Length of stay has minimal impact on the cost of hospital admission¹

Article Outline

• Abstract
• Methods
• Results
• Discussion
• References
• Copyright

Abstract 

Background: Hospital cost containment, cost reduction, and alternative care delivery systems continue to preoccupy health care providers, payers, employers, and policy makers throughout the United States. The universal metric for gauging the success of these efforts is hospital length of stay (LOS). Reducing the LOS purportedly yields large cost savings. The purpose of this study is to assess precisely how much hospitals save by shortening LOS.

Study Design: We reviewed the cost-accounting records of all surviving patients (n = 12,365) discharged from our academic medical center during fiscal year 1998 with LOS of 4 days or more. Actual costs were identified through the University of Michigan cost-accounting system. Individual patient costs were broken out on a daily basis and then decomposed further into variable direct, fixed direct, and indirect categories. The population was analyzed by determining the incremental resource cost of the last full day of stay versus the total cost for the entire stay. The data were also stratified by LOS and by surgical costs. An analysis of all trauma patients was then performed on all patients discharged from the hospital’s adult level I trauma center (n = 665). Costs were determined on specific days, including admission day, each ICU day, day of discharge from the ICU, and each of the last 2 days before the discharge day.

Results: The incremental costs incurred by patients on their last full day of hospital stay were $420 per day on average, or just 2.4% of the $17,734 mean total cost of stay for all 12,365 patients. Mean end-of-stay costs represented only a slightly higher percentage of total costs when LOS was short (eg, 6.8% for patients with LOS of 4 days). Even when the data were stratified to focus on patients without major operations, the $432 average last-day variable direct cost was only 3.4% of the $12,631 average total cost of care. A focus on the trauma center helps to explain this phenomenon. For our trauma center, variable direct costs accounted for 42% of the mean total cost per patient of $22,067. The remaining 58% was hospital overhead (fixed and indirect costs). The median variable direct cost on the first day of admission is $1,246, and the median variable direct cost on discharge is $304. Approximately 40% of the variable costs are incurred during the first 3 days of admission.

Conclusions: For most patients, the costs directly attributable to the last day of a hospital stay are an economically insignificant component of total costs. Reducing LOS by as much as 1 full day reduces the total cost of care on average by 3% or less. Going forward, physicians and administrators must deemphasize LOS and focus instead on process changes that better use capacity and alter care delivery during the early stages of admission, when resource consumption is most intense.

The American health care system is hemorrhaging red ink at nearly every level of health care delivery, including hospitals, outpatient providers, insurers, and integrated health care systems. Meanwhile, physicians earn less, employers pay higher premiums, and patients and their political representatives express dissatisfaction with the service they receive and the underlying quality of their care. Two decades of unrelenting pressure to deliver more cost-effective care have put enormous strains on the system while yielding only limited benefits. What has gone wrong?

In the ongoing reform effort, much attention and many resources have gone toward encouraging physicians and providers to shift care as much as possible away from costly inpatient hospital stays toward less expensive outpatient treatment. Among the most important metrics for gauging the success of this endeavor is the hospital length of stay (LOS). The premise has been that by discharging patients more quickly, hospitals reduce overall health care costs, even if patients continue to receive care on an outpatient basis, because such care is assumed to be less expensive.1

Because of the tremendous attention LOS has received, this premise represents an obvious starting point for inquiries into the shortcomings of health care reform. Yet the only evidence shedding any light on the empiric link between reduced LOS and cost savings is indirect, and it suggests that the link is weak. Hospitals are highly capital-intensive enterprises, and many studies have documented the high administrative and overhead costs of running them.1, 2, 3, 4, 5, 6, 7, 8 The incremental (or “marginal”) cost of admitting one more patient to a hospital may be much less than the total cost, which includes overhead. One recent study has shown that in the short run, the marginal cost of admitting one more patient may be as little as 13% of the total cost,8 and another report has shown that the marginal cost of an emergency room visit is only a small percentage of the total cost.6 These studies call into question the wisdom behind hospitals’ often Herculean efforts to keep patients out and to strictly ration their care once they are admitted.

Figures are needed that show precisely how much hospitals can save by discharging patients sooner. There is some speculation that the marginal cost of keeping a patient in the hospital 1 additional day is much less than the total cost,7 but there are no hard facts published anywhere and only limited grounds for educated guesses. There are good reasons, though, to suspect that the marginal cost of 1 more day may be remarkably small. Suppose that the incremental cost of a typical hospital admission is approximately 40% or less of the total cost and that most patients incur many more expenses early in their hospital stays than they do at the end. It follows that the incremental costs incurred during the last 10% of the average hospital stay must be well below 4%. But how much less?

