William O. Cleverley
James O. Cleverley
Using a new measure of hospital volume that is more realistic than other metrics commonly used today, healthcare finance leaders can better compare activity and costs-and identify savings potential.
At a Glance
- Equivalent patient units is a more reliable measure of a hospital's patient volume than adjusted discharges or adjusted patient days because it better accounts for both inpatient and outpatient volumes.
- Three elements are required to calculate equivalent patient units: equivalent discharges, equivalent visits, and the payment ratio.
- All of these elements are available through publicly available data, making it possible for hospitals to immediately adopt this new metric and, thereby, better understand their potential for savings.
For at least the past 50 years, the predominant measures of hospital volume have been either adjusted patient days or adjusted discharges. The computation of either metric involves dividing outpatient charges by the average inpatient charge for either patient days or discharges.
For example:
Adjusted Discharges = Discharges + Discharges x (Outpatient Charges /Inpatient Charges)
Healthcare finance leaders and policy analysts regularly express displeasure with these two volume metrics. When inpatient charges represented 90 percent of total volumes, using either one made perfect sense. However, outpatient volumes have continued to trend upward. In 2009, inpatient charges accounted for only about 50 percent of hospital volumes. Yet for lack of an acceptable alternative, adjusted patient days and adjusted discharges remain the most commonly used volume metrics.
We propose a better metric: equivalent patient units. Cost per equivalent patient unit appears to be a far more reliable measure than cost per adjusted discharge or adjusted patient days. Benchmarking costs is a crucial exercise for healthcare finance leaders looking for opportunities to reduce expenses in response to declining reimbursements. However, accurate benchmarking requires accurate metrics. In making the case that equivalent patient units is a more reliable metric than either adjusted patient days or adjusted discharges, we need refer only to adjusted discharges, because adjusted discharges seems to be the primary metric currently employed and because the same argument applies equally to both of the currently predominant metrics.
What's Wrong with Adjusted Discharges?
The major increase in outpatient activity in recent decades does not, in and of itself, make adjusted discharges an invalid metric of overall hospital activity. The bigger problem: The metric has become an unreliable measure for benchmarking volumes and costs.
The primary purpose for the adjusted discharge metric is to permit some type of comparative analysis, usually looking at cost or revenue divided by adjusted discharges for one or multiple hospitals covering one or multiple time periods. For example, the analysis could compare changes in cost from one year to the next at one facility-or it could compare the cost per adjusted discharge across multiple hospitals within a given time period. For several reasons, however, adjusted discharges no longer works well as a comparative metric.
Pricing effects. Ratios of cost to charge (RCCs) are not uniform across inpatient and outpatient procedures and are changing at different rates. Changing RCCs-resulting from pricing strategies-create a strong bias in the measurement of adjusted discharges. The exhibit below illustrates this pricing impact. In the example, the only change is an overall price increase of 10 percent, but the entire price increase is allocated to outpatient procedures. The result: A 10 percent increase in adjusted discharges (from 20 to 22).
Exhibit 1
Historical RCC data for acute care hospitals suggest that charge increases are being shifted from inpatient to outpatient and ancillary areas, as shown in the exhibit below. Many hospitals are raising outpatient procedure prices at rates greater than inpatient procedure prices because of the greater incidence of discount-from-billed charge payments in the outpatient arena. Given no other changes, this shift creates an increase in the number of adjusted discharges, which would overstate the actual increase in facility volume.
Exhibit 2
Further complicating the application of adjusted discharges, every hospital has experienced different rates of change for inpatient and outpatient RCCs, which makes benchmarking among hospitals unreliable.
Case mix complexity. Many hospitals modify the adjusted discharge metric to recognize case mix complexity by multiplying the adjusted discharge value by the hospital's inpatient case mix index (CMI). (An alternative is dividing a cost-per-adjusted discharge metric by the hospital's CMI.)
This approach may be fine for the inpatient component, but it includes an implicit assumption that outpatient complexity is equivalent, which means, for example, that a hospital with an inpatient CMI of 2.0 would have an outpatient case mix complexity factor also about double the average value. Our research suggests that this assumption is incorrect (Cleverley, W.O., Cleverley, J.O., and LaFortune, J.M., State of the Hospital Industry-2010 Edition). It shows that larger hospitals often have higher inpatient CMI values than do smaller hospitals, but the relative degree of outpatient complexity is not of the same magnitude. As a result, larger hospitals would appear to have more adjusted discharges (and lower cost-per-adjusted discharge) than smaller hospitals when inpatient case mix adjustments are included.
A Possible Solution: Equivalent Patient Units
Creating a hospitalwide output measure requires a method for adding inpatient volumes to outpatient volumes. There are two widely accepted case-mix-adjusted volume measures available for both inpatient and outpatient areas.
