Acute treatment costs of intracerebral hemorrhage and ischemic stroke in Argentina Introduction In Argentina, stroke is the third most common cause of death and one of the most important causes of adult disability (1). Every year, approx- imately 196 per 100,000 individuals experience a stroke, and 52 per 100,000 die because of the disease (2, 3). In a large, international, comparative study on stroke mortality between 1968 and 1994 using data from the World Health Organization, Acta Neurol Scand 2009: 119: 246���253 DOI: 10.1111/j.1600-0404.2008.01094.x Copyright �� 2008 The Authors Journal compilation �� 2008 Blackwell Munksgaard ACTA NEUROLOGICA SCANDINAVICA Christensen MC, Previgliano I, Capparelli FJ, Lerman D, Lee WC, Wainsztein NA. Acute treatment costs of intracerebral hemorrhage and ischemic stroke in Argentina. Acta Neurol Scand 2009: 119: 246���253. �� 2008 The Authors Journal compilation �� 2008 Blackwell Munksgaard. Background and purpose ��� Stroke is the third leading cause of death in Argentina, yet little information exists on the acute treatment provided for stroke or its costs. This study estimates the national costs of the acute treatment of first-ever intracerebral hemorrhage (ICH) and ischemic stroke (IS) in Argentina. Methods ��� Retrospective hospital- based inception study design using data on resource use and costs from high-volume stroke centers in Argentina, and published population- based incidence data. Treatment provided at two large urban hospitals were evaluated in all patients admitted with a first-ever stroke between 1 January 2004 and 31 August 2006, and costs were assigned using appropriate unit cost data for all resource use. Cost estimates in Argentinian pesos were converted to US dollars ($) using the 2005 purchasing power parity index. National costs of acute treatment for incident strokes were estimated by extrapolation of average costs estimates to national incidence data. Assumptions of the average cost of stroke treatment on a national scale were examined in sensitivity analysis. Results ��� The acute care of 167 patients with stroke was thoroughly evaluated from hospital admission to hospital discharge. Mean length of hospital stay was 35.4 days for ICH and 13.0 days for IS. Ninety-one percent of the patients with ICH and 68% of the patients with IS were admitted to an ICU for a mean length of stay (LOS) of 12.9 20.3 and 3.6 5.9 days respectively. Mean total costs of initial hospitalization were $12,285 (SD 14,336) for ICH and $3888 (SD 4018) for IS. Costs differed significantly by Glasgow Coma Scale (GCS) score at admission, development of pneumonia and infections during hospitalization, and functional outcome at hospital discharge. Aggregate national healthcare expenditures for acute treatment of incident ICH were $194.2m (range 97.1���388.4) and $239.9m for IS (range 119.9���479.7). Conclusion ��� The direct hospital costs of incident ICH and IS in Argentina are substantial and primarily driven by stroke severity, in-hospital complications and clinical outcomes. With the expected increase in the incidence of stroke over the coming decades, these results emphasize the need for effective preventive and acute medical care. M. C. Christensen1, I. Previgliano2, F. J. Capparelli3, D. Lerman2, W. C. Lee4, N. A. Wainsztein3 1Global Development, Novo Nordisk A ��� S, Bagsvaerd, Denmark 2Hospital de Agudos J.A. Fern��ndez, Buenos Aires, Argentina 3Institute for Neurological Research, FLENI, Buenos Aires, Argentina 4HERQuLES ��� Abt Bio-Pharma Solutions, Inc., Bethesda, MD, USA Key words: intracerebral hemorrhage ischemic stroke treatment hospital costs outcomes Argentina Michael C. Christensen, Global Development, Novo Nordisk A ��� S, Krogshoejvej 55, 2880 Bagsvaerd, Denmark Tel.: +45 44420557 Fax: +45 44427390 e-mail: mcrc@novonordisk.com Accepted for publication July 17, 2008 246

