| JOURNAL HOME | CME HOME | THIS MONTH | PAST ISSUES | ETOC | COLLECTIONS |
| AUTHORS | REVIEWERS | EDITORIAL BOARD | FEEDBACK | RSS | HELP |
 
|
|
|
||
|
|
© 2001 International Anesthesia Research Society
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
Academic anesthesiology departments provide clinical services for surgical procedures that have longer-than-average surgical times and correspondingly increased anesthesia times. We examined the financial impact of these longer times in three ways: 1) the estimated loss in revenue if billing were done on a flat-fee system by using industry-averaged anesthesia times; 2) the estimation of incremental operating room (OR) sites necessitated by longer anesthesia times; and 3) the estimated potential gain in billed units if the hours of productivity of current anesthesia time were applied to surgical cases of average duration. Health Care Financing Administration average times per anesthesia procedure code were used as industry averages. Billing data were collected from four academic anesthesiology departments for 1 yr. Each claim billed with ASA units was included except for obstetric anesthesia care. All clinical sites that do not bill with ASA units were excluded. Base units were determined for each anesthesia procedure code. The mean commercial conversion factor (US$45 per ASA unit) for reimbursement was used to estimate the impact in dollar amounts. In all four groups, anesthesia times exceeded the Health Care Financing Administration average. The loss per group in billed ASA units if a flat-fee billing system were used ranged from 18,194 to 31,079 units per group, representing a 5% to 15% decrease (estimated billing decrease of US$818,719 to US$1,398,536 per group). The number of excess OR sites necessitated by longer surgical and anesthesia times ranged from 1.95 to 4.57 OR sites per group. The potential gain in billed units if the hours of productivity of current anesthesia time were applied to surgical cases of average duration was estimated to be from 13,273 to 21,368 ASA units. Longer-than-average anesthesia and surgical times result in extra hours or additional OR sites to be staffed and loss of potential reimbursement for the four academic anesthesiology departments. A flat-fee system would adversely affect academic anesthesiology departments.
IMPLICATIONS: We examined the economic impact of longer-than-average anesthesia times on four academic anesthesiology departments in three ways: the estimated loss in revenue under a flat-fee system, the excess operating room sites staffed, and the potential gain in revenue if the surgeries were of average length. These results should be considered both in productivity measurements and strategies for operating room management.
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Academic anesthesiology departments are believed to provide anesthesia care for surgical procedures of longer-than-average duration, in part because surgical and anesthesiology residents work less rapidly than fully trained physicians and perhaps because patients referred to teaching hospitals may have greater medical complexity. Longer surgical times inevitably increase the total anesthesia duration.
The current payment and billing system for anesthesia care takes into consideration variations of surgical duration. Unlike other specialties that receive flat-fee reimbursement for procedures, anesthesia care is billed as a combination of a base charge for the complexity of anesthesia care (Base units) plus the total time required (Time units). Because this system differs from that of other specialties, payers would prefer that anesthesia care be billed and reimbursed by a flat-fee system that combines average anesthesia times and Base units for specific procedures (1). The effect of this system would be to reward anesthesia groups that provide coverage for surgical procedures of shorter duration and to penalize anesthesia groups that provide coverage for surgical procedures of longer duration.
In addition, longer-than-average surgical times require that, to provide anesthesia for the same numbers and types of procedures, academic groups must work more hours than groups that provide anesthesia for procedures with shorter durations. By inference, anesthesiology departments would then incur increased costs because either more operating room (OR) sites or longer workdays would be required in comparison to groups providing care for surgical cases of average duration. Also by inference, if longer-than-average surgical times were reduced to average durations and if the anesthesiology group still worked the same amount of time, then the number of billed units would increase because more surgical cases would be performed and the total number of Base units would increase (2).
