 2024 and July 2024, just in time for PY1 start date of July 2024.
As always, it’s important to remember that completing an application for the program is not a legally-binding contract; that legal commitment will come towards the end of the onboarding process just before the beginning of the performance year.
Performance Years & Overall Duration
As promised in the model’s announcement, MCP will have a duration of 10.5 years in length, which brings us to December 2034.
Performance Year (PY) 1 will begin on July 1, 2024, and this will be the half year that was referenced earlier in the model explanations. Model requirements, payments, and participation requirements will all begin in PY1, but quality and cost performance won’t count until PY2 (CY 2025.)
Eligible Entities and Regions
To be eligible for the program, participants must be Medicare-enrolled organizations that provide primary care services–this can include solo primary care practices, group practices, health systems, eligible Indian Health Programs, and Federally Qualified Health Centers that operate at least 51% of their business footprint (physical locations) in the following states: Colorado, Massachusetts, Minnesota, New Mexico, New Jersey, New York, North Carolina, and Washington.
Interestingly, for New Yorkers, the model will not be available state-wide–in fact, in what we might call a “clarification,” CMS listed the specific counties in which the model is available; that list omits the following seven counties: Bronx, Kings, New York, Queens, Richmond [these 5 counties make up New York City], and Nassau, Suffolk [these 2 counties make up Long Island], and Westchester. No direct explanation is provided for why these counties are carved out, but any tea-leaves-reader might stare at that and ponder what other models might CMMI have up their sleeve for the Greater New York City area, for which carving those geographies out of the model would support another policy goal. No other state has any sub-geographies carved out for participation.
For organizations that are eligible, the model participation is defined at the organization level, which simply means it’s a TIN-based program, not an NPI-based program. This means you cannot have only some providers who work at a specific clinic in the program, all primary care providers under that TIN will be participants.
While specialty care providers play a role in the clinical design of this program, no specialty providers will be considered “participants,” unlike how in ACO REACH, an ACO can have both participating providers (primary care providers) and preferred providers (specialists) contracted in their ACOs.
Attribution
This model utilizes a prospective attribution model, meaning the list of Medicare patients for which the participant will be held accountable for will be provided ahead of the performance period. This list will be updated each quarter and provided to the participant.
The model will prioritize voluntary alignment, meaning the physician that the beneficiary selects via the Medicare.gov website. Voluntary alignment will supersede other alignment methodologies. Secondarily, the model will prioritize claims-based attribution, meaning CMS will look across claims data to see what primary care provider provided the plurality of primary care visits and/or eligible Chronic Care Management (CCM) services, or if one of the providers billed the patient’s most recent Annual Wellness Visit or Welcome to Medicare Visit.
Additionally, patients who have not received a primary care service from a MCP participant within the last 24 months will not be eligible for attribution. Patients with End Stage Renal Disease (ESRD) or enrolled in hospice will also not be eligible for attribution.
Three Track Options
Core to the organizing framework of the model, participants will select into one of three tracks, and time spent in each track will be limited–once they expire against their time in the track, they will move onto the next track. The track framework is highly similar to other CMMI models, such as MSSP, but the time-boundaries are baked into the model from the starting line, in part due to the lessons learned CMMI has accumulated from other models, i.e. participants in other models have “stalled out” in certain tracks and failed to move towards taking on more risk or accountability for outcomes.
Tracks 1 and 2 are both meant for participants who have no or limited experience in value based care arrangements, and each of those tracks will have a 2-year maximum time limit. Due to the off-set start date of the model (model implementation is July 2024), participants will have an extra 6 months in whichever track they select into in CY2024. Participants who begin in Track 3 will remain in that track during the entire period of the program.
Each track is essentially set up as a cohort; participants will not have the option to accelerate into track levels before the end of that time-limited period–for example, if a participant begins in Track 1, it must spend the first 2.5 years in Track 1 before moving on to Track 2. There will also be no option for participants to move backwards in tracks–once you’re in Track 2 or 3, there’s no moving backwards to Track 1 or 2.
Track 1 is the only track where applicants will have to prove that they have had no experience in value based care, as defined quite strictly by CMS as:
“The applicant has participated in performance-based Medicare initiatives (including Primary Care First (PCF), Comprehensive Primary Care Plus (CPC+), Next Generation AGO (NGACO), Direct Contracting, Accountable Care Organizations Realizing Equity Access and Community Health (ACO REACH), AND/OR has been part of a Medicare Shared Savings Program (SSP) ACO that has not deferred its entry into a second agreement period under a two-sided model under § 425.200(e) in the five most recent performance years prior to the start of the agreement. This includes scenarios where 60% or more of the TIN’s NPIs or CCNs meet the aforementioned criteria.”
