Published online before print September 30, 2004, doi: 10.1161/01.STR.0000143453.78005.44
(Stroke. 2004;35:2477.)
© 2004 American Heart Association, Inc.
Original Contributions |
Immediate Computed Tomography Scanning of Acute Stroke Is Cost-Effective and Improves Quality of Life
From the Division of Clinical Neurosciences (J.M.W., S.K., S.L., P.S.), Western General Hospital, Edinburgh, UK; and the Health Economics Research Unit (J.S., J.C.), University of Aberdeen, Foresterhill Hospital, Aberdeen, UK.
Correspondence to Dr J.M. Wardlaw, Division of Clinical Neurosciences, Western General Hospital, Crewe Rd, Edinburgh, EH4 2XY, UK. E-mail jmw{at}skull.dcn.ed.ac.uk
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
Background and Purpose— Stroke is very common, but computed tomography (CT) scanning, an expensive and finite resource, is required to differentiate cerebral infarction, hemorrhage, and stroke mimics. We determined whether, and in what circumstances, CT is cost-effective in acute stroke.
Methods— We developed a decision tree representing acute stroke care pathways populated with data from multiple sources. We determined the effect of diagnostic information from CT scanning on functional outcome, length of stay, costs, and quality of life during 5 years for 13 alternative CT strategies (varying proportions and types of patients and rapidity of scanning).
Results— For 1000 patients aged 70 to 74 years, the policy "scan all strokes within 48 hours" cost £10 279 728 and achieved 1982.3 quality-adjusted life years (QALYs). The most cost-effective strategy was "scan all immediately" (£9 993 676 and 1982.4 QALYs). The least cost-effective was "scan patients on anticoagulants and those in a life-threatening condition immediately and the rest within 14 days" (£12 592 666 and 1931.8 QALYs). "Scan no patients" reduced QALYs (1904.2) and increased cost (£10 544 000).
Conclusion— Immediate CT scanning is the most cost-effective strategy. For the majority of acute stroke patients, increasing independent survival by correct early diagnosis, ensuring appropriate subsequent treatment and management decisions, reduced costs of stroke and increased QALYs.
Key Words: cerebrovascular disorders • computed tomography • cost-benefit analysis • decision analysis • intracerebral hemorrhage • stroke
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Common medical conditions such as stroke require streamlined readily available services. A typical hospital (catchment population 500 000) admits 1 to 2 patients with suspected stroke every day and assesses 10 to 15 as outpatients per week.
Eighty percent of strokes are ischemic and 15% hemorrhagic. Clinical examination alone cannot distinguish ischemic from hemorrhagic stroke; computed tomography (CT) or magnetic resonance brain imaging is required.1 Brain imaging determines management decisions such as antiplatelet or thrombolytic drugs for acute stroke2,3 and antithrombotic treatment for secondary prevention.4,5 However, CT scanning is expensive and, in some countries, a limited resource. Thrombolysis only applies to some patients in some hospitals. The effect of aspirin on long-term functional outcome, although universally available, is marginal, and some patients with hemorrhagic stroke have received some aspirin without apparent adverse effect.2
Increasing evidence points to a high rate of recurrent primary intracerebral hemorrhage (PICH) and death after initial PICH,6 possibly worsened by antithrombotic therapy.7 Thus, although long-term antiplatelet therapy might reduce the risk of ischemic vascular diseases,5 the balance of risk and benefit for patients with PICH is uncertain and possibly worsened by these drugs.8
The existing studies of cost-effectiveness of CT in stroke either were performed 
20 years ago9,10 or contained little information on how costs were derived.10,11 Thus, it was unclear whether CT scanning was actually cost-effective in the management of most strokes, and if not, whether there were particular groups of patients for whom it was cost-effective and how rapidly they should be imaged. Because a randomized controlled trial to determine the cost-effectiveness of CT in stroke would now be unethical,2 we used modeling techniques to estimate the expected benefits and costs associated with different CT scanning policies.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Model Design and Decision Tree
We developed a decision tree showing key steps in the pathways of care after stroke (Figure) and populated it with data reflecting true proportions of patients with ischemic or hemorrhagic stroke or stroke mimics. We did not examine subarachnoid hemorrhage. The tree incorporated key decisions and events in the pathway of stroke care and their effects on outcome for patients admitted to the hospital with first-ever stroke during a 5-year period (Figure). Development of the decision tree and identification of data at each node were performed blind to knowledge of the effect of treatments on length of stay, cost, or quality-adjusted life years (QALYs). By examining the effect of CT diagnosis on treatment, and hence, the effect on functional outcome, we were able to translate functional outcome into length of stay in hospital and therefore the cost of care. Given the complex clinical scenarios and data limitations, we constructed a simple, custom-designed, deterministic model12 rather than a more complex Markov model and used a conventional approach to decision analysis.
