Inpatient vs Outpatient Hospitalization

Status determinations (outpatient versus inpatient) for hospitalized patients have become a routine part of patient care in the United States. Under the guidance provided by the Medicare Benefits Policy Manual, hospitalized Medicare beneficiaries are assigned 1 of these 2 statuses. The status assignment does not affect the care a patient can receive, but rather how the hospital services provided are billed to Medicare. Hospital services provided under inpatient status are billed under Medicare Part A. Hospital services provided under outpatient status, which includes all patients receiving observation services (commonly referred to as under observation), are billed under Medicare Part B. Whether hospital services are billed under Part A or Part B is important to hospitals and Medicare beneficiaries, as both the hospital reimbursement and beneficiary liability can vary greatly depending on whether services are billed under Part A versus Part B. Hospitals are generally reimbursed at a higher rate for services provided as an inpatient (Part A). The Office of the Inspector General (OIG) recently found that Medicare paid nearly three times more for a short inpatient stay than an [outpatient] stay for the same condition.[1] Medicare beneficiary liability also varies based on status. First, beneficiaries hospitalized as inpatients are subject to a deductible under Part A ($1,216 in 2014) for hospital services associated with that hospitalization and any future inpatient hospitalization beyond 60 days of discharge.[2] Beneficiaries hospitalized as outpatients are subject to the Medicare Part B deductible ($147 in 2014), and then a 20% copay on each individual outpatient hospital service, with no cumulative limit.[2, 3] In addition, hospital pharmacy charges for Medicare beneficiaries hospitalized as inpatients are covered under Medicare A. However, for Medicare patients hospitalized as outpatients, many medications are not covered by Medicare Part B benefits. Finally, time spent hospitalized as an outpatient does not count toward the Medicare 3‐day medically necessary inpatient stay requirement to qualify for the skilled nursing facility care benefit following discharge.

HISTORY AND INTENT OF INPATIENT AND OUTPATIENT STATUS DETERMINATIONS

Prior to October 1, 2013, the Centers for Medicare & Medicaid Services (CMS) stated that physician judgment and an expectation of at least an overnight hospitalization should determine inpatient status of hospitalized Medicare beneficiaries. Guidance as to when inpatient services were covered was found in the Medicare Benefits Policy Manual (MBPM)[4]:

For a subset of patients who are hospitalized under outpatient status, billing for observation services is allowed. CMS defines observation as a well defined set of services, that should last less than 24 hours and in only rare and exceptional casesspan more than 48 hours.[5] Many providers recognize the utility of a few additional hours of hospital care and/or testing in a hospital setting to determine whether a patient can go home or needs additional evaluation, monitoring, and/or treatment that can only be provided in a hospital, consistent with the CMS definition of observation.[6] It is important to note that although observation and outpatient are frequently used interchangeably, only outpatient is technically a CMS status. Patients in observation or under observation are, in fact, a subset of patients who are hospitalized under an outpatient status.

Outpatient status may also be appropriate for patients who require hospitalization for routine and expected overnight monitoring following a procedure. These patients are often not eligible for billing of observation services or as an inpatient because alternative methods of billing for the recovery time following the procedure exist. When determining the appropriate status of a Medicare beneficiary for a hospitalization following a procedure, physicians need to be aware of whether a specific procedure appears on the Medicare inpatient‐only procedures list.[7] Per CMS, procedures designated as inpatient only are reimbursed only when the patient is admitted as an inpatient at the time the procedure is performed.[8] Conversely, outpatient status for an overnight hospitalization associated with a procedure not on the inpatient‐only list is generally appropriate. Therefore, patients hospitalized for a procedure that appears on this list should always be hospitalized under inpatient status, regardless of the amount of time that the patient is expected to be hospitalized following the procedure, including those cases for which the hospitalization is expected to be only overnight.[7, 8] Only a limited number of Current Procedural Technology (CPT) codes, mostly surgical, automatically qualify for inpatient status and do not have outpatient prospective payment system eligibility. Although most procedures on the inpatient‐only list are associated with a hospitalization that commonly span at least 2 midnights, such as coronary artery bypass grafting, some potentially overnight stay cases, such as cholecystectomy (CPT 47600) appear on the 2014 inpatient‐only list.[9]

As noted above, prior to October 1, 2013, the Medicare definitions governing outpatient versus inpatient status included a 24‐hour benchmark. However, the MBPM also notes that: Admissions of particular patients are not covered or non‐covered solely on the basis of the length of time the patient actually spends in the hospital.[10]

In practice, status determination was ultimately dependent on physician or other practitioner's complex medical judgment as specified by CMS. To validate this judgment, CMS recommended that reviewers use a screening tool as part of their medical review. This screening tool could include practice guidelines that are well accepted by the medical community but did not require or identify a specific criteria set.[11] Not surprisingly, there was and continues to be great variability in the application of outpatient versus inpatient status across hospitals in actual practice.[1, 12, 13] The ambiguity in the definition of a hospitalized patient's status helped spawn commercial clinical decision tools, such as InterQual (McKesson Corporation, San Francisco, CA) and MCG (formally known as Milliman Care Guidelines; MCG Health, LLC, Seattle, WA), to help define inpatients versus outpatients.[14, 15] However, these guidelines are complex, can be difficult to interpret and apply, and have been criticized for poor predictive value and attempting to replace physician judgment.[16, 17, 18] Furthermore, CMS has never formally endorsed any specific decision tool.

INPATIENT AND OUTPATIENT PAYMENTS AND THE RECOVERY AUDIT CONTRACTOR PROGRAM

In 2000, CMS started using Ambulatory Payment Classifications for hospital services, which made inpatient care more financially favorable for hospitals. In response to concerns that hospitals would be incentivized to overuse inpatient status, CMS made a number of changes to their payment system, including the creation of the Recovery Audit Program in 2003. This program was originally called the Recovery Audit Contractor (RAC) Program and continues to be most commonly referred to as the RAC program. The RAC program, tasked with finding and correcting improper claims to the Medicare program, began as a demonstration required in the Medicare Prescription Drug Improvement and Modernization Act of 2003 (MMA), and subsequently became a nationwide audit program under the Tax Relief and Health Care Act of 2006. Under this program, private contractors review hospital and billing records of Medicare patients and are paid on a contingency fee (8%12.5%) for all underpayments and overpayments that are identified and corrected.[19] Importantly, the RACs are not subject to any financial penalties for cases improperly denied.

RACs initially targeted many overnight inpatient stays for recoupment. These cases were attractive audit targets because the RACs could argue that the inpatient hospital services were delivered in the improper status based solely on the length of stay, without having to consider in their audit the complexity of decision making or medical necessity of the services provided. However, it is worth noting that with improvement in efficiency and advancements in medical technology, hospitals and physicians have been increasingly able to safely evaluate and treat medically complex and severely ill patients quickly, sometimes with just an overnight stay. As perspective, in 1965, the average length of stay for a Medicare patient was 13 days; in 2010, it was 5.4 days, with over one‐third of hospitalizations lasting <3 days.[20]

Concurrent with the increased RAC denials for services provided in an inpatient status, the use of observation services changed significantly from 2007 to 2012. The average length of stay for Medicare patients under outpatient status with observation services exceeded 24 hours in 2007, was 28.2 hours by 2009,[21] and grew to 29 hours by 2012.[22] Between July 2010 and December 2011, at the University of Wisconsin Hospital, 1 in 6 observation stays lasted longer than 48 hours, suggesting that long observation stays were no longer rare and exceptional as stated in CMS' own definition.[23] This same University of Wisconsin study also found that observation services were not well defined, with 1141 distinct diagnosis codes used for these services.[23]

Additionally, a Medicare Payment Advisory Commission (MedPAC; described on their website, www.medpac.gov, as a nonpartisan legislative branch agency that provides the US Congress with analysis and policy advice on the Medicare program) 2014 Report to Congress showed that from 2006 to 2012, outpatient services increased 28.5% whereas inpatient discharges decreased 12.6% over the same time period.[22]

Hospitals have also expressed concern that the RAC contingency fee payment model and a lack of penalty for improper denials promotes overzealous auditing.[24, 25] RAC recoupment has increased from approximately $939 million in 2011, to $2.4 billion in 2012, to $3.8 billion in 2013.[26, 27, 28] Given the money now at stake, it is not surprising that hospitals have become very active in appealing RAC denials. Self‐reported data submitted to the American Hospital Association (AHA) for the months January 2014 to March of 2014 show that hospitals now appeal 50% of RAC denials and win 66% of these appeals.[29] The AHA data also show that 69% of self‐reporting hospitals spent over $10,000 to manage their audit and appeals process over this same 3‐month time period, with 11% spending more than $100,000.

