NHS Safety Thermometer

NHS Safety Thermometer National Data Report 2012-14 Falls VTE Catheters and UTIs Pressure Ulcers A new mindset in patient safety improvement

Contents Authors & Acknowledgments... 2 Preface... 3 It s not just counting...it s caring!... 4 Background... 6 Key Messages... 7 Impact... 8 Headline Results... 8 NHS Safety Thermometer at a glance... 9 Data Summary... 10 Run Charts Of Data Over Time...13 Interpreting Charts...13 Pressure Ulcers... 14 Falls... 18 Catheters and UTI... 21 VTE... 26 Harm Free Care... 30 Cautions With Data Interpretation... 32 Operational Definitions... 33 Harm Free Care in Quality Accounts... 34 Resources... 36 Participation Tables... 38 1

Authors Dr Mike Durkin Director of Patient Safety, NHS England Professor Maxine Power NHS Safety Thermometer National Lead and Director of Innovation and Improvement Science, Salford Royal NHS Foundation Trust Dr Matt Fogarty Head of Patient Safety Policy and Strategy, NHS England Abigail Harrison Programme Manager, Haelo Kate Cheema Head of Service and Specialist Information Analyst, Quality Observatory John Madsen Head of Productivity & Efficiency Data and Information Services, HSCIC Kurt Bramfitt Project Manager, Haelo Acknowledgments We would like to thank all those who have contributed to this report. Professor Roopen Arya Clinical Lead, National VTE Prevention Programme, Chair of VTE Prevention Board, NHS England Karen Conway Patient Safety Lead (Mental Health), NHS England (North) Debby Gould Programme Manager, NHS England (North) Karen Handscomb Quality & Patient Safety Manager Herts & South Midlands Area Team Dr Frances Healey Senior Head of Patient Safety Intelligence, Research and Evaluation, NHS England Yvonne Higgins Quality & Safety Lead at Birmingham, Solihull and the Black Country Area Team Karen Sobey Hudson Patient Safety Projects Manager, NHS England (London) Sally Kingsland Lead Nurse, Infection Prevention and Control, NHS England (London) Caroline Lecko Patient Safety Lead, NHS England Helen Morrison Manager, National VTE Prevention Programme, NHS England Lloyd Provost Improvement Advisor & Statistician, Associates in Process Improvement David Shackley Consultant Urologist, Salford Royal Hospital NHS Foundation Trust Siobhan Teasdale Quality and Assurance Manager, NHS England (South) Vikki Tweddle Quality & Safety Lead at Arden, Herefordshire and Worcestershire Area Team Professor Charles Vincent Professor of Psychology, University of Oxford Julie Windsor Patient Safety Lead (Older Persons and Falls), NHS England 2

Preface Over four million patients have been surveyed using the NHS Safety Thermometer, a measurement system developed by the NHS for the NHS. This development has been an unprecedented, courageous and positive step for England and has given us new insights into harm. For example, we now know how many of our patients receive harm free care, as defined by the absence of the four harms the Safety Thermometer measures, and we are seeing steady improvements. We are delighted to share the key messages from data analysis of over one thousand organisations. Individual case studies can be found on the harm free care website www.harmfreecare.org and raw data are publically available on the Health and Social care Information centre website www.hscic.gov. uk and the Safety Thermometer Webtool www. safetythermometer.nhs.uk. In the last two years the NHS Safety Thermometer data has come of age and is now a national statistic. This status, combined with open access to the public, patients and providers marks a significant step in our commitment to transparency and nothing about me, without me. In the words of Peter Drucker if you can t measure it, you can t improve it. As a nation we are now uniquely positioned to improve care. Our commitment to measurement of these four harms in all NHS settings means that we now have data from patient s homes, community settings and nursing homes as well as the more traditional hospital settings. Moreover, the focus on frontline teams as data stewards has continued to grow ownership and spawned spontaneous improvements, creating increased situational awareness of the risk of harm and compelling clinicians to take action. Over the last three years we have seen steady incremental reductions in each of the four harms (pressure ulcers, falls, urine infection in patients with catheters and venous thromboembolism). Whilst this is a positive change, too many patients are still harmed in NHS care and there is much more work Dr Mike Durkin Prof Maxine Power to be done. This year we will build on our success. A national CQUIN incentivises reduction in harm, measured by the NHS Safety Thermometer according to a locally determined improvement goal. We have proposed that organisations choose to focus on pressure ulcer prevalence and suggested they aim to achieve a 50% reduction by March 2015. This improvement alone, if delivered, could protect 41,250 from this harm and deliver considerable savings given that each new pressure ulcer is estimated to cost an additional 4,500 in care costs alone. The case is compelling and we know what to do. We must act swiftly. In addition to the CQUIN, the NHS has a series of important policy developments and investments to support safety improvement at scale. The patient safety collaborative programme, being developed as a partnership between NHS England and NHS Improving Quality implemented across each Academic Health Science Network area, will bring together communities of interest, creating an engine room for change. A safety fellows programme will help recruit a new generation of safety leaders who will ignite the change process locally. A national campaign, Sign up to Safety will recruit NHS organisations and individuals committed to improving care. We believe the NHS Safety Thermometer will be an essential component of their toolkit in the war on harm, including the Next Generation of Safety Thermometers which are being developed for Medications, Maternity, Mental Health and Children and Young People. The NHS Safety Thermometer data collection and the improvements we have seen have only been possible thanks to the leadership, commitment and hard work of NHS staff. Our heartfelt thanks go out to each and every one of you. Dr Mike Durkin Director of Patient Safety, NHS England Prof Maxine Power NHS Safety Thermometer National Lead 3

