Trust Board Meeting: Wednesday 12 November 2014
TB2014.115
Title
Board Quality Report
Status
For Information
History
This is a monthly report, presented alternately to the Trust Board
or to the Quality Committee
Board Lead(s)
Dr Tony Berendt, Medical Director
Key purpose
TB2014.115 Quality Report
Strategy
Assurance
Policy
Performance
Page 1 of 24
Oxford University Hospitals
TB2014.115
Executive Summary
1. The Board Quality Report (BQR) presents information that is as contemporary as
possible and where possible includes data from the last calendar month.
2. In relation to key quality metrics:
• For 13 of the 53 quality metrics, pre-specified targets were not fully achieved in the
last relevant data period. For selected metrics, trend data are provided along with
brief exception reports.
• 46 of the 53 quality metrics report performance for the month of September 2014.
This is the most contemporaneous and validated information available.
3. In relation to patient safety and clinical risk:
• 4 Serious Incidents Requiring Investigation (SIRI) were reported in October 2014.
• For October 2014, five SIRI reports were recommended to the Oxfordshire Clinical
Commisisoning Group for closure.
4. In relation to Quality Walk Rounds:
• There were 5 Quality Walk Rounds during October 2014.
5. In relation to clinical effectiveness:
• The Summary Hospital-Level Mortality Indicator (SHMI) was updated on 23
October 2014. The current SHMI is 0.99, which is within expected range.
6. Patient Experience:
• Patient experience information is presented in a dashboard format, including Family
and Friends Test data, complaints, activity, PALS and compliments.
7. Recommendation
The Trust Board is asked to receive this report.
TB2014.115 Quality Report
Page 2 of 24
Oxford University Hospitals
TB2014.115
Board Quality Report
1.
Purpose
1.1. This paper aims to provide the Trust Board with information on the quality of
care provided within the organisation, and on the measures being taken in
relation to quality assurance and improvement.
1.2. This Board Quality Report will be received for information by relevant Trust
Committees (Clinical Governance Committee) following the meeting of the Trust
Board.
2.
Key Quality Metrics
2.1. A suite of fifty three key quality metrics is currently used for consideration by the
Board, these are listed in dashboard format in Table 1.
2.2. These metrics have been chosen as they are considered to be linked to the
quality of clinical care provided across the organisation, and data quality is
deemed satisfactory.
2.3 Trend graphs and exception reports in relation to selected metrics, where
specified thresholds have not been met (‘red-rated’) or are amber-rated having
been green-rated in the previous period, are included. Thresholds are drawn
from a mixture of sources (national benchmarking, commissioner-generated
and internally-derived).
2.4
Data provided are as contemporaneous as is possible. As a result of the
reporting timeframes associated with the Board Quality Report, there are a
number of metrics which may not have been validated. The detailed sections of
the Board Quality Report outline October information, however October
information is not available for the dashboard metrics. October information is
provided in the body of the report where the information has been available
before month end.
TB2014.115 Quality Report
Page 3 of 24
Oxford University Hospitals
TB2014.115
Table 1
BQR ID
Rating
Rating Descriptor
Last
Period
Period
PS01
97.78%
Green
Safety Thermometer (% patients receiving
Amber care free of any newly acquired harm) [one
month in arrears]
Sep 14
PS02
93.8%
Green
Safety Thermometer (% patients receiving
Amber care free of any harm - irrespective of
acquisition) [one month in arrears]
PS03
94.67%
Red
Red
Amber
Internal
95%
97%
Sep 14
Internal
91%
93%
VTE Risk Assessment (% admitted patients
receiving risk assessment)
Sep 14
National
95%
95.25%
PS04
2
N/A
Serious Incidents Requiring Investigation
(SIRI) reported via STEIS
Sep 14
N/A
N/A
PS05
31
Green
Green
Number of cases of Clostridium Difficile > 72
hours (cumulative year to date)
Sep 14
National
35
N/A
PS06
0
Green
Green
Number of cases of MRSA bacteraemia > 48
hours (cumulative year to date)
Sep 14
National
1
N/A
Internal
85%
88%
PS07
Antibiotic prescribing - % prescriptions where
indication and duration specified [most
Green
recently available figure, undertaken
quarterly]
Jul 14
86.6%
Amber
98%
Green
Antibiotic prescribing - % compliance with
Green antimicrobial guidelines [most recently
available figure, undertaken quarterly]
Jul 14
Internal
93%
95%
PS08
PS09
80.91%
Amber
Amber
% patients receiving stage 2 medicines
reconciliation within 24h of admission
Sep 14
Internal
75%
85%
PS10
96.55%
Green
Amber
% patients receiving allergy reconciliation
within 24h of admission
Sep 14
Internal
94%
96%
PS11
1887
N/A
N/A
N/A
PS12
3.87%
Green
6.5%
5%
PS13
42
N/A
N/A
N/A
PS14
1
Green
Green Falls leading to moderate harm or greater
PS15
0
Green
Number of hospital acquired thromboses
Green identified and judged avoidable [two months
in arrears]
PS16
46.27%
N/A
PS17
2.09%
Green
PS18
98.59%
Green
PS19
8
N/A
PS20
0
Green
CE01
CE02
Red
Total number of incidents reported via Datix
Green
Threshold
Source
Sep 14
% of incidents associated with moderate
harm or greater
Sep 14
Total number of newly acquired pressure
ulcers (category 2,3 and 4) reported via Datix
Sep 14
Internal
Sep 14
Internal
8
7
Aug 14
Internal
1
0
N/A
N/A
Cleaning Score - % of inpatient areas with
initial score > 92%
Sep 14
Green
% 3rd and 4th degree tears in obstetrics
[C&W Division]
Sep 14
Internal
5%
N/A
Green
% radiological investigations achieving 5
day reporting standard [CSS Division]
Aug 14
Commissioner
95%
98%
N/A
N/A
1
N/A
Number of CAS alerts received
Sep 14
CAS alerts breaching deadlines at end of
month and/or closed during month beyond
deadline
Sep 14
0.99
N/A
Standardised Hospital Mortality Ratio (SHMI)
[most recently published figure, quarterly
reported as a rolling year ending in month]
Sep 14
N/A
N/A
189
Crude Mortality
Sep 14
N/A
N/A
Red
TB2014.115 Quality Report
Internal
Page 4 of 24
Oxford University Hospitals
TB2014.115
N/A
67.2%
Red
National
80%
90%
Red
Dementia - % patients aged > 75 admitted
as an emergency who are screened [one
month in arrears]
Sep 14
CE03
CE04a
79.4%
Red
Red
Statutory and Mandatory Training - %
required modules completed
Sep 14
Internal
85%
95%
CE05
93.61%
Amber
Green
ED - % patients seen, assessed and
discharged / admitted within 4h of arrival
Sep 14
National
85%
95%
CE06
85.25%
Green
Amber
Stroke - % patients spending > 90% of
admission in specialist stroke environment
Sep 14
National
70%
80%
CE07
84.75%
Amber
Amber
Stroke - % patients accessing specialist
stroke environment within 4h of arrival
Sep 14
National
75%
85%
CE08
502
N/A
N/A
N/A
93.74%
Green
% of elective paediatric day cases managed
Green as such (Did not result in an overnight stay)
[C&W Division]
Aug 14
Internal
70%
75%
CE09
5.9
Amber
Vascular - Mean length of stay for patients
Amber undergoing elective AAA repair (3 month
rolling period) [NOTSS Division]
Aug 14
Internal
8
5
CE10
CE11
2.63%
Green
Green
Aug 14
Internal
5%
3%
94.44%
Green
Cardiology - % patients receiving primary
Green angioplasty within 60 minutes of arrival at
hospital [MRC Division]
Sep 14
Internal
85%
90%
CE12
2.7
Amber
Cardiology - Mean number of days from
Amber referral to admission to cardiology at tertiary
centre [MRC Division]
Sep 14
Internal
3
2
CE13
0%
Green
Cardiac surgery-% rate of patients with
Green organ space infections following cardiac
surgery via the sternum [MRC Division]
Sep 14
Internal
1%
0.5%
CE14
CE15
1.56%
Green
Green
Cardiac Surgery - % mortality following
elective primary CABG [MRC Division]
Aug 14
Internal
6%
4%
CE16
0
Green
Red
Number of unscheduled returns to theatre
within 48 hours [NOTSS Division]
Sep 14
Internal
2
1
100%
Green
Rheumatology - % relevant patients who
Green have their DAS28 score documented
[NOTSS Division]
Sep 14
Internal
95%
98%
CE17
CE18
0
Green
Green
Number of unscheduled returns to theatre in
gynaecology [C&W Division]
Sep 14
Internal
2
1
CE19
549
N/A
Number of patients admitted to SEU wards
from SEU triage [S&O Division]
Sep 14
N/A
N/A
75.64%
Red
% patients having their operation within the
time specified according to their clinical
categorisation [CSS Division]
Sep 14
Internal
90%
95%
CE20
3.01%
Amber
Neuroscience Intensive Therapy Unit (NITU)
Amber readmission rate within 48 hours of
discharge [NOTSS Division]
Sep 14
Internal
4%
2%
CE21
80.47%
Green
% fractured NOF patients who receive
Green surgery within 36 hours of admission
