In the fast-paced environment of a hospital’s labor and delivery unit, planning and coordinating the many aspects of patient care can be a never-ending series of split-second decisions on how to optimize scheduling and allocate tasks to various members of the healthcare team.
Typically, these decisions are the responsibility of a resource nurse, whose role includes monitoring the nursing team’s workload, pairing nurses with each laboring patient, reserving beds or operating rooms for patients and assigning scrub technicians to operating rooms. The resource nurse is also responsible for calling in additional support, reassigning nurses who are performing other roles and rescheduling inductions or cesarean sections if the floor becomes too busy. While making these staffing and scheduling decisions, the resource nurse also has to try to predict how long labor will take and which patients may experience complications requiring additional procedures.
Intrigued by the complexity of this role, researchers from the Massachusetts Institute of Technology (MIT) Computer Science and Artificial Intelligence Laboratory (CSAIL) in Cambridge, Mass., conducted a research study to determine if a robot could assist nurses in this type of decision helping to ensure efficient scheduling for optimal patient care.
“Nurses scheduling the labor and delivery unit have one of the hardest jobs in healthcare because there is so much to keep track of and so much uncertainty involved in childbirth,” said Neel Shah, M.D., an obstetrician at Beth Israel Deaconess Medical Center in Boston, Mass., who co-authored the study. “In many industries we have decision-support tools to help people with these kinds of complex jobs, and we wanted to see if we could use artificial intelligence to develop similar assistance for our nurses.”
The CSAIL researchers created a special robot (nicknamed “Ginger” and pictured above) that processed data obtained from seven resource nurses. The data was pulled from observing the nurses in real-life situations, which helped inform a scheduling policy allowing the robot to make high-quality recommendations based on assessing all possible actions. The policy allows the robot to continue to learn how to respond to new scenarios without someone having to manually input more data.
“The technology we developed is similar to what Google does when you type something into the search field. You could type out the whole thing you are searching for but usually Google will suggest what you want once you begin typing just to make things a little bit easier. The robot aims to make helpful suggestions in a similar way,” Shah explained.
Once the robot was trained, researchers brought it to the labor and delivery unit at Beth Israel Deaconess Medical Center, where the resource nurse often coordinates 10 nurses, 20 patients and 20 rooms at the same time, according to the study's authors.
Seventeen physicians and registered nurses participated in the experiment, including resource nurse Kristen A. Jerrier, RN, BSN, CNIV.
“Working with the robot was particularly interesting because the goal was to impact the way I am able to provide care by allowing me to spend more time with my patients,” Jerrier said.
The study found that the nurses developed a high degree of trust with the robot. Their results indicated that the robot made suggestions the doctors and nurses carried out 90 percent of the time. While researchers are looking to expand their research to more hospitals, several nurses who participated in the study noted that robots like this could be useful as an educational tool for new nurses adapting to the complexities of the resource role.
“Whether the robot actually becomes a part of our everyday practice or not, it was nice to know that researchers recognize the important, complex responsibilities of nurses,”Jerrier said. “It was very rewarding to see how well the robot learned from the nurses, and I enjoyed adding my years of experience to the robot’s understanding of the role of a resource nurse.”
To learn more about the study, please visit the MIT Press Room.
*Image courtesy of the Massachusetts Institute of Technology (MIT) Computer Science and Artificial Intelligence Laboratory (CSAIL).