It s not just counting...it s caring! Here are just a few of those who are committed to improving patient safety and have been involved in the NHS Safety Thermometer classic and next generation. Royal Devon and Exeter NHS Foundation Trust representing Harm free Care with their #HFCselfie #research Sharon Bennett testing the Medications Safety Thermometer at Central Manchester University Hospitals NHS Foundation Trust Soline Jerram from Brighton & Hove CCG make care safer for all Burton Hospitals NHS Foundation Trust are caring not counting Emma and Sue at Royal Bolton Hospitals NHS Foundation Trust Connie Sharrock on a Safety Thermometer webex from Wrightington, Wigan and Leigh NHS Foundation Trust Andrea and Wendy from Sussex Partnership Trust testing the Mental Health Safety Thermometer Steve Williams first #HFCselfie from University Hospitals South Manchester NHS Foundation Trust Hayley Peters #inspirational nurse leaders making a difference #proud at Taunton and Somerset NHS Foundation Trust District nurses from Tameside and Glossop Community Healthcare Business Group - trying out the medications Safety Thermometer Berni George great exec support for matrons with #safetythermometer today @ RDE the matrons value it at Royal Devon and Exeter NHS Foundation Trust Pauline Gilroy and colleagues at a Safety Thermometer workshop in Leicester #HFCSelfie 4

Medicines Safety Pharmacist, Helen at Tameside Hospitals NHS Foundation Trust Andrew Alldred and colleagues #HFCSelfie make medicines safer Great work today from the Matrons Getting in the harm free care spirit at the Safety Thermometer workshop Debra Vidler and colleagues #HFCSelfie Kent and Medway working together for harm free care Saba Rifat, pharmacist out testing the medications Safety Thermometer Trio from Royal Berkshire NHS Foundation Trust enjoying the Safety Thermometer event Justine Heywood from University Hospital Southampton NHS Foundation Trust is passionate about creating a safe and just culture Kate Mellor and co from Berkshire Healthcare NHS Foundation Trust learning more about harm free care Two orthopaedic matrons v. proud with their tests today @RDEhospital Lee from Salford Royal NHS Foundation Trust testing out the ipad for data collection Great day with Sheffield Teaching Hospitals NHS Foundation Trust safe care group talking about using data for improvement 5

Background Why was the NHS Safety Thermometer developed? In 2010 the QIPP safe care national programme sought to improve quality whilst reducing cost. The focus, on reducing harm from pressure ulcers, falls, urinary infections (in patients with a catheter) and venous thromboembolism, necessitated real time data. Whilst we had some data in our administrative and incident reporting systems, issues with coding and variability in reporting made the data unreliable and difficult to interpret. Quite simply, we did not have a system that could be used to measure improvement over time. The NHS Safety Thermometer (NHS ST) was developed, in collaboration with over 160 NHS organisations, to support the QIPP safe care programme, the Safety Express Collaborative and the harmfreecare movement. How is the NHS Safety Thermometer helping us understand harm from the patient s perspective? The NHS ST is innovative in its approach. Data are collected wherever the patient is being cared for including acute and community hospitals, nursing homes and patient s own homes. This is the first time we have had significant data on harm outside of acute care. Community patients now make up over half of the data collected. Data are also collected at the patient level so for the first time we are able to know whether a patient has been affected by more than one harm and the proportion of patients who receive harm free care (the absence of all four harms). How does the NHS Safety Thermometer contribute to our collective goal to measure outcomes as well as process? The NHS ST has built on national programmes, such as the VTE risk assessment programme, and taken the next step in incentivising the measurement of outcomes. This aligns with our current policy focus and enables us to measure the impact of our improvement work as well as the effectiveness of our systems of care. We have seen data from the NHS ST used in board reports, quality accounts page 34, the Keogh review, CQC inspections and the National Trust Development Agency (NTDA) pipeline. CCGs are using these data to agree local improvement goals, NHS England Area Teams are using it to measure harm across health economies, providers are using it to work across organisational boundaries to understand and reduce harm across pathways and frontline teams are using the data to measure the impact of their improvement work. How reliable are the NHS Safety Thermometer data? We have taken a systematic approach to the development of clear operational definitions for each measure 1. Resources are available to train local teams in high quality data collection [appendix 9.2] and many organisations have introduced training and data collection systems which are robust. However, we know that there is still some variation in application of the definitions and the method used to collect data. We need to continue to examine our data collection systems and work together to drive out variation. In this report you will see that there is significant variation between organisations. Some of this will be explained by data quality or case mix and locally we encourage organisations to learn from this variation. Despite this on-going challenge of improving data quality, the size of the data collection nationally means that we can conclude that some of the variation we see within and between organisations is indeed true variation in outcomes which can only be explained by variations in the quality of care. 1 Power M, Fogarty M, Madsen J, Fenton K, Stewart K, Brotherton A, Cheema K, Harrison A, Provost L. Learning from the design and development of the NHS Safety Thermometer. International Journal for Quality in Health Care 2014; 26(3): 287-97. 6

Key Messages How much impact has the NHS Safety Thermometer had? In 2012 we mobilised the healthcare systems to implement the NHS ST at scale on 100% of patients in NHS funded care on one day each month. We have now surveyed over four million patients, a globally unprecedented commitment to measuring harm. For the first time we know the proportion of patients harmed and are able to detect change over time at a local, regional and national level. This has taken commitment and courage from senior leaders and frontline teams and is a significant achievement for the NHS in England. Has the NHS Safety Thermometer changed our perceptions of harm? The NHS ST is designed to focus on the patient and not on attribution (whose fault is it) or avoid-ability of harm. It accepts that not all harm is avoidable but works on the premise that a significant amount is and that users are working towards a goal of defining the possible in their system. Attribution is used only as a key to system learning. The NHS ST is an attempt to shift our focus from blame to learning. Data are collected at the point of care, meaning that each month frontline staff are made aware of these issues as harms and the national CQUIN, combined with a policy focus 2, has brought these issues to the attention of senior leaders and commissioners. This has resulted in a rise in awareness of these four harms, a growing acceptance of them as harm and an ever increasing commitment to reducing them. Are we improving? We have demonstrated improvement in the harm free care composite measure and each individual harm, for example, we have seen a 13% reduction in pressure ulcer prevalence from July 2012 to June 2013. Most of the improvement achieved has been because people are intrinsically motivated to do this work. However, recognition must be given to NHS England who have signalled their strong commitment to safety and the NHS ST data collection through the national CQUIN. Improvement doesn t just happen. The change we see is a consequence of a plethora of national, regional and local improvement programmes including the national harm free care programme, the stop the pressure campaign in NHS Midlands and East, the Open and Honest Care programme in the North, the Quality Improvement and Safety Initiative in the South and the Quality Improvement programme in London. What next? Almost without exception, previous attempts to improve services have been thwarted by inadequate baseline data against which improvement can be measured. This year we are in the unique position of being able to measure improvement over time from a baseline using NHS ST data from locally, in regions and nationally. We also need to focus our improvement efforts on the processes of care which underpin these outcomes 3. A companion suite of process measures have been produced to guide organisations looking to improve 4. What are we learning? For the first time we are able to say that 7.2% of patients have one or more of these harms. 11,646 patients surveyed had two harms, 284 patients had three harms, and 4 patients surveyed had all four harms. A high level summary of the key measures can be seen on page 10. Key messages for each of the harms including proportions of harm, difference by setting, severity and variation between organisations can be seen from page 14 to 31. 2 The NHS Outcomes Framework 2014/15 3 Harm free care resources (see page 36) 4 Harm free care: learning from the Safety Express Pilot (see resources page 37) 7

