Background and Purpose To provide estimates of the prevalence of stroke and stroke-related disability for international comparisons and for planning purposes.
Methods Estimates of prevalence were derived from two population-based studies conducted 10 years apart in Auckland, New Zealand. The first, carried out in 1981, included information on survival and stroke-related disability to 14 years after stroke, and the second, undertaken in 1991 to 1992, included this information up to 3 years after stroke. An actuarial model was developed that took into account changes in incidence, long-term survival, and population structure.
Results Overall, it was estimated that 7491 people (3793 men and 3698 women) living in Auckland (total population 945 000) in 1991 had experienced a stroke at some stage in the past. This represents an age-standardized rate of 833 per 100 000 (991 per 100 000 in men and 706 per 100 000 in women) in the population aged 15 years and older. When only those who have made an incomplete recovery are considered, prevalence falls to 461 per 100 000. Of this group, one third (173 per 100 000 population 15 years and older) required assistance in at least one self-care activity.
Conclusions Usual estimates of stroke prevalence, which include all people who have ever experienced a stroke, may overestimate by almost twofold the prevalence of stroke-related disability, since many have either recovered or have no continuing dependency related to stroke. Overall prevalence does not provide information with sufficient precision for planning and purchasing ongoing services for stroke patients.
Stroke is an important cause of disability and use of health services. Knowledge of the prevalence of a condition, that is the proportion of the population affected by a given disease at a given time, is essential for planning purposes. Many estimates of the prevalence of stroke have been published from a wide range of countries but international comparisons are fraught with difficulties. Estimating the prevalence of stroke is difficult because of its relatively low occurrence especially in younger people; three quarters of all new stroke events occur in people aged 65 years and older.1 The average age-standardized prevalence of stroke worldwide is between 5 and 10 per 1000 population.2 The most likely explanations for the wide variation include differences in the methods of measurement, in the age structure of the population studied, and in the definition of stroke. Differences in prevalence may also reflect the decline in case fatality over various time periods of observation leading to an increased pool of survivors.
Traditionally, prevalence of stroke is determined by cross-sectional surveys of the population which obtain self-reported history of stroke.3 4 5 6 7 8 9 Alternatively, estimates are based on stroke incidence studies that have followed survivors over a number of years, with prevalence representing the product of incidence and duration.10 The disadvantage of the former is that it is difficult to achieve truly representative samples at older ages that include people living in institutions where many strokes occur. The disadvantage of the latter is that population changes and changes in incidence and/or survival over time are likely to misrepresent the true prevalence of stroke.
Prevalence is influenced by three dynamic forces: incidence (the number of first-ever in a lifetime stroke events occurring within a population); duration (a function of the time each such person remains alive); and changes within age and sex strata of the population, since stroke impacts on different subsets in different ways. Under these conditions an actuarial approach is the best practicable means of estimating period prevalence. This article describes the results of such a model, which was developed to take into account dynamic changes in incidence, long-term survival, and population structure using two population-based registers of stroke patients to derive varying estimates of stroke prevalence.
Acute stroke events experienced by Auckland residents were recorded at two periods in time, the first during the year ending March 1, 1982,11 and the second 10 years later during the year ending March 1, 1992.12 Stroke was defined according to the World Health Organization criteria as “rapidly developing signs of focal (or global) disturbance of cerebral function, leading to death or lasting longer than 24 hours, with no apparent cause other than vascular.”13 This definition includes spontaneous subarachnoid hemorrhage but excludes subdural and extradural hematomas and transient ischemic attacks. The analysis is based on first-ever events, that is, people who have not previously experienced a stroke.
The methods of identifying all stroke events for both studies have been described in detail elsewhere.11 12 14 15 Briefly, hospital events were identified through daily searches of hospital admission lists of all public hospitals; a wide range of conditions was investigated to ensure that all strokes that met the criteria were included. All death certificates and autopsy reports were checked on a regular basis, and any mention of stroke as a cause of death was followed up to determine whether the death had been preceded by an acute event. To locate eligible people who neither died nor were admitted to a public hospital, a validated cluster sampling method based on the accumulated practice populations of general practitioners practicing in Auckland was used.12 Patients with acute stroke events who were registered with one of a 25% random sample of all general practitioners in Auckland were included if cared for either at home, in a rest home, or in a private hospital. In the calculation of rates, these nonfatal patients treated outside the hospital were weighted by a factor of four to represent the total nonfatal, out-of-hospital treated population as defined by the sampling frame.
