IMPACT OF INTERNAL MEDICINE POINT-OF-CARE ULTRASOUND ON LENGTH OF STAY, COSTS, AND FORMAL IMAGING: A PROPENSITY-SCORE MATCHED PROSPECTIVE COHORT STUDY

Background: Point-of-care ultrasound (POCUS) is being adopted by hospitalists and internal medicine (IM) residencies due to its impact on procedural safety, diagnostic accuracy and efficiency, provider and patient satisfaction, and its role as a teaching adjunct in medical education. Studies demonstrating impact on system efficiency and cost are few and mostly limited to the emergency department (ED) and intensive care settings. Randomized trials allocating the known diagnostic benefits of POCUS to one group and withholding from another are unlikely within a controlled setting and provider group. The goal of this study was to evaluate the impact of POCUS within an inpatient IM setting on system and patient metrics.

Methods: Prospective cohort of 12,399 admissions at a large tertiary care teaching hospital assigned to hospitalists with or without POCUS available for use during a given hospitalization. Regression and propensity score matched analyses compared length of stay (LOS), hospitalization cost, radiology volume and costs, patient satisfaction (HCAP scores), and readmission between patients cared for with and without POCUS available, as well as for patients that actually received at least one POCUS exam during hospitalization vs. those that did not. Subgroup analyses by admission diagnosis (pneumonia, acute heart failure and acute kidney injury) were also performed.

Results: POCUS was available for 9,985 and unavailable for 2,414 patients. Hospitalist attendings cared for patients in both POCUS availability categories. Patients cared for with POCUS available vs. POCUS unavailable had: lower total and per-day hospitalization costs ($17,474 vs. $21,803; p< 0.0001 and $2,805.88 vs. $3,557.53; p< 0.0001, respectively), lower total and per-day radiology cost ($705.41 vs. $829.12; p< 0.0001 and $163.11 vs. $198.53; p< 0.0001, respectively), fewer total chest radiographs (1.31 vs. 1.55; p=0.005), but a modest increase in total chest CT (0.22 vs 0.15; p=0.001). Mean length of stay was 5.77 (5.63-5.91) days in the POCUS available group vs. 6.08 (5.66-6.51) in the unavailable group (p=0.143). Of those for whom POCUS was available, 994 (10%) patients had at least one POCUS exam performed during hospitalization. These 994 patients underwent 5,353 total POCUS exams evaluating 8,110 body areas during their hospitalizations. When metrics were analyzed using propensity score matching (covariates: age, race, acute kidney injury, acute heart failure, pneumonia, severity of illness index, LOS, and presence of resident physicians on the team) for patients that received an exam(s) (n=869) vs. eligible patients for whom the provider chose not to use POCUS (n=3,476), significant differences were observed for total ($15,082 vs. $15,746; p< 0.0001) and per-day ($2,685 vs. $2,753; p=0.042) hospitalization costs.There was no significant difference in quantity of formal ultrasound exams or echocardiograms, 30-day readmission, post-discharge ED visits, or pre-selected HCAPS scores (targeting respect, listening, provider explanation and patient understanding of diagnosis, and overall care rating) for any group or subgroup analysis.

Conclusions: Availability and the selected use of POCUS by hospitalist providers trained within a robust IM POCUS program significantly, and meaningfully, reduced total and per-day costs of hospitalization and chest radiographs at the expense of a slight increase in chest CT scans during the early years of an IM residency-based POCUS program.