Operating room (OR) workflow and process time optimization has become an important focus in recent years. In many instances, the problems in the workflow are easy to identify but hard to change, because those who have to change do not see any direct benefit from it. A major challenge in changing processes in the OR is therefore to construct a situation in which all participants benefit from the optimization process (a win-win situation). In our hospital, an internal transfer pricing system (ITPS) for anesthesia services was introduced in 2001 to optimize perioperative workflow. The basic principle of an ITPS, which is a well established method of economic steering in manufacturing industries, is that every organizational unit that receives services or goods from another organizational unit of the same company has to pay for it according to a predefined billing system, rather than receiving this internal service for free. The basic rationale for an ITPS is that anyone who has to pay for something will try to use it as efficiently as possible. In our tertiary-level university hospital in northern Germany, an ITPS for anesthesia services was introduced in 2001. The system was designed to give incentive to the surgical services and the anesthesia department to optimize the perioperative process and, hence, to reduce costs. More precisely, the surgical departments should have an incentive to reduce procedure times and thereby reduce anesthesia service times, but in turn the department of anesthesiology should have an incentive to reduce anesthesia process times because they are an important part of the OR turnover time. In this study, we describe the method of an ITPS for anesthesia services and analyze the effects of the ITPS for anesthesia services on process times in our anesthesia department.
Before 2001, the anesthesia department received a budget from the hospital management to pay for personnel and material costs. This budget was mainly based on the budget of the previous years and was taken out of the general hospital overhead. Since the introduction of the ITPS in 2001, the anesthesia department receives its budget according to the delivered anesthesia time. The payer is no longer the hospital, but the departments that request anesthesia services. Accordingly, the overhead costs previously used to fund the anesthesia service were distributed among all departments that request anesthesia services.
Because the main aim in introducing the ITPS was to optimize the perioperative process, a specific new time unit was introduced to motivate both the delivering and requesting departments to be as efficient as possible. Instead of paying for the total anesthesia time or the total case time, the surgical departments pay only for the time they can influence by their own work: from the end of anesthesia induction, when surgical maneuvers can be started, until the surgical work is finished and the extubation of the patient can be started, which is in most cases the end of dressing and repositioning the patient. This time frame is called “perioperative” time in our hospital (Fig. 1) and equals the surgically controlled time as defined by Dexter et al. (1). When the surgically controlled time is used as the basis for reimbursement, the anesthesiologist has a strong incentive to keep preparation and induction, but also extubation times and recovery room transfer, as short as possible, because the department does not receive any money for these time spans. Only during the surgically controlled time is revenue generated. However, the surgeon can save considerable amounts of money if the surgically controlled time is kept short. As soon as the patient is ready for surgery, the surgical department is charged, and all delays after this point—e.g., the surgeon is not available or radiographs are missing—go directly out of the bottom line of the surgical department.
For this study, we compared anesthesia process times for the 3 yr before the introduction of the ITPS and the 3 yr after the introduction of the ITPS. After approval from our ethics committee, we extracted the relevant process times from the central anesthesia record database (Medlinq®; Hamburg, Germany) generated from the scanned anesthesia records. Because 22,000–25,000 anesthesia cases are performed each year in our department, we decided to include only the anesthesia cases from the first 6 mo of every year in our analysis. Approximately 94% of all anesthesia cases are performed for patients from 10 surgical services: general surgery (including visceral and thoracic surgery, pediatric surgery, hepatic surgery, and liver transplantation); ear, nose, and throat (including pediatric) surgery; obstetrics and gynecology; oral and facial surgery; ophthalmology (retina and non-retina); traumatology; neurosurgery (including interventional neuroradiology); cardiovascular surgery (including pediatric cardiac surgery); urology; and orthopedics. This study focused on cases performed for these 10 surgical subspecialties. Smaller and infrequently served subspecialties, such as interventional cardiology or endoscopy, pediatric radiology, and cases treated in the emergency room, were not included in the analysis, because these cases are mostly performed outside the OR and its workflow. For the same reason, epidural anesthesia for obstetric cases was also excluded. The analysis focused only on the anesthesia services. The two intensive care units (ICUs) and the pain clinic also run by our department are not part of the analysis.
