1. Australian and New Zealand Emergency Laparotomy Audit – Quality Improvement
An overview
May 2018
This PowerPoint presentation is being sent to hospitals and departments that have expressed an interest in joining ANZELA-QI. It will also be available on the ANZELA-QI web site <
The aim is to provide some background to ANZELA-QI. You are free to show it and pass it on as you wish. ANZELA-QI would be pleased to have comments that might help improve future versions.

2. ANZELA-QI pilot audit
Co-led by Royal Australasian College of Surgeons (RACS) and Australian and New Zealand College of Anaesthetists (ANZCA) with collaboration from:
General Surgeons Australia
New Zealand Association of General Surgeons
Australian Society of Anaesthetists
New Zealand Society of Anaesthetists
Australasian College for Emergency Medicine
College of Intensive Care Medicine
In late 2017 the ANZELA-QI working group submitted a business case for a 12-month pilot bi-national EL audit to the Councils of the RACS and ANZCA.
This was supported, and ANZELA-QI is being co-led by the Royal Australasian College of Surgeons and the Australian and New Zealand College of Anaesthetists, with close collaboration from four specialty societies; General Surgeons Australia, New Zealand Association of General Surgeons, Australian Society of Anaesthetists and the New Zealand Society of Anaesthetists and the College of Intensive Care Medicine and the Australian College of Emergency Medicine.  It is also being strongly supported by individual surgeons and anaesthetists.

3. Aims of the ANZELA-QI pilot
To agree an Australian and New Zealand EL data set and test it
To focus on Quality Improvement from the outset by providing hospitals with frequent data reports showing performance against KPIs
Provide EL data to support a funding application for a five year definitive project
The principal aims of the ANZELA-QI pilot is outlined in this slide.
The pilot ANZELA-QI dataset is a slimmed down version of the NELA dataset and includes some questions of local relevance. Those responsible for NELA have been very supportive and helpful during the establishment of ANZELA-QI.
Note that each year NELA has made revisions to its dataset which is still being redefined on the basis of their experience. Doubtless the pilot ANZELA-QI dataset will also need to be refined. So all feedback will be welcome.
A very important aim is to obtain data that can been used to support funding for a longer term project. NELA was funded for four years and this has been extended for another five years. ANZELA-QI would like to obtain funding for five years in the first instance. This is likely to be more successful if we can demonstrate ‘proof of concept’.

4. ANZELA-QI time lines
In June 2017 Dave Murray, then Chair of National Emergency Laparotomy Audit in England and Wales (NELA), was the invited guest at the ANZCA NSW meeting. In anticipation of his presence a number of clinicians in Australia and New Zealand contacted the conference convener, David Elliot, seeking an informal meeting with Dave Murray to discuss emergency laparotomy in Australia and New Zealand. This took place in the Westmead hospital two days later. At that meeting it became clear a number of hospitals and individuals were considering some form of local emergency laparotomy audit.
In August the joint WA/NT/SA RACS state meeting in Perth was used an opportunity to hold a planning meeting and symposia. In order to progress this in an orderly manner the attendees agreed to establishing an ANZELA-QI working party. The aim of the working party was to undertake the background work that would permit ANZELA-QI to be formally established with its Governance committee with its own Terms of Reference at the 2018 joint ASC in Sydney.
The complexities of a project that involves multiple Colleges has meant this has been slightly slower than anticipated and so the ANZELA-QI working party will remain in place for an additional couple of months.

5. The rationale: limited EL data
No published Australian or New Zealand data prior to 2017
Publications
1 published prospective multi-hospital audit
1 retrospective single hospital audit
1 retrospective administrative data state audit
Other data
2017 & 2018 ASC abstracts and presentations
Two year national administrative data from the Independent Hospital Pricing Authority
With the exception of NELA almost all international studies published before ~2014 suggested a 30-day mortality of ~15% and >25% in those over 80 years. There was also wide variation and poor compliance with evidence based standards. NELA currently reports a 30-day mortality around 11%.
There is very limited published EL data in Australia or New Zealand. Prior to 2017 the only data appeared to be a one line comment in the discussion of a single paper. During 2017 and 2018 data from a small number of Australian EL studies has become available.
Taken together these studies suggest that contemporary 30-day mortality in Australia may be slightly lower than that reported overseas. There appears to be the same inter-hospital variations and poor compliance with evidence based standards noted overseas. The available Australian EL data suggests there may be some important differences when compared to overseas. These differences include the large proportion of patients transferred (15% to 30%) as part of their EL care and the greater proportion undertaken in the private sector (~15%).
However, it is clear that the quality of this Australian data is in no way comparable to that now available overseas. This is unsatisfactory given that EL is a high risk operation that is associated with substantial cost.

