This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via Web of Science (3) Citing Articles via Google Scholar Google Scholar Articles by Jackson, T. Articles by Myles, P. S. Search for Related Content PubMed PubMed Citation Articles by Jackson, T. Articles by Myles, P. S.

---

INTRAVENOUS ANESTHESIA

Part II: Total Episode Costs in a Randomized, Controlled Trial of the Effectiveness of Four Anesthetics

Terri Jackson, MA, PhD^*, and Paul S. Myles, MB BS, MPH, MD, FFARCSI, FANZCA

*Hospital Services Research Group, Monash University Health Economics Unit, West Heidelberg; and Department of Anaesthesia and Pain Management, Alfred Hospital, Prahan, Australia

Address correspondence and reprint requests to Terri Jackson, MA, PhD, Hospital Services Research Group, Monash Health Economics Unit, PO Box 477, W. Heidelberg, VIC 3081, Australia. Address e-mail to terri.jackson{at}buseco.monash.edu.au

	Abstract

Top Abstract Introduction Methods Results Discussion References

 
Newer anesthetics promise improved clinical outcomes, but usually come at a higher price per dose. Previous studies have found few economic benefits in the immediate postoperative period, but have hypothesized that earlier recovery may lead to lower costs for the whole episode of hospitalization. This study uses cost data for patients enrolled in a randomized, controlled clinical trial comparing four anesthetics to test whether the higher costs of the newer anesthetics would be offset against decreased use of other hospital resources. Five hundred general surgery patients were randomly assigned to one of four anesthetic regimens. Estimates from the hospital’s patient costing system were used, with validated cost records for a subset of 360 patients. Five patients admitted to the intensive care unit or requiring prolonged hospitalization skewed the distribution of costs, but none of these complications could be attributed to anesthesia. No significant differences were found on length of stay, mean episode cost, operating room costs, ward costs, or readmission rate within 3 mo. The study was not powered to sufficiently show differences in intensive care unit admission or other uncommon outcomes. Patient quality of recovery did not vary among groups, but neither patient willingness-to-pay nor satisfaction were directly measured.

Implications: Propofol and sevoflurane do not offer any significant economic advantages over thiopental and isoflurane in adults undergoing elective inpatient surgery.

	Introduction

Top Abstract Introduction Methods Results Discussion References

 
The development of shorter-acting anesthetics has been one of a number of medical advances to profoundly change surgical practice in the past two decades. Along with minimal access surgery and early ambulation, safer and shorter-acting anesthetics have made same-day surgery possible for a range of surgical procedures, and have shortened hospital stay for many others.

It is frequently argued that differences in anesthetic outcomes might influence the entire episode of hospitalization (1–6), but no studies of the newer anesthetics have attempted to document these more distant economic outcomes (3,4). Our study reports a comparison of the costs of care for a large cohort of patients randomized to one of four regimens for the induction and maintenance of anesthesia (thiopental-isoflurane, propofol-isoflurane, propofol induction and maintenance—total IV anesthesia [TIVA], or sevoflurane induction and maintenance). Clinical outcomes for these patients with planned stays of one or more nights are presented elsewhere (7), and showed no difference among the four anesthetics, including no patient-reported benefits as assessed by a quality of recovery survey instrument (8), for the more costly anesthetics.

The clinical study provides a good basis for economic evaluation, as pragmatic trials using the typical range of presenting surgical cases are more directly applicable to routine clinical practice (9), as is the focus of the study on patients admitted for at least an overnight stay. Despite the lack of significantly different clinical outcomes, we sought to test whether the hypothesized benefits of the newer anesthetics in reduced hospital length of stay (LOS) and reduced ward nursing requirements would be realized as cost savings over the whole episode of inpatient care, and/or would result in reduced rates of readmission in the 3 mo after discharge. The perspective of the analysis is that of the cost-conscious anesthesiologist, drug formulary committee, or health care provider, and only in-hospital costs are considered.

	Methods

Top Abstract Introduction Methods Results Discussion References

 
Elective surgery patients who were expected to require at least overnight hospital care (n = 500) were recruited for participation in a randomized, controlled trial of four anesthetic regimens. Index admissions for the period July 29, 1997, through May 28, 1999, were analyzed, with additional analysis of readmissions to the hospital in the 3 mo after the index admission. Forty-seven cases were subsequently excluded from the trial because of changes in treatment plan, including canceled surgery, administration of an alternative anesthetic, and discharge on the day of surgery (same-day discharge rather than overnight care).