Before the operation, the patient is transferred to the OR suite from the ward or the outpatient clinic and remains in the OR holding area. When the previous operation is about to end, the patient is transferred by the anesthesia nurse to the induction room that is adjunct to the OR. When the anesthesiologist is available after the end of the previous operation, the anesthesia preparation is started. This is defined as the starting point of the anesthesiologist presence (Fig. 1). The following anesthesia process times are documented in fractions of 5 min by the anesthesiologist in the anesthesia record along the time line of the anesthesia chart during and after the operation (Fig. 1): preparation and positioning of the patient by the anesthesiologist before anesthesia induction (preparation time); anesthesia induction period, which ends after all anesthesia-related interventions, such as catheter placement, are completed and the patient is ready to be positioned for the operation on the OR table by the surgical staff (induction time); the surgical positioning and preparation before incision (surgical positioning and preparation time); the pure surgical time from incision to closure (incision to closure time); dressing and repositioning of the patient (dressing time); the period after the end of the operation until the end of the anesthesia (extubation time); and the transfer time to the recovery room (recovery room transfer time). Approximately 10% to 12% of the patients were admitted to the ICU after surgery. In these cases, the transfer time to the ICU was taken instead of the recovery room transfer time. The time definitions used are identical in the entire hospital, and the time spans and time points have been agreed on by all services and the hospital administration. The time documentation in the anesthesia chart is the central documentation for OR management and is also used for statistical and billing purposes. For this analysis, only the four time spans that sum up to the anesthesia-controlled time (ACT) were used.
For obvious reasons, anesthesia process times are strongly influenced by the type of anesthesia and the techniques used. Changes in the ACT can be achieved by changes in the performance of specific techniques or by changes in the portfolio of techniques used. For example, average anesthesia induction time can be reduced either because the same sorts of inductions are performed quicker or because intrinsically shorter inductions—e.g., fewer placements of central venous catheters (CVCs)—are used. To separate both levers, we grouped the anesthesia cases in one of the following 10 categories: conscious sedation, face-mask ventilation, laryngeal mask airway (LMA), endotracheal tube airway, endotracheal tube airway with CVC, endotracheal tube airway with placement of CVC and arterial line, endotracheal tube airway with placement of CVC, arterial line and pulmonary artery catheter and/or transesophageal echocardiography, endotracheal tube airway with placement of CVC, arterial line and epidural catheter, spinal anesthesia (SP), and brachial plexus block (BPB). Using these categories, we focused on the most important constellations, each of which had more than 500 cases in the entire study population. We excluded all cases that did not fit exactly into one of these categories. For example, cases were excluded if a CVC was placed in a patient with an LMA or SP or if an epidural catheter, but no CVC, and arterial line were placed in a patient with an endotracheal tube airway. Additional exclusion criteria included a fiberoptic intubation technique, the placement of double-lumen endotracheal tubes, combined spinal/epidural anesthesia, the use of spinal catheters, and monitored anesthesia care cases. Because of these rigid exclusion criteria, we assumed that the analyzed groups were fairly homogenous, at least if larger numbers of cases for each category were examined.
The data were extracted from the hospital's central anesthesia record database (Medlinq®) and analyzed in Microsoft Excel 2002 (Microsoft, Redmond, WA) and SPSS 11.5 (SPSS Inc., Chicago, IL). If not otherwise stated, the mean ± sd is displayed. For comparison of the overall development of the anesthesia process times from 1998 to 2003, the Kruskal-Wallis test was used. For the subgroups, the 1998 versus the 2003 value of the ACT was compared by using a Mann-Whitney U-test. A P value <0.05 was considered significant.
During the examined six consecutive first half-years from 1998 to 2003, a total of 63,893 cases were performed in the previously described services. A total of 8117 cases (12.7%) were excluded on the basis of the exclusion criteria described in the Methods section, which left 55,776 cases in the analysis. The age, ASA physical status, and emergency status of the patients and the distribution of excluded cases can be seen in Table 1. In Figure 2, the overall development of the fractions of ACT are displayed. The duration of the induction period remained virtually unchanged, but the time needed for preparation before anesthesia and recovery room transfer decreased significantly. The reduction in extubation time was significant, but very small. The subgroup analysis is displayed in Figure 3, in which the mean ACT for 1998 is compared with 2003. Overall, the mean ACT increased with the complexity of the induction, as expected. In most anesthesia subgroups there was a significant decrease in ACT. No reductions were achieved in general anesthesia with the placement of CVCs and arterial lines; in general anesthesia with CVC, arterial lines, and invasive cardiac monitoring; or in BPBs. The ACT reduction ranged from 0 to 18 min for the different subgroups. The magnitude of reduction was used to calculate the total ACT saved by the improved performance (Table 2). Saved ACT was defined as the difference between the expected total ACT in 2003 (2003 case number multiplied by the average ACT in 1998) and the actual ACT in 2003 (2003 case number multiplied by the average ACT in 2003). Because the subgroups were of very different sizes, the relative contributions to the total ACT saved were very different. The largest effect came from the general anesthesia cases with endotracheal tube airways and no further instrumentation.