Care Delivery Model
Now we get to the good stuff–how care for patients will fundamentally change at primary care sites!
First, we revisit the care model goals: “building participant capacity to deliver equitable, team-based care and improve outcomes over time on key metrics like hypertension and diabetes control, depression, emergency department (ED) visits, and total cost of care.”
The program has three domains along which it is hoping for care delivery transformation: Care Management, Care Integration, and Community Connection. Across each of these domains, the model has specific clinical care delivery requirements that providers in the program must implement–the full list is available in Appendix C (pg. 75), but the CMS summary illustrates the broad strokes of what they’re looking for:
“Overall, in Track 1 participants will build the foundation to implement advanced primary care services through activities such as risk-stratifying their population, developing workflows for care management, chronic disease management, and behavioral health and HRSN screenings. •
“Overall, in Track 2 participants continue to meet the requirements of Track 1 while also expanding and integrating the services available to their patients (e.g. once patients are risk stratified and care management workflows established, implementing chronic care management for high-risk patients).
“Overall, in Track 3 participants will continue to meet and build upon the requirements of Tracks 1 and 2, to further optimize and expand care delivery and specialty care integration (e.g. once patients are risk stratified, chronic care management for high risk patients is established, taking this further and ensuring there are individualized care plans for all high risk patients aligned to their chronic health needs as well as linkages to community based supports).
Care Management Domain
This domain is probably the most straightforward and unexciting of the bunch; it’s care management as you’ve heard it before. CMMI wants participating providers to empanel and risk stratify its patients, identify workflows for chronic and episodic care management, implement chronic care self-management programs, and offer individualized care plans, group education and connections to community-based resources. Nothing exciting to see here.
Care Integration Domain
This domain focuses on the connection between the primary care provider and specialty care providers, found outside the walls of the primary care practice. CMS is very clearly signaling in the RFA and its supporting documents that it wants to see primary care providers more in control of the spending that is spent in specialty care. To do so, CMS is looking for a number of care delivery integrations:
Integrating with behavioral health: starting with evidence-based screening and evaluation of all patients for behavioral health conditions, with a specific focus on depression and substance use disorder screening
Understand the value of specialists: CMS is committing to providing data to PCPs to understand the performance of specialists (Against cost? Against quality? Not too many details provided here) in the provider’s region, in an effort to identify “high-quality” specialist providers
Creating collaborative care arrangements (CCAs): Once in Track 2, the participants will be asked to create one (1) documented collaboration between the PCPs and a specialist from the following specialty areas: cardiology, orthopedics, pulmonology. These three areas were picked for focus due to CMS finding “large Medicare spending” and the hypothesis that improved access to these specialists and services may reduce total cost of care.
Access to a new e-Consult Code: participating PCPs will have access to a new code (MEC), designed to address issues that PCPs have with utilizing current e-consult and interprofessional consult codes. The MEC code will be priced at $40, before geographical adjustment is applied. (Available in Track 2 & 3 only)
Access for specialists to a new ambulatory co-management code (ACM for short): This new code is aimed at supporting the co-management of patients who requires both longitudinal primary and speciality care where there is a short period of time where both PCP and specialist need to collaborate more closely to stabilize a chronic condition. This is different from today’s often-used Complex Chronic Care Management (CCCM) codes as the ACM code use will relax the definition around requiring the condition to place the patient at “significant risk of death, acute exacerbation, or functional decline” and also requires the specialist to identify themselves as co-managing the patient during that period of time. The ACM code will be priced at $50, before geographical adjustment is applied. (Available in Track 3 only)
Community Connection Domain
Highly similar to other CMS efforts around health-related social needs, participants in the model will be required to screen all patients for social needs and connect them to community-based supports. The use of staff (such as community health workers) with lived experience is being centered in this model and close partnership with social care providers is also a key element. No payment for social care services is included in this model.
Each of the care delivery requirements has to be implemented at the site-level–i.e. For larger providers, you must ensure the care model is implemented across all clinical locations, not just one site. Participants will be given the first 12 months in any of the given tracks to implement the care delivery requirements associated with that track.