The total population of Auckland was 854 000 in 1981 census and 945 000 in the 1991 census. All survivors for the 1981-82 study were followed up at varying intervals through the first 14 years, whereas survivors of the 1991-92 study were followed up for 3 years. Information was complete for 98.5% of the survivors at 14 years from the 1981-82 cohort, and 99.8% for the survivors at 3 years from the 1991-92 cohort. Information at follow-up was sought directly from the patient whenever possible, but in situations in which this was inappropriate (largely because of speech and language problems and confusion), from the main caregiver.
An actuarial approach was used together with quinquennial age- and sex-specific census data for Auckland for the 30-year period 1961 to 1991 to provide annual age- and sex-specific population data from age 15 years. For years prior to 1981, the 1981 incidence rate was assumed; and for years between 1981 and 1991 an average of the 1981 and 1991 incidence rates was applied. Survival probabilities to 14 years after stroke for each age-sex grouping were determined by applying Kaplan-Meier survival analysis16 to the data collected in the two cohort studies. New Zealand life tables were then used to extend further survival probabilities beyond 14 years up to 30 years after the stroke event.17 This set of survival probabilities was used to “age” the stroke cases. For each year and age and sex strata, the results were summed to give estimated prevalent numbers as of 1991 in each age and sex group.
The mean survival time or duration was determined using the relationship between duration, prevalence, and incidence. Specifically, the mean survival time for each age-sex group was calculated from the actuarial model by dividing the number of prevalent cases at 1991 by the number of incident cases for that year. The 1981-82 and 1991-92 cohort data were used to determine actual median survival times.
An estimate of the prevalence of the ongoing consequences after stroke was calculated by including only those survivors who regarded their recovery from stroke as not complete. A further estimate was restricted to those people who reported needing help with at least one activity of daily living (washing, dressing, bathing, feeding, transferring, or toileting) to provide a better indication of the heavy users of health and community services. The percentage of cases in both categories was determined using the combined follow-up data from each of the two studies, 14 years for the 1981-82 study and 3 years for the 1991-92 study.
The point estimate of the rates for each age-sex group was determined at the median survival time. These incidence estimates were used in the actuarial process described above in order to “age” the cases and obtain prevalence estimates for stroke-related disability.
Overall incidence rates in Auckland in 1981-82 (162 per 100 000 in men and 113 per 100 000 in women) did not differ significantly from those reported 10 years later (148 per 100 000 in men and 124 per 100 000 in women)12 and have been reported at similar levels elsewhere.18 19 Substantial improvement in case fatality at 1 month occurred between the 1981-82 and 1991-92 studies for both men and women (from 27% to 21% in men and 38% to 25% in women).
The total number of new stroke events in the population 15 years and older was 1044 (516 men and 528 women) in 1981-82 and 1270 (570 men and 700 women) in 1991-92. The estimated total number of Auckland residents in this age range who had experienced a stroke at some time in the past was 7491 (3793 men and 3698 women). However, many survivors believed they have made a complete recovery and perceived their health as normal for their age. Table 1⇓ presents various estimates of prevalence based on different criteria. Rates are higher in men than in women at every age group, and the age-standardized rate in men (991 per 100 000) is 40% higher than in women (706 per 100 000).
Prevalence of survivors who reported incomplete recovery from their initial stroke suggests that approximately 460 per 100 000 of the adult population live with the ongoing consequences of stroke (493 per 100 000 in men and 448 per 100 000 in women). This estimate of prevalence is conservative as it includes people who are independent in daily activities. It also indicates that men (50%) are more likely to make a complete recovery than women (37%).
When only those survivors who require care in at least one activity of daily living are included, the rates fall to 173 per 100 000 (156 per 100 000 in men and 188 per 100 000 in women). The relationship between this dependent group of people with stroke-related disability requiring ongoing care and the overall prevalence is shown in the Figure⇓. Again, it is apparent that among stroke survivors more women (27%) are dependent on others for self-care activities than men (16%).
The estimated mean and median years of survival after a stroke are shown in Table 2⇓. Stroke patients live, on average, 5.9 years after the onset of a stroke, with men living, on average, 1.6 years longer than women. Survival is highly dependent on age, with the average survival declining sharply with increasing age. In contrast, the median survival, also shown in Table 2⇓, is only 2.2 years (2.6 years for men and 1.9 years for women), reflecting the high early case fatality during the first year, especially in the older age groups. The median duration of survival, perhaps a more useful measure for planning purposes, is less than 2 years for men and women over the age of 75 years.