Cost containment in anesthesia has become a major concern in recent years (2), and there have been reports on cost-containment efforts for drugs and supplies (3–7) and process optimization efforts (8,9) from different anesthesia departments. These efforts relied mainly on information and education—efforts that might have only transient effectiveness. In our study, we describe the introduction of a permanent incentive system for anesthesiologists to improve their participation in the OR workflow. The key learning from the first three years of experience with the new system is that the important time reductions came from the preparation period before anesthesia induction and the recovery room transfer period. In contrast, the induction period, which is often considered the prime marker for anesthesia performance, remained unchanged overall. Our interpretation of this finding is that the process times of pre- and afterwork are easier to reduce by effort, whereas the pace with which inductions are performed might change according to the skill level of the anesthesiologist but cannot be reduced simply by will. The ACT was reduced for almost all anesthesia techniques examined, but more complex techniques were less likely to have reductions in ACT. The reason might be that in an academic institution with residents in training performing most cases, it is more difficult to reduce process times for complex techniques. In contrast, there has been a large reduction in cases with epidural catheters used in combination with general anesthesia. This might be explained by the renaissance of epidural catheters placed in combination with general anesthesia for postoperative pain therapy in recent years in our hospital (10). The findings are consistent with our clinical impression that with more epidurals being placed for this purpose, preparation time has decreased, and with catheters being used more often during the operation, the extubation time has been reduced.
For methodological reasons, we did not quantify the second lever that might have been used to optimize process times: the use of anesthesia techniques with faster process times. As can be seen in Table 2, there has been an increased use of LMA. Another example would be the reduced placement of arterial lines for specific indications. This second lever of practice change might have had a profound effect on overall process performance. However, to quantify this effect is very difficult. We cannot exclude the possibility that a change in case mix led to the reduction of invasive monitoring or the increased use of LMA. We have therefore refrained from taking this effect into account and looked only at changes in process times of identical anesthesia techniques. There are also changes in treatment regimens that increase process times in the anesthesia department, such as the increased number of epidural catheters. Besides the benefit resulting from excellent postoperative pain management, there was also a clear trend to a reduced admission to the ICU in specific indications, such as pancreatic surgery. Increased costs in the anesthesia department are more than outweighed by reduced overall hospital costs. It is important to track these changes in treatment protocols, because adjustment in the ITPS might become necessary if costs are shifted from one department to another.
Another interesting issue is the effect of the described ITPS for anesthesia services on the surgical process times. The results from our hospital have been recently published (11). The surgical times are, of course, highly dependent on the type of surgery; therefore, the authors used a complex methodological adjustment based on the case-mix index. The key finding in this study was that the adjusted surgical durations were reduced by 12% after the introduction of the ITPS. However, because of the complex adjustment of the large varieties of different surgical procedures, the study was limited to the year before and two years after the introduction of the ITPS. Our study, covering three years before and three years after introduction of the ITPS, suggests that it takes some time until the full effect is reached.
There are important methodological limitations in our study. Because we used retrospective data, it is, of course, difficult to prove the causality of the introduction of the ITPS and the reduction of the process times. Indeed, many other reasons may have influenced the results, such as changes in the staff skills, changed materials or drugs, or new treatment protocols. However, there have been no structural changes in the staff composition; i.e., the percentage of anesthesiologists in training remained unchanged. The ICU transferal rates in our study ranged from 10.2% to 12.2% during the study period, with a trend to more frequent ICU transferal rates in the more recent years, probably because of a change in the case mix. However, because ICU transferal is more time consuming than recovery room transfer, this increased transfer rate led to an increase in the average ACT. With increased ICU transferal rates, the benefit calculated in our study is therefore under-estimated. For historical reasons, our hospital does not have a central OR suite. Instead, several independent OR suites with two to three ORs are located in different buildings, and each OR suite has its own recovery room. Only one recovery room serves five ORs and is staffed with 2 nurses. In all other recovery rooms, only one recovery room nurse is in charge. There has been no major change in staffing of the recovery rooms or in the transfer process.
