The task of providing care for acutely ill, hospitalized patients has become increasingly challenging with each subsequent year.1 Resident duty hours restrictions, service caps, curriculum requirements, and limitations on the numbers of full-time equivalent (FTE) house officers have impaired the ability of teaching hospitals to care for a growing volume of patients.2–5 For many hospitals, the answer to this crisis has been (1) to hire hospitalists6 and midlevel practitioners7–11 to run nonteaching services and (2) to pay for their services with increased reimbursement from the additional patients.12,13
Hospitals with a payer mix that favors case-based, rather than per diem, reimbursement accrue additional benefit from the reduced length of stay (LOS) associated with hospitalist services.14,15 In contrast, our public hospital (Olive View–UCLA Medical Center, Sylmar, California) receives no direct reimbursement for approximately one-half of all inpatients, because they are indigent and have no third-party payer source. The vast majority of patients who do have insurance are funded by Medi-Cal (California's Medicaid program). For most hospitalized patients, Medi-Cal provides per diem reimbursement, after retrospective chart review to determine the patient's appropriateness for hospitalization. Medi-Cal denies payment for some inpatient days because of issues with documentation or delays in care. By focusing our efforts on reducing these payment denials for inpatient services rendered, we planned to care for increasing internal medicine (IM) admissions in a cost-effective manner by instituting an academic hospitalist service. The purpose of this study was to analyze and quantify the revenue impact and cost-effectiveness of our new hospitalist service by using historical comparisons with Medi-Cal reimbursement of inpatient IM services at our institution.
We instituted our daytime admitting hospitalist service in December 2007, in response to inpatient admissions that had increased beyond the capacity of our ward teams. Initially, our team consisted of one attending hospitalist who worked approximately 50 hours per week and one nurse practitioner (NP) who worked approximately 40 hours per week. The service admitted four to six patients per weekday, or more as needed. In July 2008, we added one resident physician to the service at no extra cost, by combining the hospitalist rotation with our general internal medicine (GIM) consult rotation. This additional physician allowed for up to eight hospitalist service admissions per weekday. We projected the total annual cost for the first year of the service to be approximately $270,000 for physician and NP salary/benefits (1.2 FTE physicians and 1 FTE NP).
To justify this expenditure, we searched our utilization review database for diagnoses associated with large numbers of Medi-Cal payment denials. We found that, in 2005, Medi-Cal patients admitted with low- and intermediate-risk chest pain stayed for an average of 2.48 days, and Medi-Cal subsequently denied payment for 413 inpatient days (“denied days”). We estimated that, by asking our hospitalist service to focus on optimizing and expediting the care of most of these patients, we could reduce their LOS to less than two full days and could cut in half the number of denied days associated with low-risk chest pain. At an estimated Medi-Cal reimbursement rate of approximately $1,400 per day, such avoidance of 206 denied days per year would reduce our annual loss by about $289,000, which would potentially offset the cost of the hospitalist service. Thus, we tracked the diagnosis of low-risk chest pain, as well as all IM admissions, for denied days, LOS, and Medi-Cal denial rate (i.e., the percentage of days for which Medi-Cal retrospectively denied payment). By using written surveys mailed to randomly selected inpatients after discharge, we also tracked patient satisfaction scores by physician and by service. An outside vendor (Press Ganey Associates, Inc., South Bend, Indiana) created the survey design, administered the survey, and reported the results (including outside comparator groups).
For the first full year of the hospitalist service, we paid approximately $310,000 in salary and benefits to support the service. Those costs were apportioned as follows: one contract hospitalist physician to staff the service for 36 weeks received $150,000; payments to contract and part-time physicians to cover the ward service, which allowed a full-time faculty member to cover the hospitalist service for 16 weeks, cost $40,000; and the salary and benefits for one full-time NP totaled $120,000. Thus, the cost of the service was $40,000 (15%) higher than our initial estimate.
In its first year, the hospitalist service admitted approximately 24 patients per week and carried an average daily census of 14 patients. Thus, our per-day cost of physician/NP care on this service was approximately $60, which compared favorably with our teaching service average of approximately $70 per inpatient day for the same period.
