Projected stroke volumes to provide a 10-year direction for New Zealand stroke services
View Article PDF
Stroke is a leading cause of death and disability worldwide.1 In New Zealand, it has been estimated that around 8–9,000 people have a stroke each year and 50,000 people live with the consequences of stroke.2 The current annual cost has been estimated to sit at $NZ 700 million.3 Reducing the burden of stroke is a key goal to improve health outcomes in New Zealand.
Global and New Zealand stroke epidemiological studies have found important trends in stroke incidence and mortality over the past several decades.4 When comparing New Zealand to other high-income countries, stroke incidence is relatively high. For example, the age-adjusted incidence rate is reported to be 119 per 100,000 in Auckland compared with 76 per 100,000 in Adelaide.4 Mortality was also higher in New Zealand compared to other countries such as Australia. However, other less developed countries have incidence rates of 250 per 100,000.
The longitudinal Auckland Regional Community Stroke (ARCOS) studies have explored stroke incidence and mortality in the greater Auckland region over the past three decades and provide the most up-to-date high-quality local data.5 Overall stroke incidence has decreased by 23% and stroke mortality by 62% from 1981 to 2012. Māori and Pacific groups had a slower rate of decline and continue to experience stroke at a significantly younger age (mean ages 60 and 62 years) compared with New Zealand Europeans (mean age 75 years). The 28-day stroke case fatality has also dropped by 14% and this was seen across all ethnic groups.
It is reassuring that modern day public health initiatives, primary care preventive strategies, and secondary and tertiary level acute health interventions have made a difference. However, in light of a growing and ageing population the burden of stroke is likely going to increase despite these advances and initiatives. In addition, just like there are international variations in stroke incidence one can anticipate that there are geographic variations within New Zealand not readily captured by the ARCOS data.
This paper uses health administrative data to define the current stroke incidence across New Zealand, exploring geographic variation, and will provide future projections based on epidemiological trends. Potential solutions to address future challenge will be explored and discussed.
Methods
Building a model
The New Zealand demand for stroke services is primarily based on the incidence and prevalence of stroke in New Zealand. These in turn are influenced by demographic changes, lifestyle changes, risk factor management and efficacy of post-stroke intervention over time. Other contributing factors that influence demand include baseline population growth. Complexity and availability of diagnostics and interventions also impacts demand. Supply of stroke services depends largely on the workforce availability, which depends on training, immigration and attrition/retirement, but is also influenced by use of technology and shifting models of care to potentially more efficient approaches to patient management. Figure 1 presents a graphic model summarising the factors that impact stroke volumes.
Figure 1: Proposed model for stroke service impact estimates.
Data sources
Up-to-date nationwide New Zealand stroke audit or epidemiological data is not available. The analyses presented are based on Ministry of Health (MoH) National Minimal Dataset (NMDS) hospital discharge data. Over 95% of patients with stroke are now hospitalised, making the NMDS reasonably comprehensive, however data collection relies on hospital coders and does not undergo validation for accuracy introducing some limitation.5 For this reason, data from the high-quality ARCOS studies will be referenced at times.5 ARCOS data is limited to greater Auckland and is now several years old (last data collection 2011/12), making it unsuitable as the main data source for this project. International data is referenced also, but does not readily apply to New Zealand due to significant inter-country variations in stroke incidence.4
MoH data for the period from 1 July 2015 through 30 June 2016 (latest full financial year data available at the time of this analysis) was used to determine ‘current’ stroke volumes. Data is grouped by district health board (DHB) population, age and ethnicity.
Populations figures are based on the most recent (2013) national census data. Where available, New Zealand statistics projection estimates for 2015/16 were used (50th percentile figures). DHB-stratified population data was only available for 2013 without published projections and thus the 2013 actual numbers had to be used for the DHB-based rate estimates.
The ARCOS data included for comparison purposes was collected 1 March 2011 through 29 February 2012.6
Stroke diagnoses included in this analysis are ICD-10-AM I61 (intracerebral haemorrhage), I63 (cerebral infarction) and I64 (stroke unspecified). Subarachnoid haemorrhage is excluded, as this is generally managed by neurosurgical rather than acute stroke services. TIA is excluded because many TIA patients are managed as outpatients and thus not captured in discharge data, resulting in too many missing data points to allow for accurate modelling.7
Modelling for future stroke services
Future stroke volume projections were modelled by adjusting for population growth by domicile and age. First, all of New Zealand data was calculated applying both the 2011/12 ARCOS stroke incidence rate and the 2015/16 raw MoH incidence rate by New Zealand population projections in five-year brackets until 2068. Then the current rate of strokes under versus over 65 years of age were calculated, thus arriving at current age-adjusted incidence rates. This allowed projection of stroke rates incorporating New Zealand statistics age bracket prediction for future years. This data was then compared to the ARCOS age-adjusted incidence figures.
Population growth will not occur evenly across New Zealand. To account for such geographic variation, locality specific predictions presented by New Zealand statistics over the next 20 years were explored. For each locality, growth over the next 10 and 20 years is expressed as a factor of 1.0. If the number is less than 1.0 it indicates a drop in population and if greater than 1.0 growth. The decimal digits indicate percentage change (for example 1.15 indicates a 15% growth and 0.95 a 5% drop).
Figures were normalised around 1.0 to depict relative growth difference between DHBs. These figures were then combined with age-adjusted growth to arrive at individual DHB projections for stroke admission volumes in 2028 and 2038. Because growth data by urban area was not available for DHBs, some assumptions had to be made. For example, it was assumed that the main urban centre in each DHB is the main driver of population change based on highest population volume. Generally speaking, the spread across localities within DHBs was minimal (eg, Palmerston North 1.06 and Feilding 1.05 in MidCentral DHB). Overall growth was normalised across all DHBs, meaning the total number of strokes remains accurate. However, it is possible that estimates are off by a few patients in either direction for each DHB.
Modelling future secondary stroke services
The estimates of hospital bed occupancy were derived by multiplying the stroke volumes per DHB by average length of stay (LOS) for acute (6 days)8,9 and rehabilitation (21 days).10,11 The rehabilitation figure was multiplied by 0.28 to reflect that on average only 28% of stroke patients are transferred to inpatient rehabilitation (2016/17 MoH DHB reporting data). Aged residential care bed requirements were estimated by multiplying the total number of strokes by 0.15 to reflect an average discharge rate of 15% to these facilitates11–13 multiplied by the average duration patients reside at ARC facilities. A figure for this is not available specifically for stroke patients. General median LOS prior to dying or discharge was between 5 and 25 months depending on publication.14 For the purposes of this analysis a figure of 12 months was used.
Final annual figures were divided by 365 days and multiplied by 0.15 to allow for optimum patient flow. This is a conservative figure with international work indicating that optimal patient flow to maximise cost-efficiency and patient safety is achieved by inflating bed numbers by up to 40%.15 Flow is less critical in longer-stay facilities and ARC bed days were increased by only 5% accordingly.
Staffing levels can be estimated by looking at the number of patients admitted per year averaged per day, proportion of patients seen by a given clinician, duration of consultations, number of assessments per week, shift length and coverage (ie, number of shifts covered per day/week). Non-clinical/administrative time also needs to be factored in. There are available online stroke unit staff calculators available to assist with this (eg, http://www.stroke-education.org.uk/staff-calculators/).
Modelling for future tertiary services
Much of stroke care provided at New Zealand tertiary centres mimics the care provided at secondary hospitals. A notable exception is endovascular clot retrieval (ECR). This is a key new intervention provided by acute stroke services that is anticipated to have a significant impact on the future of stroke care provision. Clot retrieval modelling is complex due to several key features: (1) clot retrieval is time critical, needing to be commenced within hours of symptom onset;16 (2) clot retrieval is technically challenging, requiring highly skilled interventionalists;17 (3) clot retrieval needs to be available 24 hours per day to avoid treatment gaps requiring a 1:3–1:4 roster; and (4) clot retrieval volumes are relatively low, making it unfeasible to currently offer this service safely at more than three centres in New Zealand.
To achieve equitable provision of clot retrieval, the National Stroke Network has recently produced a National Clot Retrieval Strategy that identifies Auckland, Wellington and Christchurch as the three tertiary centres to offer this service until volumes increase to support a fourth centre. This paper explores projected clot retrieval numbers into the future.
Results
Current stroke volumes
As of 2015, the estimated New Zealand population overall was 4,567,000 people. Per the National Minimum Dataset data there were 8,450 stroke discharges (ICD10-AM I61, I63 and I64) in 2015, resulting in a hospital admission rate for stroke (primary discharge diagnosis) of 185/100,000. This data differs from the ARCOS data, which found a stroke incidence of 147/100,000 new stroke in 2011/2012 in the greater Auckland region. This discrepancy could be explained by several factors. Auckland is not representative of all of New Zealand; the discharge data contains coding errors, captures some people with stroke admissions that include readmissions for the same stroke event, or the ARCOS data missed some patients.
Because this report aims to provide an all of New Zealand perspective with a primary focus on health service utilisation to assist with service modelling, the MoH figures will be used although ARCOS figures will be referred to at times. The limitations are acknowledged and future studies should aim to provide better data to improve the accuracy of such models as are presented here.
Strokes are not evenly distributed across New Zealand. Table 1 shows the stroke admissions per 100,000 by DHB and Figure 2 shows ethnic distribution of stroke admissions by DHB.
Table 1: Current stroke volumes by DHB.
*NMDS figures for 12 months starting 1 July 2015.
†Statistics NZ figures for DHB population based on 2013 census data.
Figure 2: Ethnic distribution by DHB.
The above regional differences correlate with high number of non-New Zealand European populations and likely social deprivation, but this was not explored in detail for this project.
Future stroke volume projections
Per New Zealand statistics, the population will grow to 5,389,700 by 2028, 5,769,800 by 2038, and 6,060,500 by 2048. If stroke incidence remains static, the associated volumes of stroke admissions per annum respectively can thus be anticipated: 9,972, 10,675, 11,213.
However, the population is ageing (see Figure 3) and accordingly one would expect an associated disproportionate increase over time rather than a static incidence across the population.
Figure 3: New Zealand Population age estimated 2008–2068.
Adjusting for a disproportionate increase of patients >65 years of age and in light of the fact that 76% of stroke occur in this age group, a more accurate estimate would be the following admission volumes for 2028, 2038 and 2048 respectively: 11,828, 14,282 and 15,532. A breakdown of anticipated stroke volumes by age is depicted in Table 2.
Table 2: Projected stroke volumes adjusted by age.
There is also an anticipated growth of Māori and Pacific sub-populations. As the stroke incidence is higher in these ethnic groups (with earlier age of onset) these groups will become increasingly important to consider with an anticipated growth from 836 Māori and 470 Pacific stroke patients in 2015 to 976 Māori and 589 Pacific in 2028 and 1,195 Māori and 725 Pacific patients in 2038. Interestingly the ARCOS data suggests an even greater disparity with an anticipated 1,700 Māori and 1,180 Pacific stroke sufferers in 2028. However, there is no particularly disproportionate growth within this sub-population and the ethnicity spread was thus not further included in subsequent modelling.
Population growth is not evenly distributed. Auckland will experience the greatest growth with up to 72% increase in population in some localities over the next 20 years. Other areas will stay, by comparison, relatively stable (see Table 3).
Table 3: Population growth by geographic location over next 20 years.
While some especially large urban centres will experience disproportionate population growth, the effect of ageing of the population represents a more significant driver of projected stroke volumes than other factors such as immigration. Table 4 projects DHB-based stroke volumes in 2028 adjusting for both age- and locality-specific anticipated growth—some of which is driven by immigration.
Table 4: Stroke volumes (age adjusted) by DHB accounting for regional variation in projected population growth.
*Locality adjusted = volumes adjusted by age and population growth taking into account variation in population growth across the country.
Projected stroke-related resource needs
Based on these projected figures, estimates on anticipated bed requirements and staffing levels are possible. Table 5 shows some staffing estimates for 2028. This list is not exhaustive, focuses on only some professional groups limited to acute inpatient care, and is based on untested assumptions around the time spent with patients per day. This is illustrated as a potential tool to develop comprehensive workforce strategies, but requires further validation before using this data to guide service implementation.
Table 5: Projected institutional stroke resource needs by 2028.
*FTE = Full time equivalents based on a 40-hour work week = 1.0 FTE; ARC = aged residential care; CNS = clinical nurse specialist; PT = physiotherapist.
Projected regional tertiary stroke volumes
Regional services will have to be formalised by 2028 to ensure New Zealanders can access endovascular clot retrieval.
Clot retrieval treatment volumes can best be estimated by looking at intravenous alteplase treatment rates. This is because approximately one-third to one-fourth of alteplase-treated patients are eligible for clot retrieval. This number needs to be increased by 10–20% to account for the patients who are not eligible for alteplase but were still included in some of the studies (eg, patients presenting after the 4.5 hour intravenous alteplase window or who were taking dabigatran).
Intravenous thrombolysis treatment numbers in 2015 were 125 in the Northern, 101 in the Midland, 103 in the Central, and 76 in the South Island Regions.18 Based on these figures, and using the above assumptions, one would expect 50 patients to be treated with clot retrieval in the Northern region, 40 patients in the Midland and Central regions and 30 in the Southern region, per year, without any change in current thrombolysis rates.
These numbers will increase over the next 10 years for at least two reasons. Firstly, as presented above, stroke volumes will increase. Second, with greater regional organisation, access to acute therapies including thrombolysis and clot retrieval will increase as was demonstrated in the Telestroke Pilot evaluation.19 See Table 6 for projections. This is corroborated by data from a South Australian study,20 that estimates that 7% of ischaemic stroke patients are eligible. Based on current volumes this would mean about 400 clot retrieval-eligible patients in New Zealand now and 500 in 2028.