The purpose of this study was to examine and quantify the economic effect of longer-than-average anesthesia times on four academic anesthesiology departments in three ways: 1) estimated loss in revenue if a flat-fee system were used; 2) estimated excess OR sites staffed as compared with average anesthesia times; and 3) estimated potential gain in billed units if the hours of productivity of current anesthesia time were applied to surgical cases of average duration. Additionally, we correlated anesthesia times with surgical times to address the hypothesis that longer anesthesia times reflect longer surgical times.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Billing information for one calendar year (1999) was collected from four academic anesthesiology departments (Groups A through D) (Table 1). Departments recorded data on individual claims by using either surgical procedure codes or anesthesia procedure codes (APT) and time of anesthesia care in minutes. For those claims that recorded only surgical procedure codes, the corresponding APT code was determined by using a database that cross-references surgical and anesthesia codes (3). Time units were calculated as 15-min increments. Base units for each APT were determined from a commercially available reference (3). All billed claims using ASA units were included except for obstetric anesthesia (anesthesia for vaginal or cesarean delivery). All clinical sites (e.g., intensive care units, pain clinics, or preoperative screening clinics) that do not bill with ASA units were excluded. Average anesthesia times per APT used were from the ITIMER data from the Health Care Financing Administration (HCFA) Web site (4). Two of the departments (A and D) also provided data specifying which patients were insured under Medicare. The reported mean reimbursement for commercial or managed care payers per ASA unit (i.e., commercial conversion factor of US$45 per ASA unit) was used (5). For each group, the number of OR sites staffed daily was estimated as the 12-mo mean of the number of sites listed on the OR schedule at the beginning of the day on the 10th of each month. If the 10th occurred on a weekend or holiday, then the 20th of the month was substituted. If remote sites were staffed, then all remote sites (non-OR rooms) were combined and counted as one site. Labor and delivery suites were excluded (6).
|
Total ASA units actually billed were calculated for each claim by adding the Time units (15-min Time units) and Base units (3). The estimated total ASA units that a flat-fee system would assign for each claim were calculated for each claim from the same Base units for the APT and the HCFA average Time units (15-min Time units). The total Base and Time units were added for all claims for each group for both actual billed data and calculated units assuming a flat-fee system (Table 1 for abbreviations and calculations). Means and SD for Time units, Base units, and total ASA units were calculated. In addition, the ratio of Time units to total ASA units and the total ASA units per hour of anesthesia care (tASA/h) were calculated as noted in Table 1. For example, tASA(act) represents the actual total ASA billed by a group, and tASA(HCFA) represents total ASA units that were estimated to be billed under a flat-fee system.
Impact 1: Estimated Loss of Revenue if Using a Flat-Fee Reimbursement System
For the estimate of reimbursement under a flat-fee system, tASA(HCFA) was used, as defined in Table 1. The estimated loss in units was calculated as the difference between tASA(HCFA) and tASA(act). The loss in billed charges (in US dollars) was estimated by multiplying the loss in units by US$45. The percentage of lost revenue in comparison to actually billed charges was also determined (Table 2).
|
Impact 2: Estimated Excess OR Sites Staffed Compared with Average Anesthesia Times
Because the Base units are the same for the calculations of both actual and flat-fee billing, the difference in billed units is accounted for by the difference between actual Time units and average anesthesia times for each claim and APT. This difference represents the excess time of care provided because of longer anesthesia times. Dividing excess time of care by the Time units per OR site provided an estimate of the number of excess OR sites required to be staffed (Table 2).
Impact 3: Potential Gain in Billed Units if Using the Same Time Units Worked Performing Anesthesia for Cases of Average Duration
tASA/h(HCFA) describes the number of ASA units that would be billed for every hour of anesthesia care if average anesthesia times were used with the same case mix of surgeries. Multiplying tASA/h(HCFA) by the actual hours of care provided [hours of care = TU(act)/4] determined the estimated potential billed units. The difference between potential billed units and actual billed units represents the potential gain if the anesthesia times were at average levels (Table 2).
Impact 1 was also determined for the subset of patients (from Groups A and D) for whom Medicare was recorded as the financial class. Surgical times were defined as the time from surgery start to surgery finish, whereas the anesthesia time was from the start of anesthesia care to the finish of anesthesia care (7). Surgical time from Group A’s OR database was collected and compared with the anesthesia time for the calendar year 1999. Because remote sites and cases performed at the ambulatory surgicenter were excluded from this analysis, not all billed cases were analyzed from Group A’s OR database. Pearson’s correlation was used to define the relationship between the variables. A P value <0.05 was considered significant.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
All four groups provided a total of 64,634 cases, with a range of 11,742 to 24,369 (Table 3). The Base units, Time units, and tASA actually billed and estimated by using average HCFA times are shown for each group as an aggregate in Table 3 and mean per case in Table 4. Groups with longer Time units per case showed a higher ratio of Time units to total ASA units and a corresponding lower tASA/h.
