Treatment Pathways, Resource Use, and Costs of Endovascular Coiling Versus Surgical Clipping After aSAH
Background and Purpose— The International Subarachnoid Aneurysm Trial (ISAT) reported that endovascular coiling yields better clinical outcomes than surgical clipping at 1 year. The high cost of the consumables associated with the endovascular coiling procedure (particularly the coils) led health care purchasers to conclude that coiling was a more costly procedure overall. To examine this assumption and provide evidence for future policy, accurate and comprehensive data are required on the overall resource usage and cost of each strategy.
Methods— We provide detailed results of patient treatment pathways, resource utilization, and costs up to 24 months postrandomization for endovascular and neurosurgical treatment of aSAH. We report data on costs related to initial and subsequent procedures (ward days, ITU, equipment, staff, consumables, etc), adverse events, complications, and follow up. The data are based on a subsample of all patients randomized in ISAT, containing all patients across 22 UK centers (n=1644).
Results— There was a nonsignificant difference −£1740 (−£3582 to £32) in the total 12-month cost of treatment in favor of endovascular treatment. Endovascular patients had higher costs than neurosurgical patients for the initial procedure, for the number and length of stay of subsequent procedures, and for follow-up angiograms. These were more than offset by lower costs related to length of stay for the initial procedure. In the following 12- to 24-month period, costs for subsequent procedures, angiograms, complications, and adverse events were greater for the endovascular patients, reducing the difference in total per patient cost to −£1228 (−£3199 to £786) over the first 24 months of follow-up.
Conclusions— No significant difference in costs between the endovascular and neurosurgery groups existed at 12- or 24-month follow up.
Aneurysmal subarachnoid hemorrhage (aSAH) is a major cause of death and severe disability in relatively young patients (mean age 52), with an annual incidence of 6 to 12 per 100 000 in Western countries.1 The natural history of the disease suggests that 30% of patients will die within 24 hours of the bleed and a further 25% to 30% of patients will rebleed and die over the next 4 weeks if no treatment is adopted.2 50% of survivors will be dead by 6 months without treatment.3
Until 1992, when the Guglielmi detachable platinum coil (GDC) was introduced to Europe, there was no alternative to craniotomy and neurosurgical clipping to secure the aneurysm and prevent rebleeding. In 2002 ISAT, a large MRC-funded prospective randomized controlled trial, reported that endovascular coiling yields better clinical outcomes than surgical clipping at 1 year, and these results have significantly changed clinical practice in both UK and overseas centers.4,5
The high cost of the consumables associated with the endovascular coiling procedure (particularly the coils) led hospitals and health care purchasers to conclude that coiling was a much more costly procedure overall. To examine this assumption and provide a better evidence base for future policy, accurate and comprehensive data were required on the overall resource usage and cost of each strategy. Non–health care costs associated with each procedure are also likely to be important: long-term care for dependent survivors may be a substantial cost driver given the high prevalence of relatively young patients with severe disabilities in the treatment population, whereas any differences between treatments in employment status after aSAH may have wider social importance and may influence patient attitudes. Finally, detailed cost information is an essential prerequisite to calculating the incremental cost-effectiveness of coiling compared with clipping.
Previously published studies of the costs of aSAH concentrated on estimating hospital inpatient days using retrospective designs,6–9 and provide reasonable length of stay estimates but no information on other resource use items such as staffing, consumables, complications, readmissions, and rehabilitation. We now report the first detailed comparison of treatment pathways, resource utilization, and costs associated with endovascular coiling and surgical clipping in patients with ruptured aSAH, using data collected prospectively in the UK during the ISAT multi-country randomized trial.
