Lumbar Imaging with Reporting of Epidemiology (LIRE) Jeffrey - - PowerPoint PPT Presentation

Lumbar Imaging with Reporting

• f Epidemiology (LIRE)

Jeffrey (Jerry) Jarvik, M.D., M.P.H.

Director, Comparative Effectiveness, Cost and Outcomes Research Center

Bryan A. Comstock, MS

Operations Director, Center for Biomedical Statistics

Brian Bresnahan, PhD

Health Economist, Dept. of Radiology

Nick Anderson, PhD

Associate Director, Bioinformatic Core, ITHS

Key People

UW

• Jerry Jarvik, MD, MPH- PI
• Katie James, PA-C, MPH-

Project Director

• Bryan Comstock, MS- Biostats
• Nick Anderson, PhD-

Bioinformatics

• Brian Bresnahan, PhD- Health

Economist

• Patrick Heagerty, PhD- Biostat
• Judy Turner, PhD-

Psychologist/Pain expert Non-UW

• Rick Deyo, MD, MPH-OHSU
• Dan Cherkin, PhD-GHRI
• Rene Hawkes- GHRI
• Safwan Halabi, MD-HFHS
• Dave Nerenz, PhD- HFHS
• Dave Kallmes, MD- Mayo
• Jyoti Pathak, PhD- Mayo
• Patrick Luetmer, MD- Mayo
• Andy Avins, MD, MPH-KPNC

Disclosures

• Physiosonix (ultrasound company)

–Founder/stockholder

• Healthhelp (utilization review)

–Consultant

• Springer: Evidence-based Neuroradiology

–Co-Editor

• GE Healthcare: CER Advisory Board (past)

–Consultant

Background and Rationale

• Lumbar spine imaging frequently

reveals incidental findings

• These findings may have an

adverse effect on:

–Subsequent healthcare utilization –Patient health related quality of life

Prevalence of Disc Degeneration s LBP

Modality Author/ Year Age Range Prev

MR Boden/ 1990 20-60 60-80 44% 93% MR Stadnik/ 1998 17-60 61-71 52% 80% MR Weishaupt/ 1998 20-50 72-100% MR Jarvik/ 2001 35-70 91%

Disc Degeneration in Asx

Conceptual Model

Diagnostic Test Normal TN: Reassurance FN: False Reassurance Abnormal TP: Anxiety FP (including incidental): Needless Anxiety

Conceptual Model

Diagnostic Test Normal TN: Reassurance FN: False Reassurance Abnormal TP: Anxiety FP (including incidental): Needless Anxiety

LIRE target

Therapeutic Value of Diagnostic Test

(Sox et al Ann Int Med 1981)

• Pts with non-cardiac chest pain

randomized to ECG+CPK vs. no tests

• Pts getting tests showed less short

term disability

• Conclusion: testing can directly

improve HRQOL via reassurance

Natural History of Low Back Pain and Radiculopathy- Modic et al:

Radiology 2005: 235;297

• 246 subjects from primary care and ER

w/in 2 wks sx

–150 LBP / 96 radiculopathy –Random allocation

• imaging info (115)
• no imaging info (131)

SF-36 General Health

p=0.07 *p=0.001

Conclusion from Modic et al: Radiology 2005

• Effect of imaging likely mediated

through anxiety produced by findings

• Testing can directly worsen HRQOL

Dx Testing Consequences

Diagnostic Test Normal TN: Reassurance (TVDT) FN: False Reassurance Abnormal TP: Anxiety FP (including incidental): Needless Anxiety

Sox et al

Dx Testing Consequences

Diagnostic Test Normal TN: Reassurance (TVDT) FN: False Reassurance Abnormal TP: Anxiety FP (including incidental): Needless Anxiety

Probability of any lumbar spine finding >90%

Sox et al Modic et al

Martin Roland, Maurits van Tulder

Disc degeneration: Approximately 80%-100% of people without back pain have this, so finding may not be related to patient’s pain.

Lumbar Spine Macro

The following findings are so common in people without low back pain that while we report their presence, they must be interpreted with caution and in the context of the clinical situation (Reference-Jarvik et al,

Spine 2001):

Finding (prevalence in pts without low back pain) Disc degeneration (91%) Disc signal Loss (83%) Disc height loss (56%) Disc bulge (64%) Disc protrusion (32%) Annular fissure (38%)

Support for Clinical Decision Support

• Blackmore et al, JACR 2011

–Used evidence-based decision support tool –Showed sustained decrease of

• 23% for lumbar spine MR for LBP
• 23% for brain MRI for headache
• 27% for sinus CT

LIRE Preliminary Data

• Starting 12/2005, we made the

macro available to insert into reports

• Arbitrary for which patients the macro

was incorporated

• 2/~10 attendings used the macro
• Not randomized, but arbitrary

Hypothesis

• The benchmark information will

influence subsequent management

• f primary care patients with LBP

–Fewer subsequent imaging tests –Fewer referrals for minimally invasive pain treatment –Fewer referrals to surgery –Less narcotic use

Results: Subsequent Imaging Within 1 Yr (retrospective pilot)

p=0.14 OR*=0.22 1/71 12/166

* Adjusted for imaging severity

Results: Subsequent Narcotic Rx Within 1 Yr (retrospective pilot)

p=0.01 OR*=0.29 5/71 37/166

Possible Confounding by Severity

• Arbitrary assignment of macro

shouldn’t be related to severity

• Controlled for age, race,

insurance status, deg severity by imaging (>mod central or foraminal sten, extrusion)