|
|
Groups A, B, and C showed similar values for Time units per OR site, implying that each OR site was staffed for similar amounts of time, whereas Group D’s Time units per OR site was higher, implying that OR sites in that group were staffed longer in the evening or on weekends (Table 5). This comparison is consistent with the clinical impressions of each of the groups’ representatives.
|
All four groups had longer-than-average anesthesia times (Table 5). The loss per group in billed units under a flat-fee billing system was estimated as 18,194 to 31,079 ASA units, which would represent a decrease of 5% to 15% of total billings per group. For the groups studied, the loss in revenue ranged from US$818,719 to US$1,398,536 per group. The number of excess OR sites staffed because of longer anesthesia times ranged from 1.95 to 4.57 OR sites per group. The increase in potential billed units if the same amount of time were used to provide anesthesia care for cases of average duration was estimated as 13,273 to 21,368 ASA units; this would represent a potential gain in revenue of US$597,274 to US$961,561. Only two groups (A and D) provided claim data with financial class identified (Table 6); in these groups, 16% of patients in Group A and 29% of patients in Group D were insured by Medicare. Although the Base units per case were higher than in the overall four-group analysis of all financial classes, the Time units per case were lower for Medicare patients in both groups, as were average Time units estimated from HCFA data. Because both actual and HCFA data show shorter anesthesia times, the difference in anesthesia times most likely results from a difference in case mix (numbers and types of APT) among all patients and the subset of Medicare patients. Despite this, the effect of lost revenues dealing with Medicare patients is similar to that found with all patients (Group A, -11% for Medicare-only patients and -12% for all patients; Group D, -5% and -4%, respectively). For 12,663 cases from Group A, surgical times correlated significantly with the anesthesia times (r2 = 0.85, P < 0.0001) (Fig. 1).
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Confirming our clinical impression, this study demonstrated that anesthesia times highly correlate with surgical times. Because surgical times are not recorded in most anesthesiology databases, the use of anesthesia times as a surrogate for surgical times is justified by this high correlation. Longer-than-average anesthesia times strongly influence each of the academic anesthesiology departments studied by increasing the costs associated with staffing additional OR sites or additional hours in individual sites and by decreasing charges per hour of anesthesia care. Unlike the current reimbursement system, which considers both Base units and Time units, a flat-fee system would further decrease reimbursement to these academic departments in comparison to anesthesiologists providing care for surgical cases of average or even shorter duration.
Under a flat-fee system, there would be winners and losers. The loss of revenue by academic departments would be offset by a gain in revenue by groups providing anesthesia for cases of shorter-than-average duration. Inequities in reimbursement could be amplified if the times used to develop a flat-fee reimbursement schedule failed to accurately reflect average times of specific types of cases (1). In the current HCFA database, multiple surgical procedures of considerably varying complexity can be classified within one APT code. Under a flat-fee system, it would be essential to increase the number of APT codes to be more specific and correspond more accurately to the surgical procedures. Further problems in reimbursement could arise if third-party payers were to undervalue the conversion factor of a flat-fee system.
In this study, we chose a methodology that did not look at overall times for specialty surgeries but compared times for each individual APT. We were concerned that if we simply compared the overall times, we would not make an accurate evaluation. To illustrate our concern, we could have asked participating groups to provide us with the average minutes per case. In this case, we were concerned that we would not know whether the difference in anesthesia times was related to longer surgery duration or to different surgical complexity. Because of the same concerns, we considered but also rejected the option of asking groups to provide us with average minutes per case by specialty, e.g., orthopedic surgery, gynecology, plastic surgery, etc. Because we chose to perform a detailed comparison of times based on APT, we required groups to provide the data at an individual case level rather than as aggregate data. Additionally, by choosing to perform the analysis by APT, we were able to compare the groups, not only with each other but also with actual HCFA averages (which are based on APT). By using this methodology, we were able to estimate the losses if a flat-fee reimbursement system were implemented.
Unfortunately, by choosing a methodology that required a detailed level of data, we found that only academic groups were willing to participate because of the time and cost of producing the data. Inclusion of private-practice groups presumably would have demonstrated that the losses of academic departments under a flat-fee system would have been offset by gains by private groups.