This appeals process is not only costly to hospitals, it is also lengthy. As of January 2014, the average wait time for an appeal hearing with an administrative law judge (level 3 appeal) exceeded 16 months.[30] In fact, the appeals process has become so backlogged that hospitals' rights to assignment of level 3 (administrative law judge) appeals have been temporarily suspended.[30] In August 2014, CMS offered a $0.68 on the dollar partial payment for hospitals willing to settle all eligible outstanding appeals in an attempt to relieve the appeals backlog.[31] In addition, the AHA currently has a suit against the US Department of Health & Human Services over the RAC appeals backlog.[32]

Increased use of outpatient status may be driven by pressures from the RAC program and, potentially, by improvements in the efficiency of care. Because hospitals are paid less for care provided under outpatient status than they are for the identical care provided under inpatient status, hospitals faced both potential financial penalty for improvements in efficiency and the threat of RAC audits.

THE 2‐MIDNIGHT RULE: A FIX?

Given the challenges in defining inpatient versus outpatient hospitalization, the increasing use of outpatient status and the increasing length of stay of outpatient hospitalizations with observation services, in 2013, CMS responded with new policies to define the visit status for hospitalized patients. On August 2, 2013, CMS announced the fiscal year 2014 hospital Inpatient Prospective Payment System final rule (IPPS‐2014) to become effective October 1, 2013. This document was formally issued as part of the Federal Register on August 19, 2013.[33] Central to the CMS IPPS‐2014 was a 2‐midnight benchmark that offered a major change in how physicians were to determine the status (inpatient vs outpatient) of hospitalized patients. With this 2‐midnight benchmark, now informally known as the 2‐midnight rule, CMS finalized its proposal to generally consider patients that are expected by a practitioner (with knowledge of the case and with admitting privileges) to need hospitalization that will span 2 or more midnights as inpatient. The IPPS‐2014 also finalized the converse of this: hospitalizations expected to span <2 midnights are to be regarded as outpatient with 2 exceptions:

1. If the hospitalization is associated with a procedure appearing on the previously described Medicare inpatient‐only procedures list, or
2. A rare and unusual circumstance in which an inpatient admission would be reasonable regardless of length of stay. Currently, unanticipated mechanical ventilation initiated during the hospitalization visit is the only rare and unusual circumstance that qualifies as such an exception.[7]

CMS' stated goals and expectations for the 2‐midnight benchmark were:

1. Reduce the growing number of prolonged hospitalizations (>48 hours) for Medicare beneficiaries under outpatient status.
2. Decrease billing disputes between hospitals and Medicare auditors, especially RACs, by establishing more clearly defined, time‐based status criteria.
3. Reduce the number of outpatient encounters overall. Because CMS expected the rule to convert a net increase of cases from outpatient to inpatient, resulting in higher payments to hospitals, CMS included a 0.2% payment cut in hospital reimbursement in the IPPS‐2014 as an offset.[33, 34]

Although unrelated to the goals and expectations above, the IPPS‐2014 also included a requirement that:

CMS allowed for an authentication (generally regarded as a cosignature that is timed and dated) of the inpatient admission order by an attending physician with admitting privileges, done prior to discharge, in cases where the inpatient order had been placed by a practitioner (such as a resident, fellow, or physician assistant) without admitting privileges. Attending physician authentication of the inpatient admission order must be done prior to discharge [a]s a condition of payment for hospital inpatient services under Medicare Part A.[35]

From the August 2, 2013 announcement until the effective date of October 1, 2013, hospitals had just 2 months to interpret and comply with the IPPS‐2014, a complex 546‐page document that required hospitals to make extensive changes to admission procedures, workflows, and electronic health records (EHRs). In addition, extensive physician, provider, and administrator education was needed. During these 2 months, hospitals continued to request additional information and clarification from CMS regarding many aspects of the IPPS‐2014, including basic questions that included (1) how to apply the 2‐midnight benchmark to patients who were transferred from 1 hospital to another and (2) when the clock started for hospital services in determining a patient's expected length of hospitalization.

Despite concerns voiced by Congress and medical organizations, the new policy went into effect as scheduled.[36, 37] However, just days prior to October 1, 2013, CMS issued a 3‐month limited suspension of auditing and enforcement of the 2‐midnight rule by the RACs that was subsequently extended by CMS 2 more times, first through March 31, 2014 and then again through September 30, 2014. Other audits to be performed by RACs and all other government audits, including those performed by Medicare Administrative Contractors (MACs) were allowed to continue.[38] In particular, the MACs were instructed to conduct patient status reviews using a probe and educate strategy, which, via educational outreach efforts, would instruct hospitals how to adapt to the new rule. On April 1, 2014, the Protecting Access to Medicare Act of 2014 was signed into law, which, under section 111 of this law, permitted CMS to continue medical review activities under the MAC probe and educate process through March of 2015, and prohibited CMS from allowing RACs to conduct inpatient hospital status reviews on claims with these same dates of admission, October 1, 2013 through March 31, 2015.

The MACs were created by the MMA of 2003, which mandated that the Secretary of Health & Human Services replace Part A Fiscal Intermediaries and Part B carriers with Medicare Administrative Contractors (MACs).[39] As established by CMS, MACs are multi‐state, regional contractors responsible for administering both Medicare Part A and Medicare Part B claims and serve as the primary operational contact between the Medicare Fee‐For‐Service program, and approximately 1.5 million health care providers enrolled in the program.[39]

THE IPPS‐2014 AND CMS' STATED GOALS AND EXPECTATIONS

In the analysis that accompanied the IPPS‐2014, Medicare expected the use of outpatient services to decrease overall, as the new rules would effectively eliminate almost all outpatient hospitalizations >48 hours. Although no official data are yet available from CMS, our early experience under the 2‐midnight rule has suggested that long observation stays have declined in frequency, a favorable outcome of the new policy. However, as designed, the new 2‐midnight IPPS rule most predominately affects 1‐day stays, or more accurately, 1‐midnight stays. This is because many hospitalizations that previously met inpatient criteria (as defined by commercially available products such as MCG or InterQual), but spanned <2 midnights would have been classified as inpatient prior to October 1, 2013. However, since October 1, 2013, these same hospitalizations are now classified as outpatient. An example of such a case is a patient who presents to an emergency department with symptoms of a transient ischemic attack and has a high ABCD (age 60 years, blood pressure 140/90 mm Hg at initial evaluation, clinical features, duration of symptoms, diabetes score).[40] Prior to the 2‐midnight rule, this patient, based on the severity of the signs and symptoms upon presentation, could have been appropriately hospitalized as an inpatient.

Now, under the current IPPS and the ability of many hospitals to efficiently evaluate and treat such patient in <2 midnights, the patient should be categorized as an outpatient, at least initially, despite the severity and high risk of his/her presentation. In fiscal year 2013, The Johns Hopkins Hospital had 1791, 1‐day inpatient stays for Medicare beneficiaries, representing 15.2% of all Medicare admissions. Similarly, in the 12 months just prior to the 2‐midnight rule (October 1, 2012 to September 30, 2013), 10.4% (1280) of all Medicare encounters at the University of Wisconsin were 1‐day inpatient stays under previous criteria. Because of implementation of the 2‐midnight rule in October 2013, Medicare outpatient hospitalization for 1‐day stays at The Johns Hopkins Hospital increased by 49%, from an average of 117 patients/month to 174 patients/month. Nationally, it is possible that a reduction in long observation stays could be offset by an increase in 1‐day‐stay outpatient hospitalization encounters.

A second key expectation and goal of IPPS‐2014 was, by shifting to a more concrete, time‐based definition of inpatient, to decrease the disagreement between hospitals and auditors regarding patient status (inpatient vs outpatient). As noted earlier, many disputes with auditors for hospitalizations prior to October 2013 did not involve the need or type of hospital services provided, but rather the status under which the care was provided. However, the new time‐based criterion hinges not on actual length of hospitalization, but the expected length of hospitalization as determined by a practitioner with admitting privileges and knowledge of the patient. Accurately and consistently predicting the length of hospitalization has proven to be challenging, even for the most experienced practitioners. Since October 2013, for patients hospitalized at The Johns Hopkins Hospital through its emergency department, the admitting physicians' expectation of whether a patient would require 1 versus 2 or more midnights of necessary hospitalization was correct only half of the time. Given past experience, the RACs may challenge the medical judgment that lead practitioners to expect a hospitalization of 2 or more midnights without having to challenge whether the care provided was medically necessary.

