for Critical Result Notification

Transcription

1 Development of a Push Alert System for Critical Result Notification Challenges, Successes, and Lessons Learned Dr. Michael Feldman Assistant Director, Clinical Informatics Medical Director, Sunnybrook Centre for Prehospital Medicine Sunnybrook Health Sciences Centre Toronto, Ontario

2 Conflicts None to declare 2

3 Ontario Pay for Results Variable Incentive Funding Ontario s Wait Time Strategy to improve performance of hospital emergency departments CTAS 1-3 patients 90th percentile ED length of stay <8 hours CTAS 4-5 patients 90th percentile ED length of stay <4 hours 3

4 Push Alert Notification Broadly felt that time to disposition (treatment, admission, discharge) was subject to bottlenecks Delivery of results - one of the bottlenecks? Questionnaire completed by ED physicians identified critical laboratory results, imaging results, and normal cardiac enzymes represented decision points relating to patients disposition or treatment 4

5 Utility of Critical Result Notification Audit of critical result notification in 121 US institutions Median time to notification of a licensed health care provider 7 minutes Median time for specimen collection and processing 48 minutes Median time for initiation of therapy 108 minutes Valenstein et al, Arch Pathol Lab Med. 2008;132:

6 Patient Safety Recommendations Massachusetts Coalition for Patient Safety (2005) Who: ordering provider, or a provider directly responsible for the patient, or who can take action What: prioritize list of critical values that require accelerated notification When: appropriate time frame for notification How: e.g. computerized push alert system; verbal notifications; backup or fail-safe system Uniform process should be developed for all diagnostic studies Tracking, support, and maintenance 6

7 Computerized Notifications Adoption of computerized push system for critical result notification in Padua, Italy Decreased mean time for notification from 30 minutes to 11 minutes Increased rate of successful notifications (i.e. correct physician within 1 hour) Piva et al, Am J Clin Pathol, 2009;131:

8 System Design, Part I Lab results sent to port on existing ED server If ordering physician is an ED physician on EDIS, the HL7 data is parsed If critical value is found, a message is sent to a handheld device belonging to the responsible ED physician Accurate list of MRP is key to delivering the information to the right person 8

9 System Design II Separate system for preliminary crosssectional imaging reports Worked with IT to add function to IMPAX Radiology resident must push a button to send the result Relationships with IMPAX IT personnel 9

10 System Design III How to deliver messages? Text messaging, s, web page Which devices? Considered data only vs. data + voice functionality (no more than one device on your belt) Considered iphone, Blackberry, Android platforms Final choice is platform-independent uses s Hospital IT supports iphones 10

11 System Design Laboratory Information System Voice EDIS WebER server IMPAX Apps: Sanford Guide Tarascon PdiSTAT Pedi STAT MedCalc

12 Security and Privacy Hospital Privacy Office assessment (get to know your Privacy Officer) Sending information via internet Receipt of information can only be hospital accounts Loss or misuse of handheld device Hospital IT security assessment Required iphones only MobileIron : controls device access to hospital network; allows remote wipe 12

13 Device Management Physicians responsible for security of device; IT responsible for maintenance

14 Evaluation Identification of key colleagues interested in research/assessment Before/after? Observational study? Randomized trial? Questionnaire? Important t to study this to maintain i ongoing funding for program 14

15 Real-world Performance Nuisances vs. useful alerts Number/frequency of alerts Workflow Clinical significance of critical value (K+ 5.9 mmol/l vs. 7.1 mmol/l) Device problems Failure to connect to hospital network Failure to synchronize s Poor cellular network coverage of department 15

16 Lessons Learned Clinician-oriented design Preference for single voice/notification device with apps Corporate-managed devices Lock down settings Responsiveness to clinicians Close contact with IT team, Privacy Officer Frequent updates to clinicians Develop relationships with colleagues in IT, Privacy Office, Medical Imaging departments 16

17 Successes Devices work as planned Databases track performance so all significant errors have been traced to users Significant ifi attention ti has been generated these will be a target of study Stay tuned! 17