Improving the Quality of Stroke Care Tony Rudd

“it is the duty of the physician to explain to the patient, or to his friends, that the condition is past relief, that medicines and electricity will do no good, and that there is no possible hope of cure” William Osler

St Thomas’ Hospital

St Thomas’ Hospital Stroke Care in 1988  Patients admitted under care of any of 17 general physicians to any one of 15 wards  Very little happened acutely  Brain scans difficult to obtain and therefore rarely done  Referred to geriatricians for rehabilitation – long wait  No stroke specialist service either in hospital or community

(Adjusted incremental costs (US$/PPP) and hazard ratios (with 95% CI) by centre (Riga (Latvia) as reference) The EC BIOMED 1 Stroke Project) The EC BIOMED 1 & 2 Stroke Project Grieve R et al. Stroke 2001

What are the problems that still need solving?  Unacceptable variation in the quality of care between hospitals  Variable quality of clinical and managerial leadership  Variable resources provided for stroke care  UK slow to adopt new technologies e.g. Thrombolysis  Still a belief amongst many people (professionals and public) that stroke untreatable disease of old people

Early Stroke Audit Results (1998/9)  18% of patients through stroke unit  23% cognitive assessment  44% visual fields recorded  55% rehabilitation goals set  41% G.P. contacted within 3 days of discharge

Key Forces for Change  Research  Guidelines summarising the research evidence and clinical consensus (RCP Guidelines and NICE Guidelines on Acute Stroke and TIA and NICE Technology Appraisal on alteplase)  National Audit  Sentinel audit  SINAP  Carotid endarterectomy audit  National Audit Office report  National Stroke Strategy

Key Forces for Change  NICE Quality Standards  Commissioning Outcomes Framework (COF) standards  Performance standards set by Care Quality Commission  Stroke Improvement Programme and Stroke and Cardiac Networks  Stroke Research Network  Public opinion  Media reports  Voluntary sector campaigning

Stroke Programme at the RCP  Guidelines  NICE Guidelines on Acute care and TIA  Intercollegiate Guidelines on the rest  4th edition to be published Sept 2012  National Audit  Sentinel audit ( )  Carotid interventions audit  Acute Continuous Stroke Audit (SINAP)  SSNAP  Change management  Presentations  Workshops  Peer Review  Politics/lobbying  Stroke Improvement Programme links

History of Stroke Audit in the UK  1997 Department of Health commissioned national stroke audit  Intercollegiate stroke working party  Audits conducted every 2 years  Structure  Process  (Outcome)  Patient experience (Picker survey)  Primary care audit  Now funded by HQIP

Features of Audit 1 100% participation Run by clinicians Exceptional level of data quality and completeness Detailed analysis centrally to allow tailored interrogation of data Performed every 2 years allowing time for implementation of change Rapid production of results

Individual detailed hospital reports with results benchmarked against national/regional averages Reports to Strategic Health Authorities, Healthcare Commission, Networks, Department of Health and Parliament Extensive media coverage because public data of key indicators Features of Audit 2

Other sources of data  Primary care – Quality Outcomes Framework (QOF)  Vital signs data  Accelerated metrics for SIP  Routine Hospital Statistics (HES). Used by Dr Foster

Results: Stroke unit provision – comparison over time Stroke unit in hospital73%79%91%92% Median (IQR) stroke beds 20 (14-27)20 (15-29)24 (16-30)25 (20-34) Specialist community/ domiciliary rehabilitation team 31%27%32%70%

Sentinel Stroke Audit RCP London

Nationally 1Patients treated for 90% of stay in a Stroke Unit Screened for swallowing disorders within first 24 hours of admission Brain scan within 24 hours of stroke70.5 4Commenced aspirin by 48 hours after stroke Physiotherapy assessment within first 72 hours of admission Assessment by an Occupational Therapist within 4 working days of admission Weighed at least once during admission89.2 8Mood assessed by discharge Rehabilitation goals agreed by the multi-disciplinary team by discharge 97.3 Average for 9 indicators for Key Process Indicators

Number of 9 Key Indicators Achieved Only 32% of patients who were eligible for all 9 indicators received all 9.