Further, the IPPS‐2014 has not been accompanied by any significant changes to the payment scheme for auditors. RACs continue to be paid a percentage of any monies they determine to have been improperly paid by CMS, but with no penalty for cases that are overturned on appeal. Historically, the vast majority of RAC recovery fees have been due to determination of overpayments by CMS.[41, 42] Despite the 2‐midnight rule, RACs will continue to have a financial incentive to allege overpayment. In the initial probe and educate audits by MACs under the new 2014‐IPPS, despite inpatient admission orders being authenticated and certified by an attending physician, claims are being denied because the documentation does not support an expectation for a 2‐midnight hospitalization. Namely, auditors are continuing to challenge not the medical necessity of the services that hospitals provide, but rather the status in which those services were provided. Thus far, the IPPS‐2014 does not appear to fully remedy the auditing conflict that existed prior to October 2013.

As noted above, the IPPS‐2014 also requires, as of October 1, 2013, as a condition of payment for hospital services under Part A, that the inpatient admission order must be either entered by a practitioner with admitting privileges or authenticated prior to discharge by an attending physician involved in the care of the patient in cases in which the inpatient admission order was entered by a practitioner without admitting privileges (eg, resident, physician assistant, or fellow).[43] The requirement of an attending physician's cosignature has involved major changes to physician workflow and the electronic heath record (EHR) framework at The Johns Hopkins and the University of Wisconsin Hospitals, and does not keep up with modern healthcare systems in which patients are admitted 24 hours a day by a variety of providers (eg, residents, nurse practitioners) who otherwise may write stand‐alone orders. These changes have proven to be time‐consuming, costly, and have not, to our knowledge, improved patient care or utilization of resources.

The new visit status rules have also led to confusion among clinicians. A recent large survey of hospitalists conducted by the Society of Hospital Medicine demonstrated that more than half of respondents disagreed that the 2‐midnight rule improved hospitalist workflow compared to prior observation policy.[44] In addition, only 40% of hospitalists reported confidence in how to apply the rule.[44] Thus, the intent to clarify visit status policy with the IPPS‐2014 has not translated to clear and useful rules for frontline clinicians.

FUTURE DIRECTIONS

After over a year under the 2‐midnight rule, although long observation stays may be reduced, it seems unlikely these new regulations will achieve 2 of CMS' stated goals: (1) decreasing the use of outpatient status for hospitalizations and (2) resolving status disputes between auditors and hospitals. In addition, attempts at compliance with the new rules and regulations have diverted large amounts of physician time and hospital resources away from patient care. There is a clear need to reform both the hospitalization status policy and the RAC programs that enforce these rules.

One path Congress and CMS could consider is to reform the current Medicare reimbursement paradigm for hospital services to eliminate the need to distinguish inpatient from outpatient status. For example, H.R. 1179Improving Access to Medicare Coverage Act of 2013,[45] of the 113th Congress, if reintroduced, would decouple the link between the qualification for skilled nursing facility benefits from visit status by allowing time spent hospitalized as an outpatient to count toward the 3‐day benchmark. The overarching goals of any visit status policy reform should be to: (1) simplify or eliminate the 2‐track status process for hospitalized patients, (2) stop or minimize the threat of audits based on status, and (3) maintain budget neutrality. Two additional options for consideration would be to: (1) create a low‐acuity modifier for use with patients anticipated to have short stays and low resource use and (2) preselect specific Diagnosis Related Groups based on historical data and create designations for those diagnoses of lesser intensity. Accountable care organizations contracts, a new model for healthcare payment, could potentially be structured to eliminate or simplify payment based on visit status for hospitalized patients. With bundled payments on the horizon and the possible phase‐out of fee‐for‐service reimbursement, the issue may become less paramount in the coming years. No solution will be perfect and must balance costs, ease of administration, and beneficiary protection.

There are reasons to be optimistic that change may soon be realized. CMS is currently considering significant hospitalization status policy reform. In the proposed IPPS‐2015, CMS asked for input on payment for short‐stay hospitalizations and, in the final IPPS‐2015 released August 4, 2014, CMS indicated its willingness to continue to work with stakeholders in revising these policies.[46] Additionally, CMS has responded to hospitals on 3 separate occasions by delaying RAC audits pertaining to the 2‐midnight rule. Further, the current MAC probe and educate audits focus on education with respect to 2‐midnight rule implementation rather than threatening hospitals with major financial penalties.[47] Congress has also been responsive in this area. In addition to the 3 delays announced by CMS, Congress passed legislation that mandated an additional delay to RAC audits that pertain to the 2‐midnight rule. Moreover, the Subcommittee on Health of the House Ways and Means Committee held hearings that included the 2‐midnight rule and RAC reform in May 2014, and the Senate Special Committee on Aging held hearings on the impact of visit status on Medicare beneficiaries in July 2014.[48, 49] Additionally, the House Ways and Means Health Subcommittee recently issued a draft bill to address Medicare hospital issues.[50] The OIG has also been responsive to hospital concerns regarding the current RAC program with a recent report recommending that CMS develop additional performance evaluation metrics to improve RAC performance and ensure that RACs are evaluated on all contract requirements.[51] Additionally, MedPAC has been considering several short‐stay payment reform options, modifying the need for a 3‐day inpatient hospitalization to qualify for postdischarge skilled nursing facility benefits and adjusting RAC contingency fees based on overturn rates.[52, 53] These actions by CMS, Congress, and the OIG, as well as the options under consideration by MedPAC, demonstrate a degree of regulatory and legislative responsiveness to hospital and provider concerns in the area of visit status determination.

The Medicare program is vital to tens of millions of disabled and elderly Americans. Fraud and abuse of the Medicare program should not be tolerated. Yet, the current system of assigning, monitoring, and auditing outpatient versus inpatient hospital care is in need of reform. It will be up to CMS and Congress to continue to work with hospitals and physicians to find an improved way to appropriately and fairly compensate hospitals for hospital services in a way that that does not depend on a poorly defined and contentious status of a patient. Such reform must include the RAC program. It is our hope that both CMS and Congress will prioritize status determination and payment reform so that Medicare beneficiaries, physicians, and hospitals all have a sustainable, fair, and transparent process.

Status determinations (outpatient versus inpatient) for hospitalized patients have become a routine part of patient care in the United States. Under the guidance provided by the Medicare Benefits Policy Manual, hospitalized Medicare beneficiaries are assigned 1 of these 2 statuses. The status assignment does not affect the care a patient can receive, but rather how the hospital services provided are billed to Medicare. Hospital services provided under inpatient status are billed under Medicare Part A. Hospital services provided under outpatient status, which includes all patients receiving observation services (commonly referred to as under observation), are billed under Medicare Part B. Whether hospital services are billed under Part A or Part B is important to hospitals and Medicare beneficiaries, as both the hospital reimbursement and beneficiary liability can vary greatly depending on whether services are billed under Part A versus Part B. Hospitals are generally reimbursed at a higher rate for services provided as an inpatient (Part A). The Office of the Inspector General (OIG) recently found that Medicare paid nearly three times more for a short inpatient stay than an [outpatient] stay for the same condition.[1] Medicare beneficiary liability also varies based on status. First, beneficiaries hospitalized as inpatients are subject to a deductible under Part A ($1,216 in 2014) for hospital services associated with that hospitalization and any future inpatient hospitalization beyond 60 days of discharge.[2] Beneficiaries hospitalized as outpatients are subject to the Medicare Part B deductible ($147 in 2014), and then a 20% copay on each individual outpatient hospital service, with no cumulative limit.[2, 3] In addition, hospital pharmacy charges for Medicare beneficiaries hospitalized as inpatients are covered under Medicare A. However, for Medicare patients hospitalized as outpatients, many medications are not covered by Medicare Part B benefits. Finally, time spent hospitalized as an outpatient does not count toward the Medicare 3‐day medically necessary inpatient stay requirement to qualify for the skilled nursing facility care benefit following discharge.

HISTORY AND INTENT OF INPATIENT AND OUTPATIENT STATUS DETERMINATIONS

Prior to October 1, 2013, the Centers for Medicare & Medicaid Services (CMS) stated that physician judgment and an expectation of at least an overnight hospitalization should determine inpatient status of hospitalized Medicare beneficiaries. Guidance as to when inpatient services were covered was found in the Medicare Benefits Policy Manual (MBPM)[4]:

For a subset of patients who are hospitalized under outpatient status, billing for observation services is allowed. CMS defines observation as a well defined set of services, that should last less than 24 hours and in only rare and exceptional casesspan more than 48 hours.[5] Many providers recognize the utility of a few additional hours of hospital care and/or testing in a hospital setting to determine whether a patient can go home or needs additional evaluation, monitoring, and/or treatment that can only be provided in a hospital, consistent with the CMS definition of observation.[6] It is important to note that although observation and outpatient are frequently used interchangeably, only outpatient is technically a CMS status. Patients in observation or under observation are, in fact, a subset of patients who are hospitalized under an outpatient status.

