TIM O CONNELL M ENG M D INFORMATICS SOLUTIONS

TIM O’CONNELL, M. ENG, M. D. INFORMATICS SOLUTIONS TO REAL-WORLD PROBLEMS IN EMERGENCY RADIOLOGY

DISCLOSURES • Speakers fees, Siemens Healthcare • Training sponsorship, Toshiba Canada • President, Resolve Radiologic Ltd

GOALS FOR TODAY’S TALK • a. Learn about some of the safety issues around the time -sensitive workflows in emergency radiology • b. Understand how visual control systems can be used to improve radiologist awareness of workflow and safety issues in emergency radiology, including inter physician communication • c. Learn how customized informatics solutions can be deployed in the radiology department settings to improve the quality and safety of care

GOALS FOR TODAY’S TALK • Learn what is possible • I’m not saying “this is what you should do” • I’m saying “some of this you can do”

DEFINITION: VISUAL CONTROL • Visual control is a business management technique employed where information is communicated by using visual signals • This entails making problems, abnormalities, or deviations from standards visible to everyone http: //en. wikipedia. org/wiki/Visual_control

BACKGROUND: VISUAL CONTROL

BACKGROUND: VISUAL CONTROL

ER RADIOLOGY: 3 BIG CHALLENGES • Get the right imaging for the right patient • Get the right results to the right person • Do both of the above in the right timeframe

SIMPLE? Patient Radiologist ER Physican If all we had were 3 people, it would still be hard

BUT IN ANY GIVEN CASE… Patient Radiologist ERP Unit Clerk Tech Rad ER Consult. Trainee Porter Nurse Consult. Family Staff MD

WHAT WE NEED VS WHAT WE HAVE • Need: • Easy to use, reliable, safe tools for communication of our complex workflow between the people who need information • Have: • Pagers - 1950 • Fax machines - 1966 • Paper forms - 2 nd century BC

AND OUR PROCESSES ARE SIMPLE, RIGHT? Ordering Interpretation Results

NOT REALLY… Reiner, B. I. (2014). Hidden Costs of Poor Image Quality: A Radiologist's Perspective. Journal of the American College of Radiology : JACR, 11(10), 974– 978.

COMPLEXITY = RISK • Taking too long to image a case • Imaging a case incorrectly • Taking too long before interpreting a case • Communicating incorrect preliminary results • Misunderstanding/miscommunication of results • Distribution of results to wrong people

NOT LIKE THE GM PLANT • Traditional business analysis and solutions don’t apply • If a car assembly line gets backed up, the cars don’t spontaneously explode • But in ER radiology, delays to diagnosis or poor communication can mean increased morbidity or mortality

PROBLEM LIST • Order management • Image study technique/monitoring • Radiation Dose and Protocol Management • Report management/workload • Interpretation Quality and Efficiency • Critical results communication

ORDER MANAGEMENT

ORDER MANAGEMENT Reiner, B. I. (2014). Hidden Costs of Poor Image Quality: A Radiologist's Perspective. Journal of the American College of Radiology : JACR, 11(10), 974– 978.

CURRENT STATE IN VGH ER: ORDER MANAGEMENT • Requisition printer in the ER Reading Room • All CT/US/MR studies require a telephone call from the EP • After call, we get requisition, protocol it, and get it to CT/MR tech or trainee to perform US

PROBLEM LIST: ORDER MANAGEMENT • EPs who forget to call us • EPs who call us but forget to enter the order • When we forget to protocol/deliver requisition

SOLUTION: SCHEDULED STUDIES DISPLAY 50” Monitor and mini PC - $1400 186 lines of code - $600 (free!) Access to Modality Worklist Server

SOLUTION: ORDERED/SCHEDULED STUDIES DISPLAY • Future state is to display both ordered and scheduled studies • Pending access to RIS database • Will avoid risk where EP forgets to call us

STUDY TECHNIQUE / DOSE & PROTOCOL QA

STUDY TECHNIQUE/DOSE AND PROTOCOL QA Reiner, B. I. (2014). Hidden Costs of Poor Image Quality: A Radiologist's Perspective. Journal of the American College of Radiology : JACR, 11(10), 974– 978.

CURRENT STATE: STUDY TECHNIQUE Real-time monitoring of ER CT and MR scan consoles 50” monitor - $800 DVI Splitter - $160 DVI-over-Cat 5 Converter - $300 50’ Cat 5 Run - $400 50” monitor - $800 DVI Splitter - $160 DVI-to-H. 264 Converter - $1300 Network Jack - $250 Total Cost: ~$4500

BENEFITS: REAL-TIME MONITORING • Appendix checking with MRI tech for MR Appendicitis • Position of ROI for CT Bolus Tracking, CT Scan Range • Correct CT Protocol Selection • Real-time Diagnoses we have seen: • Acute intracranial hemorrhage • Aortic dissection • AAA Rupture • Massive PE • Pneumothorax

DOSE AND PROTOCOL QA • Two tools: • Real-time CT Dose Database • 70” display in reading room • Offline CT Protocol Analyzer • Interactive, web-based tool

Irradiation Event Time Name of Irradiation Event Study Description Bodypart CT Protocol k. V Patient Size

Effective Pitch SSDE Dose Modulation CTDIvol DLP Irradiation Duration m. As Scan Length Difference in dose due to: 1. Longer (Z-axis) topogram 2. Larger patient

Identical protocols, with different names - time to clean up the scanner!

