To determine the costs of each anesthetic technique from the perspective of the chief financial officer of the ambulatory surgery center, total costs were calculated by summing the costs of drugs, nursing labor, and resources used in the PACU for managing postoperative pain and nausea. Nursing labor costs were prorated on the basis of the actual time spent by the nurse with a patient and divided by the number of patients cared for by the nurse. The mean institutional costs for nursing salaries and benefits were provided by the hospital administration. Separate analyses were performed to account for drug wastage and nursing labor. Cost determinations from the perspective of the OR pharmacy director were limited to the acquisition (direct) costs of drugs.
An a priori power analysis indicated that 28 patients needed to be enrolled in each group to obtain an 80% chance of detecting a 20% mean reduction in the total costs for a 1-h anesthetic from $55 to $43 at the P <or=to 0.05 level of significance. For the power analysis, assumptions of the mean (+/- SD) costs were taken from previously published data [12,13]. One-way analysis of variance was used to compare the continuous variables among the four treatment groups. If a significant difference was noted, a Newman-Keuls multiple comparison test was used to determine intergroup differences. Categorical variables were analyzed using the chi squared test with Yates' continuity correction or Fisher's exact test, as appropriate. A P value <0.05 was considered statistically significant. Data are presented as mean values +/- SD, numbers, or percentages.
There were no significant differences among the four study groups in terms of patient weight, gender, history of previous surgery, motion sickness, or prior PONV, as well as the type of surgical procedures (Table 2). Patients in the methohexital-desflurane group were slightly younger than those in the propofol-sevoflurane group. However, there were no significant differences in the duration of surgery and anesthesia, volume of IV fluids administered, total intraoperative doses of sufentanil, or use of inhaled anesthetics as expressed in minimum alveolar anesthetic concentration (MAC)-minutes (product of the end-tidal concentration in multiples of the MAC and duration of administration). There were also no differences in the EEG-BIS, MAP, or HR values after induction, at skin incision, at completion of the surgical procedure, and on awakening.
Early recovery times (time from discontinuation of anesthetic drugs to eye opening, following commands, and being oriented to person, place, and date of birth) were similar among the four groups. There were no differences in the time from discontinuation of anesthesia to the time when patients could sit up, ambulate, retain oral fluids, achieve discharge criteria, and be discharged from the hospital ambulatory center (Table 3). There were also no differences among the four groups in the need for postoperative analgesic medications, incidence or severity of PONV, number of repeat episodes of PONV, and need for rescue antiemetic medication. The use of propofol (versus methohexital) for the induction of anesthesia was not associated with decreased PONV (49% vs 53%; P = 0.84). The difference in the number of patients who developed nausea or vomiting after receiving sevoflurane or desflurane did not achieve statistical significance. However, in this study, rescue antiemetics were administered more frequently when sevoflurane was used for maintenance of anesthesia (40% vs 20%; P < 0.05).
The cost of using propofol for the induction of anesthesia was higher than that for methohexital. However, the costs of inhaled drugs and IV adjuvants required during the maintenance period did not differ among the four treatment groups (Table 4). If FGF rates of <or=to1 L/min were used, there would have been significantly lower costs in the methohexital-desflurane group compared with the other three anesthetic groups. Even with FGF rates of 2 L/min, the costs of intraoperative anesthetic drugs were significantly lower in the methohexital-desflurane group compared with the propofol-sevoflurane group (P < 0.05). This cost difference would have been more pronounced had the costs of wasted drugs been considered in the overall costs (Table 4).
In the PACU, the incidence of pain and the need for analgesic medications did not differ among the four groups. Although the higher incidence of emesis in the sevoflurane (versus desflurane) groups did not achieve statistical significance, there was a significantly increased use of PACU resources in patients receiving the propofol-sevoflurane combination. The costs of nursing labor did not differ among the four groups, reflecting similar recovery times. When costs associated with each anesthetic technique were compared, they were significantly lower in patients who received methohexital with either desflurane or sevoflurane compared with patients who received propofol with desflurane or sevoflurane. These cost differences were mainly due to differences in the acquisition costs of the drugs used for induction.
