Royal College of Surgeons in Ireland Coláiste Ríoga na Máinleá in Éirinn Usability and Clinical Decision Support Systems Oxford, 29th September 2010 Derek Corrigan, Borislav D. Dimitrov, Tom Fahey PHS / Department of General Practice Overview • What is a Clinical Decision Support System? • What are the technical barriers to be considered? – “usability” issues • What role can Clinical Predication Rules (CPRs) play in a CDSS? • What form might a state of the art CPR based system look like? • What are the challenges to achieving this state of the art? PHS / Department of General Practice Clinical Decision Support Systems • Clinical decision support system (CDSS) – I.T. based systems that are designed to improve clinical decision making by suggesting possible clinical actions/interventions based on an underlying clinical evidence base • CDSS Types – Diagnostic systems – diagnose appendicitis – Reminder systems – generate patient letters – Disease management systems – diabetes management – Drug dosing/prescribing – insulin, warfarin PHS / Department of General Practice CDSS - Barriers to Implementation • Human Issues – As previously discussed - knowledge, attitudes, behavior • Technical Issues – Need to consider usability issues of CDSS and how we can potentially use CPRs as part of them PHS / Department of General Practice CDSS - Usability Issues 10 Commandments for Effective CDSS – (Bates et al 2003) • CDSS must fit into the clinical workflow – logical sequence and steps – Primary care diagnosis tool on laptop or PC in practice as part of consultation – Secondary care drug administration as part of round system may need something more mobile and compact – palm held application – No point developing tools that can only be accessed outside of the clinical process workflow – they are then seen as an additional overhead • Speed is everything – users will not tolerate slow systems! – No matter how useful an application is, it will not be used if performance is prohibitively slow – need to prototype and user test during development PHS / Department of General Practice CDSS - Usability Issues (contd.) • Deliver needs in real time – relevant information presented as needed - user should not need to go searching for information – In drug prescription systems, one action may imply another associated action that the system can automatically prompt for e.g. drug interactions, corollary orders or order sets – Do not wait until end of the entire process to highlight issues to the user • Physicians will resist stopping – offer alternatives and overrides – Users will not accept blanket restrictions – result in lack of confidence and frustration – Provide alternative suggestions justified by clinical evidence and always allow override of recommended actions – ultimately the clinician must be trusted PHS / Department of General Practice CDSS - Usability Issues (contd.) • Changing direction in a system is easier than stopping – – – – • Can modify clinician behavior with beneficial results Change default medication dosages based on evidence Change default frequency of dosage based on evidence Bates et al detail a change resulting in cost savings of $250,000 per year Monitor impact of CDSS and usage patterns and revise – – – – Understand how clinicians are using your system Track underlying usage of rules, alerts and overrides Large levels of overrides indicate potential problem Amend evidence base where necessary and/or provide more detailed justification PHS / Department of General Practice CDSS - Usability Issues (contd.) • Simple interventions work best – CPRs rather than books of guidelines – – • Simple rules may offer easiest gains e.g. Ottawa Ankle Rule Working clinicians will not have time to review pages of guidelines – these should be provided as supporting documentation where necessary Usability testing – include the users! – – No point having state of the art evidence base with a bad user interface Get user feedback through prototyping early in the design process PHS / Department of General Practice CDSS - Usability Issues (contd.) • Manage your knowledge base – needs to be easily updated – • Ensure that formal evidence base review and update is done regularly based on latest clinical evidence Provide patient specific recommendations based on electronic patient record (EPR) – Ideally integrate CDSS as part of EPR – Use individual patient data to generate individual recommendations PHS / Department of General Practice CDSS - Clinical Prediction Rules • Evidence Base – Potentially can use Clinical Prediction Rules (CPRs) as part of a clinical guideline to drive evidence based care in a CDSS • Issues to consider – Level of CPR validation of potential CPRs – Type of clinical setting where validation of CPR has been done – Impact assessment of CPR – does it indicate a high quality CPR with impact for patient, physician or process? PHS / Department of General Practice CDSS – Maintaining the Evidence Base • Huge potential evidence base in EHRs in primary care e.g. General Practice Research Database in UK (GPRD) • Create clinical trials/ epidemiological studies from primary care data • May be used to amend an existing CPR i.e. the inclusion of additional symptoms/signs identified by data analysis as potential diagnostic cues • May be used to create new CPR’s based on new diagnostic cue combinations • Evidence base becomes “self-learning” and adds or improves the list of CPR’s used as the evidence base PHS / Department of General Practice The TRANSFoRm Project PHS / Department of General Practice TRANSFoRm Services 1 CPR Repository Clinical Prediction Rules Service 2 Distributed GP EHRs With CDSS 3 Research Study Designer Study Criteria Design CP Rules Manager CP Classifier 5 CPR Data Mining and Analysis CPR Analysis & Extraction Tool Find Eligible Patient 4 Research Study Management Recruit Eligible Patient Study Data Management PHS / Department of General Practice CDSS - Challenges The EPR – Still a huge amount of work to be done just to create usable EPRs System Interoperability – consistent representation of clinical data using technologies – Health Level 7 (HL7), OpenEHR – pool and share clinical data by allowing systems to ‘talk together’ Semantic Interoperability – being able to interpret clinical concepts contained within data in a consistent way – Clinical Coding Terminologies such as SNOMED-CT, ICPC, ICD PHS / Department of General Practice In Summary • CPRs have huge potential as a form of evidence in CDSS • Need to consider technical usability issues as well as human issues • Future directions - develop the Electronic Patient Record and incorporate research tools to generate and update the evidence base – “self-learning” PHS / Department of General Practice Thank You Discuss! PHS / Department of General Practice