Observation “Services” and Observation “Care”—One Word Can Mean a World of Difference

Observation care, observation status, and observation stays have recently become the focus of considerable policy, provider, and patient attention. A key component of this public debate has focused on whether hospitals are shifting costs onto patients by admitting Medicare beneficiaries under “observation status” to avoid readmission penalties or CMS Recovery Audit Contractor claim denials (Baugh and Schuur 2013). Absent from this discussion has been the opportunity to step back and distinguish important nuances amid the confusing terminology and assess the role for hospital observation across the nation.

In this issue of Health Services Research, Dr. Wright and colleagues identify patient, hospital, and geographic predictors of hospital observation service use in a 100 percent Medicare administrative claims file. They found observation utilization varied at the patient level according to race and age and at the hospital level according to location, bed size, and ownership type.

Throughout, the authors are careful to use the term “observation stay,” and not necessarily “observation care,” let alone “care within a dedicated observation unit,” because a hospital bill provides no additional detail as to what the term means. Ross et al. (2013) recently proposed categorizing observation stays along a continuum ranging from Type 1, which best describes “observation care,” to Type 4, which refers to discretionary care that is billed as observation services. In this typology, Type 1 observation services are based on specific clinical protocols and delivered in a dedicated observation unit—these services most closely parallel the past two decades of research showing that observation care can be more efficient and effective than inpatient hospitalization for clinical conditions such as chest pain, asthma, and atrial fibrillation. Type II observation services, like Type I, are delivered in a dedicated observation unit but do not follow a predefined clinical protocol and tend to be used for conditions such as mild dehydration or for elderly patients who suffer a fall without major injury. Type III observation services use specific clinical protocols that are unlikely to be different from Type I observation services; however, they are not delivered in a dedicated unit and therefore may carry some of the operational inefficiencies of traditional inpatient care. Finally, Type IV services are neither protocol-based or delivered in a dedicated unit and most likely represent hospital care delivered in parallel to traditional inpatient hospital care for patients who do not meet institutional or policy-driven criteria for inpatient reimbursement. Unlike intensive care unit services, which are distinctly reflected in administrative claims data, observation care delivered in a dedicated unit is billed identically to discretionary observation services delivered in any hospital bed. Current proposals by the Centers for Medicare and Medicaid Services seek to modify hospital reimbursement for observation and inpatient hospitalizations; however, no proposal to date is designed to expand the number of or increase the granularity of existing billing codes to enable researchers or policy makers to distinguish between each type of observation service.

Previous studies suggest that increased Type 1 observation stays (actual “observation care”) could save the US health care system billions annually and simultaneously improve quality (Baugh et al. 2012; Baugh and Schuur 2013), while Type 4 observation stays have become the subject of an Office of the Inspector General inquiry seeking to improve the efficiency and appropriateness of Medicare payments for hospital “services.” The current work by Wright et al. (2014) uses administrative claims and, as such, captures any kind of observation stay, not necessarily only Type 1 “observation care” (Wright 2013). In fact, the authors found that 63 percent of observation stays for Medicare beneficiaries lasted longer than 24 hours, which exceeds both the duration expected by Medicare for most observation stays and the benchmark used in protocol-driven dedicated observation units. The distinction between “observation services” that may closely mirror inpatient hospitalization and protocol-driven “observation care” is important, and it carries marked implications for the interpretation of these findings by providers and policy makers.

As expected, the authors also found marked variation between hospitals in the use of observation stays. However, the finding that 18 percent of hospitals in the sample deliver no observation services to Medicare beneficiaries and that the mean number of hospital observation stays was only 14.2 per 1,000 outpatient visits (compared to an inpatient admission rate of 312 per 1,000 in the Medicare population nationally) may come as a surprise amidst the substantial policy attention. Smaller hospitals and critical access hospitals (often the same) seem to be far less likely to bill for observation services. Is it because these hospitals are too small and lack the sufficient acute care volume to justify the fixed costs of a dedicated observation unit and staff for Type 1 observation care, and if so, do patients therefore get less efficient hospitalizations? Or, could it be that patients with longer observation stays at critical access hospitals are billed as inpatient hospitalizations? The authors' findings of significantly shorter observation length of stay at both smaller and critical access hospitals would support this premise. As a result, we should ask whether these observed differences between hospitals represent true differences in the observation care provided or just differences in observation billing strategies. For many hospitals, the use of observation stays may markedly improve hospital efficiency and create a safer environment for all patients regardless of the duration or type of observation care. Before policy makers pass judgment on observation stays as a whole, we should be careful to distinguish billing patterns from clinical care. To improve patient care, future research must move beyond the limitations of administrative claims data by capturing hospital-level details about how observation stays are organized and provided before we can conclude whether observed variations truly represent meaningful differences in the quality of hospital care.

The influences on hospitals to use (or not use) observation services are wide-ranging and complex; Wright et al.'s (2014) analysis reinforces this by demonstrating patient, hospital, and state-level influences on the prevalence, penetration, and duration of observation stays among Medicare beneficiaries. The finding that most observation stays for Medicare beneficiaries are consolidated around several prevalent diagnoses in hospitals with disproportionately more emergency department visits reflects a stark reality of modern acute care. Emergency department and hospital overcrowding have garnered numerous national calls for attention, yet the continuing trend of hospital closure and increasing use of the ED for all acute, unscheduled care continues unabated (Morganti et al. 2013). Observation stays at many larger hospitals provide a “release valve” to the capacity bottleneck between the emergency department and inpatient hospital. Just as the type and value of each observation visit cannot be gleaned from administrative claims alone, health services researchers must also assess the aggregate effects of these observation stays on overall hospital efficiency and outcomes, which may carry just as much significance for patients as out-of-pocket costs. As C. S. Lewis famously wrote: “For what you see and hear depends on where you are standing: it also depends on what sort of person you are.” In the case of hospital observation that could be the difference between a “service” and “care.”

Supporting Information

Additional supporting information may be found in the online version of this article:

Appendix SA1: Author Matrix.