Using a cost-minimization analysis, methohexital was found to be less costly than propofol for the induction of general anesthesia in the ambulatory setting. The most surprising finding was that the induction of anesthesia with methohexital, followed by maintenance with desflurane, was associated with decreased total costs from the perspectives of both the pharmacy and the healthcare institution, particularly when low flow rates (<or=to1 L/min) were used. The recommendation by the United States Food and Drug Administration that a minimal FGF rate of 2 L/min be used with sevoflurane increases the costs of this regimen. In this study, we examined the total costs of an anesthetic regimen, including the costs of nursing labor and wastage of drugs. A proper pharmacoeconomic analysis should not be limited to the direct cost of drugs used but should also encompass the indirect costs due to delayed awakening, prolonged stay in the PACU, and postoperative complications such as pain and PONV [5,14].
We demonstrated that using methohexital for the induction of anesthesia in healthy outpatients is associated with an early recovery profile comparable to that of propofol. The respective time intervals to eye opening, responsiveness to commands, tracheal extubation, orientation to time and place, and fulfillment of home-ready criteria were similar. However, using more sensitive indices of cognitive recovery (e.g., Maddox wing, p-deletion, Digit Symbol Substitution Test) might have demonstrated the purported advantages of propofol compared with methohexital. In a pharmacoeconomic analysis [5,14], the time spent by patients in the PACU and the resources consumed have been shown to be of greater importance than scores on cognitive and/or psychomotor tests. Similarly, using rigid protocols in comparative studies of different anesthetics has reduced their usefulness in economic analyses. Examples include studies in which the concentration of the volatile anesthetics was maintained constant at a specified MAC equivalent from the start to the end of the surgical procedure, then abruptly discontinued. Although this study design may elucidate which volatile anesthetic is associated with the most rapid awakening, it is not helpful in economic evaluations because anesthesiologists do not normally administer these drugs in this manner. In this study, the anesthesiologists altered the volatile anesthetic concentrations as necessary to maintain an adequate depth of anesthesia. All changes in the delivered anesthetic concentrations and total FGF rates were recorded in order to calculate the amount of volatile anesthetic used. There were no differences in the number of MAC-minutes of exposure to the volatile anesthetics or in the intraoperative BIS values, which suggests that anesthesia was maintained at a similar depth during the operation.
One of the controversies in pharmacoeconomic analyses of anesthetic regimens is the inclusion of PACU nursing labor costs. Cost-minimization studies of anesthetic drugs often ignore staffing costs because it is assumed that nursing costs are comparable if the drugs have similar recovery profiles [12,13]. Because nursing labor costs constitute a major fraction of the total costs in the management of surgical patients [5,14], most anesthesia-related studies have used a linear model that assumes that costs were directly proportional to the time spent in the OR complex and that any time saved was associated with a proportionately decreased cost [14,15]. This is based on the concept of "opportunity costs," in which the assumption is made that the time spent by an employee in a particular patient-care activity obviates other tasks from being performed, necessitating the hiring of another person to perform that work. This assumption is deemed correct only if the nurses in the PACU take care of one patient at a time, are not present before the arrival of the patient in the PACU, are sent home after that patient is discharged, and are only paid for the time actually spent working in the Phase I recovery area. This does not reflect the true situation in a typical healthcare facility. Nursing labor costs are semifixed rather than variable because nurses usually work a fixed shift in one recovery area at hospital-based ambulatory centers. Dexter and Tinker  concluded that the major determinant of PACU costs was the peak number of patients admitted to the PACU at any one time, and a reduction in the time to discharge had a minimal impact on overall PACU costs. Therefore, we presented comparative data including and excluding nursing labor costs (Table 4).