[NOTSS Division]
Sep 14
Commissioner
70%
72%
CE22
CE23
24.71%
Amber
Amber % deliveries by C-Section [C&W Division]
Sep 14
Commissioner
33%
23%
1.02%
Green
7 day admission rate following assessment
Green on (and discharge from) paediatric CDU
[C&W Division]
Sep 14
Internal
4%
2%
CE24
Transfer Lounge Usage
Red
Vascular - % mortality following elective AAA
repair [NOTSS Division]
TB2014.115 Quality Report
Sep 14
Page 5 of 24
Oxford University Hospitals
PE01
73
Green
Green
Friends & Family - Net Promoter Score [one
month in arrears]
PE02
94.1%
Green
Friends & Family - proportion extremely likely
Green or likely to recommend [one month in
arrears]
PE03
82
Amber
Green Complaints Received
PE04
0
Green
Green
PE05
254
N/A
PE06
0
Green
PE07
70.94%
Green
PE08
78.99%
N/A
PE09
0
Green
PE10
58.67%
Amber
PE11
9
N/A
Number of complaints received initially
graded as RED
PALS contacts made
Green Single sex breaches
Amber % patients EAU length of stay < 12h
Red
TB2014.115
Sep 14
Internal
63
70
Sep 14
Internal
90%
94%
Sep 14
Internal
90
80
Sep 14
Internal
2
1
N/A
N/A
Sep 14
Sep 14
National
3
2
Sep 14
Internal
65%
70%
N/A
N/A
% Complaints upheld or partially upheld
[Quarterly in arrears]
Sep 14
Number of legal claims received / inquests
opened initially graded as RED
Mar 14
Internal
2
N/A
Sep 14
Internal
45%
60%
N/A
N/A
% patients returning feedback forms in
Green specialist surgery outpatients [NOTSS
Division]
Number of reopened complaints
TB2014.115 Quality Report
Sep 14
Page 6 of 24
Oxford University Hospitals
ORBIT
Reporting
TB2014.115
Board Quality Report
How to interpret charts
Data are presented in this report in a number of different ways – including statistical process control (SPC) charts,
line charts (without confidence intervals / control limits), histograms and cumulative histograms. Graphics have
been selected in order to encourage the analysis of trends and to identify when a change in relation to the historical
position is likely to be ‘real’ or statistically significant.
SPC charts show a trend line and allow easy reference to the historical mean for that metric at a time at which it
was stable and ‘within control’. Where shown, the mean is displayed as a horizontal orange line. In addition,
warning limits and control limits are shown where appropriate, above and below the mean. Warning limits are
placed at two standard deviations (2SD – dashed black line) and control limits at three standard deviations (3SD –
solid black line). If a data point is found beyond the control limit (3SD from the mean) in either direction, the change
is statistically significant and is very unlikely to have occurred simply by chance.
There are other patterns within the data that are likely to reflect real change as opposed to random fluctuation –
these patterns are known as special cause variations. They include:
2 consecutive points lying beyond the warning limits (unlikely to occur by chance)
7 or more consecutive points lying on the same side of the mean (implies a change in the mean of the process)
5 or more consecutive points going in the same direction (implies a trend)
TB2014.115 Quality Report
Page 7 of 24
Oxford University Hospitals
TB2014.115
Patient Safety
PS03 VTE Risk Assessment (% admitted patients receiving risk assessment)
Narrative
The deterioration in VTE risk
assessments coincides with
switching off of assessment
functionality within Casenotes
system on 09 April. EPR is now
the only route for assessment
(apart from NOC site where
paper is still used).
The thromboprophylaxis team
has provided quick ‘user guides’
on how to complete a VTE risk
assessment quickly. These
have been given out to all
wards where adult in-pts reside.
All clinical leads have been sent
reminders about the VTE risk
assessment target and our fall
in April. We have also asked
EPR to provide an alert at 6 hrs
of admission for all nonassessed patients. The Trust is
on track to achieve compliance
with this indicator as planned by
30 November 2014
The chart shows the proportion of inpatients within the Trust risk assessed for VTE (either individually or as part of a cohort).
The data point for the most recent calendar month may improve up until submission to NHS England as further cohorted
patients are identified following clinical coding. Earlier figures are those submitted to NHS England. [Owner: N Curry].
Patient Safety
PS07 Antibiotic prescribing - % prescriptions where indication and duration
specified [most recently available figure, undertaken quarterly]
Narrative
This is a quarterly report. No
new data available for the
quarter at time of reporting.
Each antimicrobial prescription has to have a clinical reason as to why it is prescribed along with the length of the course
written in days/doses. [Owner: Infection Control team].
TB2014.115 Quality Report
Page 8 of 24
Oxford University Hospitals
TB2014.115
Patient Safety
PS09 % patients receiving stage 2 medicines reconciliation within 24h of admission
Narrative
Introduction of an enhanced
pharmacy weekend service
for medical wards on the John
Radcliffe site has seen a 33%
improvement in compliance
with results in the mid-90%.
Of all the patients admitted on
these wards over 90% have
had medicines reconciliation
completed within 24 hours.
Recent approval via winter
pressure monies will see the
extension of this service on
the John Radcliffe and Horton
sites in October. It is expected
to have the same effect as the
introduction on the medical
wards.
The chart shows the proportion of inpatients for whom second stage pharmacy-led medicines reconciliation is completed
within 24 hours of admission. Spot check audit by pharmacy staff once per month. Approximately 600 patients are included in
the audit Trust-wide. Please note that this audit was not performed in May 2013 due to capacity issues in pharmacy. [Owner: P
Devenish].
Clinical Effectiveness
CE03 Dementia - % patients aged > 75 admitted as an emergency who are screened
[one month in arrears]
Narrative
Development and
implementation of a
Dementia Steering Group,
Development and
implementation of a
Dementia Strategy.
The Dementia Strategy
includes:
•
Education and training
for all staff,
•
the National and Local
context,
•
reporting and
monitoring, specialist
staff considerations.
Elderly patients admitted on a non-elective basis should be screened for dementia using a screening question and / or a simple
cognitive test. Performance shown in this graph reflects figures submitted monthly to NHS England. These figures are derived
from both EPR and local paper-based systems.
TB2014.115 Quality Report
Page 9 of 24
Oxford University Hospitals
TB2014.115
Clinical Effectiveness
CE04a Statutory and Mandatory Training - % required modules completed
Narrative
Individual Divisional results
and compliance are are
monitored through Divisional
performance meetings.
A paper with planned actions
and strategies has been
presented to the Qualtiy
Committee in October 2014.
Clinical Effectiveness
CE05 ED - % patients seen, assessed and discharged / admitted within 4h of arrival
Narrative
Indicator results monitored
through Divisional
performance meetings.
For further detail on
measures being taken
regarding this indicator
please see the IBP.
% Patients attending ED who are discharged or admitted within 4 hours of arrival. [Owner: EMT]
TB2014.115 Quality Report
Page 10 of 24
Oxford University Hospitals
TB2014.115
Clinical Effectiveness
CE07 Stroke - % patients accessing specialist stroke environment within 4h of arrival Narrative
This indicator is subject to
variations in the total number of
stroke admissions and the
degree of diagnostic uncertainty
on admission. The stroke team
are focussing on direct access
to the ward at weekly breech
meetings with the Emergency
department.
A specialist nurse role is being
developed to proactively identify
patients with a possible stroke
in ED to arrange early transfer
to the unit.
Other strategies employed
include options to increase
capacity on the acute stroke
unit.
Clinical Effectiveness
CE10 Vascular - Mean length of stay for patients undergoing elective AAA repair (3
month rolling period) [NOTSS Division]
Narrative
The mean length of stay
following elective AAA surgery
is below 6 days.
Information collected from ORBIT and based on the primary procedure coded and elective admission method.
TB2014.115 Quality Report
Page 11 of 24
Oxford University Hospitals
TB2014.115
Clinical Effectiveness
CE13 Cardiology - Mean number of days from referral to admission to cardiology Narrative
at tertiary centre [MRC Division]
Exception report not included in
the most recent report to the
Clinical Governance Committee.
