Search

OBJECTIVE: To assess the feasibility of conducting a randomized controlled trial of an instructional and educational stroke DVD and determine the feasibility and preliminary efficacy of this intervention in a multinational context. DESIGN: Non-funded, pilot randomized controlled trial of intervention versus usual care. SETTING: International, multicentre, community-based. PARTICIPANTS: Community-living adults up to three years post stroke with moderate to severe disability and their nominated informal caregivers. INTERVENTIONS: Intervention patients viewed and practised rehabilitation techniques demonstrated in the DVD over six weeks. MAIN MEASURES: Trial feasibility by number of active recruitment sites, recruitment efficiency, randomization and follow-up. Intervention feasibility by patient and caregiver impressions. Preliminary efficacy by the quality of life - 5-level EuroQol-5D (EQ-5D) health status measure, General Health Questionnaire and Centre for Epidemiological Studies-Depression at two months. RESULT(S): In total, 14 recruitment sites were established across eight countries. Recruitment was achieved at nine (64%) sites. Over 16months, 66 participants were recruited (mean (SD) age=63.5 (12.47)years) and randomized to intervention ( n=34) and control ( n=32) groups. In total, 54 (82%) completed a follow-up assessment. Patient and/or caregiver comments about the benefits and barriers to accessing the intervention were mixed. There were no significant between-group differences in outcomes at two months ( P>0.05). CONCLUSION(S): Conducting a multinational trial of a stroke DVD requires full funding. The intervention was acceptable to some patients and their caregivers, yet a generalized education approach did not fully meet their needs and/or expectations. A more individualized method may be required to meet peoples' changing needs during stroke recovery.

Background and Aims: The Very Early Rehabilitation of SpEech after stroke (VERSE) trial recently showed no benefit of more therapy (26 hours) compared to usual care (13 hours) delivered in the first 12 weeks of stroke recovery. This paper examines the question: What therapy frequency and session length is associated with recovery in individual aphasia therapy regimens in the first 12 weeks after stroke? Methods: Participants with aphasia, recruited from 17 Australia/New Zealand acute-stroke units were randomised to one of three treatment arms. This secondary analysis used linear mixed models to investigate the effects of therapy frequency and session length on the Western Aphasia Battery-Aphasia Quotient (AQ) at 12 weeks, after controlling for therapy amount. Result(s): Participants (n=214) received aphasia therapy (mean=1377 minutes; sd=735) in 36 sessions (sd=19) with mean session length: 39 minutes (sd=11) in the first 12 weeks. The mean AQ was 41.2 (sd=28.0) at baseline and 67.7 (sd=29.5) at 12 weeks. After controlling for baseline stroke and aphasia severity, site and therapy amount, the session length had a significant effect (b=0.33, p=0.024) of increasing AQ at 12 weeks by 0.33 points per minute increase in session length. On average, 1350 minutes of therapy provided in 30 sessions of 45 minutes each, can be expected to achieve 5 points more on AQ at 12 weeks, than if 45 sessions of 30 minutes each was provided. Conclusion(s): Sessions of longer duration were associated with better recovery than a larger number of shorter sessions which achieved the same total amount of therapy.

Background and Aims: In developed countries, women appear to receive evidence-based care less often than men. Whether there are sex differences in stroke care and outcomes in developing countries is uncertain. Method(s):We included consecutive strokes admitted to the 115 People's Hospital - a major metropolitan public hospital in Vietnam during Jun 2017-Mar 2018. The data were collected using the REgistry of Stroke Care Quality (RES-Q). Baseline characteristics, clinical care provided in hospital (e.g. stroke unit access, thrombolysis), and outcomes at discharge (mortality and functional outcome assessed by the ability to walk without assistance) were compared between sexes. Result(s): Data were available for 6,601 strokes (44% women, 81% ischemic stroke). Compared to men, women were older (mean age 65 vs 60 years), less often current smokers (16% vs 63%) and had more atrial fibrillation detected by screening (12% vs men 8%; all p values<0.001). More women had assessment for rehabilitation needs <=72 hours (35% vs 31%, p=0.022) or carotid investigation <=7 days after hospital admission (4% vs 6%, p=0.041). There were no differences in receipt of other clinical care between sexes (Figure). Women had somewhat greater inhospital mortality 3.4% vs 2.7%; p=0.139) and lesser ability to walk on discharge (47% vs 44%; p=0.031) than men, but the differences were small and not influenced by the clinical care received. Conclusion(s): Generally, men and women with acute stroke received equivalent hospital care. Outcomes at discharge did not differ by sex. Nonetheless, uncertainty exists over the sex differences in long-term outcomes, requiring longer-term follow-ups. (Figure Presented).