Table 6: Projected stroke and clot retrieval volumes by DHB in 2028.
†IV tPA = intravenous tissue plasminogen activator refers to ‘thrombolysis’ depicting projected volumes assuming a 10% rate (current national target).
†† Thrombolysis rates will increase through improved access via use of telestroke and other changes in model of care resulting in a likely much higher rate of around 20% in 2028 used in the subsequent column to provide a less conservative projections of anticipated clot retrieval volumes for 2028. These are not actual thrombolysis volumes. Each DHB is aware of their own current volumes and can use the above table to estimate current clot retrieval volumes based on their actual thrombolysis rate.
Furthermore, very recent studies already indicate that patient inclusion criteria will expand over time, resulting in more patients being eligible for this intervention.21 The impact of this remained uncertain at the time of submission of this paper and accordingly was not included in the below table.
As explained in the Methods section, centralisation of clot retrieval services in New Zealand is unavoidable. It is vital therefore that regional hyperacute stroke pathways are established so that there is equitable provision of acute stroke therapy across the country, rather than simply being available to populations living close to large tertiary hospitals. Achieving this will require systems reorganisation across all district health boards, in partnership with ambulance service providers.
New stroke models of care
Technological and scientific advances in the past three decades have seen the development of more specialised inter-disciplinary stroke units, although it is not always possible to offer these in small rural centres.
A number of studies have shown that decentralised thrombolysis services (telestroke, mobile stroke units, local stroke-ready health centres) can achieve similar thrombolysis rates and door-to-needle times for rural populations as a centralised system can achieve for an urban population.19,22 However, especially in the metropolitan setting when one includes other key therapeutic components and considers both patient outcomes and cost-efficiency, centralised systems tend to fare better.23
An example of this was the London Hyper-Acute Stroke Model, which consolidated the treatment of all early-phase (first 72 hours) acute stroke patients in London previously treated at 30 local hospitals into eight specialised high-volume centres designated hyperacute stroke units, or “HASUs”. Paramedic crews would no longer bring suspected acute stroke patients to their local hospital by ambulance, but instead to the nearest HASU, driving past multiple local hospitals that previously would have received them and even bypassing emergency departments.
An evaluation found that this new model of care resulted in a 25% reduction in mortality rates, marked improvement in processes of care, a rise in thrombolysis rate from 4% in 2010 to 11% in mid-2012, all while the costs of treating each stroke patient were reduced by 6%.24
Despite London’s unpredictable traffic conditions, 98% of patients arrive at the HASU within 30 minutes of being picked up by ambulance. Reports from patients about their experience with the new pathway have been largely positive even considering being cared for ‘further away from home’.24,25
A similar model is currently being implemented at night and on weekends in Auckland. Additional infrastructure and coordination resources are required to achieve this change in model of care, but it is well justified based on available data. Another area within New Zealand that would likely benefit from such collaborative centralisation is the Wellington metropolitan area by combining Hutt Valley and Capital & Coast DHB stroke services.
The value of centralisation in more rural areas remains uncertain. Because care is becoming more complex, access to specialists is becoming more important and the use of telemedicine for both acute stroke and rehabilitation will help bridge this gap as has already been demonstrated.19 Reducing clinician and patient travel times is also of benefit from an economic perspective, freeing up resources for other activities as patient volumes increase.
Use of registries and mobile devices can also help connect clinicians, monitor patient outcomes and service performance, and assist with ongoing service improvement to drive positive change. This will become increasingly important as volumes and patient complexity rise.
Stroke rehabilitation services have had limited attention over the past decades. To improve rehabilitation efficiency, inter-disciplinary teams need to be fully staffed and include all required professional groups to offer comprehensive and timely therapy. Similarly, investing more resources into community stroke services to support earlier and better support of patients in the community will help to reduce rehabilitation bed demand. New models of care, including the use of telerehabilitation to link patients in their homes to practitioners, alternate providers (eg, therapy assistants) and increasingly, self-care strategies, need to be developed and implemented in the stroke rehabilitation space.
It has been suggested that 90% of the world’s stroke burden is attributable to modifiable risk factors and that achieving control of behavioural and metabolic risk factors could avert more than three-quarters of strokes worldwide. In Oceania, the top five modifiable risk factors contributing to the stroke burden were high blood pressure, high body mass index, diet low in fruits, smoking and diet low in vegetables in order of impact.26 National policy changes and public awareness campaigns as well as new ideas as to how to best address these issues are vitally needed.
Finally, with increased complexity of care, care pathways, integration of care, and clear and open discussions around ceilings of care and advanced care planning are harder investments to quantify but should not be forgotten.
Discussion
Based on the available data, stroke incidence and mortality has reduced over the past three decades although internationally New Zealand stroke outcome performance lags behind other OECD countries. The New Zealand population is growing and ageing, and if further reductions in stroke incidence are not achieved stroke volumes can be anticipated to increase by 40% from 7,231 in 2015 to 10,112 in 2028.
Care will be further complicated by a disproportionate growth in volumes in rural areas with high degrees of social deprivation and disadvantaged ethnic population. There will also be an increasing need to centralise stroke services due to the complexity of modern stroke therapies.
Centralisation is inevitable in metropolitan areas and requires technology support in geographically dispersed areas to ensure that the disparity gap does not widen further.
Improving access to acute intervention and rehabilitation services is essential to ensure mortality and disability continue to decline and minimise the societal (and individual) impacts of the New Zealand’s increasing stroke burden. Upfront investments in these interventions, including workforce development, will maximise the possibility of containing health service resource requirements within capacity.
Investments in prevention will become increasingly important. We need to achieve a 30% reduction in stroke incidence to keep stroke volumes stable. This will require public policy consideration and massive public awareness campaigns.
This report has several limitations. The greatest limitation is that the main data source is ICD-10-AM hospital coding data that has not been validated. Coding errors are common in these types of data sources and thus the data has to be interpreted keeping this in mind. Prospectively collected proper epidemiological or registry data is not available in New Zealand beyond the Auckland region and even that data is now six years old. International data is too dissimilar to be used in its place. Also, data obtained from New Zealand statistics population data does not always correlate well with the NMDS data as regards time frames or localities, and potential inaccuracies are possible introducing further errors. Much of the projections are based on assumptions and hypothetical scenarios. New Zealand data on models of care is limited. Telestroke and stroke unit care have been found to be effective.19,27 However, there is no data around centralisation of stroke care, comparative analyses of different models of stroke unit care, or service models such as early supportive discharge in New Zealand. This makes extrapolations to the New Zealand setting tenuous and better New Zealand data is needed.
Conclusion
New Zealand stroke care improvements are not as prominent as in other OECD countries, but significant improvement has certainly been observed and more recent efforts will not be captured in the presented data. Nonetheless, while incidence rates have reduced, the anticipated population growth and ageing over the next decade will see a 40% increase in stroke admissions unless stroke incidence is further reduced through effective national primary prevention programmes. Long-term stroke burden due to disability and aged residential care requirements could be reduced through improved acute and rehabilitation stroke care, but this will require upfront service investments.
Summary
Abstract
Aim
Stroke incidence and mortality are declining due to effective public health initatives and modern healthcare advances. However, due to population growth and ageing, the burden of stroke continues to rise worldwide. This paper aims to provide stroke volume projections for the next 1-2 decades and explores potential solutions to anticipated challenges.
Method
Health administrative, where available epidemiological, and New Zealand Statistics data was used to model stroke service demand up to 2028.
Results
Despite improvements in stroke prevention and management, stroke volumes are projected to increase by 40% by 2028 due to population growth and, more importantly, ageing. Associated with this increase will be a need for more hospital beds and staff resources.
Conclusion
Efforts to optimise stroke service performance and, increasingly, stroke prevention are required to ensure that the New Zealand Health Service will be able to manage the increased volumes of patients. Better data is required to validate the presented figures, which are largely based on unvalidated health administrative data.
Author Information
Acknowledgements
Correspondence
Correspondence Email
Competing Interests
- Johnston SC, Mendis S, Mathers CD. Global variation in stroke burden and mortality: estimates from monitoring, surveillance, and modelling. Lancet Neurol 2009; 8:345–54.
- Brown P. Economic burden of stroke in New Zealand. Three decades of Auckland regional Community Stroke (ARCOS) studies: What have we learned and what is next for stroke care and stroke research? [PhD Thesis]. Auckland, New Zealand: AUT University; 2009.
- Brown P. Economic burden of stroke in New Zealand. In: Feigin VL, ed. Three decades of Auckland regional Community Stroke (ARCOS) studies: What have we learned and what is next for stroke care and stroke research? Auckland: AUT University; 2009 16.
- Thrift AG, Thayabaranathan T, Howard G, et al. Global stroke statistics. Int J Stroke 2017; 12:13–32.
- Feigin VL, Krishnamurthi RV, Barker-Collo S, et al. 30-Year Trends in Stroke Rates and Outcome in Auckland, New Zealand (1981–2012): A Multi-Ethnic Population-Based Series of Studies. PloS one 2015; 10:e0134609.
- Krishnamurthi R, Jones A, Barber PA, et al. Methodology of a population-based stroke and TIA incidence and outcomes study: the Auckland Regional Community Stroke Study (ARCOS IV) 2011–2012. Int J Stroke 2014; 9:140–7.
- Lavin TL, Ranta A. Transient ischameic attack/stroke electronic decision support: a 14-month safety audit. Journal of stroke and cerebrovascular diseases: the official journal of National Stroke Association 2014; 23:267–70.
- Child N, Barber PA, Fink J, Jones S, Voges K, Vivian M. New Zealand National Acute Stroke Services Audit 2009: organisation of acute stroke services in New Zealand. N Z Med J 2011; 124:13–20.
- Health Roundtable: New Zealand Stroke Report, 2017. http://www.healthroundtable.org/
- Awatea K. Supporting life after stroke: Counties Manukau DHB; 2015.
- McNaughton H, DeJong G, Smout RJ, Melvin JL, Brandstater M. A comparison of stroke rehabilitation practice and outcomes between New Zealand and United States facilities. Arch Phys Med Rehabil 2005; 86:S115–S20.
- Cadilhac DA, Kilkenny MF, Andrew NE, et al. Hospitals admitting at least 100 patients with stroke a year should have a stroke unit: a case study from Australia. BMC Health Serv Res 2017; 17:212.
- Tooth L, McKenna K, Goh K, Varghese P. Length of stay, discharge destination, and functional improvement: utility of the Australian National Subacute and Nonacute Patient Casemix Classification. Stroke 2005; 36:1519–25.
- Steventon A, Roberts A. Estimating length of stay in publicly-funded residential and nursing care homes: a retrospective analysis using linked administrative data sets. BMC Health Serv Res 2012; 12:377.
- Monks T, Worthington D, Allen M, Pitt M, Stein K, James MA. A modelling tool for capacity planning in acute and community stroke services. BMC Health Serv Res 2016; 16:530.
- Goyal M, Menon BK, van Zwam WH, et al. Endovascular thrombectomy after large-vessel ischaemic stroke: a meta-analysis of individual patient data from five randomised trials. Lancet 2016; 387:1723–31.
- Adamczyk P, Attenello F, Wen G, et al. Mechanical thrombectomy in acute stroke: utilization variances and impact of procedural volume on inpatient mortality. Journal of stroke and cerebrovascular diseases : the official journal of National Stroke Association 2013; 22:1263–9.
- Liu Q, Ranta AA, Abernethy G, Barber PA. Trends in New Zealand stroke thrombolysis treatment rates. N Z Med J 2017; 130:50–6.
- Ranta A, Lanford J, Busch S, Providence C, Iniesta I, Rosemergy I, Wilson A, Cariga P, Richmond V, Gommans J. Impact and implementation of a sustainable regional telestroke network. Intern Med J. 2017 Jul 25. doi: 10.1111/imj.13557
- Chia NH, Leyden JM, Newbury J, Jannes J, Kleinig TJ. Determining the Number of Ischemic Strokes Potentially Eligible for Endovascular Thrombectomy: A Population-Based Study. Stroke 2016; 47:1377–80.
- Nogueira RG, Jadhav AP, Haussen DC, et al. Thrombectomy 6 to 24 Hours after Stroke with a Mismatch between Deficit and Infarct. N Engl J Med 2017; 378:11–21.
- Lahr MM, Luijckx GJ, Vroomen PC, van der Zee DJ, Buskens E. Proportion of patients treated with thrombolysis in a centralized versus a decentralized acute stroke care setting. Stroke 2012; 43:1336–40.
- Lahr MM, van der Zee DJ, Luijckx GJ, Vroomen PC, Buskens E. Centralising and optimising decentralised stroke care systems: a simulation study on short-term costs and effects. BMC Med Res Methodol 2017; 17:5.
- Hunter RM, Davie C, Rudd A, et al. Impact on clinical and cost outcomes of a centralized approach to acute stroke care in London: a comparative effectiveness before and after model. PLoS One 2013; 8:e70420.
- Moynihan B, Paul S, Markus HS. User experience of a centralized hyperacute stroke service: a prospective evaluation. Stroke 2013; 44:2743–7.
- Feigin VL, Roth GA, Naghavi M, et al. Global burden of stroke and risk factors in 188 countries, during 1990–2013: a systematic analysis for the Global Burden of Disease Study 2013. The Lancet Neurology. 2016; 15(9):913–924.
- Te Ao BJ, Brown PM, Feigin VL, Anderson CS. Are stroke units cost effective? Evidence from a New Zealand stroke incidence and population-based study. Int J Stroke 2012; 7:623–30.