Resource use data were collected prospectively alongside the Medical Research Council International Subarachnoid Aneurysm Trial (ISAT) (ISRCTN49866681), details of which are reported elsewhere.4,10–11
Before the trial began, all categories of hospital resource use likely to be important were identified. Trial case record forms (CRFs) were then designed to record the volume of resources used by each patient in each category, including type and number of hospitalizations and procedures, procedure duration in minutes, number of coils used in endovascular procedures, and lengths of stay in intensive care and in general wards. Hospital readmissions for further treatment or after complications or adverse events were documented, as was any follow up angiography or imaging. Information was also recorded on patient discharge to rehabilitation. Employment-related information was collected and is reported separately from health service costs. Some information was recorded on the domestic circumstances of patients, but not on informal care received; these data are excluded from this article.
The unit costs used in this analysis are summarized in Table 1; full details are available for downloading at www.herc.ox.ac.uk/downloads.
It was not feasible to collect detailed information for every patient on type and number of staff involved in each procedure, specific equipment, and every consumable item used during the interventions. Instead, a detailed observational study was undertaken at the Oxford center to identify the main cost drivers for each procedure. Staffing types and numbers, unit cost of coils, and cost per inpatient day were identified as the main cost drivers. A questionnaire was then designed, asking all UK centers to provide local information on the normal numbers and types of staff present for both procedures, mean cost of coils, and cost per inpatient day in an intensive therapy unit (ITU), neurology ward, and neurosurgery ward for that center. Eight of the 22 (36%) centers completed and returned the questionnaire; 2 centers subsequently merged, resulting in 7 independent responses, and these centers represented a cross-section of all the UK based ISAT centers. No significant differences between centers were found in reported theater staffing, cost of coils, or cost per inpatient ward or ITU day, and so an average was computed across all centers. Unit costs for staff time were obtained from national salary tables,12,13 and equipment costs were provided by the manufacturers. The cost per inpatient day in a district general hospital (DGH) and the cost per day in a rehabilitation clinic were taken from standard publications.12
The unit costs of equipment and consumables used by both procedures were provided by manufacturers and by the Oxford Radcliffe Hospitals NHS Trust finance department. The equivalent annual cost for equipment was calculated using standard methods at an annual rate of 5%.14
For the endovascular procedure, consumables used were assumed to be the same for all patients, but patient specific data were available on the use of thrombolytic agent and the number of coils. For the neurosurgical procedure, consumables and the cost of anesthetics were assumed to be the same for all patients and were included as a lump sum, but were varied according to whether temporary clips were used or not. The average cost for anesthetics was calculated separately for endovascular and neurosurgical patients and included with the cost of consumables.
In summary, for the endovascular procedure, theater costs were calculated as the product of procedure duration and cost per minute of staff, equipment costs, the product of number of coils used per patient and cost per coil, the cost of thrombolytic agent if used, and a cost for all other consumables. The cost per patient for neurosurgery was calculated as the product of theater time and cost per minute of staff and equipment, plus the cost of consumables. If patients underwent a second procedure, this was costed using the same methods.
The unit cost of an angiogram was provided by the hospital finance department; the cost of a hospital bed day case was included for each angiogram. These are shown in Table 1, along with the unit cost of a CT and MRI scan, taken from national sources.15
The total health care cost per patient was calculated by the volume of health service resources used multiplied by the relevant unit cost. This was then averaged across each trial arm to obtain a mean cost per endovascular or neurosurgical patient. Uncertainty around difference in mean cost was handled through nonparametric 95% confidence intervals. Proportions between groups were compared using standard χ2 test. 98% of the data were complete and therefore no attempt to impute the remaining 2% was considered. We compared baseline characteristics for patients with missing items and those with complete data and found no differences.
Patients were followed up at 2 months postintervention and on an annual basis thereafter. Long-term follow-up data continue to be collected. In this article we present 12- and 24-month postrandomization treatment pathways, resource use, and costs. All analyses are by intention to treat.
Costs are expressed in 2004 £UK with prices inflated to this base using the Hospital and Community Health Service Inflation Index where appropriate. All costs incurred over the 12- to 24-month period were discounted at a rate of 3.5%, in line with common practice.
A total of 2143 patients were recruited to the ISAT trial by 42 neurosurgical centers. In this article we focus on UK patients only, comprising 1644 patients, with 809 randomized to the endovascular group and 835 to the neurosurgery group.