Background and Aims: We aimed to examine sex differences in disease profiles, management, and survival at 1 and 5 years after ischemic stroke (IS) among people with atrial fibrillation (AF). Method(s): Individual participant data on covariates and survival were obtained from 9 population-based stroke incidence studies conducted in Australasia, Europe, and South America (1993-2014). The presence of AF at the time of stroke onset was self-reported (2 studies) or confirmed by ECG/medical records (7 studies). The mortality rate ratio (MRR) between women and men was estimated using Poisson modelling. Study-specific unadjusted and adjusted MRRs were combined using random-effects meta-analysis. Result(s): There were 8645 participants (mean age: 72.6; 51.8% women). The pooled AF prevalence was 26.6% in women and 20.8% in men. Among the 1862 IS patients with AF, women were older than men. There were no significant sex differences in prescription of antihypertensive, antiplatelet and anticoagulation agents at admission. Crude pooled 1-year mortality was greater for women (30.4%) than men (24.5%) (1- year MRR 1.24; 95% CI, 1.01-1.51; 5-year MRR 1.28; 0.98-1.66). However, the sex difference was greatly attenuated after accounting for age, pre-stroke function and stroke severity (1-year MRR 1.09; 95% CI 0.97-1.22; 5-year MRR 0.81; 0.60-1.09). Conclusion(s): AF was more prevalent after IS among women than men. Among those with AF and IS, stroke management was similar irrespective of sex. Greater mortality in women with AF after IS was mostly attributable to pre-stroke factors. The data are very encouraging in that sex differences in management and outcomes were absent. .

Background Urinary incontinence has negative effects for patients and health systems. In recent Stroke Foundation audits, nationally 35% of acute and 41% of rehabilitation inpatients had urinary incontinence, but only half had associated management plans. Aim To determine if broader implementation (multiple Health Districts, across phases of care) of our previously-piloted multidisciplinary Structured urinary Continence Assessment and Management Plan (SCAMP) interventions (including clinically-applicable tools, processes and education; audit and feedback) improves inpatient post-stroke urinary incontinence assessment, diagnosis and management. Methods Design: before- and after-implementation study at six acute stroke/ medical units, two comprehensive stroke and six rehabilitation services in three New South Wales Local Health Districts.Primary outcome: change in the proportion of incontinent patients who have a continence management plan.Secondary outcomes: change in the proportion of patients who have a urinary continence assessment, documented incontinence diagnosis and receive continence education. We will also determine the intervention effect on clinician knowledge, skills and confidence, and potential cost-effectiveness (hospital perspective).Data collection: medical record audits, clinician questionnaires and site-specific teams identifying local barriers and enablers to continence management. Data will be collected for 3 months before and after the 7-month implementation period. Sustainability will be evaluated 16 months after implementation commences.Results Ethics is approved. Before-implementation data collection commenced January 2019. The project will be completed by July 2020. Discussion Our SCAMP interventions can potentially operationalise high-level concepts outlined in Australian stroke guidelines into effective clinical care across hospitals a

Background and Aims:Urinary incontinence is a common, costly but often overlooked stroke complication. We aim to determine if implementing our SCAMP intervention (includes previously-piloted clinically-applicable tools and processes and education, audit and feedback) improves urinary incontinence assessment and management for in-hospital patients following stroke. Methods:Design: pragmatic, before- and after-implementation study at six acute stroke/ medical units, two comprehensive stroke and six rehabilitation services in NSW, Australia. Primary outcome: change in the proportion of incontinent patients who have a continence management plan. Secondary outcomes: change in the proportion of patients who have a urinary continence assessment, recorded diagnosis of incontinence type/s, and receive continence education. We will also determine the intervention effect on clinician knowledge, skills and confidence, and the potential cost-effectiveness from a hospital perspective. Data collection: medical record audits, clinician questionnaires and site-specific teams identifying local barriers and enablers to continence management. Data will be collected for 3 months before and after the 7-month implementation period. To assess sustainability, outcomes will again be measured 16 months after implementation commences. Fifteen consecutive medical record audits (incontinent patients) per site per month will provide >80% power to detect a 20% absolute increase (from before intervention) in the proportion of incontinent patients with an incontinence plan (type 1 error rate of 5%). Results:Ethics has been approved. Before-implementation data collection is underway. The project will be completed by July 2020. Conclusions:We will translate high-level concepts outlined in clinical guidelines into effective and efficient continence care. SCAMP has the potential to be read

