IHI Expedition Eliminating Overuse in Medical Imaging Jim Duncan, - - PDF document

2/18/2014 1

IHI Expedition

Eliminating Overuse in Medical Imaging

Wednesday, February 19, 2014

These presenters have nothing to disclose

Jim Duncan, MD, PhD Kelly McCutcheon Adams, LICSW

Expedition Coordinator

2

Kayla DeVincentis, CHES, Project Coordinator, Institute for Healthcare Improvement, currently manages web-based Expeditions and the Executive Quality Leaders Network. She began her career at IHI in the event planning department and has since contributed to the State Action

• n Avoidable Rehospitalizations (STAAR) Initiative, the

Summer Immersion Program, and IHI’s efforts for Medicare-Medicaid enrollees. Kayla leads IHI’s Wellness Initiative and has designed numerous activities, challenges, and educational opportunities to improve the health of her fellow staff members. In addition to implementing the

• rganization’s first employee health risk assessment, Kayla

is certified in health education and program planning. Kayla is a graduate of Northeastern University in Boston, MA, where she obtained her Bachelors of Science in Health Science with a concentration in Business Administration.

2/18/2014 2

WebEx Quick Reference

• Welcome to today’s

session!

• Please use chat to “All

Participants” for questions

• For technology issues only,

please chat to “Host”

• WebEx Technical Support:

866-569-3239

• Dial-in Info: Communicate /

Join Teleconference (in menu)

3

Raise your hand Select Chat recipient Enter Text

4

When Chatting…

Please send your message to All Participants

2/18/2014 3

Expedition Support

All sessions are recorded Materials are sent one day in advance Listserv address: medicalimaging@ls.ihi.org

– Sends an email to all participants and faculty – Use only for questions relevant to all participants – To add yourself or colleagues, email us at info@ihi.org

5

Where are you joining from?

2/18/2014 4

Expedition Director

7

Kelly McCutcheon Adams, LICSW has been a Director at the Institute for Healthcare Improvement since 2004. Her primary areas of work with IHI have been in Critical Care and End of Life Care. She is an experienced medical social worker with experience in emergency department, ICU, nursing home, sub- acute rehabilitation, and hospice settings. Ms. McCutcheon Adams served on the faculty of the U.S. Department of Health and Human Services Organ Donation and Transplantation Collaboratives and serves on the faculty of the Gift of Life Institute in

• Philadelphia. She has a B.A. in Political Science from

Wellesley College and an MSW from Boston College.

Today’s Agenda

8

Your stories about overuse and good decisions Measuring overuse Action period assignment

2/18/2014 5

Expedition Objectives

At the end of this Expedition, participants will be able to: List common examples of medical imaging

• veruse

Explain strategies for reducing overuse in medical imaging Plan tests to make changes in own environment Utilize tools to assess what changes generate improvement

9

Schedule of Calls

Session 1 – The Harm Caused by Overuse in Radiology Date: Wednesday, February 5, 1:00 PM – 2:30 PM ET Session 2 – Measuring Overuse Date: Wednesday, February 19, 1:00 PM – 2:00 PM ET Session 3 – Strategies for Eliminating Overuse Date: Wednesday, March 5, 1:00 PM – 2:00 PM ET Session 4 – Measuring What Changes Lead to Improvement Date: Wednesday, March 19, 1:00 PM – 2:00 PM ET Session 5 – Sustaining the Gains Date: Wednesday, April 2, 1:00 PM – 2:00 PM ET

10

2/18/2014 6

Faculty

11

Jim Duncan, MD, PhD, is a Professor of Radiology and the Chief Quality and Safety Officer for the Mallinckrodt Institute of Radiology at Washington University School

• f Medicine. He maintains a clinical practice in

interventional radiology and divides his time between St. Louis Children's Hospital and Barnes-Jewish Hospital in

• St. Louis. Dr. Duncan works on multiple quality and

safety improvement initiatives for both local and national

• rganizations. He has a BS from the University of

Michigan as well as an MD and PhD in Cellular and Molecular Biology from Washington University. He completed the IHI Improvement Advisor Professional Development Program in 2012.

Session 2: Measuring Overuse

Choosing a focus for your improvement effort

– Your examples of overuse in imaging

Measurement strategies

– Start simple: Leverage current data sources – Using current performance as a baseline

12

2/18/2014 7

13

“All Teach, All Learn”

What examples of overuse did you find in your organization? How is your organization measuring CT Dose?

13

Tell us about your frontline teams – how are they successfully eliminating overuse

• r measuring CT dose?

Starting to Measure

Dose per image

– Data source: CT Dose Reports

14

Images per study

– Dual Scans – Data source

– Billing records

2/18/2014 8

Poll Questions

How are you capturing CT dose information at your site?

– Not yet capturing information on CT doses – Manual methods – Mixture of manual and automated methods – Fully automated process – Not sure

15

Capturing CT Doses

Steps

– CT scan performed

– Scanner calculates

dose metrics

– CTDIvol, DLP

– Copy these values

into the Radiology report

– Which value do we

enter into the report?