For the PDF of this article,
contact nzmj@nzma.org.nz
Projected stroke volumes to provide a 10-year direction for New Zealand stroke services
View Article PDF
Stroke is a leading cause of death and disability worldwide.1 In New Zealand, it has been estimated that around 8–9,000 people have a stroke each year and 50,000 people live with the consequences of stroke.2 The current annual cost has been estimated to sit at $NZ 700 million.3 Reducing the burden of stroke is a key goal to improve health outcomes in New Zealand.
Global and New Zealand stroke epidemiological studies have found important trends in stroke incidence and mortality over the past several decades.4 When comparing New Zealand to other high-income countries, stroke incidence is relatively high. For example, the age-adjusted incidence rate is reported to be 119 per 100,000 in Auckland compared with 76 per 100,000 in Adelaide.4 Mortality was also higher in New Zealand compared to other countries such as Australia. However, other less developed countries have incidence rates of 250 per 100,000.
The longitudinal Auckland Regional Community Stroke (ARCOS) studies have explored stroke incidence and mortality in the greater Auckland region over the past three decades and provide the most up-to-date high-quality local data.5 Overall stroke incidence has decreased by 23% and stroke mortality by 62% from 1981 to 2012. Māori and Pacific groups had a slower rate of decline and continue to experience stroke at a significantly younger age (mean ages 60 and 62 years) compared with New Zealand Europeans (mean age 75 years). The 28-day stroke case fatality has also dropped by 14% and this was seen across all ethnic groups.
It is reassuring that modern day public health initiatives, primary care preventive strategies, and secondary and tertiary level acute health interventions have made a difference. However, in light of a growing and ageing population the burden of stroke is likely going to increase despite these advances and initiatives. In addition, just like there are international variations in stroke incidence one can anticipate that there are geographic variations within New Zealand not readily captured by the ARCOS data.
This paper uses health administrative data to define the current stroke incidence across New Zealand, exploring geographic variation, and will provide future projections based on epidemiological trends. Potential solutions to address future challenge will be explored and discussed.
Methods
Building a model
The New Zealand demand for stroke services is primarily based on the incidence and prevalence of stroke in New Zealand. These in turn are influenced by demographic changes, lifestyle changes, risk factor management and efficacy of post-stroke intervention over time. Other contributing factors that influence demand include baseline population growth. Complexity and availability of diagnostics and interventions also impacts demand. Supply of stroke services depends largely on the workforce availability, which depends on training, immigration and attrition/retirement, but is also influenced by use of technology and shifting models of care to potentially more efficient approaches to patient management. Figure 1 presents a graphic model summarising the factors that impact stroke volumes.
Figure 1: Proposed model for stroke service impact estimates.
Data sources
Up-to-date nationwide New Zealand stroke audit or epidemiological data is not available. The analyses presented are based on Ministry of Health (MoH) National Minimal Dataset (NMDS) hospital discharge data. Over 95% of patients with stroke are now hospitalised, making the NMDS reasonably comprehensive, however data collection relies on hospital coders and does not undergo validation for accuracy introducing some limitation.5 For this reason, data from the high-quality ARCOS studies will be referenced at times.5 ARCOS data is limited to greater Auckland and is now several years old (last data collection 2011/12), making it unsuitable as the main data source for this project. International data is referenced also, but does not readily apply to New Zealand due to significant inter-country variations in stroke incidence.4
MoH data for the period from 1 July 2015 through 30 June 2016 (latest full financial year data available at the time of this analysis) was used to determine ‘current’ stroke volumes. Data is grouped by district health board (DHB) population, age and ethnicity.
Populations figures are based on the most recent (2013) national census data. Where available, New Zealand statistics projection estimates for 2015/16 were used (50th percentile figures). DHB-stratified population data was only available for 2013 without published projections and thus the 2013 actual numbers had to be used for the DHB-based rate estimates.
The ARCOS data included for comparison purposes was collected 1 March 2011 through 29 February 2012.6
Stroke diagnoses included in this analysis are ICD-10-AM I61 (intracerebral haemorrhage), I63 (cerebral infarction) and I64 (stroke unspecified). Subarachnoid haemorrhage is excluded, as this is generally managed by neurosurgical rather than acute stroke services. TIA is excluded because many TIA patients are managed as outpatients and thus not captured in discharge data, resulting in too many missing data points to allow for accurate modelling.7
Modelling for future stroke services
Future stroke volume projections were modelled by adjusting for population growth by domicile and age. First, all of New Zealand data was calculated applying both the 2011/12 ARCOS stroke incidence rate and the 2015/16 raw MoH incidence rate by New Zealand population projections in five-year brackets until 2068. Then the current rate of strokes under versus over 65 years of age were calculated, thus arriving at current age-adjusted incidence rates. This allowed projection of stroke rates incorporating New Zealand statistics age bracket prediction for future years. This data was then compared to the ARCOS age-adjusted incidence figures.
Population growth will not occur evenly across New Zealand. To account for such geographic variation, locality specific predictions presented by New Zealand statistics over the next 20 years were explored. For each locality, growth over the next 10 and 20 years is expressed as a factor of 1.0. If the number is less than 1.0 it indicates a drop in population and if greater than 1.0 growth. The decimal digits indicate percentage change (for example 1.15 indicates a 15% growth and 0.95 a 5% drop).
Figures were normalised around 1.0 to depict relative growth difference between DHBs. These figures were then combined with age-adjusted growth to arrive at individual DHB projections for stroke admission volumes in 2028 and 2038. Because growth data by urban area was not available for DHBs, some assumptions had to be made. For example, it was assumed that the main urban centre in each DHB is the main driver of population change based on highest population volume. Generally speaking, the spread across localities within DHBs was minimal (eg, Palmerston North 1.06 and Feilding 1.05 in MidCentral DHB). Overall growth was normalised across all DHBs, meaning the total number of strokes remains accurate. However, it is possible that estimates are off by a few patients in either direction for each DHB.
Modelling future secondary stroke services
The estimates of hospital bed occupancy were derived by multiplying the stroke volumes per DHB by average length of stay (LOS) for acute (6 days)8,9 and rehabilitation (21 days).10,11 The rehabilitation figure was multiplied by 0.28 to reflect that on average only 28% of stroke patients are transferred to inpatient rehabilitation (2016/17 MoH DHB reporting data). Aged residential care bed requirements were estimated by multiplying the total number of strokes by 0.15 to reflect an average discharge rate of 15% to these facilitates11–13 multiplied by the average duration patients reside at ARC facilities. A figure for this is not available specifically for stroke patients. General median LOS prior to dying or discharge was between 5 and 25 months depending on publication.14 For the purposes of this analysis a figure of 12 months was used.
Final annual figures were divided by 365 days and multiplied by 0.15 to allow for optimum patient flow. This is a conservative figure with international work indicating that optimal patient flow to maximise cost-efficiency and patient safety is achieved by inflating bed numbers by up to 40%.15 Flow is less critical in longer-stay facilities and ARC bed days were increased by only 5% accordingly.
Staffing levels can be estimated by looking at the number of patients admitted per year averaged per day, proportion of patients seen by a given clinician, duration of consultations, number of assessments per week, shift length and coverage (ie, number of shifts covered per day/week). Non-clinical/administrative time also needs to be factored in. There are available online stroke unit staff calculators available to assist with this (eg, http://www.stroke-education.org.uk/staff-calculators/).
Modelling for future tertiary services
Much of stroke care provided at New Zealand tertiary centres mimics the care provided at secondary hospitals. A notable exception is endovascular clot retrieval (ECR). This is a key new intervention provided by acute stroke services that is anticipated to have a significant impact on the future of stroke care provision. Clot retrieval modelling is complex due to several key features: (1) clot retrieval is time critical, needing to be commenced within hours of symptom onset;16 (2) clot retrieval is technically challenging, requiring highly skilled interventionalists;17 (3) clot retrieval needs to be available 24 hours per day to avoid treatment gaps requiring a 1:3–1:4 roster; and (4) clot retrieval volumes are relatively low, making it unfeasible to currently offer this service safely at more than three centres in New Zealand.
To achieve equitable provision of clot retrieval, the National Stroke Network has recently produced a National Clot Retrieval Strategy that identifies Auckland, Wellington and Christchurch as the three tertiary centres to offer this service until volumes increase to support a fourth centre. This paper explores projected clot retrieval numbers into the future.
Results
Current stroke volumes
As of 2015, the estimated New Zealand population overall was 4,567,000 people. Per the National Minimum Dataset data there were 8,450 stroke discharges (ICD10-AM I61, I63 and I64) in 2015, resulting in a hospital admission rate for stroke (primary discharge diagnosis) of 185/100,000. This data differs from the ARCOS data, which found a stroke incidence of 147/100,000 new stroke in 2011/2012 in the greater Auckland region. This discrepancy could be explained by several factors. Auckland is not representative of all of New Zealand; the discharge data contains coding errors, captures some people with stroke admissions that include readmissions for the same stroke event, or the ARCOS data missed some patients.
Because this report aims to provide an all of New Zealand perspective with a primary focus on health service utilisation to assist with service modelling, the MoH figures will be used although ARCOS figures will be referred to at times. The limitations are acknowledged and future studies should aim to provide better data to improve the accuracy of such models as are presented here.
Strokes are not evenly distributed across New Zealand. Table 1 shows the stroke admissions per 100,000 by DHB and Figure 2 shows ethnic distribution of stroke admissions by DHB.
Table 1: Current stroke volumes by DHB.
*NMDS figures for 12 months starting 1 July 2015.
†Statistics NZ figures for DHB population based on 2013 census data.
Figure 2: Ethnic distribution by DHB.
The above regional differences correlate with high number of non-New Zealand European populations and likely social deprivation, but this was not explored in detail for this project.
Future stroke volume projections
Per New Zealand statistics, the population will grow to 5,389,700 by 2028, 5,769,800 by 2038, and 6,060,500 by 2048. If stroke incidence remains static, the associated volumes of stroke admissions per annum respectively can thus be anticipated: 9,972, 10,675, 11,213.
However, the population is ageing (see Figure 3) and accordingly one would expect an associated disproportionate increase over time rather than a static incidence across the population.
Figure 3: New Zealand Population age estimated 2008–2068.
Adjusting for a disproportionate increase of patients >65 years of age and in light of the fact that 76% of stroke occur in this age group, a more accurate estimate would be the following admission volumes for 2028, 2038 and 2048 respectively: 11,828, 14,282 and 15,532. A breakdown of anticipated stroke volumes by age is depicted in Table 2.
Table 2: Projected stroke volumes adjusted by age.
There is also an anticipated growth of Māori and Pacific sub-populations. As the stroke incidence is higher in these ethnic groups (with earlier age of onset) these groups will become increasingly important to consider with an anticipated growth from 836 Māori and 470 Pacific stroke patients in 2015 to 976 Māori and 589 Pacific in 2028 and 1,195 Māori and 725 Pacific patients in 2038. Interestingly the ARCOS data suggests an even greater disparity with an anticipated 1,700 Māori and 1,180 Pacific stroke sufferers in 2028. However, there is no particularly disproportionate growth within this sub-population and the ethnicity spread was thus not further included in subsequent modelling.
Population growth is not evenly distributed. Auckland will experience the greatest growth with up to 72% increase in population in some localities over the next 20 years. Other areas will stay, by comparison, relatively stable (see Table 3).
Table 3: Population growth by geographic location over next 20 years.
While some especially large urban centres will experience disproportionate population growth, the effect of ageing of the population represents a more significant driver of projected stroke volumes than other factors such as immigration. Table 4 projects DHB-based stroke volumes in 2028 adjusting for both age- and locality-specific anticipated growth—some of which is driven by immigration.
Table 4: Stroke volumes (age adjusted) by DHB accounting for regional variation in projected population growth.
*Locality adjusted = volumes adjusted by age and population growth taking into account variation in population growth across the country.
Projected stroke-related resource needs
Based on these projected figures, estimates on anticipated bed requirements and staffing levels are possible. Table 5 shows some staffing estimates for 2028. This list is not exhaustive, focuses on only some professional groups limited to acute inpatient care, and is based on untested assumptions around the time spent with patients per day. This is illustrated as a potential tool to develop comprehensive workforce strategies, but requires further validation before using this data to guide service implementation.
Table 5: Projected institutional stroke resource needs by 2028.
*FTE = Full time equivalents based on a 40-hour work week = 1.0 FTE; ARC = aged residential care; CNS = clinical nurse specialist; PT = physiotherapist.
Projected regional tertiary stroke volumes
Regional services will have to be formalised by 2028 to ensure New Zealanders can access endovascular clot retrieval.
Clot retrieval treatment volumes can best be estimated by looking at intravenous alteplase treatment rates. This is because approximately one-third to one-fourth of alteplase-treated patients are eligible for clot retrieval. This number needs to be increased by 10–20% to account for the patients who are not eligible for alteplase but were still included in some of the studies (eg, patients presenting after the 4.5 hour intravenous alteplase window or who were taking dabigatran).
Intravenous thrombolysis treatment numbers in 2015 were 125 in the Northern, 101 in the Midland, 103 in the Central, and 76 in the South Island Regions.18 Based on these figures, and using the above assumptions, one would expect 50 patients to be treated with clot retrieval in the Northern region, 40 patients in the Midland and Central regions and 30 in the Southern region, per year, without any change in current thrombolysis rates.
These numbers will increase over the next 10 years for at least two reasons. Firstly, as presented above, stroke volumes will increase. Second, with greater regional organisation, access to acute therapies including thrombolysis and clot retrieval will increase as was demonstrated in the Telestroke Pilot evaluation.19 See Table 6 for projections. This is corroborated by data from a South Australian study,20 that estimates that 7% of ischaemic stroke patients are eligible. Based on current volumes this would mean about 400 clot retrieval-eligible patients in New Zealand now and 500 in 2028.
