National Heart Failure Audit Steering Group meeting minutes 3 April 2014, 13.30-15.30 rd Boardroom, 3 floor 170 Tottenham Court Road, London W1T 7HA 1. Apologies and introductions 1.1 Present: Gemma Baldock-Apps, Janine Beezer, Henry Dargie, Gethin Ellis, Dawn Lambert, Suzanna Hardman, Theresa McDonagh (chair), Richard Mindham, Polly Mitchell, Jim Moore, Julie Sanders, Kathy Simmonds. 1.2 In attendance: Morag Cunningham 1.3 Apologies: John Cleland, Jackie Austin. TM introduced Gethin Ellis, a new member of the Steering Group who will be representing Cardiology in Wales and the Welsh Cardiac Networks. 2. Minutes of the last meeting and matters arising 3 (AP1): PM confirmed that Marion Standing had added life status information to the export, and would be updating the online reports over the next few weeks. The reports and export format would be circulated around the Steering Group before being release more widely to participating hospitals. Action 1: PM to circulate draft mortality data releases to Steering Group prior to release. 3. Annual report 2013/14 analysis plan This has not been circulated yet, as it was discussed immediately prior to the current meeting. TM noted that in 2013/14 the report would be shorter, with fewer Kaplan-Meier graphs, and with more detailed analysis reserved for peer reviewed articles and online appendices. She gave a summary of proposed changes to the analysis plan: Basic demographic and descriptive analyses will remain broadly similar, with a few pieces of analysis being cut. No cardiovascular or heart failure mortality (except perhaps headline figures) as these showed little difference from all cause last year. KM curves (1 year and 5 year mortality): LVSD v. HFPEF, cardiology v. other wards, cardiology and HFNS follow-up, therapies for LVSD and additive benefit of therapies. Readmission (30 day and 1 year): All-cause, CV and HF readmission. Possible inclusion of analysis on relationship between length of stay and readmission. Anonymised mortality and readmission by hospital (similar to the length of stay charts in previous years) Trends over 5 years (e.g. for in-patient mortality, prescription and referral rates). GE notes that it would be very useful for Welsh and English figures for key indicators to be published separately. He noted that this would be very useful for promoting data entry and quality improvement locally. It was agreed that this would be easy to include alongside hospital-level analysis. Action 2: PM to write up proposed revisions to annual report analysis and circulate to Steering Group for comment. SH suggested that more copies of the annual report could be printed for the BSH. PM noted that the number of reports produced for the BSH had been increased last year due to increased demand, and that the BSH had kindly donated funding to the audit towards the costs of production. SH suggested that PM contact Andrew Clarke now regarding continues financial input from the BSH. TM noted that the PDF of the report could be included in the electronic revalidation appraisal tool. TM noted that the audit slides from the 2013 BSH meeting were not yet on the website. Action 3: SH to remind BSH secretariat to include audit presentation slides on BSH website. PM suggested that the audit could produce a set of slides alongside the annual report, which could be used or adapted by hospitals for local presentations. 4. Dataset revision PM reported that the dataset revision had gone ahead with only minor glitches, which were under control. Philips, who develop CVIS (previously known as TOMCAT) clinical system have asked whether we can extend the reporting period to 12 months, and have reported that their HF module will not be ready for use until later in the year. They plan to release four upgrades (heart failure, CRM, EPS ablation and BCIS) at once, and also are pushing for hardware upgrades in centres at the same time. TM noted that CVIS is mostly used in cath labs, and PM confirmed that not many centres (under 5) actually used CVIS to upload data to the HF audit. Many centres using CVIS found it easier to directly enter data to the audit, with GBA and East Sussex Trust being a case in point. PM said she had told Philips that extending the parallel reporting period for 12 months would be unlikely to be possible – the audit is in the process of developing a risk model, and hospitals not including data for new risk fields will be in danger of being excluded from this model, resulting in no hospital-level mortality being published for them, and potentially skewing the model. PM asked whether the Steering Group would consider extending the parallel reporting period at all. It was noted that to make allowances for Philips would be unfair on all other companies and hospitals who had managed to upgrade their systems in time for the new dataset roll-out. The Steering Group was generally not happy to extend the period of running the two datasets in parallel. Action 4: PM to find out which hospitals are using the CVIS HF module, and which are using it to upload data to the audit. 5. Best practice tariff TM and PM have met with Monitor and the Department of Health twice to discuss a proposed best practice tariff for heart failure. Best practice tariffs pay a higher rate when centres meet certain quality indicators, and a lower rate when they fail to do so. The aim is to give a financial incentive for good quality care, much like CQUIN payments or QOF in primary care. An initial proposal has been made for a heart failure best practice tariff to be based on audit data. TM noted that she had initially been enthusiastic, assuming that there would be an incentive offered for good performance, and considering a link between audit participation and good performance with financial reimbursement to be a good way of improving HF services. However it had emerged that there was no additional money for the scheme, so there would be a penalty for non-compliant Trusts, such that if they don’t meet targets they could lose over 10% of their heart failure funding. There is a workshop on 16 April, with PM, TM and SH will attend, along with other HF professionals, including HF nurses and cardiologists. It was noted that we want to improve HF care, but don’t want to unintentionally take resources out of HF and compound problems in poorly performing centres. Commissioning of services is done locally, so it may be difficult or impossible to reapportion money centrally at the end of the year – instead two