Most of the progress that providers have made in reducing LOS occurred during the mid-1980s, and since then LOS has declined little despite a continued emphasis on that objective.1 If further LOS reductions would yield few cost savings and could not be accomplished without jeopardizing the quality of patient care, then it is time for providers to look for other avenues to save money. The purpose of our study is to evaluate the pervasive belief that health care costs can be curtailed significantly by reducing hospital LOS.

Back to Article Outline

Methods 

We reviewed the cost-accounting records of all surviving patients discharged from our hospital during fiscal year 1998 with LOS of 4 days or more (n = 12,365). Costs were measured using the University of Michigan’s general ledger and activity-based cost-accounting system, which is a conventional framework that uses software from market leader TSI (Eclipsys; Delray Beach, FL). Patient costs were tracked on a daily basis and further decomposed into three categories. Variable direct costs measure expenditures that can be identified directly with the care of individual patients on a particular day, such as laboratory tests, radiographs, and disposable supplies. Fixed direct costs capture expenditures that can be identified with a specific hospital department but not with a particular patient, and indirect costs lie entirely outside individual departments. Examples of fixed direct costs include equipment and medical devices used to care for trauma patients. Examples of indirect costs include the admissions area and the chief executive officer’s salary. Taken together, fixed direct and indirect costs are collectively referred to as “hospital overhead.”

Indirect overhead costs are allocated as follows. After being identified as such, they are grouped into pools of like funds (eg, buildings, facilities, administrative). Hospital finance proceeds to determine an allocation metric (eg, square footage in the case of buildings) and the direct departments designated to receive this overhead. The indirect costs are then allocated to individual departments by a simultaneous equations method, so that the order of allocation is immaterial. Once the departmental allocation is determined, that pool of expenses is allocated to specific patient-billed services based upon direct costs. In the chemical laboratory, for example, if the direct cost for a glucose test is $1 and the direct cost for a liver enzyme test is $6, then the liver enzyme service is assigned six times more overhead expense than the glucose. This method of activity-based cost accounting is conventional.

This study scrutinizes the physician-controllable variable direct costs per patient because in the short run, overhead is beyond physicians’ immediate control. Unlike the Cook County Hospital analysis,8 our study categorizes nursing as a variable direct cost. We include nursing among the categories that physicians can influence. Because nursing constitutes the majority of end-of-stay variable direct costs, treating nursing expenditures as a fixed cost (which they may be over the short term) would effectively halve the results that follow.

Individual patient costs were broken out on a daily basis. We omitted all patients who died or were discharged against medical advice (on the grounds that physicians did not determine the timing of these patients’ discharges) and those patients with LOS of 3 days or less. The population that remained was then stratified by LOS and subsequently divided into subgroups of patients who had major surgery and patients who did not.

To gain a more focused perspective, we also examined all 665 patients discharged from the hospital’s level I trauma center. Within this group, nine activities together accounted for more than 99% of total costs: nursing (42.8% of total cost), surgical services (12.7%), laboratory (9.4%), radiology (8.4%), pharmacy (8.3%), emergency services (7.9%), respiratory and pulmonary (5.1%), rehabilitation services (3.3%), and supplies (1.3%). The entire patient population was first analyzed by looking at the first 3 days and the last 3 days of the patients’ hospital stays. We then dropped patients who died or were discharged against medical advice. Within this group, separate analyses were conducted for patients who spent at least 7 days in the hospital and had a minimum of 3 days in the ICU, and for patients with LOS of 4 days or more.

Statistical analyses were performed using the two-tailed Student’s t-test with p values denoting levels of statistical significance (p < 0.05). Mean ± SD is reported.

Back to Article Outline

Results 

The University of Michigan Medical Center discharged 29,036 patients during fiscal year 1998. Table 1 provides descriptive statistics on all surviving patients discharged during this period who had LOS of 4 days or more and were not discharged against medical advice (n = 12,365). Patients who fit this description had mean and median LOS of 10.5 days and 7 days, respectively. Mean (± SEM) and median total costs per case were $17,734 ± $229 and $10,289, respectively, and the mean (± SEM) and median variable direct costs of the last full day before discharge were $420 ± $7 and $310. These results show that hospital-wide, the last full day before discharge represented 9.5% of the mean 10.5-day LOS, and yet only 2.4% of the total cost of care can be identified specifically with this interval. Table 1 also stratifies patients by LOS. In all cases, last-day resource consumption was remarkably light; mean variable direct costs incurred on that day were $458 or less. For LOS of 4 days, for example, the last full day represented 25% of the total LOS, and yet only 6.8% of the total cost of care was attributable to this period.