Impact Between July 2012 and March 2014, 4,071,418 people have been surveyed. Data collection between January 2012 and June 2012 has been excluded as pilot data. Since July, approximately 200,000 patients have been surveyed each month. Data are collected in hospitals, care homes, patients own homes across the country by 829 organisations who regularly submit data. Number of patients surveyed and organisations submitting Number of patients surveyed 250,000 200,000 150,000 100,000 50,000 0 900 800 700 600 500 400 300 200 100 0 Number of organisations submitting May-12 Mar-12 Jan-12 Patients surveyed Organisations submitting Headline Results (P.3) Proportion of patients with a pressure ulcer: 5.16% (F.2) Proportion of patients with harm from a fall: 0.91% (C.2) Proportion of patients with a urine infection (and a catheter): 0.98% (V.3) Proportion of patients being treated for a new VTE: 0.35% (HFC.1) Proportion of patients with harm free care: 92.67% Operational Definitions of each harm measure can be found on page 33. 8

NHS Safety Thermometer at a glance 4,071,418 829 PEOPLE SURVEYED ORGANISATIONS COLLECTING DATA 5.16% 0.91% 0.98% 0.35% PRESSURE ULCERS FALLS WITH HARM (In the last 72 hours in a care setting) URINE INFECTIONS (in patients with a catheter) NEW VTE Harm free care 92.67% 9

Data Summary Key findings from the data have been summarised in a series of tables which illustrate: 7.1. The numbers of patients surveyed and the proportion of patients harmed in acute care 7.2. The numbers of patients surveyed and the proportion of patients harmed in non acute care 7.3. Percentage change and extrapolated benefit in acute care 7.4. Percentage change in non acute care 7.5. Different sources of data for measuring these harms where available Representation of the data over time and further measures for each harm can be seen from page 14. 7.1. Acute care settings This table shows the number of patients surveyed for each harm measure, the number of patients who were found to have a harm on the day of the survey and the proportion of patients harmed in acute care settings 1. The proportion of patients who received harm free care (HFC.2) (the absence of all four harms) is 92.75%. Indicator Number of patients surveyed Number of patients harmed Proportion of patients harmed 2 P2: Pressure ulcers of new origin 1,977,258 21,504 1.08% P3: Pressure ulcers of any origin 1,977,258 90,734 4.56% F2: Falls with harm 1,977,258 15,258 0.78% C3: New UTI with a catheter 1,977,258 12,147 0.62% V3: VTE treatment (new DVT or PE) 1,977,258 10,464 0.54% Whilst it is challenging to accurately estimate from proportions the number of patients harmed in the NHS in England, logic can be applied to the percentages to give a crude estimate. If we take HES admissions for this time period (23 million) we can use these proportions to estimate the number of patients harmed nationally. For example if we take the proportion of patients with a pressure ulcer (4.56%) and divide the population figure by 2 to account for length of stay, we can estimate that 535,000 people (178,502-1,071,013) had a pressure ulcer (July 12 March 14). 10

7.2. Non acute care settings This table shows the number of patients surveyed for each harm measure, the number of patients who were found to have a harm on the day of the survey and the proportion of patients harmed in non-acute care settings 3. The proportion of patients who received harm free care (HFC.2) (the absence of all four harms) is 92.59% Indicator Number of patients surveyed Number of patients harmed Proportion of patients harmed 2 P2: Pressure ulcers of new origin 2,094,160 26,477 1.24% P3: Pressure ulcers of any origin 2,094,160 119,069 5.79% F2: Falls with harm 2,094,160 21,790 1.24% C3: New UTI with a catheter 2,094,160 6,741 0.32% V3: VTE treatment (new DVT or PE) 2,094,160 3,981 0.22% 7.3. Change over time and extrapolated benefit in acute care settings This table shows the percentage change in acute care settings 1 from the baseline period (2012-13) to year one (2013-14). The final column estimates how many patients have been protected from harm across NHS England with a range 4 with the improvements seen to date.. Indicator Baseline (2012-13) Year One (2013-14) Amplitude of change Number of patients protected (in sample) Number of patients protected (in population) 5 P2: Pressure ulcers of new origin 1.20% 1.01% -15.83% 1,679 10,593 (3,531-21,186) P3: Pressure ulcers of any origin 4.80% 4.46% -7.08% 9,255 54,622 (18,207 109,243) F2: Falls with harm 0.92% 0.67% -27.17% 3,288 19,359 (6,453-38,719) C3: New UTI with a catheter 0.69% 0.57% -17.39% 3,459 20,269 (6,756 40,537) V3: VTE treatment (new DVT or PE) 0.63% 0.47% -25.40% 2,616 15,459 (5,153 30,917) HFC1: Harm free care 92.16% 93.21% 1.14% 27,154-5 1 Acute care settings includes acute hospital inpatient wards. 2 This is the median from July 2012 to March 2014. 3 Non acute care settings includes community hospitals, own home (district nursing caseloads), nursing homes, residential care homes and other community settings. 4 Estimated figures are based on full year HES () admissions for England and NHS ST national data for July 2012 to March. The percentage reduction in the NHS ST is taken from the HES data and then divided by the average length of stay (2 days). In order to account for the varied length of stay of patients who may have one of these four harms the range is then produced by dividing the population figure by length of stay ranging from 1 to 6 days (6 days representing the 90th percentile of the overall length of stay distribution). 5 It is not possible to extrapolate for the composite measure 11

