Nurse Assistant Robots Could Improve ER Efficiency and Care
I pitched my editor at New Scientist an update of a story I did a little more than a year ago on a team of scientists at Vanderbilt University in Nashville, Tenn., that is developing a robot to do triage in the ER. (Check out that piece here.) Unfortunately, he advised me to hold off until the project matures a bit. But, since I talked to 2 of the leaders on this project, I thought I’d share what they told me in terms of how the project has evolved over the last 13 months.
ERs in the U.S. are very overcrowded, said Alan Storrow, an associate professor of emergency medicine at Vanderbilt, and vice chairman for research and academic affairs in the emergency department at Vanderbilt University Medical Center. The advantage to having a robotic “nurse assistant” in the emergency department is that it can free doctors and nurses up for more important tasks, such as patient evaluation and treatment. In addition, it could also reduce patients’ length of stay in the ER while simultaneously improving ER efficiency.
In early November, the researchers applied for a grant from the National Science Foundation’s National Robotics Initiative. The grant is for $1.25 million over 5 years, said Storrow, and the team expects to hear back from the NSF in early 2012. He said that the grant will probably be funded through the National Institute of Nursing Research at the National Institutes of Health, a partner of the NSF in the foundation’s robotics initiative.
Navigating the Uncanny Valley
“We are being very clear throughout this whole process that this is an assistant to the healthcare workers,” said Karen Miller, director of clinical research operations in the ER at Vanderbilt University Medical Center. Miller continued by saying that the robot will not replace healthcare workers; it is meant to extend their capabilities for monitoring and evaluating patients. “We want the robot to look like a robot, and the healthcare workers want the robot to look like a robot and not like a person,” said Miller.
Both patients and healthcare providers must accept the presence of a robot in the ER, so the team is bringing stakeholders, including patients, nurses, MDs, and other ER staff, together with Vanderbilt’s engineering, bioinformatics, computer science, and biostatistics departments, in a working group. The stakeholders will provide feedback to the Vanderbilt engineering and computer teams about what they think the robot should look like and what they think the robot should do.
For example, “A concern for us upfront is if the patients are going to be comfortable interacting with a robot, and if not, why not?,” said Storrow. Is it because of the bot’s appearance? Is it because they do not trust the robot? He said that the researchers want to address these issues straight away, “rather than trying to develop something, putting it in the ER, and seeing how it goes.”
One piece of feedback concerned the size of the bot. ER patients and visitors said that they would find the robot intimidating if it was too large (say, 5 or 6 feet tall), so the researchers decided to make it about 4 or 4.5 feet high.
Storrow and Miller did mention that, perhaps because they are more used to interacting with technology in a medical setting, ER staffers seem more comfortable with the idea of a robot in the ER. The healthcare workers seemed more accepting for whatever reason about the prospect of an assistant who can gather information and assist with some of the automated tasks, said Miller. She added that, “I think a nurse put it the best when she said that she can interact with the patient, and do more of the nursing assessment, without having to do the more mundane tasks such as registration and vital signs.”
Said Storrow, “That makes sense to me. The prospect of having help, even through technology, is appealing to and appreciated by the healthcare worker.”
In terms of the patients and visitors to the ER, Miller said that the positive remarks the researchers got indicate that patients would appreciate having something to do while they wait, and start the registration and information gathering process earlier, so that they would not be asked the same questions many times. The thought that the bot has the potential to decrease ER length of stay is a positive from a patient perspective, Morrow said.
A Concept Evolves
The team’s original idea was to develop a system of robots called the TriageBot system. Then the researchers decided to integrate the components of that system into one robot that would be more flexible and perform a variety of tasks, “which is really what you need in an assistant,” said Miller.
The team is now using local funding to collaborate with computer scientists and engineers at Vanderbilt on the current version of the robot, dubbed Triage Interactive Nurse Assistant2, or TINA2. The future incarnation of the bot, called Mobile Interactive Nurse Assistant, or MINA, would be covered under the NSF grant. MINA could be used in other areas of the ER, say in a patient room, and to perform other tasks in addition to triage, said Storrow.
For example, MINA could provide patient education, such as showing patients a video providing discharge instructions for post-treatment care. MINA could also be integrated with and communicate with the hospital record systems.
Storrow said that the ER gets very busy, but “there are certain labs you would like the results of right away, because you would like to act upon those results as soon as possible.” So if the bot was integrated with the hospital record systems, it would be able to monitor the availability of lab results, collate a few pieces of information, and then page the MD or the nurse about the results. “It may allow us to act upon the results in a more expedient way and ultimately improve patient clinical outcomes,” Storrow said.
Miller said that, once the robot is perfected, she definitely sees such a robot being adopted in ERs across the country. Widespread use is, however, many years away.
Below: The concept graphic for the MINA robot.