|
CT Strategies
We devised CT scanning "strategies" to reflect different priorities and scan availabilities. We used "scan all patients within 48 hours of stroke" as the base comparator. This is the standard in UK guidelines,13 which, although not always achieved,14 nonetheless represents "best current practice." Twelve alternative CT scanning strategies were drawn (Table 1) to reflect typical patterns of CT use with differing CT availability and use and availability of acute treatments. These were designed to be relevant during at least the next 5 years.
|
Data required for the base case analysis of the main comparator included parameter estimates for each node in the decision tree, clinical and final outcomes, effect of treatments, and costs. Data were obtained from many sources (Table 2), including systematic reviews of: (1) the accuracy of clinical diagnosis of stroke;15 (2) CT scan diagnosis (stroke versus not stroke15 and infarct from hemorrhage16); (3) antithrombotic drugs for primary treatment2,4 and secondary prevention of ischemic stroke5 and after intracranial hemorrhage;8 (4) and thrombolysis.3 Estimates for outcomes of PICH and treatments17 (stopping antithrombotic treatment18 or surgical removal of the hematoma19) and the few patients presenting with tumors, infections, and other stroke mimics were also determined.15
|
Outcomes
Stroke category at admission was incorporated using the Oxfordshire Community Stroke Project (OCSP) Classification:20 total anterior circulation syndrome (TACS), partial anterior circulation syndrome, posterior circulation syndrome, or lacunar syndrome (LACS). Patients with PICH were classified as TACS or non-TACS. Outcomes were quantified using the modified Rankin scale (mRS) as alive and independent, dependent, or dead21 at 6, 12, and 24 months after stroke.
Length of stay for first episode of care by OCSP category at admission and functional outcome at 6, 12, and 24 months were obtained from an urban teaching hospital Stroke Registry22 of data on all patients with suspected stroke, transient ischemic attack, or retinal ischemic symptoms collected between November 1990 and May 1999 (catchment population 500 000).
Five-year hospital length of stay information for patients with first stroke in the Stroke Registry was obtained using record linkage performed by the Information and Statistics Division of the Scottish Executive using the General Acute Inpatient and Day Case Discharge data set (SMR01).23 This yielded data on length of stay by all episodes of care and different hospital settings for 1778 patients (96%) from the Stroke Registry data set.15
Life years were estimated up to 5 years after first-ever stroke in 6 monthly intervals for 1854 patients from Stroke Registry survival data22 using Cox’s proportional hazards regression analysis, taking into consideration the age distribution for stroke patients aged 45 to 89 in 5-year age bands and OCSP classification. This resulted in, for example, a mean survival of 3.997 years for a patient with TACS aged 45 to 49 years compared with 1.457 years at 75 to 79 years; for a patient with LACS aged 45 to 49, the mean survival was 4.764 years compared with 3.764 years at ages 75 to 79.
