Lumbar Imaging with Reporting of Epidemiology (LIRE) Update Jeffrey (Jerry) Jarvik, M.D., M.P.H. Professor of Radiology, Neurological Surgery and Health Services Adjunct Professor Orthopedic Surgery & Sports Medicine and Pharmacy Director, Comparative Effectiveness, Cost and Outcomes Research Center (CECORC) Kari Stephens, Ph.D. Assistant Professor, Psychiatry & Behavioral Sciences Adjunct Assistant Professor, Biomedical Informatics & Medical Education
Acknowledgements • NIH: UH2 AT007766-01; UH3 AT007766 • AHRQ: R01HS019222-01; 1R01HS022972-01 • PCORI: CE-12-11-4469 Disclosures (Jarvik) • Physiosonix (ultrasound company) – Founder/stockholder • Healthhelp (utilization review) – Consultant • Evidence-Based Neuroimaging Diagnosis and Treatment (Springer) – Co-Editor
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 in Normals Modality Author/ Age Prev Year Range MR Boden/ 20-60 44% 1990 60-80 93% MR Stadnik/ 17-60 52% 1998 61-71 80% MR Weishaupt/ 20-50 72-100% 1998 MR Jarvik/ 35-70 91% 2001
Disc Degeneration in Asx
Intervention Text The following findings are so common in normal, pain-free volunteers, that while we report their presence, they must be interpreted with caution and in the context of the clinical situation. Among people between the age of 40 and 60 years, who do not have back pain, a plain film x-ray will find that about: • 8 in 10 have disk degeneration • 6 in 10 have disk height loss Note that even 3 in 10 means that the finding is quite common in people without back pain.
UH3 Hypothesis • For patients referred from primary care, inserting epidemiological benchmark data in lumbar spine imaging reports will reduce: – subsequent cross-sectional imaging (MR/CT) – opioid prescriptions – spinal injections – surgery.
Participating Systems Name # Primary # PCPs Care Clinics (Randomized) (Randomized) Kaiser Perm. 20 865 N. California Henry Ford 26 228 Health System, MI Group Health 19 245 Coop of Puget Sound Mayo Health 36 345 System Total 101 1683
Stepped Wedge RCT
Wave 1 Implementation Site Sub-site Wave 1 Started April 1 st , 2014 Group Health April 1 st , 2014 Henry Ford Mayo April 10 th , 2014 La Crescent, Prairie du Chien April 24 th , 2014 St. James, Austin, Waseca August 27 th , 2014 Plainview June 25 th , 2014 Kaiser
Problems Encountered • People – Wrong skills – Lack of buy-in – Personality fit (or lack thereof) – Political/leadership issues • Structure/System – Multiplicity of data systems – Distributed administration vs. centralized
People: Example #1 • Implementation problems resolved when IT project manager replaced – Solutions rapidly found to implementation problems – Improved communication – Improved buy-in
People: Example #2 • Sudden regionalizing of radiology reporting • Randomization by clinic impossible • UW, site-PI and local leadership found technical solution
People- Lessons • Leverage pre-existing good relationships • Need familiarity w/data systems + personalities • Find team members who are a better fit ASAP • Work with local stakeholders to identify possible interference on horizon
Structure/System: Example #1 Distributed vs. Centralized • Distributed – Clinic autonomy standardization for implementation difficult (e.g. multiple RIS) • Centralized – Standardization efforts can also interfere with implementation (e.g. initiative to standardize radiology reporting)
Structure/System: Example #2 • Dynamic rendering vs. permanent part of EMR – Only way to implement in a timely manner – Required manual verification – For Wave 2, programmer was able to permanently insert intervention into EMR – Uncovered 2 nd problem: intervention tied to where report accessed vs. where order originated
Structure/System: Example #3 • Small Wave2 clinic closed with 2 MDs Wave1 clinic • Stepped-wedge design complicates impact: timing determines exposure
Structure/System Lessons • Centralized vs. Distributed – More centralized systems started on-time – Consider longer start-up for distributed/complex systems • Communication key in learning about and remedying problems (dynamic rendering, system regionalization) • Build on existing relationships
Semantic Alignment Kari Stephens, PhD • Making sure information (data) from multiple sources can be combined to conduct research
Semantic Alignment Longitudinally Time 2 3 4 1 • Now: Planning for pulling data repeatedly over time – Clear and frequent communication with sites – Same data file format repeated, test with index files – Document validation process • Long term: repeat data extractions – Conduct validation checks between extractions – Document process to create library of procedures (who / what / how) – Determine validation best practice methods
Semantic Alignment Time 1 within Site Site 4 2 • Now: multiple systems of care within sites 3 – e.g. proprietary radiology report codes – Staff turnover increases potential error and effort – Validation with primary / centralized research team • Long term: replicability – Track and document process for extraction and alignment; difficult to maintain post funding – Stabilize methods within sites as much as possible
Semantic Alignment Site 1 between Sites 4 2 • Now: defining variables 3 – Outcome variables: NLP for reports, RVUs (BOLD) • Review of index files – ↑ sites and variability = ↑ time / effort / complexity – Validate that independent variables mean the same thing (i.e., orders, PCP, clinic, gender, age, etc.) – Stepped wedge design reduces burden • Long term: usable dataset for analyses – Adjust analytic plan for variability
LIRE Update/Forecast • Wave 1: moderate choppy seas • Wave 2: light headwinds • Wave 3-5: smooth sailing • Data quality check 10/15/14
Henry Ford Mayo UW Safwan Halabi, MD- site PI Dave Kallmes, MD-site PI Jerry Jarvik, MD MPH-PI Dave Nerenz, PhD- site PI Beth Connelly Zoya Bauer, MD, PhD Jim Ciarelli Kevin Erdal Brian Bresnahan, PhD Bryan Macfarlane Patrick Luetmer, MD Brooke Wessman Jyoti Pathak, PhD Bryan Comstock, MS Rachel Blair Todd Sheley Janna Friedly, MD DeShawn Mahone Dan Waugh Laurie Gold, PhD Todd Wohlers Group Health Patrick Heagerty, PhD Kaiser Dan Cherkin, PhD-site PI Katie James, PA-C, MPH Andy Avins, MD MPH-site PI Heidi Berthoud Luisa Hamilton Dwipen Bhagawati Sean Rundell, PT, PhD Mike Matza Kristin Delaney Kari Stephens, PhD John Rego, MD Lawrence Madziwa Judy Turner, PhD Cliff Sweet, MD Camilo Estrada OHSU Mary Muth Rick Deyo, MD, MPH Patrick Chang
Why Pragmatic Trials Are Important
Recommend
More recommend