Measuring Success in the Model: Performance Measure Set
As all other CMMI models, program participants will be expected to perform against a standard measure set, which CMS is attempting to align across other programs via the CMS Universal Measure Set.
As always, the devil is in the details on measuring quality: for all claims-based measures will be assessed against the participant’s MCP-attributed patients (FFS Medicare members attributed to the providers in the model), but for non-claims-based measures, those will be measured against a participant’s total patient population, not just Medicare. CMS claims to be doing this to support “total practice transformation,” which you can choose to believe or not.
The full measurement matrix outlining which measures are applicable to which track is available on page 27 of the RFA, but include:
Clinical/Social:
Controlling High Blood Pressure (reporting modality: eCQM)
Diabetes: Hemoglobin A1c Poor Control (>9%+) (reporting modality: eCQM)
Colorectal Cancer Screening (reporting modality: eCQM)
Screening for Depression and Follow Up Plan (reporting modality: eCQM)
Depression Remission at 12 months (reporting modality: eCQM)
Person-centered Primary Care Measure: (reporting modality: TBD)
Screening for Social Drivers of Health (reporting modality: TBD)
Utilization/Cost:
Total Per Capita Cost: (reporting modality: claims) (an average of per capita costs across all attributed beneficiaries and includes all Medicare FFS Parts A and B standardized allowable charges incurred by each attributed beneficiary in the quarter. This calculation does not include any of the model payments, including Enhanced Services Payment, Performance Incentive Payment, and Upfront Infrastructure Payments.)
Emergency Department Utilization: (reporting modality: claims)
Total Per Capita Cost Continuous Improvement [non-FQHCs & non–Indian Health Programs only]: (reporting modality: claims)
Emergency Department Utilization Continuous Improvement [FQHCs & Indian Health Programs only]: (reporting modality: claims)
For the measures related to utilization/cost, CMS will use regional benchmarks to measure performance against. For the clinical measures, CMS will use national benchmarks.
There will also be tiered levels of performance achievements for participants to earn partial credit for meeting these measures, set for both the clinical and cost measures; in Tracks 1 & 2, the lower threshold will be set at the 50th percentile with the upper threshold set at the 70th percentile. In Track 3, the upper threshold will be set at the 80th percentile.
Payment Model
Upfront Infrastructure Payment
This CMS upfront investment will be available to Track 1 participants only. The use of these funds is meant to fall within three specific categories: increased staffing, social determinants of health strategies, or health care infrastructure, such as electronic health record systems, IT systems, etc. To be eligible for this investment, the participant will have to provide that their either (1) do not have an e-consult technology solution or (b) meet the “low revenue” threshold of 35%. The UIP will be made in two lump sum payments of $72,000 each, at the beginning of PY1 and one year later.
Prospective Primary Care Payment (PPCP)
The bulk of the payment design of the model lies here, in the PPCP, which is essentially a capitation payment for primary care services delivered by the participant. Over time, the goal is to move away from Fee For Service (FFS) based payments into full capitation for primary care services. In Track 1, 100% of billing will remain in FFS; in Track 2, providers will experience a 50/50 split between capitation and FFS billing; and in Track 3, participants will move to 100% prospective payments.
The prospective payment will be based on two years of historical claims data for the Primary Care Services for the attributed Medicare FFS beneficiaries; meaning the payment will be tailored to the participant and its patients. Over time however, CMS plans to move to a methodology that takes into account regional primary care spending, to standardize primary care payments across participants. CMS commits to spending the first three years of the model understanding variation in primary care spending, and using that insight to reform the PPCP methodology for PY3. This revised methodology will only apply to non-FQHC participants, and CMS clearly warns participants that they should anticipate that the revised methodology may lower their PPCP payments, if they were indeed higher than regional spending trends.
A full list of the Prospective Primary Care Payment Services that will be part of the capitation is available on page 82 of the RFA.
Enhanced Services Payment (ESP)
This payment is a per-beneficiary-per-month (PBPM) payment that is clinically and socially-risk adjusted to account for the beneficiary’s level of need; it is available to participants across all three tracks. This payment will be paid quarterly and is intended to support care management, patient navigation, behavioral health or other enhanced care coordination services. Since these goals overlap with some existing Medicare covered services under the Physician Fee Schedule, there will be non-payment of claims that align to these duplicative services (see Appendix F pg. 81 of the RFA).