Outpatient status may also be appropriate for patients who require hospitalization for routine and expected overnight monitoring following a procedure. These patients are often not eligible for billing of observation services or as an inpatient because alternative methods of billing for the recovery time following the procedure exist. When determining the appropriate status of a Medicare beneficiary for a hospitalization following a procedure, physicians need to be aware of whether a specific procedure appears on the Medicare inpatient‐only procedures list.[7] Per CMS, procedures designated as inpatient only are reimbursed only when the patient is admitted as an inpatient at the time the procedure is performed.[8] Conversely, outpatient status for an overnight hospitalization associated with a procedure not on the inpatient‐only list is generally appropriate. Therefore, patients hospitalized for a procedure that appears on this list should always be hospitalized under inpatient status, regardless of the amount of time that the patient is expected to be hospitalized following the procedure, including those cases for which the hospitalization is expected to be only overnight.[7, 8] Only a limited number of Current Procedural Technology (CPT) codes, mostly surgical, automatically qualify for inpatient status and do not have outpatient prospective payment system eligibility. Although most procedures on the inpatient‐only list are associated with a hospitalization that commonly span at least 2 midnights, such as coronary artery bypass grafting, some potentially overnight stay cases, such as cholecystectomy (CPT 47600) appear on the 2014 inpatient‐only list.[9]

As noted above, prior to October 1, 2013, the Medicare definitions governing outpatient versus inpatient status included a 24‐hour benchmark. However, the MBPM also notes that: Admissions of particular patients are not covered or non‐covered solely on the basis of the length of time the patient actually spends in the hospital.[10]

In practice, status determination was ultimately dependent on physician or other practitioner's complex medical judgment as specified by CMS. To validate this judgment, CMS recommended that reviewers use a screening tool as part of their medical review. This screening tool could include practice guidelines that are well accepted by the medical community but did not require or identify a specific criteria set.[11] Not surprisingly, there was and continues to be great variability in the application of outpatient versus inpatient status across hospitals in actual practice.[1, 12, 13] The ambiguity in the definition of a hospitalized patient's status helped spawn commercial clinical decision tools, such as InterQual (McKesson Corporation, San Francisco, CA) and MCG (formally known as Milliman Care Guidelines; MCG Health, LLC, Seattle, WA), to help define inpatients versus outpatients.[14, 15] However, these guidelines are complex, can be difficult to interpret and apply, and have been criticized for poor predictive value and attempting to replace physician judgment.[16, 17, 18] Furthermore, CMS has never formally endorsed any specific decision tool.

INPATIENT AND OUTPATIENT PAYMENTS AND THE RECOVERY AUDIT CONTRACTOR PROGRAM

In 2000, CMS started using Ambulatory Payment Classifications for hospital services, which made inpatient care more financially favorable for hospitals. In response to concerns that hospitals would be incentivized to overuse inpatient status, CMS made a number of changes to their payment system, including the creation of the Recovery Audit Program in 2003. This program was originally called the Recovery Audit Contractor (RAC) Program and continues to be most commonly referred to as the RAC program. The RAC program, tasked with finding and correcting improper claims to the Medicare program, began as a demonstration required in the Medicare Prescription Drug Improvement and Modernization Act of 2003 (MMA), and subsequently became a nationwide audit program under the Tax Relief and Health Care Act of 2006. Under this program, private contractors review hospital and billing records of Medicare patients and are paid on a contingency fee (8%12.5%) for all underpayments and overpayments that are identified and corrected.[19] Importantly, the RACs are not subject to any financial penalties for cases improperly denied.

RACs initially targeted many overnight inpatient stays for recoupment. These cases were attractive audit targets because the RACs could argue that the inpatient hospital services were delivered in the improper status based solely on the length of stay, without having to consider in their audit the complexity of decision making or medical necessity of the services provided. However, it is worth noting that with improvement in efficiency and advancements in medical technology, hospitals and physicians have been increasingly able to safely evaluate and treat medically complex and severely ill patients quickly, sometimes with just an overnight stay. As perspective, in 1965, the average length of stay for a Medicare patient was 13 days; in 2010, it was 5.4 days, with over one‐third of hospitalizations lasting <3 days.[20]

Concurrent with the increased RAC denials for services provided in an inpatient status, the use of observation services changed significantly from 2007 to 2012. The average length of stay for Medicare patients under outpatient status with observation services exceeded 24 hours in 2007, was 28.2 hours by 2009,[21] and grew to 29 hours by 2012.[22] Between July 2010 and December 2011, at the University of Wisconsin Hospital, 1 in 6 observation stays lasted longer than 48 hours, suggesting that long observation stays were no longer rare and exceptional as stated in CMS' own definition.[23] This same University of Wisconsin study also found that observation services were not well defined, with 1141 distinct diagnosis codes used for these services.[23]

Additionally, a Medicare Payment Advisory Commission (MedPAC; described on their website, www.medpac.gov, as a nonpartisan legislative branch agency that provides the US Congress with analysis and policy advice on the Medicare program) 2014 Report to Congress showed that from 2006 to 2012, outpatient services increased 28.5% whereas inpatient discharges decreased 12.6% over the same time period.[22]

Hospitals have also expressed concern that the RAC contingency fee payment model and a lack of penalty for improper denials promotes overzealous auditing.[24, 25] RAC recoupment has increased from approximately $939 million in 2011, to $2.4 billion in 2012, to $3.8 billion in 2013.[26, 27, 28] Given the money now at stake, it is not surprising that hospitals have become very active in appealing RAC denials. Self‐reported data submitted to the American Hospital Association (AHA) for the months January 2014 to March of 2014 show that hospitals now appeal 50% of RAC denials and win 66% of these appeals.[29] The AHA data also show that 69% of self‐reporting hospitals spent over $10,000 to manage their audit and appeals process over this same 3‐month time period, with 11% spending more than $100,000.

This appeals process is not only costly to hospitals, it is also lengthy. As of January 2014, the average wait time for an appeal hearing with an administrative law judge (level 3 appeal) exceeded 16 months.[30] In fact, the appeals process has become so backlogged that hospitals' rights to assignment of level 3 (administrative law judge) appeals have been temporarily suspended.[30] In August 2014, CMS offered a $0.68 on the dollar partial payment for hospitals willing to settle all eligible outstanding appeals in an attempt to relieve the appeals backlog.[31] In addition, the AHA currently has a suit against the US Department of Health & Human Services over the RAC appeals backlog.[32]

Increased use of outpatient status may be driven by pressures from the RAC program and, potentially, by improvements in the efficiency of care. Because hospitals are paid less for care provided under outpatient status than they are for the identical care provided under inpatient status, hospitals faced both potential financial penalty for improvements in efficiency and the threat of RAC audits.

THE 2‐MIDNIGHT RULE: A FIX?

Given the challenges in defining inpatient versus outpatient hospitalization, the increasing use of outpatient status and the increasing length of stay of outpatient hospitalizations with observation services, in 2013, CMS responded with new policies to define the visit status for hospitalized patients. On August 2, 2013, CMS announced the fiscal year 2014 hospital Inpatient Prospective Payment System final rule (IPPS‐2014) to become effective October 1, 2013. This document was formally issued as part of the Federal Register on August 19, 2013.[33] Central to the CMS IPPS‐2014 was a 2‐midnight benchmark that offered a major change in how physicians were to determine the status (inpatient vs outpatient) of hospitalized patients. With this 2‐midnight benchmark, now informally known as the 2‐midnight rule, CMS finalized its proposal to generally consider patients that are expected by a practitioner (with knowledge of the case and with admitting privileges) to need hospitalization that will span 2 or more midnights as inpatient. The IPPS‐2014 also finalized the converse of this: hospitalizations expected to span <2 midnights are to be regarded as outpatient with 2 exceptions:

1. If the hospitalization is associated with a procedure appearing on the previously described Medicare inpatient‐only procedures list, or
2. A rare and unusual circumstance in which an inpatient admission would be reasonable regardless of length of stay. Currently, unanticipated mechanical ventilation initiated during the hospitalization visit is the only rare and unusual circumstance that qualifies as such an exception.[7]

CMS' stated goals and expectations for the 2‐midnight benchmark were:

1. Reduce the growing number of prolonged hospitalizations (>48 hours) for Medicare beneficiaries under outpatient status.
2. Decrease billing disputes between hospitals and Medicare auditors, especially RACs, by establishing more clearly defined, time‐based status criteria.
3. Reduce the number of outpatient encounters overall. Because CMS expected the rule to convert a net increase of cases from outpatient to inpatient, resulting in higher payments to hospitals, CMS included a 0.2% payment cut in hospital reimbursement in the IPPS‐2014 as an offset.[33, 34]

Although unrelated to the goals and expectations above, the IPPS‐2014 also included a requirement that:

CMS allowed for an authentication (generally regarded as a cosignature that is timed and dated) of the inpatient admission order by an attending physician with admitting privileges, done prior to discharge, in cases where the inpatient order had been placed by a practitioner (such as a resident, fellow, or physician assistant) without admitting privileges. Attending physician authentication of the inpatient admission order must be done prior to discharge [a]s a condition of payment for hospital inpatient services under Medicare Part A.[35]

From the August 2, 2013 announcement until the effective date of October 1, 2013, hospitals had just 2 months to interpret and comply with the IPPS‐2014, a complex 546‐page document that required hospitals to make extensive changes to admission procedures, workflows, and electronic health records (EHRs). In addition, extensive physician, provider, and administrator education was needed. During these 2 months, hospitals continued to request additional information and clarification from CMS regarding many aspects of the IPPS‐2014, including basic questions that included (1) how to apply the 2‐midnight benchmark to patients who were transferred from 1 hospital to another and (2) when the clock started for hospital services in determining a patient's expected length of hospitalization.

