Page 1 of 8 Previous Full Text References (20) Next Full Text An Economic Evaluation of Trauma Care in a Canadian Lead Trauma Hospital Jean Séguin, BSc; Bryan G. Garber, MD, FRCSC, MSc; Douglas Coyle, MA, MSc; Paul C. Hébert, MD, FRCPC, MSc From the Departments of Surgery (JS, BGG) and Medicine (PCH), University of Ottawa, and the Clinical Epidemiology Unit, Loeb Institute (BGG, DC, PCH), Ottawa, Canada. THE JOURNAL OF TRAUMA: INJURY, INFECTION, AND CRITICAL CARE 1999;47:S99- S103 Background: The objective was to determine the average cost per quality-adjusted life year (QALY) gained of treating trauma victims at a tertiary trauma hospital and to determine the cost-effectiveness of trauma care at this center. The setting was a tertiary trauma center in the province of Ontario, Canada. The study population consisted of consecutive trauma admissions with ISS > 12 from April, 1994 to April, 1996. The study was of a retrospective cohort design with a cross-sectional survey. Methods: The hospital perspective was taken. Costs were determined from a retrospective cohort using a hospital-based case-costing system. Utility estimates for calculation of QALYs gained were obtained using a cross-sectional survey design. Cost-effectiveness was determined by estimating the incremental cost/qaly attributable to treatment at the trauma center. Sensitivity analysis was employed to vary assumptions about the proportion of costs and increased survival. Results: 484 patients with a median age of 39 years and a median ISS of 22 were studied. The average cost per QALY was $1,721, with a maximum value of $3,861. The increase in cost per QALY gained for treatment in a tertiary care center as opposed to a nontrauma center was $4,303, assuming a 20% increase in survival and assuming that the existence of the center increased the cost of care by 50%. The incremental cost/qaly ranged from $191 to $15,492 in the sensitivity analysis varying assumptions about the increased proportion of costs and survival attributable to care at the tertiary trauma center. Conclusions: This is the first economic evaluation of tertiary trauma care which includes both costs as opposed to charges as well as estimates of the QALYs gained. The results suggest that tertiary trauma care is cost-effective and less costly than treatment programs for other disease conditions when the quality-adjusted life years gained are included in the evaluation. Key Words: Trauma; Economic evaluation; Quality-adjusted life years. Trauma is the number one killer of Canadians younger than 45 years. In Canada, the average injury admission rate nationally for 1995-1996 has been estimated at 72 per 10,000 population, and these injury-related admissions accounted for 2,187,305 hospital days. 1 Regional trauma care systems and specialized trauma centers were developed to reduce the mortality attributable to severe injuries.
Page 2 of 8 Several studies have shown that trauma care systems can reduce the preventable trauma death rate by as much as 30% for patients who reach the hospital alive. 2-6 The cost-effectiveness of these systems has never been properly evaluated. Although a number of published studies have sought to characterize the costs of trauma, 7-9 none of these studies conformed to standard economic models for the evaluation of medical technologies. 10-13 The estimation of quality-adjusted life years (QALYs) gained from trauma care was uniformly absent. The QALY is an important parameter because it measures both the number of years gained by the health intervention and the person's preference for the achieved health state. 14-16 Most of these studies reported hospital charges instead of costs, and the former may be substantially inaccurate. 17 Therefore, we performed an economic evaluation of trauma care at a single lead center in the province of Ontario. This involved assessing the average cost per QALY for care of that trauma patient and then modeling the potential incremental cost-effectiveness of trauma care at a lead trauma center. PATIENTS AND METHODS Study Population The patient cohort for our study consisted of consecutive trauma admissions to the Ottawa General Hospital with an Injury Severity Score 18 (ISS) > 12 from April, 1994 to April, 1996. The Ottawa General Hospital is the lead trauma center for the Ottawa-Carleton area, servicing a regional population of 1.2 million. On the average, there are 240 admissions per year with an ISS > 12. A data analyst tracks all such admissions and trauma codes through the emergency department until discharge. The computerized trauma database provided the hospital cost for all of these patients. Determination of Utilities Utilities were estimated by performing a cross-sectional mail survey of all trauma victims who were discharged alive with an ISS > 12 during the study period. Surviving trauma victims were followed by direct contact, by tracing medical records, or by checking death records kept by the Deputy Registrar General of Ontario. Patients were randomly allocated to receive and self-complete either the EuroQol 19 or the Health Utilities Index. 20 Both of these are generic instruments designed to measure utility, the patient preference for the current health status. Surveys were administered using a modified Dillman 21 technique. Nonresponders were sent a reminder card at 2 weeks, followed by a second survey questionnaire. Those who had failed to respond by that point were contacted by telephone to complete the questionnaire by telephone interview. Surrogate completion was attempted for incompetent respondents. This was required for only one patient. Determination of QALYs The utility values obtained from the cross-sectional survey were incorporated into calculation of QALYs using a standard formula (Appendix). 22 Survivors at follow-up were assumed to have a life expectancy of 90% of age- and gender-specific life expectancy, a stable future, and a stable future quality of life. 15 The estimated survival was discounted at a rate of 5% annually to reflect the greater value society places on immediate, rather than long-term, benefits. 23,24 Patients who survived to discharge but were lost to follow-up were assumed to have had a quality of life equal to the mean