Background: Discharge care planning that includes person-centred goal setting is recommended, but is inconsistently undertaken in practice. Aim(s): To develop and field test a goal setting package for use in comprehensive discharge care planning following stroke. Method(s): A multidisciplinary, national expert working group of clinicians and academics (N=15) was convened. Panel consensus was obtained using an iterative approach to develop a discharge goal setting package based on a previously developed menu. Field testing occurred in acute and subacute settings. Following training, behavioural observation methods were used to document clinical use of the package. Feedback was obtained from clinician focus groups with reference to the Theoretical Domains Framework. Result(s): The final package included a 34-item menu aligned to a manual containing: guideline summaries; common goals; SMART goal metrics; evidence-based strategies; and worked examples. Twenty-three clinicians from two sites attended training and five were observed by the researcher. Observations revealed that: the package could be incorporated into practice; an appropriate mix of person-centred goals were set; and the menu provided opportunities to raise additional issues. The process took 30-60 minutes. Focus group respondents (N=8) indicated that the package: was clinically useful and ensured goals were person-centred. Enablers included: ability to incorporate the package into existing processes; and beliefs that it promoted better patient care. Barriers included: additional time taken. Conclusion(s): Following refinements the goal setting package will provide a much-needed tool to facilitate more effective person-centred discharge planning for those with stroke. Further testing in a larger trial should confirm its utility.

Background: The first national survey of acute hospital services was conducted in 1999 by the Stroke Foundation to provide a snapshot of stroke service provision and attitudes towards interventions for the management. A subsequent survey was performed in 2004 which preceded the establishment of the National Audit program in 2007. Aim(s): To describe changes in services and resources for acute stroke over time. Method(s): The identification of acute public hospitals invited to participate varied from survey to survey, with eligibility primarily based on stroke admissions. All data was self-reported by a clinical representative considered appropriate to describe stroke care within their hospital. Survey data from 1999, 2004, 2007, 2009, 2011, 2013, 2015 and 2017 were mapped for consistency to enable reliable comparisons. Private hospitals were excluded from this analysis. Result(s): In 1999, 36/163 hospitals (22%) had an acute stroke unit (SU) or combined acute/rehabilitation SU, 93 (57%) had access to computed tomography (CT) within 24 hours, and 7 (4%) had access to magnetic resonance imaging (MRI) within 24 hours. In 2017, 89/121 hospitals (74%) had an acute SU, 117 (97%) reported access to CT (within 3 hours), and 86 (71%) had MRI access within 24 hours. Rapid assessment outpatient clinics for stroke or transient ischaemic attack (TIA) were available in 6% of sites in 1999 and in 2017, 50% had rapid access TIA clinics or admitted all patients with TIA. Conclusion(s): Access to stroke units and other best-practice stroke care has progressed.

Background and Objectives: The T3 cluster randomised trial aimed to improve Triage, Treatment and Transfer (T3) of patients with acute stroke in Emergency Departments (EDs). Method(s): Our prospective, multicentre, parallel group trial with blinded outcome assessment, randomised EDs 1:1 to receive either the T3 intervention or no support. The intervention targeted: (1) Triage: suspected stroke patients to be assigned to patients with suspected acute stroke triaged as Australian Triage Scale 1 or 2; (2) Treatment: screening for tPA eligibility and administration where applicable; management of fever, hyperglycaemia and swallowing; and (3) rapid Transfer to the stroke unit. The implementation strategy comprised: (i) workshops to determine barriers and solutions; (ii) education; (iii) clinical opinion leaders; (iv) reminders. Primary Outcome: 90-day death or dependency (mRS>2). Secondary Outcomes: 90-day: health status (SF-36), functional dependency (Barthel Index); and 11 in-hospital quality-of-care outcomes. Result(s): 26 eligible sites from three states and one territory participated involving 1879 patients (preimplementation n = 574; post-implementation n = 1305 including 751 patients in intervention sites and 554 in control sites). There were no significant differences between groups for primary outcome of 90-day mRS or secondary patient outcomes. There were no significant differences between groups for 11 quality of care outcomes. Conclusion(s): This large trial used an evidence-based implementation strategy previously shown to be highly effective in stroke units but it did not change ED clinician behaviour. ED is a complex, chaotic environment with competing priorities. Alternative ways to support evidence implementation into ED warrant investigation.