fixed tariffs would be used. This could potentially mean that overall a significant amount of funding is taken out of heart failure, if many hospitals fail to meet the target. This would penalise underperforming hospitals, and potentially make matters worse for them. It was noted that an overspend/underspend one year could be balanced out the next, to avoid long term financial loss. SH said that there was the possibility of staggering the introduction of the penalty/incentive so that Trusts have time to change their practice without being hit with a large financial shortfall. The scheme would work alongside existing incentive projects such as CQUIN, and would not replace them. JM noted that QOF has been very successful in improving HF management in primary care. If audit data was used to evidence best practice, participation in the audit could be used initially, along with, or followed by two or three other measures (e.g. specialist input, prescription rates, referral rates). Ideally we need to see examples of BPTs for other chronic conditions, to see what measures they have used and how/whether they have worked. This would have a potential benefit for the audit in terms in increased participation and data quality. JM reported that in Gloucestershire, discussions about participation in the HF audit had centred on whether there were any financial penalties for non-participation. PM noted that the commissioners would need the data from NICOR, and it had been suggested that it would be required every quarter. This would require hospitals to have quarterly deadlines for data submission, and for NICOR to link to ONS and produce hospital-level analysis four times per year. This was deemed unachievable. JB noted that CQUIN payments were calculated on an annual basis. Action 5: PM to pass JM’s email on to Department of Health contact, for suggested invitation to workshop to discuss best practice tariff. 6. Primary care project PM and JM outlined a proposed primary care audit project, using data from JM’s community HF audit in Gloucestershire. Gloucestershire Care Services has carried out an extensive audit for 10 years into the care delivered by their community heart failure service. The data captured by this audit covers the diagnosis, medication and referrals for all patients using the service, and collects very similar information to the National Heart Failure Audit dataset. The National Heart Failure Audit has been considering for some time whether and how to extend into primary care. Exploring the potential for extending the audit into primary care is one of the contracted deliverables from HQIP. The first stage of this proposed project would involve the Gloucester Care Services data being imported into the National Heart Failure Audit database, using the main audit dataset and data definitions. This would enable linkage of the primary care records to ONS and HES data, in order to obtain longer term outcomes data for these patients, at minimal additional cost to the audit. This could result in an initial research project to explore whether the long term outcomes of these patients. This would be a platform for the linkage between the community data and regional secondary care data from Gloucestershire Hospitals Trust – this could explore whether patients are admitted to hospital, either before or after their inclusion in the community audit, and track their medicines management across these two services. Gloucestershire Care Services have already agreed that they would be keen to be part of the project, and to share their data with the audit. Gloucestershire Hospitals Trust has also expressed interest, and is in the process of getting an SpR involved in the National Audit tot retrospectively enter data on HF patients admitted to the Trust. Furthermore, there is the possibility of linking with non-specialist primary care data for the Gloucestershire region, i.e. those heart failure patients who are not treated by the specialist heart failure service, using QOF data. JM has spoken to the primary care audit group attached to the local CCG, who have indicated that they are keen to do this, but the practicalities and information governance needed to be explored. PM noted that there were a few IG issues that needed to be explored, including whether NICOR is currently allowed to hold primary care data under the existing S251 approval. Action 6: Organise TC in the next couple of weeks with JM, TM, PM and representatives from Gloucestershire Hospitals Trust to discuss primary care linkage project. 7. Research update HD gave an overview of the ongoing research projects using NHFA data: Bristol University project looking at the measurement of BNP in primary and secondary care. Oxford University project looking at the variation in hospital care of heart failure patients that can be attributed to characteristics of the hospital’s organisational structure. Project led by TM, undertaking a propensity analysis, matching patients based on demographic and aetiological characteristics, to investigate the impact of specialist care on outcomes. In-house analysis for Novartis, looking at the admission characteristics of heart failure patients, and their eligibility for serelaxin. An abstract has been accepted for the ESC (European Society of Cardiology) conference in Barcelona for this project. The MAPP-HF project, which looks to analyse the progression of HF across primary and secondary care, and identify earlier opportunities for prevention and treatment. Initial discussions with John Deanfield about a programme of work looking at diabetes and HF programme. TM raised the issue of charging researchers for data extracts, which has been discussed at the NICOR PLG (professional liaison group) meeting, and asked JS for an update of the NICOR position on this. JS noted that NICOR does not charge for the data itself, but rather covers the costs of extracting and cleaning data, preparing extracts, and managing the data sharing process. A NICOR policy was devised in 2011 to charge academic research groups £5,000 per extract, and an additional £5,000 if linkage to another dataset was required. Commercial research groups would be charged a minimum of £10,000. Under the HQIP contract, NICOR is contractually obliged to charge for research extracts. In February, all audit academic groups were reminded to use the existing NICOR charging policy, as it had previously been applied sporadically. At the same time, NICOR is reviewing the charging