The results from the Auckland stroke studies provide an accurate estimate of overall prevalence of stroke by building on the strengths of two comprehensive population-based stroke registers in which all stroke events have been prospectively identified and validated. Estimates of the overall prevalence of stroke from the Auckland stroke studies are within the range of previous estimates of the prevalence of stroke.
Two recent studies provide age-specific estimates of the prevalence of stroke for the age groups 55 years and older, one in a suburb of Rotterdam20 and the other in North Yorkshire, England.6 Estimates of prevalence from the Rotterdam study were based on a cohort study of 7983 subjects aged 55 years and older, and information on stroke history was obtained by self-report; only 67% of positive responses were confirmed by medical information and of these, the event led to hospital admission in only 50%. The North Yorkshire study used a two-stage postal questionnaire sent to a 10% random sample of the population.6 A high response rate was achieved overall, although this was age dependent and older people were less likely to respond. Age-standardized rates were calculated from published age-specific rates for the age groups 55 to 64, 65 to 74, 75 to 84, and 85 years and older using the Segi world population. In all three studies, prevalence was higher in men (42.8, 40.9, and 36.6 per 1000 population in Rotterdam, North Yorkshire, and Auckland, respectively) than in women (29.7, 28.8, and 27.1 per 1000 population, respectively) as shown in Table 3⇓.
In addition to estimates of overall prevalence, this article establishes more useful and realistic estimates that take into account recovery and care needs for those with ongoing impairments. In contrast to previous estimates, stroke-related disability rates are shown to be well below the traditional planning levels.
Our estimate includes all current stroke survivors from all types of stroke incurred at any age, acknowledges the combined effect of age- and sex-specific changes in incidence rates, and takes account of improvements in case fatality over the previous decade. Short- to intermediate-term survival has the greatest impact on the calculation of duration because annual survival probabilities are multiplicative. However, stroke patients have a persistent and statistically significant excess risk of death for several years after their first-ever stroke.21 Since mortality experience only to 14 years after stroke was available for these estimations, the survival probabilities that are not specific to the stroke population could overestimate duration and therefore prevalence. Stroke incidence is lowest in those age strata in which stroke-specific survival probabilities are least available and in which early, rather than later, stroke death is more common. The margin for error is substantially contained by the fact that from age 65 years, three quarters of all stroke subjects in every age-sex group had died within 10 years of their stroke event.
Comparisons with other studies are fraught with methodological difficulties and the failure to publish age-specific rates using standard age groups. Two recent studies that present age-specific rates for age groups above 55 years allow useful comparisons with the Auckland study and demonstrate a surprising degree of concordance. However, as demonstrated, for purchasing authorities and those planning services for stroke survivors and their family caregivers, overall prevalence is less useful than estimates that are limited to patients with incomplete recovery. Neither study presented rates that accounted for recovery patterns, although it was stated in the North Yorkshire study that one quarter had recovered from their strokes. The low percentage of people who made a complete recovery may reflect the lower response rate to the postal questionnaire in older people. A similar proportion of prevalent cases of stroke in people over the age of 45 years in a study in the United States had a persisting limitation of activity, reflecting the selection of study subjects, none of whom was in an institutional setting.7
Our study, which includes all stroke survivors including those living in institutional care, suggests that one half of the survivors of an acute stroke are completely recovered, confirming the findings of a large survey in the United Kingdom that measured care needs of stroke survivors.22 Approximately 30% will make an incomplete recovery, although they will not necessarily require assistance with usual care activities. A further 20% will require assistance with at least one activity, the majority (60%) of whom will require institutional care. It is significant that women tend to fare less well than men, although this is, to a large extent, explained by the open-ended upper age group (85 years and older) in which women predominate.
