The Medicare IPPS 40 Years Later

Lessons Learned and What to Do Next

Abstract

The 1983 implementation of the Medicare Inpatient Prospective Payment System (IPPS) was successful in controlling Medicare inpatient costs because it was designed as a clinically credible management tool that facilitated real behavior change and performance improvement. The next phase of IPPS should expand the inpatient payment bundle to a hospital episode-of-care performance bundle that explicitly links episode cost and quality. A uniform, comparable, and transparent episode performance bundle that highlights the tradeoffs between episode cost and quality can expand the incentives to control costs and provide hospitals the management information to improve performance.

FACING the imminent insolvency of the Medicare Part A hospital insurance trust fund (^{Board of Trustees of the United States Federal Hospital Trust Fund, 1982}), Congress passed legislation in 1983 that implemented the Medicare Inpatient Prospective Payment System (IPPS). The IPPS was designed to provide hospitals with the financial incentive to control costs. There were no new taxes and no reduction in payment levels to hospitals. Implemented as a budget neutral system, IPPS reformed how hospitals were paid. As a direct result of IPPS, Medicare annual expenditures for hospital care in 1990 were $18 billion lower ($37 billion in today's dollars) than originally projected at the time of IPPS implementation, representing a 20% decrease in projected annual hospital Medicare expenditures (^{Russell & Manning, 1989}). These expenditure reductions were achieved without a negative impact on patient outcomes (^{Kahn et al., 1990}).

As the 40-year anniversary of IPPS approaches, the Medicare Part A hospital insurance trust fund is once again approaching insolvency (^{Congressional Budget Office, 2020}). What are the lessons learned from the implementation of IPPS and can those lessons be used to expand the scope of IPPS? This article has 3 main objectives: to review the intent and assumptions of the original IPPS research; to analyze the reasons for the success of IPPS; and to propose the next phase in the evolution of IPPS so that it can provide hospitals with expanded incentives to control costs and improve performance while helping to safeguard the Medicare trust fund.

FOUNDATIONAL RESEARCH

The IPPS financial incentive was created by paying hospitals an all-inclusive fixed price for each type of inpatient admission. Although this fixed-price payment bundle established the underlying financial incentive for efficiency, the success of IPPS was dependent on hospitals making substantive and sustainable behavior changes in both the administrative and clinical management of hospitals. Because hospital resource requirements vary on the basis of the types of patients being treated, the IPPS payment bundle had to explicitly reflect the patient case mix of a hospital. During the 1970s, core research on the classification of hospital inpatients was done at Yale University and published in a series of 3 foundational articles (^{Fetter et al., 1980; Mills et al., 1976; Thompson et al., 1979}). The Yale research defined the types of patients treated by hospitals based on a new patient classification methodology: the Diagnosis-Related Groups (DRGs) (^{Fetter et al., 1980}). The DRGs were explicitly designed to document the relationship between medical and administrative decisions and to motivate physicians to use hospital resources efficiently. The DRG methodology was developed as a management tool that allowed industrial methods of cost and quality control to be adapted and applied to the hospital industry. The objective was to create the basis for “control of the production process” (^{Mills et al., 1976}). The decision to make DRGs the unit of payment in IPPS and the ability of hospitals to use the DRGs as an operational management tool turned out to be crucial to the ultimate success of IPPS.

For DRGs to be an effective management tool, it was necessary to develop a DRG-based “case mix cost accounting system” that could provide hospitals with an “integrated picture of the financial consequences of providing care to individual patients” (^{Thompson et al., 1979}). The case mix cost accounting system was essential to IPPS implementation because it provided a foundation for a new financial approach to hospital management and the basic building blocks for a DRG-based payment system. However, when Yale researchers presented data on the resource use and cost of treating patients in each DRG to hospital administrators, it was met with polite disinterest. In retrospect, the reason for this reaction is clear. At the time, hospitals operated under a payment system that was cost based, so improved efficiency meant less revenue. Unfortunately, without a payment system that rewarded efficiency, hospitals were not motivated to use DRGs as a management tool for performance improvement.