The fact that multiple surgical procedures are included under the rubric of one anesthesia code is also a concern in the methodology of this study. The anesthesia billing information, including anesthesia minutes billed per APT, is available in participating departments’ billing databases. The surgical times are not usually found in the anesthesia billing database and are found only in the OR information system. Therefore, collection of actual surgical times and more specific surgical procedures from multiple sites would require substantial additional investment by participating departments and may not have been possible for all the groups. We believe that the approach used in this study is more appropriate for examining anesthesia billing issues, although the use of specific surgical procedures and times would be a more accurate method of precisely characterizing longer surgical times. Of course, in terms of the financial performance of an academic department, there is little difference between longer times for similar cases and longer times for more complex cases that are classified under the same anesthesia code. Furthermore, the methodology was designed to allow any anesthesiology group, even one working in multiple hospitals, to reproduce the work from the group’s own database.
For similar reasons, when the anesthesia times and surgical times were evaluated for correlation, we were able to examine only one group. Because the OR information system database for each group would have to be accessed and related to the anesthesia billing database, this was not possible for all the groups. If the investigators at each site were required to collect data on each of the 64,634 cases from multiple databases and format the data in a uniform way, the expense and effort would have been prohibitive. In Group A’s case, these data had already been accessed and related.
Another methodological issue is our choice of a unit reimbursement of US$45. This is the mean commercial conversion factor for reimbursement that has been reported (5). Most anesthesiologists bill larger amounts per unit, whereas many payers, including Medicare, reimburse at a lower rate. Our analysis, therefore, understates lost charges but could overestimate actual lost reimbursement, on the basis of the payer mix in each of the participating groups.
We also used the HCFA database in this study to determine average anesthesia times for various surgical procedures, despite the fact that this database is based on the experience with Medicare patients, who represent a minority of our total sample. However, we believe that these average times are applicable to non-Medicare patients because, as shown in Table 6, in comparison to the average times, Medicare patients show percentage decreases in total ASA units billed similar to those of patients in the total sample.
The limitations of the data collected most affected the estimate of the excess ORs staffed and the associated costs of longer-than-average surgical times (Impact 2). In this study, the results estimated the total Time units or minutes beyond average time that anesthesia care was provided. Group D had the shortest anesthesia times of the four groups analyzed, but those times were still longer than the HCFA average. As a result, Group D required only 8% additional work because of longer surgical times. Group A had the largest number of excess OR sites, at 4.57, and Group B had the highest percentage of excess OR sites (Table 5). These results may overestimate the excess OR sites staffed because a simple decrease in surgical time may not directly translate to a reduction in OR sites (8). Nevertheless, the results demonstrate that a group that quantifies the costs of staffing an OR site can estimate the costs associated with longer surgical times. This estimate supports the argument that a portion of the staffing costs of an academic anesthesiology department should be allocated to institutional sources. Similarly, in estimating Impact 3, we understood that academic hospitals are unlikely to ever have times comparable to nonacademic hospitals, but this loss in potential revenue demonstrates an additional method to illustrate the financial effect on anesthesiology departments of longer-than-average times.
Further, a more detailed estimate of costs must include an evaluation of cases performed during regular hours as compared with after hours. Anesthesia provided after hours costs more in both staffing and opportunity costs (9). In other words, it is more costly to perform anesthesia care at 9:00 PM than it would at 2:00 PM. Therefore, a reduction in costs should be associated with a reduction in after-hours care even if the same number of OR sites is open during weekdays.
The results have direct implications on clinical productivity measurements. Measuring clinical productivity has become important for strategic planning and management of a department or group, as well as justification of budgets to university administrators. The results of this study show that tASA/h is lower in groups with longer-than-average anesthesia times when compared with groups using the same staffing mix and providing anesthesia for cases of average duration. Furthermore, in comparison to groups with anesthesia times that are shorter than average, this difference may be even greater. In addition, the results showed that additional OR sites required staffing because of the longer-than-average anesthesia times. These two findings imply that simply examining tASA billed or tASA per full-time equivalent (FTE) may be inadequate to allow for accurate comparisons between anesthesiology groups (6). In examining individual clinical productivity measurements within a group, the duration of surgery was one of the factors that led to the conclusion that "clinical days worked per clinical FTE" is the most useful and accurate measurement (2).
In this study, we did not attempt to define the reasons behind longer-than-average anesthesia times in this group of academic programs. Three factors are most likely to contribute to longer anesthesia times: 1) residents in surgical training programs; 2) residents in anesthesiology training programs; and 3) more medically complex patients presenting for the same procedures. Quantifying the contribution of each of these factors is beyond the scope of this article and is likely to vary, as does the overall excess duration of cases, from department to department.