7.4. Change over time in non acute care settings This table outlines the percentage change in community care settings 1 from the baseline period (2012-13) to year one (2013-14). It is not possible to estimate the number of patients protected from harm across NHS England in non acute settings as we have no denominator. Indicator Baseline (2012-13) Year One (2013-14) Amplitude of change P2: Pressure ulcers of new origin 1.45% 1.26% -13.10% P3: Pressure ulcers of any origin 6.59% 5.69% -13.66% F2: Falls with harm 1.14% 1.04% -8.77% C3: New UTI with a catheter 0.38% 0.32% -15.79% V3: VTE treatment (new) (DVT, PE) 0.24% 0.19% -20.83% HFC1: Harm free care 91.41% 92.57% 1.27% 7.5. Other data sources for acute care This table shows data from other data sources where available for the NHS ST measures. We recommend that NHS ST data are understood in the context of other data sources for measuring harm whilst recognising the varied purpose and challenges with each data source 2. These data cannot be directly compared due to differences in methodology, definitions and data collection systems. They can be used to understand whether the same trends are being seen over time. These data displayed over time can be seen from page 14. Indicator NHS ST Proportion of patients () Hospital Episode Statistics (HES)() Adverse Incident Reports (number of reports per month) Performance Data () P2: Pressure ulcers of new origin 1.09% - 3 - P3: Pressure ulcers of any origin 4.59% 0.51% 0.46% - F2: Falls with harm 0.77% 3.02% 4.1% - C1: Catheterisation 18.83% - - - C3: New UTI with a catheter 0.61% 0.02% - - V1: VTE risk assessment 88.60% - - 95.79% (UNIFY) V2: VTE prophylaxis 81.93% - - - V3: VTE treatment (new) (excl. other category) 0.53% 0.58% - - HFC1: Harm free care 92.79% - - - 1 Non acute care settings includes community hospitals, own home (district nursing case-loads), nursing homes, residential care homes and other community settings. 2 Power M, Stewart K, Brotherton A What is the NHS Safety Thermometer? Clinical risk 2012;18(5):163-9 3 A dash (-) signals that no aggregate national data are available from this data source. 12

Run Charts of Data Over Time The following charts show data on the 4,071,418 patients surveyed between July 2012 and March 2014. Most charts are displayed as run charts, showing the proportion of patients reported with a particular harm, or as harm free on a monthly time series. In all cases the vertical axis is a percentage. Scales vary according to the data presented. Unless specified otherwise the horizontal axis is time in months. Each data point shown typically represents the aggregate data for over 200,000 patients. The data presented is the aggregate data derived from the data submitted from all organisations. Non random patterns of variation were identified a priori to determine changes in the system which are unlikely to occur through chance alone (known as special cause variation) including: o An astronomical data point o Five consecutive data points ascending or descending (a run) o Six consecutive data points above or below the median (a shift) Median lines are presented on each chart and are re-set after each non-random pattern signals system change. Each median line is labelled with the percentage and a percentage reduction has been calculated based on the amplitude of change between the two median values. Interpreting Charts Median The median is a measure of central tendency, similar to a mean, and describes the middle number of a ranked series of numbers. For example, in the series 1, 1, 2, 3 and 4 the median is 2. When ranked, this is the middle number. Using Microsoft Excel it is easy to calculate the median of a set of numbers by ranking them and using the =MEDIAN() function. Common cause (or normal) variation This is variation inherent to the process being measured. If only common cause variation can be identified, the process is considered stable and predictable. Common cause variation isn t good or bad by itself; a process can be stable but still unacceptable and in need of change. Special cause variation This is variation that occurs when the source of variation is unusual and not inherent to the process itself. It means the process is unstable and unpredictable. When special cause variation is identified, its root cause can be investigated and if it is desirable (to the good) it can be incorporated, or eliminated if not. Introducing a change can be considered as creating a special cause variation. Shifts A shift is indicated by the presence of six or more data points on one side of the median. Points actually on the median line don t count; they neither break the string of six points nor add to it. Trends Five or more consecutive data points increasing or decreasing in the same direction indicate special cause variation in the form of a trend. Flat line segments don t count, either to break a trend or to count towards it. Calculate the revised median Take all the data points from the beginning of the shift onwards. Using only these data, calculate a new median to plot on the chart. There are two methods for doing this. A spreadsheet program (such as Microsoft Excel; the worksheet function is =MEDIAN()) can be used. Alternatively, calculate the median manually by ordering all the data points in ascending order and finding the middle value, or, if you have an even number of points, the average of the two middle points. 13

Pressure Ulcers On average each month 26,000 patients will be found to have a pressure across England (1) Improvement has largely occurred in prevention of category 2 pressure ulcers which account for 65% of all pressure ulcers (4) On average 6000 are newly acquired in hospital each month (1) In acute care the NHS ST finds 4.59% of patients with a pressure ulcer. This can now be triangulated with other data sources (HES and NRLS) (9) 14

1. The proportion of patients with a pressure ulcer Proportion of patients 7.0% 6.0% 5.0% 4.0% 3.0% 2.0% 1.0% 0.0% 5.43% 4.72% 1.28% 1.07% Any pressure ulcers (P3) New pressure ulcers (P2) Median The proportion of patients with a pressure ulcer (P3) has reduced by 13.08% from 5.43% to 4.72%. The change (special cause variation) occurred in June 13. The proportion of patients with a new pressure ulcer (P2) has reduced by 16.41% from 1.28% to 1.07%. The change occurred in June 13. 2. Pressure ulcers by care setting Proportion of patients 8.0% 7.0% 6.0% 6.03% 5.0% 4.88% 4.0% 3.0% 2.0% 1.0% 0.0% 5.06% 4.38% Acute settings Non acute settings Median The proportion of patients with a pressure ulcer (P3) in an acute setting has reduced by 10.24% from 4.88% to 4.38%. The change occurred in June 13. The proportion of patients with a pressure ulcer (P3) in non acute settings has reduced by 16.08% from 6.03% to 5.06%. The change occurred in June 13. 3. Non acute setting breakdown Setting July 2012 to March 2014 Patients surveyed Community (nationally) 5.93% 1,310,812 Community Hospital Ward 8.30% 192,880 Hospice 11.05% 8,359 Mental Health Community 0.80% 19,578 Mental Health Ward 1.18% 82,300 Nursing Home 4.48% 150,248 Other 1.85% 40,769 Own Home 5.69% 261,503 Residential Care Home 3.42% 27,711 This table shows the proportion of patients with a pressure ulcer (P3) in non acute settings split by the setting options within non acute. This varies from 0.8% in mental health community settings to 11.05% in hospice settings, however due to case mix and sample size these different settings cannot be directly compared. 15