Under the existing cost-based payment system there were no performance standards, which resulted in wide variation in hospital costs for treating similar patients. For example, researchers discovered a 6-fold variation in the average amount Medicare was paying to individual hospitals for the treatment of an acute myocardial infarction (heart attack) with no plausible justification for that level of variation (^{Schweiker, 1982}). Such variation meant that opportunities for improvement existed, thereby creating the potential for cost savings. That led to the realization that DRG costs could be used to establish reimbursement standards for each DRG that were balanced to “reflect the interests of both the public and the [hospital] industry” (^{Fetter et al., 1980}). Basing payment on the national cost per DRG standardized payment across hospitals and essentially eliminated variation in payments that had no plausible justification.

The 1982 US Department of Health & Human Services (HHS) Report to Congress proposing IPPS succinctly summarized the objective of a DRG-based payment system as setting “a reasonable price for a known product” (^{Schweiker, 1982}). The combined effect of the IPPS financial incentive for efficiency created by a prospective price and the linkage of the clinical and financial aspects of care inherent in a DRG unit of payment revolutionized how hospitals were managed. In a 2001 review of the key reasons for the success of IPPS, the Centers for Medicare & Medicaid Services (CMS) concluded:

The success of any payment system that is predicated on providing incentives for cost control is almost totally dependent on the effectiveness with which the incentives are communicated. (Health Care Financing Administration [HCFA], U.S. Department of Health & Human Services, 2001)

The DRGs created a clinically credible language that enabled the effective communication of cost containment incentives across the entire hospital. The language value of DRGs was essential in engaging physicians, who control the majority of hospital resources and, therefore, are critical to any effort at cost control. The IPPS set clinically credible performance standards (the DRG prices) by type of patient (the DRG) with variation from the DRG price resulting in a profit or loss that was directly proportional to a hospital's performance relative to the DRG price. This easily understood “product with a price” design, communicated in a clinically credible manner (the DRGs), was the foundation for the success of IPPS.

The crucial lesson from IPPS implementation is that an incentive-based payment system can be effective only if it is designed as a clinically credible management tool that facilitates real behavior change and performance improvement. The corollary to this lesson is that any incentive-based regulations must be designed to provide clinically credible information that facilitates an effective response to the incentives.

IPPS IMPLEMENTATION

During the late 1970s and early 1980s, the HHS funded the implementation of a series of state-level hospital Medicare payment reform projects (^{Schweiker, 1982}). Based on the extensive research done at Yale, the State of New Jersey implemented an all-payer DRG-based prospective payment system for hospitals. Although the basic concept of prospective DRG payment rates is straightforward, there are numerous individual design components that can significantly affect the incentives and effectiveness of the system (^{Averill & Kalison, 1991}). New Jersey had to address the full array of implementation details such as outlier policies and adjustments for hospital teaching status and hospitals with a disproportionate share of low-income patients. The lessons learned from the New Jersey implementation were crucial to the design of IPPS and its ultimate success.

The New Jersey DRG implementation provided HHS a regulatory framework and operational experience. Thus, the decision by HHS to propose IPPS was not based on an untested theory. The New Jersey implementation allowed the legislation implementing IPPS to be very detailed in terms of the key design issues of a DRG-based IPPS. The detailed nature of the IPPS legislation allowed CMS to focus on implementation issues without the need to research design alternatives. As a result, IPPS was operational in October 1983, only 5 months after the passage of the IPPS legislation in April of 1983. The IPPS simultaneously provided a means of equitably distributing Medicare payments across hospitals and creating the incentive to reduce the growth of Medicare hospital payments. So, IPPS was effective in both distributing and controlling Medicare hospital payments. The initial IPPS savings have been permanently incorporated into the underlying cost of hospital care. After IPPS implementation, Medicare hospital payments never returned to the double-digit inflation rates experienced prior to IPPS.

The basic structure of IPPS has remained largely unchanged for nearly 40 years thereby allowing hospitals to manage and plan within a known and stable payment environment. Most hospital administrators today have only known hospital management under a DRG-based pricing system. The stability of IPPS has been both positive and negative. The rapid success of IPPS in controlling Medicare hospital expenditures had the unintended consequence of removing the urgency and political will to expand the scope of the IPPS payment bundle and to adopt the basic IPPS payment system design across more types of providers and more aspects of provider performance.

