Vesela Kovacheva snuggles with her oldest daughter, Kalina, moments after her birth at the Brigham.

As an obstetric anesthesiologist and a mother of two daughters, Vesela Kovacheva, MD, PhD, of the Department of Anesthesiology, Perioperative and Pain Medicine, has a keen understanding of what many of her patients experience during childbirth.

Most mothers undergoing a C-section receive a spinal anesthetic prior to the procedure — allowing them to remain comfortable while still being able to see and hear their newborns right away. But in nearly 75 percent of patients, it has an undesirable side effect of causing the mother’s blood pressure to drop. To counteract this, patients also typically receive another type of medication, a vasopressor, to elevate blood pressure.

This push and pull between the two drugs is a delicate dance as anesthesiologists work constantly to keep a mother’s blood pressure at a normal, controlled level before, during and after a C-section. If the numbers move too far in one direction or the other, mothers can experience nausea, vomiting, light-headedness and, in rare cases, stroke — events that may also have consequences for their baby’s well-being.

To keep patients safe and healthy, anesthesiologists monitor patients’ vital signs throughout the procedure, manually calculating and adjusting the medication dosage on an infusion pump, as needed, on a minute-by-minute basis — among multiple other tasks during surgery.

“The delivery of their baby is a very special time for our patients. It’s something they’re going to remember for the rest of their lives,” Kovacheva said. “Yet it’s also a very busy time for us as clinicians. We’re doing all of these things simultaneously while taking care of an anxious, awake patient and preparing for major abdominal surgery. In a couple of minutes, the math becomes too complex for any human to do, and we may under- or overtreat the patient.”

There must be a better way to do this, Kovacheva recalled thinking while recovering from her own C-section following the birth of her younger daughter at the Brigham two years ago.

It wasn’t long before she got to work on identifying and developing a possible solution using machine learning, a type of artificial-intelligence technology in which computers can be trained to identify patterns and make predictions after analyzing massive amounts of data.

Now, in collaboration with colleagues across the Brigham and Partners HealthCare specializing in anesthesiology, data science and pharmacology, Kovacheva and her team are working to develop a software algorithm that could monitor, anticipate and, under physician supervision, react to changes in maternal blood pressure in real time.

“We envision this as a physician superpower — something that augments, not substitutes, the doctor’s expertise. There will always be a need for a physician to tell the machine what to do,” she said. “Our hope is that this software will delegate a tedious, error-prone task to the machine so that I, as an anesthesiologist, could hold my patient’s hand, talk to them and focus on other aspects of their care.”

Training the Machine

While clinical guidelines provide a starting point for vasopressor dosage based on the particular clinical situation, every patient is unique and requires a personalized approach, Kovacheva explained. Patients who share similar traits may react to the same medication dose in different ways, and arriving at the right level is sometimes a matter of trial and error, she added.

The technology Kovacheva and her colleagues are developing would analyze a patient’s real-time blood pressure and heart rate and compare that against a large database of similar scenarios. In turn, the software will recognize when the medication needs to be adjusted and alert an anesthesiologist about the patient’s status and recommended dosage change.

Vesela Kovacheva

“The way machine learning works is that it utilizes previous patient records and ‘teaches’ itself what should be done, based on what we’ve done in the past,” she said. “The most challenging part is that we need a very large number of patient records to train the software. We currently have about 700 records, and we’ll be doing a lot of internal testing to make sure that it’s safe and effective before we test it in people.”

If successful, Kovacheva hopes the technology could be useful to providers in more remote areas, where anesthesiologists are often in short supply, as well as in other types of surgeries and procedures.

For now, she is eager to do whatever she can to ensure childbirth is a joyous time and goes smoothly for more patients, families and providers alike.

“I feel that I have an idea that will hopefully make this experience better for not only moms and babies but also for the people caring for them,” Kovacheva said.

Home