4. Pressure ulcers by category Proportion of patients 5.0% 4.0% 3.0% 2.0% 1.0% 0.0% 3.61% 3.08% 1.17% 1.05% 0.63% Category 2 Category 3 Category 4 Median The proportion of patients with a category 2 pressure ulcer (P3) has reduced by 14.68% from 3.61% to 3.08%. The change occurred in June 13. The proportion of patients with a category 3 pressure ulcer (P3) has reduced by 10.26% from 1.17% to 1.05%.The change occurred in September 13. The proportion of patients with a category 4 pressure ulcer (P3) is 0.63%. 5. Pressure ulcers by category (acute settings) Proportion of patients 5.0% 4.0% 3.0% 2.0% 1.0% 0.0% 3.52% 0.89% 0.42% 3.06% The proportion of patients with a category 2 pressure ulcer (P3) in acute settings has reduced by 13.07% from 3.52% to 3.06%. The change occurred in May 13. The proportion of patients with a category 3 pressure ulcer (P3) in acute settings is 0.89% and category 4 is 0.42%. Category 2 (P3) Category 3 (P3) Category 4 (P3) 6. Pressure ulcers by category (non acute settings) Proportion of patients 5.0% 4.0% 3.0% 2.0% 1.0% 0.0% 4.21% 3.61% 3.05% 1.42% 1.22% 0.83% Category 2 (P3) Category 3 (P3) Category 4 (P3) The proportion of patients with a category 2 pressure ulcer (P3) in non acute settings has reduced by 27.5% from 4.21% to 3.05%. The change occurred in November 12 and July 13. The proportion of patients with a category 3 pressure ulcer (P3) in non acute settings has reduced by 14.08% from 1.42% to 1.22%. The change occurred in September 13. The proportion of patients with a category 3 pressure ulcer (P3) in non acute settings is 0.83%. 16

7. The burden of pressure ulcers in different specialties 16,000 14,000 12,000 10,000 8,000 6,000 4,000 2,000 0 GENERAL MEDICINE GERIATRIC MEDICINE GENERAL SURGERY TRAUMA & RESPIRATORY CRITICAL CARE MIXED SPECIALTY STROKE MEDICINE REHABILITATION ACCIDENT & GASTROENTEROLOGY CLINICAL ONCOLOGY SPINAL INJURIES NEUROLOGY UROLOGY VASCULAR SURGERY EMERGENCY MEDICAL ONCOLOGY CARDIOTHORACIC INFECTIOUS DISEASES DIABETIC MEDICINE PALLIATIVE MEDICINE CLINICAL OLD AGE PSYCHIATRY COLORECTAL SURGERY PLASTIC SURGERY 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Pareto charts can be used to determine where the largest opportunity for improvement can be found. The pareto analysis shows that 80% of all pressure ulcers are found in the 8 specialties highlighted by the red box. (time period is July 12 to March 14). The remaining 20% is made up of 26 specialties. This demonstrates where pressure ulcers have been found, both old and new. The opportunity for prevention of old pressure ulcers found in these specialties may be elsewhere. 8. Variation between organisations Rate per 1,000 100 90 80 70 60 50 40 30 20 10 0 0 10000 20000 30000 40000 50000 60000 70000 Number of patients surveyed Organisations Lower and upper control limits Funnel plots can be used to highlight variation between organisations and identify outliers. Variation may be due to case mix, different collection methods and variation in performance. This funnel plot shows variation between acute organisations, each dot represents one organisation for pressure ulcer prevalence (P3). (Over time period July 12 to March 14). 9. Pressure Ulcers recorded in other Data Sources Hospital Episodes Statistics Incident Reporting 1.0% 7000 Propoprtion of admissions 0.8% 0.6% 0.4% 0.2% 0.0% 0.53% Number of incidents with a pressure ulcer recorded 6000 5000 4000 3000 2000 1000 0 5138 476 Acute Non Acute Median The proportion of patient admissions with a pressure ulcer coded in HES (Hospital Episode Statistic) is 0.53%. There has been no change signalled over time. The median number of incidents related to a pressure ulcer reported to NRLS each month is 5138 in acute settings and 476 in non acute settings. There has been no change signalled over time. If we use HES admissions as a denominator this gives us 0.51%. 17

Falls There has been a 25% reduction in falls (in the last 72 hours) (10) There has been a 35% reduction in falls with harm. This is the single largest change seen in all of the measures (10) 1,508 patients surveyed in acute care were found to have severe harm from falls (13) For the first time we have consistent data on harm from falls in non acute settings (12) 18

10. The proportion of patients with evidence of a fall in a care setting in the last 72 hours Proportion of patients 3.0% 2.5% 2.0% 1.5% 1.0% 0.5% 0.0% 2.42% 1.04% 2.04% 0.92% 0.78% Falls (F1) Falls with harm (F2) Median The proportion of patients with evidence of fall in the last 72 hours (F1) has reduced by 15.70% from 2.42% to 2.04%. The change occurred in May 13. The proportion of patients with evidence of harm from a fall in the last 72 hours (F2) has reduced by 25.00% from 1.04% to 0.78%. The change occurred in January 13 and September 13. 11. Falls in the last 72 hours (acute settings) Proportion of patients 1.2% 1.0% 0.8% 0.6% 0.4% 0.2% 0.0% 0.94% 0.81% 0.61% The proportion of patients with harm from a fall in the last 72 hours (F2) in acute settings has reduced by 35.10% from 0.94% to 0.61%. The change occurred in January 13 and August 13. Fall with harm (F2) Median 12. Falls in the last 72 hours (non acute settings) July 2014 March 2014 Proportion of patients 1.6% 1.4% 1.2% 1.0% 0.8% 0.6% 0.4% 0.2% 0.0% 1.07% Falls with harm (F2) 0.95% Median Setting All Falls Falls with harm Patients Surveyed Community 1.85% 1.05% 1,310,812 Community Hospital Ward 3.23% 1.03% 192,880 Hospice 3.83% 1.23% 8,359 The proportion of patients with harm from a fall in the last 72 hours (F2) in non acute settings has reduced by 11.21% from 1.07% to 0.95%. The change occurred in August 13. This table shows the proportion of patients with harm from a fall in the last 72 hours (F2) in non acute settings split by the setting options within non acute. This varies from 2.07% in mental health community settings to 1.23% in hospice settings, however due to case mix and sample size these different settings cannot be directly compared. Mental Health Community 3.54% 2.07% 19,578 Mental Health Ward 2.98% 1.18% 82,300 Nursing Home 3.75% 1.23% 150,248 Other 2.26% 0.90% 40,769 Own Home 1.55% 0.83% 261,503 Residential Care Home 3.17% 1.28% 27,711 19