Of the remaining 453 patients, costing information was not available for 38, and an additional 55 cost records were judged to be unreliable as they had no recorded operating theater costs, leaving a fully costed sample of 360 patients. Missing cost data were attributed to staff turnover in the clinical costing unit of the hospital, but no systematic pattern was found among patients with missing cost records. Omitted patients were evenly distributed among the four trial groups and across types of surgery. Five patients were found to have very large total episode costs (>US$ 9000) compared with the all-cases mean of US$ 2245, and results are reported both for the whole sample and omitting these five anomalous patients.

Information on the costs of the episode of care were extracted from the hospital’s clinical costing system (Transition 1^TM; Eclipsys Corp., Delray Beach, FL). Records were identified by unit record number and date of surgery from the clinical trial data set, and subsequent admissions within a 3-mo period were identified by using these same variables. For each admission, cost fields were extracted aggregated to the following categories: imaging, pathology, ward (nursing and associated ward costs), operating theater (including anesthesiologists, surgeons, theater and recovery nurses, and anesthetics), medical (pre- and postoperative medical care), intensive care unit (ICU; nursing and medical), pharmacy (other than anesthetics), and allied health services.

Clinical costing builds a patient-specific cost profile based on recorded use of hundreds of hospital "products": a particular pathology test, a dependency-weighted day of nursing care in Ward X (10). Many of these estimates rely on relative value scales (for example, a schedule of pathology tests done in the laboratory). To derive a price per test, records of tests done in a specific period are weighted by the relative value scale being used, and a standard cost per test is estimated by dividing total departmental expenditure by the number of weighted tests. This method more closely approximates resources used in the production of patient care than a charge-based methodology, as estimates do not reflect product cross-subsidies or market-related charge patterns (5,11).

Nursing costs were estimated by using a standard patient dependency system weighted for the staffing profile of specific wards. Although this method of measurement does not guarantee that savings in nursing costs are realized, it does at least apportion dependency-related shares of nursing cost to individual patients for each day on a particular ward. In the ICU, both medical and nursing labor costs were estimated by using a dependency instrument. For other inpatient medical care, sessional payments to clinical departments are apportioned on a per diem basis, with additional weighting for days of higher medical input (day of admission, discharge, and the first postoperative day).

Perioperative costs are aggregated to an operating theater cost center. Purchase costs of each anesthetic agent are allocated on the basis of the number of units administered to each patient (but with wastage averaged across all patients on the theater list). Staff expenditures (nursing and technical salaries, and surgical and anesthesiologist sessional payments) are apportioned to patients in a defined period on the basis of recorded theater times (in 15-min blocks) weighted by a relative value scale of operations to take account of different staffing profiles for each operation. Postanesthesia care unit (PACU) nursing salaries are more crudely allocated by using 30-min blocks which are summed for each patient and each theater list, but these are not weighted by either patient dependency or numbers of patients being monitored. Costs of other pharmaceuticals administered in theater are assigned to the pharmacy cost center, with imprest-system use allocated to patients on the basis of the relative value scale of surgical procedures. Capital costs of equipment to administer these drugs are not included in the cost estimates, and were not separately estimated, as per-patient difference between them at normal theater volumes would not be great.

General hospital overhead costs (administration, utilities, etc.) are assigned to intermediate departments after first having been allocated to each other (by using simultaneous equations to record the proportion of payroll office costs to be allocated to the housekeeping department and the proportion of housekeeping department costs to be allocated to the payroll office) (12). Costs reported are thus fully absorbed costs, and totals from the clinical costing system are reconciled to patient care expenditures reported in the hospital’s financial accounts, and patient numbers to the official health department patient returns (13).

An exchange rate of 0.6 (A$:US$) is used to report costs in US dollars. Compared with the United States, Australia spends a much smaller proportion of its gross domestic product per capita on health care (14), with the consequence that episode costs will appear low by US standards. Price relativities for the anesthetics evaluated here were found to be similar to those reported elsewhere, and Australian differences in other treatment costs are unlikely to distort cost relativities among the interventions.

Data were analyzed by using SPSS 9.0.1 for Windows (Chicago, IL). Cost analysis was initially blinded to the anesthetics used in the four intervention groups. The nonparametric Kruskal-Wallis one-way analysis of variance test was used to identify differences in the ranks of mean costs because of the skewed nature of the cost data. Mean costs (in US$) are reported; however, in preference to median or other nonparametric measures, following recommendations for the reporting of economic evaluations (15). The ² test was used to test for differences in the proportion of next-day discharges and in readmission rates across the four treatment groups. P values of <0.05 were considered significant.