Utility weights were derived by functional outcome for the Stroke Registry cohort using data24 from the EuroQol (EQ)5D25 administered at 18 months after stroke. Values were very similar to those obtained in a subsequent systematic review.26 Utility weights estimated for the EQ5D states using the time trade-off technique (TTO) for the UK population aged >60 years were assigned to each functional outcome state.27 TTO weights were alive and independent, mean 0.78 (95% CI, 0.73 to 0.82); and alive and dependent, mean 0.34 (95% CI, 0.23 to 0.45).15
QALYs were calculated by multiplying the number of life years resulting from the scanning strategy by TTO weights to determine quality of life associated with being independent or dependent.15,24
Current Access to CT Scanning
We obtained data on access to CT for stroke from all Scottish Radiology departments,15 including normal hours of scanning and rapidity of access to CT for stroke within and outside normal office hours. This enabled us to ensure that the 12 scanning strategies reflected correctly the range of current access for CT scanning and provided an estimate of the likely additional resources required to move from a slower to a faster scanning strategy. However, the data were not directly used in the model.15
Costs were determined by the National Health Service in year 2000 prices. CT scan costs within and outside normal working hours were estimated in detail for a large teaching and 2 general hospitals (urban and rural) in Scotland to provide a range of suitably stylized exemplars against which other hospitals could compare their own costs and CT service.15 The mean "low" and "high" CT costs, based on minimum and maximum times to scan a stroke patient, in a large teaching hospital, for example, were: normal working hours £43 ("low" £30 to "high" £72); and after hours £79 (£55 to £173).15 Inpatient stay costs were derived from government-published data by mean cost per bed day for a teaching hospital (£239), large general hospital (£217), and a long in-hospital stay (£116) using the relevant inpatient acute specialties costs (medical and neurology),23,28 including general investigations and treatments. Again, these costs provide exemplars against which other stroke service providers can judge the likely cost-effectiveness of their CT scanning approaches.
Implementing the Model and Analyses
The primary analysis was conducted for a cohort of 1000 patients aged 70 to 74 years and repeated for 1000 patients aged 60 to 64 years and 80 to 84 years in teaching urban and rural general hospitals.15 The cost-effectiveness of the 12 CT scanning strategies was estimated by assessing the incremental costs and outcomes of each strategy compared with the main comparator.
Sensitivity Analysis
One-way sensitivity analysis was undertaken on: CT scan costs (by type of hospital), the proportion of suspected stroke proven to have a stroke, the proportion of patients with PICH causing TACS, the sensitivity and specificity of CT scans, and the costs of inpatient care.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Average length of stay for first stroke was closely related to functional outcome at 6 months: alive and independent (mRS 0 to 2) 14 days, alive but dependent (mRS 3 to 5) 51 days, and dead (mRS 6) 33 days.15
The proportion of patients scanned during normal working hours ranged from none to 90% and after hours from none to 76%, depending on the scanning strategy. The cost of CT scanning 1000 patients aged 70 to 74 years for the comparator at a large teaching hospital was £46 728, varying with the scanning strategy from £0 (S12 [see Table 1 for definitions of S1–S12]) to £70 676 (S1).
The total cost of length of stay per 1000 patients aged 70 to 74 years by scanning strategy ranged from £9 923 000 for S1 to £12 546 000 for S5. In general, the cost of length of stay for the comparator (£10 233 000) was lower than strategies that scanned patients at later times (S12, S11, S8, S4, S10, S9, or S5).
The comparator achieved 1982.3 QALYs for 1000 patients aged 70 to 74 years at a cost of £10 279 728 (Table 3). "Scan all patients immediately" (S1) increased QALYs by 0.1 to 1982.4 and reduced cost (£9 993 676). Scanning at later times generally increased costs and reduced QALYs. The greatest loss of QALYs occurred when patients were not scanned at all (S12), resulting in a loss of 78.1 to 83.3 QALYs.
|
Ranking the scanning strategies according to increasing total cost (Table 3) shows that S1 (scan all patients immediately) is the least costly strategy, closely followed by S6 and S2, which involved scanning the majority of patients within 24 hours. S1 remained the dominant strategy when the analysis was repeated for 60- to 64-year-olds and for 80- to 84-year-olds.
Sensitivity analyses of: (1) the proportions of suspected stroke that proved to be a stroke; (2) patients with ischemic or hemorrhagic stroke or PICHs that were TACS; (3) the utility weights per outcome; (4) CT scanning cost (minimum and maximum); (5) and the effect of delay to CT scanning on time to starting aspirin did not change the ranking of strategies. However, a lower cost per day changed the order of the cost-effectiveness of scan strategies but only if inpatient costs per day were at least 10% lower than our estimate, when the costs of the comparator and SI became similar. If the cost per day was even lower, S1 ceased to cost less than the comparator, and because of the very small expected difference in QALYs, the incremental cost per QALY of "scanning all immediately" rose rapidly. If the cost of the delayed scanning options was significantly lower than assumed currently in the model (eg, if delayed scanning was not associated with significantly longer inpatient stays), then delayed scanning strategies could cost less than strategy S1.