Despite concerns voiced by Congress and medical organizations, the new policy went into effect as scheduled.[36, 37] However, just days prior to October 1, 2013, CMS issued a 3‐month limited suspension of auditing and enforcement of the 2‐midnight rule by the RACs that was subsequently extended by CMS 2 more times, first through March 31, 2014 and then again through September 30, 2014. Other audits to be performed by RACs and all other government audits, including those performed by Medicare Administrative Contractors (MACs) were allowed to continue.[38] In particular, the MACs were instructed to conduct patient status reviews using a probe and educate strategy, which, via educational outreach efforts, would instruct hospitals how to adapt to the new rule. On April 1, 2014, the Protecting Access to Medicare Act of 2014 was signed into law, which, under section 111 of this law, permitted CMS to continue medical review activities under the MAC probe and educate process through March of 2015, and prohibited CMS from allowing RACs to conduct inpatient hospital status reviews on claims with these same dates of admission, October 1, 2013 through March 31, 2015.

The MACs were created by the MMA of 2003, which mandated that the Secretary of Health & Human Services replace Part A Fiscal Intermediaries and Part B carriers with Medicare Administrative Contractors (MACs).[39] As established by CMS, MACs are multi‐state, regional contractors responsible for administering both Medicare Part A and Medicare Part B claims and serve as the primary operational contact between the Medicare Fee‐For‐Service program, and approximately 1.5 million health care providers enrolled in the program.[39]

THE IPPS‐2014 AND CMS' STATED GOALS AND EXPECTATIONS

In the analysis that accompanied the IPPS‐2014, Medicare expected the use of outpatient services to decrease overall, as the new rules would effectively eliminate almost all outpatient hospitalizations >48 hours. Although no official data are yet available from CMS, our early experience under the 2‐midnight rule has suggested that long observation stays have declined in frequency, a favorable outcome of the new policy. However, as designed, the new 2‐midnight IPPS rule most predominately affects 1‐day stays, or more accurately, 1‐midnight stays. This is because many hospitalizations that previously met inpatient criteria (as defined by commercially available products such as MCG or InterQual), but spanned <2 midnights would have been classified as inpatient prior to October 1, 2013. However, since October 1, 2013, these same hospitalizations are now classified as outpatient. An example of such a case is a patient who presents to an emergency department with symptoms of a transient ischemic attack and has a high ABCD (age 60 years, blood pressure 140/90 mm Hg at initial evaluation, clinical features, duration of symptoms, diabetes score).[40] Prior to the 2‐midnight rule, this patient, based on the severity of the signs and symptoms upon presentation, could have been appropriately hospitalized as an inpatient.

Now, under the current IPPS and the ability of many hospitals to efficiently evaluate and treat such patient in <2 midnights, the patient should be categorized as an outpatient, at least initially, despite the severity and high risk of his/her presentation. In fiscal year 2013, The Johns Hopkins Hospital had 1791, 1‐day inpatient stays for Medicare beneficiaries, representing 15.2% of all Medicare admissions. Similarly, in the 12 months just prior to the 2‐midnight rule (October 1, 2012 to September 30, 2013), 10.4% (1280) of all Medicare encounters at the University of Wisconsin were 1‐day inpatient stays under previous criteria. Because of implementation of the 2‐midnight rule in October 2013, Medicare outpatient hospitalization for 1‐day stays at The Johns Hopkins Hospital increased by 49%, from an average of 117 patients/month to 174 patients/month. Nationally, it is possible that a reduction in long observation stays could be offset by an increase in 1‐day‐stay outpatient hospitalization encounters.

A second key expectation and goal of IPPS‐2014 was, by shifting to a more concrete, time‐based definition of inpatient, to decrease the disagreement between hospitals and auditors regarding patient status (inpatient vs outpatient). As noted earlier, many disputes with auditors for hospitalizations prior to October 2013 did not involve the need or type of hospital services provided, but rather the status under which the care was provided. However, the new time‐based criterion hinges not on actual length of hospitalization, but the expected length of hospitalization as determined by a practitioner with admitting privileges and knowledge of the patient. Accurately and consistently predicting the length of hospitalization has proven to be challenging, even for the most experienced practitioners. Since October 2013, for patients hospitalized at The Johns Hopkins Hospital through its emergency department, the admitting physicians' expectation of whether a patient would require 1 versus 2 or more midnights of necessary hospitalization was correct only half of the time. Given past experience, the RACs may challenge the medical judgment that lead practitioners to expect a hospitalization of 2 or more midnights without having to challenge whether the care provided was medically necessary.

Further, the IPPS‐2014 has not been accompanied by any significant changes to the payment scheme for auditors. RACs continue to be paid a percentage of any monies they determine to have been improperly paid by CMS, but with no penalty for cases that are overturned on appeal. Historically, the vast majority of RAC recovery fees have been due to determination of overpayments by CMS.[41, 42] Despite the 2‐midnight rule, RACs will continue to have a financial incentive to allege overpayment. In the initial probe and educate audits by MACs under the new 2014‐IPPS, despite inpatient admission orders being authenticated and certified by an attending physician, claims are being denied because the documentation does not support an expectation for a 2‐midnight hospitalization. Namely, auditors are continuing to challenge not the medical necessity of the services that hospitals provide, but rather the status in which those services were provided. Thus far, the IPPS‐2014 does not appear to fully remedy the auditing conflict that existed prior to October 2013.

As noted above, the IPPS‐2014 also requires, as of October 1, 2013, as a condition of payment for hospital services under Part A, that the inpatient admission order must be either entered by a practitioner with admitting privileges or authenticated prior to discharge by an attending physician involved in the care of the patient in cases in which the inpatient admission order was entered by a practitioner without admitting privileges (eg, resident, physician assistant, or fellow).[43] The requirement of an attending physician's cosignature has involved major changes to physician workflow and the electronic heath record (EHR) framework at The Johns Hopkins and the University of Wisconsin Hospitals, and does not keep up with modern healthcare systems in which patients are admitted 24 hours a day by a variety of providers (eg, residents, nurse practitioners) who otherwise may write stand‐alone orders. These changes have proven to be time‐consuming, costly, and have not, to our knowledge, improved patient care or utilization of resources.

The new visit status rules have also led to confusion among clinicians. A recent large survey of hospitalists conducted by the Society of Hospital Medicine demonstrated that more than half of respondents disagreed that the 2‐midnight rule improved hospitalist workflow compared to prior observation policy.[44] In addition, only 40% of hospitalists reported confidence in how to apply the rule.[44] Thus, the intent to clarify visit status policy with the IPPS‐2014 has not translated to clear and useful rules for frontline clinicians.

FUTURE DIRECTIONS

After over a year under the 2‐midnight rule, although long observation stays may be reduced, it seems unlikely these new regulations will achieve 2 of CMS' stated goals: (1) decreasing the use of outpatient status for hospitalizations and (2) resolving status disputes between auditors and hospitals. In addition, attempts at compliance with the new rules and regulations have diverted large amounts of physician time and hospital resources away from patient care. There is a clear need to reform both the hospitalization status policy and the RAC programs that enforce these rules.

One path Congress and CMS could consider is to reform the current Medicare reimbursement paradigm for hospital services to eliminate the need to distinguish inpatient from outpatient status. For example, H.R. 1179Improving Access to Medicare Coverage Act of 2013,[45] of the 113th Congress, if reintroduced, would decouple the link between the qualification for skilled nursing facility benefits from visit status by allowing time spent hospitalized as an outpatient to count toward the 3‐day benchmark. The overarching goals of any visit status policy reform should be to: (1) simplify or eliminate the 2‐track status process for hospitalized patients, (2) stop or minimize the threat of audits based on status, and (3) maintain budget neutrality. Two additional options for consideration would be to: (1) create a low‐acuity modifier for use with patients anticipated to have short stays and low resource use and (2) preselect specific Diagnosis Related Groups based on historical data and create designations for those diagnoses of lesser intensity. Accountable care organizations contracts, a new model for healthcare payment, could potentially be structured to eliminate or simplify payment based on visit status for hospitalized patients. With bundled payments on the horizon and the possible phase‐out of fee‐for‐service reimbursement, the issue may become less paramount in the coming years. No solution will be perfect and must balance costs, ease of administration, and beneficiary protection.