A report to be presented to
November CGC outlining
strategies for sustainable
improvement.
Directorate goal is that patients are transferred within 2 days of referral.
Clinical Effectiveness
CE20 % patients having their operation within the time specified according to
their clinical categorisation [CSS Division]
Narrative
The data is derived from JR
emergency theatre patients. The
timing is from the time the surgeon
first asks for theatre space to the
time the patient is anaesthetised.
This data therefore reflects the full
spectrum of delays from lack of
bloods, scans, portering, theatre
time and availability of staff. The
Jan to Sept 14 data shows that
peak breaches times are between
06:00 to18:00 hrs accounts for
79% of cases.
The number of breaches has risen
since the loss of theatres 9 &10 in
Autumn 2013 which resulted in the
loss of 1 urgent bookable list per
week. The CSS Division has
been exploring options to improve
access to theatres for emergency
patients working with colleagues
in SUON Division. One of the
newly appointed Consultant
Surgeons is developing a new
protocol for access to emergency
and urgent bookable theaters.
Additional urgent bookable
capacity will be put in place from
1st October – 2 lists = 10 hours of
operating time. An action plan is
being drawn up to include
anesthetist led prioritization of lists
TB2014.115 Quality Report
Page 12 of 24
Oxford University Hospitals
TB2014.115
and validation of patients.
Clinical Effectiveness
CE21 Neuroscience Intensive Therapy Unit (NITU) readmission rate within 48 hours of Narrative
discharge [NOTSS Division]
Patient level information
suggests that the returns to
Theatre were justifiable
based on the clinical
condition of the patients.
One would not expect patients to be readmitted to NITU following discharge. The measure aims to highlight whether patients
are discharged too early. Data collected at local level and presented as number of readmissions against number of discharges.
Clinical Effectiveness
CE23 % deliveries by C-Section [C&W Division]
TB2014.115 Quality Report
Narrative
Page 13 of 24
Oxford University Hospitals
TB2014.115
This indicator result is
primarily due to the CS rate at
the JR as red for 3 months .
The average for Quarter 2 is
25.7%. The Clinical leads in
Delivery suite are reviewing
instrumental/SVD deliveries in
July in relation to the impact
this may have on the
Caesarean section rate.
This performance indicator is
regularly reviewed and
monitored through Divisional
performance and governance
structures.
Proportion of deliveries by C-Section. Other routes include normal delivery (NSVD) and instrumental delivery. Denominator is
number of women delivered (as opposed to number of neonates).
Patient Experience
PE03 Complaints Received
Narrative
See detailed narrative regarding
the details of this indicator in
section 7 of the Board Quality
Report.
The chart shows the number of new complaints received and logged by the corporate complaints department [Owner: K
Harris].
Patient Experience
TB2014.115 Quality Report
Page 14 of 24
Oxford University Hospitals
TB2014.115
PE10 % patients returning feedback forms in specialist surgery outpatients [NOTSS
Division]
Narrative
See detailed narrative regarding
the details of this indicator in
section 7 of the Board Quality
Report.
Feedback forms are available to all patients in specialist surgery outpatient departments and patients are encouraged to
complete. Data, both positive and negative, are reviewed to identify area of good practice and areas for improvement. Data
are collected locally on number of forms completed and returned against total number of outpatients per month.
3.
Patient Safety and Clinical Risk
3.1. Information relating to patient safety and clinical risk is provided within the key
quality metrics.
3.2. 5 Serious Incidents Requiring Investigation (SIRI) reports were recommended
to Oxfordshire Clinical Commissioning Group (OCCG) for closure during
October 2014.
3.3. Following internal closure of a SIRI report, the report is presented to the OCCG
for agreement and endorsement of both the level and quality of the investigation
and the appropriateness of the recommendations to prevent a re-occurrence.
3.4. Based on theNHS England Serious Incident Framework, SIRI investigaitons are
categorised into two levels. A category 2 investigation cannot be closed until
such time as all the recommended actions have been completed.
3.5. Table 2 below outlines the SIRI’s that have been provided to the OCCG for
closure of the investigation and report.
Table 2
SIRI ref
Division
Description
2014/036
MRC
Guidewire of Uncertain origin
2014/031
S & O/NOTSS
Elective Hip Surgery
2014/033
NOTSS
Hospital Acquired Cat 3 Pressure Ulcer
TB2014.115 Quality Report
Page 15 of 24
Oxford University Hospitals
TB2014.115
2014/034
MRC
Hospital Acquired Cat 3 Pressure Ulcer
2014/035
MRC
Hospital Acquired Cat 3 Pressure Ulcer
3.6. Themes and key learnings from closed SIRI reports are presented to the Quality
Committee and the Clinical Governance Committee on a monthly basis.
3.7. Table 3 below provides a list of the 4 SIRI’s that have been declared during
October 2014. Investigations have commenced and will be reported in due
course.
Table 3
4.
SIRI ref
Division
Description
2014/042
MRC
Medication incident HGH
2014/043
MRC
Hospital Acquired Cat 3 Pressure Ulcer
2014/044
NTOSS
Hospital Acquired Cat 3 Pressure Ulcer
2014/045
MRC
Hospital Acquired Cat 3 Pressure Ulcer
Quality Walk Rounds
4.1. There were 5 quality walk rounds in October 2014. These are detailed in figure
3 below.
Figure 3.
Hospital Site
John Radcliffe Hospital
Nuffield Orthopaedic Centre
Areas Visited
SEU Ward 6F
West Wing Theatres
Maternity Assessment Unit and Observation Area
Heart Centre Outpatients Department
E Ward
4.2. Key issues with the potential to affect quality or patient experience identified
during the Quality Walk Rounds included concerns regarding recruitment and
retention, transport & transfer between hospital sites, funding for new
equipment and the condition of the environment
4.3. All issues have actions associated with them and these will be monitored
through Divisional governance processes.
5.
Clinical Effectiveness
5.1. The Clinical Audit Committee did not meet in the reporting month. This
Committee meets bi-monthly therefore a meeting is scheduled for early
November 2014.
TB2014.115 Quality Report
Page 16 of 24
Oxford University Hospitals
TB2014.115
5.2. The latest Summary Hospital-level Mortality Indicator (SHMI) was published on
the 23 October 2014. This SHMI value, for the data period April 2013 to March
2014, is 0.99. This value is described as ‘as expected’ using the Health and
Social Care Information Centre (HSCIC) 95% confidence intervals adjusted for
over-dispersion. When compared to the previous SHMI release, for the data
period January 2013 to December 2013, this is an increase of 0.01.
5.3. The latest Hospital Standardised Mortality Ratio (HSMR) was published by Dr
Foster on the 10th October 2014. This HSMR value, for the data period July
2013 to June 2014, is 90. Dr Foster has advised that the anticipated rebasing
of data has been delayed to November 2014.
5.4. The current advisory from Dr Foster is that from November 2014 the data will be
rebased on a monthly basis and Dr Foster will be moving from using the
Secondary Uses Services (SUS) to Hospital Episodes Statistics (HES) data.
5.5. The in-depth analysis of the SHMI and HSMR is currently being completed by
the Clinical Governance team. A report will be submitted to the Mortality Review
Group for review at their meeting on the 20 November 2014.
5.6. The Clinical Outcomes Review Committee met on the 10 October 2014. The
Outcome audits presented were:
• National Adult Cardiac Surgery audit and consultant outcomes
(April 2010 - March 2013 data) Findings: Overall consultant outcomes
were good in the context of increasing complexity of patient case mix and
procedures.
•
Oxford Bone Marrow Transplant outcomes Findings: There were
improved outcomes across all procedures. There were no concerns with
mortality rate. There were issues highlighted with theatre utilisation and
availability of intensive care beds. The action plan was acknowledged by
the Committee and it was recognised that the action plan would need
Divisional support. The unit has started weekly Mortality and Morbidity
meetings to ensure that there is a contemporaneous review of cases.
•
Inflammatory Bowel Disease - Inpatient care and experience (adult
JR and HGH), January - December 2013 data Organisational audit
(adult), 3rd February - 31st March 2014 data Biologics (adult JR and
HGH), 12th September 2011 to 28th February 2014 data Findings: There
were issues with data submission due to a lack of administration support
and an imposed requirement to enter c.200 data points per patient into
the web tool.
The Committee recommended that the local data
spreadsheet be submitted directly to the national clinical audit organisers,
since this contains the data required but has been populated in an
efficient and simple manner compared to the high demands of using the
web tool. In other respects, excellent compliance with national standards
was noted.