The ESP will also be risk-adjusted through a variety of methods; (1) whether the beneficiary is enrolled in the Medicare Part D low-income subsidy (LIS), (2) the Area Deprivation Index (ADI) score based on the beneficiary’s residence compared to a regional reference population, and (3) the beneficiary’s most recent CMS-Hierarchical Condition Categories (HCC) risk score. Then, the ESP will be adjusted for geographic price differences. There will be a risk score growth cap to protect against HCC upcoding.
The stratification of ESP payments begins with identification of enrollment in Low-Income Subsidy; for those patients, the ESP PBPM payment will be at the maximum ESP of $25. For all other patients, CMS will calculate the ESP based on HCC score percentile and ADI percentile (see page 38 in the RFA for complete table). This ultimately produces a spread of ESP that varies between $2 PBPM and $25 PBPM. CMS estimates that the average ESP amount will be $15 PBPM in Track 1; $10 in Track 2, and $8 in Track 3.
Over time and advancement through the tracks in the model, the ESP will convert into a Performance Incentive Payment (PIP). Average ESP amounts will decrease by 50% for each Track level (i.e. Track 3 ESP payments are 50% less than Track 2 payments; Track 2 payments are 50% less than Track 1 payments). However, for beneficiaries who are either enrolled in LIS or in the top quartile of HCC and EDI, their ESP amounts will not be decreased over time, ensuring that the bump in ESP payment does not reduce over time, even as the participant moves across the progressive tracks.
Performance Incentive Payment (PIP)
The PIP is an upside-only payment adjustment that is calculated as a percentage of the sum of FFS and PPCP amounts to each participant for the PPCP services for their beneficiaries–i.e. It’s a % of primary care services, however it's being paid FFS or capitation or hybrid. The PIP will be calculated on an annual basis and begin in PY2 (CY2025). The PIP will be paid in two lump sums; the first in the first quarter of the performance year (based on expected average PIP); and the second payment in the third quarter of the year following the performance year.
In Track 1, participants will be eligible for a PIP of a maximum of 3%; in Track 2, the maximum will be 45%; and in Track 3, the maximum will be 60%.
To be eligible at all to pull down the maximum PIP, providers in tracks 2 and 3 must meet or exceed the 30th percentile nationally for Total Per Capita Cost. Then, providers must meet certain percentages per measure for each track (full table is available on pages 46 & 47 of the RFA). Each of these measures are weighted and maximum weight is placed on utilization/cost measures.
Health Equity in the Model
As with other models under the Biden-Harris administration, there are several components of Making Care Primary that align to the administration’s goals of advancing health equity.
The use of social risk adjustment in payments: the Enhanced Services Payments (ESP) portion of the fiscal model will be socially risk adjusted, utilizing the Area Deprivation Index (more information about that here) and other income-based measures.
Health Equity Plans: an element of the ACO REACH program being pulled over here is the requirement that all participants create a Health Equity Plan, which follows a CMS template to identify sub-populations, their needs, and interventions to help close equity gaps
Reducing Cost-Sharing for Beneficiaries: Participants will be able to opt-into a program to reduce cost-sharing for certain services, helping to broaden access to certain Medicare beneficiaries.
Collecting demographic data on beneficiaries: participants will be required to submit annual patient-reported demographic data including race, ethnicity, geography, disability status, etc. Participants will also collect health-related social needs data, as described in the Care Delivery Model section.
Outstanding Areas of Grey
CMS continues to be vague about multi-payer alignment to this model, expanding a bit on it in the RFA but mostly including language around Medicaid-alignment models, the vague hope of providing useful and clean data to practices (moonshot!), “goal-oriented and data-driven convening” with payers (which sounds suspiciously like a conference to me) and an email address for commercial payers to drop them a line and learn more about how they align to the model. It’s not that the goals outlined by CMMI aren’t laudable, it’s just that they are paper-thin on explaining how they’re going to draw in the multi-payer alignment they so desire.
Are you interested in learning more about the Making Care Primary model or how to apply for a CMMI value based model? HSG routinely leads clients through successful CMMI program applications. Reach out to learn more about how we can help your organization be successful in this or other CMMI programs.
About the Author: Juliette Price is the Chief Solutions Officer at Helgerson Solutions Group. Follow her on Twitter and connect with her on LinkedIn.