	Results

Top Abstract Introduction Methods Results Discussion References

 
No differences on any outcome variable were found to be significant (Table 1 ). The mean cost of hospitalization for these patients was US$1977, ranging from US$1850 for the thiopental-isoflurane group to US$2102 for the propofol TIVA group. When the five high-cost cases are also considered, the mean episode cost is increased to US$2245, ranging from US$1959 for the sevoflurane group to US$2657 for the thiopental-isoflurane group (a difference of 36% between the high and low groups).

Mean LOS was 2.87 (median = 2) days across all groups, with a 9% difference between the means of the highest and lowest groups (thiopental/isoflurane 2.76 days versus propofol/isoflurane 3.02). The thiopental-isoflurane group had a slightly higher than expected proportion of next-day discharge (56%) when compared with the sample mean (47%), but again the difference was not statistically significant. Taking the five high-cost cases into account, mean LOS ranged from 2.78 (propofol TIVA) to 3.78 (thiopental-isoflurane), with an all-cases mean of 3.21.

Ward costs (including nursing, and other ward costs such as meals, ward clerks, and porters) and ICU costs are apportioned per dependency-weighted day of stay in the designated ward. Typical admissions averaged US$231 for nursing costs, ranging from US$199 for the propofol-isoflurane group to US$277 for the sevoflurane group, a difference of 39%. The addition of the five high-cost cases changed relativities, with an all-cases mean of US$347, and a range from US$231 for propofol TIVA to US$592 for the thiopental-isoflurane group.

Operating theater costs averaged more than half of total episode costs (54%) at US$1072. This is similar to the proportion of combined operating room/anesthesiologist costs reported from the United States (5) for elective surgery patients. Across intervention groups, the two propofol interventions were the highest cost (US$1194 and US$1155 for propofol/isoflurane and TIVA, respectively), with sevoflurane recorded with the lowest theater cost of US$944. Again, no differences were significant. The operating theater totals included purchase costs of anesthetic drugs allocated by units of anesthetic administered to each patient, although these could not be separately identified in the aggregated theater cost data. Relativities among intervention groups remained the same when the five higher-cost cases were included, but the mean across all cases increased to US$1117.

Thirty-seven patients (10%) were readmitted in the 3 mo after the index operation, with a total of 83 readmissions in the period. The proportion of readmissions varied from 8% for the thiopental-isoflurane group to 12.5% for the propofol TIVA group, but differences were not statistically significant by using a ² test. Review of medical record abstracts for patients with multiple readmissions showed these to be condition-related rather than related to anesthesia.

	Discussion

Top Abstract Introduction Methods Results Discussion References

 
Previous studies of the economics of the newer anesthetics have been limited in a number of ways. Most have concentrated on day-only surgery [where claimed advantages in recovery times would have the largest economic impact (4,16,17)], but provide little information about use for elective procedures requiring postoperative inpatient care. Most concentrate on a single operative procedure (2,6) or surgical specialty (18–20) so as to reduce potential confounding (18), but with the result that extrapolation to a broad range of surgical procedures is reduced.

Similarly, most seek to standardize recovery time measurement by maintaining anesthesia to a defined point, rather than titrating the anesthetic to minimize recovery time, as is more commonly done in clinical practice (7). Finally, these economic evaluations typically focus on the use and costs of drugs only (1,2,21), follow patients only to discharge from the PACU (6,19,22–24), and/or limit their costing to the perspective of an anesthesia department (6), rather than the institution or higher level decision-makers.

Most have suggested that it is important that formulary and other purchasing decisions be based on total resource use rather than price or operating theater and PACU resource use only. However, no previous study has documented changes in such total resource use for the newer anesthetics (3). Finally, many investigators mention the importance of patient satisfaction with the experience of anesthesia and recovery, but only a few have reported significant differences (1,16), while other investigators have found no differences (18,21) between newer and older anesthetics.

This study built on a pragmatic clinical trial with patient-reported (quality of recovery) end-points (7). Recruitment of patients was from the range of elective general and specialist surgical procedures typically encountered by anesthesiologists in general hospital practice. Day-surgery was excluded so as to estimate the impact on total hospital resource use for inpatients. Normal anesthetic practice was followed in reducing levels of anesthesia late in the procedure so as to minimize recovery time. Cost estimation covered the entire episode of hospitalization and was sufficiently sensitive to identify meaningful differences between the groups were they present.