Equivalent discharges for inpatient areas. By using Medicare-severity diagnosis-related group (MS-DRG) case weights, we can determine that a hospital with 10 discharges and an average CMI of 2.0 has 20 equivalent discharges.
Equivalent visits for outpatient areas. By using relative weights (RWs), we can determine that a hospital with 500 outpatient visits and an average RW of 3.0 has 1,500 equivalent visits.
These two metrics can be combined to get equivalent patient units. How? The answer may be found in Medicare payment rates. For example, let's assume that Medicare pays a hospital $5,000 for an inpatient discharge with a CMI of 1.0, and $50 for an outpatient visit with a RW of 1.0. With this information, we can take the relative payment ratio of outpatient to inpatient payment of 0.01 ($50/$5,000) and create an equivalent patient unit:
Equivalent Patient Units = Equivalent Discharges + (Payment Ratio X Equivalent Visits)
For example, if a hospital has 15 equivalent discharges and 1,500 equivalent visits, we can create the equivalent patient units as follows:
15 + (0.01 x 1,500) = 30
The concept appears to make sense. Assuming that Medicare is trying to relate payment to underlying reasonable cost, which is a stated objective, the use of relative Medicare payment rates may be valid. It is not perfect, but it should represent a major improvement from the adjusted discharge measure of hospital output.
Case Example
As is often stated, "The proof of the pudding is the eating." So we will demonstrate the use of equivalent patient units with a real-world example constructed using publicly available data for two not-for-profit hospitals in the same geographic area (see the exhibit below). To illustrate differences, we will compute two overall, facilitywide cost metrics: cost per adjusted discharge, case mix adjusted, and cost per equivalent patient unit (which reflects inpatient and outpatient case intensity adjustment).
Exhibit 3
Calculating cost per adjusted discharge, case mix adjusted. Six values are required to compute this metric, as the exhibit shows. The case mix adjusted discharge measure is based on the multiplication of adjusted discharges by the Medicare CMI. For example, for Hospital A in the exhibit, the cost per adjusted discharge, CMI adjusted, is calculated as follows:
Operating Expenses ($768,986,760) /Adjusted Discharges (34,676) x Medicare CMI (1.8553) = $11,953
The advantage of using a Medicare CMI is simply its availability. The Medicare CMI can be used for virtually every U.S. acute care hospital. The drawback to the Medicare CMI is that it does not
adequately reflect patients in nontraditional Medicare areas (primarily obstetrics). If a hospital has a large maternity area, its Medicare CMI would likely be overstated when compared with its all-payer CMI. As discussed earlier, however, the application of any CMI metric to adjusted discharges poses benchmarking challenges. We now turn our attention to equivalent patient units to see if the metric can truly overcome the pitfalls of adjusted discharges.
Calculating cost per equivalent patient unit. As our formula illustrates, there are three requirements to calculating equivalent patient units: equivalent discharges, equivalent visits, and the payment ratio. All of these elements are available through publicly available data, but they must all be calculated.
The number of equivalent discharges is calculated by dividing total gross inpatient charges by the hospital's average Medicare inpatient charge per discharge (CMI adjusted). As shown in the exhibit, this calculation yields 65,375 equivalent inpatient discharges for Hospital A.
Ideally, we would use the all-payer average charge and CMI; however, these values are not publicly available, and the Medicare values do not pose significant issues. Although the average charge for a Medicare patient is likely to differ from that for a non-Medicare patient (the charge for the Medicare patient will likely be higher for the same MS-DRG), this relationship would exist uniformly across all hospitals. Moreover, although the Medicare data do not adequately reflect nontraditional Medicare areas (obstetrics, as discussed), the price-to-weight relationship is the sole focus in the equivalent discharge methodology. We are simply concerned with an average charge per case mix of 1.0-a value that should show little variation whether it is based on a large or small volume of obstetric cases (provided the mark up across inpatient areas is similar).
Equivalent outpatient visits are calculated by dividing total gross outpatient charges by the hospital's Medicare outpatient charge per visit (relative weight adjusted). Thus, for Hospital A, total gross outpatient charges of $601,621,000 were divided by $211.70 to yield 2,841,919 equivalent outpatient visits. Again, a primary advantage of the equivalent patient unit methodology is the application of an outpatient-specific case mix (relative weight) to outpatient-specific volume-something the adjusted discharge formula does not permit.