Figure 1 shows the flow diagram of resource use consumption and the missing data items up to 24 months postrandomization.
Full information was available on 792 of 809 patients in the endovascular group (97.9%) and 819 of 835 patients in the surgery group (98%) at 1 year follow-up. At 2 years follow-up, complete data were available for 786 of 809 (97.1%) in the endovascular group and 816 of 835 (97.7%) in the surgery group. These constitute the final samples used in the cost analysis. However, we also report resource use results by specific resource category for all patients for whom information is available in that category.
From the original 809 randomized to endovascular treatment, 791 underwent the allocated treatment with 10 cross-overs to surgery. Of the remaining 8, 7 died before the procedure and 1 never had a procedure (Figure 1). From the original 835 randomized to surgery, 776 had the allocated treatment and 32 crossed-over to coiling. Of the remaining 27, 19 died before receiving any treatment and 8 never had a procedure.
Resource Use Associated With the First Episode of Care
Table 2 shows the resource use associated with the first episode of care, over the first 12 months of follow-up.
Mean resource use (SD) averaged over the number of patients using the resource category is reported. The mean time spent in the angiogram suite for the first endovascular intervention was estimated to be 137 (56) minutes with an average of 3.8 (2.5) coils per intervention. The average duration in the angiogram suite among crossover patients was calculated as 140 (51) minutes with a mean number of 5.1 (2.5) coils inserted.
The mean (SD) theater duration of those receiving surgery as first treatment was estimated to be 214 (77) minutes with 213 patients receiving temporary clips. Mean theater duration for cross-overs was estimated to be 207 (78) minutes with 3 patients receiving temporary clips.
Resource Use Associated With Subsequent Procedures, Complications, Adverse Events, and Follow-Up Angiograms
Table 3⇓ presents the number of patients and resource utilization for patients requiring subsequent procedures or follow-up angiograms or after an adverse event or complication, for the period from postoriginal discharge up to 1-year follow-up.
Employment-Related Resource Use
Occupational status before aSAH could not be collected for patients who were dead at 1 year follow-up (139 patients). Seventeen patients in the endovascular group and 22 in the neurosurgery group who were known to be alive at 12 months follow-up had some items of missing employment information. At 24 months follow-up these figure increased to 63 patients in the endovascular group and 84 in the neurosurgery group with some items of missing employment information.
Some patients who returned to work after aSAH had to stop within the 24-month period because of health reasons, whereas others returned to work with modified hours. To capture any possible variation in work status we calculated the number of days in paid employment since randomization for the patient. For those patients in full-time employment, it was assumed that the total number of days at work was 22 days per month (52 weeks divided by 12 months). For patients in less than full-time employment, we made the assumption that this equated to 11 days per month.
Figure 2 shows the proportion of patients who had returned on any particular day during the 24-month postrandomization period. It also presents χ2 results testing for differences at 12 and 24 months follow-up. The number of patients returning to work was significantly higher in the endovascular group compared with the neurosurgery group at 12 months follow-up, χ2=6.29, P<0.012. The difference was no longer significant at 24 months follow-up.
The mean (SD) number of days in paid employment since randomization was estimated to be 59(84) days in the endovascular group and 43(73) days in the neurosurgery group, a significant difference of 16 days (10 to 25, P<0.01) within the first year. At 24 months follow-up these figure were 148 (195) days in the endovascular group and 119 (180) in the neurosurgery group, a significant difference of 29 days (10 to 47, P<0.01) favoring the endovascular group.
Similar, but greater differences were found in the subgroup of patients who were in paid employment before the aSAH: at 12 months follow-up the difference was 24 days (13 to 35, P<0.01) whereas at 24 months follow-up the difference was estimated to be 39 days (13 to 35, P<0.01).
