Developing a Sustainable Care Delivery Payment Model for Children With Medical Complexity Free

The original HCIA-SNP grant concept was to expand care coordination and medical comanagement services on the basis of Gordon and Colby’s care model: A tertiary–primary care partnership based on the premise that each child is best served by a community primary care provider (PCP) working in conjunction with a tertiary complex care team to ensure a complete medical home.²  Eligibility criteria for the program are shown in Table 1.

The complex care team is composed of a pediatric nurse case manager working in a dyad with a care coordinator assistant (CCA), along with a complex care medical provider for each child and family.^7,16  The nurse–CCA dyad function as the CMC family’s principal care coordination contact. The CCA role and dyad concept were newly created during the project. CCAs provide most of the nonmedical care coordination services, freeing up nurses to address medical needs and expanding the number of children a nurse can serve. Team job descriptions are listed in Table 2.

Children and families are enrolled at an initial 90-minute visit attended by the nurse–CCA dyad, a complex care medical provider, and others on the basis of anticipated needs of the child and family. At the initial visit, program function is explained, families are informed that participation is voluntary, and consent for enrollment is documented if all involved feel the service is a good match for the child’s needs. A comprehensive care coordination summary is the product of the initial visit. The care coordination summary provides a roadmap for the child’s care: Their past medical and social history, their current status, and the family’s/patient’s goals for the future. The complex care team reviews the care coordination summary with the family, the child’s PCP, specialists, and involved community agencies. After the enrollment visit, patients are seen for care coordination updates at least every 6 months, and by the complex care inpatient consult service during all admissions and emergency department visits when able. Families and care colleagues can also access the complex care program providers 24 hours a day via the on-call line.

Before the HCIA-SNP grant, the Milwaukee program was serving ∼180 families with full-time equivalent staff of 4 nurses, 1 physician, and 3 advanced practice nurses (APN). The initial HCIA-SPA expansion plan projected an enrollment of 1800 CMC at Children’s Wisconsin and 200 at AFCH, with a parallel expansion of staff to serve the increased population. A new post-NICU segment was also started. The HCIA-SNP budget provided the funds to expand both programs. As the principal investigator for the HCIA-SNP, Medicaid approved all budget items, giving Medicaid a direct measure of program personnel costs.

A staff time-study instrument (Fig 1) was designed to enable documentation of work effort for each staff position per child enrolled while distinguishing historical reimbursed from nonreimbursed services delivered by the programs. Taken together, the time-study data and program cost information derived from the grant budget allowed Medicaid to estimate the personnel cost for delivering services to each child in the program, a process similar to time-driven activity-based costing.^17,18 

Each week, a 2-hour grant meeting took place between Medicaid HCIA-SNP project managers and representatives from each program to review progress, including specifics regarding the onboarding of new staff and patient enrollment. Medicaid project management team members also made site visits to both programs, observing clinical team huddles and attending outpatient visits and inpatient rounds. These site visits proved extremely informative because clinicians shared firsthand examples of how the program worked and why components deserved funding; Medicaid staff had the opportunity to experience the health care return their payments were producing. Semiannual presentations were also made to Medicaid leadership; the incorporation of patient/family stories was encouraged to help illustrate the programs’ purpose and impact. These leadership presentations offered yet another opportunity for the payer to see the impacts funding was having on Wisconsin CMC and their families.

Hospital and medical school finance and billing personnel were also updated on the project monthly. Each of these entities held a financial position in the project; the hospital as the employer of the nurses and CCAs and the medical school being financially responsible for the providers, administrative assistants, and research coordinator staff. Transaction personnel focused on the potential need to build different billing infrastructure to support the new payment approach, and when to start working directly with their counterparts at Medicaid.