TB2014.115 Quality Report
Page 17 of 24
Oxford University Hospitals
TB2014.115
5.7. The Clinical Outcomes Review Committee considered the results of a review of
the national clinical audit external data submissions at OUH. This will be
forwarded to Divisional and Clinical Directors for review and validation, and
actions taken forward through Clinical Governance Committee. The Committee
is investigating options to make collation of clinical audit data more simple and
robust.
5.8. The Mortality Review Group met on the 13 October 2014. The Group reviewed
the findings of the Dr Foster SHMI Insight Report (January to December 2013
data).
5.9. The Mortality Review Group is undertaking work to standardise and improve the
OUH Mortality Review process. Findings of this work will be brought to Clinical
Governance Committee and Quality Committee in November and December.
5.10. The Mortality Review Group discussed the Faculty of Public Health report
(Appendix 1) on hospital mortality rates and commended the report to the Trust
Board for information. In the report the Faculty advises that hospital mortality
rates should not be used to:
• Compare the quality of one hospital to another, for example, in league
tables.
•
•
Attribute ‘preventable deaths’ to individual hospitals.
Falsely assume that a low or ‘within expected limits’ mortality ratio implies
good quality of care and overlooks clinical or organisational failings that
are causing harm to patients.
•
Obsessively search for lessons learnt from ‘preventable deaths’ at the
expense of not learning from the many more cases where patients
survived.
•
Assume that there are such things as ‘good’ hospitals and ‘bad’ hospitals.
Most hospitals are large complex organisations with both good and bad
elements across different departments and sites.
•
It was noted that hospital mortality rates ignore external pressures.
•
Further, Trusts should:
o Use an evidence-based, structured approach to review all hospital
deaths.
o Pay more attention to staffing levels.
o Pay more attention to community health and social care factors to
ensure there are adequate means of caring for terminally ill
patients, and relieve pressure on the ‘pinch points’ of accident and
emergency and elderly care wards, and to prevent unnecessary
admission in the first instance.
5.11. As agreed by Trust Management Executive in July 2014, a review of the Terms
of Reference for the meetings addressing the Clinical Effectiveness and
Outcomes function of the Quality Strategy has been completed. Revised Terms
TB2014.115 Quality Report
Page 18 of 24
Oxford University Hospitals
TB2014.115
of Reference of key sub-committees of the Clinical Governance Committee are
currently progressing through the committee structures with a view to approval
at the CGC in December 2014.
6.
Quality Account and Quality Strategy
6.1. Progress against the organisational quality priorities as set out in the 2013/14
Quality Account was reported to the Quality Committee in October. A review
and refresh of the Quality Strategy, the relationship between Trust Quality
Priorities and the Strategy is proposed. This will be reported to the Trust Board
in January 2015.
7.
Experience of Patients
7.1. The Patient Experience Team has produced a dashboard for Trust Board
(Appendix 2). This includes Friends and Family Test (FFT) data, complaints
activity, management of complaints, PALS and compliments. The intention of
the dashboard is to provide a Trust wide overview to support divisional analysis.
7.2. Friends and Family Test. Inpatient, ED and Maternity response rates:
•
The national comparator FFT results for September 2014 were not
available at the time of writing this report. This means that comparison
with Shelford Group could not include results for September 2014.
•
Inpatients: Overall there is limited change in the FFT score and the
percentage of extremely likely and likely responses. The Children’s and
Womens (C&W) division score has returned to the previous level
following a variance. This fluctuation is likely to can be attributed to a low
number of respondents, which can affect reliability of the data.
•
Emergency Department: The FFT score for this area has identified that
the percentage of patients who are either extremely likely or likely to
recommend the service has decreased in September.
•
Maternity: The FFT score of patients either extremely likely or likely to
recommend the service for maternity remains similar to the previous
report. The response rate has increased in August and the improved
response rate has been maintained in September. This has resulted in
the Maternity Service receiving the star of the month for patient
experience reporting.
7.3. Friends and Family Test. CQUIN status:
•
All targets have been achieved in this financial year.
FFT was
implemented in Outpatients and Day cases ahead of the 1 October 2014
TB2014.115 Quality Report
Page 19 of 24
Oxford University Hospitals
TB2014.115
deadline which has resulted in achievement of the associated CQUIN.
The Trust is recording feedback on paper questionnaires initially but will
implement text messaging for some services.
•
The patient experience team have developed a project plan to increase
response rates in the emergency departments and inpatients by quarter
4. The aim is to achieve the target response rate for emergency
departments is 20% over quarter 4 and the target response rate for
inpatients is 30%. The strategies to improve uptake include team walk
through and a social media campaign which will publicise the feedback
received and changes made.
7.4. Friends and Family Test. Complaints:
•
S&O received the highest number of formal complaints in September, this
increase relates to access primarily in the Renal, Transplant and Urology
Directorate.
•
MRC received the second highest number of formal complaints in
September, there has been a sustained improvement in the number of
NOTSS complaints over the last 4 months.
•
The issue of access to Outpatients continues to be a recurrent theme for
complaints received by all Divisions however there have been some local
improvements in both eye and ENT outpatients related to the day of
appointment. Care/nursing care also features highly as a theme for all
clinical divisions.
7.5. Managing complaints
•
The Trust continues to meet the target of 95% for acknowledgement of
complaints and this month is the first time the trust has achieved 100% in
September 2014 for acknowledgement of the complaint within 3 days.. In
addition the Trust achieved 97% for responding to complainants within 25
working days or agreed timescale against a target of 95%.
7.6. National context
•
NHS England’s Patient Experience Team have been asked to explore the
options available in terms of surveys of complainants in the NHS. This
work forms part of the Department of Health’s Programme Board for Hard
Truths (relating to complaints).
•
The Trust is participating in the national complaints benchmarking
scheme jointly run by the NHS Benchmarking network and the Patients
Association. The NHS Benchmarking Network was established in 1996 in
TB2014.115 Quality Report
Page 20 of 24
Oxford University Hospitals
TB2014.115
response to a need for NHS organisations to work together to improve
services rather than to continually “reinvent the wheel”.1
8.
Infection Control
8.1. Legionella continues to be identified from outlets on the John Radcliffe. It must
be noted that all high risk outlets (such as showers) where legionella counts are
reported are immediately taken out of use.
8.2. The positive counts continue in the JR2 building where refurbishment works
continue on the showers. Blenheim head and Neck ward in the Churchill
retained Estate now only has one shower out of use due to persistent positive
counts, even though it has been refurbished. In view of this, OUH Estates are to
trial a water treatment system within this area to try and eradicate the persistent
counts.
8.3. Clostridium Difficile - The OUH Trust has an upper limit of 67 cases for 2014 –
2015. The OUH Trust had 6 cases of C. diff from samples taken after three
days of admission. All cases are reviewed at the Monthly Health Economy
meeting, which is held with the OCCG, PHE. Oxford Health and OUH in
attendance.
8.4. Table 4 below outlines the number of cases per month that are apportioned to
the OUH Trust.
Table 4
JR,
Churchill,
NOC
Monthly Limit
Cumulative total
Cumulative limit
April 2014
May 2014
June 2014
July 2014
August
2014
September
2014
1
6
7
6
3
6
5
1
5
5
7
10
5
14
15
6
20
21
6
23
27
6
29
33
8.5. MRSA Bacteraemia & Screening Compliance
•
The OUH has a limit of 0 avoidable MRSA Bacteraemia for 2014 – 2015,
there have been zero avoidable MRSA bacteraemia apportioned to the
OUH financial year to date.
•
There was 1 MRSA Bacteraemia apportioned to the OUH in September
2014, however it was agreed at the Post-Infection Review meeting with
the OCCG that this was Unavoidable.
1
Source http://www.nhsbenchmarking.nhs.uk/partnership-projects/ComplainantsSurvey.php
TB2014.115 Quality Report
Page 21 of 24
Oxford University Hospitals
•
TB2014.115
The trust achieved an average of 61% compliance with MRSA screening,
81% for elective admissions and 55% for emergency admissions. Clinical
areas with high turnover of patients have lower compliance with
screening of emergency admissions.
8.6. Cleaning Standards
•
Clinical areas are required to achieve a minimum 92% Compliance with
the monthly cleaning audit Table 5 below outlines the cleaning audit
compliance results for September 2014.
Table 6
Division
Quality Assurance Team
audits
Neurosciences,
Orthopaedics, Trauma
& Specialist Surgery
Medicine,
Rehabilitation &
Cardiac
Children’s and
Women’s
Surgery & Oncology
Clinical Support
Services
OUH total
9.
