The Impact of Federal Value-Based Primary Care Programs on Participating Oregon Practices: A Snapshot

Abstract

This analysis examined data on quality, utilization, and cost from the 62 Oregon primary care practices that participated in the Comprehensive Primary Care Initiative (CPC Classic) and continued with the Comprehensive Primary Care Plus (CPC+) program in 2017, and compared these with Oregon primary care practices that participated in CPC Classic only or did not participate in either program. In 2017, practices participating in both programs performed better across all payer types than practices in the comparison group on 24 of 26 quality measures, including breast and cervical cancer screening and several chronic disease measures. Practices that participated in both programs also showed positive trends in emergency department utilization and avoidable emergency department utilization across all payer types. In addition, CPC-participating practices had lower per member per month adjusted claims-based costs among commercially insured members for numerous service types. The results indicate positive impacts of CPC program participation and lend continued support for CPC+ and other value-based payment programs in Oregon.

Introduction

Comprehensive Primary Care Plus (CPC+) is a federal initiative that seeks to strengthen primary care through care delivery transformation and multi-payer payment reform. Oregon was one of 18 regions selected by the Centers for Medicare & Medicaid Services (CMS) to join the program, which began in 2017. CPC+ builds upon the Comprehensive Primary Care Initiative (CPCI, or CPC Classic), in which Oregon also participated.

CPC+ is consistent with many of Oregon’s delivery system innovations, which include greater investments in primary care along with an emphasis on increased adoption of value-based payment models. Both the CPC Classic and CPC+ models require primary care practices to change the way they deliver care, with a focus on key functions such as care management, population health, and comprehensiveness and coordination.¹ The CPC+ model also aligns with the contractual value-based payment requirements for Oregon’s Coordinated Care Organizations.

The Oregon CPC+ Payer Group² came together to meet the CPC+ participation requirements and to advance and spread value-based payment programs in the state. Since 2017, participating Oregon payers have met monthly to identify opportunities for alignment, collaboration, and shared learning. The group’s goals are to:

• Support sustainable primary care transformation;
• Identify and share payer and clinic best practices to achieve program care delivery and payment model goals;
• Reduce fragmentation, seek simplification, and leverage existing resources; and
• Understand and demonstrate the value of CPC+.

The Importance of Analysis and ROI

The Payer Group recognized that primary care practices need sufficient data on quality, utilization, and cost to meet care delivery requirements, conduct quality improvement activities, and reduce costs. The group knew that reports that synthesize multiple payers’ data — rather than present slices of data from each payer — were a more comprehensive approach to performance reporting.

During CPC Classic, Oregon had explored a shared data aggregation solution but, for a variety of reasons, it didn’t come to fruition. Under CPC+, payers were eager to try again. To build consensus for data aggregation, the Payer Group undertook a multistep process that included creating a framework and principles for decision-making. They outlined a set of agreements and considerations, including:

• Recognition that payers were in different stages of collecting clinical and claims data;
• Agreement that progress required payer commitment and leadership;
• Acknowledgment of the importance of as many payers as possible participating;
• Understanding of the challenges and opportunities of collecting clinical and administrative data; and
• Affirmation that any solution should support payer evaluation of the payment model and whether/how to sustain and spread it.

In 2018, the majority of participating payers committed to aggregating claims data with Comagine Health to help payers and practices learn from the CPC+ experience.³ They wanted Oregon-specific data to help stakeholders across the state understand how the model impacted primary care and the potential for spread beyond participating payers and providers. Comagine Health works with Oregon’s largest health insurers plus the Oregon Health Authority and CMS to develop a comprehensive claims database — the Oregon Data Collaborative — which includes claims from 2015 to present, representing data for over 3 million covered Oregonian lives. Our database includes 80% of the fully insured population, 23% of the self-insured population, 100% of the Medicaid population, and 87% of the Medicare population in Oregon. Claims data are submitted quarterly to Comagine Health’s data services vendor, who cleans and aggregates the data, calculates measures, and populates results into Comagine Health’s secure online reporting portal. At no charge, primary care and women’s health provider organizations in Oregon are able to securely access quality, utilization, and expenditure measure results for their organizations and drill down to practice, provider, and patient-level results.

Aggregated data opened the door to examining whether and how the CPC payment models made a difference for participating practices on quality, cost, utilization, and outcomes across payers. The Oregon CPC+ Payer Group, in collaboration with Comagine Health, launched Data Bytes, a series of brief, infographic-heavy publications highlighting key analytic findings to share progress across participants following a multiyear collaborative process. This report summarizes two Data Bytes published by the Payer Group in 2020 measuring quality, utilization, and cost.