12 Key Process Indicators (2010) This round we have added four additional indicators and removed one (rehab goals agreed by discharge)

Number of 12 Key Indicators Achieved Only 16% of patients who were eligible for all 12 indicators received all nine.

How are the data used to influence change?  Workshops  Slide toolkits  Publicity  “I’ve been trying to get the trust to offer scanning for stroke patients for 5 years, within a day of receiving the audit report the chief executive had convened a meeting with stroke service and radiology” A stroke physician after publication of performance indicators 2004 audit  Influencing policy at a national level  Influencing policy at SHA level

Transforming Stroke care in London:Case for change 90 Target Below Target Above Target London Stroke Providers against Sentinel Audit 12 key indicators 2006 Change in London Stroke Providers against Sentinel Audit 12 key indicators 2006 vs 2004 scores London Stroke Units Sentinel Audit Comparison 2004 and 2006

The scale of the problem of stroke in London Second biggest killer and most common cause of disability Population >8 million 11,500 strokes a year in London – 2,000 deaths

30-minute blue light ambulance travel time from the hyper-acute stroke units The green area shows the areas that are within 30 minutes travel time (under ambulance blue light conditions) of a proposed HASU

London Stroke Strategy  Additional £20m per year for stroke care but only paid if hospitals delivering the required quality  Centralise hyperacute (hyperacute stroke units HASU) care into 8 units situated to provide easy access to the whole population (no more than 30 minutes by ambulance)  All acute stroke patients admitted to HASU. This is not just a thrombolysis service  Further 20 stroke units for on going rehabilitation  Improve community care and longer term rehabilitation  Neurovascular services for patients with TIA

London SHA Stroke Strategy  Bidding process to provide care  London Clinical Director  Regular inspections to ensure quality of care maintained  Obliged to submit continuous audit

Prophets of Doom Predictions  Not possible to implement major system reorganisation in London for a condition as complex as stroke  Staffing requirements unachievable (400 nurses and 100 therapists) – Recruitment – where will staff come from? – Training – how will staff develop the necessary skills? – Leadership – who can provide the necessary leadership? – There is a risk that the available workforce will be consumed by early implementers, leaving later implementers unable to recruit to posts.

Prophets of doom predictions  Patients will not accept being taken to a hospital that is not local to them  Not possible to transport people within 30 minutes to a HASU  Repatriation will fail and HASUs will quickly become full  Even if get acute services working it will fail because impossible to change community services  Unsustainable

1 year outcomes % of patients spending 90% of their time on a dedicated SU

1 year outcomes Average length of stay

1 year outcomes 3.5% 10% 12% Feb-July 2009AimFeb-July 2010 Thrombolysis rates 14% Jan-March 2011

London Stroke Survival vs Rest of England Hazard ratio for survival in London %CI p<0.001

Cost-effectiveness analysis of the London Stroke Service: Results based on 6438 strokes per annum Differences inUnadjustedAdjusted Differences in total costs at 30 days3,307,6773,763,472 Differences in total deaths at 30 days Differences in total QALYs at 30 days5144 Incremental cost per death averted at 30 days15,45155,371 Incremental cost per QALY gained at 30 days64,47886,106 Differences in total costs at 90 days-5,393,533-3,544,210 Differences in total deaths at 90 days Differences in total QALYs at 90 days11286 Incremental cost per death averted at 90 daysDominant Incremental cost per QALY gained at 90 daysDominant Differences in total costs at 10 years-21,318,180-22,786,954 Differences in total QALYs at 10 years4,4923,886 Incremental cost per QALY gained at 10 yearsDominant Professor Steve Morris et al

Cost-effectiveness analysis of the London Stroke Service: Results based on 6438 strokes per annum Differences inUnadjustedAdjusted Differences in total costs at 30 days3,307,6773,763,472 Differences in total deaths at 30 days Differences in total QALYs at 30 days5144 Incremental cost per death averted at 30 days15,45155,371 Incremental cost per QALY gained at 30 days64,47886,106 Differences in total costs at 90 days-5,393,533-3,544,210 Differences in total deaths at 90 days Differences in total QALYs at 90 days11286 Incremental cost per death averted at 90 daysDominant Incremental cost per QALY gained at 90 daysDominant Differences in total costs at 10 years-21,318,180-22,786,954 Differences in total QALYs at 90 days4,4923,886 Incremental cost per QALY gained at 10 yearsDominant