There are reasons to be optimistic that change may soon be realized. CMS is currently considering significant hospitalization status policy reform. In the proposed IPPS‐2015, CMS asked for input on payment for short‐stay hospitalizations and, in the final IPPS‐2015 released August 4, 2014, CMS indicated its willingness to continue to work with stakeholders in revising these policies.[46] Additionally, CMS has responded to hospitals on 3 separate occasions by delaying RAC audits pertaining to the 2‐midnight rule. Further, the current MAC probe and educate audits focus on education with respect to 2‐midnight rule implementation rather than threatening hospitals with major financial penalties.[47] Congress has also been responsive in this area. In addition to the 3 delays announced by CMS, Congress passed legislation that mandated an additional delay to RAC audits that pertain to the 2‐midnight rule. Moreover, the Subcommittee on Health of the House Ways and Means Committee held hearings that included the 2‐midnight rule and RAC reform in May 2014, and the Senate Special Committee on Aging held hearings on the impact of visit status on Medicare beneficiaries in July 2014.[48, 49] Additionally, the House Ways and Means Health Subcommittee recently issued a draft bill to address Medicare hospital issues.[50] The OIG has also been responsive to hospital concerns regarding the current RAC program with a recent report recommending that CMS develop additional performance evaluation metrics to improve RAC performance and ensure that RACs are evaluated on all contract requirements.[51] Additionally, MedPAC has been considering several short‐stay payment reform options, modifying the need for a 3‐day inpatient hospitalization to qualify for postdischarge skilled nursing facility benefits and adjusting RAC contingency fees based on overturn rates.[52, 53] These actions by CMS, Congress, and the OIG, as well as the options under consideration by MedPAC, demonstrate a degree of regulatory and legislative responsiveness to hospital and provider concerns in the area of visit status determination.

The Medicare program is vital to tens of millions of disabled and elderly Americans. Fraud and abuse of the Medicare program should not be tolerated. Yet, the current system of assigning, monitoring, and auditing outpatient versus inpatient hospital care is in need of reform. It will be up to CMS and Congress to continue to work with hospitals and physicians to find an improved way to appropriately and fairly compensate hospitals for hospital services in a way that that does not depend on a poorly defined and contentious status of a patient. Such reform must include the RAC program. It is our hope that both CMS and Congress will prioritize status determination and payment reform so that Medicare beneficiaries, physicians, and hospitals all have a sustainable, fair, and transparent process.

Status determinations (outpatient versus inpatient) for hospitalized patients have become a routine part of patient care in the United States. Under the guidance provided by the Medicare Benefits Policy Manual, hospitalized Medicare beneficiaries are assigned 1 of these 2 statuses. The status assignment does not affect the care a patient can receive, but rather how the hospital services provided are billed to Medicare. Hospital services provided under inpatient status are billed under Medicare Part A. Hospital services provided under outpatient status, which includes all patients receiving observation services (commonly referred to as under observation), are billed under Medicare Part B. Whether hospital services are billed under Part A or Part B is important to hospitals and Medicare beneficiaries, as both the hospital reimbursement and beneficiary liability can vary greatly depending on whether services are billed under Part A versus Part B. Hospitals are generally reimbursed at a higher rate for services provided as an inpatient (Part A). The Office of the Inspector General (OIG) recently found that Medicare paid nearly three times more for a short inpatient stay than an [outpatient] stay for the same condition.[1] Medicare beneficiary liability also varies based on status. First, beneficiaries hospitalized as inpatients are subject to a deductible under Part A ($1,216 in 2014) for hospital services associated with that hospitalization and any future inpatient hospitalization beyond 60 days of discharge.[2] Beneficiaries hospitalized as outpatients are subject to the Medicare Part B deductible ($147 in 2014), and then a 20% copay on each individual outpatient hospital service, with no cumulative limit.[2, 3] In addition, hospital pharmacy charges for Medicare beneficiaries hospitalized as inpatients are covered under Medicare A. However, for Medicare patients hospitalized as outpatients, many medications are not covered by Medicare Part B benefits. Finally, time spent hospitalized as an outpatient does not count toward the Medicare 3‐day medically necessary inpatient stay requirement to qualify for the skilled nursing facility care benefit following discharge.

HISTORY AND INTENT OF INPATIENT AND OUTPATIENT STATUS DETERMINATIONS

Prior to October 1, 2013, the Centers for Medicare & Medicaid Services (CMS) stated that physician judgment and an expectation of at least an overnight hospitalization should determine inpatient status of hospitalized Medicare beneficiaries. Guidance as to when inpatient services were covered was found in the Medicare Benefits Policy Manual (MBPM)[4]:

For a subset of patients who are hospitalized under outpatient status, billing for observation services is allowed. CMS defines observation as a well defined set of services, that should last less than 24 hours and in only rare and exceptional casesspan more than 48 hours.[5] Many providers recognize the utility of a few additional hours of hospital care and/or testing in a hospital setting to determine whether a patient can go home or needs additional evaluation, monitoring, and/or treatment that can only be provided in a hospital, consistent with the CMS definition of observation.[6] It is important to note that although observation and outpatient are frequently used interchangeably, only outpatient is technically a CMS status. Patients in observation or under observation are, in fact, a subset of patients who are hospitalized under an outpatient status.

Outpatient status may also be appropriate for patients who require hospitalization for routine and expected overnight monitoring following a procedure. These patients are often not eligible for billing of observation services or as an inpatient because alternative methods of billing for the recovery time following the procedure exist. When determining the appropriate status of a Medicare beneficiary for a hospitalization following a procedure, physicians need to be aware of whether a specific procedure appears on the Medicare inpatient‐only procedures list.[7] Per CMS, procedures designated as inpatient only are reimbursed only when the patient is admitted as an inpatient at the time the procedure is performed.[8] Conversely, outpatient status for an overnight hospitalization associated with a procedure not on the inpatient‐only list is generally appropriate. Therefore, patients hospitalized for a procedure that appears on this list should always be hospitalized under inpatient status, regardless of the amount of time that the patient is expected to be hospitalized following the procedure, including those cases for which the hospitalization is expected to be only overnight.[7, 8] Only a limited number of Current Procedural Technology (CPT) codes, mostly surgical, automatically qualify for inpatient status and do not have outpatient prospective payment system eligibility. Although most procedures on the inpatient‐only list are associated with a hospitalization that commonly span at least 2 midnights, such as coronary artery bypass grafting, some potentially overnight stay cases, such as cholecystectomy (CPT 47600) appear on the 2014 inpatient‐only list.[9]

As noted above, prior to October 1, 2013, the Medicare definitions governing outpatient versus inpatient status included a 24‐hour benchmark. However, the MBPM also notes that: Admissions of particular patients are not covered or non‐covered solely on the basis of the length of time the patient actually spends in the hospital.[10]

In practice, status determination was ultimately dependent on physician or other practitioner's complex medical judgment as specified by CMS. To validate this judgment, CMS recommended that reviewers use a screening tool as part of their medical review. This screening tool could include practice guidelines that are well accepted by the medical community but did not require or identify a specific criteria set.[11] Not surprisingly, there was and continues to be great variability in the application of outpatient versus inpatient status across hospitals in actual practice.[1, 12, 13] The ambiguity in the definition of a hospitalized patient's status helped spawn commercial clinical decision tools, such as InterQual (McKesson Corporation, San Francisco, CA) and MCG (formally known as Milliman Care Guidelines; MCG Health, LLC, Seattle, WA), to help define inpatients versus outpatients.[14, 15] However, these guidelines are complex, can be difficult to interpret and apply, and have been criticized for poor predictive value and attempting to replace physician judgment.[16, 17, 18] Furthermore, CMS has never formally endorsed any specific decision tool.