stroke mortality rates were twice as high in Argentina, Chile and Uruguay as in the USA and Europe (4). Intracerebral hemorrhage (ICH), which is associated with a substantially worse prognosis than ischemic stroke (IS), accounts for 20���30% of all strokes in Argentina (2, 5). Despite the significant consequences of stroke to the public health in Argentina, there is very little data published on the exact form of acute treatment provided or the hospital costs associ- ated with its treatment. A number of studies in Argentina have examined ethnic differences in stroke patterns and major risk factors for the different stroke subtypes one study has examined prevalence and incidence of stroke (2, 6, 7) and two studies have examined the clinical outcomes of IS and ICH (8, 9). Evidence of the actual treatment provided for stroke and its economic implications is, however, critical for judging the adequacy of current treatment practices, the potential for new therapeutic interventions and hence optimal decision making in healthcare management and resource allocation. From a public health perspective, estimates of the differ- ences in treatment costs of patients with hemor- rhagic and ischemic stroke are particularly important as they form the basis for understand- ing the future economic challenges presented by an increasing incidence and prevalence of the two major types of stroke. To better understand current treatment practices for stroke and their costs, this study considers both of these parameters in patients admitted to a public or private hospital for a first-ever stroke (ICH or IS) in Argentina between 2004 and 2006. Subjects and methods Patient population As no national stroke registry exists in Argentina on the annual number of hospital admissions for stroke, their treatment or their treatment costs, we designed a retrospective hospital-based incep- tion study retrieving data from hospital medical records at two Argentinian hospitals receiving a high volume of patients with stroke. On the basis of the resource utilization observed at these hospitals, we assigned appropriate unit costs to estimate average hospital costs per patient with stroke, and on the basis of published literature on the incidence of stroke in Argentina, we extrapolate average treatment costs to the national level. We studied patients with a first- ever diagnosis of ICH or IS, between 1 January 2004 and 31 August 2006 admitted to Hospital de Agudos J.A. Fernandez and FLENI hospital in Buenos Aires, Argentina. We included patients aged 21 years or older who were admitted for acute treatment with a confirmed diagnosis of ICH or IS through computed tomography (CT) scan or magnetic resonance imaging (MRI). ICH and IS were defined according to the ICD-10 codes I60���I69. To assess treatment patterns and costs in an incident stroke population, we excluded patients with ICH and IS with a previous stroke. The institutional ethics commit- tees at the two hospitals approved the study. Data collection The following baseline data were collected from medical records: age, gender, recognized risk factors for stroke (arterial hypertension, alcohol intake, smoking, diabetes mellitus and dyslipide- mia) and presence of co-morbidities. Once admit- ted to the emergency room (ER), glucose level, systolic ��� diastolic blood pressure, body tempera- ture, GCS score and CT scan findings were also recorded. Neuroradiological findings were deter- mined in the initial CT scan and classified according to location and volume of hematoma, and midline shift (displacement of the septum pellucidum across midline). Neurological altera- tions (diplopia, right ��� left hemiparesia, cephalea, right ��� left hemiplegia, aphasia and dysphasia) at hospital admission were also recorded. Data were collected on the treatment provided during initial hospitalization, including patient disposition and length of stay [ER, ICU and general ward (GW)], diagnostic tests performed at admission and during the hospital stay (CT scan, MRI, echo- cardiography, digital angiography, carotid echo Doppler and chest X-ray), surgical interventions, intensive care procedures, radiology examina- tions, laboratory tests, rehabilitation and medical therapy. Clinical outcomes were assessed as in- hospital mortality, functional outcome at hospital discharge and discharge destination. Functional outcome was assessed using the modified Rankin Scale (mRS) and discharge destination was recorded for each patient as either death or transfer to another hospital, nursing home, reha- bilitation facility, the patient��s home or other. We attempted to minimize any potential missing values at the data collection stage by differenti- ating the true absence of values (i.e. no resource utilization) from missing values (i.e. no data available, although there is evidence for resource utilization). On a few rare occasions where missing values were present, mean values were used to replace missing values. Hemorrhage and ischemic stroke in Argentina 247