LIRE, The RCT

A pragmatic, cluster randomized trial

Proposed Study Flow

Primary Care Clinics With LBP Patients Randomize Clinics Macro with Epi Info Outcomes Assessment No Macro with Epi Info Outcomes Assessment

LIRE Sites

• Kaiser Permanente

Northern California

– Andy Avins, MD MPH

• Henry Ford Health

System

– Safwan Halabi, MD

• Group Health

Research Institute/GHC

– Dan Cherkin, PhD

• Mayo Clinic Health

System

– Dave Kallmes, MD

4+1 Working Groups and Leaders

• 1. Refinement of benchmark text

Jerry Jarvik

• 2. Implementation of cluster randomization

Bryan Comstock, MS

• 3. Spine intervention intensity measure

Brian Bresnahan

• 4. Electronic data capture

Nick Anderson

• 5. Katie’s WG of 1: IRB, Protocols, Subcontr

LIRE, the RCT UH2 Aims/Working Groups

• Aim 1/WG1: Refine the information to

be included in the radiology report so that it is specific for imaging modality and patient age.

WG1- Refining the Message

• Have identified the most recent

literature

• Abstracted prevalence data that is

modality and age specific

• On target to finish by ~March

2013

Aim/Working Group 2

Bryan Comstock- Biostatistician, Center for Biomedical Statistics, UW

• Develop site-specific deployment

methods for the stepped wedge, cluster randomization scheme.

Choice of Study Design

Stepped Wedge Design

Stepped Wedge Design

• A one-way cluster, randomized

crossover design

• Temporally spaces the intervention
• Assures that each participating

clinic eventually receives the intervention

Advantages of SW Design

• Controls for external temporal trends
• Assures all sites receive intervention
• Participation more palatable for

interventions viewed as desirable

WG2- Progress

• Sites have identified clinics (units of

randomization) and number of primary care providers at each clinic.

• Working with site health system

programmers for placement and timing of benchmark info

Aim/Working Group 3

Brian Bresnahan, PhD- Health Economist

• Develop/validate a composite

measure of spine intervention intensity-a single metric of overall intensity of resource utilization for spine care

Aim/WG 3 (cont.)

• Will convert CPT codes to RVUs as
• ur primary metric of back-related

utilization

• Will validate CPT conversion by

directly pulling RVUs from one site

• Will explore RVU as proxy metric by

examining correlation with disability, pain and HRQOL in BOLD registry

Aim/WG 3 Progress

• Working with site programmers to

pull CPT and RVU data

• Already established data pulls for 2

sites

• Have initial BOLD data for RVU-PRO

analysis

Aim/Working Group 4

Nick Anderson, PhD- Bioinformatics Core, ITHS

• Develop/validate electronic data

methods and tools to capture

• utcomes of interest (subsequent

diagnostic testing, opioid prescriptions, spinal injections, spine surgeries).

Aim 4 Progress

• Already established data pulls from 2

sites for BOLD (Kaiser N. CA and Henry Ford)

• Working with site programmers for

direct EMR pulls

• Considering using VDW at HMORN

sites

Key Aspects of Pragmatic Trial

• Broad inclusion criteria
• Waiver of consent
• Simple, easily implementable

intervention

• Passive collection of outcomes

Key Challenge- IRB Waiver of Consent

• KPNC, HFHS and GHC/GHRI-

–Initial conversations with IRBs reason for optimism for waiver

• Mayo- greater challenge
• UW- full committee review

Key Challenge- IRB Consolidation

• KPNC likely willing to cede to another

HMORN site (GHRI)

• HFHS has apparently never ceded

(there’s always a first time…)

• Mayo- greater challenge
• UW- has cooperative agreement

with GHRI

Key People

UW

• Jerry Jarvik, MD,MPH- PI
• Katie James, PA-C, MPH-

Project Director

• Bryan Comstock, MS- Biostats
• Nick Anderson, PhD-

Bioinformatics

• Brian Bresnahan, PhD- Health

Economist

• Patrick Heagerty, PhD- Biostat
• Judy Turner, PhD-

Psychologist/Pain expert Non-UW

• Rick Deyo, MD, MPH-OHSU
• Dan Cherkin, PhD-GHRI
• Rene Hawkes- GHRI
• Safwan Halabi, MD-HFHS
• Dave Nerenz, PhD- HFHS
• Dave Kallmes, MD- Mayo
• Jyoti Pathak, PhD- Mayo
• Patrick Luetmer, MD- Mayo
• Andy Avins, MD MPH-KPNC

Questions for Audience

1. Any experience with using RVUs as a metric for patient reported outcomes? 2. We want to collect pain NRS from the clinical record. What experience with missing data do people have for clinically collected variables, such as the BPI? 3. What experience do people have with getting HMORN and non-HMORN sites to cooperatively review protocols?

Health Care Systems Research Collaboratory Grand Rounds: Lumbar Imaging with Reporting of Epidemiology

Jeffrey (Jerry) Jarvik, M.D., M.P .H. Bryan A. Comstock, MS Brian Bresnahan, PhD Nick Anderson, PhD January 25, 2013

A Virtual Home for Knowledge about Pragmatic Clinical Trials using Health Systems: www.theresearchcollaboratory.org