INPATIENT AND OUTPATIENT PAYMENTS AND THE RECOVERY AUDIT CONTRACTOR PROGRAM

In 2000, CMS started using Ambulatory Payment Classifications for hospital services, which made inpatient care more financially favorable for hospitals. In response to concerns that hospitals would be incentivized to overuse inpatient status, CMS made a number of changes to their payment system, including the creation of the Recovery Audit Program in 2003. This program was originally called the Recovery Audit Contractor (RAC) Program and continues to be most commonly referred to as the RAC program. The RAC program, tasked with finding and correcting improper claims to the Medicare program, began as a demonstration required in the Medicare Prescription Drug Improvement and Modernization Act of 2003 (MMA), and subsequently became a nationwide audit program under the Tax Relief and Health Care Act of 2006. Under this program, private contractors review hospital and billing records of Medicare patients and are paid on a contingency fee (8%12.5%) for all underpayments and overpayments that are identified and corrected.[19] Importantly, the RACs are not subject to any financial penalties for cases improperly denied.

RACs initially targeted many overnight inpatient stays for recoupment. These cases were attractive audit targets because the RACs could argue that the inpatient hospital services were delivered in the improper status based solely on the length of stay, without having to consider in their audit the complexity of decision making or medical necessity of the services provided. However, it is worth noting that with improvement in efficiency and advancements in medical technology, hospitals and physicians have been increasingly able to safely evaluate and treat medically complex and severely ill patients quickly, sometimes with just an overnight stay. As perspective, in 1965, the average length of stay for a Medicare patient was 13 days; in 2010, it was 5.4 days, with over one‐third of hospitalizations lasting <3 days.[20]

Concurrent with the increased RAC denials for services provided in an inpatient status, the use of observation services changed significantly from 2007 to 2012. The average length of stay for Medicare patients under outpatient status with observation services exceeded 24 hours in 2007, was 28.2 hours by 2009,[21] and grew to 29 hours by 2012.[22] Between July 2010 and December 2011, at the University of Wisconsin Hospital, 1 in 6 observation stays lasted longer than 48 hours, suggesting that long observation stays were no longer rare and exceptional as stated in CMS' own definition.[23] This same University of Wisconsin study also found that observation services were not well defined, with 1141 distinct diagnosis codes used for these services.[23]

Additionally, a Medicare Payment Advisory Commission (MedPAC; described on their website, www.medpac.gov, as a nonpartisan legislative branch agency that provides the US Congress with analysis and policy advice on the Medicare program) 2014 Report to Congress showed that from 2006 to 2012, outpatient services increased 28.5% whereas inpatient discharges decreased 12.6% over the same time period.[22]

Hospitals have also expressed concern that the RAC contingency fee payment model and a lack of penalty for improper denials promotes overzealous auditing.[24, 25] RAC recoupment has increased from approximately $939 million in 2011, to $2.4 billion in 2012, to $3.8 billion in 2013.[26, 27, 28] Given the money now at stake, it is not surprising that hospitals have become very active in appealing RAC denials. Self‐reported data submitted to the American Hospital Association (AHA) for the months January 2014 to March of 2014 show that hospitals now appeal 50% of RAC denials and win 66% of these appeals.[29] The AHA data also show that 69% of self‐reporting hospitals spent over $10,000 to manage their audit and appeals process over this same 3‐month time period, with 11% spending more than $100,000.

This appeals process is not only costly to hospitals, it is also lengthy. As of January 2014, the average wait time for an appeal hearing with an administrative law judge (level 3 appeal) exceeded 16 months.[30] In fact, the appeals process has become so backlogged that hospitals' rights to assignment of level 3 (administrative law judge) appeals have been temporarily suspended.[30] In August 2014, CMS offered a $0.68 on the dollar partial payment for hospitals willing to settle all eligible outstanding appeals in an attempt to relieve the appeals backlog.[31] In addition, the AHA currently has a suit against the US Department of Health & Human Services over the RAC appeals backlog.[32]

Increased use of outpatient status may be driven by pressures from the RAC program and, potentially, by improvements in the efficiency of care. Because hospitals are paid less for care provided under outpatient status than they are for the identical care provided under inpatient status, hospitals faced both potential financial penalty for improvements in efficiency and the threat of RAC audits.

THE 2‐MIDNIGHT RULE: A FIX?

Given the challenges in defining inpatient versus outpatient hospitalization, the increasing use of outpatient status and the increasing length of stay of outpatient hospitalizations with observation services, in 2013, CMS responded with new policies to define the visit status for hospitalized patients. On August 2, 2013, CMS announced the fiscal year 2014 hospital Inpatient Prospective Payment System final rule (IPPS‐2014) to become effective October 1, 2013. This document was formally issued as part of the Federal Register on August 19, 2013.[33] Central to the CMS IPPS‐2014 was a 2‐midnight benchmark that offered a major change in how physicians were to determine the status (inpatient vs outpatient) of hospitalized patients. With this 2‐midnight benchmark, now informally known as the 2‐midnight rule, CMS finalized its proposal to generally consider patients that are expected by a practitioner (with knowledge of the case and with admitting privileges) to need hospitalization that will span 2 or more midnights as inpatient. The IPPS‐2014 also finalized the converse of this: hospitalizations expected to span <2 midnights are to be regarded as outpatient with 2 exceptions:

1. If the hospitalization is associated with a procedure appearing on the previously described Medicare inpatient‐only procedures list, or
2. A rare and unusual circumstance in which an inpatient admission would be reasonable regardless of length of stay. Currently, unanticipated mechanical ventilation initiated during the hospitalization visit is the only rare and unusual circumstance that qualifies as such an exception.[7]

CMS' stated goals and expectations for the 2‐midnight benchmark were:

1. Reduce the growing number of prolonged hospitalizations (>48 hours) for Medicare beneficiaries under outpatient status.
2. Decrease billing disputes between hospitals and Medicare auditors, especially RACs, by establishing more clearly defined, time‐based status criteria.
3. Reduce the number of outpatient encounters overall. Because CMS expected the rule to convert a net increase of cases from outpatient to inpatient, resulting in higher payments to hospitals, CMS included a 0.2% payment cut in hospital reimbursement in the IPPS‐2014 as an offset.[33, 34]

Although unrelated to the goals and expectations above, the IPPS‐2014 also included a requirement that:

CMS allowed for an authentication (generally regarded as a cosignature that is timed and dated) of the inpatient admission order by an attending physician with admitting privileges, done prior to discharge, in cases where the inpatient order had been placed by a practitioner (such as a resident, fellow, or physician assistant) without admitting privileges. Attending physician authentication of the inpatient admission order must be done prior to discharge [a]s a condition of payment for hospital inpatient services under Medicare Part A.[35]

From the August 2, 2013 announcement until the effective date of October 1, 2013, hospitals had just 2 months to interpret and comply with the IPPS‐2014, a complex 546‐page document that required hospitals to make extensive changes to admission procedures, workflows, and electronic health records (EHRs). In addition, extensive physician, provider, and administrator education was needed. During these 2 months, hospitals continued to request additional information and clarification from CMS regarding many aspects of the IPPS‐2014, including basic questions that included (1) how to apply the 2‐midnight benchmark to patients who were transferred from 1 hospital to another and (2) when the clock started for hospital services in determining a patient's expected length of hospitalization.

Despite concerns voiced by Congress and medical organizations, the new policy went into effect as scheduled.[36, 37] However, just days prior to October 1, 2013, CMS issued a 3‐month limited suspension of auditing and enforcement of the 2‐midnight rule by the RACs that was subsequently extended by CMS 2 more times, first through March 31, 2014 and then again through September 30, 2014. Other audits to be performed by RACs and all other government audits, including those performed by Medicare Administrative Contractors (MACs) were allowed to continue.[38] In particular, the MACs were instructed to conduct patient status reviews using a probe and educate strategy, which, via educational outreach efforts, would instruct hospitals how to adapt to the new rule. On April 1, 2014, the Protecting Access to Medicare Act of 2014 was signed into law, which, under section 111 of this law, permitted CMS to continue medical review activities under the MAC probe and educate process through March of 2015, and prohibited CMS from allowing RACs to conduct inpatient hospital status reviews on claims with these same dates of admission, October 1, 2013 through March 31, 2015.

The MACs were created by the MMA of 2003, which mandated that the Secretary of Health & Human Services replace Part A Fiscal Intermediaries and Part B carriers with Medicare Administrative Contractors (MACs).[39] As established by CMS, MACs are multi‐state, regional contractors responsible for administering both Medicare Part A and Medicare Part B claims and serve as the primary operational contact between the Medicare Fee‐For‐Service program, and approximately 1.5 million health care providers enrolled in the program.[39]

THE IPPS‐2014 AND CMS' STATED GOALS AND EXPECTATIONS

In the analysis that accompanied the IPPS‐2014, Medicare expected the use of outpatient services to decrease overall, as the new rules would effectively eliminate almost all outpatient hospitalizations >48 hours. Although no official data are yet available from CMS, our early experience under the 2‐midnight rule has suggested that long observation stays have declined in frequency, a favorable outcome of the new policy. However, as designed, the new 2‐midnight IPPS rule most predominately affects 1‐day stays, or more accurately, 1‐midnight stays. This is because many hospitalizations that previously met inpatient criteria (as defined by commercially available products such as MCG or InterQual), but spanned <2 midnights would have been classified as inpatient prior to October 1, 2013. However, since October 1, 2013, these same hospitalizations are now classified as outpatient. An example of such a case is a patient who presents to an emergency department with symptoms of a transient ischemic attack and has a high ABCD (age 60 years, blood pressure 140/90 mm Hg at initial evaluation, clinical features, duration of symptoms, diabetes score).[40] Prior to the 2‐midnight rule, this patient, based on the severity of the signs and symptoms upon presentation, could have been appropriately hospitalized as an inpatient.

