Hand-carried Ultrasound Echocardiography in the Hospital

This was an unblinded, parallel-group, superiority trial that randomized participants in a balanced (1:1) allocation ratio. We made no changes to our protocol after we began enrollment. Nor did we conduct any interim analyses. Each morning of the study period, including weekends and holidays, investigators listed all consecutive patients who were newly referred for SE by participating hospitalists. After hospitalists completed their initial evaluation of each patient, which included a history, physical exam, and review of medical records, investigators asked them to choose their indications for SE from a closed list. Given these indications, investigators then asked if HCUE could help them to assess their patients; if not, investigators excluded such patients. Finally, hospitalists were asked for their current management plans, which included medication regimens, diagnostic tests, specialty consultations, and timing of hospital discharge. If hospitalists remained masked to any preliminary SE results, investigators assigned participants who gave written informed consent to an intervention group. We used a computer-generated list of random numbers for the allocation sequence, and we concealed the sequence in an on-site computer system that only released allocations to investigators after they entered participant identifiers.

Hospitalists either awaited SE results for participants randomized to usual care or they performed HCUE for participants randomized to HCUE-guided care. Hospitalists performed HCUE with a MicroMaxx 3.4 hand-carried ultrasound machine equipped with a cardiology software package and a 1 to 5 MHz P17 cardiac transducer (Sonosite, Inc., Bothell, WA). The details of the HCUE protocol are described elsewhere. In brief, without assistance from nursing staff, hospitalists positioned participants on standard hospital beds and recorded 7 best-quality images up to 6 seconds in duration. Then hospitalists replayed the images for immediate interpretation. They made linear measurements with a caliper held directly to the ultrasound monitor and translated them into qualitative assessments based on simplified standard values.

The primary outcome was the length of stay on the short stay unit or general medicine ward service of the referring hospitalist. The duration began when the patient was admitted to the hospitalist service and ended when the patient was discharged home or transferred to a different service in the hospital. We chose this duration, rather than the length of stay of the entire hospitalization, because we wanted the primary outcome to reflect how HCUE affected hospitalists' management decisions in the hospital, including the transfer of participants to other services.