Abstract
Rationale
Because of its rapidly-growing and aging populations, the global burden of stroke will be felt most strongly in Asia. Data from Asia are scarce. Even rarer are data from multiethnic populations.
Aim
The Singapore Stroke Registry is a countrywide registry of risk factors, stroke subtypes, management and outcome of incident and recurrent stroke in multiethnic Singapore.
Methods and design
Using a cold-pursuit design, potential subjects are identified from medical claims, hospital discharge summaries and death certificates. Standardized methods are used for case ascertainment and data recording by trained staff into standardized case report forms. All ages and both genders are included.
Study outcome
Stroke incidence, subtypes and mortality will be calculated. Trend data will be obtained. Inter-ethnic differences will be explored.
Discussion
The Registry has many features of an ‘ideal’ incidence study. Previously unavailable countrywide data on an Asian population will be acquired. The unique data on inter-ethnic differences will help fill knowledge gaps in stroke epidemiology.
Introduction and rationale
Stroke is a major cause of death and disability, with 16·9 million first stokes, 5·9 million stroke-related deaths and 102 million Disability-Adjusted Life-Years lost in 2010 (1). Stroke incidence is the best measure of the impact of primary prevention strategies; stroke outcome reflects stroke severity and stroke services effectiveness (2). While good quality data are available in the developed world, such data are lacking Asia outside of China, Japan and Korea (3). With its rapidly growing and aging populations, the bulk of the challenges due to stroke in the coming years will be in Asia. There are few studies in multiethnic populations (4–8), none in Asia. Few countries can obtain nationwide data (9).
Singapore is an economically-developed, small, tropical island city-state with a population of 5·5 million people. Of the 3·9 million residents (citizens and permanent residents), 74·3% are Chinese, 13·3% Malays, 9·1% Indians and 3·3% Other Races (10). Stroke is a major health-care problem in Singapore (11). It is the fourth leading cause of death, highest among Malay women (12). In 2004, it was the second leading cause of premature mortality, third cause of disease burden, and seventh cause of disability (13). Even after rehabilitation, many stroke survivors have poor physical function and health-related quality of life (14).
Community-based surveys revealed an age and gender-standardized stroke prevalence among adults ≥50 years old of 3·65%, lowest among Malay women (15), and incidence of 1·8/100 person-years, highest among Malay women (16). In single-hospital studies, 24% of strokes were due to intraparenchymal haemorrhage, 2% subarachnoid haemorrhage and 74% ischemic stroke (11); subtypes and risk factors for ischemic stroke differed between Chinese and Indians (17,18). In a three-hospital study, there was higher early mortality in haemorrhagic stroke, but higher recurrence and late mortality in ischemic stroke (19).
The Singapore Stroke Registry is a countrywide project by Singapore's Ministry of Health(MOH) to obtain robust data on stroke. It will measure the size of the problem, separate the incidence of first from recurrent stroke, determine risk factors and subtypes, assess prognosis including early and late case-fatality, and uncover geographical and inter-ethnic variations of stroke patterns. Data on the utility of investigations, location of care, and treatment will serve as an audit tool. Time trends will inform the effectiveness prevention measures, and aid stroke services planning, and special-needs group identification (20).
Methods
Design
This is a countrywide, prospective, ‘cold pursuit' study of stroke patients in Singapore.
Patient population — inclusion and exclusion criteria
Inclusion criterion:
Satisfied the World Health Organization (WHO) definition of stroke modified for the WHO Monitoring Trends and Determinants in Cardiovascular Disease, World Health Organization(MONICA) project: ‘rapidly developed signs of focal (or global) disturbance of cerebral function lasting >24 h (unless interrupted by surgery or death), with no apparent nonvascular cause; this category included patients presenting with clinical signs and symptoms suggestive of subarachnoid haemorrhage, intracerebral haemorrhage, or cerebral ischemic infarction’ (21,22).
Exclusion criteria (22) are the following:
Transient ischemic attack (TIA)
Silent brain infarction
Events associated with trauma, blood disease, or malignancy
Neonates
Venous strokes
Case ascertainment
Potential cases are identified from medical claims to the government (Mediclaims listings) for stroke. Case finding is supplemented by Hospital In-patient Discharge Summary review from acute care and rehabilitation/community hospitals and emergency room attendances of all Singapore hospitals, and Death Registry data from the Ministry of Home Affairs. Codings used to identify patients, based on the International Classification of Diseases 9th Revision-Clinical Modification(ICD-9CM), covers ICD-9 codes 430, 431–437 while excluding 432·1 (subdural haemorrhage), 435(transient cerebral ischemia) and 438(late effects of cerebrovascular disease). The name lists from these multiple and overlapping sources are merged using the national registration identity card (NRIC) number, which is unique to every individual in Singapore, to obtain the master patient list.
The patient lists for the respective hospitals are generated from the master list. Case notes are traced from the hospital's medical record offices, and cases verified and information extracted on standardized case report forms (CRFs) by registry coordinators(RCs) who have worked as qualified nurses in neurology/general medicine wards. Pertinent records from previous admissions and other medical services are scrutinized. Death certificates are used for fatal cases.