Applying the core IPPS approach to other aspects of provider performance requires that clinically credible performance standards be set for each type of patient with payment levels adjusted in direct proportion to the variation from the performance standard. To effectively expand the IPPS design and structure more broadly across today's health care delivery system, one must understand the details of why IPPS was so effective.

ROLE OF RISK ADJUSTMENT

The severity of illness and complexity of a provider organization's patient population (case mix) impact its resource requirements and cost of operations. The method of accounting for the impact of case mix is referred to as risk adjustment. Any payment system in which providers assume financial risk for the care and treatment of a patient population must be risk adjusted. Failure to have adequate risk adjustment can result in unfair levels of payment and potentially result in incentives for providers to avoid treating complex and costly patients. In IPPS, DRGs are the method of risk-adjusting hospital payments. Payment by the DRG is inherently risk adjusted because each DRG payment amount reflects the complexity and costliness of the patients assigned to the DRG.

The categorical nature of DRGs permits a separation of the computation of the relative payment weights (prices) and the definition of the DRG categories (products). Such a separation is an inherent by-product of the categorical nature of DRGs. In noncategorical systems, such as those based on linear or logistic regression, the clinical model and payment weights (coefficients) are interdependent and there is no separate definition of the clinical model. As noted by the CMS, developing the clinical model (DRGs) separate from the payment weights was a critical factor in the success of IPPS:

The separation of the clinical and payment weight methodologies allows stable clinical methodology to be maintained while the payment weights evolve in response to changing practice patterns. (^{Health Care Financing Administration [HCFA], U.S. Department of Health & Human Services, 2001})

The DRG clinical model has remained relatively stable, thereby creating a consistent and powerful communications tool. However, the payment weights have fluctuated to reflect changing practice patterns and new technology.

The method of risk adjustment is crucial to the success of any incentive-based payment system because it constitutes the basic building blocks for communicating the system's incentives. Since the IPPS risk categories (the DRGs) are clinically credible, the IPPS incentives are expressed and communicated in a clinically credible and understandable manner. Risk adjustment has 2 primary purposes: to predict the impact of a provider's case mix and to effectively communicate the incentives in the payment system. An effective risk adjustment system must both predict and communicate.

THE IPPS CORE DESIGN PRINCIPLE

The IPPS sets a clinically credible performance standard (the DRG price) for each type of patient (the DRG) with variation from the DRG price resulting in a profit or loss that is directly proportional to a hospital's performance relative to the DRG price. This underlying IPPS design principle requires that a clinically credible performance standard be set for each type of patient with performance determined on the basis of variation from the standard. The IPPS incentives impacted the entire hospital, not just a few types of patients (eg, only cardiovascular patients) or isolated examples (eg, patients with only infectious disease complications). Substantive and sustainable behavior change requires changes to organizational culture and should be made organization-wide (^{Walker & Soule, 2017}).

The IPPS core design principle can also be applied to other aspects of performance such as quality of care or delivery system effectiveness (eg, readmission rate). The evaluation of any measure of quality of care or delivery system effectiveness must be limited to those patients whose clinical circumstances indicate that there is reasonable likelihood that the quality problem or delivery system failure could have been prevented (eg, readmission for a postoperative wound infection following orthopedic surgery). Hospitals must not be held accountable for patient outcomes that are clearly beyond their control and for which they have no realistic ability to prevent (eg, readmission for appendicitis following orthopedic surgery). If hospitals are held accountable for patient outcomes beyond their control, the quality improvement process breaks down and becomes ineffective.

The IPPS design principle is straightforward and represents the general structure of a well-designed payment system that can also serve as a management tool for hospitals and other providers (^{Averill et al., 2011}). The success of IPPS has clearly demonstrated that provider behavior will respond to incentives that adhere to the IPPS design principle. Although the IPPS savings have been permanently incorporated into the underlying cost of hospital care, the savings largely occurred during the initial decade following the 1983 implementation; IPPS has not produced substantial new savings in subsequent decades. To achieve new savings, the IPPS incentives must be expanded to incorporate more services and more aspects of performance.