13. Severity of harm from falls Proportion of patients 0.9% 0.8% 0.7% 0.6% 0.5% 0.4% 0.3% 0.2% 0.1% 0.0% 0.69% 0.28% 0.04% 0.56% 0.20% Low Moderate Severe Median The proportion of patients with low harm from a fall in the last 72 hours (F2) has reduced by 18.84% from 0.69% to 0.56%. This change occurred in September 13. The proportion of patients with moderate harm from a fall in the last 72 hours (F2) has reduced by 28.57% from 0.28% to 0.20%. This change occurred in January 13. The proportion of patients with severe harm from a fall (F2) is 0.04%. 14. Variation between organisations Rate per 1,000 40 35 30 25 20 15 10 5 0 0 5000 10000 15000 20000 25000 30000 35000 40000 Number of patients surveyed Organisations Upper and lower control limits Funnel plots can be used to highlight variation between organisations and identify outliers. Variation may be due to case mix, different collection methods and variation in performance. This funnel plot shows variation between acute organisations for falls with harm (F2), each dot is one organisation. (Over time period July 12 to March 14). 15. Falls recorded in other Data Sources Hospital Episodes Statistics Incident Reporting Propoprtion of admissions 4.0% 3.5% 3.0% 2.5% 2.0% 1.5% 1.0% 0.5% 0.0% 3.05% Number of incidents with a slip, trip or fall recorded 20,000 15,000 10,000 5,000 0 15,986 3745 Acute Non Acute Median The proportion of patient admissions with a fall coded in HES (Hospital Episode Statistic) is 3.05%. There has been no change signalled over time. The median number of incidents with a slip, trip or fall reported to NRLS is 15,986 in acute settings and 3745 in non acute settings. There has been no change signalled. If we use HES admissions as a denominator this gives us 1.4%. 20

Catheters and UTI 1 in 5 patients in acute care have a catheter and just under 1 in 10 patients in non acute care (18) There has been a More 35% reduction women have in a falls catheter with harm. and a UTI This than is the men single but there largest has change been a 32% seen improvement in all of the measures (22) (10) 8.23% of patients with a catheter are also being treated for a UTI (19) Reduction For the first in time UTIs we does have not consistent appear to data be on driven harm by from falls reduction in non acute in catheter settings use (12) (16) 21

16. The proportion of patients with a catheter Proportion of patients 16.0% 14.0% 12.0% 10.0% 8.0% 6.0% 4.0% 2.0% 0.0% 12.90% The proportion of patients with an in dwelling urethral urinary catheter present on the day of survey or removed in the last 72 hours (C1) is 12.90%. This median, however, is misleading as there is a significant difference between acute and non acute settings as shown in the two following charts. Catheter (C1) Median 17. Catheters in acute settings 25.0% The proportion of patients with a catheter (C1) in acute settings is 18.90% Proportion of patients 20.0% 15.0% 10.0% 5.0% 0.0% 18.90% Catheter (C1) Median 18. Catheters in non acute settings Proportion of patients 10.0% 8.0% 6.0% 4.0% 2.0% 0.0% 7.90% 7.30% The proportion of patients with a catheter in non acute settings has reduced by 7.59% from 7.90% to 7.30%. The change occurred in January 13. Catheter (C1) Median 22

19. Catheters and UTIs Proportion of patients 0.8% 0.6% 0.4% 0.2% 0.0% 0.55% 0.47% Catheter and new UTI (C3) 0.39% Median The proportion of patients with a catheter and receiving treatment for a new UTI (C3) has reduced by 29.09% from 0.55% to 0.39%. The change occurred in December 12 and October 13. If we use patients with a catheter as the denominator we can see that 8.23% of patients with a catheter are also being treated for a UTI (old and new) 20. Catheters and UTIs in acute settings Proportion of patients 1.0% 0.8% 0.6% 0.4% 0.2% 0.0% 0.76% 0.63% 0.51% The proportion of patients with a catheter and receiving treatment for a new UTI (C3) in acute settings has reduced by 32.89% from 0.76% to 0.51%. The change occurred in November 12 and October 13. Catheter and new UTI (C3) Median 21. Catheters and UTIs in non acute settings Proportion of patients 0.6% 0.5% 0.4% 0.3% 0.2% 0.1% 0.0% 0.38% 0.32% 0.29% The proportion of patients with a catheter and receiving treatment for a new UTI (C3) in non acute settings has reduced by 23.68% from 0.38% to 0.29%. The change occurred in January 13 and October 13. Catheter and new UTI (C3) Median 23

22. Catheters and UTIs by gender Proportion of patents 1.0% 0.8% 0.6% 0.4% 0.2% 0.0% 0.72% 0.41% 0.59% 0.49% 0.33% Males Females Median The proportion of male patients with a catheter and receiving treatment for a new UTI (C3) has reduced by 19.51% from 0.41% to 0.33%. The change occurred in October 13. This measure for females has reduced by 31.94% from 0.72% to 0.49%. The change occurred in November 12 and August 13. 23. Catheters by time in situ The proportion of patients with a catheter with a time in situ of: 10.0% 8.98% 1-28 days is 8.98%. Proportion of patents 8.0% 6.0% 4.0% 2.0% 0.0% 3.26% 0.70% >28 days is 3.26% Days not known is 0.70% 1-28 days >28 days Days not known Median 24. Proportion of patients with a catheter and being treated for a UTI by time in situ Proportion of patents 0.5% 0.4% 0.3% 0.2% 0.1% 0.0% 0.38% 0.16% 0.03% 0.31% 0.14% 0.25% 0.11% 0.02% The proportion of patients with an in dwelling catheter present on the day of survey or removed in the last 72 hours with a time in situ of 1-28 days and receiving treatment for a new UTI has reduced by 34.21% from 0.38% to 0.25%. This measure for >28 days has reduced by 31.25% from 0.16% to 0.11%. 1-28 days >28 days Days not known Median This measure for days not known has reduced by 33.33% from 0.03% to 0.02% For both 1-28 days and >28 days the change occurred in November 12 and October 13. The change occurred for days not known in October 13. 24