Table 6: Projected stroke and clot retrieval volumes by DHB in 2028.
†IV tPA = intravenous tissue plasminogen activator refers to ‘thrombolysis’ depicting projected volumes assuming a 10% rate (current national target).
†† Thrombolysis rates will increase through improved access via use of telestroke and other changes in model of care resulting in a likely much higher rate of around 20% in 2028 used in the subsequent column to provide a less conservative projections of anticipated clot retrieval volumes for 2028. These are not actual thrombolysis volumes. Each DHB is aware of their own current volumes and can use the above table to estimate current clot retrieval volumes based on their actual thrombolysis rate.
Furthermore, very recent studies already indicate that patient inclusion criteria will expand over time, resulting in more patients being eligible for this intervention.21 The impact of this remained uncertain at the time of submission of this paper and accordingly was not included in the below table.
As explained in the Methods section, centralisation of clot retrieval services in New Zealand is unavoidable. It is vital therefore that regional hyperacute stroke pathways are established so that there is equitable provision of acute stroke therapy across the country, rather than simply being available to populations living close to large tertiary hospitals. Achieving this will require systems reorganisation across all district health boards, in partnership with ambulance service providers.
New stroke models of care
Technological and scientific advances in the past three decades have seen the development of more specialised inter-disciplinary stroke units, although it is not always possible to offer these in small rural centres.
A number of studies have shown that decentralised thrombolysis services (telestroke, mobile stroke units, local stroke-ready health centres) can achieve similar thrombolysis rates and door-to-needle times for rural populations as a centralised system can achieve for an urban population.19,22 However, especially in the metropolitan setting when one includes other key therapeutic components and considers both patient outcomes and cost-efficiency, centralised systems tend to fare better.23
An example of this was the London Hyper-Acute Stroke Model, which consolidated the treatment of all early-phase (first 72 hours) acute stroke patients in London previously treated at 30 local hospitals into eight specialised high-volume centres designated hyperacute stroke units, or “HASUs”. Paramedic crews would no longer bring suspected acute stroke patients to their local hospital by ambulance, but instead to the nearest HASU, driving past multiple local hospitals that previously would have received them and even bypassing emergency departments.
An evaluation found that this new model of care resulted in a 25% reduction in mortality rates, marked improvement in processes of care, a rise in thrombolysis rate from 4% in 2010 to 11% in mid-2012, all while the costs of treating each stroke patient were reduced by 6%.24
Despite London’s unpredictable traffic conditions, 98% of patients arrive at the HASU within 30 minutes of being picked up by ambulance. Reports from patients about their experience with the new pathway have been largely positive even considering being cared for ‘further away from home’.24,25
A similar model is currently being implemented at night and on weekends in Auckland. Additional infrastructure and coordination resources are required to achieve this change in model of care, but it is well justified based on available data. Another area within New Zealand that would likely benefit from such collaborative centralisation is the Wellington metropolitan area by combining Hutt Valley and Capital & Coast DHB stroke services.
The value of centralisation in more rural areas remains uncertain. Because care is becoming more complex, access to specialists is becoming more important and the use of telemedicine for both acute stroke and rehabilitation will help bridge this gap as has already been demonstrated.19 Reducing clinician and patient travel times is also of benefit from an economic perspective, freeing up resources for other activities as patient volumes increase.
Use of registries and mobile devices can also help connect clinicians, monitor patient outcomes and service performance, and assist with ongoing service improvement to drive positive change. This will become increasingly important as volumes and patient complexity rise.
Stroke rehabilitation services have had limited attention over the past decades. To improve rehabilitation efficiency, inter-disciplinary teams need to be fully staffed and include all required professional groups to offer comprehensive and timely therapy. Similarly, investing more resources into community stroke services to support earlier and better support of patients in the community will help to reduce rehabilitation bed demand. New models of care, including the use of telerehabilitation to link patients in their homes to practitioners, alternate providers (eg, therapy assistants) and increasingly, self-care strategies, need to be developed and implemented in the stroke rehabilitation space.
It has been suggested that 90% of the world’s stroke burden is attributable to modifiable risk factors and that achieving control of behavioural and metabolic risk factors could avert more than three-quarters of strokes worldwide. In Oceania, the top five modifiable risk factors contributing to the stroke burden were high blood pressure, high body mass index, diet low in fruits, smoking and diet low in vegetables in order of impact.26 National policy changes and public awareness campaigns as well as new ideas as to how to best address these issues are vitally needed.
Finally, with increased complexity of care, care pathways, integration of care, and clear and open discussions around ceilings of care and advanced care planning are harder investments to quantify but should not be forgotten.
Discussion
Based on the available data, stroke incidence and mortality has reduced over the past three decades although internationally New Zealand stroke outcome performance lags behind other OECD countries. The New Zealand population is growing and ageing, and if further reductions in stroke incidence are not achieved stroke volumes can be anticipated to increase by 40% from 7,231 in 2015 to 10,112 in 2028.
Care will be further complicated by a disproportionate growth in volumes in rural areas with high degrees of social deprivation and disadvantaged ethnic population. There will also be an increasing need to centralise stroke services due to the complexity of modern stroke therapies.
Centralisation is inevitable in metropolitan areas and requires technology support in geographically dispersed areas to ensure that the disparity gap does not widen further.
Improving access to acute intervention and rehabilitation services is essential to ensure mortality and disability continue to decline and minimise the societal (and individual) impacts of the New Zealand’s increasing stroke burden. Upfront investments in these interventions, including workforce development, will maximise the possibility of containing health service resource requirements within capacity.
Investments in prevention will become increasingly important. We need to achieve a 30% reduction in stroke incidence to keep stroke volumes stable. This will require public policy consideration and massive public awareness campaigns.
This report has several limitations. The greatest limitation is that the main data source is ICD-10-AM hospital coding data that has not been validated. Coding errors are common in these types of data sources and thus the data has to be interpreted keeping this in mind. Prospectively collected proper epidemiological or registry data is not available in New Zealand beyond the Auckland region and even that data is now six years old. International data is too dissimilar to be used in its place. Also, data obtained from New Zealand statistics population data does not always correlate well with the NMDS data as regards time frames or localities, and potential inaccuracies are possible introducing further errors. Much of the projections are based on assumptions and hypothetical scenarios. New Zealand data on models of care is limited. Telestroke and stroke unit care have been found to be effective.19,27 However, there is no data around centralisation of stroke care, comparative analyses of different models of stroke unit care, or service models such as early supportive discharge in New Zealand. This makes extrapolations to the New Zealand setting tenuous and better New Zealand data is needed.
Conclusion
New Zealand stroke care improvements are not as prominent as in other OECD countries, but significant improvement has certainly been observed and more recent efforts will not be captured in the presented data. Nonetheless, while incidence rates have reduced, the anticipated population growth and ageing over the next decade will see a 40% increase in stroke admissions unless stroke incidence is further reduced through effective national primary prevention programmes. Long-term stroke burden due to disability and aged residential care requirements could be reduced through improved acute and rehabilitation stroke care, but this will require upfront service investments.
Summary
Abstract
Aim
Stroke incidence and mortality are declining due to effective public health initatives and modern healthcare advances. However, due to population growth and ageing, the burden of stroke continues to rise worldwide. This paper aims to provide stroke volume projections for the next 1-2 decades and explores potential solutions to anticipated challenges.
Method
Health administrative, where available epidemiological, and New Zealand Statistics data was used to model stroke service demand up to 2028.
Results
Despite improvements in stroke prevention and management, stroke volumes are projected to increase by 40% by 2028 due to population growth and, more importantly, ageing. Associated with this increase will be a need for more hospital beds and staff resources.
Conclusion
Efforts to optimise stroke service performance and, increasingly, stroke prevention are required to ensure that the New Zealand Health Service will be able to manage the increased volumes of patients. Better data is required to validate the presented figures, which are largely based on unvalidated health administrative data.
Author Information
Acknowledgements
Correspondence
Correspondence Email
Competing Interests
- Johnston SC, Mendis S, Mathers CD. Global variation in stroke burden and mortality: estimates from monitoring, surveillance, and modelling. Lancet Neurol 2009; 8:345–54.
- Brown P. Economic burden of stroke in New Zealand. Three decades of Auckland regional Community Stroke (ARCOS) studies: What have we learned and what is next for stroke care and stroke research? [PhD Thesis]. Auckland, New Zealand: AUT University; 2009.
- Brown P. Economic burden of stroke in New Zealand. In: Feigin VL, ed. Three decades of Auckland regional Community Stroke (ARCOS) studies: What have we learned and what is next for stroke care and stroke research? Auckland: AUT University; 2009 16.
- Thrift AG, Thayabaranathan T, Howard G, et al. Global stroke statistics. Int J Stroke 2017; 12:13–32.
- Feigin VL, Krishnamurthi RV, Barker-Collo S, et al. 30-Year Trends in Stroke Rates and Outcome in Auckland, New Zealand (1981–2012): A Multi-Ethnic Population-Based Series of Studies. PloS one 2015; 10:e0134609.
- Krishnamurthi R, Jones A, Barber PA, et al. Methodology of a population-based stroke and TIA incidence and outcomes study: the Auckland Regional Community Stroke Study (ARCOS IV) 2011–2012. Int J Stroke 2014; 9:140–7.
- Lavin TL, Ranta A. Transient ischameic attack/stroke electronic decision support: a 14-month safety audit. Journal of stroke and cerebrovascular diseases: the official journal of National Stroke Association 2014; 23:267–70.
- Child N, Barber PA, Fink J, Jones S, Voges K, Vivian M. New Zealand National Acute Stroke Services Audit 2009: organisation of acute stroke services in New Zealand. N Z Med J 2011; 124:13–20.
- Health Roundtable: New Zealand Stroke Report, 2017. http://www.healthroundtable.org/
- Awatea K. Supporting life after stroke: Counties Manukau DHB; 2015.
- McNaughton H, DeJong G, Smout RJ, Melvin JL, Brandstater M. A comparison of stroke rehabilitation practice and outcomes between New Zealand and United States facilities. Arch Phys Med Rehabil 2005; 86:S115–S20.
- Cadilhac DA, Kilkenny MF, Andrew NE, et al. Hospitals admitting at least 100 patients with stroke a year should have a stroke unit: a case study from Australia. BMC Health Serv Res 2017; 17:212.
- Tooth L, McKenna K, Goh K, Varghese P. Length of stay, discharge destination, and functional improvement: utility of the Australian National Subacute and Nonacute Patient Casemix Classification. Stroke 2005; 36:1519–25.
- Steventon A, Roberts A. Estimating length of stay in publicly-funded residential and nursing care homes: a retrospective analysis using linked administrative data sets. BMC Health Serv Res 2012; 12:377.
- Monks T, Worthington D, Allen M, Pitt M, Stein K, James MA. A modelling tool for capacity planning in acute and community stroke services. BMC Health Serv Res 2016; 16:530.
- Goyal M, Menon BK, van Zwam WH, et al. Endovascular thrombectomy after large-vessel ischaemic stroke: a meta-analysis of individual patient data from five randomised trials. Lancet 2016; 387:1723–31.
- Adamczyk P, Attenello F, Wen G, et al. Mechanical thrombectomy in acute stroke: utilization variances and impact of procedural volume on inpatient mortality. Journal of stroke and cerebrovascular diseases : the official journal of National Stroke Association 2013; 22:1263–9.
- Liu Q, Ranta AA, Abernethy G, Barber PA. Trends in New Zealand stroke thrombolysis treatment rates. N Z Med J 2017; 130:50–6.
- Ranta A, Lanford J, Busch S, Providence C, Iniesta I, Rosemergy I, Wilson A, Cariga P, Richmond V, Gommans J. Impact and implementation of a sustainable regional telestroke network. Intern Med J. 2017 Jul 25. doi: 10.1111/imj.13557
- Chia NH, Leyden JM, Newbury J, Jannes J, Kleinig TJ. Determining the Number of Ischemic Strokes Potentially Eligible for Endovascular Thrombectomy: A Population-Based Study. Stroke 2016; 47:1377–80.
- Nogueira RG, Jadhav AP, Haussen DC, et al. Thrombectomy 6 to 24 Hours after Stroke with a Mismatch between Deficit and Infarct. N Engl J Med 2017; 378:11–21.
- Lahr MM, Luijckx GJ, Vroomen PC, van der Zee DJ, Buskens E. Proportion of patients treated with thrombolysis in a centralized versus a decentralized acute stroke care setting. Stroke 2012; 43:1336–40.
- Lahr MM, van der Zee DJ, Luijckx GJ, Vroomen PC, Buskens E. Centralising and optimising decentralised stroke care systems: a simulation study on short-term costs and effects. BMC Med Res Methodol 2017; 17:5.
- Hunter RM, Davie C, Rudd A, et al. Impact on clinical and cost outcomes of a centralized approach to acute stroke care in London: a comparative effectiveness before and after model. PLoS One 2013; 8:e70420.
- Moynihan B, Paul S, Markus HS. User experience of a centralized hyperacute stroke service: a prospective evaluation. Stroke 2013; 44:2743–7.
- Feigin VL, Roth GA, Naghavi M, et al. Global burden of stroke and risk factors in 188 countries, during 1990–2013: a systematic analysis for the Global Burden of Disease Study 2013. The Lancet Neurology. 2016; 15(9):913–924.
- Te Ao BJ, Brown PM, Feigin VL, Anderson CS. Are stroke units cost effective? Evidence from a New Zealand stroke incidence and population-based study. Int J Stroke 2012; 7:623–30.