September 2014
Domestic audit
scores
Nursing audit
scores
86%
93%
97%
90%
92%
96%
91%
91%
94%
94%
99%
92%
93%
90%
92%
93%
94%
96%
Safe Staffing Levels
9.1. The Trust is required to comply with NICE guidance, which includes providing
reports to the Trust Board on the levels of nursing and midwifery staffing on a
ward by ward/shift by shift basis.
9.2. The first report to the Board on staffing levels was presented in March 2014 and
this has since been refined to be included within the Board Quality Report.
9.3. National reporting
•
The summary of the figures submitted to NHS Choices via the Unify
platform for August and September 2014. This report includes the actual
hours worked against the planned roster hours for nursing and midwifery
staff. The report also separates registered nurses and unregistered Care
Support Workers. Table 7 below reflects the figures submitted to NHS
Choices.
Table 7
Registered Staff
Un- registered Staff
TB2014.115 Quality Report
August
91.80%
92.17%
September
93.68%
94.16%
Page 22 of 24
Oxford University Hospitals
TB2014.115
•
The levels in August reflect the impact of the holiday period when
temporary staff availability is less.
•
The report submitted by the trust can be viewed on the NHS Choices
website which is linked to the Trust’s own website (Nursing & Midwifery
Safe Staffing). The information is by hospital site, day and night shifts,
and includes all staff that has worked i.e. permanent and temporary
staff. http://www.ouh.nhs.uk/about/saferstaffinglevels.aspx
9.4. Trust safe staffing management
•
The Trust reviews staffing levels via the ward monitoring system which is
reported and scrutinised daily. A Red, Amber, Green rating (RAG) is
applied and reviewed at twice daily staff and bed capacity meetings held
on each hospital site. These meetings address ‘short notice’ shortfalls
and addresses the needs of areas of increased acuity and activity.
•
The meeting ensures awareness and review of shifts that require
‘escalation’ in order to allow for mitigation and support to those clinical
areas. During the August and September reporting period there has been
an increase in the number of escalation and minimal safe shifts due to
lack of availability of temporary staff to fill shifts during the holiday period.
Those reported as minimal safe shifts do indicate less resilience against
issues such as sick leave and rapid changes in acuity if
sustained.(Appendix 3).
•
There are a number of vacant posts within in-patient areas which has
affected levels of staffing. A coordinated recruitment program is in
progress nationally and internationally. Retention of staff is proving
challenging within in some ward areas with high turnover. Strategies are
being progressed reflecting the feedback from the staff survey and the
Recruitment and Retention Summit held in July 2014 in the Trust.
9.5. Acuity and Dependency Review
•
The Acuity and Dependency review of staffing levels in May/June 2014
was presented to the September 2014 Trust Board. This review is
undertaken six monthly, those areas of highest priority were presented to
the Trust Management Executive and agreed in principle and will be
subject to a business case.
10. Recommendation
10.1. The Board is asked to receive this report.
Dr Tony Berendt, Medical Director
Catherine Stoddart, Chief Nurse
Report prepared by:
TB2014.115 Quality Report
Page 23 of 24
Oxford University Hospitals
TB2014.115
Annette Anderson, Head of Clinical Governance
29 October 2014
TB2014.115 Quality Report
Page 24 of 24
Hospital mortality rates
Position statement
Implications for public health
The Faculty of Public Health (FPH) is committed to ensuring that
hospital care is safe, effective and provides the highest possible
standard of clinical care. The first attempt to categorise hospitals by
their mortality rates and make imputations about the quality of care
delivered in them based on how many patients die, dates back to the
seminal work of Florence Nightingale (best known for dramatically
decreasing hospital deaths in the Crimea by improving sanitation) and
the statistician William Farr. Together, they compiled the first ever
‘league table’ of hospital mortality rates in 1863. Their efforts were
widely criticised at the time, and many of the problems with their
methodology have never been satisfactorily resolved. It is essential to
understand the controversies around hospital mortality data in order to
prevent the significant harm that can come from their misuse.
Counting deaths and creating mortality ratios
The question of which deaths to count is more complex than it appears at first
glance. There is no agreement about the time frame that should be used (eg.
only deaths while in hospital or also for some period after discharge), which
† Words in bold are defined on page 4
departments within the hospital to include, whether deaths from all conditions
or from selected conditions should be counted, and which hospital the death
should be attributed to when a patient dies after transferring between hospitals.
KEY CONTROVERSIES
Furthermore, because hospitals are of such widely varying size and complexity,
comparing absolute numbers of deaths is meaningless. As hospitals which tend
to admit sicker patients will have higher crude mortality rates (†), most
1. Counting deaths and creating
mortality ratios
published statistics today present a standardised mortality ratio, which is
2. Shortcomings in risk adjustment
calculated as the number of ‘observed deaths’ divided by the number of
3. Can hospital mortality statistics
be used to identify hospitals with
poor quality care?
‘expected deaths’ after taking into account differences in age, sex and disease
across hospitals (a statistical process called ‘risk adjustment’). However, just as
there is no agreement over which deaths to count, neither is there any ‘gold
standard’ formula to calculate the number of expected deaths. Various
organisations choose different ways to count observed deaths and use different
methods of risk adjustment, for example Dr Foster Intelligence produces the
4. Can use of mortality statistics
cause harm?
5. How should we use hospital
mortality data?
hospital standardised mortality ratio (HSMR), Health and Social Care
Information Centre (HSCIC) the summary hospital mortality indicator (SHMI)
and Caspe Healthcare Knowledge Systems (CHKS) the Risk Adjusted Mortality
Index (RAMI); not surprisingly they can produce wildly different estimates of
mortality rates. One study, which compared four different methods across 83
hospitals in America, found that of 28 identified as the ‘worst’ mortality
hospitals by one company, 12 appeared in the ‘best’ category when other
methods were used1.
Patients die in hospital for many reasons, of which quality of hospital care is
only one. Others include lack of availability of hospice beds, inadequate primary
care, inadequate care-home provision and even patient choice. The hospital has
no control over any of these external factors yet they can all result in increased
numbers of deaths, which increases the ratio of observed to expected deaths,
and hence the HSMR or SHMI. Changes in community health services can have
a big effect on ‘hospital’ mortality rates – for instance when Walsall opened a
1
hospice, the nearby hospital’s HSMR and SHMI fell
classification codes. For instance using blood test results,
sharply2.
blood pressure, weight etc. might create a fuller picture of
the patient's severity of disease on admission to hospital,
Shortcomings in risk adjustment
Whenever a patient is discharged from or dies in hospital,
data about their diseases and any operations performed
are summarised using classification codes and submitted
as hospital episodes statistics (HES data) to a national
database. Major problems are lack of consistency in coding
and lack of reliability, because all we have in order to
make predictions about a patient’s likelihood of dying,
based on similar patients in the past, are a handful of
codes summarising the reason for their admission, eg.
“J18.0 Broncho-pneumonia, unspecified”. Studies of
clinical coders have found considerable variation in
choice of codes and coding depth. This means expected
deaths, and hence mortality ratios, can vary substantially
depending on how patients have been coded3. Dr Foster
has acknowledged that a hospital recording an average of
and potentially their likelihood of dying.
However, this introduces other problems: firstly we don't
fully know which data are best at predicting severity of
disease; secondly different hospitals can collect different
and non-comparable data, making comparisons between
hospitals impossible; thirdly information systems are often
not set up to draw data from many different sources
together. The more data that is required to predict
patients’ severity of disease, the greater the likelihood that
certain bits of information are not collated.
In the case of children's heart surgery in Leeds, the simple
failure to consistently supply children's weight to the
calculations affected these to such an extent that the unit’s
mortality rate appeared to be double the national average.
This incorrect calculation was then used as part of the
justification of a proposal to close the unit9.
2.5 codes per patient would have an HSMR around 15-20
points higher than one recording 5.5-6 codes per patient4.
There are perverse financial incentives to increase coding
depth too, since if more codes are recorded hospitals can
charge their commissioners more as their patients will
appear to be sicker. A similar phenomenon of over-coding
has been observed in the United States5.
Patients whose admission includes a palliative care code
are effectively considered ‘very likely to die’ by Dr Foster’s
calculations, and therefore these patients have a profound
effect on HSMR. Some patients receive palliative care even
when they are not at imminent risk of death, yet the
palliative care code, and hence HSMR, cannot distinguish
these two categories of patients. (SHMI calculations of
‘expected deaths’ are not nearly so influenced by palliative
care coding but have other limitations). National guidance
Can hospital mortality statistics be used
to identify hospitals with poor quality
care?
