Building an HMI for the Medical IoT

Building on the Integrated Clinical Environment (ASTM F2761) architecture, DocBox has created a Medical Internet of Things (IoT) platform that enables the development and use of clinical apps. This talk describes the technical and design characteristics of the DocBox Human Machine Interface (HMI) for the patient bedside, and the potential for data driven healthcare.

Transcript

1. Building an HMI for the Medical IoT Jereme Givens-Lamothe, Senior Software Architect at DocBox Inc
2. Agenda • About DocBox Inc • DocBox Platform • Clinical Flowsheet App • Patient Bedside HMI 2
3. About DocBox Inc • Boston-based • Founded in 2007 • Creating a Medical IoT platform to... • Improve patient safety • Improve efficiency & quality of healthcare • Enable data driven hospitals 3
4. DocBox Platform 4
5. DocBox Platform - Bedside • Integrated Clinical Environment • ASTM F2761 • System Apps • Clinical Apps • Clinical Decision Support • Workflow Automation • Medical Devices • ICE-Compliant Medical Devices • Driver support for legacy devices 5
6. DocBox Platform - Software Stack • Ubuntu GNU/Linux • Qt Framework • Qt Application Manager • RTI DDS Connext • IIoT Connectivity Framework • ICE SDK • Patient-Centric Data Model 6 Ubuntu GNU/Linux Qt RTI DDS Application Manager App Device ICE SDK Equipment
7. DDS - Data Distribution Service • Connectivity Framework for the IIoT • (Industrial Internet of Things) • Data-Centric Publish-Subscribe • OMG Standard • RTI Connext DDS • Used in industrial, military and medical applications 7 https://www.rti.com/products/dds
8. Clinical Flowsheet App • Provides workflow automation, clinical decision support for nursing staff • Displays events, patient trends, and is used for clinical documentation • Integrate DocBox's patient-centric data model into a modern, easy-to-use UI • Built using Qt Quick • Custom Controls, UI style • Item Models expose data from ICE 8
9. Patient Trends • Use Qt Charts to display 18 hour history of multiple patient vitals • LineSeries represents a metric on a device • OpenGL backend supports many points • One sample a second for 18 hours means 64k samples per LineSeries! • Multi-touch gestures for timeline navigation • PinchArea enables zooming in timeline • MouseArea enables forward / backward navigation through time 9
10. Events Visualization • Label Series - Display adjacent text • Custom Qt Charts Series • Callout - Display detail • Respond to user touch events • Point Merging • Calculate a merge threshold in JS • Custom Item Model does the work • Synchronized to Trends date time axis 10
11. DocBox On-Screen Keyboard • DocBox has a multi-touch capacitive screen • No physical keyboard • Significantly customized layouts • Custom style, to match look ’n feel of our apps • Expanded functionality • Always present full screen • Present the question or field name • Multiple input fields 11
12. Clinical Flowsheet - GOMS Analysis • Nurses spend significant time completing documentation • DocBox wanted to quantify any workflow improvements introduced by Clinical Apps and report generation • Preliminary GOMS Analysis of ICU Nurse workflow • Comparing handwritten to digital report • On average, reduced documentation time by 75.8%! • Primarily due to automation of data collection 12
13. HMI for the Medical IoT • HMI uses the Qt Application Manager • Implements Qt Wayland Compositor • EGLFS platform plugin on Ubuntu / Intel • System UI has two major responsibilities • Window Management / Navigation • Lock Screen 13
14. HMI for the Medical IoT • DocBox Apps are "native" executables • Apps communicate w/ ICE over DDS • System Apps comprise a standard core • Clinical Apps can be installed per patient bed • Clinical Apps are medical devices! • D-Bus vs. DDS? 14
15. Recap / Conclusion • Many apps, many devices make a Medical IoT • DocBox is enabling data driven hospitals • Apps are built w/ Qt, communicate via DDS • … and speak a common data model! • Qt enabled a small team to implement big ideas 15
16. Thank you! email - jlamothe@docboxinc.com web - http://www.docboxmed.com twitter - @DocBoxMed