Abstract
In its most recent report on health care costs, the federal government estimated that total health care spending in the United States in 2016 totaled $3.3 trillion, accounting for 17.9% of gross domestic product and resulting in average per capita health care spending of $10,348. 1 Although the annual rate of increase in health care spending has fluctuated over the past decade because of a variety of factors, concern about the unsustainability of health care cost increases on the part of federal and state governments, employers, and ordinary people abound and echo similar concerns expressed by the same entities for decades.
Numerous strategies to slow the rate of rise of health care costs have been tried. Two of the main strategies that have been pursued over the past decade have been to change the way health care providers are paid, with an eye toward paying for “value” rather than “volume” and changes in insurance benefit designs that make individuals more accountable financially for the care they receive.
The key obstacle to changing from a volume- to a value-oriented payment system is that most providers do not feel confident that they can make the change without undermining their financial viability. Forces for change vary markedly market by market and the United States, as a whole, is still in the early stages of “health care transformation,” with fee for service (FFS) remaining the dominant form of payment throughout the country. Although providers are aware that many of the services they deliver provide little or no value, they continue to provide those services because their current business model incentivizes delivery of as high a volume of services as possible.
This paper examines the types of changes in payment and benefit design that are being employed currently in an effort to control health care costs, the actions providers should be taking to maximize their chances of success in the new health care environment, and metrics delivery system leaders can use to gauge how prepared they are to take on financial risk.
Attempts to Rein in Costs
There are 3 main problems with FFS from cost and quality perspectives: (1) it creates an incentive to provide more services than are necessary, (2) it does not provide an incentive to be efficient in care delivery, and (3) it does not tie payment to quality or outcomes of care. As a result, FFS has driven substantial increases in health care costs, as well as unjustified variations in virtually every aspect of practice and in clinical and financial outcomes.
Numerous payment innovations have been introduced over the past decade that (1) provide an incentive to improve quality of care, (2) increase financial risk borne by health care providers, and (3) increase the share of costs that are borne by patients. One result has been a plethora of new terminology to describe new forms of payment and new types of insurance benefit designs.
Payment innovations
Foremost among new payment terminology are performance-based payment (PBP), value-based payment (VBP), and risk-based payment (RBP). Under PBP, some portion of payment is based on whether or not particular performance metrics are satisfied. Commonly used performance metrics include measures of quality of care and achievement of certain cost targets, use of particular types of personnel (eg, case managers integrated into primary care practices), and reporting of certain types of information.
Most, but not all PBP programs are “upside only”; that is, payment of bonuses for achieving established performance targets without any “downside financial risk” in the sense of assumption of responsibility for a portion or all of the difference between financial targets and actual financial performance when the latter is worse than financial targets. PBP programs, however, sometimes impose financial penalties based on worse than target quality or safety of care.
There are many types of PBP, including “Bundled Payments” (payment for specified episodes of care or for all care over a specified period of time [ie, capitation]); Accountable Care Organization shared savings programs; Pay for Performance; the Centers for Medicare & Medicaid Services (CMS) Hospital Value-Based Purchasing Program; the CMS Hospital Readmissions Reduction Program; CMS’ Value Modifier Program (also called the Physician Value-Based Modifier); and CMS’ Hospital-Acquired Conditions Program.
VBPs are, for all practical purposes, the same as PBPs. The term VBP is used more frequently than PBP when the payment arrangement involves some downside financial risk or when providers have to exceed a target level of financial performance in order to share in the savings they generate. Shared savings programs usually have a “quality gate” (ie, providers have to achieve certain quality targets in order to be entitled to share savings). Not all VBP arrangements involve true downside financial risk for providers, however.
RBPs always involve downside financial risk for the provider. Performance targets typically include both cost and quality of care targets, with achievement of the quality targets being required in order to share in savings.
The Health Care Payment Learning and Action Network (LAN), a public–private partnership, recently refined an earlier CMS framework for classifying alternative payment models (APMs), (ie, alternatives to traditional FFS). 2 Table 1 maps the LAN framework to the more commonly used PBP, VBP, and RBP terminology.
Alternate Payment Model Framework and Common Payment Descriptors
APM, Alternative Payment Model; LAN, Health Care Payment Learning and Action Network; FFS, fee for service; PBP, performance-based payment; VBP, value-based payment; RBP, risk-based payment.