Table 1. Costs for Surviving Patients for the Hospital at Largelegend, legend

Table 2 examines subgroups of this population who had major surgery (n = 4,163) and those who did not (n = 7,195). The purpose of this stratification is to determine whether the results in Table 1 might be driven primarily by patients with large up-front expenses followed by long periods of convalescence (with low attendant resource costs). The left-hand columns of Table 2 describe patients who incurred $2,000 or more in surgical costs during their hospital stays; the right-hand columns include those patients who incurred less than $1,000 in such costs. There were relatively few patients (n = 1,007) who had total surgical costs in the $1,000 to $2,000 range, and so this group was omitted to ensure a stark distinction between the categories. By design, this stratification isolates patients who likely had large expenditures early in their hospital stays and juxtaposes them against lower-cost patients who could conceivably consume hospital resources more uniformly over the entire LOS. As one might expect, cases involving major surgery entailed higher mean total costs. The $36 difference in the last-day variable direct costs ($396 for patients with major surgery and $432 for patients without major surgery), although statistically significant (p = 0.020 by the two-tailed Student’s t-test), is small. Consequently, variable direct costs incurred on the last full day before discharge constituted only 1.5% of the $26,547 average total cost for patients having major surgery, compared with 3.4% of the $12,631 average total cost for patients not having major surgery. The stratification by LOS in Table 2 reinforces the results in Table 1. What is striking is that last-day costs as a percentage of total costs are remarkably small even for patients who undergo no major surgery and have LOS of 4 or 5 days.

Table 2. Comparison of Patients With Major Surgery and Patients Without Major Surgerylegend, legend

To gain perspective and detail, we conducted a subgroup analysis on all patients discharged from our level I trauma center during this period (n = 665). Table 3 provides descriptive statistics on this population. The table indicates that the $9,177 mean variable direct cost per patient is only 42% of the mean total cost per patient of $22,067. The remaining 58% is hospital overhead. The initial analysis of the population revealed that 46 patients died (6.9%) and 4 were discharged against medical advice (0.6%). There were 287 patients (43.2%) who spent at least 1 day in the ICU and 468 patients (70.4%) who stayed in the hospital 3 days or more. For the entire population, mean and median LOS were 8.3 days and 5 days, respectively. Total variable direct costs were $6,102,551 and total costs were $14,674,355. In aggregate, 19.4% of the total variable direct costs were incurred on the first day of hospitalization, 11.1% were incurred on the second day, and 7.7% were assessed on the third day. On our trauma service, 57.6% of all variable direct costs were incurred within the patients’ first 7 days of stay. The mean and median variable direct costs and total costs per patient for the first 3 days are identified in Table 4.

Table 3. Descriptive Statistics for the Trauma Service in Fiscal Year 1998 (n = 665) legend

Table 4. Variable Direct Costs Per Day for the First 3 Days of Trauma Hospitalization

Table 5 identifies the same figures for the last 3 days of these patients’ hospital stays. A comparison of the two tables reveals that these last 3 days are much less expensive than the first few days. For example, the median variable direct cost of $1,246 on the first day of admission is more than four times the $304 median variable direct cost on the day of discharge.

Table 5. Variable Direct Costs Per Day for the Last 3 Days of Trauma Hospitalization

For various reasons, the $614 mean variable direct cost incurred on the discharge day overstates the true cost of the last day of a patient’s hospital stay. First, this figure includes the unusually high costs incurred by the 46 nonsurvivors in the population. The left side of Table 6 omits these 46 patients, along with 4 patients who were discharged against medical advice. Second, this $614 figure includes 242 patients who stayed in the hospital for 3 days or less. For these patients, the costs incurred in the last 3 days are also costs incurred in the first 3 days. To isolate the costs incurred at the end of the patients’ hospital stays, the right-hand side of Table 6 also omits these 242 patients (along with the 50 patients who were omitted in the left-side calculations). These figures identify the end-of-stay variable direct costs of what one might consider a “representative” surviving trauma-service patient in fiscal year 1998. The mean variable direct costs of the last 3 days of stay were $487, $396, and $333, and the median costs were $334, $278, and $257, respectively.

Table 6. End-of-Stay Variable Direct Costs Per Day for Trauma Subgroups legend

Further refinement is possible by stratifying this population. Table 7 represents patients who spent at least 3 days in the ICU and 7 days in the hospital. These are the most seriously injured patients, and the table shows that these patients were costly to care for even after they were discharged from the ICU. The first 3 post-ICU days had mean variable direct costs of $1,017, $1,054, and $894, respectively. Yet by the time they were discharged, these patients’ variable direct costs of care fell below $500 per day, as was the case for all other subgroups analyzed.