Key Findings

Comagine Health examined 2017 data from the 62 Oregon primary care practices that participated in CPC Classic and continued with the CPC+ program in 2017 (CPC-participating practices). Comagine Health compared these with Oregon primary care practices that participated in CPC Classic only or did not participate in either program. The practices participating in CPC+ only were not included in either the comparison group or the participating practices group because 2017 was the first year of the CPC+ model. For each measure, Comagine Health took the average of all practices in each group where the practice had at least 30 attributed primary care patients in the measure denominator.

CPC-Participating Practices Performed Better Across All Payer Types on Most Health Quality Measures

The first analysis shared with practices included 2017 results for the quality measures that Comagine Health computes for its adult cost-of-care reporting, across all payers, as well as by payer type: commercial, Medicaid, and Medicare (fee-for-service and Advantage combined).⁴ (Measure descriptions are available in Comagine Health’s Technical Appendix.) Participating practices performed better on 24 of 26 quality measures (Table 1).

Below, we highlight six of the 26 quality measures, starting with the findings that practices participating in both CPC+ and CPC Classic performed better than comparison practices across all payer types on Breast Cancer Screening and Cervical Cancer Screening quality measures (Figure 1). Practices that participated in both CPC Classic and CPC+ also had higher performance rates than comparison practices across all payers in Comprehensive Diabetes Care (comprised of three measures) and Annual Monitoring for Patients on Persistent Medications quality measures (Figure 2). Practices with fewer than 30 patients were excluded from this analysis, which is for the 2017 calendar year.

We focused on these six quality measures because they:

• Showed statistically significant results (P<0.05) in at least the all-payer combined category;
• Were selected as priority measures by the CPC+ Payer Group as well as by commercial payers and Medicaid; and
• Fit into one of two groupings (preventive screenings or care of patients with chronic conditions) that reflect the focus areas of the CPC+ program.⁶

Breast Cancer Screening and Comprehensive Diabetes Care Measures include Medicare fee-for-service data that were made available through the Medicare Qualified Entity program, while other measures include Medicare Advantage data only.

CPC-Participating Practices Showed Positive Trends in Utilization and Costs

The second analysis highlighted utilization and cost measures. Utilization measures include commercial, Medicaid, Medicare Advantage, and Medicare fee-for-service data; cost measures include commercial data only, due to restrictions on sharing noncommercial cost data.⁷ For each measure, Comagine Health took the average of all practices in each group where the practice had at least 30 attributed primary care patients in the measure denominator. Practices with fewer than 30 patients were excluded from this analysis. Medicare fee-for-service data were made available through the Medicare Qualified Entity program.

Participating practices had significantly more adult outpatient visits than comparison practices across all payer types (Figure 3). These findings are consistent with the CPC program emphasis on enhancing primary care and care coordination to improve patient outcomes.

Participating practices had lower adult emergency department (ED) utilization and fewer avoidable ED visits than comparison practices. Significance varied by payer type (Figure 4).

Practices participating in both CPC Classic and CPC+ had lower per member per month adjusted claims-based costs of adult emergency services, inpatient services, and outpatient facility costs among commercially insured patients (Figure 5). Differences were statistically significant across all three service areas (Figure 5).

Although CPC-participating practices receive non-fee-for-service funds through care management fees and performance-based incentive payments, Comagine Health did not try to compare professional costs between practices because this additional professional reimbursement is not captured in available claims data. This analysis reflects only claims-based facility costs, which are not expected to be directly affected by the additional CPC payments.

Cost measurements were adjusted for practices’ average risk score. Practices participating in both CPC Classic and CPC+, on average, had adult patient populations that were more likely to be hospitalized or become high-resource users (were sicker) than the comparison group. In the commercial population, the average adjusted risk of the participating practices was 1.06 and the comparison practices’ average risk score was 0.96.

Implications

While we cannot claim causation, we found that participation in CPC Classic and CPC+ is associated with positive outcomes in cost, quality, and utilization compared to nonparticipating clinics. Combining all payer types frequently resulted in statistically significant differences when an individual payer type’s results may not have been statistically significant or sometimes showed better performance in nonparticipating practices. For example, the measure Annual Monitoring for Patients on Persistent Medications – Diuretics, a subset of the “Total” measure, showed statistically significantly better performance among participating practices when all payer types were combined. However, within Medicaid, participating practices performed worse on this measure than nonparticipating practices (not statistically significant). Also, only the Medicare result was statistically significant among payer types when examined alone, but when all payer types were combined, the result was significant. This is a demonstration of the value of aggregated data; if we were to assess these measures using only one payer type, the overall impact on practices’ patient populations would not be visible.

Since the Payer Group published the Data Bytes, stakeholders across the state and the nation, including CMS, have expressed interest in learning more. Payers in the state are considering how this data informs decisions about value-based payment efforts; a number of regions have asked about the process the Payer Group used to get agreement on data aggregation and publication. The policy implications remain to be seen, but we expect that the results will shore up support for the CPC+ payment model and for the continued shift to implementation of aligned value-based payment models across the state and, perhaps, nationally.