Patients cared for by hospitalist attendings in 2008 had higher satisfaction scores across all physician-specific dimensions than did those in other comparator groups, as detailed below. The aggregate of physician-specific satisfaction scores for the hospitalists' patients (86.1) compares favorably to the aggregate for patients on other medical services (83.5), all inpatients at our facility (83.6), and patients at other U.S. teaching hospitals with 200 to 399 beds (84.9). Patients rated hospitalists' services higher on the dimension of “the [amount of] time [that the] physician spent with you,” despite the hospitalist team's lack of interns and medical students, who are the ones traditionally spending more time at the bedside. Our hospital also tracks Center for Medicare and Medicaid Services quality indicators, but we did not have enough patients in the time period with core measure diagnoses to allow meaningful comparisons across services.
Patients with chest pain
The LOS for Medi-Cal beneficiaries admitted with low- and intermediate-risk chest pain decreased from 2.48 days in 2005 to 1.92 days in 2008. The number of Medi-Cal-denied days associated with this diagnosis fell from 413 in 2005 to 190 in 2008, which represented a decrease of $312,200 in unreimbursed care for the year (i.e., 223 days at approximately $1,400 per day). Total reimbursement per admission of a patient with chest pain did fall somewhat, from $1,948 in 2005 to $1,831 in 2008, but reimbursement per inpatient day for those admissions rose from $787 in 2005 to $955 in 2008 (Table 1). Sixty-nine fewer Medi-Cal beneficiaries with low- and intermediate-risk chest pain were admitted in 2008 than in 2005, and thus aggregate reimbursement for this diagnosis was $36,504 lower than the amount expected on the basis of LOS and denial rates similar to those from 2005. However, the reduction in LOS and denied days in 2008 resulted in 148 fewer unpaid days and, thus, an opportunity for additional reimbursements of up to $207,200 for other Medi-Cal inpatients, on the basis of increased bed availability and increased service throughput, as detailed below.
Inpatient IM services as a whole
Since 2004, the first full calendar year of Accreditation Council for Graduate Medical Education (ACGME) duty hours limits, the total number of patients admitted to IM services at our hospital has risen steadily, from 7,226 in 2004 to 8,643 in 2008. Our average daily inpatient IM census has climbed from 87.9 to 107.1 patients over the same period. Given regulations with respect to team admissions, caps on patient numbers, duty hours, days off, outpatient block rotation and consult rotation requirements,16 and total housestaff FTEs, our maximum admission rate without additional hospitalist services would be 22.7 patients per day, or 8,300 admissions per year; beyond that maximum, the service would be in a state of ever-increasing overflow. Given our overall average LOS of 4.5 days, the corresponding average daily census would be 102.2 patients. In 2007, the final year before initiation of the hospitalist service, IM admitted 8,069 patients, for an average daily census of 97.7 patients. Thus, the addition of hospitalist services in 2008 allowed our IM services to accommodate 574 more admissions than in the previous year, without breaking ACGME rules. This change resulted in an average daily census in 2008 that was 9.4 patients higher than that in 2007; those additional patients could not have been accommodated with the ward team staffing that existed before 2008.
Given both a payer mix that included 54% Medi-Cal coverage and an overall 73% approval rate for per diem payment, approximately $1,870,000 in Medi-Cal revenue for 2008 may be attributed to those additional 9.4 inpatients per day. In addition, the overall Medi-Cal denial rate for all of IM fell from 29.0% in 2007 to 27.4% in 2008, a drop that resulted in reimbursement for 322 additional inpatient days, or an additional $450,800 in revenue for 2008. Therefore, through increased inpatient census and turnover, as well as increased operational efficiency (as illustrated by the analysis of chest pain patients above), our IM service generated 1,653 more paid Medi-Cal inpatient days in 2008, for $2.31 million more in reimbursement than in 2007 (Table 2).
If one attributes the 2007 payer mix and denial rate to existing teaching services (even when they are running at maximum capacity), less than $880,000 in revenue can be ascribed to the higher census on our traditional teaching services. Therefore, an additional $1.43 million in annual revenue is attributable entirely to the work of our hospitalist service. When the physician and NP salary outlays were subtracted, we found that a net income of $1,120,000 accrued to our hospital during 2008 because of the hospitalist service.