For the PDF of this article,
contact nzmj@nzma.org.nz
Projected stroke volumes to provide a 10-year direction for New Zealand stroke services
View Article PDF
Stroke is a leading cause of death and disability worldwide.1 In New Zealand, it has been estimated that around 8–9,000 people have a stroke each year and 50,000 people live with the consequences of stroke.2 The current annual cost has been estimated to sit at $NZ 700 million.3 Reducing the burden of stroke is a key goal to improve health outcomes in New Zealand.
Global and New Zealand stroke epidemiological studies have found important trends in stroke incidence and mortality over the past several decades.4 When comparing New Zealand to other high-income countries, stroke incidence is relatively high. For example, the age-adjusted incidence rate is reported to be 119 per 100,000 in Auckland compared with 76 per 100,000 in Adelaide.4 Mortality was also higher in New Zealand compared to other countries such as Australia. However, other less developed countries have incidence rates of 250 per 100,000.
The longitudinal Auckland Regional Community Stroke (ARCOS) studies have explored stroke incidence and mortality in the greater Auckland region over the past three decades and provide the most up-to-date high-quality local data.5 Overall stroke incidence has decreased by 23% and stroke mortality by 62% from 1981 to 2012. Māori and Pacific groups had a slower rate of decline and continue to experience stroke at a significantly younger age (mean ages 60 and 62 years) compared with New Zealand Europeans (mean age 75 years). The 28-day stroke case fatality has also dropped by 14% and this was seen across all ethnic groups.
It is reassuring that modern day public health initiatives, primary care preventive strategies, and secondary and tertiary level acute health interventions have made a difference. However, in light of a growing and ageing population the burden of stroke is likely going to increase despite these advances and initiatives. In addition, just like there are international variations in stroke incidence one can anticipate that there are geographic variations within New Zealand not readily captured by the ARCOS data.
This paper uses health administrative data to define the current stroke incidence across New Zealand, exploring geographic variation, and will provide future projections based on epidemiological trends. Potential solutions to address future challenge will be explored and discussed.
Methods
Building a model
The New Zealand demand for stroke services is primarily based on the incidence and prevalence of stroke in New Zealand. These in turn are influenced by demographic changes, lifestyle changes, risk factor management and efficacy of post-stroke intervention over time. Other contributing factors that influence demand include baseline population growth. Complexity and availability of diagnostics and interventions also impacts demand. Supply of stroke services depends largely on the workforce availability, which depends on training, immigration and attrition/retirement, but is also influenced by use of technology and shifting models of care to potentially more efficient approaches to patient management. Figure 1 presents a graphic model summarising the factors that impact stroke volumes.
Figure 1: Proposed model for stroke service impact estimates.
Data sources
Up-to-date nationwide New Zealand stroke audit or epidemiological data is not available. The analyses presented are based on Ministry of Health (MoH) National Minimal Dataset (NMDS) hospital discharge data. Over 95% of patients with stroke are now hospitalised, making the NMDS reasonably comprehensive, however data collection relies on hospital coders and does not undergo validation for accuracy introducing some limitation.5 For this reason, data from the high-quality ARCOS studies will be referenced at times.5 ARCOS data is limited to greater Auckland and is now several years old (last data collection 2011/12), making it unsuitable as the main data source for this project. International data is referenced also, but does not readily apply to New Zealand due to significant inter-country variations in stroke incidence.4
MoH data for the period from 1 July 2015 through 30 June 2016 (latest full financial year data available at the time of this analysis) was used to determine ‘current’ stroke volumes. Data is grouped by district health board (DHB) population, age and ethnicity.
Populations figures are based on the most recent (2013) national census data. Where available, New Zealand statistics projection estimates for 2015/16 were used (50th percentile figures). DHB-stratified population data was only available for 2013 without published projections and thus the 2013 actual numbers had to be used for the DHB-based rate estimates.
The ARCOS data included for comparison purposes was collected 1 March 2011 through 29 February 2012.6
Stroke diagnoses included in this analysis are ICD-10-AM I61 (intracerebral haemorrhage), I63 (cerebral infarction) and I64 (stroke unspecified). Subarachnoid haemorrhage is excluded, as this is generally managed by neurosurgical rather than acute stroke services. TIA is excluded because many TIA patients are managed as outpatients and thus not captured in discharge data, resulting in too many missing data points to allow for accurate modelling.7
Modelling for future stroke services
Future stroke volume projections were modelled by adjusting for population growth by domicile and age. First, all of New Zealand data was calculated applying both the 2011/12 ARCOS stroke incidence rate and the 2015/16 raw MoH incidence rate by New Zealand population projections in five-year brackets until 2068. Then the current rate of strokes under versus over 65 years of age were calculated, thus arriving at current age-adjusted incidence rates. This allowed projection of stroke rates incorporating New Zealand statistics age bracket prediction for future years. This data was then compared to the ARCOS age-adjusted incidence figures.
Population growth will not occur evenly across New Zealand. To account for such geographic variation, locality specific predictions presented by New Zealand statistics over the next 20 years were explored. For each locality, growth over the next 10 and 20 years is expressed as a factor of 1.0. If the number is less than 1.0 it indicates a drop in population and if greater than 1.0 growth. The decimal digits indicate percentage change (for example 1.15 indicates a 15% growth and 0.95 a 5% drop).
Figures were normalised around 1.0 to depict relative growth difference between DHBs. These figures were then combined with age-adjusted growth to arrive at individual DHB projections for stroke admission volumes in 2028 and 2038. Because growth data by urban area was not available for DHBs, some assumptions had to be made. For example, it was assumed that the main urban centre in each DHB is the main driver of population change based on highest population volume. Generally speaking, the spread across localities within DHBs was minimal (eg, Palmerston North 1.06 and Feilding 1.05 in MidCentral DHB). Overall growth was normalised across all DHBs, meaning the total number of strokes remains accurate. However, it is possible that estimates are off by a few patients in either direction for each DHB.
Modelling future secondary stroke services
The estimates of hospital bed occupancy were derived by multiplying the stroke volumes per DHB by average length of stay (LOS) for acute (6 days)8,9 and rehabilitation (21 days).10,11 The rehabilitation figure was multiplied by 0.28 to reflect that on average only 28% of stroke patients are transferred to inpatient rehabilitation (2016/17 MoH DHB reporting data). Aged residential care bed requirements were estimated by multiplying the total number of strokes by 0.15 to reflect an average discharge rate of 15% to these facilitates11–13 multiplied by the average duration patients reside at ARC facilities. A figure for this is not available specifically for stroke patients. General median LOS prior to dying or discharge was between 5 and 25 months depending on publication.14 For the purposes of this analysis a figure of 12 months was used.
Final annual figures were divided by 365 days and multiplied by 0.15 to allow for optimum patient flow. This is a conservative figure with international work indicating that optimal patient flow to maximise cost-efficiency and patient safety is achieved by inflating bed numbers by up to 40%.15 Flow is less critical in longer-stay facilities and ARC bed days were increased by only 5% accordingly.
Staffing levels can be estimated by looking at the number of patients admitted per year averaged per day, proportion of patients seen by a given clinician, duration of consultations, number of assessments per week, shift length and coverage (ie, number of shifts covered per day/week). Non-clinical/administrative time also needs to be factored in. There are available online stroke unit staff calculators available to assist with this (eg, http://www.stroke-education.org.uk/staff-calculators/).
Modelling for future tertiary services
Much of stroke care provided at New Zealand tertiary centres mimics the care provided at secondary hospitals. A notable exception is endovascular clot retrieval (ECR). This is a key new intervention provided by acute stroke services that is anticipated to have a significant impact on the future of stroke care provision. Clot retrieval modelling is complex due to several key features: (1) clot retrieval is time critical, needing to be commenced within hours of symptom onset;16 (2) clot retrieval is technically challenging, requiring highly skilled interventionalists;17 (3) clot retrieval needs to be available 24 hours per day to avoid treatment gaps requiring a 1:3–1:4 roster; and (4) clot retrieval volumes are relatively low, making it unfeasible to currently offer this service safely at more than three centres in New Zealand.
To achieve equitable provision of clot retrieval, the National Stroke Network has recently produced a National Clot Retrieval Strategy that identifies Auckland, Wellington and Christchurch as the three tertiary centres to offer this service until volumes increase to support a fourth centre. This paper explores projected clot retrieval numbers into the future.
Results
Current stroke volumes
As of 2015, the estimated New Zealand population overall was 4,567,000 people. Per the National Minimum Dataset data there were 8,450 stroke discharges (ICD10-AM I61, I63 and I64) in 2015, resulting in a hospital admission rate for stroke (primary discharge diagnosis) of 185/100,000. This data differs from the ARCOS data, which found a stroke incidence of 147/100,000 new stroke in 2011/2012 in the greater Auckland region. This discrepancy could be explained by several factors. Auckland is not representative of all of New Zealand; the discharge data contains coding errors, captures some people with stroke admissions that include readmissions for the same stroke event, or the ARCOS data missed some patients.
Because this report aims to provide an all of New Zealand perspective with a primary focus on health service utilisation to assist with service modelling, the MoH figures will be used although ARCOS figures will be referred to at times. The limitations are acknowledged and future studies should aim to provide better data to improve the accuracy of such models as are presented here.
Strokes are not evenly distributed across New Zealand. Table 1 shows the stroke admissions per 100,000 by DHB and Figure 2 shows ethnic distribution of stroke admissions by DHB.
Table 1: Current stroke volumes by DHB.
*NMDS figures for 12 months starting 1 July 2015.
†Statistics NZ figures for DHB population based on 2013 census data.
Figure 2: Ethnic distribution by DHB.
The above regional differences correlate with high number of non-New Zealand European populations and likely social deprivation, but this was not explored in detail for this project.
Future stroke volume projections
Per New Zealand statistics, the population will grow to 5,389,700 by 2028, 5,769,800 by 2038, and 6,060,500 by 2048. If stroke incidence remains static, the associated volumes of stroke admissions per annum respectively can thus be anticipated: 9,972, 10,675, 11,213.
However, the population is ageing (see Figure 3) and accordingly one would expect an associated disproportionate increase over time rather than a static incidence across the population.
Figure 3: New Zealand Population age estimated 2008–2068.
Adjusting for a disproportionate increase of patients >65 years of age and in light of the fact that 76% of stroke occur in this age group, a more accurate estimate would be the following admission volumes for 2028, 2038 and 2048 respectively: 11,828, 14,282 and 15,532. A breakdown of anticipated stroke volumes by age is depicted in Table 2.
Table 2: Projected stroke volumes adjusted by age.
There is also an anticipated growth of Māori and Pacific sub-populations. As the stroke incidence is higher in these ethnic groups (with earlier age of onset) these groups will become increasingly important to consider with an anticipated growth from 836 Māori and 470 Pacific stroke patients in 2015 to 976 Māori and 589 Pacific in 2028 and 1,195 Māori and 725 Pacific patients in 2038. Interestingly the ARCOS data suggests an even greater disparity with an anticipated 1,700 Māori and 1,180 Pacific stroke sufferers in 2028. However, there is no particularly disproportionate growth within this sub-population and the ethnicity spread was thus not further included in subsequent modelling.
Population growth is not evenly distributed. Auckland will experience the greatest growth with up to 72% increase in population in some localities over the next 20 years. Other areas will stay, by comparison, relatively stable (see Table 3).
Table 3: Population growth by geographic location over next 20 years.
While some especially large urban centres will experience disproportionate population growth, the effect of ageing of the population represents a more significant driver of projected stroke volumes than other factors such as immigration. Table 4 projects DHB-based stroke volumes in 2028 adjusting for both age- and locality-specific anticipated growth—some of which is driven by immigration.
Table 4: Stroke volumes (age adjusted) by DHB accounting for regional variation in projected population growth.
*Locality adjusted = volumes adjusted by age and population growth taking into account variation in population growth across the country.
Projected stroke-related resource needs
Based on these projected figures, estimates on anticipated bed requirements and staffing levels are possible. Table 5 shows some staffing estimates for 2028. This list is not exhaustive, focuses on only some professional groups limited to acute inpatient care, and is based on untested assumptions around the time spent with patients per day. This is illustrated as a potential tool to develop comprehensive workforce strategies, but requires further validation before using this data to guide service implementation.
Table 5: Projected institutional stroke resource needs by 2028.
*FTE = Full time equivalents based on a 40-hour work week = 1.0 FTE; ARC = aged residential care; CNS = clinical nurse specialist; PT = physiotherapist.
Projected regional tertiary stroke volumes
Regional services will have to be formalised by 2028 to ensure New Zealanders can access endovascular clot retrieval.
Clot retrieval treatment volumes can best be estimated by looking at intravenous alteplase treatment rates. This is because approximately one-third to one-fourth of alteplase-treated patients are eligible for clot retrieval. This number needs to be increased by 10–20% to account for the patients who are not eligible for alteplase but were still included in some of the studies (eg, patients presenting after the 4.5 hour intravenous alteplase window or who were taking dabigatran).
Intravenous thrombolysis treatment numbers in 2015 were 125 in the Northern, 101 in the Midland, 103 in the Central, and 76 in the South Island Regions.18 Based on these figures, and using the above assumptions, one would expect 50 patients to be treated with clot retrieval in the Northern region, 40 patients in the Midland and Central regions and 30 in the Southern region, per year, without any change in current thrombolysis rates.
These numbers will increase over the next 10 years for at least two reasons. Firstly, as presented above, stroke volumes will increase. Second, with greater regional organisation, access to acute therapies including thrombolysis and clot retrieval will increase as was demonstrated in the Telestroke Pilot evaluation.19 See Table 6 for projections. This is corroborated by data from a South Australian study,20 that estimates that 7% of ischaemic stroke patients are eligible. Based on current volumes this would mean about 400 clot retrieval-eligible patients in New Zealand now and 500 in 2028.