Moving Forward

During the last 18 months of CPC+, the Oregon Payer Group has committed to continue convening regularly, with a focus on gathering, analyzing, and disseminating information that demonstrates the benefits of value-based payments via the CPC+ model. The Payer Group will work with Comagine Health to create at least five more Data Bytes. Comagine Health and the Payer Group are exploring the following topics: primary care cost, quality and utilization trends over time; behavioral health integration; specialty care; care coordination and care transitions; and emergency department and inpatient utilization. Future reports will include a three-way comparison of practices that participated in both CPC+ and CPC classic; practices that participated in CPC+ only; and those that participated in neither program. In an encouraging development, CMS has expressed interest in funding additional Data Bytes and will make its decision to do so over the next year.

The COVID-19 pandemic has strained the nation’s health care system, especially primary care. The rapid changes in health care utilization have further demonstrated the inadequacies of a fee-for-service payment environment. As a result, many are turning away from fee-for-service and looking to a system focused on value-based payment models that allow for consistent, flexible payment structures. Understanding the opportunities and challenges of value-based payment, highlighted through longitudinal data aggregation efforts, is an integral part of this journey. We hope the Oregon experience will serve as a launching pad for other states to analyze and evaluate their value-based payment experiences so we can continue to learn from and share with each other.

How Comagine Health Conducted This Study

The data source is Comagine Health’s Oregon Data Collaborative, a voluntary all-payer claims database in Oregon. All measures are from calendar year 2017, except the comprehensive diabetes care measures. Due to data integrity issues with those three measures in the calendar year 2017 data, they are based on the 12 months ending June 30, 2017. Medicaid data were not available for those measures for that period.

For each measure, Comagine Health took the average of all practices in each group where the practice had at least 30 attributed primary care patients in the measure denominator. Analysts used generalized linear model regression to compare rate scores and percent scores by CPC program participation status (participating vs. comparison) for all measures and stratified by payer type. Statistical significance was set at P<0.05. Mean values were generated by measure for each CPC status. All analyses were conducted using SAS Software 4.0 (SAS Institute Inc., Cary, NC).⁸

The Oregon Data Collaborative does not have claims information for all the self-insured population in the state, national commercial insurers, Tricare, or services provided by the Indian Health Service. This point-in-time study did not assess practice performance upon joining a CPC program and cannot rule out selection bias among practices joining a CPC program.

NOTES

¹ Centers for Medicare & Medicaid Services. CMS.gov. Comprehensive Primary Care Plus. https://innovation.cms.gov/innovation-models/comprehensive-primary-care-plus. Accessed November 2, 2020.
² The Oregon CPC+ Payer Group is comprised of: Advanced Health, AllCare Health, CareOregon, InterCommunity Health Network CCO, Moda/Eastern Oregon CCO Oregon Health Authority, PacificSource Health Plans, Primary Health, Providence Health Plan and Providence Health Assurance, Trillium Community Health, UnitedHealthcare, Willamette Valley Community Health, Yamhill Community Care, Centers for Medicare & Medicaid Services. Note: the organizations in italics are not participating in the CPC+ Payer Group as of September 2020. Primary Health and Willamette Valley Community Health no longer operate as coordinated care organizations or payers in Oregon; Trillium Community Health does not contract with any CPC+ providers; and Advanced Health decided not to participate in the final 18 months of the convening contract.
³ Comagine Health is one of the three co-conveners of the Oregon CPC+ Payer Group. Their role as the data aggregator remains separate.
⁴ Medicare fee-for-service data is available through Comagine Health’s participation in the Medicare Qualified Entity program.
⁵ In two cases in our overall analysis, the combined rate is in the opposite direction of all the individual payer results. The measures Antidepressant Medication Management: Effective Acute Phase Treatment and Antidepressant Medication Management: Continuation Phase Treatment showed lower rates for every payer type (commercial, Medicaid, and Medicare Advantage) when looked at separately, but higher rates when all payers were combined. This is because CPC-participating practices have a lower proportion of Medicaid members than nonparticipating practices, and Medicaid has a much lower compliance than other payer types on these two measures. Within Medicaid, there was no statistically significant difference on these two measures between CPC-participating practices and nonparticipating practices, although within the commercial payer type the difference was statistically significant.
⁶ Centers for Medicare & Medicaid Services. CMS.gov. Comprehensive Primary Care Plus. Available at: https://innovation.cms.gov/initiatives/comprehensive-primary-care-plus/. Accessed November 2, 2020.
⁷ Medicare fee-for-service data is available through Comagine Health’s participation in the Medicare Qualified Entity program.
⁸ CPC+ Payer Group. Resources. http://cpcplusoregon.org/resources. Accessed November 2, 2020.