Recurrence of stroke after 28 days of the preceding recorded stroke episode is counted as another episode (22).
Data collected
The following data are collected:
Demographics — name, date of birth, gender, ethnic group
Admission data — admitting hospital, date/time of hospital arrival, date/time of stroke onset; stroke aetiology based on neuroimaging — ischemic, haemorrhagic(subarachnoid/parenchymal), unknown (no scan done)
Risk factors either known or on treatment at time of admission — previous stroke, TIA, hypertension, diabetes mellitus, hyperlipidemia, ischemic heart disease, atrial fibrillation/flutter, valvular heart disease, peripheral vascular disease; smoking status(current, ex-, never)
Investigations — blood glucose, cholesterol profile; electrocardiogram performance; date/time of first brain scan and scan type
Treatments/Interventions — antiplatelets, anticoagulants, thrombolytics, surgical procedures, endovascular therapies; delivery of stroke education; assessment for rehabilitation; is the patient in a clinical trial or stroke unit
Course and outcome — use of stroke pathway; complications(pneumonia, urinary tract infection, other sepsis, deep venous thrombosis, angina/acute myocardial infarction, recurrent stroke within 28 days); date of hospital discharge; discharge destination
The data are cross-matched with the Death Registry to determine cause of death and timing with respect to stroke onset.
Data integrity, monitoring
Data are entered into a password-protected database by the RCs; only RCs and statisticians of the National Registry of Disease Office (NRDO) have data access. CRFs are stored under lock and key, and destroyed after data entry. In 2007, an online CRF was developed, avoiding the necessity to enter data into hard copy CRFs and shortening the data entry process.
A separate team of staff performs yearly audits on data collected by the RCs to ensure data accuracy of ≥95% of data entered.
Sample size estimates
Based on available hospital data, approximately 6000–10 000 stroke cases are expected annually (23).
Statistical analyses
The following outcomes will be determined:
Epidemiology — stroke incidence, recurrence, subtype, risk factors, length of hospital stay, 30-day case fatality and late morality rates; of first, recurrent and all stroke; also by age, gender and subtype
Geographical and secular variations — stroke patterns, including inter-ethnic differences
Utilization of investigations, treatments (including rehabilitation) and place of care (including stroke units)
Outcomes — complications and mortality
Trends in all above outcomes
Means and their 95% confidence intervals will be calculated for normally-distributed continuous variables; proportions will be used for categorical variables. Age-standardized rates will be calculated using the SEGI world population. The chi-square test will used to compare categorical variables, Student's t-test for normally distributed continuous variables. Logistic regression will be used with all relevant covariates for categorical outcomes while liner regression will be used for numerical outcomes. STATA version 11·0 will be used for analyses.
Study organization
The NRDO is a statutory board funded by MOH (24). The Registry is overseen by a multi-disciplinary Honorary Advisory Committee comprising stroke physicians from the public, university and private sectors.
Ethical considerations
The Singapore Government has enacted the National Registry of Diseases Act to allow the compilation of information on the incidence of selected non-communicable diseases, including stroke, without the need to obtain informed consent from individual patients.
Discussion
This Registry is likely to be the only countrywide multiethnic stroke registry in the world. With its large patient numbers, it will provide previously unavailable reliable data on inter-ethnic differences among Asians.
The Registry incorporates many of the features of the ‘ideal’ incidence study (20). Definitions are standardized. Data on first-ever-in-a-lifetime stroke are available. Standard methods are used to identify patients and collect data. There is complete, countrywide ascertainment based on multiple overlapping sources. The Registry is prospective, though ‘cold-pursuit’ is used. Singapore has a large, well-defined, stable population. The denominator, the national population numbers, is regularly updated by the Department of Statistics. Both genders and all ages are included.
The data obtained can be compared and combined with other similar studies and contribute data for the Global Burden of Diseases study (1). First-ever and recurrent stroke are distinguished. Survival status within 30 days and any future time point will be available. Brain scan is widely performed — stroke subtypes can be more accurately determined. A standardized data extraction form is used by trained reviewers. Trend data on the long-term will be available (22).
The Registry is a major part of the on-going effort by the Government to prevent, detect, monitor and reduce the impact of stroke in Singapore. Other registries managed by NRDO include acute myocardial infarction, cancer and renal disease. Patient data across registries can be compared and combined — patients often have more than one illness.
Summary and conclusions
The Singapore Stroke Registry is a unique opportunity to obtain reliable core data on a countrywide basis on stroke incidence, subtypes, risk factors, management and outcome. It has many features of the ‘ideal’ incidence study. Data on inter-ethnic differences in an Asian country make this registry unique. It hopes to contribute to the growing knowledge of global stroke epidemiology.
Footnotes
Acknowledgements
We thank the following who helped with the development of the Registry: Simon Shorvon; Jorgen Seldrup; Kee Seng Chia; Seang Mei Saw; C Rajasoorya; Noor Hafizah; En Yun Loy; Huili Zheng.