Percutaneous vertebroplasty among low income Medicare beneficiaries

Cohort.

To assess the patterns and predictors of percutaneous vertebroplasty use among low-income Medicare beneficiaries from 1 U.S. state.

Many observational studies have examined the effectiveness of vertebroplasty, but none has described its pattern of use in the community.

We studied a cohort of low-income Medicare beneficiaries from Pennsylvania diagnosed with a spine fracture between 2001 and 2004. Healthcare utilization data were assessed to determine receipt of a vertebroplasty. The frequency of vertebroplasty among spine fracture patients was examined by year and by county of residence. Multivariable logistic regression models were constructed to identify correlates of receiving a vertebroplasty.

We found 6392 patients who fulfilled our criteria for a spine fracture. Of these, 459 (7.2%) underwent a vertebroplasty in the subsequent 12 months. The percent of patients sustaining a fracture who underwent vertebroplasty increased from 6.8% to 8.2%, a 20% increase, over the 4-year study period. Substantial variation in use of vertebroplasty was observed across county of patient residence, with several counties having zero patients and many with >15% undergoing vertebroplasty. Several patient factors predicted use of vertebroplasty, including age 65 to 84 years (compared with 85 years and older), and more physician visits in the 12 months before the spine fracture. The model fit statistic (C statistic) for models with only patient factors was weak (c = 0.60), but this increased substantially when county of residence was also included in the model (c = 0.72).

Vertebroplasty use has generally increased over the study period. Patient characteristics determined from healthcare utilization data are weak correlates of vertebroplasty use, but a patient's county of residence is a relatively strong predictor.