From previous research (25), we had expected to find small but significant differences in effectiveness against which we could evaluate differences in costs attributable to the higher prices of propofol and sevoflurane, and any differences in total episode costs. What began as a cost-effectiveness study became a cost-minimization analysis, on the basis that clinical outcomes were not significantly different.

We found that total episode costs for elective surgery inpatients randomized to four different anesthetic regimens do not differ between the newer and older anesthetics. Although some might take this as evidence that the higher prices of propofol and sevoflurane can be recouped by decreased hospital costs throughout the stay, examination of the relativities in total cost demonstrates that, while not significant, the differences are roughly those which might be expected based on purchase price differentials.

Some caveats must be applied to this conclusion. The most serious issue is whether the sample size for the economics component of the study was large enough to detect relevant differences. Economic evaluations built on clinical trials have the distinct advantage of randomization and lower marginal cost to conduct. However, cost data are "often, but not always, more noisy (variable) than effect data and so detecting differences at the same level of inferential error will be difficult" (26). Ours was a large trial compared with others reported in the literature, but not powered to detect differences in major complications. Neither the prolonged stays for five patients, nor the differential readmission rates were found to be anesthesia-related, but a contribution of the anesthetic to cause or prevention of these events cannot be eliminated.

A second limitation is the possibility of selection bias introduced with patient attrition as a result of incomplete theater cost records. Investigation of the 360 (and subsequently, 355) patients analyzed for this costing study showed these subsamples did not differ from the larger cohort in terms of their ASA physical status or type and extent of surgery, and no significant differences in these variables were found among groups in the subsample reported here.

Nurse labor costs in the PACU were not well measured (and are intrinsically difficult to estimate because staff treat and monitor more than one patient at a time). Balanced against this is the fact that there was no difference in recorded times to fitness for PACU discharge in the clinical effectiveness study (7). Thus, unmeasured differences in nursing costs are likely to be small and, as many commentators have observed (19,27,28), represent semivariable costs which may not result in practical reallocation of resources, even if requirements for care are reduced.

New technologies often result in reappraisal of older patterns of care. Results of the current study may reflect unmeasured benefits from the introduction of the newer anesthetics on the care processes for all patients which could not be easily identified or costed.

Finally, neither patient satisfaction nor willingness to pay were measured explicitly as economic outcomes, although the lack of difference in patient-reported quality of recovery scores provides evidence that explicit patient-satisfaction or willingness-to-pay measures are unlikely to have yielded large enough differences to change the overall conclusions of this study.

Although Macario et al. (5) estimate that only approximately 3% of total inpatient surgical costs are "under immediate control of anesthesia providers," they argue that "choosing less costly alternatives can reduce such costs as long as the quality of care is not decreased." We conclude that propofol and sevoflurane do not offer any significant economic advantages over thiopental and isoflurane in adults undergoing elective inpatient surgery.

	Acknowledgments

 
This study was supported by the Abbott/Australian Society of Anaesthetists Research Grant (1996) and a research grant from the Alfred Hospital Research Trust (1997). Financial support was also provided from an unrestricted grant from Abbott Australasia Pty Ltd and AstraZeneca.

We would like to thank Lynda Butcher and Sean Downer (Alfred Hospital Finance Department), and Helen Fletcher, Jenny Hunt, and Jenny Watts for their research assistance.

	Footnotes

 
July 18, 2000.

	References

Top Abstract Introduction Methods Results Discussion References

 