EXISTING IPPS REFORMS

Over the past decade, the CMS has launched 2 major IPPS reforms aimed at expanding the IPPS incentives. The first IPPS reform targets payment adjustments for quality and includes 3 quality-related programs that adjust IPPS payments based on hospital performance: Hospital-Acquired Condition Reduction Program (HACRP) (^{Centers for Medicare & Medicaid Services [CMS], 2022b}), Hospital Readmission Reduction Program (HRRP) (^{Centers for Medicare & Medicaid Services [CMS], 2022c}), and Hospital Value-Based Purchasing Program (HVBPP) (^{Centers for Medicare & Medicaid Services [CMS], 2022d}). The HACRP and the HRRP are limited to just a few clinical areas and are based on all-cause methodologies, so readmissions and complications over which a hospital has no reasonable control are included. Both HACRP and HRRP are penalty only and can have disproportionate payment adjustments. The HVBPP has largely failed to produce significant savings or improve quality of care (^{U.S. Government Accountability Office, 2015}) because it has been based on a constantly changing multitude of performance measures, which are combined into a payment adjustment that is complex and difficult for hospitals and patients to understand. The Medicare payment adjustments related to quality are applied independently and have been criticized by hospitals as being overlapping with inconsistent scoring methods (^{Kahn et al., 2015}).

The second major CMS reform was intended to address fragmented care and minimal coordination across providers and health care settings. In the Bundled Payments for Care Improvement (BPCI) initiative, the CMS created 90-day episodes around a hospitalization (^{Centers for Medicare & Medicaid Services [CMS], 2022a}). This initial model of BPCI was a voluntary program covering a limited number of admissions. After the first BPCI model concluded in 2016, an analysis found that for admissions with a 90-day episode, the participating hospitals did not experience any reduction in overall readmission rates, emergency department (ED) visits, or mortality. Although there was some reduction in post–acute residential facility usage, Medicare experienced net losses under BPCI (^{Centers for Medicare & Medicaid Services [CMS], 2021}). The CMS subsequently implemented an Advanced BPCI program (BPCI-A) in 2018, now extended to 2025. The first year of BPCA-A resulted in only a small reduction in episode expenditures (^{Joynt Maddox et al., 2021}). Although Advanced BPCI does set a target episode expenditure level, there are no clear clinically credible targets for the aspects of performance (complications, readmissions, etc) that are the major drivers of episode expenditures.

Although HACRP, HRRP, HVBPP, and BPCI are logical extensions of the IPPS incentives, they have not dramatically changed hospital behavior. These programs are narrowly focused and do not apply to the entire hospital, so they have little impact on the overall performance improvement culture of the hospital. Their design does not adhere to the IPPS core design principle, nor do these programs create a clinically credible and uniform language of performance expectations. As a result, they are not an effective communications tool for hospitals and patients. They are independent programs that are not integrated together in any meaningful way. However, if HACRP, HRRP, HVBPP, and BPCI were to adhere to the IPPS design principle, a comprehensive integration of each program's general approach has the potential to be the basis for the next phase in the evolution of IPPS.

TRANSFORMING IPPS TO AN EPISODE PERFORMANCE BUNDLE

A hospital episode-of-care payment bundle would expand the scope of the IPPS payment bundle to include care decisions and services beyond the inpatient hospital stay. Although a hospital episode prospective price creates a strong incentive for episode cost efficiency, it must be combined with performance information on the core drivers of episode cost. Cost, quality of care, and delivery system effectiveness are inextricably connected. An excess volume of hospital-related services is often the result of quality-of-care problems and delivery system ineffectiveness. Hospital episode cost effectiveness and quality performance need to be combined into an episode performance bundle that is clinically credible and meaningful as a management tool for hospitals. An episode performance bundle can provide a holistic view of the care provided during a hospital episode of care.

A hospital episode-based system integrated with measures of quality performance requires a categorical risk adjustment model such as DRGs that allows separate standards for each measure of performance to be determined on the basis of the same underlying clinical model. The current DRGs used in IPPS are focused on efficiency and do not meet this requirement. However, there are widely used variations of DRGs available that can risk adjust both episode cost and measures of quality of care and delivery system effectiveness (^{Averill et al., 2002; 3M Clinical and Economic Research, 2023}).