6. The basics Pilot database REDCap (Vanderbilt University)
Accessible from any web enabled device
Australia - National HREC ethics approval (not Tasmania and NT)
granted in March 2018
waiver of consent
Site Specific Approval (governance) required
New Zealand - National HDEC ethics approval
each participating site will require local ethics approval
Organisation audit in preparation in Australia and underway in NZ
Private hospitals, IHPA and AIHW differ …..
Pilot hospitals
Approx. 50 expressed an interest at present
The ANZELA-QI Working Party has agreed to a pilot dataset and data dictionary. This pilot project will be built on REDCap, a secure database platform developed by Vanderbilt University, and hosted on the RACS server in Melbourne along with other audits that the College is responsible for. REDCap is widely used in numerous countries, including Australia and New Zealand, for medical research projects.
Ethical consent has been obtained in both Australia and New Zealand. However, hospitals in both will still need to obtain local governance approval. Local hospital governance processes are highly variable and hospitals wishing to contribute to the ANZELA-QI pilot would be wise to commence this process as soon as possible. Please contact the RACS RAAS office in Adelaide if you require further details.
An organisational audit is in the advanced stages of preparation in Australia, and actively underway in New Zealand. All participating hospitals will be asked to complete this.
Hospitals may wish to commence data collection whilst awaiting local governance approval. They must do so in a manner that is compatible with the ANZELA-QI dataset. There are two options:-
To collect the data by hand/paper, and once local governance approval has been granted, enter that collected data themselves on the the ANZELA-QI REDCap. If the number of EL is small that is likely to be the easiest way. Please note that the RAAS office does not have the support to enter paper records sent to them.
To collect the data electronically using a template that will be provided by the RAAS. This can then be sent to the RAAS for downloading into REDCap. Please note that the RAAS office does not have the support to clean electronic records and unless sent in the proscribed format it will not be entered onto REDCap. The RAAS will also assess the viability of an upload before accepting it, giving precedence to those with larger volume and greater commonality of fields/values. There will be a limit to how many sites can be accommodated for upload, so please speak with this RAAS office if you would like to contribute in this way.
Once hospitals start collecting prospective data they may then wish to add in earlier data, albeit retrospective, so they have a comparative base line. This is perfectly acceptable. But note 1 and 2 above!

7. ANZELA-QI versus NELA - differences
Inclusion/exclusion criteria
Emergency Laparotomy not just bowel surgery
Quality Improvement focus from outset
Futility of surgery
Frailty score (Canadian)
Goals of Care documentation
Include patients who die without an EL & those with acute abdomens but not operated on
Transfers: 25% to 30% in Australia versus ~3% in NELA
Private sector: ~15% in Australia
Although modelled on NELA there will be differences between the two audits. The most significant are summarised on this slide.
Some details about QI are on the next slide.
There is increasing recognition that frailty may be a major determinate of risk. In recognition of this NELA has just added a field to its dataset. ANZELA-QI will use the Rockwood scale that will be available in the dataset for reference.
Some patients presenting with an acute abdomen may not undergo surgery. NELA did not record these. This has the potential to bias the outcome as a hospital that has a ‘proactive’ approach to avoiding futile surgery will by definition exclude high risk cases and so have a lower mortality. Recording these will not be easy as there is no natural point of data collection.
The number and proportion of patients transferred during the course of their EL care will be much greater in Australia than in the UK or New Zealand. Contrary to what might be expected the majority of transfers appear to be between metro hospitals and not from rural or remote hospitals.
It is estimated that 15% of emergency laparotomies take place in the private sector. Many will be for a complication of an elective operation and these are known to have a worse outcome. Many will also be transferred to a public hospital at some point in their care.