1. Smith I, Terhoeve P, Hennart D, et al. A multicentre comparison of the costs of anaesthesia with sevoflurane or propofol. Br J Anaesth 1999; 83: 564–70.[Abstract/Free Full Text]
2. Suttner A, Boldt J, Schmidt C, et al. Cost analysis of target-controlled infusion-based anesthesia compared with standard anesthesia regimens. Anesth Analg 1999; 88: 77–82.[Abstract/Free Full Text]
3. Fazi L, Watcha M. The economics of newer anaesthetic drugs: should we take the Rolls-Royce or the bicycle today? Paediatr Anaesth 1999; 9: 181–5.[Web of Science][Medline]
4. Tagliente T. Pharmacoeconomics of propofol in anesthesia. Am J Health Syst Pharm 1997; 54: 1953–62.[Abstract/Free Full Text]
5. Macario A, Vitez TS, Dunn B, McDonald T. Where are the costs in perioperative care: analysis of hospital costs and charges for inpatient surgical care. Anesthesiology 1995; 83: 1138–44.[Web of Science][Medline]
6. 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.[Abstract]
7. Myles PS, Hunt JO, Fletcher H, et al. Propofol, thiopentone, sevoflurane and isoflurane: a randomized controlled trial of effectiveness. Anesth Analg 2000; 91: 1163–9.[Abstract/Free Full Text]
8. Myles P, Hunt J, Nightingale C, et al. Development and psychometric testing of a quality of recovery score after general anesthesia and surgery in adults. Anesth Analg 1999; 88: 83–90.[Abstract/Free Full Text]
9. Schwarz D, Lellouch J. Explanatory and pragmatic attitudes in therapeutic trials. J Chron Dis 1967; 20: 637–48.[Web of Science][Medline]
10. Jackson T, Watts J, Lane L, Wilson R. Data comparability in patient level clinical costing systems. Casemix Q 1999; 1: 36–45.
11. Finkler S. The distinction between cost and charges. Ann Intern Med 1982; 96: 102–9.
12. Jackson TJ. Cost estimates for hospital inpatient care in Australia: evaluation of alternative sources. Aust N Z J Pub Health 2000; 24: 233–40.
13. Jackson T, Watts J, Wilson R, et al. Cost weights for inpatient and outpatient care: final report to the Victorian Department of Human Services—April 1999. Melbourne: Hospital Services Research Group, Monash Health Economics Unit; 1998.
14. Altman S, Jackson T. Health care in Australia: lessons from Down Under. Health Affairs 1991; 10: 129–44.[Medline]
15. Briggs A, Gray A. The distribution of health care costs and their statistical analysis for economic evaluation. J Health Serv Res Policy 1998; 3: 233–45.[Medline]
16. Tang J, Chen L, White P, et al. Recovery profile, costs, and patient satisfaction with propofol and sevoflurane for fast-track office-based anesthesia. Anesthesiology 1999; 91: 253–61.[Web of Science][Medline]
17. Fulton B, Goa KL. Propofol: a pharmacoeconomic appraisal of its use in day case surgery. PharmacoEconomics 1996; 9: 168–82.[Web of Science][Medline]
18. Alhashemi J, Miller D, O’Brien H, Hull K. Cost-effectiveness of inhalational, balanced and total intravenous anaesthesia for ambulatory knee surgery. Can J Anaesth 1997; 44: 118–25.[Web of Science][Medline]
19. Hsu S, Shalansky S. Pharmacoeconomics of propofol versus thiopental for induction of anaesthesia in short procedures. Can J Hosp Pharm 1995; 48: 208–13.[Medline]
20. Wagner B, O’Hara D. Cost analysis of propofol versus thiopental induction anesthesia in outpatient laparoscopic gynecologic surgery. Clin Therapeutics 1995; 17: 770–6.[Web of Science][Medline]
21. Smith I, Thwaites A. Target-controlled propofol vs sevoflurane: a double-blind, randomised comparison in day-case anaesthesia. Anaesthesia 1999; 54: 745–52.[Web of Science][Medline]
22. Kurpiers E, Scharine J, Lovell S. Cost-effective anesthesia: desflurane versus propofol in outpatient surgery. J Am Assoc Nurse Anesth 1996; 64: 69–75.
23. Nathan N, Rezzoug A, Dolan P, et al. Propofol en perfusion ou isoflurane en circuit ferme: etude du cout. Ann Fr Anesth Reanim 1993; 12: 571–4.[Web of Science][Medline]
24. Sun R, Watcha MF, White P, et al. A cost comparison of methohexital and propofol for ambulatory anesthesia. Anesth Analg 1999; 89: 311–6.[Abstract/Free Full Text]
25. Myles P, Hendrata M, Bennett A, et al. Postoperative nausea and vomiting: propofol or thiopentone: does choice of induction agent affect outcome? Anaesth Intensive Care 1996; 24: 355–9.[Web of Science][Medline]
26. Drummond MF, O’Brien B, Stoddart GL, Torrance GW. Methods for the economic evaluation of health care programmes. Oxford: Oxford University Press, 1997.
27. Dexter F, Coffin S, Tinker J. 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]
28. Dexter F, Tinker J. Analysis of strategies to decrease postanesthesia care unit costs. Anesthesiology 1995; 82: 94–101.[Web of Science][Medline]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via Web of Science (3) Citing Articles via Google Scholar Google Scholar Articles by Jackson, T. Articles by Myles, P. S. Search for Related Content PubMed PubMed Citation Articles by Jackson, T. Articles by Myles, P. S.