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The study indicates that of 13 possible imaging strategies, a policy of "CT scan all patients immediately" is dominant. Although the costs of CT scanning are highest for this strategy because of more scanning occurring after hours, these higher costs are offset by savings in the length of inpatient stay because many management decisions and better outcomes depend on accurate early diagnosis of stroke. The costs of after-hours scanning would have to rise markedly (well above the current maximum costs) to outweigh the cost savings in length of stay on current bed occupancy cost figures.
The results were sensitive to a fall in the cost of inpatient days. The unusual sensitivity of the incremental cost-effectiveness estimates is largely a product of the very small difference in outcome between a strategy of "scan all immediately" and 1 of "scan all within 48 hours of admission to hospital." Because the majority of patients have cerebral infarction, the main treatment is aspirin, and there is no good evidence of a time dependency of the effect of aspirin up to 48 hours after stroke.2 It is perhaps not surprising that the difference between S1 and the comparator is sensitive to the cost of inpatient care.
Have we underestimated the cost of inpatient care for stroke? This is unlikely because: (1) the standard costs of inpatient care used do not include the intensive nursing attention often required for stroke; and 2) the length of stay data came from a population with a larger proportion of milder strokes than may be typical in other large hospitals. We studied first stroke only because data on recurrent stroke and outcomes were lacking. Although we did not include other costs of patient care like family doctor visits or costs to the patients and their families, these all simply increase the cost of stroke care and will be higher the greater the dependency (eg, after recurrent stroke). Thus, if anything, it is likely that we have underestimated the cost-effectiveness of CT by not including these costs.
What are the limitations? Much of the data in the model were obtained in Scotland, but Scotland has a geographically defined population, centralized hospital statistics, and good record linkage, making large-scale epidemiological studies feasible and accurate. The range of CT scanning for stroke provided in Scotland is similar to other parts of Europe29 and better on average than in the rest of the United Kingdom (C Squire, Royal College of Radiologists Audit, personal communication, 2002). The costs of CT scanning and hospital bed occupancy may differ from other parts of the world, but it is unlikely that the ratio of CT-to-bed-occupancy cost will be very different, and it is the latter that is important. Also, sensitivity analyses showed that even if CT costs were 4x higher (eg, normal working hours average cost of £42.90 versus after-hours "high" cost of £173.46 in a teaching hospital), "scan all immediately" was still the dominant strategy. Finally, the costs we used are given so that other health care providers can use these as exemplars against which to compare their own CT scanning-to-inpatient costs. Data on long-term survival and health outcomes after stroke were lacking. Therefore, we were only able to examine survival during 5 years rather than the life spans of patients. This is important for patients <65 years of age. Further long-term studies on health outcomes and costs of care after stroke are needed.
We have assumed that 4% of patients would reach the hospital in time to be considered for thrombolysis (included in scanning strategies S6–S9), but if 10% of patients were eligible for thrombolysis, then the CT scanning strategy "scan all immediately" would become even more cost-effective, assuming that the estimate of thrombolysis treatment effect is correct.3
The differences in QALYs and costs between strategies may seem marginal. However, these small differences become substantial across the whole population. Because patients who achieved independence at 6 months had a much shorter average length of stay (14 days) than those who remain dependent (51 days); even a marginal shift in the proportion of patients from dependency to independency (eg, 13 of 1000 following aspirin) would reduce length of stay duration by 445 days and cost by £106 355 per 1000 patients, or £12 762 600 saved per year in the United Kingdom alone. The balance is such that failure to use imaging judiciously could considerably increase the cost of stroke care. Our survey of radiology departments in Scotland identified that the single most major deficit to improving access to CT for stroke was a shortage of radiologists, closely followed by too few radiographers, rather than not enough CT scanners, which was the problem 10 to 15 years ago.15 Anecdotally, we understand the problem to be the same in many developed countries. Developing countries are likely to have to first overcome the problem of too few CT scanners. Countries with poor provision of CT for stroke, for whatever reason, are basically operating similar to the "scan no one" (S12) strategy in this model and so would benefit in terms of reduced costs and improved QALYs by improving access to CT. The cost of installing CT scanners (or training more radiologists) would quickly be recouped in reduced costs of stroke to health services.
