A, All observation stays. B, All inpatient stays. Cost numbers reflect total cost of care per encounter; reimbursement reflects payment to the hospital for the encounter. For 2 midnights and longer observation stays, estimated reimbursement per encounter is $2,639 more than current hospital reimbursement. For shorter than 2 midnight inpatient stays, estimated reimbursement per encounter is $3,050 less than current hospital reimbursement.
A, Shorter stays (<2 midnights). B, Longer stays (≥2 midnights). C, All encounters. Reimbursement reflects payment to the hospital for the encounter. Dashed line indicates per-encounter average weighted inpatient and observation current reimbursement.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Sheehy AM, Graf BK, Gangireddy S, Formisano R, Jacobs EA. “Observation Status” for Hospitalized Patients: Implications of a Proposed Medicare Rules Change. JAMA Intern Med. 2013;173(21):2004–2006. doi:10.1001/jamainternmed.2013.9382
Hospitalized patients in the United States are increasingly placed on “observation status.”1 The Centers for Medicare & Medicaid Services (CMS) state that observation should rarely exceed 48 hours.2 At our hospital, however, observation stays frequently exceed 48 hours, particularly for adult general medicine patients, as our research group recently reported in this journal.3 Reimbursement has not covered the costs, resulting in operating losses.3
After our study was completed, CMS proposed a rules change,4 assigning inpatient status and Medicare Part A coverage to all hospital stays of 2 midnights (“2 days”) or longer, assuming medical necessity supported a 2 midnights-or-longer stay. With few exceptions, stays shorter than 2 midnights would be considered observation and therefore not be covered by Medicare Part A. The skilled nursing facility benefit under Medicare Part A would continue to require a 3-midnight (“3 day”) stay. After payment of the specified deductible and applicable coinsurance, Medicare Part A covers hospital care and skilled nursing facility care, among other services. We used our data to assess the implications of the proposed rules change.
As previously described, we studied all observation and inpatient encounters at the University of Wisconsin Hospital between July 1, 2010, and December 31, 2011.3 The University of Wisconsin–Madison Health Sciences institutional review board approved the study. We divided existing inpatient and observation cost and reimbursement data by encounters shorter than 2 midnights (<2 midnights) and 2 midnights or longer (≥2 midnights). We then assumed unchanged costs and applied existing observation reimbursement rates for encounters shorter than 2 midnights (<2 midnights) to inpatient encounters shorter than 2 midnights (<2 midnights) and existing inpatient reimbursement rates for encounters 2 midnights or longer (≥2 midnights) to observation encounters 2 midnights or longer (≥2 midnights). Thus, stays of shorter than 2 midnights (<2 midnights), previously considered inpatient, were reclassified as observation, and stays of 2 midnights or longer (≥2 midnights), previously classified as observation, were reclassified as inpatient.
Discharges to skilled nursing facilities were evaluated by a 3-midnight cut point to model the number of current observation patients who would gain this benefit under Medicare Part A. Finally, as patients admitted shortly before midnight achieve inpatient status more quickly than those admitted just after midnight, we compared actual encounter length of stay (in hours) for encounters of fewer than 2 and 2 or more midnights.
Under the proposed rules change, per-encounter reimbursement for stays of fewer than 2 midnights that were previously classified as inpatient decreased by $3050, since they would all be reclassified as observation. Reimbursement increased by $2639 for stays of 2 midnights or longer that were previously classified as observation and that would be reclassified as inpatient (Figure 1). Given the smaller number of encounters currently classified as observation of 2 or more midnights (1211) compared with current inpatient encounters of shorter than 2 midnights (8231), total reimbursement gains for observation patients (+$3.2 million) would be dwarfed by decreased reimbursement for inpatients (−$25.1 million) (Figure 2).
Of 320 current observation patients who were discharged to skilled nursing facilities, only 112 (35.0%) had stays of 3 or more midnights that would qualify them for a Medicare Part A benefit under the proposed rules change.
Our shortest encounter of 2 midnights or longer was 26.6 hours, and our longest encounter of shorter than 2 midnights was 47.2 hours in length. Almost half (5442/11 598 [46.9%]) of our shorter-than-2-midnight encounters were between 26.6 and 47.2 hours. Thus, patients with identical lengths of stay in hours would often have different hospital status and Medicare coverage under the proposed changes, solely because of the time of their admission relative to midnight.
The CMS seems to have anticipated that the rules change would lead to higher hospital reimbursement because a 0.2% reduction in Diagnosis Related Group (DRG) payments was also proposed.5 On the contrary, our analysis suggested a $14.6 million/y reduction in reimbursement for our hospital before the additional 0.2% reduction. Also, many Medicare beneficiaries discharged to skilled nursing facilities would not gain a Part A benefit because their hospital stays would be for shorter than 3 midnights.
Although we assumed that hospitals would maintain the current proportions of shorter-than-2-midnight and 2-midnight-or-longer stays under the proposed rules, the proportions may change. Given the substantial financial incentives favoring 2 midnights or longer, more patients may remain in the hospital for that period, thus increasing both reimbursements and the cost of care. Although stays of 2 midnights or longer still would require justification based on medical necessity, such determinations are notoriously subjective and difficult to make, particularly when a difference in length of stay of only a few hours could have substantial financial implications. On August 2, 2013, CMS issued final rules for observation stays and payments for fiscal year 2014, which begins on October 1, 2013.6 Our findings suggest that CMS should continue to evaluate the changes to observation status and make further modifications as needed in the years ahead.
Corresponding Author: Ann M. Sheehy, MD, MS, Division of Hospital Medicine, Department of Medicine, University of Wisconsin School of Medicine and Public Health, 1685 Highland Ave, MFCB 3126, Madison, WI 53705 (firstname.lastname@example.org).
Published Online: August 26, 2013. doi:10.1001/jamainternmed.2013.9382.
Author Contributions: Dr Sheehy had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Sheehy, Graf, Jacobs.
Acquisition of data: Sheehy.
Analysis and interpretation of the data: Sheehy, Graf, Gangireddy, Formisano, Jacobs.
Drafting of the manuscript: Sheehy.
Critical revision of the manuscript for important intellectual content: Graf, Gangireddy, Formisano, Jacobs.
Obtaining funding: Sheehy.
Administrative, technical, or material support: Graf, Gangireddy, Formisano, Jacobs.
Conflict of Interest Disclosures: Dr Graf reports receiving royalties from Smith & Nephew for patents he has assigned to them involving orthopedic implants used in arthroscopic surgery. No other disclosures are reported.
Funding/Support: This study was supported by the University of Wisconsin Department of Medicine (Dr Sheehy). Dr Jacobs was supported by the National Institutes on Aging, grant R01 AG033172-01.
Role of the Sponsors: The sponsors had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; and preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication
Additional Contributions: We thank Daniel S. Dexter, BA, for assistance with the data set, Andrew T. LaRocque, BBA, for assistance with data presentation, and Cynthia Heidke, BA, for assistance with data presentation and editing. They received no additional compensation for their work on this study.