Page 3 of 8 quality of life of survivors in the same injury severity group. Patients who died were assigned a QALY of zero, as suggested by Torrence and Feeny. 15 Estimation of Resource Consumption The trauma database allowed a determination of the amount of resources consumed per trauma patient. This included seven groups of services: imaging, laboratory, nursing, rehabilitation and social work, respiratory therapy, operating room, and pharmaceuticals. Although information on emergency department resource consumption was not directly available, it was estimated by summing up all of the nonoperative procedures (e.g., endotracheal intubation, ventilation, IV therapy) performed during their stay in the emergency department. In addition, the number of nursing hours was not available; therefore, we estimated the hours by taking the average number of nursing hours provided for trauma victims presenting between June and July of 1996 (30 patients). Estimation of Costs Estimates of the costs of admission were obtained from a hospital-based case-costing system that allows the measurement of departmental workload, the identification of patient-specific resource use, and the calculation of unit costs for individual resource items. Following national guidelines, the direct costs of treatment (e.g. staff, consumables) were allocated directly to the treatments received, and indirect costs (e.g. overheads) were allocated to treatments. Thus, the cost per admission per patient could be calculated. The trauma center used in this study can be considered to be representative, in cost structure, of other Canadian trauma centers. Because of the fee for service payment mechanism in Ontario, it was not feasible for this study to retrospectively determine the physician costs for each of the sample patients. Thus, we determined the potential range of physician costs by considering the likely physician billings associated with a minor and a severe trauma case; this was estimated to range from $190 to $10,000 per patient. Analysis This study was performed from the hospital perspective. Descriptive analysis was performed on all variables. The total average cost/qaly was estimated for all patients. Estimates of the cost/qaly were also obtained for subsets of patients on the basis of admission to intensive care, age, and injury severity. To evaluate cost-effectiveness of trauma care at this facility, we determined the incremental cost/qaly gained assuming that the trauma care received resulted in a 20% increase in survival 2-6 and supposing that the existence of a specialized trauma center increased the cost of caring for that patient above that of a nonspecialized center by 50%. Sensitivity analysis 25 was performed to test the robustness of our estimate of the incremental cost/qaly of trauma care by varying assumptions about the proportional increase in costs and survival attributable to the trauma care provided at the lead trauma hospital. RESULTS Patient Data There were 484 patients identified over the study period that met our inclusion criteria, of which 47 died during their hospital stay. Of the 437 survivors eligible for follow-up, 251 (57.4%) responded to the questionnaire. This provided a total of 298 patients (251 survivors and 47 nonsurvivors) out of
Page 4 of 8 the original cohort in whom utility estimates could be performed; the nonsurvivors were assumed to have a utility score of 0. The median follow-up time was 11.8 months (interquartile range = 15.3). Study respondents demonstrated a median age of 39 years (interquartile range = 37.5), and 69% were male. They had a median ISS of 22.0 (interquartile range = 12.75), and the majority of patients had sustained multiple injuries (76.2%) as a result of blunt trauma (95.6%). Those patients who responded to the cross-sectional survey and provided utility estimates were similar to nonresponders in age, gender, injury category and length of stay (Table 1). TABLE 1. Characteristics of the entire cohort of trauma patients as well as those contacted at discharge and those lost to follow up Utility Estimates The mean utility score for all survivors at discharge was 0.67 (SD = 0.30). The mean utility score for all patients, including nonsurvivors, was 0.54 (SD = 0.37). The mean discounted QALY for all patients was 8.3. QALYs were highest in the younger group and diminished with increasing age. The same was found with the ISS grouping; that is, the patients with a low ISS had the greatest number of QALYs and the group with a high ISS had the lowest number of QALYs (Table 2). TABLE 2. Cost in Canadian dollars per quality-adjusted life years (QALYs) for trauma patients treated at our tertiary trauma hospital Costs The total costs for trauma care of the entire cohort of 484 patients was estimated to be $8,258,517 Canadian dollars, and the average cost per trauma admission was $14,115. The majority of hospital costs related to trauma care were attributable to nursing (62%) with the remainder almost equally distributed between radiology, laboratory investigations, operating room and recovery, pharmacy, rehabilitation and social work, respiratory therapy, physician costs and the emergency department. Estimation of the Average Cost per QALY The average total cost per QALY was $1,721 on the basis of the average cost per patient admission of $14,115 and a mean discounted QALY per patient of 8.3 (Table 2). The maximum value of the cost/qaly was estimated by including the costs of those patients who were lost to follow-up and by assuming that they had a mean utility score of 0. This gave a mean cost per QALY of $3,861.17 for the entire cohort of 484 trauma patients. The cost per QALY was higher in those patients who required admission to an intensive care unit and in those older than 60 years (Table 2). There appeared to be no trends between ISS and cost per QALY.