Now, under the current IPPS and the ability of many hospitals to efficiently evaluate and treat such patient in <2 midnights, the patient should be categorized as an outpatient, at least initially, despite the severity and high risk of his/her presentation. In fiscal year 2013, The Johns Hopkins Hospital had 1791, 1‐day inpatient stays for Medicare beneficiaries, representing 15.2% of all Medicare admissions. Similarly, in the 12 months just prior to the 2‐midnight rule (October 1, 2012 to September 30, 2013), 10.4% (1280) of all Medicare encounters at the University of Wisconsin were 1‐day inpatient stays under previous criteria. Because of implementation of the 2‐midnight rule in October 2013, Medicare outpatient hospitalization for 1‐day stays at The Johns Hopkins Hospital increased by 49%, from an average of 117 patients/month to 174 patients/month. Nationally, it is possible that a reduction in long observation stays could be offset by an increase in 1‐day‐stay outpatient hospitalization encounters.

A second key expectation and goal of IPPS‐2014 was, by shifting to a more concrete, time‐based definition of inpatient, to decrease the disagreement between hospitals and auditors regarding patient status (inpatient vs outpatient). As noted earlier, many disputes with auditors for hospitalizations prior to October 2013 did not involve the need or type of hospital services provided, but rather the status under which the care was provided. However, the new time‐based criterion hinges not on actual length of hospitalization, but the expected length of hospitalization as determined by a practitioner with admitting privileges and knowledge of the patient. Accurately and consistently predicting the length of hospitalization has proven to be challenging, even for the most experienced practitioners. Since October 2013, for patients hospitalized at The Johns Hopkins Hospital through its emergency department, the admitting physicians' expectation of whether a patient would require 1 versus 2 or more midnights of necessary hospitalization was correct only half of the time. Given past experience, the RACs may challenge the medical judgment that lead practitioners to expect a hospitalization of 2 or more midnights without having to challenge whether the care provided was medically necessary.

Further, the IPPS‐2014 has not been accompanied by any significant changes to the payment scheme for auditors. RACs continue to be paid a percentage of any monies they determine to have been improperly paid by CMS, but with no penalty for cases that are overturned on appeal. Historically, the vast majority of RAC recovery fees have been due to determination of overpayments by CMS.[41, 42] Despite the 2‐midnight rule, RACs will continue to have a financial incentive to allege overpayment. In the initial probe and educate audits by MACs under the new 2014‐IPPS, despite inpatient admission orders being authenticated and certified by an attending physician, claims are being denied because the documentation does not support an expectation for a 2‐midnight hospitalization. Namely, auditors are continuing to challenge not the medical necessity of the services that hospitals provide, but rather the status in which those services were provided. Thus far, the IPPS‐2014 does not appear to fully remedy the auditing conflict that existed prior to October 2013.

As noted above, the IPPS‐2014 also requires, as of October 1, 2013, as a condition of payment for hospital services under Part A, that the inpatient admission order must be either entered by a practitioner with admitting privileges or authenticated prior to discharge by an attending physician involved in the care of the patient in cases in which the inpatient admission order was entered by a practitioner without admitting privileges (eg, resident, physician assistant, or fellow).[43] The requirement of an attending physician's cosignature has involved major changes to physician workflow and the electronic heath record (EHR) framework at The Johns Hopkins and the University of Wisconsin Hospitals, and does not keep up with modern healthcare systems in which patients are admitted 24 hours a day by a variety of providers (eg, residents, nurse practitioners) who otherwise may write stand‐alone orders. These changes have proven to be time‐consuming, costly, and have not, to our knowledge, improved patient care or utilization of resources.

The new visit status rules have also led to confusion among clinicians. A recent large survey of hospitalists conducted by the Society of Hospital Medicine demonstrated that more than half of respondents disagreed that the 2‐midnight rule improved hospitalist workflow compared to prior observation policy.[44] In addition, only 40% of hospitalists reported confidence in how to apply the rule.[44] Thus, the intent to clarify visit status policy with the IPPS‐2014 has not translated to clear and useful rules for frontline clinicians.

FUTURE DIRECTIONS

After over a year under the 2‐midnight rule, although long observation stays may be reduced, it seems unlikely these new regulations will achieve 2 of CMS' stated goals: (1) decreasing the use of outpatient status for hospitalizations and (2) resolving status disputes between auditors and hospitals. In addition, attempts at compliance with the new rules and regulations have diverted large amounts of physician time and hospital resources away from patient care. There is a clear need to reform both the hospitalization status policy and the RAC programs that enforce these rules.

One path Congress and CMS could consider is to reform the current Medicare reimbursement paradigm for hospital services to eliminate the need to distinguish inpatient from outpatient status. For example, H.R. 1179Improving Access to Medicare Coverage Act of 2013,[45] of the 113th Congress, if reintroduced, would decouple the link between the qualification for skilled nursing facility benefits from visit status by allowing time spent hospitalized as an outpatient to count toward the 3‐day benchmark. The overarching goals of any visit status policy reform should be to: (1) simplify or eliminate the 2‐track status process for hospitalized patients, (2) stop or minimize the threat of audits based on status, and (3) maintain budget neutrality. Two additional options for consideration would be to: (1) create a low‐acuity modifier for use with patients anticipated to have short stays and low resource use and (2) preselect specific Diagnosis Related Groups based on historical data and create designations for those diagnoses of lesser intensity. Accountable care organizations contracts, a new model for healthcare payment, could potentially be structured to eliminate or simplify payment based on visit status for hospitalized patients. With bundled payments on the horizon and the possible phase‐out of fee‐for‐service reimbursement, the issue may become less paramount in the coming years. No solution will be perfect and must balance costs, ease of administration, and beneficiary protection.

There are reasons to be optimistic that change may soon be realized. CMS is currently considering significant hospitalization status policy reform. In the proposed IPPS‐2015, CMS asked for input on payment for short‐stay hospitalizations and, in the final IPPS‐2015 released August 4, 2014, CMS indicated its willingness to continue to work with stakeholders in revising these policies.[46] Additionally, CMS has responded to hospitals on 3 separate occasions by delaying RAC audits pertaining to the 2‐midnight rule. Further, the current MAC probe and educate audits focus on education with respect to 2‐midnight rule implementation rather than threatening hospitals with major financial penalties.[47] Congress has also been responsive in this area. In addition to the 3 delays announced by CMS, Congress passed legislation that mandated an additional delay to RAC audits that pertain to the 2‐midnight rule. Moreover, the Subcommittee on Health of the House Ways and Means Committee held hearings that included the 2‐midnight rule and RAC reform in May 2014, and the Senate Special Committee on Aging held hearings on the impact of visit status on Medicare beneficiaries in July 2014.[48, 49] Additionally, the House Ways and Means Health Subcommittee recently issued a draft bill to address Medicare hospital issues.[50] The OIG has also been responsive to hospital concerns regarding the current RAC program with a recent report recommending that CMS develop additional performance evaluation metrics to improve RAC performance and ensure that RACs are evaluated on all contract requirements.[51] Additionally, MedPAC has been considering several short‐stay payment reform options, modifying the need for a 3‐day inpatient hospitalization to qualify for postdischarge skilled nursing facility benefits and adjusting RAC contingency fees based on overturn rates.[52, 53] These actions by CMS, Congress, and the OIG, as well as the options under consideration by MedPAC, demonstrate a degree of regulatory and legislative responsiveness to hospital and provider concerns in the area of visit status determination.

The Medicare program is vital to tens of millions of disabled and elderly Americans. Fraud and abuse of the Medicare program should not be tolerated. Yet, the current system of assigning, monitoring, and auditing outpatient versus inpatient hospital care is in need of reform. It will be up to CMS and Congress to continue to work with hospitals and physicians to find an improved way to appropriately and fairly compensate hospitals for hospital services in a way that that does not depend on a poorly defined and contentious status of a patient. Such reform must include the RAC program. It is our hope that both CMS and Congress will prioritize status determination and payment reform so that Medicare beneficiaries, physicians, and hospitals all have a sustainable, fair, and transparent process.