These data have several implications. First, surgical time must be considered, along with turnover time, day-of-surgery cancellations, and on-time starts, in any analysis of OR efficiency. Previous studies have shown that more rapid turnover times cannot increase the number of cases performed per OR day (8). Because surgical duration greatly exceeds anesthesia-controlled time, our data suggest that decreasing the duration of surgery could possibly increase the number of cases performed. Second, although this analysis focused on loss of revenue as well as increased operating costs as applied to anesthesiology departments in hospitals with longer-than-average anesthesia times, there are also hospital costs that should be quantified in a complete analysis of the indirect costs of surgical training programs.
In conclusion, this study demonstrates the negative financial effect of longer-than-average anesthesia and surgical times on four academic anesthesiology departments. In particular, costs and lost revenue are estimated. The evidence that the duration of surgery strongly influences the total ASA units billed gives further credence to avoiding this as a measurement of individual anesthesiologists’ clinical productivity or as an isolated measurement of departmental clinical productivity.
|
Acknowledgments |
|---|
The authors wish to thank Karin Bierstein from the American Society of Anesthesiologists for reviewing the manuscript and providing editorial guidance; Ann Tate, Financial Affairs Manager at the University of Texas Southwestern Medical Center at Dallas, for assisting in data collection; and Jordan Kicklighter in the Department of Anesthesiology at The University of Texas Medical Branch for preparing and editing this manuscript.
|
Footnotes |
|---|
Presented in part at the 2001 Annual Meeting of the International Anesthesia Research Society, Ft. Lauderdale, FL, March 20, 2001 and at the 2001 Annual Meeting of the American Society of Anesthesiologists, New Orleans, LA, October 15, 2001.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
- Bierstein K. Practice management: ASA analyzes commercial relative value system based on average time. ASA Newslett June 1997. Available at: http://www.asahq.org/NEWSLETTERS/1997/06_97/Pract_Mgmt_June.html. Accessed April 4, 2001.
- Abouleish AE, Zornow MH, Levy RL, et al. Measurement of individual clinical productivity in an academic anesthesiology department. Anesthesiology 2000; 93: 1509–16.[Web of Science][Medline]
- American Society of Anesthesiologists. Crosswalk: a guide for surgery/anesthesia CPT codes. Park Ridge, IL: American Society of Anesthesiologists, 1999.
- Health Care Financing Administration. Physician time associ-ated with work RVUs used in creating practice expense rela-tive values. 1999. Available at: http://www.hcfa.gov/stats/resource.htm. Accessed April 4, 2001.
- Bierstein K. Practice management: fees paid for anes-thesia services—survey results. ASA Newslett August 1999. Available at: http://www.asahq.org/NEWSLETTERS/1999/08_99/practmgnt0899.html. Accessed April 4, 2001.
-
Abouleish AE, Prough DS, Zornow MH, et al. Designing meaningful industry metrics for clinical productivity for anesthesiology departments. Anesth Analg 2001; 93: 309–12.
[Abstract/Free Full Text] - Association of Anesthesia Clinic Directors. Procedural times glossary. Available at: http://aacdhq.org/glossary.htm. Accessed January 4, 2001.
- Dexter F, Coffin S, Tinker JH. Decreases in anesthesia-controlled time cannot permit one additional surgical operation to be reliably scheduled during the workday. Anesth Analg 1995; 81: 1263–8.[Abstract]
- Strum DP, Vargas LG, May JH. Surgical subspecialty block utilization and capacity planning: a minimal cost analysis model. Anesthesiology 1999; 90: 1176–85.[Web of Science][Medline]
This article has been cited by other articles:
|
|
 |
|
  A. E. Abouleish, D. S. Prough, C. W. Whitten, and L. A. Conlay Increasing the Value of Time Reduces the Lost Economic Opportunity of Caring for Surgeries of Longer-Than-Average Times Anesth. Analg., June 1, 2004; 98(6): 1737 - 1742. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. E. Abouleish, D. S. Prough, C. W. Whitten, and M. H. Zornow The Effects of Surgical Case Duration and Type of Surgery on Hourly Clinical Productivity of Anesthesiologists Anesth. Analg., September 1, 2003; 97(3): 833 - 838. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. E. Abouleish, D. S. Prough, S. J. Barker, C. W. Whitten, T. Uchida, and J. L. Apfelbaum Organizational Factors Affect Comparisons of the Clinical Productivity of Academic Anesthesiology Departments Anesth. Analg., March 1, 2003; 96(3): 802 - 812. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