process, to ensure that it accurately reflects the costs of managing and preparing the data. TM noted that the NHFA has previously charged commercial organisations (Novartis, Servier) and grant funded academic research groups (George Centre in Oxford, and Bristol University) for analysis and data extracts. The audit has taken a more lenient approach for investigator-led research and non-grant-funded projects, and has waived the charge in these instances. TM expressed concern that the audit would be required to charge £5,000 for investigator-led projects, and suggested that this may obstruct smaller, but nonetheless valuable research projects. TM noted that the PLG had discussed the suggestion that each audit had the ability to waive the fee for three projects per year. She asked whether the audit could also use money brought in from grant funded/commercial projects to support unfunded projects. JS said that the charges would not apply retrospectively for projects where a NICOR data sharing agreement had already been signed. As part of the new charging strategy NICOR will look at each audit potentially being able to waive the fees for a fixed number of projects- perhaps 3. The possibility of audits being able to subsidise projects that are unable to source their own funding would also be discussed. TM noted that the propensity analysis is to be carried out by an external researcher, Ian Ford, who would be paid £10,000 for the work. She asked whether he or the audit would have to be charged for releasing this extract to him. JS confirmed that as long as there was no cost to NICOR, no charge would be levied, and that the audit could use industry funding to cover the costs of this analysis. JS noted that no charge would be applied for clinical audit or service improvement projects (e.g. individual clinicians carrying out local quality improvement projects). Clinical audit in this sense is to be defined as measuring data against existing standards. Anything extending beyond this should be considered research. The audit research group should make a decision as to whether or not a given project should be considered an audit or research project. SH described the importance of the BSH having access to collective audit data in order to drive change in practice, and specifically asked JS to confirm that the BSH would not be asked for a fee if they requested collective audit data for their membership (individual Trusts already have access). JS was able to confirm that this usage would not incur a fee. 8. NICOR update JS gave a brief update of NICOR-wide activities: HQIP contracts for 2014-16 will be signed this week. This will enable new staff to be recruited, specifically a data manager and another analyst. SOPs are being developed to formalise many of the processes within NICOR. UCL has awarded NICOR an impact award for a heart failure PhD project. This means that UCL provides half the funding for a 3-year PhD project (around £30,000). The rest of the funding should be provided by industry. NICOR is applying for grants to link both the National Diabetes Audit and the UK Renal Registry with the NICOR audits. A series of new cardiovascular health technology registries are being started within NICOR, as part of the NHS England ‘Commissioning Through Evaluation’ scheme. The UK TAVI registry is no longer funded by commissioners, and all funding now comes from industry. A new website had been launch – feedback is welcome. NICOR is running a patient and public engagement even in 8 May, to ascertain what information patient and the public want to see from NICOR, and how they would like us to provide it. 9. Prospective methodology pilot Exploring the possibility of a prospective audit is now one of the HQIP contracted deliverables. Prospectively identifying patients removes the need for pulling the notes for each patient, saving considerable time and effort. It also gives heart failure clinicians the opportunity to impact on audited patients’ care while they are still hospitalised – currently patients are only audited postdischarge, meaning that their management cannot be altered. The pilot will run from September to November 2014. Around 10 hospitals will be involved, and it should aim for variation in location and size of centres. The data for this three month period will be compared to the NHFA audit and HES data for these centres in the same three month period in 2013. TM noted that most patients arrive either in A&E, and are then moved to an AMAU or CCU/cardiology ward. Potential heart failure patients should be flagged up on arrival to hospital, either using their EPR, or by someone daily visiting the admission wards and identifying patients with symptoms and signs of heart failure. The symptoms should be confirmed by BNP - if used and echo, and at this point, if the echo shows significant dysfunction, the patient should be flagged for inclusion in the audit. GE noted that a person would need to check the electronic flagging of patients, as this was often inaccurate or incomplete. DL asked whether patients coded as having heart failure, but not captured by this process, should also be audited. It was deemed that this would be too much work, but acknowledged that in order for the project to be thorough these patients would need to be examined to determine whether they are actually heart failure admissions or not. TM noted that there was no additional money to fund this project, and the incentive for centres to take part was their inclusion in subsequent publications. PM confirmed that around 10 centres had already expressed interest. GE added that he would be interested in taking part as well. 10. AOB No other business was raised. Date and time of next meeting: Thursday 3 July, 13.30-15.30 Appendix: Action Points # 1 2 3 4 5 6 Action PM to circulate draft mortality data releases to Steering Group prior to release. PM to write up proposed revisions to annual report analysis and circulate to Steering Group for comment. SH to remind BSH secretariat to include audit presentation slides on BSH website. PM to find out which hospitals are using the CVIS HF module, and which are using it to upload data to the audit. PM to pass JM’s email on to Department of Health contact, for suggested invitation to workshop to discuss best practice tariff. Organise TC in the next couple of weeks with JM, TM, PM and representatives from Gloucestershire Hospitals Trust to discuss primary care linkage project. Owner PM Completed? PM SH PM PM JM, PM