For planning purposes, our results provide an estimate of stroke-related disability of 460 per 100 000 (4.6 per 1000) population aged 15 years and older, with 173 per 100 000 (1.7 per 1000 ) requiring assistance in self-care. These planning estimates are conservative, since a substantial proportion of people who experience a stroke are already handicapped before their strokes.23
In summary, our estimates of overall stroke prevalence based on an actuarial model are conservative, relatively robust, and account for all stroke survivors from all types of stroke at any age. Our results stress the importance of the inclusion of measures other than overall prevalence in estimating the burden of a chronic disease from which many people recover. A planning estimate of around 4.6 per 1000 population that takes into account recovery and ongoing care needs, however, is a more accurate reflection of the impact of stroke and is substantially less than the current guidelines of 6.0 per 1000 population used by many purchasing authorities.24
The number of people living with stroke will tend to rise in the future for two reasons: population numbers in stroke-affected age and sex groups are increasing, and stroke survivors are living longer, increasing the numbers remaining in the prevalence pool. However, first-ever stroke is more likely to kill people who are already frail and handicapped by other causes, particularly if they are elderly.23 The need for planning services relates more to the acute stages of stroke rather than to the management of ongoing disability after a stroke.23 Greater efforts at primary prevention are warranted.
We wish to acknowledge the Health Research Council of New Zealand, the National Heart Foundation of New Zealand, and the Grand Lodge of Freemasons of New Zealand for their support of this research. Assistance with the analyses by Brody Sinclair is also gratefully acknowledged.
- Received June 23, 1997.
- Revision received August 4, 1997.
- Accepted August 5, 1997.
- Copyright © 1997 by American Heart Association
Terent A. Stroke morbidity. In: Whisnant J, ed. Stroke: Populations Cohorts and Clinical Trials. Oxford, UK: Butterworth-Heinneman Ltd; 1993:37-58.
Bharucha NE, Bharucha EP, Bharucha AE, Bhise AV, Schoenberg BS. Prevalence of stroke in the Parsi community of Bombay. Stroke. 1988;19:60-62.
Aho K, Reunanen A, Aromaa A, Knekt P, Maatela J. Prevalence of stroke in Finland. Stroke. 1986;17:681-686.
Clark ID, Opit LJ. The prevalence of stroke in those at home and the need for care. J Public Health Med. 1994;16:93-96.
Geddes JML, Fear J, Tennant A, Pickering A, Hillman M, Chamberlain MA. Prevalence of self reported stroke in a population in northern England. J Epidemiol Community Health. 1996;50:140-143.
O’Mahony PG, Dobson R, Rodgers H, James OF, Thomson RG. Validation of a population screening questionnaire to assess prevalence of stroke. Stroke. 1995;26:1334-1337.
Bonita R, Beaglehole R, North JDK. Event, incidence and case fatality rates of cerebrovascular disease in Auckland, New Zealand. Am J Epidemiol. 1984;120:236-243.
Asplund K, Tuomilehto J, Stegmayr B, Wester PO, Tunstall-Pedoe H. Diagnostic criteria and quality control of the registration of stroke events in the MONICA Project. Acta Med Scand. 1988; 728(suppl):26-39.
Bonita R, Broad JB, Anderson NE, Beaglehole R. Approaches to the problems of measuring the incidence of stroke: the Auckland Stroke Study, 1991-1992. Int J Epidemiol. 1995;24:535-542.
Kalbfliesh JD, Prentice RL. The Statistical Analysis of Failure Time Data. New York, NY: John Wiley & Sons; 1980.
Demographic Trends 1995, Statistics New Zealand, Wellington, New Zealand.
Stegmayr B, Asplund K, Wester PO. Trends in incidence, case-fatality rate, and severity of stroke in Northern Sweden, 1985-1991. Stroke. 1994;25:1738-1745.
Bots ML, Looman SJ, Koudstaal PJ, Hofman A, Hoes AW, Grobbee DE. Prevalence of stroke in the general population. Stroke. 1996;27:1499-1501.
Dennis MS, Burn JP, Sandercock PA, Bamford JM, Wade DT, Warlow CP. Long-term survival after first-ever stroke: the Oxfordshire Community Stroke Project. Stroke. 1993;24:796-800.
Clark ID, Opit LJ. The prevalence of stroke in those at home and the need for care. J Public Health Med. 1994;16:93-96.
Malmgren R, Bamford J, Warlow C, Sandercock P, Slattery J. Projecting the number of patients with first ever strokes and patients newly handicapped by stroke in England and Wales. Br Med J. 1989;298:656-660.
Stevens A, Rastery J, eds. Health Care Needs Assessment. Vol 1. Oxford, UK: Radcliffe Medical Press; 1994;178:178-179.