Within the first years of the HCIA-SNP project, it became clear the programs would not be able to care for the number of CMC initially estimated. Before the grant, nurses managed panels of 35 to 45 children without the support of a CCA. The nurses’ work included a substantial amount of care coordination that did not require medical training. On the basis of the amount of time involved, we initially hypothesized that the nurse–CCA dyad could care for a panel of 90 to 120 children. CCAs are not licensed clinicians but typically have previous health care work experience, a strong understanding of the local health care system, and a passion for the program’s commitment to children and families. Regular discussions with the dyad teams elicited subjective information indicating that care quality appeared to suffer along with clinician job satisfaction as a dyad approached the 90 to 100-patient threshold. Program leadership further reduced panel sizes by applying a staff flex factor to the upper limit, reducing maximum panels by 15 patients. The flex factor was intended to ease the impact across teams if a dyad member left the program (eg, for other employment, maternity, or medical leave) the remaining teams would then have built in capacity to absorb additional patients more readily with less negative impact on care quality and team burnout until all the dyads could return to full staffing. The final panel size target for the dyads was set at 75 to 85 patients/families. The dyad figure was then used to set program capacity per the number of HCIA-SNP-funded dyad positions: Milwaukee’s program with 9 dyads, 675 to 765 children, and families; AFCH’s program with 3 dyads, 225 to 255 children, and families.

The creation of the CCA position is an example of matching clinical skills to operational duties; CCAs doubled a registered nurses’ panel size while lowering the cost of care per child served.¹⁷  Metrics must still be developed to assure the dyad approach has added value to family and patient outcomes. Subjectively, we feel care delivered to patients and families has improved and not suffered with the dyad approach.

Other position panel sizes were similarly set for the APN and physician groups. The physician academic provider panel ratios reflected a 15% to 20% reduction in clinical full-time equivalent to account for academic, non–HCIA-SNP administrative time and their responsibility in consulting with the APNs regarding challenging patient care issues. Table 2 shows the panel sizes for all clinical positions.

With the influence of the Affordable Care Act, the grant partners initially anticipated structuring a payment model around shared risk. During the second year of the grant, Medicaid changed this approach because they felt a shared risk model would not be fair to the clinical programs. The CMC population has extreme utilization risk; the projected number of CMC served by the programs would be far too small to expect predictable savings across a given time span. Under Medicaid’s direction, the teams shifted to establishing a targeted case-management (TCM) approach to pay for a portion of CMC program services.¹⁹ 

The TCM payments were directed at services that did not already have a conduit for payment or where not being reimbursed under past evaluation and management (E&M) fee-for-service approaches. As the payer for >85% of the children served by the Milwaukee program, Medicaid also had access to the historical E&M revenue generated from the care of CMC.

The time study revealed an interesting finding regarding effort spent by personnel related to the duration a child was in the program (Fig 2). More effort was expended during the first month after a child was enrolled, with subsequent months settling into a reduced effort of care that was fairly consistent month to month. These intervals offered a natural division between 2 distinct periods of care activity: Enrollment and the ongoing care time frames.

Medicaid and their actuarial consultants now had the elements needed to calculate staff program costs per patient served: Personnel costs from the grant budget, average personnel time spent per patient during program involvement, the projected number of patients able to be served by each personnel category, and the historical revenue generated by the traditional fee-for-service system. This information allowed Medicaid to calculate the reimbursement needed to sustain program staffing in the care of a CMC patient.

The new payment model design continued the E&M payment avenue along with hospital facility fees, for all face-to-face ambulatory and inpatient encounters, and added 2 new TCM codes for work historically not covered or not reimbursed under the past fee-for-service system. A 1-time enrollment fee and a monthly ongoing care payment corresponding to the distinct time-study work effort intervals.

The enrollment code pays for program staff services directed at preparing the initial enrollment visit and time-intensive work done during the calendar month of the enrollment. The main enrollment deliverable is the development of a comprehensive care plan. The ongoing care code is a monthly fee for continuing care coordination and medical comanagement services for each child in the program. To receive the ongoing care fee, the programs must document reciprocal contact with the child’s family or other personnel involved with the child’s care (eg, comanagement or care-coordination discussion with a subspecialist, PCP, or community service agent).

