Procedures Performed by the Hospitalist and Non-hospitalist

Abstract

The Authors' Reply. We thank Dr. Barsuk and colleagues their thoughtful analysis of our article. They raise important concerns about the generalizability of our findings. The SHM data is based on surveys of hospitalist groups, not individuals. Our study examined procedure rates of individual hospitalists. Although the absolute number of respondents in our dataset was small, it was obtained through a probability sample. The sampling frame was representative of the American College of Physicians (ACP) membership, stratified by geographic region of the US and year of graduation from medical school.1 We believe our findings are credible enough to warrant further examination of hospitalists’ procedural practices. Dr. Barsuk also noted that prior year performance rates for four of the nine core procedures were 50% (lumbar puncture), 49% (abdominal paracentesis), 44% (thoracentesis), and 39% (central line placement). However, we also found that hospitalists who reported doing a procedure performed a median of 5 lumbar punctures, 5 abdominal paracentesis, 5 thoracentesis, and 5.5 central line placements in the prior year, with inter-quartile ranges that varied between 2–30 procedures. These procedure volumes suggest that while some hospitalists may perform bedside procedures regularly, the majority do not. If further studies confirm these findings, hospitals, policy makers, residency accreditation committees, and professional societies should re-consider long-held assumptions about hospitalists’ procedural practice patterns and how they might affect our patients. A growing literature shows that regular performance of procedures is associated with lower complication rates.2 If most hospitalists truly do perform only a handful of bedside procedures a year, it is debatable whether they are performing a sufficient number to remain comfortable, much less proficient with them.3 Further research into how hospitalists acquire and maintain procedural competence is necessary. Simulation-based competency training is one avenue for addressing this concern.4 Other potential solutions include referring patients to sub-specialty proceduralists and developing dedicated proceduralist services within hospitalist groups5. We believe that specific interventions should be tailored to local needs and resources, so as to ensure that our patients always receive high-quality inpatient care.