Kacie McMillan, her hands joined, presses steadily and firmly into the man's chest.
The hospital patient has gone into cardiac arrest, and it's McMillan's job to mimic the heart's operation, pumping oxygen-rich blood through his body and sustaining his life.
She's getting constant encouragement and advice from a somewhat metallic-sounding female voice.
"Just a little faster."
"You're doing well."
"You can press a little faster."
"Compress a little deeper."
McMillan glances up at an observer.
"You can tell I'm getting tired," the registered nurse says.
If this were a real situation, no one would blame McMillan for getting a little annoyed at her constantly speaking coach and saying: "Do it yourself if you're so good."
But it isn't real. Her patient is a manikin, and the voice is emerging from a computer. Together, they are part of the American Heart Association's Resuscitation Quality Improvement Program - RQI for short.
The training gizmos, used quarterly by registered nurses, licensed practical nurses, nursing assistants, medical assistants and respiratory therapists, are at nearly 300 hospitals across the country, said Dr. Michael Kurz, associate professor at the University of Alabama-Birmingham Medicine and a heart association volunteer, in an email. But so far, there's only one in Minnesota: St. Luke's, which has been using five RQI "carts" for a little more than a year.
"I think we always want to be on the forefront of new technology and doing what's best for our patients," said Brittney Kurhajetz, a critical care education coordinator at St. Luke's.
Offered by the American Heart Association in partnership with Laerdal Medical of Norway and Tennessee-based HealthStream, RQI is an effort to improve outcomes when cardiac arrests occur in hospitals. In 2016, 209,000 heart attacks occurred in hospitals, according to heart association data. Among adult victims, only about one in four survived - a figure that has been relatively stable for the past five years.
Citing a variety of studies, the heart association contends more frequent training for key hospital personnel likely will result in higher survival rates. The RQI program is one strategy to achieve that.
Each mobile cart comes with an adult-sized and an infant-sized manikin and the laptop computer that runs the program. In about 10 minutes, Kurhajetz said, the nurse, assistant or therapist can complete the program and, assuming a score of 75 percent or better, keep his or her certification up to date.
The certification has to be updated every two years, Kurhajetz said. Traditionally, this is done in a biennial basic life support class, which is what St. Luke's did until October 2016, she said.
But there's a problem, Kurhajetz said, with training only once in 24 months in a skill that may not be used frequently but is the difference between life and death when it's needed.
"What we've found is that your skills, your confidence and your competence have gone down when you try to remember that class that you took two years ago," Kurhajetz said.
In its updated guidelines for 2015, the heart association reported that two-year retraining was "not optimal. More frequent training ... may be helpful for providers who are likely to encounter a cardiac arrest."
McMillan, a registered nurse in the cardiac unit at St. Luke's for two years following four years as a nurse's aide, liked the classes, she said, but likes the more frequent computer-driven training better.
"With the computer you're actually being tested, and you're watched really closely," she said. "With the class, nobody's really critiquing the depth and the quality of the compressions. Also, doing it more frequently, doing it the four times of the year, makes me more confident in knowing what I'm doing, knowing where to place my hands."
It requires a concerted effort. Kurhajetz advised a bystander, giving it a trial, to stand on a stool, thus allowing greater downward pressure. The unnamed manikin apparently is an upper-body workout buff, because even at that angle, several compressions required concerted effort from the novice.
But with the RQI manikin, as in real life, what's actually required is 30 deep compressions at a rapid tempo (15 for an infant), Kurhajetz said. (The heart association suggests pumping to the beat of the disco song "Stayin' Alive.")
That's followed by two applications with a ventilator mask to the nose and mouth, then 30 more deep compressions - at about 80 to 90 pounds of pressure, Kurhajetz said.
In the RQI version, this goes on for two minutes, and then the computer evaluates the participant's performance. In the demonstration, McMillan passed easily, with a 92 percent mark. But passing isn't a given, Kurhajetz said. Those who struggle are given the opportunity to work individually with one of the hospital's educators.
There's also an RQI version in which two people work together, one using the ventilator and one compressing the chest. They switch after two minutes. The "gold standard" is 10 seconds to make the switch, Kurhajetz said. The test is done when each has had a chance to perform each function.
In real life, of course, it can last much longer. Earlier this year, McMillan and another nurse responded to a patient who was in cardiac arrest, she said.
"I just relieved another nurse and stepped in and started doing compressions until you basically get tired, and then you switch off again," she said. "It's a group effort. There's a lot of people involved when a situation like that happens."
That went on for close to an hour, McMillan said, and by then "everyone was tired." The patient survived.
Does it work?
The RQI training method debuted just in 2014 at Texas Health Presbyterian Hospital in Dallas, and so far data on its effectiveness is mostly anecdotal. At the University of Alabama at Birmingham Health System, Kurz said, nurses shifted from a "Should I perform CPR?" mentality to "I'm doing CPR unless someone tells me not to."
Texas Health Resources, a group of 24 hospitals in the Dallas-Fort Worth area, reported a 21 percent increase in survival rates following cardiac arrest from before RQI to after it was implemented, according to a study in the June 2016 edition of the journal Nurse Leader.
Switching from training classes to the more frequent individualized computer-driven sessions also has proven to produce significant cost savings, Kurz said - $250,000 in one year at Texas Health Resources and $300,000 in one year at Seton Healthcare Family in Austin, Texas.
What Kurhajetz said she's excited about is the improved training she has seen over the first year.
"I think it's really beneficial," she said. "Instead of every two years, taking four hours to do a classroom class and losing skills over time, this shows that let's get in, let's do it, let's keep our skills up and help our patients."