8. Emergency Laparotomy Pathway Quality improvement care (ELPQuiC)
Evidence based ‘bundle of care’ introduced at 4 hospitals:
Early Warning Score
Early antibiotics
Operation < 6 hours
Post-op ICU
Goal Directed Care
Following introduction:
Adjusted mortality reduced 15.6% -> 9.6%
5.97 lives extended beyond 30 days per 100 patients treated
Huddart et al, BJS, 2014
It was agreed at the first meeting that Quality Improvement would be at the centre of ANZELA-QI; hence its name.
The initial contract granted to NELA required it to be a data assurance rather than quality improvement project. NELA’s initial remit was to collect and publish high quality comparative information which can then used to drive quality improvement. In the first three years NELA has demonstrated a reduced mortality and improved compliance with care standards. These encouraging results prompted the UK government to extend the initial four year funding for a further five years.
However, this feedback was largely via annual reports and so not timely in relation to the events it was reporting. QI techniques can provide near real time feedback. The additional five years funding NELA was granted from December 2017 will permit it to report its data using QI techniques. Initially this will be through the distribution of quarterly reports to participating hospitals.
The Emergency Laparotomy Pathway quality improvement care (ELPQuiC) audit was a direct follow on from the UK Emergency Laparotomy Network (ELN) audit. The four hospitals that took part introduced a ‘bundle of care’ package for patients undergoing EL. The bundle of care package was based on the evidence of many previous studies such as NCEPOD and the seminal report ‘The Higher Risk General Surgical Patient’.
The ELPQuiC pooled data showed an additional 5.97 patients per 100 treated survived beyond 30 days following the introduction of the ELPQuiC bundle (P<0.001). More recently a QI study from Copenhagen reported improved outcome following the introduction of a bundle of care.
The Emergency Laparotomy Collaborative (ELC) is a scaled up version of the EPLQuiC study across 28 hospitals in south east England. The NELA provided data to the ELC every three months and this was presented to participating hospitals in a way that permitted them to compare themselves to their ELC peers. The ELC has now closed and results likely to become available later in The EPOCH study is also closed and its results will become available in due course.

9. NELA has established nine KPI’s
NELA has established nine KPI’s. Note these are all process measures and not outcomes. ANZELA-QI will use eight of these.
Although outcomes make the headlines, and for EL notably mortality, it will be impossible for a one year pilot study to generate enough data that is robust enough for outcome analysis. In NELA one third of hospitals with a risk adjusted mortality that is worse than the average are above the average the next year – and vice versa. On the available data the average hospital in NELA undertakes almost double the number of emergency laparotomies of the average Australian public hospital.
The RAG rating is the Red-Amber-Green scale of compliance with the standards of ≤50%, 50% to 80%, ≥80%. Note that NELA does not provide standards; these were derived from many studies and reports in the UK as on the previous slide. NELA only measure compliances with the standards, recognising that a standard of 100% is unrealistic.

10. Pre-operative risk assessment
Options
P-POSSUM
SORT
NSQUIP
NELA
and 19 others …
Key issue is that a risk assessment is done
For many patients presenting with an acute abdomen timely treatment and surgery has a direct impact on survival and on the quality of that survival. In particular, the early identification and management of sepsis has been important. ANZELA-QI will specifically enquire as to the presence of sepsis.
ANZELA-QI will use the NELA risk assessment that is available as an app. The NELA risk algorithm is specifically designed for EL and is based on contemporary, albeit UK, data. P-POSSUM, the historical alternative, includes elective cases and is more than 20 years old. There are other differences between NELA and P-POSSUM. See <
Whether or not a pre-operative risk assessment was undertaken will be one of the important KPI’s measured in ANZELA-QI. Both prospective NELA and the prospective Perth Emergency Laparotomy Audit (PELA) showed that 30-day mortality in patients who did not have a pre-operative risk assessment was greater than those who did, in PELA almost double. Presumably this greater 30-day mortality was because the critical status of the patient was not appreciated in those who did not have a risk assessment.
It is very important that the NELA score is recorded as ANZELA-QI intends. In the ANZELA-QI it will be possible to specifically record whether the NELA score was performed and documented pre-operatively, or calculated later and so only available retrospectively after the surgery. The latter clearly can not be used to drive the timeliness of surgery which is its purpose. Ideally all patients will have a pre-operative NELA risk assessment that is documented. And note that the standard is that it is both measured and documented pre-operatively.
However, a pre-operative NELA risk assessment will not always be done. If the NELA score is not measured and documented pre-operatively, ANZELA-QI would still like the NELA score to be calculated respectively and recorded, but that this shown to have been done retrospectively. It will then be possible for ANZELA-QI to analyse the outcome for patients with the same NELA score by whether they did or did not have a pre-operative score.

11. Frailty and Goals of Care
There is increasing recognition that frailty is an important determinate of outcome. This relates to both mortality and quality of life. The latter, especially if the outcome is likely to result in a loss of independence, is very important to patients.
ANZELA-QI will use the Rockwood frailty score. Many will not be familiar with this so this diagram will be embedded in the database for reference.
End of Life matters now have a national focus and the Australian Commission in Safety and Quality in Health has recently published a national consensus statement. There are plans to introduce a national form that patients will complete to document their Goals of Care (GoC). The example shown had been trialled in Western Australia.
EL are a common operation frequently undertaken in an elderly, acutely unwell and high-risk patient. Documenting GoC is very important and ANZELA-QI will record whether this occurred and recorded in the notes.
Frailty and GoC are good examples are another example of a very important ANZELA-QI principal. The aim of ANZELA-Qi is to prospectively record these important factors that can then be used to guide patient care. If they are not measured and documented in advance then they cannot be used to influence care. When completing the ANZELA-QI database it is important to clearly differentiate assessments that were done prospectively, and so available at the time to the clinician, and those that became available after the event and so not available to the clinician and so could not have influenced care.