Background: Readmissions within 90-days post-stroke appear to be associated with comorbidities or disability-related issues based on the severity of the initial stroke. We describe patient and clinical factors associated with readmission to hospital within 90 days after discharge from stroke in the Australian setting. Method(s): We used person-level linkages between data from the Australian Stroke Clinical Registry (2009-2013), national death registrations and hospital admission data from Victoria, Queensland, New South Wales and Western Australia. Time to first readmission among patients with stroke discharged alive (excluding discharges to inpatient rehabilitation) was examined using competing risks regression accounting for death within 90-days. Covariates included age, ability to walk on admission (a measure of stroke severity), stroke type, stroke unit care, length of stay (LOS) and the Elixhauser Comorbidity Index (derived using ICD-10 coded admission data in the preceding 5-years). Result(s): Among 9,018 patients with stroke discharged (46% female; 79% ischaemic stroke), 20.7% were readmitted within 90-days. In univariable analyses readmission was associated with greater stroke severity, older age, shorter LOS and having more comorbidities, but not stroke unit care. After adjustment, factors independently associated with a greater risk of 90-day readmission were ability to walk on initial stroke presentation (HR 0.41, 95%CI:0.35, 0.48), older age (HR 1.89, 95%CI:1.77,2.02), intracerebral haemorrhage (HR 1.54 95%CI:1.31,1.80) and higher Elixhauser comorbidity scores (HR 1.40,95%CI:1.35,1.46). Conclusion(s): Older patients with higher number of comorbidities who are discharged to the community are at greater risk of early readmission. Alternative care pathways may be needed to better support this group.

Background. Memory rehabilitation can improve functional outcomes. We aimed to assess the potential cost effectiveness of a memory skills group (MSG) intervention compared to standard care for survivors of stroke with memory impairment. Method. The MSG was embedded into two Australian public health services. A decision analytic model was utilised to compare costs and outcomes. Historical control cohorts were used to collect costs of assessment and therapy during rehabilitation in a six week period after the MSG was embedded (intervention) and a six week period prior to the availability of the MSG (control). The primary outcome measure of effectiveness was attainment of memory-specific goals at six-week follow-up. These outcomes were collected prospectively for patients who were provided the MSG intervention. Outcomes of patients receiving standard care were estimated by applying outcomes obtained from an analogous cohort in a separate study. Results. There were 33 patients in the control period and 46 patients in the intervention period (12 provided the MSG intervention and 34 provided standard care). Mean age was 62 years and 61% were male. Average assessment and therapy costs were $388 per person in the control period compared to $282 per person in the intervention period. In the intervention period 39% of patients achieved their memory-specific goals, compared to 27% expected in those in the control period. Conclusions. Improvements expected in the intervention period were delivered at a lower cost compared to the control period. The MSG intervention was found to be potentially cost-effective compared to standard care.

Background: There is increasing interest in the use of electronic support for self-management after stroke, with little evidence on the effectiveness or acceptability. Aim(s): To explore participants perceptions of the acceptability, satisfaction and usefulness of electronic support for self-management after stroke. Method(s): We conducted a pilot randomised clinical trial design with mixed-methods analyses. Participants, recruited from the Australian Stroke Clinical Registry, were randomised to intervention (goal setting and supportive messages) and control (administrative messages). A satisfaction survey was completed at the end of the 4-week trial (all participants). Ten participants from the intervention group were invited to participate in a focus group (FG) interview. Qualitative data were transcribed into NVivo. An inductive approach to thematic analysis was undertaken to code and summarise the data. Result(s): At follow-up, 27/54 satisfaction surveys were completed (14/25 intervention; 13/29 control). Five intervention participants attended the FG (80% male, median age 68, median time since stroke 21 months). The menu used to set goals was considered helpful (85%), and guidance from the clinicians was described as integral in ensuring goals were realistic and individualised. Participants in the intervention group felt the electronic messages were easy to understand (92%) and relevant, but some were ''too general''. Overall, they reported the messages assisted them to achieve their goals, and 85% would recommend the project to others. Conclusion(s): Use of tailored electronic support for ongoing self-management of survivors of stroke was acceptable, and deemed useful in achieving goals. These findings are valuable to inform future trial design.