Cost-effectiveness analysis of the London Stroke Service: Results based on 6438 strokes per annum Differences inUnadjustedAdjusted Differences in total costs at 30 days3,307,6773,763,472 Differences in total deaths at 30 days Differences in total QALYs at 30 days5144 Incremental cost per death averted at 30 days15,45155,371 Incremental cost per QALY gained at 30 days64,47886,106 Differences in total costs at 90 days-5,393,533-3,544,210 Differences in total deaths at 90 days Differences in total QALYs at 90 days11286 Incremental cost per death averted at 90 daysDominant Incremental cost per QALY gained at 90 daysDominant Differences in total costs at 10 years-21,318,180-22,786,954 Differences in total QALYs at 90 days4,4923,886 Incremental cost per QALY gained at 10 yearsDominant

Sensitivity analysis Results were qualitatively unchanged after undertaking sensitivity analysis on the following: Stroke mimics LOS in the HASU Unit cost per day in the HASU LOS in ICU Neurosurgery rates Discharge destinations

Effects (e.g., deaths, QALYs) Costs (£) Cost Before Outcome Before Before Quadrant 1 Quadrant 2Quadrant 3 Quadrant 4 Cost-effectiveness plane Better outcomes Worse outcomes Higher costs Lower costs

Effects (e.g., deaths, QALYs) Costs (£) Cost Before Outcome Before Before Quadrant 1 Quadrant 2 Quadrant 3 Quadrant 4 Cost-effectiveness plane Better outcomes Worse outcomes Higher costs Lower costs

Areas where work still needed  Early supported discharge  Bed based intermediate care  Longer term rehabilitation  Vocational rehabilitation

3 Click on the relevant number on the map or below to go to your region’s results 11 North EastNorth East 22 North WestNorth West 3 Yorkshire and the Humber 44 West MidlandsWest Midlands 55 East MidlandsEast Midlands 66 East of EnglandEast of England 77 South WestSouth West 88 South CentralSouth Central 99 LondonLondon 1010 South East CoastSouth East Coast SINAP

Hospital participation and quartiles for participants

Thrombolysis Out of stroke patients Total number of patients given thrombolysis2541 Median (and IQR) number of patients per hospital given thrombolysis 12 (2-35) Percentage of all patients given thrombolysis8% Median (and IQR) percentage of patients per hospital given thrombolysis 6% (2-10%) Median (and IQR) age of thrombolysed patients72 (63-80) Number of thrombolysed patients aged 81 or over561 (22%) Median (and IQR) time from door to needle (minutes)60 (41-85) Median (and IQR) time from scan to thrombolysis (minutes)32 (20-50)

The effects of getting to a SU quickly % Compliance with KIs (n=30351) (does not include patients already in hospital at time of stroke) SU within 4 hours 53% (15946) SU within 24 hours 80% (24236) SU within 72 hours 89% (27108) Did not go to SU 6% (1822) KI 1 Scanned within 1 hour of arrival at hospital KI 4 Stroke consultant 24h KI 6 Prognosis/diagnosis discussed with relative/carer within 72h where applicable KI 7 Continence plan drawn up within 72h where applicable KI 8 Percentage of potentially eligible patients thrombolysed KI 9 Bundle 1: Seen by nurse and one therapist within 24h and all relevant therapists within 72h KI 10 Bundle 2: Nutrition screening and formal swallow assessment within 72 hours where appropriate KI 12 Bundle 4: Patient given antiplatelet within 72h where appropriate and had adequate fluid and nutrition in all 24h periods* Average Average of 12 KIs For every Key Indicator, the results are worse at each stage, showing that getting to a stroke unit quickly impacts on a range of process measures