It is sometimes claimed that, provided case mix has been
properly adjusted for, any remaining variations in mortality
ratios between hospitals must be a result of variations in
quality of care. However, studies have consistently
demonstrated that even after risk adjustment, there is no
clear relationship between a hospital’s standardised
mortality ratio and its quality of care10. This is because:
• Risk adjustment is never perfect, especially given the
limitations of HES (which was originally intended for use
in finance and planning and not for predicting mortality)
and known variations in coding practice
on when palliative care codes should be applied is
notoriously ambiguous6, and, because of this, hospitals
vary enormously in their coding practice, which ranges
7
from 0 to 44% of deaths .
Even if a universally consistent standard of coding could be
achieved, this does not eliminate the problems of adjusting
for risk as the probability of a patient dying from a given
condition is not the same throughout the country (the socalled ‘constant risk fallacy’8). In addition, patient lifestyle
factors such as smoking and alcohol are not recorded in
HES data. This means that calculating risk of dying based
on patients’ lifestyles has to use proxy measures instead,
• Hospitals are large and complex organisations that will
inevitably have a mixture of outstanding care and areas
of weakness which counterbalance one another
• Quality is a complex concept and includes many more
factors than merely the risk of death. The NHS definition
of quality in healthcare encompasses four domains:
preventing illness, safe, effective and personal11. Some
of the problems exposed in the Keogh Review and the
Francis Inquiry were serious violations of patient
dignity but by themselves were unlikely to kill any
patients.
and this is chiefly done using postcode as a (very
Careful auditing of case records has found that only
problematical) proxy for deprivation, and hence lifestyle,
around one in every 20 deaths in hospital has any factors
risk factors.
that might have impacted on the inevitability of the patient
dying, in other words whether it was a ‘preventable
One response to the inherent shortcomings of using
administrative coding data to predict the risk of dying is to
use clinical parameters as well as or instead of
2
death’12. Deaths due to failings in care reflect an incredibly
small proportion (around 0.15%, or one in six hundred) of
all admissions, and it is perfectly possible for a hospital to
have a low HSMR whilst nevertheless offering poor quality
limits have little incentive to ask themselves whether
care.
there is more they could be doing to improve patient
safety and quality of care. For example faced with a high
Some people advocate the use of hospital mortality data
as a ‘smoke alarm’ that can potentially identify hospitals
with problems. However, using this data as a screening
test to identify poor quality hospitals results in frequent
‘high mortality alerts’ that require significant resources to
investigate but which most of the time do not result in
findings of poor quality care. One study estimated that in
91% of hospitals investigated as a result of their high
mortality rates, the alert would be a false alarm13.
Meanwhile, hospitals with normal or low mortality data
provide no reassurance that the quality of care is
HSMR, Stafford Hospital first examined their coding and
realised patients were seriously under-coded. When they
increased their coding depth, their HSMR fell to below
average, prompting Dr Foster to praise them for being
one of the ‘five most improved’ hospitals in the
country16. Yet at the very same time, the Healthcare
Commission was uncovering widespread problems in
care. Similarly, Sir Bruce Keogh acknowledged that
many of the problems uncovered in the hospitals he
reviewed during 2013 could well be happening in other
hospitals too. More recent inspections of hospitals with
acceptable. If hospital mortality rates are to be likened to
a smoke alarm, they are the faulty sort that have a
apparently normal or low HSMRs appear to be
confirming this17, 18.
tendency to go off whenever the hot tap is run, but fail to
• Credibility. The expression “hospital mortality rates”
sound when the frying pan has caught fire.
sounds so integral to hospitals that explanations of high
Can use of hospital mortality statistics
cause harm?
• Institutional damage. Significant stigma can follow
14
the publication of an adverse mortality statistic . This
rates that appeal to external factors such as availability
of hospice beds or quality of primary care and nursing
home care may sound unconvincing to anyone who has
not understood how these statistics are actually
calculated.
can cause a loss of confidence in patients, demoralise
staff and inhibit recruitment. Perversely this could
potentially happen to a good quality hospital with a high
How should we use hospital mortality
data?
HSMR, resulting in a fall in quality of care. False alarms
can be expensive to investigate.
Ways in which hospital mortality statistics must not
be used:
• Distress and confusion to the general public. The
media frequently use mortality data completely
inappropriately, for example to confuse the number of
• To compare the quality of one hospital to another, eg.
league tables.
‘excess deaths’ with the number of ‘preventable deaths’
• To attribute ‘preventable deaths’ to individual hospitals,
or to imply that the mortality rate is a useful proxy
• To falsely assume that a low or ‘within expected limits’
measure for quality of care. By definition, approximately
mortality ratio implies good quality of care and overlook
half of all hospitals will have more than average deaths
clinical or organisational failings that are causing harm
and half will have fewer. To claim that the highest
to patients.
mortality hospitals have thousands of ‘excess deaths’
makes no more sense than to suggest that hospitals
with the lowest mortality rates have sent thousands of
people home alive who really should have died. It also
• To only focus attention on hospitals when attempting to
interpret hospital mortality statistics, instead of also
considering the impact of external factors such as
community pressures or hospice facilities.
conceals the reality that ‘preventable deaths’ happen
in all hospitals.
• To obsessively search for lessons learnt from
‘preventable deaths’ at the expense of not
• Conflicts of interest. A number of companies sell
analysis tools to hospitals on the premise that these
enable the hospitals to analyse which patients are
contributing to the mortality figures. The vested interest
in defending the putative link between mortality and
quality of care was acknowledged in the Francis
learning from the many more cases where patients
survived.
• To assume that there are such things as ‘good’ hospitals
and ‘bad’ hospitals. In reality, most hospitals are large
complex organisations with both good and bad elements
across different departments and sites.
Inquiry15.
• Distracting from genuine problems. Hospitals with
Better ways to investigate hospital quality of care:
high HSMRs are expected to examine their coding and
clinical systems to try to find an explanation for the high
• An evidence-based, structured approach to review all
figures, which may miss important quality problems that
hospital deaths. One example of this is being developed
harm patients but do not cause death. On the other
during 2014 by the London School of Hygiene and
hand, hospitals whose figures are low or within normal
Tropical Medicine.
3
• A much more consistent, rigorous and questioning
Appropriate use of hospital mortality data:
approach to inspecting hospitals. Every NHS hospital
should experience as thorough a review as the process
undertaken by Sir Bruce Keogh’s team. The process
conducted in the 14 Keogh-review hospitals could easily
be adopted as a template for periodic inspection of every
• FPH does not support the use of risk-adjusted mortality
statistics to compare or rank hospitals, nor does it believe
they should be used to estimate ‘preventable deaths’.
• If hospital mortality data is to continue to be published,
hospital, with economies of scale if this were
we need a far more explanatory communication strategy
undertaken.
to accompany the data and outline its serious
• Pay more attention to staffing levels, one of the factors
19
identified by the Berwick report . A consistent criticism
weaknesses and limitations.
• There is an important place for the analysis of certain
of Stafford was that it was under-staffed by doctors,
categories of deaths such as maternal deaths or peri-
nurses and other healthcare professionals.
operative deaths. However, the most effective way to
• Pay far more attention to community health and social
care factors to ensure there are adequate means of
caring for terminally ill patients, and
more generally to relieve pressure on the ‘pinch points’
analyse this data is the existing Confidential Inquiry
which seeks to learn and disseminate lessons without
‘naming and shaming’ individual hospitals or their staff.
• There may be a role for mortality data in research, for
of A&E and elderly care wards, and to prevent
instance to explore how well it conveys information
unnecessary admission in the first place.
about factors external to the hospital which might help
target resources more appropriately.
DEFINITIONS
HSCIC Health and Social Care Information Centre. Source of
HES data and calculate SHMI.
Berwick Review A review of safety in the NHS by the
American Prof Don Berwick following the Francis Inquiry.
HES data Hospital episodes statistics – the principal national
database which is used to calculate mortality statistics.
CHKS A private company that provides a hospital
benchmarking service and calculates the RAMI.
HSMR Hospital standardised mortality ratio. Dr Foster’s
calculation takes deaths in hospital other than in Accident
and Emergency departments, for the 56 most common
medical conditions, and divides by the number of deaths
it expects for the hospital, based on factors including
age, sex, other medical conditions and local deprivation.
Patients who die after being transferred between two
hospitals are counted twice over.
Classification codes Two systems are in use which classify
patients according to why they came into hospital: the
World Health Organization’s 10th International
Classification of Disease (ICD-10), which lists medical
conditions, and the Office of Population Censuses and
Surveys Classification of Interventions and Procedures
(OPCS-4) which describes operations.