There have been general changes in clinical practice during recent years that could have influenced the results of our study. The percentage of pure regional anesthesia cases remained relatively unchanged (5% to 6% of all cases in our study population). However, as in most anesthesia departments, there has been an increased use of IV anesthetics for general anesthesia cases in recent years. The cases in which anesthesia maintenance was achieved by IV anesthetics (i.e., propofol in almost all cases) increased from 33% of all cases in 1998 to 43% of all cases in 2003, whereas the cases in which volatile anesthetics were used decreased from 56% of all cases in 1998 to 48% of all cases in 2003. Within the volatile-anesthetic group, the use of sevoflurane increased from 26% to 54%, whereas the use of isoflurane decreased from 72% to 42%. The opiate most widely used in our department is sufentanil, but the use of remifentanil increased from 5% in 1998 to 16% in 2003. It is compelling to argue that changes in the drugs used, especially the increased use of IV anesthetics and fast-acting drugs such as remifentanil, are partially responsible for the reduction of anesthesia process times seen in our study. However, the effect of these practice changes would be expected to take place in the induction and extubation period and not in the preparation and recovery room transfer period.
Because the anesthesia record database we used was not established before 1997, we had to confine our study to the years 1998 to 2003. We cannot exclude the possibility that there might be a general continuous trend toward shorter process times. However, for the recovery room transfer time, the average time decreased from 1998 to 1999 but then increased again from 1999 to 2000. Only after 2000 was there a constant decrease. This makes a continuous trend toward shorter transfer times unlikely. For the average preparation time, the situation is less clear. The time had already decreased from 1999 to 2000, i.e., the year before the introduction of the ITPS. However, it remained unchanged from 1998 to 1999, and this, again, makes a continuous trend toward shorter preparation times less likely.
We did not present data on a financial effect of the decreased process times, because we measured anesthesia process times and not costs. To make conclusions from the former to the latter can be misleading. If over-utilization of ORs is a constant problem, with almost all ORs running late every day, the reduction of anesthesia process times is likely to have direct economic benefits in terms of reduced overtime payment, but this is very difficult to quantify. Overtime varies a lot and depends on many factors, such as surgical case mix, case scheduling, and OR management decisions. There might be various reasons why anesthesia process time reductions might not lead to overall improved OR performance. For example, the time gained during the ACT can get lost during the surgically controlled time. Furthermore, previous studies have concluded that the reduction of anesthesia process times is not sufficient to allow the performance of an additional case (1). In general, increasing OR efficacy and profitability is a very complex issue, and the surgical portfolio has a very large effect on OR economics (12). Our study was focused on the effect of an ITPS on anesthesia process times and not on overall OR economics. We assume that the reduction of anesthesia process times can help to improve OR performance, but only if it is integrated in a more holistic OR management approach. Many levers have been suggested to optimize OR workflow and save OR costs, ranging from structural changes (such as optimal time allocation to services) (13,14) or increasing nursing staff to permit the completion of more cases (15) to very practical, day-to-day challenges, such us reducing turnover time (16) or enforcing the timely presence of the surgical and anesthesiological staff in the OR (8,9). In our study, we describe a new and successful approach to streamline perioperative processes by using a transfer pricing system. The achieved reductions from 4 to 18 minutes per case appear to be small. However, because of the large numbers of cases, the reductions total more than 1000 hours in 6 months. In the optimal case, these savings are savings not only in anesthesia process times, but also in OR downtime, presumably also resulting in labor cost savings on the surgical and nursing side.
Even though the ITPS is not an educational program, but rather a permanent economic tool, education is a key success factor for the implementation of such a system. Only if all members of the staff (which in our case included approximately 90 anesthesiologists and 60 anesthesia nurses) are aware of the benefit of an efficient process management can reduction of process times be achieved. Therefore, there were numerous meetings and talks on the new system before and after the introduction of the ITPS. It is important to note that both physicians and nurses are fixed-salary employees of the hospital, and there is no direct monetary incentive for any individual to enhance process performance. However, there is a strong indirect incentive. In our hospital, most working contracts are time-limited. The budgetary situation of the department determines how many personnel can be employed and whether contracts can be renewed. The budgetary situation also determines how much time academic personnel can spend on research.