Background: Variation from evidence-based care contributes to increased deaths and disability after stroke. Aim(s): To determine whether iterative quality improvement interventions (individually and collectively) change clinical practice and improve stroke care. Method(s): Before-and-after design comparing 20 self-selected Queensland hospitals across four time-points: T0 [baseline:03/2004-06/2012]; T1 [financial incentives for stroke unit care: 07/2012-02/2014]; T2 [externally facilitated quality improvement program (Stroke Link) incorporating clinical performance feedback; +/- action plan; ongoing support:03/ 2014-10/2014], T3 [post-measurement:11/2014-12/2015] compared to temporal trends in 23 other Australian hospitals contributing data to the Australian Stroke Clinical Registry. Outcome(s): Primary: net change in composite score (proportion of processes achieved of 8 pre-defined processes) [T0 vs T3]. Secondary: change in individual processes; composite score change comparing hospitals that did vs did not develop action plans in T2. Multivariable analyses adjusted for patient clustering by hospital. Result(s): Patient characteristics were similar across time-points (45-47% female, median age 73-75 years) in 23,800 Queensland patients (11,894 [T0]; 4781 [T3]); and n=20,484 from comparator hospitals (5,903 [T0]; 5,188 [T3]). Primary Outcome: 18% absolute improvement in composite score (coeff 0.18 95% CI 0.15, 0.21) compared to no change in comparator hospitals (95%CI -0.03, 0.03). Overall, 6/8 individual processes improved e.g. stroke unit admission odds ratio [OR] 1.57, intravenous thrombolysis if ischaemic stroke OR 1.41; mobilised during admission OR 2.19. Hospitals developing action plans had greater improvements in composite score (coeff 0.23 95%CI 0.09, 0.37). Conclusion(s): In Queensland hospitals positive changes in clinician practice were achieved from quality improvement interventions implemented during the Stroke123 study. Application in other settings

Background: Prognostic performances of models predicting risk of recurrent events of cardiovascular disease (CVD) are not adequate for use in clinical settings. We aimed to determine whether adapting the Framingham Risk Score (FRS) to an Australian population could effectively predict recurrent cardiovascular outcomes. Method(s): Patients comprised survivors of stroke/TIA who participated in the Shared Team Approach between Nurses and Doctors For Improved Risk factor Management (STAND FIRM) trial (n=563). We used standardised anthropometric, biochemical and blood pressure data, collected at baseline, to evaluate risk factors for stroke/TIA. Cox proportional hazards regression models were used to determine the risk of recurrence of CVD-related events and deaths within 2 years after stroke/ TIA; adjudicated by two independent stroke specialists. Regression estimates were then used to recalibrate the coefficients used by the FRS, and performance of the model assessed. Result(s): In men, the recalibrated FRS model had poor discrimination (C-statistic=0.607) and appeared to better predict CVD recurrence (AUC=0.608) than the original FRS model (AUC=0.515). However in women, the recalibrated FRS model had poor discrimination (Cstatistic =0.626) and prediction of CVD recurrence (AUC=0.571) similar to the original FRS model (AUC=0.617). Conclusion(s): The original FRS and recalibrated FRS models appeared to perform poorly in Australian men and women with stroke. The identification of relevant risk factors, easily measured in a clinical setting, may help clinicians better monitor the risks of their patients and enhance secondary prevention strategies.

Background: Uncertainty exists over attributable causes for the excess mortality after stroke, and of any differences by sex. Method(s): First-ever strokes (2010-2013) admitted to 35 hospitals enrolled in the Australian Stroke Clinical Registry (AuSCR) were linked to national death registrations. For this analysis, one-year causes of death due to stroke, ischemic heart disease (IHD), other vascular conditions (e.g. hypertension), and cancer were investigated. Age and sex specific death rates by calendar year expected in the general population were derived from Australian Government data. Expected numbers of deaths were estimated for each sex by multiplying the age-specific person-years of observation by death rate. Standardised mortality rates (SMR; observed/expected deaths) were estimated for causes of death. Result(s): Among 9,441 first-ever events, 46% were women (median age 78 years) who were 6 years older than men (p<0.001) and had higher all-cause one-year mortality (25.4% vs 19.1%, p<0.001). Compared to population norms, those surviving to 30-days had >8 fold increased risk of death from stroke/recurrent events within 1 year irrespective of sex (SMR: men 8.3 vs. women 8.8). For first-ever strokes, the excess risk of death from other vascular conditions was greater in women than men (SMR 4.0 vs 2.8; p=0.04). Compared to population norms, the excess mortality after stroke was greater for IHD (SMR: men 3.5, women 3.0) and cancer (SMR: men 2.6, women 2.6). Conclusion(s): Excess mortality after stroke is mostly attributed to stroke and recurrent events, other cardiovascular conditions and cancer. Secondary stroke prevention should target vascular risk factors and co-morbidity.

