90 Does Supervised Learning Promote a High-Cost Practice Style?

Abstract

To compare resource use in teaching encounters to non-teaching encounters in a probability sample of patient visits to United States emergency departments (ED). Although cost-awareness may be taught formally as a component of systems-based practice, clinical learning occurs chiefly through observation. Only a few large-scale studies have investigated actual resource use by trainees in practice settings, with conflicting conclusions. No studies have focused on the ED as a driver of resource utilization. We analyzed the 2010 National Hospital Ambulatory Medical Care Survey ED sub-file, a nationally representative sample of both EDs and ED visits. We defined three ED types: non-teaching, minor teaching (fewer than half teaching encounters), and major teaching (more than half teaching encounters). We evaluated four resource use outcomes: hospital admission, advanced imaging (computed tomography, ultrasound, or magnetic resonance imaging), laboratory testing, and a long ED stay. Using simple and multivariable complex survey regression models, we evaluated the association of teaching encounters with these outcomes, adjusted for other potential determinants of increased resource use. Data analysis conducted in STATA. Of the estimated 129.8 million ED visits in 2010, 7.4% were teaching encounters. These were associated with higher odds for each of the four outcomes-- hospital admission, advanced imaging, laboratory testing, and a long ED stay (adjusted for visit acuity, demographic characteristics, payer type, geographic region, and urban setting. Visits to major teaching hospitals were not associated with increased resource use, except for a long ED stay. By comparison, ED visits to minor teaching hospitals were more likely to result in high resource use for each of the three remaining categories (Table). The practice environment of midsize EDs, those more likely to be minor teaching hospitals, is the main driver of overall higher resource use by trainees, rather than either large urban teaching hospitals or unique unmeasured characteristics of the teaching encounter itself.TableRelative effects of teaching visits and teaching hospital visits on ED resource use. Adjusted odds ratios with the visit as unit of analysis. Pediatric specialty hospitals are exduded. United States, 2010.Millions of visitsHospital admissionAdvanced imagingBlood testLong ED stayModel 1: Resource use relative to non-teaching visits*Non-teaching visits98.9 (91.1%)ReferentAil teaching visits9.7 (8.9%)1.31 (1.00,1.70)1.19 (0.99,1.42)1.15 (0.92,1.43)1.78 (1.34,2.36)Minor teaching role3.7 (3.4%)1.86 (1.29,2.68)1.30 (1.01,1.68)1.50 (1.12,2.00)1.53 (1.13,2.06)Major teaching role6.0 (5.5%)1.04 (0.75,1.43)1.12 (0.89,1.42)0.99 (0.76,1.28)2.93 (1.30,2.88)Total visits†108.5 (100%)Model 2: Resource use relative to non-teaching EDs*Non-teaching EDs (n=3146)‡66.2 (53%)ReferentEDs with any teaching role58.7 (47%)1.41 (1.14,1.74)1,11 (0.98,1.26)1.13 (0.95,1.35)1.83 (1.45,2.31)Minor teaching role (n =1301)49.2 (39%)1.43 (1.16,1.78)1.14 (0.99,1.30)1.13 (0.99,14.2)1.73 (1.35,2.21)Major teaching role (n=155)9.5 (8%)1.29 (0.89,1.86)0.98 (0.79,1.21)0.90 (0.70,1.16)2.30 (1.54,3.44)Total visits124.9 (100%)*Odds ration adjusted for low acuity at triage age category, sex, race/ethnicity, payer, geographic region. Accompanied by 95% CI in parantheses.†Excludes visits seen by resident/intern, associate provider, or consultant alone.‡n is the weighted total, an estimate of the number of non-pediatric EDs in the category nationally (from table 2). Open table in a new tab