The cost-effectiveness of hospitalist-based care in teaching hospitals has been shown in various settings, typically on the basis of reducing LOS and costs in either a per-case (e.g., diagnosis-related group-based)13,14 or a fixed or capitated (e.g., health maintenance organization)17 reimbursement model. We present here a case showing that hiring providers for a hospitalist service focusing on short-stay patients can be cost-effective in a public teaching hospital that relies primarily on per diem, retrospective-review-based payment (in our case, Medi-Cal). There are many nonfinancial advantages to this model as well, not the least of which for our hospital is the management of ever-increasing patient loads in this era of greater restrictions on resident duty hours and patient loads. There are several other advantages as well as potential limitations of this model, which may enhance or limit its generalizability to other institutions. The limitations of our study relate to our volume of inpatient admissions, our cost structure, and our existing teaching services' performance as the historical control group, as discussed below.
The first limitation of our study's widespread applicability is that the main drivers of the higher revenue in our model are the increases in patient throughput and in the total census afforded by the addition of the hospitalist service. We had reached the maximum inpatient volume that our housestaff body could care for under IM Residency Review Committee (RRC) regulations, and we continue to experience an upsurge in demand for inpatient services. Reducing LOS and Medi-Cal payment denials enhances operating efficiency, but revenue will not also be enhanced unless other acute inpatients are available to fill the vacant beds. Thus, hospitals primarily receiving per diem reimbursement should not expect the same fiscal benefits that we experienced from hiring a hospitalist if their IM services are not currently at their limits or if their supply of inpatients is limited or static. In those cases, our model will likely not result in any improvement in revenue.
Compared with figures from previous reports,18,19 our overall cost for physicians' salaries on this service is relatively low, which may also limit the ability of other institutions to duplicate our results. By virtue of our hospital's location in the region with the second-lowest annual mean hospitalist salary18 and its status as a teaching hospital affiliated with a major university, we have been able to hire excellent hospitalist physicians at a very reasonable rate of pay. Although the addition of an NP to the service increases our costs, it also enhances our ability to accommodate fluctuations in daily admissions and helps contribute to efficient care with high patient satisfaction, as others have reported.7,9–11
In considering the overall cost of this new service, we have not included any increases in pay to other health care workers or in allocation of other resources needed to care for additional inpatients on the service. Calculation of the marginal cost of each additional inpatient is complex and beyond the scope of our investigation, but that cost certainly affects the cost–benefit analysis for the hospital as a whole. Thus, total costs associated with hiring hospitalists will differ among hospitals. However, if one takes as inevitable the increased demand for acute inpatient services that are inherent to delivery of health care to a growing population of elderly persons,20 the hospitalist model represents an efficient way to minimize the costs and maximize the revenues associated with such changes.
Our study's generalizability may also be limited by the fact that operational efficiency in terms of LOS and payment denials, if any, may already be at optimal levels in other health care systems. Although we had made substantial improvements in both areas before the arrival of our hospitalists, we found that our hospital made further improvements along two fronts. First, as previously discussed, our hospitalists were far more efficient than had been the regular IM staff in caring for patients with short LOS, as exemplified by the low- and intermediate-risk chest pain patients. It is very difficult for a traditional teaching service to turn patients around quickly on a regular basis, given that there are multiple layers of providers, irregular schedules and patient loads, and great variability among attendings' practices. At the same time, the hospitalist service admits patients daily, enjoys a fairly regular schedule, and is optimal for ensuring that patients are able to be discharged in one day if clinically indicated. Second, the effect of offloading teaching services that were previously at (or beyond) full capacity reduces “holdovers” and improves the ability of teams to see their admitted patients more quickly, which ensures timely attending staffing and improves overall efficiency of care. Thus, our use of a historical control group rather than a prospective cohort may result in an overestimation of the effect of introducing a hospitalist program at other institutions. For example, hospitals that are not already at maximum IM service capacity or that have fewer short-stay admissions may not derive as much benefit as we did.