Background: Goal Attainment Scaling (GAS), has support in stroke rehabilitation literature as a measure of goal achievement. However, reliability depends on the quality of how goals are written, scoring procedures, and clinician experience and training. Aim(s): To determine the inter-rater agreement for goal setting and GAS using a newly developed goal setting package for stroke. Method(s): Clinicians (N=20) responsible for discharge planning from four stroke units (acute and subacute) have been recruited. All were trained in using the goal setting discharge package which consisted of a 36-item goal menu; manual; templates; and GAS scoring guides. Patients undergo goal setting twice, once with a clinician/participant and once with a research team member. Rater assessment order is randomised. Agreement on the type, and quality (0-6 scale) of goals is scored by a blinded assessor. GAS content is scored according to the number of categories that meet the GAS scoring guide criteria (0-4 scale). Result(s): Data collection concludes in April 2018. Currently 29 goals (mean 3.6 per assessment) from 4 rating pairs have been assessed. The percentage of similar goals for rating pairs (same menu item) so far is 72%. At completion findings with respect to percentage of similar goals and GAS scores will be reported as will intra class correlations between quality scores and GAS content scores for inter-rater reliability. Conclusion(s): It is hoped the findings will support the use of the goal setting package and associated training to provide a reliable and standardised approach for goal setting as an outcome tool.

Introduction: Women appear to receive evidence-based care less often than men but the reasons why are uncertain. We examined sex differences in management of stroke and factors might account for these disparities. Method(s): INSTRUCT is an individual participant data pooling study of incident strokes (1987-2013) from 13 population-based cohorts worldwide. In hospitalised people, log-binomial regression was used to estimate the prevalence ratio (PR) for women compared to men for thrombolysis, admission or discharge medications (antihypertensives, antiplatelets, anticoagulants), investigations (e.g. echocardiography) and surgical interventions (e.g. carotid endarterectomy). Study-specific multivariable models incorporating adjustment for socio-demographics, stroke-related factors and pre-stroke health were combined using random-effects meta-analysis. Result(s): Data were available for 16,054 (52.4% women) people admitted to hospital. Women, compared to men, were older (10/13 studies) and suffered more severe strokes (7/13 studies). Women were slightly more likely to be taking an antihypertensive on admission (3 studies, pooled PRcrude 1.08, 95% CI 1.01-1.16) and less likely to receive echocardiography, (3 studies, pooled PRcrude 0.90, 95% CI 0.82-0.98) or carotid investigations (8 studies, pooled PRcrude 0.89 95% CI 0.81-0.97) inhospital. Adjusting for age and stroke severity attenuated these difference making them non-significant: antihypertensives, pooled PRadjusted, 1.00, 95% CI 0.94-1.07; echocardiography, pooled PRadjusted 1.00 95% CI 0.90-1.12 and carotid investigation, pooled PRadjusted 0.95 95% CI 0.91-1.00. Conclusion(s): Women and men with stroke received equal levels of care. Sex differences were associated with age and stroke severity highlighting the importance of improving care in elderly people and managing modifiable factors for stroke severity.

Background and aims: Memory deficits are common following stroke and are associated with poorer clinical outcomes. Two approaches to rehabilitate memory have some support but have not been directly compared: compensatory memory skills group training (MSG) and computerised cognitive training (CCT). We aimed to compare the effectiveness of these approaches following stroke. Method(s): Participants were randomised into one of three conditions following blinded baseline assessment: CCT (LumosityTM), MSG or waitlist control (WC). Interventions lasted six weeks. Blinded assessments were repeated immediately post-intervention and at a six-week follow up. The primary outcome measure was attainment of personalised memory goals using Goal Attainment Scaling. Secondary outcome measures included neuropsychological measures of memory, subjective ratings of everyday forgetfulness, and self-reported strategy use. Random-effects regression was used to model outcomes. Result(s): Sixty-five survivors of stroke with subjective memory complaints (median age 63 years, 56% male) were randomised: CCT (n=22), MSG (n=24), and WC (n=19). MSG participants showed significantly greater attainment of memory goals relative to CCT and WC following intervention and at follow-up (d=1.54; 95% CI [11.57 - 25.11]). While all participants reported a significant increase in internal strategy use, this improvement was only maintained at follow-up for MSG participants (d=1.27; 95% CI [0.05 - 7.78]). There was no significant change in objective nor subjective secondary outcomes at follow up. Conclusion(s): Current findings support the use of MSG but not CCT training to improve everyday memory functioning for stroke survivors. These Phase II data will be used to inform a wider Phase III investigation.