The evolution of IPPS to an episode performance bundle needs to meet the following requirements:

• The hospital episode-of-care performance bundle should encompass care decisions and services prior to and after the inpatient hospital stay.
• The hospital episode-of-care performance bundle should combine hospital episode cost effectiveness and quality-of-care performance into a clinically credible and uniform unit of payment and performance comparison.
• The determination of cost and quality-of-care performance must adhere to the IPPS core design principle and be based on a comparison to national standards.
• The measures of quality performance and delivery system effectiveness in the episode performance bundle should reflect patient expectations for hospital-related care.
• The IPPS DRG unit of payment should be transitioned to an episode performance bundle unit of payment.
• The episode performance bundle unit of payment should apply to the majority of hospital admissions.

By meeting these requirements an episode performance bundle unit of payment can create a clinically credible language of hospital performance expectations. An integrated episode performance bundle explicitly links episode cost and quality. It highlights the tradeoffs between episode cost and quality and identifies performance improvement opportunities for hospital management.

The DRGs would not have changed hospital management without DRGs being the IPPS unit of payment. The reality is that for an episode performance bundle to have a substantial and sustainable impact on hospital management and cost, IPPS must be transformed to an episode performance bundle unit of payment.

SELECTING THE MEASURES OF QUALITY OF CARE AND DELIVERY SYSTEM EFFECTIVENESS

An episode performance bundle is not only a management tool for hospitals but should also increase patient awareness of relative hospital performance and create a more transparent, patient-engaged system. The performance measures in a hospital episode performance bundle should be relevant to patients. Patients tend to think in terms of an entire episode of care related to a specific health care event such as a hospitalization. For inpatient care, patients have some basic expectations:

• They should not experience any unnecessary hospital admissions.
• They should not suffer any avoidable complications or mortality.
• After hospital discharge, they should not have any unplanned readmissions or ED visits.
• They should not experience any unnecessary admissions to a residential post–acute care facility.

These 6 measures of unnecessary or preventable quality or delivery system failures are major drivers of episode cost and can serve as an initial set of performance measures in a hospital episode performance bundle. Patients also have expectations relative to clinical outcomes such as slowing disease progression or improved function, but such measures are highly disease specific, difficult to measure, and usually require extensive additional data collection.

Holding hospitals accountable for overall admission volume is difficult because unnecessary hospital admissions are often related to care in the community that cannot be directly controlled by the hospital. However, indirect measures such as excess admissions through the ED for nonsurgical low severity patients (^{Averill et al, 2019}) and an underuse of outpatient surgery (^{Averill et al., 2021}) are 2 performance measures that directly impact the overall rate of admissions to a hospital and are under a hospital's control. Including these 2 measures in the episode performance bundle could, in part, address hospital performance related to excess hospital admissions.

EXAMPLE OF AN EPISODE PERFORMANCE BUNDLE

To illustrate the concept of an episode performance bundle, consider an episode encompassing 3 days prior to admission and 30 days postdischarge for a patient with moderate underlying disease who has a total hip replacement. There would be a prospective price for the hip replacement episode along with the expected rate of avoidable quality and delivery system failures, such as potentially preventable inpatient complications (^{Hughes et al., 2006}), potentially preventable readmissions (^{Goldfield et al., 2008}), return ED visits (^{Goldfield et al., 2008}), 30-day postprocedure mortality (^{Averill et al., 2020}), and 4-day postdischarge admission to a postacute residential facility (^{Averill et al., 2021}). Using 2018 national IPPS data, the expected rate per 100 for a patient with moderate underlying disease who is having a total hip replacement is 7.73 for inpatient complications, 6.70 for readmissions, 5.02 for return ED visits, 3.83 for 30-day postprocedure mortality, and 44.73 for 4-day postdischarge admission to a postacute residential facility. These rates per 100 include only those patients whose clinical circumstances make the quality or delivery system failure potential preventable. Risk-adjusted variation from these expected rates provides hospitals with basic management information on the key drivers of episode cost. These rates per 100 are for patients with moderate underlying disease and would differ for patients with a greater or lesser burden of illness. For example, the national Medicare potentially preventable readmission rate for patients having a total hip replacement with a minor, moderate, major, and extreme burden of illness is 3.44, 6.70, 12.99, and 16.33, respectively.