16

2/18/2014 9

Digression: Metrics versus Intent

Intent = preventing cancer

– Should be measuring DNA mutations that involve oncogenes

Intent = preventing skin injury

– Should be measuring cell death at particular skin locations

Rare that we measure exactly what we want

– Measurements as surrogates

– Cancer = radiation dose to sensitive tissues – Skin injury = peak skin dose

Possible to learn from imperfect data

– Infer the result of interest – Waiting for better data usually just delays improvement

17

Operational Definitions

"An operational definition is a procedure agreed upon for translation of a concept into measurement of some kind." - W. Edwards Deming Example: Radiation dose for a Head CT

– Need to define process for measuring “radiation dose”

– CTDI (CT Dose Index) is “dose/CT slice” measured in a phantom – DLP (Dose Length Product) is CTDI x number of slices – Other measures such as SSDE: size specific dose estimate

– Need a process for identifying which exams are “Head CTs”

– Names within Radiology Information Systems are not standardized

– “Head CT” vs “CT Head” vs “CT Brain” vs “Std Head"

– Options: CPT codes, RADLEX, etc.

18

2/18/2014 10

Data Standards in Medical Imaging

DICOM: Digital Imaging & Communications in Medicine

– International standard: file format analogous to JPEG, TIFF,

• PNG. Includes numerical values for each pixel as well as

metadata

– Metadata fields: dose metrics, patient identifiers, date/time, etc.

– DICOM files are transferred, stored and viewed using Picture

Archiving and Communications Systems (PACS)

CPT: Current Procedural Terminology

– US standard: nomenclature for imaging procedures

– 70450 = noncontrast head CT

– CPT codes are used by Radiology Information Systems (RIS) to

• rder exams, manage workflow, organize reports and bill

procedures

19

Man vs Machine

Manual (invites human error)

– Dictation/transcription errors – Data embedded in reports – Applying rules

Assessing compliance

– Importance of feedback

– Timely – Specific

20

2/18/2014 11

Improvement Cycle

21

Improving Data Capture

Automated data capture

– Improved accuracy and reliability – Much more detailed datasets

22

2/18/2014 12

Poll Question

Where would you go to look up CT doses for a series of patients?

– Information is not available – Radiology reports – PACS (Picture Archiving and Communication System) – RIS (Radiology Information System) – Dose database – Not sure

23

Analyzing Data

“We are drowning in data, what we seek is knowledge.” Deming

24

2/18/2014 13

Data Analysis

Importance of predictive models

– Quality = conformance to expectation – Substantial deviation from expectation is an informative event

Visualizing data: CT volume at St Louis Children’s Hospital

25

Transforming Data into Knowledge

Data vs Information

– Data: raw numbers/observations – Information: determined by patterns within the data

Knowledge is the ability to predict future events

26

2/18/2014 14

Learning and Knowledge

Learning

– We learn nothing new about the inner workings of our system

when the observed results agree with our expectations

Unexpected events are informative

– To learn about a system we need to find and study anomalies – Use the new knowledge to manipulate the system so that it

fulfills our needs

27

What could have happened in 2005 that turned linear growth into linear contraction?

Using Data to Understand the System

28

Possible causes for the observed decrease in CT volume

– External forces

–

Decrease in overall volume? – Importing outside CT

studies?

–

Made possible by DICOM standard – Awareness

–

Image Gently campaign led to changes in behavior

–

Conversations about radiation doses from CT

2/18/2014 15

Communication Example

Sharing CT dose information with customers

– Patients, Referring Physicians, Medicaid, etc.

Simplify the conversation

– Quality = conformance to expectation

– Use predictions as anchors – Teach customers to monitor variation from prediction

Lab results as an analogy

“But Medicine is Different”

“You cannot apply ideas from _________* because people get hurt if we make mistakes”

*brewing beer, information theory, statistical process control, …

2/18/2014 16

Preparing for the next session

How comfortable are you and your team with:

– Explaining the different causes of variation? – Creating and interpreting run charts? – Creating and interpreting control charts?

31

Action Period Assignment

Pick a measure

– Examples

– CT Volume – Compliance with CT recording – CT Dose metrics for a common exam (e.g. CT Head) – Dual scan rate – Volume of imported studies – You pick

Assess baseline performance in your system

– Looking for volunteers to submit their data

– Next session will include tools for analyzing data

2/18/2014 17

Assignment and Next Sessions

Assignment: Baseline measurements

– Tracking 1-2 metrics of overuse in your organization

Session 3

– Testing strategies for improvement (PDSAs)

Session 4

– Determining what changes led to improvement

Session 5

– Sustaining the gains, expanding to all aspects of imaging

Questions?

34

Raise your hand Use the Chat

Just for fun: Use the chat to tell us your diagnosis for each of these two images

2/18/2014 18

5x5 pixels .03K 9x9 pixels .08K 18x18 pixels .24K 37x37 pixels 1.0K 75x75 4K 300x300 64K 600x600 256K 150x150 16K

Earth Rise Apollo 8 12/24/1968

Pattern Recognition

8x5 pixels .04K 16x11 pixels .18K 32x23 pixels .75K 64x47 pixels 3.0K 128x94 12K 512x375 192K 1024x750 768K 256x187 48K

Jackson Pollock Number 1 circa 1947

Random Pattern

2/18/2014 19

Expedition Communications

Listserv for session communications: medicalimaging@ls.ihi.org To add colleagues, email us at info@ihi.org Pose questions, share resources, discuss barriers or successes

37

Next Session

Session 3 – Strategies for Eliminating Overuse Date: Wednesday, March 5, 1:00 PM – 2:00 PM ET

38