To date, through its Medicare and Medicaid programs, the federal government has been a major driver toward VBP. In January of 2015, Secretary of Health and Human Services (HHS) Sylvia Burwell announced ambitious VBP goals, namely that 30% of all Medicare provider payments would be tied to value through APMs by the end of 2016 and that 50% would be by the end of 2018. 3 It is unlikely that this 50% target will be achieved, however, given that the federal government's movement to VBP programs has slowed under the Trump Administration. HHS Secretary Price canceled mandatory participation in 3 bundled payment programs. Secretary Azar and CMS Administrator Verma have signaled a willingness to test mandatory payment models, but feel that VBP programs should be evaluated further to determine if they are achieving their goals before they are expanded. They have not established specific VBP targets. In 2017, providers reported that 26% of their patients (and 23% of their net patient revenue) were in VBP arrangements, but they expect that in 2020 those numbers will double to 52% (and 48%). 4
New types of insurance benefit designs
A second strategy for reining in health care costs has focused on new insurance benefit designs that increase individuals' sensitivity to the cost of health care services. Changes in insurance benefit design have included increasing deductibles, co-payments, and coinsurance; implementation of tiered and preferred provider networks with lower or no patient co-pay or coinsurance when preferred providers are used; and reference pricing, under which beneficiaries' choice of a provider for a particular service, such as a total hip replacement, is not limited, but the amount their insurance will pay for that service is fixed. Reference pricing has been shown not only to drive beneficiaries to lower cost providers, but also to lead higher cost providers to reduce their fees. 5
Requirements for Success Under VBP and RBP
In order to be successful under VBP/RBP, providers need several capabilities. The first is strong administrative and clinical leadership that is committed to making the changes that are required for success under VBP/RBP. These leaders need to articulate a clear vision of where the organization is heading and a clear rationale for why it needs to go in that direction. Organizational leaders also need to create or reinforce a culture that prioritizes quality, efficiency, and the best interests of patients over personal and organizational income.
The second requirement is a set of capabilities that are needed to proactively and effectively manage the care of all people for whom a provider is responsible (ie, population health management [PHM]), rather than simply responding to patients' needs when they reach out because of an acute problem. Several types of PHM capabilities/tools are essential. The first relates to data analytics. Health care is like anything else – you can't manage it if you can't measure it. In the case of health care data analytics, providers need to have strong data governance in place, and be able to collect and/or import multiple types of data (eg, clinical, insurance claims, patient-reported outcomes) from multiple sources, including electronic health records (EHRs), insurers, and patients. Some of the most valuable data are stored in EHRs as free text, rather than in structured, analyzable fields. Providers should invest in making it easy to capture key data elements, particularly those related to quality metrics, in analyzable fields or invest in natural language processing software that can extract such information from free text. Providers also need to be able to perform quality assurance (QA) on data, and clean, integrate, normalize and house data securely, all of which require technology and substantial expertise.
Although meaningful data analyses require effective QA, integration and normalization of data, it is the analyses themselves that are essential to assess current performance, identify actionable opportunities to improve performance and patient outcomes, drive provider and patient behavior change, and monitor performance over time. Many providers assume that the purchase of a data management and analysis solution will provide the data analyses, reports, and dashboards that are needed to improve and monitor performance over time. Unfortunately, that often is not the case. As a result, providers often purchase a data warehouse/analytics platform solution, only to find that they have underinvested in the design and production of reports and dashboards for key stakeholders. Providers should identify their key internal customers for data analytics and determine the information they need in order to satisfactorily perform their responsibilities. Managers need actionable data in order to manage, and different stakeholders/users need different types of information. Actionable analyses include segmentation of a population by risk (ideally, by modifiable risk), identification of gaps in care, identification of unjustified variations in care and opportunities to improve performance, prediction of undesirable events, and monitoring of the performance of physicians, practice sites, hospitals, and systems as a whole compared to benchmarks. Health systems also need to make sure that they package the results of analyses in ways that are easy for internal stakeholders to digest, and to make them available in a user-friendly way. To have the most impact on clinician behavior, actionable information should be integrated into clinical work flows (eg, in the EHR), rather than simply being provided through a periodic stand-alone report.
Provider organizations and health systems need to be able to report on the metrics that will be used to assess their performance under each of their performance-/risk-based contracts. Given the plethora of quality metrics in use, providers should decide on and monitor their performance on a set of metrics that they want to use to “market” themselves to insurers, employers, and patients and to serve as the basis for assessment of their performance under value-based contracts.
A lot of attention is being given to “big data” (ie, being able to leverage new types of data and data on more individuals) and to use of artificial intelligence in data analyses. Collection and use of data on social determinants of health, patient-reported outcomes, and patient preferences would be of particular value. However, despite the hype about big data and artificial intelligence, providers likely will not benefit from investment in them at this time, compared to investing in making sure that a more standard data analytic foundation is in place.