Clinical coders Hospital administrators who summarise a
patient’s medical history and reasons for admission into
codes which are entered into a national database for
purposes of health service finance and planning.
Coding depth The number of clinical codes entered per
patient. Can be up to about a dozen but in practice 3-6.
Hospitals that enter more codes per patient may be more
diligent but this can result in patients appearing sicker
and thus more expensive to treat and considered more
likely at die in risk adjustment formulae, which can then
underestimate the hospital’s standardised mortality ratio.
Crude mortality rate The number of deaths over a period of
time divided by the number of patients admitted over the
same period. It fails to account for differences in patient
characteristics between hospitals.
Dr Foster Intelligence The company that calculates HSMR.
Excess death A statistical calculation of the number of
additional deaths relative to the national average that
occur in hospitals whose standardised mortality ratio is
higher than average. One could imagine a hypothetical
opposite, negative deaths, or excess deaths which did not
happen in hospitals with lower than average mortality
rates.
Francis Inquiry A series of two major inquiries, the first
independent, the second public, into failings in care at
Stafford Hospital.
4
Keogh Review Commissioned after the second Francis
Inquiry was published, this was a detailed inspection of
the quality of care in 14 hospitals identified as having
either a high HSMR or a high SHMI.
Perioperative death A death that occurs during or shortly
after surgery.
Preventable death In contrast to ‘excess death’, a
preventable death is one that is considered, after careful
review, to be a result of failings in medical care.
Preventable deaths can occur in any healthcare setting
and there is little evidence to suggest they are more
common in hospitals with a high standardised mortality
ratio.
RAMI Risk Adjusted Mortality Index – another method of
expressing standardised hospital mortality ratios,
calculated by CHKS.
SHMI Summary hospital mortality indicator. Calculated by
HSCIC, this includes all deaths in patients admitted to
non-specialist hospitals and deaths within 30 days of
discharge, divided by the number of expected deaths
adjusting eg. for age, sex and diagnosis, but does not
make adjustments for deprivation or palliative care.
Standardised mortality ratio The number of ‘observed
deaths’ over a period of time divided by the number of
‘expected deaths’ that might potentially have occurred
had the hospital had the same pattern of deaths as the
national average for the type of patients admitted.
Acknowledgements:
• Finally, hospital mortality data can be used for internal
quality control and audit, since this is not coloured by
Author: Dr David Pitches
internal system variations such as coding. Unfortunately
With thanks to FPH’s Health Services Committee
this sounds less compelling than “your hospital has a
higher than expected mortality rate, so there must be a
May 2014
problem and you need to go and fix it”20, but this was
how Florence Nightingale succeeded in improving
outcomes for soldiers in the Crimea, whereas her later
attempt at putting hospitals into a ‘league table’ was
fundamentally flawed.
REFERENCES
1 Shahian DM, Wolf RE, Iezzoni LI, Kirle L et al. 2010.
Variability in the Measurement of Hospital-wide Mortality
Rates. New England Journal of Medicine. 363,2530-9.
2 Mott MacDonald. 2013. Independent Review of Mortality
Rates at the Manor Hospital – Final Report.
3 Audit Commission. 2008. PbR Data Assurance
Framework 2007/08.
4 Aylin P, Bottle A, Jarman B. 2009. Monitoring hospital
mortality – a response to the University of Birmingham
report on HSMRs. The Dr Foster Unit at Imperial.
5 Simborg DW. 1981. DRG creep: a new hospital-acquired
disease. New England Journal of Medicine. 304(26), 16024.
6 Health and Social Care Information Centre. 2013. The
Use of Palliative Care Coding in the Summary Hospitallevel Mortality Indicator.
7 HSCIC. 2014. Dataset: Percentage of provider spells
with palliative care coding, July 2012 – June 2013.
Available at http://tinyurl.com/jwafr47
8 Nicholl J. 2007. Case-mix adjustment in nonrandomised observational evaluations: the constant risk
fallacy. Journal of Epidemiology and Community Health
61, 1010-1013.
9 Does Leeds General Infirmary have a death rate "twice
the national average"? Available at
http://tinyurl.com/nzcdstn
10 Pitches DW, Mohammed MA, Lilford RJ. 2007. What is
the empirical evidence that hospitals with higher risk
adjusted mortality rates provide poorer quality care?
Biomed Central Health Services Research. 7, 91.
11 Darzi. 2008. High Quality Care For All. NHS Next Stage
Review Final Report.
12 Hogan H, Healey F, Neale G et al. 2012. Preventable
deaths due to problems in care in English acute hospitals:
a retrospective case record review study. BMJ Quality and
Safety. 21, 737–745.
13 Girling AJ, Hofer TP, Wu J, Chilton PJ et al. 2012.
Case-mix adjusted hospital mortality is a poor proxy for
preventable mortality: a modelling study. BMJ Quality and
Safety. 21(12), 1052-6.
14 Lilford R, Mohammed MA, Spiegelhalter D, Thomson R.
2004. Use and misuse of process and outcome data in
managing performance of acute medical care: avoiding
institutional stigma. Lancet. 363(9415), 1147-54.
15 Francis R. Report of the Mid Staffordshire NHS
Foundation Trust Public Inquiry Volume 1: Analysis of
evidence and lessons learned (part 1).
16 BBC News. 2009. Top ranking for criticised trust.
Available from
http://news.bbc.co.uk/1/hi/england/staffordshire/8385834
.stm
17 BBC News. 2014. Bradford Royal Infirmary “cannot
provide safe care”. Available from
http://www.bbc.co.uk/news/uk-england-leeds-25648677
18 BBC News. 2014. Wexham Park Hospital “failed to
protect patients’ safety”. Available from
http://www.bbc.co.uk/news/uk-england-berkshire25648481
19 National Advisory Group on the Safety of Patients in
England, Department of Health. 2013. A promise to learn
– a commitment to act. Improving the Safety of Patients
in England.
20 Shojania KG. 2012. Deaths due to medical error:
jumbo jets or just small propeller planes? BMJ Quality and
Safety. 21, 709-712.
PRODUCED BY THE FACULTY OF PUBLIC HEALTH
The Faculty of Public Health (FPH) is the standard-setting body for specialists in public health in the UK. FPH is the
professional home for more than 3,200 professionals working in public health. Our members come from a range of
professional backgrounds (including clinical, academic and policy) and are employed in a variety of settings, usually
working at a strategic or specialist level.
FPH is a joint faculty of the three Royal Colleges of Public Health Physicians of the United Kingdom (London,
Edinburgh and Glasgow). In addition, FPH advocates on key public health issues and provides practical information
and guidance for public health professionals, aiming to advance the health of the population through three key
areas of work: health promotion, health protection and healthcare improvement.
PRODUCED BY: Faculty of Public Health, 4 St Andrew’s Place, London NW1 4LB • t: 020 3696 1452
• e: policy@fph.org.uk • w: www.fph.org.uk • Registered charity no: 263894
TB2014.115 Appendix 2 Patient experience dashboard
Friends and Family Test (FFT): Trust overview
FFT: Inpatients by division
OUH and Shelford Group FFT scores
FFT Scores: inpatients by division
80
70
60
73
90
67
85
58
85
80
80
76
75
70
50
66
65
40
Apr-14
May-14
Jun-14
Jul-14
Aug-14
60
Sep-14
OUH inpatients
Shelford Group inpatients
OUH ED
Shelford Group ED
OUH maternity
Shelford Group Maternity
55
50
Apr-14
The FFT score is calculated as follows: Proportion of respondents who would be extremely
likely to recommend (response category: “extremely likely”) MINUS Proportion of
respondents who would not recommend (response categories: “neither likely nor unlikely”,
“unlikely” & “extremely unlikely”). Likely responses are included in the denominator but not
the numerator.
May-14
Jun-14
Jul-14
Aug-14
Sep-14
FFT inpatients: Response rates by division
45%
40%
OUH and Shelford Group Response rates
35%
50%
25%
25%
21%
20%
30%
40%
25%
30%
28%
23%
13%
20%
10%
20%
15%
10%
5%
0%
Apr-14
May-14
OUH inpatients
Jun-14
Jul-14
Aug-14
0%
Sep-14
Apr-14
Shelford Group inpatients
OUH ED
Shelford Group ED
OUH maternity
Shelford Group Maternity
Higher response rates mean data are more reliable: we can be more confident that the
scores are representative of the population. Response rates in maternity increased in
August and maintained in September.