The advantages of instituting a hospitalist service have reached beyond financial considerations. Our hospitalists help to enhance patient flow, provide backup for bedside procedures on the inpatient service, screen transfer requests, participate in patient safety and performance improvement activities, and teach our housestaff and students on the inpatient services. The hospitalists (and NPs) also increase patient satisfaction, as observed by analysis of the results of our patient satisfaction survey. We were fortunate to be able to combine our GIM consult service with the hospitalist service to give housestaff a firsthand experience of that model of care and to allow them exposure to the increasingly popular career choice of hospitalist medicine. We are hopeful that we will be able to deepen this experience for our residents in the future, as we look to expand hospitalist services in response to continually increasing patient loads.
Establishing a hospitalist service was a cost-effective and efficient intervention for our hospital as we faced increases in patient load beyond RRC limits for housestaff. Although this result has also been observed in other locations, our findings are unique in that we rely almost solely on per diem reimbursement and thus do not directly benefit from reducing LOS. We believe that public teaching hospitals that share similar challenges in their patient census and reimbursement structures can benefit from adding hospitalists to their inpatient IM service.
The authors wish to acknowledge William Loos, MD, and Melinda Anderson for their administrative support and guidance in launching the hospitalist program.
The institutional review board of Olive View–UCLA Medical Center approved the study as exempt from full review.
The opinions expressed in this report are those of the authors and are not intended to be representative of the Los Angeles County Department of Health Services or the University of California.
1 Bion JF, Heffner JE. Challenges in the care of the acutely ill. Lancet. 2004;363:970–977.
2 Thorpe KE. House staff supervision and work hours: Implications of regulatory change in New York State. JAMA. 1990;263:3177–3181.
3 Weinstein DF. Duty hours for resident physicians: Tough choices for teaching hospitals. N Engl J Med. 2002;347:1296–1302.
4 Charap M. Reducing resident work hours: Unproven assumptions and unforeseen outcomes. Ann Intern Med. 2004;140:814–815.
5 Nuckols TK, Bhattacharya J, Wolman DM, Ulmer C, Escarce JJ. Cost implications of reduced work hours and workloads for resident physicians. N Engl J Med. 2009;360:2202–2215.
6 Wachter RM, Flanders S. The hospitalist movement and the future of academic general internal medicine. J Gen Intern Med. 1998;13:783–785.
7 Goksel D, Harrison CJ, Morrison RE, Miller ST. Description of a nurse practitioner inpatient service in a public teaching hospital. J Gen Intern Med. 1993;8:29–30.
8 Howie JN, Erickson M. Acute care nurse practitioners: Creating and implementing a model of care for an inpatient general medical service. Am J Crit Care. 2002;11:448–458.
9 Rudy EB, Davidson LF, Daly B, et al. Care activities and outcomes of patients cared for by acute care nurse practitioners, physician assistants, and resident physicians: A comparison. Am J Crit Care. 1998;7:267–281.
10 Genet CA, Brennan PF, Ibbotson-Wolff S, et al. Nurse practitioners in a teaching hospital. Nurse Pract. 1995;20:47–54.
11 Stetler C, Effken J, Frigon L, Tiernan C, Zwingman-Bagley C. Utilization-focused evaluation of acute care nurse practitioner role. Outcomes Manag Nurs Pract. 1998;2:152–160.
12 Diamond HS, Goldberg E, Janosky JE. The effect of full-time faculty hospitalists on the efficiency of care at a community teaching hospital. Ann Intern Med. 1998;129:197–203.
13 Brown MD, Halpert A, McKean S, Sussman A, Dzau VJ. Assessing the value of hospitalists to academic health centers: Brigham and Women's Hospital and Harvard Medical School. Am J Med. 1999;106:134–137.
14 Gregory D, Baigelman W, Wilson IB. Hospital economics of the hospitalist. Health Serv Res. 2003;38:905–918.
15 Rifkin WD, Holmboe E, Scherer H, Sierra H. Comparison of hospitalists and nonhospitalists in inpatient length of stay adjusting for patient and physician characteristics. J Gen Intern Med. 2004;19:1127–1132.
17 Craig DE, Hartka L, Likosky WH, Caplan WM, Litsky P, Smithey J. Implementation of a hospitalist system in a large health maintenance organization: The Kaiser Permanente experience. Ann Intern Med. 1999;130:355–359.
20 Etzioni DA, Liu JH, Maggard MA, Ko CY. The aging population and its impact on the surgery workforce. Ann Surg. 2003;238:170–177.