12. Acute abdomen, no surgery
What is the true denominator?
PELA
6.5% 30-day mortality
proportion ≥80 years and risk ≥10% less than NELA
43% of all presenting with an acute abdomen and who die did not have an EL
WAASM data (2 years)
4% of all acute abdomens die without an EL
6% die after an EL
Some patients presenting with an acute abdomen who might normally be expected to undergo an EL do not undergo surgery. The reasons for this are many and varied. These patients were not included in NELA although there was a pilot study reported at the November 2018 NELA conference in Birmingham.
The 30-day mortality in the Perth Emergency Laparotomy Audit (PELA) and NELA III were 6.5% and 11% respectively. However, the PELA had fewer high risk cases.
As part of the PELA all deaths under a General Surgeon in the Western Australian Audit of Surgical Mortality (WAASM) were reviewed. Those who presented with an acute abdomen that appeared to meet the PELA criteria but who did not have an EL were identified and reviewed. Of those presenting with an ‘eligible’ acute abdomen almost half died without undergoing an EL. A larger (as yet unpublished) study over two years suggests that of all patients presenting with an acute abdomen and who die about 4% do not undergo an EL and 6% do. When added together this gives a 30-day mortality that is similar to NELA.
The avoidance of ‘futile’ surgery has been a recurring theme reported by the WAASM. It is possible the lower 30-day mortality in PELA is in part secondary to the exclusion of high risk, or potentially futile, EL.
The true number not undergoing an EL will almost certainly be greater as some will not be under the care of a General Surgeon and so not reported to WAASM.
The ANZELA-QI pilot wishes to capture those presenting with an acute abdomen who meet the inclusion criteria but who do not undergo an EL. ANZELA-QI recognises this may not be easy as these patients do not go to theatre and so do not have a clearly identifiable event, and they may not be under a surgeon, or even on a surgical ward.

13. Some practical tips
The universal experience is that data to that point should be collected/completed in theatre
Most will enter the hospitals system via ED and a sepsis assessment should be undertaken there
If a pre-operative risk assessment is more likely to be prospectively completed if if is embedded in the theatre booking process
Post-operative rounds are ideal to check all the data has been collected
Full case ascertainment will require the hospital PI to check theatre lists each week to ensure all EL have been captured.
The quality of any audit is greatly reduced if there is missing data. ANZELA-QI will be reporting data quality from the outset. This will include the number of fields completed, when the data was captured etc.
Hospitals in both the UK and Australia that have collected prospective EL data have identified the same common themes that facilitate data capture, and so ensure more complete capture. Some of these common themes are summarised in the slide.
The surest way to ensure the pre-operative NELA risk assessment is undertaken is to embed it in to the theatre booking system such that a booking for an EL cannot be finalised until completed. As hospitals increasing move to electronic booking system this will be more easily achieved.
The universal experience of those who have been collecting prospective EL data in the UK and Australia is that the best time to collect the data is in theatre when the operation is being completed. Most data is there and easily recalled. It takes <7 minutes. If not done at that time the data is often not reliably recallable, and even not available. It also takes much, much longer.
The post-take round can be used to check that any EL’s have been entered and the fields to that point completed. As in the previous slide ANZELA-QI is very keen to collect data on patients who have an acute abdomen and die without a EL and the post-take round is an ideal time to check that.
Hospitals will need to undertake regular case ascertainment to ensure all EL’s are captured recognising there is no code for an EL. The local PI will need to determine how that is best done. A weekly check of the theatre register is likely to be the best way.

14. Important issues Data collected to hospitals level only
No clinician data being collected
Benchmarking with peer hospitals on an identifiable basis
No Qualified Privilege
Aggregated data open for research
Trainee Research Collaborative
The NELA and ELC experience is that dissemination of data to local hospitals who act in collaboration has been a powerful driver of change. Hospitals have more in common with the other hospitals in same town/area than those the other side of the country, in a different environment etc. Further, local hospitals can easily meet and discuss the strengths and weakness of their data and so learn from each other.
ANZELA-QI will not collect clinician level data. The management of an EL depends on a hospital system and not an individual. Data will be published at the named hospital level. This is as in NELA and recognises this is the level of openness and transparency now expected.
ANZELA-QI will encourage research projects that use de-identified aggregated data. The Trainee Research Collaborative have used NELA as the basis of a large number of highly successful and productive projects.