FFT: % Extremely likely and likely
100%
May-14
Jun-14
Jul-14
Aug-14
Sep-14
FFT inpatients: % extremely likely and likely by
division
98%
96%
96%
94%
94%
88%
Apr-14
Maternity
May-14
Jun-14
ED
Jul-14
inpatients
Aug-14
Sep-14
Trust
“I was amazed with the genuine care and support I
received. Your nursing staff and care workers are the best
I have ever come into contact with. I couldn’t ask for
more. If I could hand out gold stars to them all I would.”
Upper Gastro-Intestinal Ward
Spotlight
The patient experience team have introduced a monthly award which is selected
based on of the following criteria: excellent feedback; improved feedback;
innovative solutions or continued commitment to improving patient experience; or
high or improved response rates on surveys delivered by staff.
86%
Apr-14
100%
98%
94%
92%
90%
89%
Positive patient experience
97%
96%
99%
97%
95%
93%
91%
89%
87%
85%
FFT CQUIN
Outpatient and day case FFT project status:
FFT was implemented in Outpatients and Day cases ahead of the 1 October
2014 deadline. The Trust is recording feedback on paper questionnaires
initially but will implement text messaging for some services.
The patient experience team have developed a project plan to increase
response rates in the emergency departments and inpatients by quarter 4.
The target response rate for emergency departments is 20% over quarter 4;
and the target response rate for inpatients is 30% for the quarter overall
and 40% for March 2015.
The team plan to visit wards and talk with staff to support them, in addition
to a social media campaign which will publicise the feedback received and
changes made.
FFT trend analysis
Inpatients:
The FFT score and the percentage of extremely likely and likely responses
remain constant. The Children’s and Womens (C&W) division score has
increased again: the fluctuation is likely to be attributable to the low
number of respondents, which can affect reliability. The inpatient response
rate has not changed overall but there has been an increase in response rate
in the Children’s & Women’s division and a decrease in Surgery and
Oncology.
ED: The FFT score and the percentage of patients either extremely likely or
likely to recommend the service decreased in September. The response rate
stayed the same.
Maternity: The FFT score and the percentage of patients either extremely
likely or likely to recommend the service for maternity stayed about the
same. The response rate increased in August and the improved response
rate has been maintained in September.
May-14
Jun-14
Jul-14
Aug-14
Sep-14
The Trust’s Maternity Services are the first to receive this award for the improved
response rate in August which has been maintained in September.
Complaints
New complaints
New PALS enquiries
New complaints opened
40
0.60%
35
% PALS against Finished Consultant Episodes (FCE)
20
0.50%
30
25
25
21
17
20
15
7
6
6
10
5
15
0.40%
Feb-14 Mar-14
Apr-14 May-14 Jun-14
Jul-14
Aug-14
5
0.20%
0.17%
0.12%
0.06%
0.10%
0.12%
0.09%
0.08%
C&W
Jul-14
0.05%
0.04%
0.04%
0.02%
0.01%
Feb-14 Mar-14 Apr-14 May-14 Jun-14
Jul-14
Aug-14 Sep-14
Complaints by severity grading, July- September 2014/15
100
95%
94%
82%
80%
79%
81%
79%
79%
Annual 2013/14
•
93%
Quarter 3
(2013/14)
•
Quarter 4
(2013/14)
Quarter 1
(2014/15)
Quarter 2
(2014/15)
% complaints acknowledged within 3 days (Trust-level)
Quarter 1 (2014/15)
Quarter 2 (2014/15)
•
•
•
S&O received the highest number of formal complaints in September,
seeing an increase particularly in the Renal Transplant Directorate
MRC received the second highest number of formal complaints in
September, with NOTSS seeing a reduction over 4 months in the number
they received.
No red complaints were received in September.
Access to Outpatients continues to be a recurrent theme for complaints
received by all Divisions. Care/nursing care also features highly as a theme
for all clinical Divisions. The complaints received by corporate services
included car parking and hotel services.
100%
0
C&W
MRC
NOTSS
S&O
CSS
Corporate
Top 3 complaints themes by division, July - September
2014/15
99%
98%
•
97%
97%
96%
Target 95%
95%
94%
50
93%
92%
0
Corporate
77%
40
100
CSS
Complaints overview
94%
91%
20
S&O
96%
93%
60
NOTSS
0
78%
92%
80
1
% complaints upheld or partially upheld
97%
96%
0.00%
MRC
81%
98%
97%
1
Aug-14 Sep-14
% complaints investigations completed within agreed
timescales
0.07%
0.06%
3
0.02%
Managing complaints
0.10%
3
2
0
This includes all PALS enquiries and issues: positive, negative, or mixed
feedback; issues for resolution; and advice or information requests.
% Complaints against Finished Consultant Episodes (FCE)
0.14%
Feb-14 Mar-14 Apr-14 May-14 Jun-14
Sep-14
10
0.32%
0.30%
0.00%
0
Closed complaints
Reopened complaints: September
91%
C&W
MRC
NOTSS
S&O
Care/nursing care
Access
Hotel services
Parking
Attitude
other
CSS
Corporate
Communication
90%
Feb-14 Mar-14 Apr-14 May-14 Jun-14
Jul-14
Aug-14 Sep-14
•
Managing complaints
The Trust continues to meet the target of 95% for acknowledgement of
complaints and surpassed this by achieving 100% in September 2014.
The Trust achieved 97% for responding to complainants within 25 working
days or agreed timescale against a target of 95%.
Appendix 3
Inpatient Safe staffing dashboard
Trust staffing levels by shift
Inpatient areas including ED.
Early Shift Staffing Levels by Shift
n
August – September
2014
Staffing
Level
Assessment
Surplus
Staffing levels by ward
Inpatient Areas Including ED.
Early Shift Staffing Levels by Ward August – September 2014
Nurse Sensitive Indicators
Selected inpatient areas excluding ED. The data
presented is accurate as of the date the data was
retrieved from Datix, on the 8th of the month.
Medicine Administration
Error or Concerns
80
70
60
50
40
Number
of shifts
27
2430
Minimum
1442
20
Escalation
66
10
3965
Late Shift Staffing Levels by Ward August – September 2014
25
Agreed
Level
2354
Minimum
1477
Escalation
Grand Total
109
Surplus
Agreed
Level
1
Aug-14
98
75
76
Aug-14
Sep-14
Night Shift Staffing Levels by Ward August – September 2014
5
2
1
Vacancies/Sickness/Maternity Leave
Vacancy as a Percentage of
establishment WTE
13.0%
11.5%
Sep-14
202
83.7
80.0
80.0
75.0
190
Jul-14
Aug-14
Jul-14
Sep-14
Patient Falls with Moderate,
Major & Catastrophic Harm
Aug-14
10
3
5
4
Sep-14
Sickness Perecentage in month
6.0%
5.1%
5.5%
5
Sep-14
90.7
90.0
85.0
200
Aug-14
Maternity/Adoption WTE
95.0
216
200
This is the first time the data have been presented as a dashboard format, and will continue to evolve.
The data highlights include:
Staffing levels Trust wide by shift and by ward.
This demonstrates the levels of shifts that required escalation (red) in order to mitigate the levels of staff to
meet increased activity, acuity of patients and/or short notice sickness absence/inability to fill shifts with
temporary staff.
The number of shifts running at minimal levels reflects the number of vacancies in the Trust as well as the
reduced availability of temporary staff during August and early September which were holiday periods.
This dashboard includes the presentation of both incident and Human Resources metrics for Q2 as Nurse
Sensitive Indicators.
11.6%
Jul-14
Aug-14
12.4%
11.0%
All Patient falls
3027
12.9%
12.0%
0
Jul-14
Sep-14
Inpatient areas excluding ED. The data presented
was accurate as of the date the data was
th
retrieved, on the 20 of the month.
12.5%
4
Aug-14
Outcome of Surgical & Medical
Care
Communication, Information &
Consent
Appointment Admission Discharge
Sep-14
10
210
3965
0
4
Avoidable Grade 3-4 Hospital
Acquired Pressure Ulcers
27
58
5
Jul-14
Jul-14
220
Escalation
Sep-14
11
100
90
80
70
Number
of Shifts
853
Aug-14
All Hospital Acquired
Pressure Ulcers
3965
Minimum
Grand Total
10
Jul-14
Night Shift Staffing Levels by Shift
August – September 2014
Staffing
Level
Assessment
49
0
Number
of Shifts
Surplus
15
Extravasation incidents
Late Shift Staffing Levels by Shift
August – September 2014
Staffing
Level
Assessment
75
Jul-14
Agreed
Level
Grand Total
76
20
Complaints
Top 3 Trust Formal
Complaints: Selected Inpatient
Areas
5.0%
4.7%
4.6%
Jul-14
Aug-14
4.5%
4.0%
0
Jul-14
Aug-14
Sep-14
Sep-14