Another capability that is required for success in PHM is having personnel who are skilled in configuring/programming whatever EHR(s) is/are in place. Many providers do not take full advantage of their EHR because they don't have personnel who can configure workflows based on best practices, or integrate the results of data analyses and clinical decision support into the EHR to enable physicians to work efficiently and help them deliver high-quality care. Investment in a small, dedicated cadre of people with the requisite expertise in configuring/programming the EHR will produce substantial improvements in efficiency and quality and reduce provider frustration.
A third critically important PHM capability is care management (ie, utilization management, case management, disease management). To be most efficient, these programs should leverage a diverse team that includes nurses, community health workers, social workers, and pharmacists working in a coordinated fashion. These programs also need software that drives clinically appropriate workflows and helps to manage personnel administratively (eg, assess individual care manager performance, balance workloads), as well as a messaging platform through which care managers can communicate with patients.
A fourth important PHM capability is effective provider network development and management. Provider organizations entering into VBP/RBP contracts need to make sure that the number, type, and location of their network providers are well matched to the size, location, and clinical needs of the population for which they would be taking risk. To be successful under VBP/RBP, provider groups and health systems need to attract and retain strong performers and exclude poor performers. Given the increasing information technology, reporting, and other administrative demands on providers, one strategy for recruiting and retaining strong physicians in a network is to provide them with the technical and administrative support they need. Management of a high-performing network of physicians and facilities under risk sometimes requires exclusion of poor performers – something that many health systems are reluctant to do out of fear of losing the patients those poor performers refer to in-network facilities. If poor performers cannot improve their performance sufficiently, they should be excluded from at-risk provider networks and encouraged through other incentives to continue to use network facilities.
Effective network management also requires incentives to physicians and others in your network who are aligned with those of the group or system by which they are employed or with which they are affiliated. Network managers need to ensure that network goals and incentives are clear and understood.
A fifth capability that will increase the likelihood of success under VBP/RBP relates to actuarial and contracting expertise. Insurers have substantial actuarial and contracting expertise and experience. Providers may inadvertently enter into disadvantageous contractual arrangements unless they can bring similar expertise to the negotiating table. Providers thus need to make sure that they enter into value-/risk-based contracts (VBCs/RBCs) on terms that provide a reasonable chance of earning bonuses/shared savings/unshared savings relative to a benchmark or capitated payment rate.
In all VBCs/RBCs, providers should insist, as a condition for entering into the contract, that the payer will provide claims data for all patients covered by the contract. Those data are essential for PHM and for monitoring performance under the contract. Providers also should try to drive the selection of quality measures that will be used to evaluate their performance or prioritize VBCs/RBCs that will employ performance metrics that the provider is used to using and managing. Providers are overwhelmed by a growing number of quality measures or similar measures with varying technical specifications. It is not uncommon, for example, for a provider to be evaluated on the basis of 100 quality measures under various VBCs/RBCs. This is untenable to manage and track. Providers should advocate for use of measures on which they perform the best.
The Importance of Volume
One of the most important requirements for success under VBP/RBP is a sound strategy for attracting new patients and holding on to existing patients. Typically, advocates of VBP talk about the need “to transform from volume to value.” To be successful under VBP or RBP, however, a provider organization needs to transform from volume to “value plus volume.” The key to increasing patient volume in a value-oriented world is the delivery of value. That means delivering high-quality care at a lower cost. It also means improving patient experience by providing outstanding service and convenience, including enabling patients to get same-day appointments, and investing in telehealth/telemedicine and other ways of interacting with patients outside of face-to-face encounters (eg, home monitoring, “tele-visual visits” with physicians, digital physical therapy). Unless provider organizations are able to increase the number of patients for whom they care as part of their transition to value, they are unlikely to be successful financially under new forms of payment. Delivery of higher value care needs to be tied to a strategy for attracting incremental patient volume.
For example, providers need to make sure their performance is good enough for them to be included in narrow networks and preferred provider networks established by payers. In the future, inclusion in such networks will be critical to increasing patient volume, as well as maintaining it.
Another way of increasing volume and improving financial performance is to reduce “leakage” to out-of-network providers. Under VBP/RBP, leakage of patients to out-of-network hospitals or physicians not only reduces revenue, but typically increases costs substantially by virtue of the high fees out-of-network providers typically charge.
Finally, the hospitalizations that PHM is designed to prevent tend to be for common chronic conditions (eg, for heart failure, asthma, diabetes), which are low-margin hospitalizations. Therefore, delivery systems need to develop a plan that will enable them to replace these preventable low-margin hospitalizations with higher margin, elective hospitalizations (eg, by becoming a preferred referral site for high-margin procedures, by developing a program/Center of Excellence for management of selected types of complex patients, such as those with complex cardiac problems).