Table 6: Projected stroke and clot retrieval volumes by DHB in 2028.
†IV tPA = intravenous tissue plasminogen activator refers to ‘thrombolysis’ depicting projected volumes assuming a 10% rate (current national target).
†† Thrombolysis rates will increase through improved access via use of telestroke and other changes in model of care resulting in a likely much higher rate of around 20% in 2028 used in the subsequent column to provide a less conservative projections of anticipated clot retrieval volumes for 2028. These are not actual thrombolysis volumes. Each DHB is aware of their own current volumes and can use the above table to estimate current clot retrieval volumes based on their actual thrombolysis rate.
Furthermore, very recent studies already indicate that patient inclusion criteria will expand over time, resulting in more patients being eligible for this intervention.21 The impact of this remained uncertain at the time of submission of this paper and accordingly was not included in the below table.
As explained in the Methods section, centralisation of clot retrieval services in New Zealand is unavoidable. It is vital therefore that regional hyperacute stroke pathways are established so that there is equitable provision of acute stroke therapy across the country, rather than simply being available to populations living close to large tertiary hospitals. Achieving this will require systems reorganisation across all district health boards, in partnership with ambulance service providers.
New stroke models of care
Technological and scientific advances in the past three decades have seen the development of more specialised inter-disciplinary stroke units, although it is not always possible to offer these in small rural centres.
A number of studies have shown that decentralised thrombolysis services (telestroke, mobile stroke units, local stroke-ready health centres) can achieve similar thrombolysis rates and door-to-needle times for rural populations as a centralised system can achieve for an urban population.19,22 However, especially in the metropolitan setting when one includes other key therapeutic components and considers both patient outcomes and cost-efficiency, centralised systems tend to fare better.23
An example of this was the London Hyper-Acute Stroke Model, which consolidated the treatment of all early-phase (first 72 hours) acute stroke patients in London previously treated at 30 local hospitals into eight specialised high-volume centres designated hyperacute stroke units, or “HASUs”. Paramedic crews would no longer bring suspected acute stroke patients to their local hospital by ambulance, but instead to the nearest HASU, driving past multiple local hospitals that previously would have received them and even bypassing emergency departments.
An evaluation found that this new model of care resulted in a 25% reduction in mortality rates, marked improvement in processes of care, a rise in thrombolysis rate from 4% in 2010 to 11% in mid-2012, all while the costs of treating each stroke patient were reduced by 6%.24
Despite London’s unpredictable traffic conditions, 98% of patients arrive at the HASU within 30 minutes of being picked up by ambulance. Reports from patients about their experience with the new pathway have been largely positive even considering being cared for ‘further away from home’.24,25
A similar model is currently being implemented at night and on weekends in Auckland. Additional infrastructure and coordination resources are required to achieve this change in model of care, but it is well justified based on available data. Another area within New Zealand that would likely benefit from such collaborative centralisation is the Wellington metropolitan area by combining Hutt Valley and Capital & Coast DHB stroke services.
The value of centralisation in more rural areas remains uncertain. Because care is becoming more complex, access to specialists is becoming more important and the use of telemedicine for both acute stroke and rehabilitation will help bridge this gap as has already been demonstrated.19 Reducing clinician and patient travel times is also of benefit from an economic perspective, freeing up resources for other activities as patient volumes increase.
Use of registries and mobile devices can also help connect clinicians, monitor patient outcomes and service performance, and assist with ongoing service improvement to drive positive change. This will become increasingly important as volumes and patient complexity rise.
Stroke rehabilitation services have had limited attention over the past decades. To improve rehabilitation efficiency, inter-disciplinary teams need to be fully staffed and include all required professional groups to offer comprehensive and timely therapy. Similarly, investing more resources into community stroke services to support earlier and better support of patients in the community will help to reduce rehabilitation bed demand. New models of care, including the use of telerehabilitation to link patients in their homes to practitioners, alternate providers (eg, therapy assistants) and increasingly, self-care strategies, need to be developed and implemented in the stroke rehabilitation space.
It has been suggested that 90% of the world’s stroke burden is attributable to modifiable risk factors and that achieving control of behavioural and metabolic risk factors could avert more than three-quarters of strokes worldwide. In Oceania, the top five modifiable risk factors contributing to the stroke burden were high blood pressure, high body mass index, diet low in fruits, smoking and diet low in vegetables in order of impact.26 National policy changes and public awareness campaigns as well as new ideas as to how to best address these issues are vitally needed.
Finally, with increased complexity of care, care pathways, integration of care, and clear and open discussions around ceilings of care and advanced care planning are harder investments to quantify but should not be forgotten.
Discussion
Based on the available data, stroke incidence and mortality has reduced over the past three decades although internationally New Zealand stroke outcome performance lags behind other OECD countries. The New Zealand population is growing and ageing, and if further reductions in stroke incidence are not achieved stroke volumes can be anticipated to increase by 40% from 7,231 in 2015 to 10,112 in 2028.
Care will be further complicated by a disproportionate growth in volumes in rural areas with high degrees of social deprivation and disadvantaged ethnic population. There will also be an increasing need to centralise stroke services due to the complexity of modern stroke therapies.
Centralisation is inevitable in metropolitan areas and requires technology support in geographically dispersed areas to ensure that the disparity gap does not widen further.
Improving access to acute intervention and rehabilitation services is essential to ensure mortality and disability continue to decline and minimise the societal (and individual) impacts of the New Zealand’s increasing stroke burden. Upfront investments in these interventions, including workforce development, will maximise the possibility of containing health service resource requirements within capacity.
Investments in prevention will become increasingly important. We need to achieve a 30% reduction in stroke incidence to keep stroke volumes stable. This will require public policy consideration and massive public awareness campaigns.
This report has several limitations. The greatest limitation is that the main data source is ICD-10-AM hospital coding data that has not been validated. Coding errors are common in these types of data sources and thus the data has to be interpreted keeping this in mind. Prospectively collected proper epidemiological or registry data is not available in New Zealand beyond the Auckland region and even that data is now six years old. International data is too dissimilar to be used in its place. Also, data obtained from New Zealand statistics population data does not always correlate well with the NMDS data as regards time frames or localities, and potential inaccuracies are possible introducing further errors. Much of the projections are based on assumptions and hypothetical scenarios. New Zealand data on models of care is limited. Telestroke and stroke unit care have been found to be effective.19,27 However, there is no data around centralisation of stroke care, comparative analyses of different models of stroke unit care, or service models such as early supportive discharge in New Zealand. This makes extrapolations to the New Zealand setting tenuous and better New Zealand data is needed.
Conclusion
New Zealand stroke care improvements are not as prominent as in other OECD countries, but significant improvement has certainly been observed and more recent efforts will not be captured in the presented data. Nonetheless, while incidence rates have reduced, the anticipated population growth and ageing over the next decade will see a 40% increase in stroke admissions unless stroke incidence is further reduced through effective national primary prevention programmes. Long-term stroke burden due to disability and aged residential care requirements could be reduced through improved acute and rehabilitation stroke care, but this will require upfront service investments.
Summary
Abstract
Aim
Stroke incidence and mortality are declining due to effective public health initatives and modern healthcare advances. However, due to population growth and ageing, the burden of stroke continues to rise worldwide. This paper aims to provide stroke volume projections for the next 1-2 decades and explores potential solutions to anticipated challenges.
Method
Health administrative, where available epidemiological, and New Zealand Statistics data was used to model stroke service demand up to 2028.
Results
Despite improvements in stroke prevention and management, stroke volumes are projected to increase by 40% by 2028 due to population growth and, more importantly, ageing. Associated with this increase will be a need for more hospital beds and staff resources.
Conclusion
Efforts to optimise stroke service performance and, increasingly, stroke prevention are required to ensure that the New Zealand Health Service will be able to manage the increased volumes of patients. Better data is required to validate the presented figures, which are largely based on unvalidated health administrative data.
Author Information
Acknowledgements
Correspondence
Correspondence Email
Competing Interests
- Johnston SC, Mendis S, Mathers CD. Global variation in stroke burden and mortality: estimates from monitoring, surveillance, and modelling. Lancet Neurol 2009; 8:345–54.
- Brown P. Economic burden of stroke in New Zealand. Three decades of Auckland regional Community Stroke (ARCOS) studies: What have we learned and what is next for stroke care and stroke research? [PhD Thesis]. Auckland, New Zealand: AUT University; 2009.
- Brown P. Economic burden of stroke in New Zealand. In: Feigin VL, ed. Three decades of Auckland regional Community Stroke (ARCOS) studies: What have we learned and what is next for stroke care and stroke research? Auckland: AUT University; 2009 16.
- Thrift AG, Thayabaranathan T, Howard G, et al. Global stroke statistics. Int J Stroke 2017; 12:13–32.
- Feigin VL, Krishnamurthi RV, Barker-Collo S, et al. 30-Year Trends in Stroke Rates and Outcome in Auckland, New Zealand (1981–2012): A Multi-Ethnic Population-Based Series of Studies. PloS one 2015; 10:e0134609.
- Krishnamurthi R, Jones A, Barber PA, et al. Methodology of a population-based stroke and TIA incidence and outcomes study: the Auckland Regional Community Stroke Study (ARCOS IV) 2011–2012. Int J Stroke 2014; 9:140–7.
- Lavin TL, Ranta A. Transient ischameic attack/stroke electronic decision support: a 14-month safety audit. Journal of stroke and cerebrovascular diseases: the official journal of National Stroke Association 2014; 23:267–70.
- Child N, Barber PA, Fink J, Jones S, Voges K, Vivian M. New Zealand National Acute Stroke Services Audit 2009: organisation of acute stroke services in New Zealand. N Z Med J 2011; 124:13–20.
- Health Roundtable: New Zealand Stroke Report, 2017. http://www.healthroundtable.org/
- Awatea K. Supporting life after stroke: Counties Manukau DHB; 2015.
- McNaughton H, DeJong G, Smout RJ, Melvin JL, Brandstater M. A comparison of stroke rehabilitation practice and outcomes between New Zealand and United States facilities. Arch Phys Med Rehabil 2005; 86:S115–S20.
- Cadilhac DA, Kilkenny MF, Andrew NE, et al. Hospitals admitting at least 100 patients with stroke a year should have a stroke unit: a case study from Australia. BMC Health Serv Res 2017; 17:212.
- Tooth L, McKenna K, Goh K, Varghese P. Length of stay, discharge destination, and functional improvement: utility of the Australian National Subacute and Nonacute Patient Casemix Classification. Stroke 2005; 36:1519–25.
- Steventon A, Roberts A. Estimating length of stay in publicly-funded residential and nursing care homes: a retrospective analysis using linked administrative data sets. BMC Health Serv Res 2012; 12:377.
- Monks T, Worthington D, Allen M, Pitt M, Stein K, James MA. A modelling tool for capacity planning in acute and community stroke services. BMC Health Serv Res 2016; 16:530.
- Goyal M, Menon BK, van Zwam WH, et al. Endovascular thrombectomy after large-vessel ischaemic stroke: a meta-analysis of individual patient data from five randomised trials. Lancet 2016; 387:1723–31.
- Adamczyk P, Attenello F, Wen G, et al. Mechanical thrombectomy in acute stroke: utilization variances and impact of procedural volume on inpatient mortality. Journal of stroke and cerebrovascular diseases : the official journal of National Stroke Association 2013; 22:1263–9.
- Liu Q, Ranta AA, Abernethy G, Barber PA. Trends in New Zealand stroke thrombolysis treatment rates. N Z Med J 2017; 130:50–6.
- Ranta A, Lanford J, Busch S, Providence C, Iniesta I, Rosemergy I, Wilson A, Cariga P, Richmond V, Gommans J. Impact and implementation of a sustainable regional telestroke network. Intern Med J. 2017 Jul 25. doi: 10.1111/imj.13557
- Chia NH, Leyden JM, Newbury J, Jannes J, Kleinig TJ. Determining the Number of Ischemic Strokes Potentially Eligible for Endovascular Thrombectomy: A Population-Based Study. Stroke 2016; 47:1377–80.
- Nogueira RG, Jadhav AP, Haussen DC, et al. Thrombectomy 6 to 24 Hours after Stroke with a Mismatch between Deficit and Infarct. N Engl J Med 2017; 378:11–21.
- Lahr MM, Luijckx GJ, Vroomen PC, van der Zee DJ, Buskens E. Proportion of patients treated with thrombolysis in a centralized versus a decentralized acute stroke care setting. Stroke 2012; 43:1336–40.
- Lahr MM, van der Zee DJ, Luijckx GJ, Vroomen PC, Buskens E. Centralising and optimising decentralised stroke care systems: a simulation study on short-term costs and effects. BMC Med Res Methodol 2017; 17:5.
- Hunter RM, Davie C, Rudd A, et al. Impact on clinical and cost outcomes of a centralized approach to acute stroke care in London: a comparative effectiveness before and after model. PLoS One 2013; 8:e70420.
- Moynihan B, Paul S, Markus HS. User experience of a centralized hyperacute stroke service: a prospective evaluation. Stroke 2013; 44:2743–7.
- Feigin VL, Roth GA, Naghavi M, et al. Global burden of stroke and risk factors in 188 countries, during 1990–2013: a systematic analysis for the Global Burden of Disease Study 2013. The Lancet Neurology. 2016; 15(9):913–924.
- Te Ao BJ, Brown PM, Feigin VL, Anderson CS. Are stroke units cost effective? Evidence from a New Zealand stroke incidence and population-based study. Int J Stroke 2012; 7:623–30.