Table 4⇓ reports details of the costs associated with the first episode of care and subsequent care up to 12 months follow-up. Staff costs associated with interventions in the first episode of care were significantly lower in the endovascular group (£1450 per patient) than in the neurosurgical group (£2108 per patient), a difference of £658 (−£730, −£588) in favor of the endovascular group. As Table 1 showed, the cost per hour of staff input in theater was almost identical for the 2 procedures, and hence the difference in total staff costs is mainly attributable to shorter operating times rather than different types or numbers of staff present.
However, consumable costs were significantly higher in the endovascular group (£2627) than in the neurosurgical group (£901), a difference of £1726 (£1637, £1819). This difference was primarily attributable to the cost of coils.
The cost of hospital stay after operative procedures in the first episode of care was significantly lower in the endovascular group (£11 547) than in the neurosurgical group (£15 311), a difference of £3764 (−£5482, −£2099), reflecting the fact that the neurosurgical group spent more days in inpatient wards, ITU, and rehabilitation clinics compared with the endovascular patients.
After the first episode of care, the endovascular group incurred significantly higher costs related to additional procedures and check angiograms, but there was no significant difference between groups in the costs of complications and adverse events.
The mean (SD) overall total cost per patient at 12 months follow-up was estimated to be £18 436 (£15 849) and £20 176 (£21 984) in the endovascular and neurosurgery groups respectively, a nonsignificant difference of −£1740 (−£3582, £32) in favor of the endovascular patients.
Supplemental Tables I and II⇓ (available online at http://stroke.ahajournals.org) summarize the resource use and costs of subsequent care incurred from 12 to 24 months follow-up. The mean (SD) total follow-up costs from 12 to 24 months were estimated to be £613 (£4727) in the endovascular group and £131 (£895) in the surgery arm, a significant difference of £482 (£221, £888) in favor of the surgical group.
Increasing importance is attached to accurately identifying treatment pathways, resource utilization, and costs of health care interventions in all countries. Simplistic approaches which only look at narrow aspects of this, such as including only inpatient stays and omitting other important resource use categories such as subsequent procedures, readmissions, rehabilitation, and days lost from work, can result in incorrect policy decisions being made. Moreover, accurate detailed estimates of resource use and costs are necessary precursors to the calculation of cost-effectiveness of interventions, in this case, treatments for aSAH.
This article provides the first detailed estimates of patient treatment pathways, resource use, and costs up to 24 months postrandomization for the endovascular and neurosurgical treatment of ruptured aSAH. We found a nonsignificant difference (−£1740, −£3582 to £32) in the total 12-month cost of treatment for aSAH in favor of endovascular treatment. Endovascular patients had higher costs related to the intervention for both the first and any subsequent procedures, but this was more than offset by lower costs related to length of stay for the first procedure and fewer costs associated with complications and adverse events. In the following 12- to 24-month period, the costs related to subsequent procedures, angiograms, complications, and adverse events were greater for the endovascular group of patients, reducing the difference in total per patient cost to −£1228 (−£3199 to £786) for the period up to 24 months follow-up.
Our findings are in line with but significantly extend those suggested by 2 previous articles16,17 reporting on limited resource use and cost information for very small patient subsamples from single centers of the ISAT study in Australia (n=22) and Canada (n=62). Both suggested that the higher costs of the endovascular compared with neurosurgical propcedure were counter-balanced by savings from reduced length of stay.
As well as providing information on the hospital costs of endovascular and neurosurgical treatment for aSAH, this article reports employment status of patients after the aSAH and its treatment. This is important as the average age of patients in the ISAT study was 51.7 years. Our results show that at 12 months postrandomization a significantly larger proportion of patients in the endovascular group had returned to work compared with patients in the neurosurgical group; this significant difference no longer held at 24 months. The endovascular patients on average worked 16 days more than the neurosurgery patients over the first year, with a continuing significant difference of 29 days in paid employment over the 2 years postrandomization.
Unfortunately the study did not permit collection of information on use of long-term nursing and informal care, and our cost results omit these potentially considerable costs.