TRANSITIONING TO AN EPISODE PERFORMANCE BUNDLE PAYMENT SYSTEM

The 1983 transition from a cost-based payment system to the fixed prospective price in IPPS created substantial financial risk for hospitals. Although not as dramatic as the IPPS transition, moving to an episode performance bundle payment system will also increase hospital financial risk. Like IPPS, where the financial risk was phased in over a 5-year period, the financial risk in an episode payment system should also be phased in. This could be accomplished in the initial years of an episode payment system by adjusting the standard IPPS inpatient payment amount based on hospital performance in the key drivers of episode cost related to unnecessary or preventable quality and delivery system failures (^{Averill & Mills, 2022}). By adjusting IPPS payments based on performance in the key drivers of episode cost, incentives that are equivalent to a hospital episode performance bundle would be created without requiring hospitals to be at risk for the full episode cost. In subsequent years, the hospital financial risk would be transitioned from the IPPS unit of payment to a full episode unit of payment by setting an all-inclusive prospective payment amount for each episode (^{Vertrees et al., 2013}). Throughout the transition, the payment adjustments related to unnecessary or preventable quality and delivery system failure would be applied.

ADOPTION BY OTHER PAYERS

Much of payment reform over the past decade has focused on transitioning to capitated payment models. It is not uncommon for managed care plans to have subcapitation arrangements with providers such as primary care practices and multispecialty physician group practices. A managed care plan that pays hospitals using an episode performance bundle unit of payment would essentially be entering into a subcapitation arrangement with a hospital because an episode performance bundle unit of payment makes the hospital financially responsible for the major drivers of hospital episode-related costs over an extended period of time. There can be substantial costs associated with care management during the post–acute care period, so an episode performance bundle unit of payment can have a disproportionate financial impact on managed care plans. For an episode performance bundle to be a practical tool for managed care plans, Medicare must implement an episode performance bundle payment system with all the operational details fully functional. Only then will managed care plans consider adoption of an episode performance bundle as a practical option.

In a multipayer system, the success of any incentive-based payment system will be influenced by the extent to which payment systems across major payers are complementary and whether the systems reinforce the incentive for performance improvement. Use of episode performance bundles within Federal payment programs should be broad based. Medicare should include the episode performance bundle in evaluating the performance of Medicare Advantage plans and require state Medicaid plans to use payment systems that include performance evaluations based on an episode performance bundle. By expanding IPPS to an episode performance bundle, Medicare can influence the direction of payment policy for all payers. Indeed, the IPPS DRG unit of payment was initially applicable only to Medicare fee-for-service but is now widely used across major payers and in many foreign countries.

DISCUSSION

Although the proposed episode performance bundle payment system focuses on inpatient care, payment systems for other care settings such as outpatient surgery and other substantive outpatient services could also be transitioned to episode-based payment. The underlying approach of bundling services around an episode of care and setting clinically credible efficiency and quality performance standards (by type of patient with performance determined on the basis of variation from that standard) readily transfers to other care settings.

CONCLUSION

The 1983 implementation of the Medicare IPPS was successful in controlling Medicare inpatient costs. After 40 years, the crucial lesson from IPPS is that to be effective, an incentive-based payment system must be designed as a clinically credible management tool that facilitates real behavior change and performance improvement. The underlying IPPS core design principle requires that a clinically credible performance standard be established for each type of patient with performance determined on the basis of variation from this standard. To achieve new savings, IPPS incentives must be expanded to incorporate more services and more aspects of performance, including care decisions and services beyond the inpatient hospital stay. This requires that the IPPS DRG unit of payment be transitioned to a hospital episode-of-care unit of payment.

Although a hospital episode prospective price would create a strong incentive for episode cost efficiency, it needs to be combined with performance information on the core drivers of episode cost. The objective is to create an episode performance bundle that combines hospital episode cost effectiveness with quality and delivery system performance into a unit of payment that is clinically credible and meaningful as a management tool for hospitals. An integrated episode performance bundle explicitly links episode cost and quality and highlights the tradeoffs between episode cost and quality. The performance measures in a hospital episode performance bundle should be relevant to patient expectations for inpatient care, creating the opportunity for greater patient engagement. A payment system based on a uniform, comparable, and transparent episode performance bundle, one that encompasses performance in terms of both efficiency and quality, is a payment policy that can be readily transferred to other care settings.