25. Variation between organisations Rate per 1,000 450 400 350 300 250 200 150 100 50 0 Variation of catherisation (C1) 0 5000 10000 15000 20000 25000 30000 35000 40000 Number of patients surveyed Funnel plots can be used to highlight variation between organisations and identify outliers. Variation may be due to case mix, different collection methods and variation in performance. Funnel plots show variation between all acute organisations, each dot represents one organisation. (Over time period Feb 13 to Jan 14). Organisations Upper and lower control limits Variation of catherisation and being treated for a new UTI (C3) 25 20 Rate per 1,000 15 10 5 0 0 5000 10000 15000 20000 25000 30000 35000 40000 Number of patients surveyed Organisations Upper and lower control limits 26. Catheters and UTIs in other Data Sources Propoprtion of admissions 0.04% 0.03% 0.02% 0.01% 0.00% 0.015% Hospital Episodes Statistics The proportion of patient admissions with a catheter associated UTI as coded in HES (Hospital Episodes Statistics) is 0.015%. There has been no change signalled over time. 25

VTE VTE prophylaxis is now being measured in addition to risk assessment (28) VTE is a significant cause of avoidable harm in adult patients and is not confined to a particular age group (32) The proportion of patients being treated for a VTE has reduced by 21% (29) 26

27. The proportion of patients with a VTE risk assessment (acute settings) Proportion of patients 92.0% 91.0% 90.0% 89.0% 88.0% 87.0% 86.0% 85.0% 84.0% 88.48% The proportion of patients with a documented VTE risk assessment (V1) in acute settings is 88.48%. The data include patients who may have been in care for less than 24 hours which could explain why this proportion is lower than that reported to UNIFY. Acute Median 28. The proportion of patients with appropriate prophylaxis (acute settings) Proportion of patients 86.0% 85.0% 84.0% 83.0% 82.0% 81.0% 80.0% 79.0% 78.0% 77.0% 76.0% 81.62% The proportion of at risk patients that are receiving appropriate prophylaxis (V2) in acute settings is 81.62%. Acute Median 29. The proportion of patients receiving treatment for a VTE Propoprtion of patients 2.0% 1.8% 1.6% 1.4% 1.2% 1.0% 0.8% 0.6% 0.4% 0.2% 0.0% 1.73% 1.61% 1.36% The proportion of patients that are receiving treatment for a clinically documented VTE event (old or new) (V3) excluding other has reduced by 21.38% from 1.73% to 1.36%. The change occurred in December 12 and June 13. Any VTE (V3) Median 27

30. Treatment for new VTE (acute settings) Propoprtion of patients 0.8% 0.7% 0.6% 0.5% 0.4% 0.3% 0.2% 0.1% 0.0% 0.64% 0.54% 0.47% The proportion of patients that are receiving treatment for a new clinically documented VTE event (V3) in acute settings excluding other has reduced by 26.56% from 0.64% to 0.47%. The change occurred in December 12 and June 13. New VTE (V3) Median 31. Treatment for new VTE (non acute settings) Propoprtion of patients 0.3% 0.2% 0.1% 0.0% 0.21% 0.19% 0.15% The proportion of patients that are receiving treatment for a new clinically documented VTE event (V3) in non acute settings has reduced by 28.57% from 0.21% to 0.15%. The change occurred in February 13 and September 13. New VTE (V3) Median 32. VTE by age group Propoprtion of patients 0.8% 0.7% 0.6% 0.5% 0.4% 0.3% 0.2% 0.1% 0.0% The proportion of patients that are receiving treatment for a new clinically documented VTE event (V3) by age group is outlined below. Age Group July 2012 July 2013 Under 18 0.09% 0.02% 18-70 0.58% 0.36% Under 18 18 to 70 Over 70 Over 70 0.45% 0.31% 28

33. Treatment for new VTE by category 0.4% The proportion of patients being treated for a DVT is 0.14% Propoprtion of patients 0.3% 0.2% 0.1% 0.0% The proportion of patients being treated for a PE is 0.17% DVT PE 34. Variation between organisations Rate per 1,000 25 20 15 10 5 0 0 5000 10000 15000 20000 25000 30000 35000 40000 Number of patients surveyed Organisations Upper and lower control limits Funnel plots can be used to highlight variation between organisations and identify outliers. Variation may be due to case mix, different collection methods and variation in performance. This funnel plot shows variation between all acute organisations for treatment for new VTE (V3), each dot represents one organisation. (Over time period July 12 to March 13). 35. VTE and VTE risk assessment in other data Sources Hospital Episodes Statistics Unify Propoprtion of admissions 1.0% 0.8% 0.6% 0.4% 0.2% 0.0% 0.58% Propoprtion of admissions 100.0% 99.0% 98.0% 97.0% 96.0% 95.0% 94.0% 93.0% 92.0% 91.0% 90.0% 94.24% 95.82% The proportion of patient admissions with a VTE coded in HES (Hospital Episode Statistic) is 0.58%. There has been no change signalled over time. The proportion of admissions with a VTE risk assessment recorded has increased by 1.68% from 94.24% to 95.82%. The change occurred in April 13. 29

Harm Free Care There has been a 16% reduction in people who experience one or more harms (36) There has been a 35% reduction in falls 11,646 with patients harm. This surveyed is the had single 2 harms largest and change 284 had seen 3 harms in all (37) of the measures (10) 93% of patients now receive harm free care (36) For the first time 4 we of have those consistent surveyed data had on all harm 4 harms from falls in (37) non acute settings (12) 30

