2. To determine deliverability of the agreed LIMS embedded algorithm by each UK LIMS supplier

Transcription

1 E Alerts for GFR Meeting Derby July 2013 Background Several sets of guidelines which link the early detection of AKI with improved outcomes,(e.g. AKIN, KDIGO) have led to an interest in developing alerts for AKI based on serum creatinine changes within a 7 day period. The Edinburgh consensus conference agreed that a group should be set up to design an algorithm that could be embedded in a Laboratory Information Management System (LIMS) bringing together work that several pioneering laboratories had been doing. This group has the support of the NHS England Clinical Directors for (Dr Richard Fluck) and Pathology (Professor Jo Martin). A recently formed Joint Programme Board between NHS England and the Renal Registry will oversee the clinical interventions that will maximise the patient benefits of early AKI detection. Purpose 1. To gather the stakeholders (Renal Physicians, Clinical Biochemists and LIMS suppliers) and agree on the design of a standardised algorithm to detect AKI that can be used in all Trusts within the UK. 2. To determine deliverability of the agreed LIMS embedded algorithm by each UK LIMS supplier 3. To ensure flexibility within the LIMS embedded algorithm such that a) changes to the definition of AKI can be incorporated in the future and b) Trusts with limitations to their data archive can be accommodated. 4. To agree a boundary for this project with regard to message transfer. 5. To consider clinical governance and quality control of algorithm performance 6. To agree a route for achieving the delivery of a specification to LIMS providers. A company limited by guarantee, Registered in England, Registration No Registered Office: Association for and Laboratory Medicine, Tooley Street, London SE1 2TU Website:

2 Meeting Outcomes 1. The clinical requirement is for both detection and staging of AKI. However the meeting recognised that the algorithm is a screening tool and cannot be expected to detect all cases of AKI. Clinical judgement will be required to distinguish between true cases of AKI and unstable CKD in cases where a previous serum creatinine measurement is not available or where the only available serum creatinine is more than 7 days prior to the episode under consideration. 2. The boundary of this project is the generation of a flag within the LIMS that can trigger a standard message or series of standard messages (e.g. HL7 series messages) for receipt by Trust reporting systems. 3. The flow chart in Appendix A (separate document) summarises the algorithm that was agreed by the meeting 4. All LIMS suppliers present at the meeting confirmed that the agreed algorithm was deliverable and saw no issues that would delay implementation beyond the industry standard requirements to configure and test the algorithm. Recommendations 1. The LIMS embedded algorithms should be user configurable to allow evidence based changes that might occur to the definition of AKI and to allow Trusts to use the standardised algorithm software as they move stepwise, if necessary, towards implementing the full recommendations. It is not intended that this flexibility should facilitate and maintain a variety of algorithms over the long term. It is likely that future versions of the algorithm software will have less flexibility as Trust systems develop. 2. LIMS often contain multiple records for individual patients. The AKI diagnosis and staging algorithm should search for any records containing the three identical identifiers: Surname, DOB, NHS number* within the last 365 days in its search for previous serum creatinine results. * Identifiers should be user configurable and conform to Lab minimum data set for patient identification 3. The further processing of e alerts generated by the LIMS should take place within a Trust integration engine or similar which would integrate inputs from e.g. the Renal Registry to identify dialysing patients and Admission Discharge Transfer feeds from Patient Administration systems to direct e alerts in real time. During the interim A company limited by guarantee, Registered in England, Registration No Registered Office: Association for and Laboratory Medicine, Tooley Street, London SE1 2TU Website:

3 period where e alert integration may not be available, alerts may take the form of report comments from the LIMS. 4. During the interim period Trusts should assess the risks and benefits of delaying the reporting of AKI diagnostic/staging comments pending manual inspection of AKI alerts e.g. to exclude renal dialysis patients. 5. Suppression of e alerts at the request of users for any reason should take place outside the LIMS and within the integration engine. During the interim period where e alert integration may not be available, suppression of e alerts can take place within the LIMS via a user configurable property of the algorithm. Each agreement with users to suppress e alerts must be documented in the laboratory s quality management system. 6. Future versions of the AKI algorithm should be agreed by a group similar to the group that produced these recommendations, reporting to the Joint AKI/Renal Registry board at NHS England. Such versions should be issued to LIMS suppliers via a change order from the Data Standards Authority. 7. Quality control should be directed at a) the performance characteristics of the serum creatinine assay as judged by External Quality Assessment; b) the performance of the algorithm with standard electronic data sets (dummy patient data); and c) audit of algorithm performance in live systems at pilot sites to determine the false positive rate. 8. Where laboratories cannot use the standardised algorithm and/or carry out these recommendations in full, they should inform their Trust s risk management team so that the Trust can take a view on the risks of not following the national recommendation and plan a route map to full compliance. Next steps 1. The outcomes and recommendations of this meeting will be sent to NHS England via Richard Fluck, Gifford Batstone and Jo Martin with a recommendation that the Data Standards Authority should be approached to assist in the design of the messaging system. 2. A questionnaire to ascertain the earliest version of each LIMS provider s software that could be upgraded to contain the algorithm will be designed and distributed to each company invited to the meeting. 3. This group may meet again to discuss the content of the HL7 message(s) and report comments if required. A company limited by guarantee, Registered in England, Registration No Registered Office: Association for and Laboratory Medicine, Tooley Street, London SE1 2TU Website:

4 4. This group may also need to consider some detailed aspects of the algorithm e.g. exactly which parts of the algorithm should be user configurable. Robert Hill 8/8/13 Appendices follow

5 Appendix A AKI Algorithm See accompanying powerpoint slide Appendix B Attendance List LIMS suppliers LIMS Supplier Attendance Apex, Telepath 2000 isoft Ltd (CSC) Winpath Clinisys Solutions Ltd LabCentre/Masterlab Clinisys Solutions Ltd Omnilab Intergrated Software Solutions Absent Swisslab Roche Diagnostics Ltd Absent TrakCare Intersystems Absent Meditech FileTek UK Ltd Absent Ultra GE Healthcare Absent Medipath LRS Health Absent MOLIS Vision4Heath Absent TD Synergy Technidata Absent Apologised Cerner UK PathNet Absent StarLIMS Starlims (Abbott Informatics Solutions) Home Grown Systems Locally Maintained System Northern HSC Trust Locally Maintained System Oxford Universities Hospital Trust Locally Maintained System Royal Berkshire

6 Appendix C Attendance List Clinical Biochemists and Renal Physicians Name Representing Attendance Anne Dawnay Bill Bartlett Christopher Thompson (London) (Dundee) (North Staffs) Finlay MacKenzie UKNEQAS Gethin Roberts Gifford Batstone John Monaghan Jonathan Kay Mike Bosomworth Nick Selby Robert Hill (All Wales LIMS) NHS England (Pathology) (Derby) NHS England (Clinical Informatics Director) (Leeds) (Derby) (ACB Director Scientific Affairs) Absent sent apologies Ron Cullen UK Renal Register Absent sent apologies Shahed Ahmed Trevor Hine Peter Drew (Liverpool) (Liverpool) (Wales) Document control Original version : E Alerts_for_AKI_meeting_statement_ Minor amendments to LIMS suppliers: E Alerts_for_AKI_meeting_statement_270813

7 Addition of IBMS logo: E Alerts_for_AKI_meeting_statement_ Removal of Appendix A (AKI algorithm flowchart) from main document Inclusion of AKI algorithm as a re designed flowchart in an accompanying powerpoint document. Inclusion of modified criterion for AKI 3 for use with children under 18 years.