Butterfly CEO says 140 clinicians at one health system bought their own ultrasound probes

Butterfly Network’s Joseph DeVivo said one health system integrated handheld ultrasound after about 140 physicians bought their own probes.

Author
HDM Staff
Save
Share
Stephanie Davis and Joseph DeVivo during the HLTH Revv Stage ultrasound discussion.
Photo credit: HLTH USA 2026.

    LAS VEGAS — Roughly 140 physicians in one health system bought their own handheld ultrasound probes before the organization had a workflow to capture the scans, Butterfly Network President and CEO Joseph DeVivo said Monday at HLTH. He said the unnamed system later worked with Butterfly to connect the devices to its records and expanded to roughly 800 or 900 probes.

    DeVivo made the remarks in An Era of Imaging Abundance, a Revv Stage session moderated by Rounds co-host Stephanie Davis. Davis asked what hospital leaders should do when clinicians start using portable ultrasound on their own and how they can keep the tools from becoming concentrated in concierge care.

    Getting the scans into the record

    DeVivo said one problem was that “all that data is not going in the EMR.” He said the system used middleware to capture scans, and he urged hospital leaders to “survey your doctors” before deciding how to build a formal program.

    The case illustrates a distinction between a clinician’s ability to acquire an image and the organization’s ability to make that image useful to the next care team. An American College of Emergency Physicians workflow paper describes routes for sending interpretations to the electronic medical record and storing images in an archive.

    Butterfly’s technical specifications describe support for DICOM imaging systems and interfaces for sending results to electronic records. Its deployment guide also calls for a designated program owner, credentialing, documentation and quality review.

    DeVivo said Butterfly’s online sales gave it one way to identify how far use had spread. He estimated that physicians buy 1,500 to 2,000 probes online each quarter, and said the company recognized that a cluster of purchases came from one city and one health system. He did not name the organization or provide an independently reported count of how many scans those clinicians performed.

    The difference matters because device purchasing cannot show whether clinicians routinely use the probes, which examinations they perform or what happens after a finding. The case DeVivo described starts with the purchase count, then follows the institution’s efforts to establish a sanctioned program. The figure he gave for the expanded fleet is a measure of deployment, while outcomes such as a faster diagnosis or a change in referral patterns would require different data.

    For a hospital, the questions begin before an image is taken. The ACEP workflow paper describes identifying the patient and operator, specifying whether an exam is diagnostic, procedural or educational, and choosing where images and the interpretation go afterward. In an encounter-based workflow, those links let another clinician find the original scan without relying on the clinician who performed it to reconstruct what happened.

    The AIUM practice parameter calls for a permanent record of the ultrasound and its interpretation, including retrievable relevant images. That has implications for who can access the images, how the facility retains them and how urgent findings reach the treating team. Those responsibilities remain when a device is small enough for clinicians to buy and carry on their own.

    Medical training and where the scan fits

    DeVivo urged administrators to build continuing medical education into a frontline scanning program. He also distinguished the focused scan from the diagnostic work that may follow in radiology. The question for a program is what a clinician can reasonably assess in that setting, what requires a second look and who follows up when a scan raises a concern.

    Davis described receiving more inquiries from direct-to-consumer and concierge companies than from health systems after featuring DeVivo on Rounds. She asked how an institution could respond to activity already taking place among its clinicians. DeVivo said leaders could start by surveying their own teams, then build formal training and a way to manage the information they collect.

    DeVivo raised home care as one setting where bringing the scanner to the patient could change the decision to escalate care. He described possibilities for nurses performing limited examinations outside the hospital. The approach he outlined would still require a protocol for who is trained, which findings need review and how the result reaches the clinician responsible for the next decision.

    For executives, the purchasing pattern DeVivo described offers a starting point for that work. An inventory can show where probes already exist, while examination logs and record audits can show which scans are reaching the care team. Taken together, those measures tell an institution more about the program than the number of devices alone.

    DeVivo described the handheld scan as an initial assessment that can flag problems for follow-up in radiology. He said formal training and a way to move images into the clinical record would help hospitals use the tools in more settings, including care in the home.

    DeVivo also described the technical changes that made a handheld scanner viable. The transition he outlined was from different crystal designs for different kinds of ultrasound to semiconductor-based imaging in a single device. The company sees that as a way to bring more kinds of focused examinations to the bedside. A device’s ability to acquire those images leaves a separate question about which examinations clinicians in a particular role should be trained to perform.

    For example, a scan performed by an emergency clinician to decide whether to escalate care need not follow the same workflow as an examination performed by an imaging department after an order. The clinical context, the operator and the purpose of the examination all need to remain clear as the image moves through the organization. Butterfly’s documentation describes both order-based and encounter-based workflows, a distinction that helps explain why the device procurement decision and the integration decision can involve different teams.

    Billing was another point Davis raised. DeVivo said that, without a sanctioned workflow, the institution had no reliable way to bill for those individual scans. Connecting documentation to a revenue-cycle system may make an eligible examination visible, but the case presented onstage did not show which exams met documentation and coverage requirements or what revenue the deployment generated.

    That makes the case most useful as a description of how adoption surfaced. A hospital could start with clinicians who already use devices and ask which examinations help answer a recurring question and which need a better route for training, archiving or confirmation.