36. Proportion of patients receiving harm free care Propoprtion of patients 95.0% 94.0% 93.0% 92.0% 91.0% 90.0% 89.0% 92.21% 93.44% The proportion of patients without a pressure ulcer, harm from a fall, a urinary infection (in patients with a catheter) or new VTE (HFC1) has increased by 1.33% from 92.21% to 93.44%. The change occurred in June 13. This represents a 15.79% reduction in people with one or more harm. Harm free care Median 37. Number of harms Propoprtion of patients 9.0% 8.0% 7.0% 6.0% 5.0% 4.0% 3.0% 2.0% 1.0% 0.0% 7.48% 0.29% 6.33% The proportion of patients with 1 harm has reduced by 15.37% from 7.48% to 6.33%. The change occurred in June 13. 11,646 patients surveyed had 2 harms and 284 had 3 harms. 4 of those surveyed had all 4 harms. 1 harm More than 1 harm Median 38. Harm free care in acute settings Propoprtion of patients 96.0% 95.0% 94.0% 93.0% 92.0% 91.0% 90.0% 89.0% 92.23% 93.49% The proportion of patients without a pressure ulcer, harm from a fall, a urinary infection (in patients with a catheter) or new VTE (HFC1) in acute settings has increased by 1.37% from 92.23% to 93.49%. The change occurred in June 13. Harm free care Median 39. Harm free care in non acute settings Propoprtion of patients 96.0% 95.0% 94.0% 93.0% 92.0% 91.0% 90.0% 92.15% 93.37% The proportion of patients without a pressure ulcer, harm from a fall, a urinary infection (in patients with a catheter) or new VTE (HFC1) in non acute settings has increased by 1.32% from 92.15% to 93.37%. The change occurred in June 13. 89.0% Harm free care Median 31

Cautions With Data Interpretation The NHS Safety Thermometer was designed to measure local improvement over time and should not be used to compare organisations. There are differences in patient mix and data collection methods that can invalidate direct comparison across organisations. Users need to be trained and understand the operational definitions. Whilst the NHS Safety Thermometer is intuitive, staff who are using it need to be trained in its use. This is particularly critical for some of the operational definitions where the classifications are complex. Not all harm is avoidable. We have no way of knowing how much of the harm detected by the NHS Safety Thermometer is avoidable. It is not appropriate to interpret the data in the NHS Safety Thermometer as avoidable harm; some of it will be but some of it won t be. The NHS Safety Thermometer should not be used for attribution of causation. The NHS Safety Thermometer is about measuring patients and their harm burden not organisations and their harm burden. We strongly recommend a health economy wide discussion about the sources of harm so all organisations can work together for the benefit of patients. Definitions of old and new refer to where the patient was harmed and not the time period; i.e. new is a harm that developed in the setting where data collection takes place. Full guidance on the definitions can be found on the Health and Social Care Information Centre website. There is a potential for patient harms to be captured in the NHS Safety Thermometer on consecutive months. As we are viewing proportions over time and not counting the number of harms this does not affect the ability to use the data to measure improvement over time but it can be uncomfortable, especially for community services who may record the same pressure ulcer more than once. 32

Operational Definitions The definitions for each of the indicators described in the following pages can be found in the tables below. Pressure Ulcers P.1 The proportion of patients with an OLD pressure ulcer (present on admission to your organisation or developed within 72 hours) documented following skin inspection. P.2 The proportion of patients with a NEW pressure ulcer (NOT present on admission to your organisation & developed after 72 hours) documented following skin inspection. P.3 The proportion of patients with ANY (new or old) pressure ulcer documented following skin inspection on the day of the survey (see ST guidance). Falls F.1 The proportion of patients with evidence of a fall in a care setting in the last 72 hours (incl. home if on a DN caseload) from discussion with the patient & review of clinical notes reviewed on the day of survey. F.2 The proportion of patients with evidence of harm from a fall in a care setting in the last 72 hours (incl. home if on a DN caseload) from discussion with the patient & review of clinical notes reviewed on the day of survey. i. Each measure can be viewed by category (II-IV) ii. This measure can be viewed by harm severity Catheters & Urine Infection C.1 The proportion of patients with an In dwelling urethral urinary catheter present on the day of survey or removed in the last 72 hours C.2 The proportion of patients with an In dwelling urethral urinary catheter also receiving treatment for ANY urinary tract infection (on the basis of notes, clinical judgement and patient feedback) C.3 The proportion of patients with an In dwelling urethral urinary catheter also receiving treatment for a NEW urinary tract infection (on the basis of notes, clinical judgement and patient feedback) iii. This measure can also be viewed by OLD UTI iv. The proportion of patients (without catheters) being treated for UTI can be also viewed VTE V.1 The proportion of patients with a documented VTE risk assessment V.2 The proportion of at risk patients receiving appropriate prophylaxis (in accordance with local guidance) V.3 The proportion of patients receiving prescribed anticoagulation treatment (heparin, warfarin or equivalent) for treatment of a clinically documented VTE event. v. Each measure can be viewed by category (DVT / PE / Other) vi. This measure can be viewed by OLD and NEW VTE. Harm free care indicator 1 (HFC 1): The proportion of patients without any documented evidence of a pressure ulcer, (ANY origin, category II-IV), harm from a fall in care in the last 72 hours, a urinary infection (in patients with urinary catheter) or new VTE (developed since admission to this organisation). (The proportion of patients without documented evidence of P3, F2, C2 or V3). Harm free care indicator 2 (HFC 2): The proportion of patients without any documented evidence of a new pressure ulcer (developed at least 72 hours after admission of this care setting, category II-IV), harm from a fall in care in the last 72 hours, a new urinary infection in patients with urinary catheter (which has developed since admission to this organisation). (The proportion of patients without documented evidence of P2, F2, C3 or V3). 33

Harm Free Care in Quality Accounts The NHS ST was developed due to a gap in measurement systems that could be used to understand the burden of harm and measure change over time. These extracts from Quality Accounts highlight how organisations are now using the NHS Safety Thermometer data and harm free care resources to deliver and measure improvement. Central Manchester University Hospitals NHS Foundation Trust Quality Account 2011/12 Bolton Royal Hospital NHS Foundation Trust Quality Account 34

Salford Royal NHS Foundation Trust Quality Account South Essex Partnership University NHS Foundation Trust Quality Account 35