15. Run chart
Run charts are one form of Statistical Control Process and have been used in industry since the 1930’s.
This is an example of the run charts NELA is now presenting and shows the performance of a hospital improving each month. A change is normally detected after events and a ‘run’ can be related to any time period, or based on a number of events. As a trial two Perth hospitals have been collecting ANZELA-QI data and changes of practice can be seen in cases.
Initially ANZELA-QI will return data every month.

16. Changing practice is not easy or quick!
These graphs show the number of NELA hospitals that have achieved 80% of the ‘reportable’ KPI’s. Although there is obvious improvement over three years there is clearly room for improvement.
This was during the period when NELA was not using QI methodology. The aim of ANZELA-Qi is to provide timely feedback and accelerate change in practice.

17. Average National Weighted Activity Unit (NWAU) cost of an Emergency Laparotomy in Australian public hospitals
Estimated national EL cost >$400 million pa
Bed day cost ~$30 million
A two year study utilising NWAU data from the Independent Hospitals Pricing Authority (IHPA) suggests the cost of an EL in Australia (excluding the 10 most expensive and cheapest hospitals) ranged from $21,463 to $39,864.
The ANZELA-QI pilot is a 12-month bi-national study being established to support a funding application for a 5-year project. This is likely to cost AU$4 to $5M. ANZELA-QI is likely to be highly cost effective. During its first three years NELA reduced the average length of stay for EL’s by 2.6 days for a saving of at least £22M per annum. If private hospitals are included each day of reduced stay following an EL in Australia represents a saving of at least $34M per annum (unpublished NWAU data from IHPA).

18. What will ANZELA-QI provide?
Reports against the KPIs, published by identified hospital and in cohorts
The reports will facilitate sharing of experiences and learnings with other similar hospitals
Access to the bi-national dataset for other research projects (subject to an approval process)
The ANZELA-QI is likely to be unique in at least two ways:-
Data will be published by identified hospital. This will permit hospitals to share experiences and learn from each other. This is likely to be the first prospective Australian multi-hospital surgical audit to publish identifiable data. It will mark a significant step toward the openness and transparency that is now expected, and indeed been available in other countries for many years.
It will use QI techniques to provide prompt, and ultimately real-time, feedback to participants. The value of QI methods is likely to be confirmed when the final results of the 28-hospital UK Emergency Laparotomy Collaborative are published shortly.
De-identified aggregated data will be available for research. Trainee Research Collaboratives have been very successful in the UK and have great potential to expand the use of data generated by ANZELA-QI.

19. Specifically for New Zealand
CADENZAA
Will report the NZ data back to RACS Adelaide
Will leverage strong IT platforms
Build QI into the system
Full collaboration with ANZELA-QI
The New Zealand arm of the project is called ‘Care Delivery in New Zealand for the Acute Abdomen’.
CADENZAA aims to capture the minimum data set of ANZELA-QI and report this back to RACS for bi-national amalgamation.
In addition, CADENZAA aims to leverage currently existing, strong IT platforms in order to widen the scope of data capture to all acute abdomens while dramatically reduce data collection burden.
Theoretically, it will also be possible to capture high quality data with ease and build quality improvement into the system. The project in NZ has its own steering group, but is closely linked with the ANZELA-QI Working Party and is in full collaboration.
National ethics is in place however each participating site will require local ethics approval.

20. What next?
RACS your expression of interest in becoming a pilot hospital.
Once RACS have new pilot sites added to national ethics, each site needs to coordinate local governance approval (SSA). Then login to ANZELA-QI can be given by RACS.
The website will be the central place for all pilot hospital resources.
Forms/templates will be available to collect data locally ahead of ethical approval to provide it to ANZELA-QI.
More information at
Call Katherine Economides/RACS on
The audit is coordinated centrally from the RACS Morbidity Audit office in Adelaide, Australia. Katherine Economides, the Manager there, is the central contact for any queries about the audit.
A login to the ANZELA-QI database will be given once all approvals are in place; firstly, each site needs to be added to the national ethical approval (coordinated by RACS), secondly each site needs local governance approval (organised by the site PI and supported by RACS). In Australia, the NT and TAS need to arrange their own local ethical approval, as they are not in the National Mutual scheme (RACS can support if needed).
Until then, sites can collect data (but not submit it to ANZELA-QI) by completing one of the templates provided on the webpage (coming soon).