Accurate costing is extremely pertinent to the financing of hospital care, notably in England where the recent transition to “payment by results” anticipates that primary care trusts will commission services using a standard national price tariff, adjusted for regional variation in wages and other costs of service delivery. Inaccurate tariffs will lead to incorrect reimbursement, inappropriate incentives, and overall inefficiencies in resource allocation. The detailed cost results reported in this article, therefore, provide useful information from which national tariffs could be calculated.18
No significant difference in costs between the endovascular and neurosurgery groups existed at 12 months follow up. The higher initial and subsequent procedure costs for endovascular treatment were more than offset by lower costs related to initial procedure inpatient stay and fewer complications. In the following 12- to 24-month period, costs for subsequent procedures, angiograms, complications, and adverse events were greater for the endovascular patients, reducing the difference in total per patient cost over the 24-month period.
Our thanks go to the staff of the Diabetes Trials Unit, University of Oxford, for the data input and management; the staff of the Clinical Trial Service Unit, University of Oxford, for their invaluable help and advice, and for providing the randomization service. We would especially like to acknowledge the support of all the patients and their relatives who agreed to participate in the study and all the medical, radiological, radiographic, and nursing staff in the many participating centers (listed in Molyneux et al 2002) who have supported the trial, especially those who are not noted by name. The strength of the ISAT findings results from their willingness to take part and to continue to provide follow-up data.
Sources of Funding
The pilot phase of the study was supported by a grant from Oxford Regional Health Authority Research and Development (1994 to 97). The main trial (1997 to date) is supported by grants from: The Medical Research Council, UK; Programme Hospitalier de Recherche Clinique 1998 of the French Ministry of Health (AOM 98150) sponsored by Assistance Publique-Hôpitaux de Paris (AP-HP); the Canadian Institutes of Health Research; and the Stroke Association of the UK (for the neuropsychological assessments).
A.J.M. has a consulting and advisory agreement with Micrus, Inc, a manufacturer of detachable platinum coils, with stock interest in the company. R.S.C.K. and A.J.M. have received support for travel to meetings from Boston Scientific.
- Received February 28, 2007.
- Revision received June 6, 2007.
- Accepted June 13, 2007.
Linn FHH, Rinkel GJE, Algra A, vanGijn J. Incidence of subarachnoid hemorrhage - role of region, year, and rate of computed tomography: A meta-analysis. Stroke. 1996; 27: 625–629.
Wermer MJH, van der Schaaf IC, Velthuis BK, Algra A, Buskens E, Rinkel GJE. Follow-up screening after subarachnoid haemorrhage: Frequency and determinants of new aneurysms and enlargement of existing aneurysms. Brain. 2005; 128: 2421–2429.
Molyneux AJ, Kerr RSC, Yu LM, Clarke M, Sneade M, Yarnold JA. International subarachnoid aneurysm trial (ISAT) of neurosurgical clipping versus endovascular coiling in 2143 patients with ruptured intracranial aneurysms: A randomised comparison of effects on survival, dependency, seizures, rebleeding, subgroups, and aneurysm occlusion. Lancet. 2005; 366: 809–817.
Reed SD, Blough DK, Meyer K, Jarvik JG. Inpatient costs, length of stay, and mortality for cerebrovascular events in community hospitals. Neurology. 2001; 57: 305–314.
International Subarachnoid Aneurysm Trial (ISAT) website. Trial protocol available at: Http://www.Surgery.Ox.Ac.Uk/nvru/isat.
Molyneux A, Kerr R. Protocol 99prt18 international subarachnoid aneurysm trial (ISAT). 2006.
Netten A, Dennett J, Knight J. Unit costs of health and social care 2004. PSSRU; 2004.
Department of Health. Advanced letter. Nurses’ and midwives pay. 2005.
Drummond MF, O’Brien B, Stoddart GL, Torrance GW. Methods for the Economic Evaluation of Health Care Programmes. Oxford: Oxford University Press; 1997.
Department of Health. Nhs reference costs: Data publications. 2004.
Department of Health. Payment by results available at: Http://www.Dh.Gov.Uk/policyandguidance/organisationpolicy/financeandplanning/nhsfinancialreforms/fs/en.;2006.