Table 7. Surviving Trauma Patients with 3 or More ICU Days and Length of Stay of 7 Days or More legend

Back to Article Outline

Discussion 

In recent years, physicians and administrators have worked tirelessly to reduce hospital stays with the understanding that LOS is a good surrogate for cost. The thinking has been that lopping 1 day off of the end of an inpatient stay reduces the overall cost by the mean cost of a day in the hospital. Unfortunately, this widely held perception is incorrect. The reason is that not all hospital days are economically equivalent. Table 1, Table 2, Table 3, Table 4, Table 5, Table 6, Table 7 show a reliable rule of thumb, at least for our hospital and our trauma service: On average, the last full day of hospitalization involves incremental resource costs in the range of just $400 to $450. Cutting this day would save the hospital only a small percentage of the total cost of care.

The basic truth is that the bulk of health care expenses take the form of overhead, or they are incurred early in patients’ hospital stays. This comes as no surprise to practicing physicians who recognize that the early phase of care involves expensive diagnosis and intervention, while the final days are essentially recuperative.

Yet, for four reasons, even these figures overstate the impact that reductions in LOS have on health care costs. Suppose, for example, that our trauma service set out to discharge every eligible patient on average 10% sooner than it does now. Would it save even 2% to 3% of total costs? The answer is almost certainly not. First, 44% of the service’s 665 patients in fiscal year 1998 simply could not have been discharged much faster, either because they died (n = 46), they were discharged against medical advice (n = 4), or they already had stays of 3 days or less (n = 242). Second, to achieve this 10% LOS reduction, the trauma service would likely focus on patients who were relatively healthy. These likely candidates for reduced LOS are also the cheapest patients to retain in the hospital. Cutting their LOS does relatively little to reduce aggregate costs. Third, discharging patients 10% faster is not equivalent to eliminating the attendant end-of-stay costs. To accomplish the 10% LOS reduction, some treatments (eg, laboratory tests) must be accelerated and other treatments (eg, pharmaceuticals) must be continued on an outpatient basis. These costs would be shifted rather than eliminated. This is especially true when patients are discharged to other facilities or to home health care.

Fourth and finally, in this study all expenses surrounding nurses directly involved in patient care were categorized as variable direct costs. This approach presumes that staff can be adjusted quickly as patient activity ebbs and flows. This categorization is appropriate if the hospital uses substantial amounts of nursing overtime (which is the case at the University of Michigan) or if there is considerable nursing turnover. But if nurses are salaried or if the time horizon for decision making is very short (meaning that the hospital is committed to its staffing levels), then it may be more appropriate to treat nursing as a fixed cost.8 Because nursing represents the majority of end-of-stay costs, shifting even a portion of these expenditures out of the category of variable direct costs would have a profound impact on the results reported here.

Our analysis raises the following question: Does LOS truly matter? Although our university hospital has a tertiary referral-based population, we strongly believe that our results generalize to most hospital settings. Within most hospitals and health systems, most costs take the form of overhead charges, and the preponderance of variable costs are incurred very early in patients’ hospital stays. Put differently, our results would not generalize only if our higher costs of care stem strictly from significantly greater overhead rather than higher costs across the board, or if we are unique among hospitals in paring end-of-stay variable direct costs. We have no reason to suspect that this is the case.

It is important to note that this analysis presumes that the hospital has excess capacity, so that retaining any given patient in the hospital does not preclude admitting other patients. Hospitals facing capacity constraints encounter what economists call “opportunity costs,” meaning that those hospitals would have to forgo the opportunity to care for new patients to keep existing patients in house longer. In the event of such capacity constraints, hospitals should explicitly factor in these opportunity costs. From a revenue perspective, when the health system is at capacity, it is more effective to take in new “high-revenue” admissions by discharging “low-revenue” patients who are in the convalescent phase of their recovery. If there are capacity constraints, the case for reducing LOS may become much more compelling, both financially and medically. One important limitation of this study is that it does not explicitly factor opportunity costs (which are notoriously difficult to measure) into the analysis.