With information captured by the time study, the new TCM codes included payment for several areas considered essential for good care coordination yet not reimbursed under the past fee-for-service model. These previously nonbillable or nonreimbursed services such as health counseling, phone communication with families, clinical team huddles, school and community advocacy, discussions with other providers, clinically relevant office assistant time, and program data organization and analysis, activities that improve care efficiency and outcomes, now have an avenue for reimbursement. The initial payment structure called for a 1-time enrollment fee of $1144.20 and a monthly ongoing care fee of $413.56 per patient¹⁹ ; in 2022, these fees were increased by 15%. The fees form a capitated payment for a wide variety of care essential to the population.

The partnership also included telehealth as an acceptable method for interacting with patients and families. The initial thought was to better accommodate CMC families who lived long distances from the medical centers. The inclusion of a remote access structure proved fortuitus during the pandemic, allowing the programs to easily pivot to telehealth, retain service to a vulnerable population, and remain financially sustainable when several pediatric traditional volume-based systems were experiencing revenue shortfalls.

The complex care programs operating with the new TCM codes have the responsibility to avoid duplication of services that are provided by other agencies. The HCIA-SNP services are viewed by Medicaid as being different from Medicaid health maintenance organization care coordination services, allowing for patient participation and payment to both entities. Use of the new TCM codes by a health care organization requires the approval of Medicaid. Programs must assure similar comprehensive complex care services as established by the HCIA-SNP to receive approval. Since development of the payment model, an additional complex care program has been added at the Marshfield Clinic Health System in Marshfield, Wisconsin.

We agree with Kuo et al that additional research is needed to identify and quantify the value new payment models generate for stakeholders.²⁰  This section reflects where we feel the new payment model has increased value to stakeholders, with the acknowledgment many areas still require further investigation.

Medicaid’s initial interest in seeking HCIA funding arose from the potential to build on the cost reduction attributed to the Milwaukee program as outlined in Gordon’s original 2007 article.²  A similar pre–post cost analysis was done with a larger sample size during the HCIA-SNP granting period (excluding post-NICU population); the analysis also revealed a cost reduction for Medicaid, with the majority of savings coming from a reduction in inpatient utilization because of a decrease in length of stay.²¹  Although pre–post methodology is subject to bias from regression to the mean (versus program’s impact),^22,23  it remains encouraging that similarly structured complex care programs conducting randomized control cost impact trials also observed cost savings primarily attributed to length of stay reductions comparable to Gordon’s analysis.^9,24,25  The programs continue to work with Medicaid to analyze changes in cost of care associated with program participation, a vital element for continued payment justification.

The project produced a unique environment for payer and clinical staff to interact and learn from one another. Firsthand experience with clinicians gave the payer a clearer perspective about why funding certain activities was critical for better care. Clinicians also better understood what was simply not possible to include in a payment model because of the payer’s regulatory or budget constraints. These types of interactions should be fostered when forming future health care payment models. Ideally, when new clinical programs are being built, payers and clinical teams work together assuring stakeholder priorities are being addressed. This approach can mitigate some of the risk clinical teams take on when investing in a new process that payers may ultimately be unwilling or unable to fund.

Over the years, payers and health care delivery systems alike have invested heavily in the infrastructure needed to support a volume-based fee-for-service system; novel approaches to care requiring new payment model transaction methods must therefore overcome a great deal of system inertia. Much of the inertia is because of the investment needed by both provider and payer institutions to accommodate new billing and payment arrangements. Anecdotally, we are aware of promising novel clinical programs that were ultimately not put into practice at least partially because of the cost of needing to implement a new transaction infrastructure to support the innovation.

The HCIA-SNP project also faced this problem but was fortunate to have institutional finance and billing personnel communicating directly with payer counterparts, building out the transaction system as the payment model was being developed. Medicaid’s decision to use TCM codes also helped reduce transaction barriers because all parties were familiar with the TCM format, producing value to both payer and provider by minimizing the costs needed to accommodate the new system. We learned “throwing the billing/payment switch” can be a project unto itself and needs to be addressed upfront when considering new payment arrangements.