Contact diana@nzma.org.nz
for the PDF of this article
Projected stroke volumes to provide a 10-year direction for New Zealand stroke services
View Article PDF
Stroke is a leading cause of death and disability worldwide.1 In New Zealand, it has been estimated that around 8–9,000 people have a stroke each year and 50,000 people live with the consequences of stroke.2 The current annual cost has been estimated to sit at $NZ 700 million.3 Reducing the burden of stroke is a key goal to improve health outcomes in New Zealand.
Global and New Zealand stroke epidemiological studies have found important trends in stroke incidence and mortality over the past several decades.4 When comparing New Zealand to other high-income countries, stroke incidence is relatively high. For example, the age-adjusted incidence rate is reported to be 119 per 100,000 in Auckland compared with 76 per 100,000 in Adelaide.4 Mortality was also higher in New Zealand compared to other countries such as Australia. However, other less developed countries have incidence rates of 250 per 100,000.
The longitudinal Auckland Regional Community Stroke (ARCOS) studies have explored stroke incidence and mortality in the greater Auckland region over the past three decades and provide the most up-to-date high-quality local data.5 Overall stroke incidence has decreased by 23% and stroke mortality by 62% from 1981 to 2012. Māori and Pacific groups had a slower rate of decline and continue to experience stroke at a significantly younger age (mean ages 60 and 62 years) compared with New Zealand Europeans (mean age 75 years). The 28-day stroke case fatality has also dropped by 14% and this was seen across all ethnic groups.
It is reassuring that modern day public health initiatives, primary care preventive strategies, and secondary and tertiary level acute health interventions have made a difference. However, in light of a growing and ageing population the burden of stroke is likely going to increase despite these advances and initiatives. In addition, just like there are international variations in stroke incidence one can anticipate that there are geographic variations within New Zealand not readily captured by the ARCOS data.
This paper uses health administrative data to define the current stroke incidence across New Zealand, exploring geographic variation, and will provide future projections based on epidemiological trends. Potential solutions to address future challenge will be explored and discussed.
Methods
Building a model
The New Zealand demand for stroke services is primarily based on the incidence and prevalence of stroke in New Zealand. These in turn are influenced by demographic changes, lifestyle changes, risk factor management and efficacy of post-stroke intervention over time. Other contributing factors that influence demand include baseline population growth. Complexity and availability of diagnostics and interventions also impacts demand. Supply of stroke services depends largely on the workforce availability, which depends on training, immigration and attrition/retirement, but is also influenced by use of technology and shifting models of care to potentially more efficient approaches to patient management. Figure 1 presents a graphic model summarising the factors that impact stroke volumes.
Figure 1: Proposed model for stroke service impact estimates.
Data sources
Up-to-date nationwide New Zealand stroke audit or epidemiological data is not available. The analyses presented are based on Ministry of Health (MoH) National Minimal Dataset (NMDS) hospital discharge data. Over 95% of patients with stroke are now hospitalised, making the NMDS reasonably comprehensive, however data collection relies on hospital coders and does not undergo validation for accuracy introducing some limitation.5 For this reason, data from the high-quality ARCOS studies will be referenced at times.5 ARCOS data is limited to greater Auckland and is now several years old (last data collection 2011/12), making it unsuitable as the main data source for this project. International data is referenced also, but does not readily apply to New Zealand due to significant inter-country variations in stroke incidence.4
MoH data for the period from 1 July 2015 through 30 June 2016 (latest full financial year data available at the time of this analysis) was used to determine ‘current’ stroke volumes. Data is grouped by district health board (DHB) population, age and ethnicity.
Populations figures are based on the most recent (2013) national census data. Where available, New Zealand statistics projection estimates for 2015/16 were used (50th percentile figures). DHB-stratified population data was only available for 2013 without published projections and thus the 2013 actual numbers had to be used for the DHB-based rate estimates.
The ARCOS data included for comparison purposes was collected 1 March 2011 through 29 February 2012.6
Stroke diagnoses included in this analysis are ICD-10-AM I61 (intracerebral haemorrhage), I63 (cerebral infarction) and I64 (stroke unspecified). Subarachnoid haemorrhage is excluded, as this is generally managed by neurosurgical rather than acute stroke services. TIA is excluded because many TIA patients are managed as outpatients and thus not captured in discharge data, resulting in too many missing data points to allow for accurate modelling.7
Modelling for future stroke services
Future stroke volume projections were modelled by adjusting for population growth by domicile and age. First, all of New Zealand data was calculated applying both the 2011/12 ARCOS stroke incidence rate and the 2015/16 raw MoH incidence rate by New Zealand population projections in five-year brackets until 2068. Then the current rate of strokes under versus over 65 years of age were calculated, thus arriving at current age-adjusted incidence rates. This allowed projection of stroke rates incorporating New Zealand statistics age bracket prediction for future years. This data was then compared to the ARCOS age-adjusted incidence figures.
Population growth will not occur evenly across New Zealand. To account for such geographic variation, locality specific predictions presented by New Zealand statistics over the next 20 years were explored. For each locality, growth over the next 10 and 20 years is expressed as a factor of 1.0. If the number is less than 1.0 it indicates a drop in population and if greater than 1.0 growth. The decimal digits indicate percentage change (for example 1.15 indicates a 15% growth and 0.95 a 5% drop).
Figures were normalised around 1.0 to depict relative growth difference between DHBs. These figures were then combined with age-adjusted growth to arrive at individual DHB projections for stroke admission volumes in 2028 and 2038. Because growth data by urban area was not available for DHBs, some assumptions had to be made. For example, it was assumed that the main urban centre in each DHB is the main driver of population change based on highest population volume. Generally speaking, the spread across localities within DHBs was minimal (eg, Palmerston North 1.06 and Feilding 1.05 in MidCentral DHB). Overall growth was normalised across all DHBs, meaning the total number of strokes remains accurate. However, it is possible that estimates are off by a few patients in either direction for each DHB.
Modelling future secondary stroke services
The estimates of hospital bed occupancy were derived by multiplying the stroke volumes per DHB by average length of stay (LOS) for acute (6 days)8,9 and rehabilitation (21 days).10,11 The rehabilitation figure was multiplied by 0.28 to reflect that on average only 28% of stroke patients are transferred to inpatient rehabilitation (2016/17 MoH DHB reporting data). Aged residential care bed requirements were estimated by multiplying the total number of strokes by 0.15 to reflect an average discharge rate of 15% to these facilitates11–13 multiplied by the average duration patients reside at ARC facilities. A figure for this is not available specifically for stroke patients. General median LOS prior to dying or discharge was between 5 and 25 months depending on publication.14 For the purposes of this analysis a figure of 12 months was used.
Final annual figures were divided by 365 days and multiplied by 0.15 to allow for optimum patient flow. This is a conservative figure with international work indicating that optimal patient flow to maximise cost-efficiency and patient safety is achieved by inflating bed numbers by up to 40%.15 Flow is less critical in longer-stay facilities and ARC bed days were increased by only 5% accordingly.
Staffing levels can be estimated by looking at the number of patients admitted per year averaged per day, proportion of patients seen by a given clinician, duration of consultations, number of assessments per week, shift length and coverage (ie, number of shifts covered per day/week). Non-clinical/administrative time also needs to be factored in. There are available online stroke unit staff calculators available to assist with this (eg, http://www.stroke-education.org.uk/staff-calculators/).
Modelling for future tertiary services
Much of stroke care provided at New Zealand tertiary centres mimics the care provided at secondary hospitals. A notable exception is endovascular clot retrieval (ECR). This is a key new intervention provided by acute stroke services that is anticipated to have a significant impact on the future of stroke care provision. Clot retrieval modelling is complex due to several key features: (1) clot retrieval is time critical, needing to be commenced within hours of symptom onset;16 (2) clot retrieval is technically challenging, requiring highly skilled interventionalists;17 (3) clot retrieval needs to be available 24 hours per day to avoid treatment gaps requiring a 1:3–1:4 roster; and (4) clot retrieval volumes are relatively low, making it unfeasible to currently offer this service safely at more than three centres in New Zealand.
To achieve equitable provision of clot retrieval, the National Stroke Network has recently produced a National Clot Retrieval Strategy that identifies Auckland, Wellington and Christchurch as the three tertiary centres to offer this service until volumes increase to support a fourth centre. This paper explores projected clot retrieval numbers into the future.
Results
Current stroke volumes
As of 2015, the estimated New Zealand population overall was 4,567,000 people. Per the National Minimum Dataset data there were 8,450 stroke discharges (ICD10-AM I61, I63 and I64) in 2015, resulting in a hospital admission rate for stroke (primary discharge diagnosis) of 185/100,000. This data differs from the ARCOS data, which found a stroke incidence of 147/100,000 new stroke in 2011/2012 in the greater Auckland region. This discrepancy could be explained by several factors. Auckland is not representative of all of New Zealand; the discharge data contains coding errors, captures some people with stroke admissions that include readmissions for the same stroke event, or the ARCOS data missed some patients.
Because this report aims to provide an all of New Zealand perspective with a primary focus on health service utilisation to assist with service modelling, the MoH figures will be used although ARCOS figures will be referred to at times. The limitations are acknowledged and future studies should aim to provide better data to improve the accuracy of such models as are presented here.
Strokes are not evenly distributed across New Zealand. Table 1 shows the stroke admissions per 100,000 by DHB and Figure 2 shows ethnic distribution of stroke admissions by DHB.
Table 1: Current stroke volumes by DHB.
*NMDS figures for 12 months starting 1 July 2015.
†Statistics NZ figures for DHB population based on 2013 census data.
Figure 2: Ethnic distribution by DHB.
The above regional differences correlate with high number of non-New Zealand European populations and likely social deprivation, but this was not explored in detail for this project.
Future stroke volume projections
Per New Zealand statistics, the population will grow to 5,389,700 by 2028, 5,769,800 by 2038, and 6,060,500 by 2048. If stroke incidence remains static, the associated volumes of stroke admissions per annum respectively can thus be anticipated: 9,972, 10,675, 11,213.
However, the population is ageing (see Figure 3) and accordingly one would expect an associated disproportionate increase over time rather than a static incidence across the population.
Figure 3: New Zealand Population age estimated 2008–2068.
Adjusting for a disproportionate increase of patients >65 years of age and in light of the fact that 76% of stroke occur in this age group, a more accurate estimate would be the following admission volumes for 2028, 2038 and 2048 respectively: 11,828, 14,282 and 15,532. A breakdown of anticipated stroke volumes by age is depicted in Table 2.
Table 2: Projected stroke volumes adjusted by age.
There is also an anticipated growth of Māori and Pacific sub-populations. As the stroke incidence is higher in these ethnic groups (with earlier age of onset) these groups will become increasingly important to consider with an anticipated growth from 836 Māori and 470 Pacific stroke patients in 2015 to 976 Māori and 589 Pacific in 2028 and 1,195 Māori and 725 Pacific patients in 2038. Interestingly the ARCOS data suggests an even greater disparity with an anticipated 1,700 Māori and 1,180 Pacific stroke sufferers in 2028. However, there is no particularly disproportionate growth within this sub-population and the ethnicity spread was thus not further included in subsequent modelling.
Population growth is not evenly distributed. Auckland will experience the greatest growth with up to 72% increase in population in some localities over the next 20 years. Other areas will stay, by comparison, relatively stable (see Table 3).
Table 3: Population growth by geographic location over next 20 years.
While some especially large urban centres will experience disproportionate population growth, the effect of ageing of the population represents a more significant driver of projected stroke volumes than other factors such as immigration. Table 4 projects DHB-based stroke volumes in 2028 adjusting for both age- and locality-specific anticipated growth—some of which is driven by immigration.
Table 4: Stroke volumes (age adjusted) by DHB accounting for regional variation in projected population growth.
*Locality adjusted = volumes adjusted by age and population growth taking into account variation in population growth across the country.
Projected stroke-related resource needs
Based on these projected figures, estimates on anticipated bed requirements and staffing levels are possible. Table 5 shows some staffing estimates for 2028. This list is not exhaustive, focuses on only some professional groups limited to acute inpatient care, and is based on untested assumptions around the time spent with patients per day. This is illustrated as a potential tool to develop comprehensive workforce strategies, but requires further validation before using this data to guide service implementation.
Table 5: Projected institutional stroke resource needs by 2028.
*FTE = Full time equivalents based on a 40-hour work week = 1.0 FTE; ARC = aged residential care; CNS = clinical nurse specialist; PT = physiotherapist.
Projected regional tertiary stroke volumes
Regional services will have to be formalised by 2028 to ensure New Zealanders can access endovascular clot retrieval.
Clot retrieval treatment volumes can best be estimated by looking at intravenous alteplase treatment rates. This is because approximately one-third to one-fourth of alteplase-treated patients are eligible for clot retrieval. This number needs to be increased by 10–20% to account for the patients who are not eligible for alteplase but were still included in some of the studies (eg, patients presenting after the 4.5 hour intravenous alteplase window or who were taking dabigatran).
Intravenous thrombolysis treatment numbers in 2015 were 125 in the Northern, 101 in the Midland, 103 in the Central, and 76 in the South Island Regions.18 Based on these figures, and using the above assumptions, one would expect 50 patients to be treated with clot retrieval in the Northern region, 40 patients in the Midland and Central regions and 30 in the Southern region, per year, without any change in current thrombolysis rates.
These numbers will increase over the next 10 years for at least two reasons. Firstly, as presented above, stroke volumes will increase. Second, with greater regional organisation, access to acute therapies including thrombolysis and clot retrieval will increase as was demonstrated in the Telestroke Pilot evaluation.19 See Table 6 for projections. This is corroborated by data from a South Australian study,20 that estimates that 7% of ischaemic stroke patients are eligible. Based on current volumes this would mean about 400 clot retrieval-eligible patients in New Zealand now and 500 in 2028.
Table 6: Projected stroke and clot retrieval volumes by DHB in 2028.