Equity of care across hospitals Compliance with KIs (median and interquartile range) 25 th percentileMedian75 th percentile KI 1Percentage of patients scanned within 1 hour of arrival at hospital KI 3Percentage of patients who arrived on stroke bed within 4 hours of hospital arrival (when hospital arrival was out of hours) KI 4Percentage of patients seen by stroke consultant or associate specialist within 24h KI 7Percentage of patients who had continence plan drawn up within 72h where applicable KI 8Percentage of potentially eligible patients thrombolysed KI 9Bundle 1: Seen by nurse and one therapist within 24h and all relevant therapists within 72h (proxy for NICE QS 5) KI 11Bundle 3: Patient's first ward of admission was stroke unit and they arrived there within four hours of hospital arrival KI 12* Bundle 4: Patient given antiplatelet within 72h where appropriate and had adequate fluid and nutrition in all 24h periods* Average of 12 KIs There is huge variation between hospitals, particularly for continence, thrombolysis and the ‘nursing/therapy bundle’

Normal HoursOut of HoursInpatient N of patients in “arrival category” (46%)15719 (49%)1762 (5%) Arrival to stroke team (median and IQR) (mins) 87 (9-315)188 (15-870)1142 ( ) Arrival to stroke bed (median and IQR) (mins) 211 ( )234 ( )1785 ( ) Arrival to scan (median and IQR) (mins) 120 (45-301)170 (50-885)801 ( ) Equity of care based on when the patient arrives at hospital

The New World of SSNAP  A wish by certain people to have prospective data collection for stroke  The ‘need’ to collect outcome data

The New World of SSNAP  A wish by certain people to have prospective data collection for stroke  The ‘need’ to collect outcome data  Need for information about the whole pathway  Need for PROMS and PREMS

Data requirements  Accelerated metrics  NICE Quality Standards  Vital signs  Local stroke and cardiac network requirements  Commissioning Outcomes Framework  Quality Outcomes Framework  National audits  HES Data  CQC

Sentinel Stroke National Audit Programme (SSNAP)  Replacing all other statutory data collection (except vital signs!). Includes data needed for:  NICE QS  NHS Outcomes Framework  Accelerated metrics  COF  Funded by HQIP

Development of SSNAP  Intercollegiate Stroke Working party  Strategic data and audit group (SIP, RCP, IC, etc)

SSNAP  Prospective data collection for all stroke admissions  Web tool for direct data entry  Good data validation systems  Facility for instant local downloads  Uploading facility from other data sets  Quarterly national reporting with benchmarking against national data  Annual public reports  6 month follow-up data entry  Linkage to ONS for mortality data  HES linkage  Option for user defined fields

SSNAP: Structure  Core data set for all patients  Comprehensive dataset (optional items)  Spotlight audits  Audits on areas not covered by the core dataset. E.g. TIA, community rehabilitation  Sprint audits  short specific audits focussing on specific areas of the pathway that are of concern e.g. Therapy intensity, intermediate care.  Organisational audit  Hospital  Community  PROMS  PREMS

SSNAP Reporting  Ability to download own data anytime  3 monthly reports benchmarked against national data  Annual public reports – ‘state of the nation’  Outcomes required by DH  Mortality at 30 days and 6 months  Modified Rankin Score at 6 months  Institutionalisation rate at 6 months

SSNAP Timetable  Some uncertainty  We hope  May 2012 Organisational audit of hospital care  August 2012 Clinical data set starts  SINAP continues until SSNAP starts  1 st Spotlight and Sprints audits in year 2  Initial funding 3 years

SSNAP Team  Intercollegiate Stroke Working Party overseeing the process  Clinicians at RCP in Associate Director Roles  Geoff Cloud, Pippa Tyrrell, Martin James, Tony Rudd  Alex Hoffman, James Campbell, Sara Kavanagh plus a statistician, web developer and admin support

SSNAP Risks  Funding  Can we agree a contract?  Currently debate/dispute over intellectual property  Participation rates  Major burden for clinicians/trusts  Freedom of information act  Technical challenges  DH want us to change our name!!

Conclusions  Stroke care has transformed over the last 20 years  Audit has been one of the factors that has driven improvements  No prospect of avoiding monitoring of quality of care that we provide  We are starting a new era of prospective audit  Huge benefits for all if everyone participates

The Face of the Future of Stroke

Acknowledgements  Alex Hoffman and whole team at RCP  James Campbell, Sarah Kavanagh, Sarah Martin, and others  Martin James, Pippa Tyrrell, Geoff Cloud  ICSWP