Leadership at the Wisconsin Department of Health also acknowledged the potential for wider benefits to the state because of the complex care programs’ activities supported by the payment model, benefits such as greater parent/family productivity and decreased need for state-funded social services (eg, Child Protective Services). These types of greater societal value beyond the direct care delivered to children and families deserve additional research. The clinical partners and Medicaid are also committed to working together in an effort to influence commercial payers to adopt a similar payment model.

The Wisconsin complex care programs are now entering their sixth year under the payment model, and the programs have moved from needing institutional gap funding to being financially self-sufficient while serving almost 4 times as many children and families. Teams now concentrate on delivering care across a spectrum of services felt to be beneficial to the CMC population, while not feeling financial pressures to focus only on specific fee-for-service volume-based activity. The new payment environment improves the flexibility clinical staff have when trying to improve patient and family outcomes, such as helping with social determinants of health; this flexibility and provider control potentially increases one’s meaning in work, leading to reduced burnout and increasing job satisfaction.²⁶ 

The programs hospitals (employer of registered nurses and CCAs) and medical schools (employer of providers and support staff) had to agree on a new revenue-sharing approach. Payments from the Medicaid program come directly to the hospital with a pass-through to the medical school to pay the physician, APN, and support staff. The legal and operational effort needed to establish a mutually acceptable distribution was substantial; however, both institutions came away with a better understanding of the working aspects needed to participate in future novel or capitated payment models where care involves both entities.

The performance of the payment model during the pandemic did not go unnoticed by the hospitals and medical schools. Many pediatric volume-based systems struggled to deliver care and generate revenue during the pandemic shutdown, whereas the HCIA-SNP model maintained uninterrupted care delivery to the CMC population and financially sustained each program.

The goal of health care is to deliver value to those we serve; quantifying the value brought to patients and families resulting from a new payment model is vital to defining success.^20,27  The HCIA-SNP project has been effective in bringing the proven concept of care coordination to more CMC and their families. Most of the program’s current metrics are, however, process in nature: Production of a comprehensive care plan with regular updates, documentation and progress toward family/patient goals, monthly contact with families, 24-hour call line access, inpatient consults for all program-admitted CMC, and follow-up after discharge. Although population health domains for CMC have been articulated by broad stakeholder groups, it remains unclear how best to measure program-influenced clinical outcomes.²⁸  Regardless, determining how complex care programs positively impact families, and how equitable the gains are, is a critical future direction for complex care research.

The programs also conduct annual family satisfaction surveys that continue to receive favorable year on year results, with >99% of families stating they would recommend the program to other CMC families. Questions remain about what program elements lead to the family satisfaction results and how the payment approach supports these activities.²⁹  Delineating these patient/family-centered quality metrics will enable complex care programs to be more efficient with their strategic activities. Not all families will need all services, with each having different burdens and capacities. Learning how to match complex care services to each family’s situation should produce equitable outcomes with greater efficiency.

The Wisconsin complex care story is an example of how working prospectively with a payer, delineating cost of care, and providing quality service to a population can produce a payment model benefiting all stakeholders. The HCIA-SNP model represents a hybrid fee-for-service/capitated system leading to sustainable program reimbursement ensuring services to participating CMC. Elements instrumental to the formation and implementation of the Wisconsin payment model:

• Determine program cost of care as the basis for reimbursement and determining future value.

• Cultivate relationships with payers, and include payers from the start when developing a new clinical approach.

• Avoid at-risk payment structures when serving small, high-utilizer populations.

• Include patient/family, primary care, and subspecialist as advocates to support the proposed model of care.

• Involve billing and coding specialists early from both the health system and payer to evaluate the need for developing new transaction infrastructure.

• Set initial care deliverables/quality metrics.

• Refine quality measures as additional patient/family-centered outcomes emerge.

• Maintain an ongoing transparent relationship with the payer to review value delivered to stakeholders and adjust, improve the care model.

The depicted model illustrates how other state Medicaid agencies can partner with complex care programs to bring value to CMC and their families across the country.