Improving Healthcare Value: Addressing the confusing costs of observation hospitalizations

In the 1980s, the Centers for Medicare and Medicaid Services (CMS) began differentiating between traditional inpatient hospitalizations and a new class of outpatient hospitalizations, commonly called observation hospitalizations. Observation hospitalizations were intended to be used for lower-acuity diagnoses that usually resolve or improve within 24 hours, allowing more time for medical decision-making than is traditionally afforded in emergency departments. The development of dedicated observation units has allowed the health care system to realize the cost-savings associated with the less intensive care of these patients (1). One model estimated that appropriate use of observation units could save the health care system more than $3 billion per year (2).

Use of observation hospitalizations has expanded significantly in number and scope since the creation of this type of status. Between 2012 and 2017, observation hospitalizations for Medicare beneficiaries increased by 32%; this growth in observation volume increased CMS spending on these services by nearly 20% (3). Additionally, rather than being confined to a specific set of simple and common diagnoses, observation hospitalizations for traditional Medicare beneficiaries may now be associated with over 1,000 different diagnosis codes (4). At present, there is no relationship between coding determinations and inpatient versus outpatient hospitalizations; as a result, many observation hospitalizations are indistinguishable from inpatient hospitalizations in terms of complexity, length of stay, and location, with even ICU patients being treated under observation (5).

CMS has attempted to clarify the appropriate use of observation hospitalizations. The Inpatient Only List (IPOL), established over 20 years ago, is a set of procedures that CMS deems suitable for inpatient status and thus Medicare Part A coverage, regardless of the expected or actual length of stay. In 2013, CMS implemented the two-midnight rule, which states that services provided in a hospital will be paid for as an inpatient if the admitting provider reasonably expects the patient’s medically necessary care to last two midnights or more. Despite these clarifications, considerable confusion and uncertainty remain around this issue. In 2020, CMS proposed eliminating the IPOL entirely, yet failed to outline under what circumstances the approximately 1,700 procedures on the list might be appropriate for inpatient versus outpatient billing (6). Additionally, in the year following the implementation of the two-midnight rule, more than 20% of the nearly 3.4 million observation hospitalizations – about 750,000 hospitalizations – lasted longer than two midnights (7).

The confusion surrounding the appropriate use of observation hospitalizations is even greater when considering Medicare Advantage (MA) plans, which in 2021, covered 42% of all Medicare beneficiaries (8). MA plans are permitted to negotiate separate contracts with individual health care organizations, may or may not adhere to CMS’ IPOL, and may not be required to follow the two-midnight rule (9). There is little transparency in the billing practices of these payors including reimbursement rates for services provided as inpatient versus outpatient. There is currently no legislative requirement that MA plans share their payment structures with CMS or with their beneficiaries. Likewise, there is no explanation for how costs shift when an MA payor retroactively denies an inpatient claim and requests the hospitalization be rebilled as observation.

Recent claims data from a single MA payor at a large academic medical center provides an example of how varying designations of inpatient versus outpatient status impacts hospital revenue. In the cases represented in the data set, the hospital billed the MA plan $147,542 for care provided. The MA plan initially approved and paid these claims as inpatient episodes, reimbursing the hospital $58,187. The care delivered under these claims was provided on inpatient wards by inpatient teams, not in emergency departments or observation units. Months after the patients had been discharged and the claims closed, the MA payor performed an audit and decided that the care qualified as outpatient rather than inpatient. The final payment authorized by the payor was $13,174, 8.9% of the original hospital claim and 22.6% of the original inpatient payment; the hospital was expected to refund the difference. Necessity of the services, location of service delivery, or quality of the treatment provided, was not disputed by the payor in the audit. The only change was that the MA payor determined, through an opaque internal process, that the care should be designated as outpatient episodes and thus paid at a significantly lower rate.

The conflicts of interest created by this two-tiered payment structure are immediately obvious. Hospitals benefit from higher inpatient payments, while payors – who are able to dictate status determination and payment rates – benefit from lower observation payments for clinically indistinguishable care. It unsurprising that, as with Medicare, the percentage of observation hospitalizations for MA plan beneficiaries is increasing. Twenty-one hospitals in California’s Kaiser Permanente system found that the number of observations in their system increased by more than 600% between 2000 and 2017, from about 16,500 to 129,215 (10). Analysis of over 150 million MA enrollees showed that their chances of having an observation hospitalization more than doubled between 2004 and 2014 (11). Observation hospitalizations for MA plan beneficiaries are also increasingly likely to last longer than 24 hours or two midnights. In 2006, only 4.5% of observation hospitalizations billed by MA plans lasted two or more days; in 2014 that number was 19.2%, an increase of over 300% (11).

Hospitals and insurers spend significant resources to manage the business of status determination. Claims are submitted and denied, decisions are appealed and reversed. Private companies have entered the market to help both hospitals and insurers, adding extra expense to maintain this two-tiered system. Administrative costs related to status determination have yet to be adequately assessed, but Reynolds et al determined that for a 12-month period between 2011 and 2012, 41% of hospital case manager workforce was directed at status determinations, rather than patient care or quality improvement efforts (12). If appropriate status determination improved the type or quality of care a patient was provided, this billing structure and its associated costs might be reasonable. It does not, and our data confirms this: the quality and type of care delivered cannot change after the fact, but the payment rate can. Indeed, using readmissions as a marker for quality of care as CMS does in its Hospital Readmissions Reduction Program, available data on traditional Medicare admissions suggests both observation and inpatient hospitalizations are associated with similar 30-day readmission rates (13).

The time has come to address this hospital status classification system as a component of avoidable waste in healthcare. Several possible solutions exist. CMS could explicitly mandate that all Medicare Advantage programs follow the two-midnight rule to determine status; eliminating ambiguity on this issue would bring consistency to the status determination process, reducing the costs of navigating the myriad rules of status determination (14). Alternatively, CMS could mandate cost transparency among Medicare Advantage programs, providing hospitals and patients more data with which to make decisions on contracting and plan selection. Observation could be eliminated entirely, in a budget-neutral manner, through the development of payment rates that represent a blend of current inpatient diagnosis-related group payments and outpatient costs (15). Ultimately, we believe a complete overhaul of the concept of an outpatient observation hospitalization is needed. Our view is that patients who are cared for on hospital wards by inpatient care teams should, except in very rare cases, be classified as exactly what they appear to be: inpatients.