Impact of Physician-Owned Office-Based Laboratories on Physician Practice Patterns and Outcomes After Percutaneous Vascular Interventions for Peripheral Artery Disease

Abstract

Percutaneous vascular interventions (PVIs) for peripheral artery disease have shifted from a hospital-based setting (HBS) to physician-owned office-based laboratories (OBLs). We sought to evaluate the changes of physicians’ treatment decisions and patients’ outcomes after the transition to OBLs. We identified patients with peripheral artery disease with PVI from 2006 to 2013 in a 20% Medicare sample and identified physicians who transitioned from predominantly HBS to OBLs (switch physicians) and compared them with those who performed procedures primarily in the HBS (control physicians). Patients’ outcomes included procedure type (balloon vs atherectomy vs stent), reintervention rates, major adverse limb events, hospitalizations, and mortality. We used a difference in difference model to control for time effects. The cohort comprised 290 switch physicians who treated 8011 patients (4373 before OBL and 3638 patients after the transition to OBL) and 3756 control physicians treating 53,629 patients (40,405 in the period before the transition and 13,224 in the period afterward). Patients’ demographics and indication for interventions are recorded in Table I. Switch physicians were observed to increase atherectomy use from 24% to 27% after the switch; however, the increase was larger for control physicians (20% to 30%; P < .001; Table II). Switching to OBL was independently associated with increased reintervention rates at 30 days and 1 year (P < .001), decreases in hospitalization rates at 30 days and 1 year (P < .001), and similar 30-day and 1-year mortality compared with control physicians. Differences for major adverse limb events were not statistically significant in controlling for patients’ demographics; however, the 30-day and 1-year amputation rates decreased at a greater rate after switching to OBL. Transition to OBLs was associated with lower amputation and hospitalization rates at the expense of increased 30-day and 1-year reintervention rates after PVI. Further study is warranted to evaluate the financial implications of OBLs to justify or to expand the current use.Table IPatients’ demographics and procedure indication by switch and control physiciansSwitch physicians (n = 290)Control physicians (n = 3756)P valueaPatients (n = 8011)Patients (n = 53,629)Age, years75.0 ± 9.675.23 ± 9.8***Male sex51.850.7NSCharlson Index5.8 ± 3.36.1 ± 3.4***ESRD10.612.0**Race Unknown<0.2<0.2*** White84.780.8 Black11.313.9 Other0.81.2 Asian0.60.8 Hispanic1.83.0 North American Native0.7<0.2Indication for intervention PAD + wound33.835.2NSESRD, End-stage renal disease; PAD, peripheral arterial disease.Categorical variables are presented as percentage. Continuous variables are presented as mean ± standard deviation.aNS, Not significant, P ≥ .05; *P < .05; **P < .01; ***P < .001. Open table in a new tab Table IIProcedure type and patients’ outcomes by switch and control physicians, stratified by periodSwitch physiciansControl physiciansP valueaBefore (n = 4373) vs after (n = 3638) switch, %Before (n = 40,405) vs after (n = 13,224) transition, %Type of first interventionb Balloon22.5 vs 23.2NS23.4 vs 19.0∗∗∗<.001 Stent53.2 vs 49.6∗∗56.7 vs 51.1∗∗∗.32 Atherectomy24.3 vs 27.2∗∗19.6 vs 29.9∗∗∗<.001Reintervention ratesc 30-day9.8 vs 14.9∗∗∗8.9 vs 8.8NS<.001 1- year28.8 vs 34.8∗∗∗25.3 vs 23.4∗∗<.001Atherectomy as subsequent procedurec 30-day3.7 vs 5.6 ∗∗∗2.5 vs 3.4 ∗∗∗.27 1-year10.6 vs 13.4∗∗∗8.3 vs 9.8∗∗∗.18MALEd 30-day5.7 vs 5.0NS5.7 vs 5.0∗∗.81 1-year15.6 vs 12.0∗∗∗14.6 vs 12.4∗∗∗.09Above-ankle amputationd 30-day2.5 vs 1.2%∗∗2.2 vs 2.2NS.002 1-year7.7 vs 4.6∗∗∗6.7 vs 5.6∗.001Hospitalizationd 30-day23.9 vs 18.0∗∗∗23.8 vs 22.9NS<.001 1-year61.0 vs 48.4∗∗∗58.9 vs 49.8∗∗∗.003MALE, Major adverse limb event.aDetermined by difference in difference analysis.bAdjusted for age, race, Charlson score, sex, wound.cAdjusted for procedure type.dAdjusted for reinterventions.NS, Not significant, P ≥ .05; *P < .05; **P < .01; ***P < .001. Open table in a new tab