Page 5 of 8 Incremental Cost/QALY The increase in the cost/qaly of treating a patient with a given injury attributable to the development and maintenance of a specialized trauma center was $4,303 (Table 3), assuming that the existence of this center increased survival of trauma victims by 20% and increased the cost of care of that patient by 50%. This value ranged from an incremental cost/qaly of $191 (assuming a 90% increase in survival and 10% increase in costs) to a maximum of $15,492 (assuming a 90% increase in costs with only a 10% increase in survival). TABLE 3. Sensitivity analysis of the incremental cost/qaly attributable to care in the trauma center. The columns represent the increase in the proportion of total cost/qaly of treating the patient attributable to the existence of the trauma center, the row and percentages represent the proportion of increased survival attributable to the trauma center. DISCUSSION In this study, we have combined estimates of the direct costs of trauma care at a lead trauma hospital with long-term quality-adjusted survival after the injury. These results indicate that trauma care is far less costly than many other health care technologies. The methods we employed outline an approach for performing future economic evaluations. Moreover, our results provide important baseline information necessary for determining the cost-effectiveness of future modifications in trauma care at this facility. Several published studies have sought to characterize the costs of trauma. 7-9 One study of 597 trauma victims found that medical charges ranged from $5,446 to $24,107 in US dollars. 7 Within each body region, both length of stay and hospital charges increased as a function of injury severity. A subsequent article estimated that the total direct and indirect costs of trauma in the United States was $11 billion per year and that the elderly consumed a disproportionate share of resources. 8 Finally, in a study analyzing 12 trauma centers and 43,219 patients in the state of New York, where 90% of trauma patients were categorized in diagnosis-related groups for comorbidities and complications, it was reported that the trauma cases cost 27.5 million dollars more than nontrauma patients in the same diagnosis-related group. 9 The implication was that the care of the injured was more costly. All of these studies have similar shortcomings. First, although all of these studies actually used charges, they reported these as costs. The use of charges may reflect inaccurate dollar amounts, as hospital charges to patients (and indirectly to insurance companies) for products or services do not relate to the actual costs of those products and services. 17 In our study, actual costs were derived from the hospital case-costing system; this gave a more accurate picture of the dollar expenditure per patient. Second, all of these studies failed to provide any measurement of the value of health that resulted from the dollar cost. The use of QALYs is an integral component of all economic models of health care evaluation. 14-16 The importance of the QALY is that the life years gained from a given intervention is weighted by the utility, which is the level of satisfaction or enjoyment the individual has with their current health
Page 6 of 8 state. This provides a composite measure of both the quality and the duration of life that is obtained from a therapeutic intervention or program. Studies that provide estimates of cost per patient alone are difficult to interpret because they assume that the health state achieved is the same for all patients. The fallacy of this assumption is demonstrated by our observation that although the cost of treating patients with ISS ranging from 25 to 40 was almost quadruple that of patients with ISS ranging from 13 to 24 ($16,337 vs 4,166, respectively), the cost per QALY difference was far less impressive ($2,130 vs 1,003, respectively). Consequently, any economic evaluation of trauma care that does not incorporate an estimate of QALYs is meaningless. The cost-effectiveness of the trauma center in this study was determined by performing an incremental analysis because we could not directly compare treatment of patients in a trauma center with treatment of patients with similar injuries in a nontrauma center. However, by making assumptions regarding the proportion of health care costs arising from the provision of trauma care and the associated survival benefit, we gained some insights into the likely cost-effectiveness of tertiary trauma care. We found that the increase in the cost per QALY of treating a patient in a specialized trauma facility was only $4,303, assuming an increased survival of 20% and assuming that the existence of the facility increased the cost of care by 50%. Even assuming a worst-case scenario where the existence of the trauma center increased the cost by 90% and only produced a 10% increase in survival, the incremental cost per QALY for the care of that patient was $15,492. Accordingly, implementation of a tertiary trauma hospital would receive a grade B recommendation based on the criterion of Laupacis et al., 26 which states that the new technology or intervention is more effective than the alternative and costs less than $20,000 per QALY gained. Our estimate of the incremental cost per QALY of trauma care compares very favorably to other interventions such as the administration of nonionic contrast media to people at high risk ($23,000/QALY) 27 and hospital hemodialysis ($65,500/QALY). 