Sean Daniel Rundell MS, Presenter: Nothing to Disclose

Jeffrey Gil Jarvik MD, MPH, Abstract Co-Author: Stockholder, PhysioSonics, Inc Research Advisory Board, General Electric Company Consultant, HealthHelp

Richard A. Deyo MD, MPH, Abstract Co-Author: Nothing to Disclose

To determine if early imaging in Medicare recipients with spinal stenosis is associated with increased risk of surgery.

This is a retrospective cohort study using nationwide Medicare claims data from 2002-2006. A 20% sample from the Standard Analytical Files and the Medicare Provider Analysis and Review claims data were used. We identified Medicare recipients 68 years or older with a diagnosis code for spinal stenosis at an index visit. Early imaging was defined as receipt of radiographs, computerized tomography scan, or magnetic resonance imaging earlier than 42 days after the index visit compared to later or no imaging. We used cumulative hazard curves and Cox proportional hazards regression to estimate the associations between risk of spinal surgery and early imaging. Covariates of age, sex, comorbidity, chronicity, and number physician visits were included in regression models. Results were stratified by surgeon and non-surgeon providers. We were unable to adjust for severity of symptoms and indications for surgery with administrative data.

Of 22,781 patients with spinal stenosis, 1074 subjects received surgery. After adjusting for covariates but not confounding due to indication, those receiving earlier imaging had 3.2 times higher hazard for receiving surgery than those with late or no imaging [95% confidence interval (CI), 2.66 - 3.80]. Patients initially seeing a non-surgeon, excluding radiologists, at the initial visit who received early imaging had a hazard rate that was 4.0 times greater than those with later or no imaging [95% CI, 3.0 - 5.2].

Potential confounding by indication when using administrative data requires any apparent association between early imaging and spinal operations to be viewed cautiously. This analysis should be replicated using a data source that includes information on symptom severity along with signs and symptoms of serious spinal pathology.

We intend to use a devloping cohort of older adults with low back pain to further investigate this association. We will use this data set to estimate the magnitude of bias in the Medicare data.

Rundell, S, Jarvik, J, Deyo, R, Surgery Frequency Associated with Early Imaging in Medicare Patients with Spinal Stenosis.  Radiological Society of North America 2012 Scientific Assembly and Annual Meeting, November 25 - November 30, 2012 ,Chicago IL. http://archive.rsna.org/2012/12025813.html