Received May 3, 2004; revision received July 10, 2004; accepted July 19, 2004.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
1. Weir CJ, Murray GD, Adams FG, Muir KW, Grosset DG, Lees KR. Poor accuracy of stroke scoring systems for differential clinical diagnosis of intracranial hemorrhage and infarction. Lancet. 1994; 344: 999–1002.[CrossRef][Medline] [Order article via Infotrieve]
2. Chen Z, Sandercock P, Pan H, Counsell C, Collins R, Liu L, Xie JX, Warlow C, Peto R. Indications for early aspirin use in acute ischemic stroke. A combined analysis of 40 000 randomized patients from the Chinese Acute Stroke Trial and the International Stroke Trial. Stroke. 2000; 31: 1240–1249.
3. Wardlaw JM, Zoppo G, Yamaguchi T, Berge E. Thrombolytic therapy versus control in acute ischaemic stroke. Cochrane Database Syst Rev. 2003; 3: CD000213.[Medline] [Order article via Infotrieve]
4. Gubitz G, Sandercock P, Counsell C. Anticoagulants for acute ischaemic stroke. Cochrane Database Syst Rev. 2004; 3: CD000024[Medline] [Order article via Infotrieve]
5. Antithrombotic Trialists Collaboration. Prevention of death, myocardial infarction and stroke by antiplatelet therapy in high risk patients. Br Med J. 2002; 324: 71–86.
6. Bailey RD, Hart RG, Benavente O, Pearce LA. Recurrent brain hemorrhage is more frequent than ischemic stroke after intracranial hemorrhage. Neurology. 2001; 56: 773–777.
7. He J, Whelton PK, Vu B, Klag MJ. Aspirin and risk of hemorrhagic stroke. A meta-analysis of randomized controlled trials. J Am Med Assoc. 1998; 280: 1930–1935.
8. Keir S, Sandercock PAG, Wardlaw J. Antithrombotic therapy in patients with any form of intracranial hemorrhage: a systematic review of the available controlled studies. Cerebrovasc Dis. 2002; 14: 197–206.[CrossRef][Medline] [Order article via Infotrieve]
9. Larson EB, Omen GS, Loop JW. Computed tomography in patients with cerebrovascular disease: impact of a new technology on patient care. Am J Roentgenol. 1978; 131: 35–40.[Abstract]
10. Britton M, Jonsson E, Marke LA, Murray V. Diagnosing suspected stroke. A cost-effectiveness analysis. Int J Technol Assess Health Care. 1985; 1: 147–158.
11. van der Meulen JH, Limburg M, van Straten A, Habbema JDF. Computed tomographic brain scans and antiplatelet therapy after stroke. A study of quality of care in Dutch hospitals. Stroke. 1996; 27: 633–638.
12. Weinstein MC, Fineberg HV. Clinical Decision Analysis. Philadelphia, Pa: WB Saunders Co Inc; 1980.
13. Department of Health. National Service Framework for Older People. Available online at http://www.doh.gov.uk/nsf/olderpeople.htm. 2001. Accessed July 4, 2004.
14. Ebrahim S, Redfern J. Stroke Care—A Matter of Chance. A National Survey of Stroke Services. London, UK: The Stroke Association; 1999.
15. Wardlaw JM, Keir SL, Seymour J, Lewis S, Sandercock PA, Dennis MS, Cairns J. What is the best imaging strategy for acute stroke? Health Technol Assess. 2004; 8: iii, ix–x, 1–180.[Medline] [Order article via Infotrieve]
16. Wardlaw JM, Keir SL, Dennis MS. The impact of delays in computed tomography of the brain on the accuracy of diagnosis and subsequent management in patients with minor stroke. J Neurol Neurosurg Psychiatry. 2003; 74: 77–81.
17. Counsell C, Boonyakarnkul S, Dennis M, Sandercock P, Bamford J, Burn J. Primary intracerebral hemorrhage in the Oxfordshire Community Stroke Project. 2. Prognosis. Cerebrovasc Dis. 1995; 5: 26–34.[CrossRef]
18. Bertram M, Bonsanto M, Hacke W, Schwab S. Managing the therapeutic dilemma: patients with spontaneous intracerebral hemorrhage and urgent need for anticoagulation. J Neurol. 2000; 247: 209–214.[CrossRef][Medline] [Order article via Infotrieve]
19. Fernandes HM, Gregson B, Siddique S, Mendelow AD. Surgery in intracerebral hemorrhage. The uncertainty continues. Stroke. 2001; 31: 2511–2516.