†IV tPA = intravenous tissue plasminogen activator refers to ‘thrombolysis’ depicting projected volumes assuming a 10% rate (current national target).
†† Thrombolysis rates will increase through improved access via use of telestroke and other changes in model of care resulting in a likely much higher rate of around 20% in 2028 used in the subsequent column to provide a less conservative projections of anticipated clot retrieval volumes for 2028. These are not actual thrombolysis volumes. Each DHB is aware of their own current volumes and can use the above table to estimate current clot retrieval volumes based on their actual thrombolysis rate.
Furthermore, very recent studies already indicate that patient inclusion criteria will expand over time, resulting in more patients being eligible for this intervention.21 The impact of this remained uncertain at the time of submission of this paper and accordingly was not included in the below table.
As explained in the Methods section, centralisation of clot retrieval services in New Zealand is unavoidable. It is vital therefore that regional hyperacute stroke pathways are established so that there is equitable provision of acute stroke therapy across the country, rather than simply being available to populations living close to large tertiary hospitals. Achieving this will require systems reorganisation across all district health boards, in partnership with ambulance service providers.
New stroke models of care
Technological and scientific advances in the past three decades have seen the development of more specialised inter-disciplinary stroke units, although it is not always possible to offer these in small rural centres.
A number of studies have shown that decentralised thrombolysis services (telestroke, mobile stroke units, local stroke-ready health centres) can achieve similar thrombolysis rates and door-to-needle times for rural populations as a centralised system can achieve for an urban population.19,22 However, especially in the metropolitan setting when one includes other key therapeutic components and considers both patient outcomes and cost-efficiency, centralised systems tend to fare better.23
An example of this was the London Hyper-Acute Stroke Model, which consolidated the treatment of all early-phase (first 72 hours) acute stroke patients in London previously treated at 30 local hospitals into eight specialised high-volume centres designated hyperacute stroke units, or “HASUs”. Paramedic crews would no longer bring suspected acute stroke patients to their local hospital by ambulance, but instead to the nearest HASU, driving past multiple local hospitals that previously would have received them and even bypassing emergency departments.
An evaluation found that this new model of care resulted in a 25% reduction in mortality rates, marked improvement in processes of care, a rise in thrombolysis rate from 4% in 2010 to 11% in mid-2012, all while the costs of treating each stroke patient were reduced by 6%.24
Despite London’s unpredictable traffic conditions, 98% of patients arrive at the HASU within 30 minutes of being picked up by ambulance. Reports from patients about their experience with the new pathway have been largely positive even considering being cared for ‘further away from home’.24,25
A similar model is currently being implemented at night and on weekends in Auckland. Additional infrastructure and coordination resources are required to achieve this change in model of care, but it is well justified based on available data. Another area within New Zealand that would likely benefit from such collaborative centralisation is the Wellington metropolitan area by combining Hutt Valley and Capital & Coast DHB stroke services.
The value of centralisation in more rural areas remains uncertain. Because care is becoming more complex, access to specialists is becoming more important and the use of telemedicine for both acute stroke and rehabilitation will help bridge this gap as has already been demonstrated.19 Reducing clinician and patient travel times is also of benefit from an economic perspective, freeing up resources for other activities as patient volumes increase.
Use of registries and mobile devices can also help connect clinicians, monitor patient outcomes and service performance, and assist with ongoing service improvement to drive positive change. This will become increasingly important as volumes and patient complexity rise.
Stroke rehabilitation services have had limited attention over the past decades. To improve rehabilitation efficiency, inter-disciplinary teams need to be fully staffed and include all required professional groups to offer comprehensive and timely therapy. Similarly, investing more resources into community stroke services to support earlier and better support of patients in the community will help to reduce rehabilitation bed demand. New models of care, including the use of telerehabilitation to link patients in their homes to practitioners, alternate providers (eg, therapy assistants) and increasingly, self-care strategies, need to be developed and implemented in the stroke rehabilitation space.
It has been suggested that 90% of the world’s stroke burden is attributable to modifiable risk factors and that achieving control of behavioural and metabolic risk factors could avert more than three-quarters of strokes worldwide. In Oceania, the top five modifiable risk factors contributing to the stroke burden were high blood pressure, high body mass index, diet low in fruits, smoking and diet low in vegetables in order of impact.26 National policy changes and public awareness campaigns as well as new ideas as to how to best address these issues are vitally needed.
Finally, with increased complexity of care, care pathways, integration of care, and clear and open discussions around ceilings of care and advanced care planning are harder investments to quantify but should not be forgotten.
Discussion
Based on the available data, stroke incidence and mortality has reduced over the past three decades although internationally New Zealand stroke outcome performance lags behind other OECD countries. The New Zealand population is growing and ageing, and if further reductions in stroke incidence are not achieved stroke volumes can be anticipated to increase by 40% from 7,231 in 2015 to 10,112 in 2028.
Care will be further complicated by a disproportionate growth in volumes in rural areas with high degrees of social deprivation and disadvantaged ethnic population. There will also be an increasing need to centralise stroke services due to the complexity of modern stroke therapies.
Centralisation is inevitable in metropolitan areas and requires technology support in geographically dispersed areas to ensure that the disparity gap does not widen further.
Improving access to acute intervention and rehabilitation services is essential to ensure mortality and disability continue to decline and minimise the societal (and individual) impacts of the New Zealand’s increasing stroke burden. Upfront investments in these interventions, including workforce development, will maximise the possibility of containing health service resource requirements within capacity.
Investments in prevention will become increasingly important. We need to achieve a 30% reduction in stroke incidence to keep stroke volumes stable. This will require public policy consideration and massive public awareness campaigns.
This report has several limitations. The greatest limitation is that the main data source is ICD-10-AM hospital coding data that has not been validated. Coding errors are common in these types of data sources and thus the data has to be interpreted keeping this in mind. Prospectively collected proper epidemiological or registry data is not available in New Zealand beyond the Auckland region and even that data is now six years old. International data is too dissimilar to be used in its place. Also, data obtained from New Zealand statistics population data does not always correlate well with the NMDS data as regards time frames or localities, and potential inaccuracies are possible introducing further errors. Much of the projections are based on assumptions and hypothetical scenarios. New Zealand data on models of care is limited. Telestroke and stroke unit care have been found to be effective.19,27 However, there is no data around centralisation of stroke care, comparative analyses of different models of stroke unit care, or service models such as early supportive discharge in New Zealand. This makes extrapolations to the New Zealand setting tenuous and better New Zealand data is needed.
Conclusion
New Zealand stroke care improvements are not as prominent as in other OECD countries, but significant improvement has certainly been observed and more recent efforts will not be captured in the presented data. Nonetheless, while incidence rates have reduced, the anticipated population growth and ageing over the next decade will see a 40% increase in stroke admissions unless stroke incidence is further reduced through effective national primary prevention programmes. Long-term stroke burden due to disability and aged residential care requirements could be reduced through improved acute and rehabilitation stroke care, but this will require upfront service investments.
Summary
Abstract
Aim
Stroke incidence and mortality are declining due to effective public health initatives and modern healthcare advances. However, due to population growth and ageing, the burden of stroke continues to rise worldwide. This paper aims to provide stroke volume projections for the next 1-2 decades and explores potential solutions to anticipated challenges.
Method
Health administrative, where available epidemiological, and New Zealand Statistics data was used to model stroke service demand up to 2028.
Results
Despite improvements in stroke prevention and management, stroke volumes are projected to increase by 40% by 2028 due to population growth and, more importantly, ageing. Associated with this increase will be a need for more hospital beds and staff resources.
Conclusion
Efforts to optimise stroke service performance and, increasingly, stroke prevention are required to ensure that the New Zealand Health Service will be able to manage the increased volumes of patients. Better data is required to validate the presented figures, which are largely based on unvalidated health administrative data.
Author Information
Acknowledgements
Correspondence
Correspondence Email
Competing Interests
- Johnston SC, Mendis S, Mathers CD. Global variation in stroke burden and mortality: estimates from monitoring, surveillance, and modelling. Lancet Neurol 2009; 8:345–54.
- Brown P. Economic burden of stroke in New Zealand. Three decades of Auckland regional Community Stroke (ARCOS) studies: What have we learned and what is next for stroke care and stroke research? [PhD Thesis]. Auckland, New Zealand: AUT University; 2009.
- Brown P. Economic burden of stroke in New Zealand. In: Feigin VL, ed. Three decades of Auckland regional Community Stroke (ARCOS) studies: What have we learned and what is next for stroke care and stroke research? Auckland: AUT University; 2009 16.
- Thrift AG, Thayabaranathan T, Howard G, et al. Global stroke statistics. Int J Stroke 2017; 12:13–32.
- Feigin VL, Krishnamurthi RV, Barker-Collo S, et al. 30-Year Trends in Stroke Rates and Outcome in Auckland, New Zealand (1981–2012): A Multi-Ethnic Population-Based Series of Studies. PloS one 2015; 10:e0134609.
- Krishnamurthi R, Jones A, Barber PA, et al. Methodology of a population-based stroke and TIA incidence and outcomes study: the Auckland Regional Community Stroke Study (ARCOS IV) 2011–2012. Int J Stroke 2014; 9:140–7.
- Lavin TL, Ranta A. Transient ischameic attack/stroke electronic decision support: a 14-month safety audit. Journal of stroke and cerebrovascular diseases: the official journal of National Stroke Association 2014; 23:267–70.
- Child N, Barber PA, Fink J, Jones S, Voges K, Vivian M. New Zealand National Acute Stroke Services Audit 2009: organisation of acute stroke services in New Zealand. N Z Med J 2011; 124:13–20.
- Health Roundtable: New Zealand Stroke Report, 2017. http://www.healthroundtable.org/
- Awatea K. Supporting life after stroke: Counties Manukau DHB; 2015.
- McNaughton H, DeJong G, Smout RJ, Melvin JL, Brandstater M. A comparison of stroke rehabilitation practice and outcomes between New Zealand and United States facilities. Arch Phys Med Rehabil 2005; 86:S115–S20.
- Cadilhac DA, Kilkenny MF, Andrew NE, et al. Hospitals admitting at least 100 patients with stroke a year should have a stroke unit: a case study from Australia. BMC Health Serv Res 2017; 17:212.
- Tooth L, McKenna K, Goh K, Varghese P. Length of stay, discharge destination, and functional improvement: utility of the Australian National Subacute and Nonacute Patient Casemix Classification. Stroke 2005; 36:1519–25.
- Steventon A, Roberts A. Estimating length of stay in publicly-funded residential and nursing care homes: a retrospective analysis using linked administrative data sets. BMC Health Serv Res 2012; 12:377.
- Monks T, Worthington D, Allen M, Pitt M, Stein K, James MA. A modelling tool for capacity planning in acute and community stroke services. BMC Health Serv Res 2016; 16:530.
- Goyal M, Menon BK, van Zwam WH, et al. Endovascular thrombectomy after large-vessel ischaemic stroke: a meta-analysis of individual patient data from five randomised trials. Lancet 2016; 387:1723–31.
- Adamczyk P, Attenello F, Wen G, et al. Mechanical thrombectomy in acute stroke: utilization variances and impact of procedural volume on inpatient mortality. Journal of stroke and cerebrovascular diseases : the official journal of National Stroke Association 2013; 22:1263–9.
- Liu Q, Ranta AA, Abernethy G, Barber PA. Trends in New Zealand stroke thrombolysis treatment rates. N Z Med J 2017; 130:50–6.
- Ranta A, Lanford J, Busch S, Providence C, Iniesta I, Rosemergy I, Wilson A, Cariga P, Richmond V, Gommans J. Impact and implementation of a sustainable regional telestroke network. Intern Med J. 2017 Jul 25. doi: 10.1111/imj.13557
- Chia NH, Leyden JM, Newbury J, Jannes J, Kleinig TJ. Determining the Number of Ischemic Strokes Potentially Eligible for Endovascular Thrombectomy: A Population-Based Study. Stroke 2016; 47:1377–80.
- Nogueira RG, Jadhav AP, Haussen DC, et al. Thrombectomy 6 to 24 Hours after Stroke with a Mismatch between Deficit and Infarct. N Engl J Med 2017; 378:11–21.
- Lahr MM, Luijckx GJ, Vroomen PC, van der Zee DJ, Buskens E. Proportion of patients treated with thrombolysis in a centralized versus a decentralized acute stroke care setting. Stroke 2012; 43:1336–40.
- Lahr MM, van der Zee DJ, Luijckx GJ, Vroomen PC, Buskens E. Centralising and optimising decentralised stroke care systems: a simulation study on short-term costs and effects. BMC Med Res Methodol 2017; 17:5.
- Hunter RM, Davie C, Rudd A, et al. Impact on clinical and cost outcomes of a centralized approach to acute stroke care in London: a comparative effectiveness before and after model. PLoS One 2013; 8:e70420.
- Moynihan B, Paul S, Markus HS. User experience of a centralized hyperacute stroke service: a prospective evaluation. Stroke 2013; 44:2743–7.
- Feigin VL, Roth GA, Naghavi M, et al. Global burden of stroke and risk factors in 188 countries, during 1990–2013: a systematic analysis for the Global Burden of Disease Study 2013. The Lancet Neurology. 2016; 15(9):913–924.
- Te Ao BJ, Brown PM, Feigin VL, Anderson CS. Are stroke units cost effective? Evidence from a New Zealand stroke incidence and population-based study. Int J Stroke 2012; 7:623–30.
Contact diana@nzma.org.nz
for the PDF of this article
Projected stroke volumes to provide a 10-year direction for New Zealand stroke services
Subscriber Content
The full contents of this pages only available to subscribers.
Login, subscribe or email nzmj@nzma.org.nz to purchase this article.