There might be many important and reasonable objections against such a radical measure. To receive their fees from surgeons (or from the hospital administration on behalf of the surgeons) might sound awkward for anesthesiologists who are used to billing the patients' health insurance independently. In Germany and other European countries, this is less of an issue, because hospitals have already taken over the billing for many years. Also, with the growing popularity of uniform billing systems, such as the diagnosis-related group system and its predecessors, one single amount of money has to be split among several departments that have been involved in the case. In Germany, this includes the costs for anesthesia services. This might not be the case in the United States (US), where anesthesia departments bill separately. However, it has been argued that with the ever-increasing overhead charges in large institutions and the sharp increases in internal subsidizing of less profitable departments, the US system is heading in a similar direction: the disconnect of service and billing. Therefore, the introduction of new systems that distribute the total hospital income according to the departments' productivity irrespective of the departments' actually generated fees has also been proposed for the US (17). Our study suggests that it might be worth considering how such a distribution system can be designed to benefit process optimization. The ITPS we describe is very specific to our system. However, the principle behind it might be applicable to other systems: how can we achieve the goal that surgeons have an incentive to use anesthesia services efficiently? Second, how can we achieve the goal that anesthesiologists have an incentive to deliver their service efficiently? Reimbursement of anesthesia according to total anesthesia time has the paradoxical effect that longer anesthesia process times, in the sense of decreased efficacy, lead to increased reimbursement, which is suboptimal from an economic point of view. Our study shows that a reimbursement system based on the surgically controlled time can lead to reduced anesthesia process times. No anesthesiologist in our department would have agreed that we voluntarily wasted time before the introduction of the ITPS. However, our experience shows that there was room for improvement. The ITPS we describe is one possible tool to motivate anesthesiologists to take part in the process optimization.
We thank Joachim A. Wagner, Dipl. Ing. (Department of Anesthesiology, University Hospital Hamburg-Eppendorf, Hamburg, Germany) for assistance in data extraction from the central anesthesia record database.
1. 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.
2. Fisher DM, Macario A. Economics of anesthesia care: a call to arms! Anesthesiology 1997;86:1118–9.
3. Berman MF, Simson AE. The effect of a drug and supply cost feedback system on the use of intraoperative resources by anesthesiologists. Anesth Analg 1998;86:510–5.
4. McNitt JD, Bode ET, Nelson RE. Long term pharmaceutical cost reduction using a data management system. Anesth Analg 1998;87:837–42.
5. Hawkes C, Miller D, Martineau R, et al. Evaluation of cost minimization strategies of anesthetic drugs in a tertiary care hospital. Can J Anaesth 1994;40:894–901.
6. Lubarsky DA, Glass PSA, Ginsberg B, et al. The successful implementation of pharmaceutical practice guidelines: analysis of associated outcomes and cost savings. Anesthesiology 1997;86:1145–60.
7. Lubarsky DA, Sanderson IC, Gilbert WC, et al. Using an anesthesia information management system as a cost containment tool. Anesthesiology 1997;86:1161–9.
8. Overdyk FJ, Harvey SC, Fishman RL, Shippey F. Successful strategies for improving operating room efficiency at academic institutions. Anesth Analg 1998;86:896–906.
9. Truong A, Tessler M, Kleimann S, Bensimon M. Late operating room starts: experience with an education trial. Can J Anaesth 1996;43:1233–6.
10. Schuster M, Gottschalk A, Freitag M, Standl T. Cost drivers in patient controlled epidural analgesia for postoperative pain management after major surgery. Anesth Analg 2004;98:708–13.
11. Kuntz L, Vera A. Effects of the introduction of transfer pricing for anaesthesia services on the efficiency of a hospital. Z Betriebswirtschaftliche Forsch. In press.
12. Dexter F, Ledolter H. Managing risk and expected financial return from selective expansion of operating room capacity: mean-variance analysis of a hospital's portfolio of surgeons. Anesth Analg 2003;97:190–5.
13. Abouleish AE, Dexter F, Epstein RH, et al. Labor costs incurred by anesthesiology groups because of operating rooms not being allocated and cases not being scheduled to maximize operating room efficiency. Anesth Analg 2003;96:1109–13.
14. Dexter F, Blake JT, Penning DH, et al. Use of linear programming to estimate impact of changes in operating room time allocation on perioperative variable costs. Anesthesiology 2003;96:718–24.
15. Dexter F, Blake JT, Penning DH, Lubarsky DA. Calculating a potential increase in hospital margin for elective surgery by changing operating room time allocation or increasing nursing staffing to permit completion of more cases: a case study. Anesth Analg 2002;94:138–42.
16. Dexter F, Abouleish AE, Epstein RH, et al. Use of operating room information system data to predict the impact of reducing turnover times on staffing costs. Anesth Analg 2003;97:1119–26.
© 2005 International Anesthesia Research Society
17. Barker S. Lord or vassal? Academic anesthesiology finances in 2000. Anesth Analg 2001;93:294–300.