28 This indicates that tertiary trauma care is cost-effective, in contrast to the conclusions of previous studies that failed to incorporate estimates of QALYs. 7-9 Our study does have several other limitations. First, estimates of utilities were based on a crosssectional survey; this resulted in patients being studied at non-uniform intervals after injury. Consequently, variations in utility estimates may have been a function of differences in recovery time. A prospective study in which all patients are assessed at uniform times during their recovery would have been optimal. Second, we could not obtain complete follow-up on all patients. Despite this limitation, we could estimate that if all of these people had a utility of zero, the mean cost per QALY for our patients would have been $3,861.17, which is not substantially higher than the current estimate. Finally, we performed the analysis from a hospital perspective, as opposed to from a societal perspective. Although this approach was justified to evaluate the cost/qaly of the trauma center, a societal perspective would provide a better picture of the costs of trauma beyond the confines of the hospital. In summary, this study was designed to demonstrate the potential benefits of using both costs and QALYs in performing economic evaluations of trauma care. Our model suggests that the cost of trauma care in specialized trauma centers is quite reasonable when the health state achieved is taken into account. This demonstrates that only by capturing the health state achieved through estimation of QALYs gained from a given health care technology can we have a meaningful denominator for the dollar cost. We need to perform prospective studies where the estimates of the health states achieved can be performed at uniform times. Studies performed from the societal perspective are also necessary to obtain a more thorough estimate of the costs incurred from outpatient service use as well as from lost employment related to the injury. Such information will be crucial when evaluating the efficacy of existing and future changes in trauma care management as well as provide necessary
Page 7 of 8 information for policy decisions about ongoing resource allocation in the area of trauma care. Acknowledgments We thank Ms. Maureen Brenan-Barnes, trauma data analyst at the Ottawa General Hospital, who was of inestimable value in the collection of the data. We also thank Ms. Dariah Parsons of the Ontario Trauma Registry for her assistance. APPENDIX The general formula for calculation of patient outcome in discounted QALY for each patient is as follows: (Equation) where: Q = Quality Score R = Discount rate (5% per annum) s = Duration of survival in years: If dead at follow up, then s = time from admission until death. If alive at follow up, then s = 0.9 (age/gender specific life expectancy at admission). Equation A1 Example: Q = 0.67, age = 20 years, s = 0.9 * (61.67 years), dqaly = 0.67 * (sum(1/1.05) i 1 ) where i = (1 to s) dqaly = 8.26. REFERENCES 1. Canadian Institute for Health Information. The National Trauma Registry Report. Hospital Injury Admissions 1995/1996. Canadian Institute for Health Information; 1998. 2. Cales RH. Trauma mortality in Orange County: the effects of the implementation of a regional trauma system. Ann Emerg Med. 1984;13:1-10. 3. Shackford SR, Mackersie RC, Hoyt DB, et al. Impact of a trauma system on outcome of severely injured patients. Arch Surg. 1987;122:523-527. 4. Stewart TC, Lane PL, Stefanits T. An evaluation of outcomes before and after trauma center designation using Trauma and Injury Severity Score analysis. J Trauma. 1995;39:1036-1040. 5. Rutledge R, Fakhry SM, Meyer A, Sheldon GF, Baker C. A analysis of the association of trauma centers with per capita county trauma death rates. Ann Surg. 1993;218:512-524. 6. Mullins RJ, Mann NC, Hedges JR, Worrall W, Jurkovich GJ. Preferential benefit of implementation of a statewide trauma system in two adjacent states. J Trauma. 1998;44:609-617. 7. MacKenzie EJ, Siegel SH, Shapiro S, et al. Functional recovery and medical costs of trauma: an analysis by type and severity of injury. J Trauma. 1988;28:281-295. 8. MacKenzie EJ, Morris JA, Smith GS, et al. Acute hospital costs of trauma in the United States: implications for regionalized systems of care. J Trauma. 1990;30:1096-1103. 9. Joy SA, Lichtig LK, Knauf RA, et al. Identification and categorization of and cost for care of trauma patients: a study of 12 trauma centers and 43,219 statewide patients. J Trauma. 1994;37:303-308.