To convert equivalent visits to equivalent patient units, it is first necessary to calculate the relative payment ratio. This step is accomplished by taking the ratio of Medicare outpatient payment per relative weight of 1.0 to Medicare inpatient payment per case weight of 1.0 for each hospital (obviously, this number can vary by facility). For Hospital A, the average Medicare payment for an inpatient case with a case mix weight of 1.0 was $8,378, and the average payment for a Medicare outpatient visit with a relative weight of 1.0 was $64.89. Thus, the payment ratio for Hospital A was calculated as follows:
Payment Ratio = $64.89 / $8,378 = 0.007745
Equivalent patient units is calculated as follows:
Equivalent Patient Units = Equivalent Inpatient Discharges + (Payment Ratio x Equivalent Visits)
This calculation for Hospital A was as follows:
Equivalent Patient Units = 65,375 + (0.007745 x 2,841,919) = 87,386
Cost per equivalent patient unit is calculated using the following formula:
Cost per Equivalent Patient Unit = Operating Expenses / Equivalent Patient Units
For Hospital A, this calculation was as follows:
Cost per Equivalent Patient Unit = $768,986,760 / 87,386 = $8,800
Comparing the two metrics. When reviewing values for these metrics, remember that the scale between them is not the same: 20,000 case mix adjusted discharges is not the same as 20,000 equivalent patient units. It is their relationship across facilities and over time that is of critical importance for benchmarking.
The relative cost between the two case example hospitals should be similar-if the cost metrics were similar in the methodology. For example, is there a high degree of correlation between cost per adjusted discharge and cost per equivalent patient unit? Does one measure suggest that a hospital is high cost relative to an industry average, while the second measure suggests that it is low cost?
If we noticed similar relative values, it would make little sense to implement a new measurement methodology. However, the two cost metrics are significantly different in this example. The cost per equivalent patient unit suggests that Hospital B is 6.4 percent more costly than Hospital A ($9,366 compared with $8,800). In contrast, the cost per adjusted discharge metric suggests that Hospital A is 128 percent more costly ($11,953 compared with $5,232).
In addition, comparing the cost with relevant U.S. averages would suggest that Hospital A is very costly using the adjusted discharge method, but only slightly above U.S. averages using the equivalent patient unit method. These are very sizable differences and would lead to different conclusions regarding cost-saving opportunities.
Finally, Hospital Cost Index® values (which incorporate specific inpatient and outpatient case-weight-adjusted measures of cost) are much more aligned with the cost per equivalent patient unit method and produce similar conclusions as comparisons with U.S. medians.
To corroborate these findings, we ran a simple correlation of the two metrics on 3,000 acute care hospitals (excluding critical access hospitals) for which data were available in 2008 and found that the two measures were positively correlated, but with substantial variation. The conclusion is obvious: The two measures will produce different conclusions on many occasions.
A More Accurate Metric
So which method provides a more accurate assessment of relative cost? We believe that the methodology used to construct equivalent patient units is far better than the adjusted discharge methodology and is better equipped to distinguish cost differences between hospitals.
We conducted some additional analysis on Hospital A and B in the case example, digging deeper into cost data for each hospital. This detailed analysis showed that the equivalent patient unit method accurately concluded that Hospital A is less costly than Hospital B. Specifically, Hospital A appeared to be less costly when departmental cost measures were reviewed. Hospital A also had a lower percentage of overhead costs than Hospital B (32 percent compared with 39 percent), and Hospital A incurred lower losses on Medicare patients, both inpatient and outpatient.
Why was there such a dramatic difference between the two cost measures? For this example, there is one primary reason: Hospital B priced its outpatient services 70 percent above inpatient services while Hospital A had a 35 percent differential. The obvious result of this pricing strategy would be to inflate adjusted discharges at Hospital B relative to Hospital A.
The equivalent patient unit methodology circumvents this central issue and more accurately represents the true relative volume of patient services provided by the hospital. Some bias may still be present to the extent that mark-ups differ within inpatient or outpatient product lines, but the magnitude of the bias is significantly reduced.
Next Steps
If equivalent patient units is a superior measure to adjusted discharges, two important questions remain. How do hospital executives convert to using this superior measure? And will there be available benchmarks for them to use for comparative purposes?
With the method presented here, hospitals should have little difficulty measuring their volume in equivalent patient units. The real obstacle is the limited availability of benchmarks on an equivalent patient unit basis. Given the utility of the equivalent patient unit method, however, it is likely that the method will be widely adopted soon, once it is publicized, and comparative data will proliferate. But even without access to data from benchmarking firms, a hospital can also relatively easily obtain public use file data for a set of peers and develop its own internal benchmarks.
William O. Cleverley, PhD, is president, Cleverley & Associates, Inc., Worthington, Ohio, and a member of HFMA's Central Ohio Chapter (bcleverley@cleverleyaassociates.com).
James O. Cleverley is a principal, Cleverley & Associates, Worthington, Ohio, and a member of HFMA's Central Ohio Chapter.
Publication Date: Tuesday, March 01, 2011