What Services Should Be Included for Measurement in Value-Based Contracts?

Blog Series: The Value of Value-Based Care v1.02 -

As discussed in our initial blog post of this series, value-based contract models can take many forms—population global budget, “bundled” payment, and centers of care excellence are examples. With each model, it’s important to consider the populations and services to be included. Those populations and services where performance can be measured well, and providers can impact outcomes should be a focus. Services where measurement is less precise, or challenges exist for improving care are not well-suited for a value-based contract (VBC). This blog provides insights to guide decisions on designing a VBC program—in particular those populations and services to be included in measurement.

We will focus on population global budget VBC models, which are increasingly used in designing a VBC in the U.S. and cover all or most healthcare services for a defined population managed by a provider. Provider organizations are rewarded for delivering better care for a population—at a lower cost. Cost of care performance under these programs is based on a specific set of measured services delivered to patients assigned or attributed to the contracted provider.

Total Cost of Care Illustration

What services to include in measurement?

Decisions to include or exclude services from measurement can be guided by a number of factors, including a provider’s ability to impact and improve care and also the ability of program performance to be measured with precision and without bias. For example, patients experience costly, rare and complex diseases and medical events such as ESRD, cancer and trauma. For many of these conditions, a contracted provider may have more limited involvement in the immediate course of care, resulting in the patients and services often being excluded from a VBC. Other factors limit the comparability of performance across provider entities and with contract targets, such as competing payment programs, differences in benefit design or new technology.

Patients & clinical scenarios challenging measurement under a VBC

Most population global budget VBCs include a risk adjustment methodology that can adjust adequately for a wide range of patients and conditions to support valid measures and comparisons—especially where the contracted population is of sufficient size. However, there are certain patients and clinical scenarios that challenge measurement under a VBC—where prevalence is low, care is complex and treatment cost is potentially high and unpredictable. Even the best risk models can fall short in fully capturing their expected impact on the cost of care. These patients and events may also introduce clinical and financial uncertainty that has little to do with the performance improvements being targeted by a VBC. If a provider organization attracts a relatively large number of these patients, their measured performance could be negatively impacted. Many VBC programs apply exclusions or adjustments to address these challenges. Below are some examples.

Patients/Clinical Scenario Description
End-stage Renal Disease (ESRD) Advanced chronic kidney disease is a progressive condition and the management and treatment of these patients is costly. Government programs exist to finance and care for these patients, including through Medicare.
Institutionalized patients Individuals residing in a nursing home or other institutional setting are atypical and often are considered for exclusion or further adjustment.
End-of-life care (EOL) Patients at the end of life consume significant healthcare resources. For example, Medicare beneficiaries comprise less than 5% of that population but are responsible for almost 25% of annual program costs.1
High cost outlier patients Most VBCs will adjust measures of performance for patients with catastrophically higher medical costs. The cost of these patients may drive the performance of a provider group—potentially masking true improvements in cost of care. Many programs apply a “stop-loss” methodology to address these patients and their costs—for example, including only the first $100,000 of annual patient costs in measurement.
COVID-19 The recent pandemic has had an outsized impact on U.S. healthcare systems. For some services, such as acute inpatient stays, the mix and intensity of services changed to meet immediate needs. For other services, such as elective surgeries and routine and follow-up care, a decrease in cost and use was evident. Adjustments and exclusions when measuring performance should be considered—in particular for VBC models where performance is based on comparisons with prior experience.

While these patients and these services may be excluded when measuring performance, an argument can be made to consider these services and their costs in some way. For example, coordinating EOL care is an important responsibility for providers managing elderly populations. It’s also arguable that practice variation exists for EOL care and there are opportunities for better care at a lower cost. If appropriate adjustments or measures can be applied to address any potential bias, incentivizing providers to adopt strategies to improve EOL care makes sense. Providing incentives to continue to manage high cost patients is another opportunity—for example, holding the provider responsible for a small percentage of costs exceeding a stop-loss threshold.

Operational, benefit and program related challenges for measurement under a VBC

Payment, benefit and program differences can also create measurement challenges for a VBC. For many VBC designs, exclusions and adjustments are indicated for these services.

Operation, Benefit, Program Description
Scope of benefits VBC designs often exclude services where a member does not have a benefit for the service with an insurer. Outpatient prescription drugs are an example. Methods are available to adjust for these and other benefit differences when measuring cost of care.
Other VBC arrangements Organizations may implement multiple, potentially overlapping, VBC programs. A bundled payment design for procedures and a global budget VBC is one example. Decisions are required on how to measure the cost of care for the services and patients covered by multiple programs.
Existing care management programs Many organizations operate programs designed to improve population health—programs for high cost complex conditions or for behavioral health patients are examples. For behavioral health programs, a VBC design may “carve” those services out of the measurement of cost—given that a different entity is responsible for managing that care.
Sensitive and limited data Regulations protecting data for certain conditions and treatments may challenge valid and consistent measurement of cost and utilization for a VBC. For example, HIV/AIDS, behavioral health and other conditions often involve privacy issues and are excluded or censored in the data available to support measurement and improvement.

Other value-based contract models

This blog focuses on a global budget VBC, but similar exclusions can be applied to other models such as bundled payment. The Bundled Payments for Care Improvement (BPCI) Advanced model implemented by the Centers for Medicare and Medicaid Services (CMS) is focused on the care surrounding an acute inpatient stay. However, BPCI exclusions are made for subsequent hospital admissions for organ transplants, major trauma, cancer-related care, and ventricular shunts. CMS also addresses any overlap in services between competing VBCs—in a program specific way. Participants in the BPCI Advanced model cannot participate in the Comprehensive Care for Joint Replacement (CJR) episode program for the same clinical area. For the Oncology Care Model (OCM), CMS allows episodes to run concurrently with BPCI Advanced episodes but adjusts performance measures for that proportion of an episode that overlaps.2

Ensuring that the same criteria is being applied to targets & benchmarks

Ultimately, the experience of a provider and their covered population will be evaluated using comparisons with other experience, including prior costs, peer performance, or both. In any of these comparisons, consistency in measurement methodologies is of utmost importance, including the exclusions and adjustments described throughout this blog.


[1] Duncan I, Ahmed T, Dove H, Maxwell TL. Medicare Cost at End of Life. Am J Hosp Palliat Care. 2019;36(8):705-710. doi:10.1177/1049909119836204

[2] Centers for Medicare and Medicaid Services Center for Medicare and Medicaid Innovation, Bundled Payments for Care Improvement Advanced Request for Applications, 01/08/18 https://innovation.cms.gov/files/x/bpciadvanced-rfa.pdf

Recent Posts