Page 8 of 8 10. Weinstein MC, Stason WB. Foundations of cost-effectiveness analysis for health and medical practices. N Engl J Med. 1977;296:716-721. 11. Detsky AS, Naglie IG. A clinician's guide to cost-effectiveness analysis. Ann Intern Med. 1990;113:147-154. 12. How to read clinical journals: VII. To understand an economic evaluation (part A). Can Med Assoc J. 1984;130:1428-1434. 13. How to read clinical journals: VII. To understand an economic evaluation (part B). Can Med Assoc J. 1984;130:1542-1549. 14. Guyatt GH, Feeny DH, Patrick DL. Measuring health related quality of life. Ann Intern Med. 1993;118:622-629. 15. Torrance GW, Feeny D. Utilities and quality adjusted life years. International Journal of Technology Assessment in Health Care. 1989;5:559-575. 16. Car-Hill RA. Background material for the workshop on QALYs. Assumptions of the QALY procedure. Soc Sci & Med. 1989;14:289-297. 17. Luce BR, Elixhauser A. Estimating costs in the economic evaluation of medical technologies. International Journal of Technology Assessment in Health Care. 1990;5:57-75. 18. Baker SP, O'Neill B, Haddon W. The injury severity score: a method for describing patients with multiple injuries and evaluating emergency care. J Trauma. 1974;14:187-196. 19. EuroQoL Group. EuroQoL - a new facility for the measurement of health related quality of life. Health Policy. 1990;16:199-208. 20. Chambers LW. The McMaster Health Index Questionnaire: an update. In: Walker SR, Rosser RM, eds. Quality of Life Assessment: Key Issues in the 1990s. Dordrecht, Netherlands: Kluwer Academic Publishers; 1993. p. 131-149. 21. Dillman, DA. Mail and Telephone Surveys: The Total Design Method. New York: Wiley; 1978. 22. Drummond MF, O'Brien B, Stoddart GL, Torrance GW. Methods for the Economic Evaluation of Health Care Programs. Oxford: Oxford University Press; 1997. 23. Coyle D, Tolley K. Discounting of health benefits in the pharmacoeconomic analysis of drug therapies. Pharmacoeconomics. 1992;2:153-162. 24. Keeler EB, Cretin S. Discounting of life-saving and other nonmonetary effects. Management Science. 1983;29:300-306. 25. Briggs A, Schulpher M, Buxton M. Uncertainty in the economic evaluation of health care technologies: the role of sensitivity analysis. Health Economics. 1994;3:95-104. 26. Laupacis A, Feeny D, Detsky AS, Tugwell PX. How attractive does a new technology have to be to warrant adoption and utilization? Tentative guidelines for using clinical and economic evaluations Can Med Assoc J. 1992;146:473-481. 27. Goel V, Deber RB, Detsky AS. Nonionic contrast media: economic analysis and health policy development. Can Med Assoc J. 1989;140:389-395. 28. Churchill DN, Lemon DC, Torrance GW. Cost-effectiveness analysis comparing continuous ambulatory peritoneal dialysis to hospital hemodialysis. Med Decis Making. 1984;4:489-500. Dr. Garber is an Assistant Professor of Surgery at the University of Ottawa. Dr. Hébert is a career scientist with the Ontario Ministry of Health. Address for reprints: Dr. B.G. Garber, Division of General Surgery, Rm K-15, The Ottawa Hospital General Site, 501 Smyth Road, Ottawa, Ontario, Canada K1H -8L6; Fax: 613-737-8371; E-mail: bgarber@ogh.on.ca. J Trauma 1999 September Supplement;47(3):S99-S103 Copyright 1999 Lippincott Williams & Wilkins All rights reserved Previous Full Text References (20) Next Full Text