Background and Aims: Rapid assessment, diagnosis and treatment for patients with acute stroke events requires communication between infield paramedics and multiple hospital clinicians (emergency, medical, neurology, radiology). Treatment delays may occur with information repeated between the interdisciplinary team from different organisations/hospital departments. Aim(s): To determine if a smartphone communication app can improve clinical care timelines for patients with suspected acute stroke. Method(s): Using a 12 month pre-post historical-control design, the PulsaraTM Stop Stroke/STEMI smartphone and tablet app (PulsaraTM) was implemented pre- (25 Ambulance Victoria branches) and within-hospital (2 hospitals in regional Victoria, Australia). PulsaraTM provides secure, simultaneous, two-way, real-time communication. Eligible patients had suspected acute stroke events assessed by paramedics or hospital clinicians. Pre-hospital and hospital assessment and treatment times were captured; data collection is ongoing. Preliminary results compare stroke care timelines if PulsaraTM was initiated (PulsaraTM) or not (no PulsaraTM) (hospital 1: 22/08/2016-21/02/2017; hospitals 1 and 2: 29/05/2017-31/ 10/2017). Result(s): There were 262 patients (no PulsaraTM/PulsaraTM: n=81/ n=185, both median 75 years, 52%/48% male) with PulsaraTM initiated by paramedics (n=115) and Emergency Departments (n=70). Faster paramedic hospital-arrival-to-departure by median 10 minutes (no PulsaraTM: 55 minutes [IQR: 43-64]; PulsaraTM 45 minutes [IQR: 35- 55]). Faster emergency department door-to-first medical review time by median 18 minutes (no PulsaraTM: 23 minutes [IQR: 6-65]; PulsaraTM 5 minutes [IQR: 1-11]) and faster door-to-CT completed by median 51 minutes (no PulsaraTM: 81 minutes [IQR: 39-145]; PulsaraTM: 27 minutes [IQR: 17-42]). All thrombolysis cases used PulsaraTM (n=38). Conclusion(s): Preliminary evidence supports improved care timelines. PulsaraTM was successfully implemented for

Background and Aims: Treatment fidelity is a complex evaluative process critical for achieving quality in intervention studies. Its objective is to determine whether an intervention was delivered as intended. Despite its benefits, treatment fidelity has been inconsistently reported in behavioural intervention studies with only 14% of aphasia treatment studies explicitly reporting on treatment fidelity. Complex interventions, such as those for post-stroke aphasia, pose particular challenges to fidelity monitoring. Method(s): The COMPARE study is an Australian-based 3-arm randomised controlled trial (planned sample size n=216) which aims to determine whether two intensive and contrasting treatments (constraint induced aphasia therapy; multi-modality aphasia therapy) for chronic post-stroke aphasia are superior to usual care. Based on the theoretical models underlying the trial interventions, we developed a fidelity protocol to monitor and enhance treatment delivered by multiple clinicians across Australia. The fidelity protocol adheres to the recommended standards outlined in the TiDIER statement (a template for reporting interventions stemming from the CONSORT 2010 and SPIRIT 2013 statements). Result(s): An overview of the protocol will be presented, including the theoretical rationale, methods, challenges and results to date. Data will be presented on the assessment measures, which include (a) standardised clinician training, (b) video recording all assessment and treatment sessions, (c) providing feedback in real-time to clinicians, and (d) monitoring for therapist drift. Conclusion(s): The COMPARE fidelity protocol offers one solution to the challenges of monitoring and enhancing treatment fidelity within behavioural change interventions. Our solution may be of utility to a wide range of trials of complex interventions.