IT CAN’T HAPPEN HERE… L-Spine CT, Brigham & Women’s Hospital, 2013 Total dose: 4683. 6 m. Gy 70 m. Sv

BUT IT DOES…

REAL-TIME DOSE DISPLAY: BENEFITS • Immediate feedback on old/improper protocol use • Education about scan protocols and dose metrics for trainees • Information for clinicians - shows real-time impacts of imaging

OFFLINE PROTOCOL ANALYZER • Interactive, web-based tool • Allows user to upload CT protocol export from scanner • User can then browse protocols

OFFLINE PROTOCOL ANALYZER Upload Screen

OFFLINE PROTOCOL ANALYZER List of uploaded protocols

OFFLINE PROTOCOL ANALYZER Scan Parameters Reconstructions

OFFLINE PROTOCOL ANALYZER: BENEFITS • Allows users to diagnose protocol issues without interrupting scanner workflow • Educational tool for trainees • For accreditation requirements for CT protocol reviews

INTERPRETATION QUALITY AND EFFICIENCY

INTERPRETATION EFFICIENCY Reiner, B. I. (2014). Hidden Costs of Poor Image Quality: A Radiologist's Perspective. Journal of the American College of Radiology : JACR, 11(10), 974– 978.

INTERPRETATION: CURRENT STATE • Always a drive to improve interpretation efficiency and quality in the ER setting • Efficiency problem: • Prior images often not available • Radiology reports can be time consuming to read • Quality problem: • No good method to follow up future studies • Self QA, learning, teaching cases

PRIOR REPORTS: NLP • NLP: Natural Language Processing • Computer-based interpretation and understanding of text NN: NOUN, SINGULAR JJ: ADJECTIVE IN: PREPOSITION NNS: NOUN, PLURAL CC: COORDINATING CONJUNCTION VBZ: VERB, 3 RD PERSON SINGULAR VBP: VERB, NON-3 RD PERSON SINGULAR RB: ADVERB VBN: VERB, PAST PARTICIPLE

+ + Open Source NLP Toolkit + + +

FUTURE REPORTS: TRACKER • In interesting or uncertain cases • Need a way to flag the case as interesting and automatically receive updates about the patient if/when future imaging occurs • Benefits: • Education • QA

• Future improvement • Track by body part • Track surgical/discharge diagnoses

REPORTING MANAGEMENT/WORKLOAD

REPORTING MANAGEMENT Reiner, B. I. (2014). Hidden Costs of Poor Image Quality: A Radiologist's Perspective. Journal of the American College of Radiology : JACR, 11(10), 974– 978.

REPORTING MANAGEMENT: CURRENT STATE • At our institution: • Clinicians use Philips i. Site • Radiologists use Agfa Impax • Clinicians have no informaiton about status of study interpretation or which radiologist is reading • Problems with ER waiting on reports • Problems with ER getting interpretations if studies read by non- ER radiologists

REPORTING MANAGEMENT: SOLUTION ER Workload Tracker - on a 70” display in ER reading room and in ER Nursing Station

ER WORKLOAD TRACKING BOARD CT CR U S MR

ER WORKLIST TRACKING BOARD

REPORTING MANAGEMENT: BENEFITS • At-a-glance views of: • Current workload status (all red = get to work!) • Who is reading studies - can redirect ER physicians • New studies from other sites

CRITICAL RESULTS COMMUNICATION

REPORTING MANAGEMENT Reiner, B. I. (2014). Hidden Costs of Poor Image Quality: A Radiologist's Perspective. Journal of the American College of Radiology : JACR, 11(10), 974– 978.

CRITICAL RESULTS: CURRENT STATE • Verbal reports from ER Radiologist or trainee for all CT/MRI/US studies • CRs - ERP can leave an impression in Impax, but most use i. Site and often do not leave anything

CRITICAL RESULTS: CURRENT STATE • Problem #1: • Poor documentation by rads and trainees about delivery of critical results • “The responsible physician was informed at approximately 3: 00 pm” • No name • No specific findings • No method of communication documented • Time may be incorrect, and may be important

CRITICAL RESULTS: CURRENT STATE • Problem #2: • EPs claiming they weren’t told about incidental findings • EPs who don’t want to know about incidental findings • Patients without primary care providers

CRITICAL RESULTS: CURRENT STATE • Best effort at present, but getting contentious with EPs • Need a way to reliably document: • What was communicated • With whom • When • How • In a way that always requires all of the above and can be queried and audited (i. e. not the radiology report)

CRITICAL RESULTS: INTERIM SOLUTION One click from PACS: Web-based Critical Notification Record

Similar to BWH ANCR

CRITICAL RESULTS: FUTURE STATE Philips Intellispace Critical Findings Module

SUMMARY • Visual Control tools are effective for communication of scan information, workload and study status in emergency radiology environment, and can improve quality and safety • Small, inexpensive informatics projects can work - you don’t need to wait for big projects! • Be wary of projects that ‘boil the ocean’

SUMMARY • Going forward: still lots to do • All of these projects require ongoing maintenance • Biggest ongoing problem is the lack of hospital IT support • Need: • Clinical decision support • Electronic protocoling • Electronic, secure result delivery

THANK-YOU • Dr. John Mayo, Department Head, Radiology, VGH/UBC Site • Dr. Savvas Nicolaou, Section Head, Emergency Radiology, VGH • Dr. Luck Louis, VGH ER Radiologist • Dr. Patrick Mc. Laughlin, VGH ER Radiologist • Dr. Aaron Sodickson, BWH ER Radiologist • Dr. Jeff Hu, PGY-1 Intern • Cornell Ra, Jay Wong, and Roger Mac. Donald, VGH PACS IT Support

Thank you!