Although we did not take into consideration the costs of relatively rare but potentially costly complications (e.g., seizures, arrhythmias), we did consider the more common problems after ambulatory anesthesia, such as PONV and pain. We noted that the use of a single-induction dose of propofol was not associated with decreased PONV compared with the barbiturate. This finding is consistent with a meta-analysis by Tramer et al.  involving 84 studies with >6000 patients. However, when propofol was used for induction and maintenance of general anesthesia, it was found to be associated with a lower incidence of PONV compared with methohexital . In contrast, when these two IV anesthetics were used for maintenance of sedation, there were no differences in their recovery profiles .
In summary, we demonstrated that the use of methohexital for the induction, followed by desflurane, for maintenance of anesthesia in the ambulatory setting is associated with decreased total costs compared with an anesthetic regimen consisting of propofol with sevoflurane or desflurane. The cost differences are primarily due to differences in the acquisition of drugs used for the induction.
1. Smith I, White PF, Nathanson M, Gouldson R. Propofol: an update on its clinical uses. Anesthesiology 1994;81:1005-43.
2. Patel SS, Goa KL. Desflurane: a review of its pharmacodynamic and pharmacokinetic properties and its efficacy in general anaesthesia. Drugs 1995;50:742-67.
3. Patel SS, Goa KL. Sevoflurane: a review of its pharmacodynamic and pharmacokinetic properties and its clinical use in general anaesthesia. Drugs 1996;51:658-700.
4. Smith I, Nathanson MH, White PF. The role of sevoflurane in outpatient anesthesia. Anesth Analg 1995;81:S67-72.
5. Watcha MF, White PF. Economics of anesthetic practice. Anesthesiology 1997;86:1170-96.
6. Pastuovic MN, Cohen ME, Burton RG. Propofol: an alternative general anesthetic for outpatient oral surgery. J Oral Maxillofac Surg 1996;54:943-8.
7. Wagner BK, O'Hara DA. Cost analysis of propofol versus thiopental induction anesthesia in outpatient laparoscopic gynecologic surgery. Clin Ther 1995;17:770-6.
8. Tramer M, Moore A, McQuay H. Propofol anaesthesia and postoperative nausea and vomiting: quantitative systematic review of randomized controlled studies. Br J Anaesth 1997;78:247-55.
9. Beskow A, Werner O, Westrin P. Faster recovery after anesthesia in infants after intravenous induction with methohexital instead of thiopental. Anesthesiology 1995;83:976-9.
10. Mokriski BK, Nagle SE, Papuchis GC, et al. Electroconvulsive therapy-induced cardiac arrhythmias during anesthesia with methohexital, thiamylal, or thiopental sodium. J Clin Anesth 1992;4:208-12.
11. Korttila K, Linnoila M, Ertama P, Hakkinen S. Recovery and simulated driving after intravenous anesthesia with thiopental, methohexital, propanidid, or alphadione. Anesthesiology 1975;43:291-9.
12. Boldt J, Jaun N, Kumle B, et al. Economic considerations of the use of new anesthetics: a comparison of propofol, sevoflurane, desflurane, and isoflurane. Anesth Analg 1998;86:504-9.
13. Rosenberg MK, Bridge P, Brown M. Cost comparison: a desflurane- versus a propofol-based general anesthetic technique. Anesth Analg 1994;79:852-5.
14. Tagliente TM. Pharmacoeconomics of propofol in anesthesia. Am J Health Syst Pharm 1997;54:1954-6.
15. Dexter F, Tinker JH. Analysis of strategies to decrease postanesthesia care unit costs. Anesthesiology 1995;82:94-101.
16. Tramer M, Moore A, McQuay H. Meta-analytic comparison of prophylactic antiemetic efficacy for postoperative nausea and vomiting: propofol anaesthesia vs omitting nitrous oxide vs total i.v. anaesthesia with propofol. Br J Anaesth 1997;78:256-9.
17. Doze VA, Westphal LM, White PF. Comparison of propofol with methohexital for outpatient anesthesia. Anesth Analg 1986;65:1189-95.
© 1999 International Anesthesia Research Society
18. Sa Rego MM, Inagaki Y, White PF. Cost-effectiveness of methohexital versus propofol for sedation during monitored anesthesia care. Anesth Analg 1999;88:723-8.