Our analysis also depends upon the accuracy and integrity of the University of Michigan’s cost-accounting system, and this is the most important limitation of the study. The most immediate concern involves the categorization of costs into the variable direct category. Whenever costs are difficult to allocate to individual patients, hospitals (like other businesses) must count them instead as overhead, even though they might actually vary directly with patient activity. Hospital finance has identified for us the three activities where expenses are allocated to the indirect-cost component even though they arguably belong in the category of variable direct costs: dietetic services ($5.7 million total cost in fiscal year 1998), housecleaning ($8.9 million), and transcription ($11.2 million). These costs are substantial, but they are not an especially large percentage (8.6%) of the $301 million in total costs of running the hospital. More important, none of these activities is likely to involve much in terms of end-of-stay costs. Finally, we believe that the University of Michigan is as aggressive as any hospital in the country in tagging costs to individual patients (which readers can confirm by asking their own administrators whether they include dietetic services, housecleaning, and transcription in their variable direct costs). As such, replicating this study at other institutions would likely yield results that, if anything, are starker than those reported here. In short, we have confidence in our hospital finance department and the accounting mechanisms they have in place.

It is also important to remain aware that from the payers’ perspective, the savings from a reduced LOS depend upon that payer’s contractual relationship with the health system. In a fixed-fee reimbursement system, the payer has no strong financial stake in a patient’s LOS. In a traditional fee-for-service system, however, reimbursement is a function of charges, which typically involve a markup over total costs. The implication is that the payers’ savings from reducing LOS by 1 day could amount to $1,000 or more. In short, although the incremental resource cost of retaining a typical patient for 1 additional day is $400 to $450, fee-for-service payers may reimburse several times this amount.7 They are understandably eager for hospitals to discharge patients quickly.

Historically, LOS may have been a relatively simple and useful surrogate for costs. It was easily obtainable, difficult to manipulate, and directly comparable across institutions. Few hospitals had accounting systems that allowed them to monitor their actual costs, and although physicians had little control over overhead and saw little or no professional incentive to pursue what are typically perceived as administrative activities, they could work successfully to reduce LOS. One way to interpret this study’s results is that the success of these endeavors has left little room for further economically significant reductions in LOS. Much of the return has been achieved. Focusing substantial physician effort on further reductions in hospital LOS will yield little value.

If LOS is no longer a source of additional cost reductions, where should physicians focus their efforts? It is critical that physicians play more active roles in two key areas. First, they must work with hospital administrators to make better use of hospital capacity and overhead, which account for the majority of the inpatient hospital costs. Such efforts must include some combination of capacity reductions; optimal use of existing capacity; and a shift in patient activities, whenever possible, to off-peak periods when beds, operating rooms, and other key facilities have historically sat idle. These efforts should also include physician participation in efforts toward networking, including building better relationships with referring physicians at other institutions. Recruiting new patients can help to control costs by enabling hospitals to amortize their overhead over a larger population. Second, physicians must work to reduce costs in the early stages of their patients’ care. Costs disproportionately arise early rather than late in any given hospital stay, and measures that curtail these costs will have economically important effects.

For their part, hospital administrators must keep their physicians well informed and provide them with incentives, including professional rewards, to encourage them to participate in these activities. Together, physicians and administrators must optimize health system strategy, operations management, payer contract negotiations, and health system finance.

Back to Article Outline

References 

1. Schwartz WB, Mendelson DN. Hospital cost containment in the 1980s (hard lessons learned and prospects for the 1990s). N Engl J Med. 1991;324:1037–1042
   • View In Article
   • MEDLINE
   • CrossRef
2. Schwartz WB, Mendelson DN. Eliminating waste and inefficiency can do little to contain costs. Health Aff (Millwood). 1994;13:224–238
   • View In Article
   • MEDLINE
   • CrossRef
3. Taheri PA, Wahl WL, Butz DA, et al.  Trauma service cost (the real story). Ann Surg. 1998;227:720–725
   • View In Article
   • MEDLINE
   • CrossRef
4. Taheri PA, Butz DA, Watts CM, et al.  Trauma services (a profit center?). J Am Coll Surg. 1999;188:349–354
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (158 KB)
   • CrossRef
5. Taheri PA, Butz DA, Greenfield LJ. How patient complexity affects costs and profit margins. Ann Surg 2000;231(6):849–859.
   • View In Article
6. Williams RM. The costs of visits to the emergency departments. N Engl J Med. 1996;334:642–646
   • View In Article
   • MEDLINE
   • CrossRef
7. Reinhardt UE. Spending more through “cost control” (our obsessive quest to gut the hospital). Health Aff (Millwood). 1996;15:145–154
   • View In Article
   • MEDLINE
   • CrossRef
8. Roberts RR, Frutos PW, Ciavarella GG, et al.  Distribution of variable vs fixed costs of hospital care. JAMA. 1999;281:644–649
   • View In Article
   • MEDLINE
   • CrossRef

• 1 No competing interests declared.