Operating in Both FFS and VBP Worlds
During the transition from “volume to value + volume,” providers need to maximize FFS revenue in ways that do not undermine the transition to a value orientation. For example, identification and closure of gaps in care can generate substantial revenue at the same time that it improves quality; more efficient and higher quality management of hospitalized Medicare patients will increase margin under Diagnosis-Related Group payments and reduce penalties for quality deficiencies; and improvement of patient transitions from inpatient to outpatient care will reduce readmissions, as well as costs for post-acute care.
Providers also should ensure that patient diagnoses and information regarding interventions that have been undertaken to manage a particular condition (in both the inpatient and outpatient settings) are accurately documented in order to receive maximum appropriate payments for their patient case mix.
Assessing Readiness to Take on Risk
Hospital chief executive officers and chief financial officers often ask how they can judge their organization's readiness to take on financial risk.
Before addressing that question, it is important to note that providers do not need to think of taking on financial risk as a binary proposition. Rather, they typically have some discretion with regard to the types of value-based contracts into which they enter. Contractual dimensions that could vary include: (1) the type(s) of payers with whom a provider contracts (eg, Medicare, Medicaid, commercial fully insured, commercial administrative services only, direct arrangements with self-insured employers); (2) the type(s) of patients for whom they take risk (eg, all, adults only, pediatrics only, only patients with particular diagnoses, only patients undergoing particular procedures); (3) the geographic areas in which the population to be cared for resides and lives; (4) the providers (ie, facilities, physicians) who are subject to the contract; and (5) the amount of financial risk they take (eg, full, up to 75%, up to 50%, up to 10%) and whether or not to purchase stop-loss insurance or reinsurance.
Even with such discretion, the leaders of provider organizations would like to have metrics they can use to assess their readiness to take risk. When thinking about such metrics, one should distinguish between metrics that (1) reflect the provider's current cost and quality of care versus those that (2) provide insight into the provider's preparedness to improve performance.
Ironically, because risk contracts often use a provider's historical performance (as compared to regional or some normative performance) as the benchmark against which future performance is judged, providers whose historical performance is less good are, everything else being equal, in a better position to succeed under a RBC than those whose performance is better. The reason is that it is easier for those with poor performance to improve. As a result, higher historical utilization (eg, hospitalizations and hospital days/1000 at-risk lives, emergency room [ER] visits/1000 at-risk lives, skilled nursing facility [SNF] admissions/100 discharges; number of high-end radiology scans per 1000 at-risk lives) and cost reflect poorer current performance, but likely indicate a higher potential to be successful under risk than lower historical utilization (at least in the short run) – unless a normative benchmark will be used in contracting.
Instead of, or in addition to, looking at actual utilization and cost, leaders of provider organizations should look at measures of actionable opportunities to improve performance, which likely provide a better indication of the likelihood of being successful under risk. Such indicators include utilization measures, such as hospitalizations for ambulatory care-sensitive conditions and 1-day hospital stays; discretionary ER visits; the cost of ER “frequent flyers”; the percent of discharges to SNFs that result in an ER visit within 30 days; out of network hospitalizations and specialty visits; and, in the case of Medicare, annual wellness visits. Actionable quality measures include gaps in care; 30- and 90-day readmission rates; total CMS Quality/Patient Safety Penalties in the most recent year; and the percent of hospitalizations and outpatient surgeries with follow-up within 7 days post discharge.
Whether potentially actionable opportunities to improve performance are, in fact, achievable depends on the strength of the provider's PHM capabilities mentioned earlier in this paper, such as administrative and clinical leadership, culture, data management, data analytics and reporting, EHR configuration and programming, care management, provider network development and management, and actuarial and contracting expertise. Each of these requirements for success can be evaluated in a qualitative, but structured fashion and monitored over time.
Conclusion
Although the pace at which health care payments are becoming more value and risk based varies markedly in different geographic areas, most stakeholders believe that the movement from FFS to more VBP/RBP is inevitable. Provider success under risk requires several capabilities that take time to put in place. Provider success in a VBP world also will depend on providers growing the number of patients whose care they manage. There are no perfect indicators of a provider's readiness to take on financial risk. Those who start with poorer cost and quality performance, but who have invested in implementing key PHM capabilities, are most likely to be successful in the short term. Longer term success will require proof that the provider can perform better than cost and quality benchmarks.
Footnotes
Author Disclosure Statement
At the time this manuscript was written, Dr. Steinberg was CEO of xG Health Solutions and received a salary from and had stock in xG Health Solutions. xG Health Solutions is a for-profit subsidiary of the Geisinger Health System, which helps health care delivery systems improve their clinical and financial performance. This manuscript was not supported by funding from any entity other than xG Health, which paid Dr. Steinberg's salary. The views presented in this manuscript are based on Dr. Steinberg's personal experience.