Background: Stroke is a medical emergency. Patients must arrive and commence treatment within hours of an ischaemic stroke for hyperacute therapy to be of most benefit: intravenous thrombolysis and/or clot endovascular clot retrieval [ECR] thrombectomy. Aim(s): To assess whether there is evidence of improvement in access to time critical treatments from 2015 (when ECR became mainstream) to 2017 in Australian acute care hospitals. Method(s): Cross-sectional patient-level audit data and organisational survey data from the National Stroke Audit Acute Services 2015 and 2017. Descriptive statistics and random effects logistic regression was used. Each hyperacute treatment was the dependent variable and the independent variable was the audit period with adjustment for patient clustering by hospital. Result(s): The organisational survey was completed by 127 hospitals in 2017 and 185 hospitals in 2015. Hospitals reporting onsite ECR services increased from 18 to 21 with 872 patients reported to have undergone ECR during 2016. 8279 audits reviewed, 51% from 2017 (mean age 76 years; 55% male). In 2017 compared to 2015, the odds of patients receiving intravenous thrombolysis improved (aOR: 1.80 95% CI 1.49, 2.17), particularly if they arrived within 4.5 hours of symptom onset (aOR 2.32 95% CI 1.79, 2.99). The time from onset to thrombolysis reduced by a median 18 minutes between 2015 and 2017 (p=0.04). No change in the proportion of patients receiving thrombolysis within 60 minutes of arrival was found (2017: 30%;2015: 26%,p=0.3). Conclusion(s): Greater access to thrombolysis was found. Improvements in access within 60 minutes of hospital arrival is needed.

Background: The comparison of outcomes between organisations treating stroke has potentially major practice and policy implications. The American Heart and Stroke Associations (AHA/ASA) published a joint statement (Stroke, 2014; 45:918-944), which provided an overview of statistical considerations for evaluation of hospital-level outcomes after stroke. It was recommended that multilevel random effect regression models that use random intercept terms to describe hospital-specific effects would overcome the failure to account for variation in case numbers across hospitals, and intra-hospital clustering effects. We introduce an alternative novel cross-evaluation approach to comparing hospital performance on stroke outcomes and compare this approach to the one recommended by AHA/ASA. Method(s): The novel cross-evaluation approach relies on generating a set of counterfactual outcomes by applying individual hospital-specific regression models to other hospitals' casemix data. For a given hospital, a ranking of all the participating hospitals in terms of their hypothetical performance on this hospital's casemix data is produced. Finally, a decision-analytic procedure that aggregates all individual hospital's rankings into a total group ranking is applied to provide an overall cross-evaluation rank for every included hospital. Result(s): The cross-evaluation approach is illustrated using data from 24 hospitals participating in the Australian Stroke Clinical Registry (2009-2013). Variables used: age at stroke, sex, type of stroke, ability to walk on admission and mortality status. There were 13,477 patients included (median age 77 years; 53% male; 63% ischaemic). Conclusion(s): We propose a novel method for comparing hospital performance on stroke outcomes using mortality as an illustration. Future work will use this approach for other stroke outcomes.

Background and Aims: All-cause mortality after stroke is greater in women than men. Our aim was to examine whether there are sex differences in the causes of death (COD) or the potential contributing factors. Method(s): 1-year COD were obtained by linking first-ever strokes (2010- 2013) admitted to 35 hospitals participating in the Australian Stroke Clinical Registry to national death registrations. COD were categorised as stroke, ischemic heart disease (IHD), other cardiovascular disease (CVD, e.g. hypertension), cancer, and other causes. Competing risk models were used to estimate female:male specific hazard ratios (sHRs) of death with adjustment for factors that differed between men and women (sociodemographics, stroke severity assessed by ability to walk on admission). Result(s): Among 9,441 events (46% women), women were 7 years older (median age 78 vs 72, p<0.001) than men, had more severe stroke (32% vs 41% able to walk on admission, p<0.001), received similar acute care therapies, and had greater all-cause mortality (25.4% vs 19.1% p<0.001). Women had greater risk of death due to stroke (sHRunadjusted 1.65, 95% CI 1.42-1.91) and other CVD (sHRunadjusted 1.65, 95% CI 1.29-2.12), but these differences were substantially attenuated after adjusting for age, stroke severity (stroke: sHRadjusted 1.19, 95% CI 1.02-1.40; other CVD: sHRadjusted 1.12, 95% CI 0.85-1.48). In contrast, women less often died from IHD (sHRadjusted 0.58, 95% CI 0.43+/-0.77), or from cancer (sHRadjusted 0.67, 95% CI 0.49+/-0.91). Conclusion(s): COD after first stroke differ between genders with women having more deaths attributed to stroke or other CVD, reinforcing the need to manage cardiovascular risk factors.