20. Bamford J, Sandercock P, Dennis M, Burn J, Warlow C. Classification and natural history of clinically identifiable subtypes of cerebral infarction. Lancet. 1991; 337: 1521–1526.[CrossRef][Medline] [Order article via Infotrieve]
21. Rankin J. Cerebral vascular accidents in patients over the age of 60. II Prognosis. Scot Med J. 1957; 2: 200–215.[Medline] [Order article via Infotrieve]
22. Counsell C, Dennis MS, McDowall M, Warlow C. Predicting outcome after acute and subacute stroke. Development and validation of new prognostic models. Stroke. 2002; 33: 1041–1047.
23. National Health Service in Scotland Information and Statistics Division (NHS, ISD). Scottish Health Service Costs, Year Ended March 31, 2000. Edinburgh, UK: ISD Publications; 2000.
24. Dorman P, Dennis M, Sandercock P; United Kingdom Collaborators in the International Stroke Trial (IST). Are the modified "simple questions" a valid and reliable measure of health related quality of life after stroke? J Neurol Neurosurg Psychiatry. 2000; 69: 487–493.
25. EuroQol Group and Rotterdam. A Measure of Health-Related Quality of Life Developed by the EuroQol Group. Rotterdam, The Netherlands: EuroQol Group; 1996.
26. Post PN, Stiggelbout A, Wakker P. Utility of health states after stroke. A systematic review of the literature. Stroke. 2001; 32: 1425–1429.
27. MVH Group. The Measurement and Valuation of Health. Final Report on the Modeling of Valuation Tariffs. York, UK: University of York, Centre for Health Economics; 1995.
28. NHS Executive. Reference Costs, 2000. Leeds, UK: NHS Executive. Available online at www.doh.gov.uk/nhsexec/refcosts.htm. Accessed July 4, 2004.
29. Weir NU, Sandercock PA, Lewis SC, Signorini D, Warlow CP. Variations between countries in outcome after stroke in the International Stroke Trial (IST). Stroke. 2001; 32: 1370–1377.
This article has been cited by other articles:
 |
|
 |
|
  R. I. Lindley Is Intraarterial tPA Within 3 Hours the Treatment of Choice for Selected Stroke Patients?: No Stroke, July 1, 2009; 40(7): 2613 - 2614. [Full Text] [PDF]
|
|
|
|
 |
|
 N. Losseff, D. Ames, G. Cloud, G. Cluckie, P. Gompertz, B. Grant, H. Markus, and M. M Brown The London stroke strategy BMJ, June 29, 2009; 338(jun29_2): b2616 - b2616. [Full Text]
|
|
|
|
 |
|
 R. B. Gunderman and P. J. Mehta Continue Drawing Back the Curtain: Educating the Community About Radiology Am. J. Roentgenol., May 1, 2009; 192(5): 1388 - 1390. [Full Text] [PDF]
|
|
|
|
 |
|
 E. M. de Lucas, E. Sanchez, A. Gutierrez, A. G. Mandly, E. Ruiz, A. F. Florez, J. Izquierdo, J. Arnaiz, T. Piedra, N. Valle, et al. CT Protocol for Acute Stroke: Tips and Tricks for General Radiologists1 RadioGraphics, October 1, 2008; 28(6): 1673 - 1687. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R I Lindley Commentary on NICE guidelines for alteplase for the treatment of acute ischaemic stroke Heart, December 1, 2007; 93(12): 1617 - 1618. [Full Text] [PDF]
|
|
|
|
 |
|
 C.J. Dickinson Intracerebral haemorrhage revisited QJM, November 1, 2007; 100(11): 715 - 719. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Markus Improving the outcome of stroke BMJ, August 25, 2007; 335(7616): 359 - 360. [Full Text] [PDF]
|
|
|
|
 |
|
 P. K. Myint, S. L. Vowler, O. Redmayne, and R. A. Fulcher Utilisation of diagnostic computerised tomography imaging and immediate clinical outcomes in older people with stroke before and after introduction of the National Service Framework for older people. A comparative study of hospital-based stroke registry data (1997-2003): Norfolk experience. Age Ageing, July 1, 2006; 35(4): 399 - 403. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Steger, J. Slany, and C. Stollberger Comment on stroke with atrial fibrillation: data from the Austrian stroke registry: reply Eur. Heart J., April 2, 2005; 26(8): 848 - 848. [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||