HCA nurses say Palantir-built Timpani honors their shift requests less often than managers did

On a recent run of shifts, Amber Retzloff was the only senior nurse working. All four of her colleagues were junior, and she said she had to delay care for the sickest patients so she could guide the newer nurses through milder cases ^[8]. Retzloff is 34, has a decade of bedside experience and is a leader in National Nurses United ^[8]. "We are overriding clinical judgment and the human-to-human piece of health care with an app," she said ^[9].

HCA co-developed Timpani with Palantir ^[3]. The chain is the customer. The nurse is the person whose week the tool arranges. HCA spokesperson Harlow Sumerford described the goal as "to have the right caregivers, with the right skills and experience, available to meet the needs of our patients while also considering our colleagues' scheduling preferences" ^[12]. In that sentence, nurses' preferences arrive after an "also." Sumerford said HCA keeps improving the tool based on nurses' feedback ^[12].

Nurses describe what they actually do with a schedule they did not ask for. They trade shifts and file appeals, and they say this takes more of their time than it did before the software arrived ^[6]. Some call out instead, though calling out more than a few times a year puts a nurse at risk of termination ^[7]. "I can't remember any issues before," said a non-union nurse in Texas who asked not to be named for fear of retaliation. "Everything came out on time." ^[20]

HCA's own figure is that Timpani, on average, schedules nurses for 1 percent of their requested days off, known as red days ^[13]. Lee Barker, an HCA nurse in Missouri, said that may look low, but working a red day was previously unheard of at their facility ^[14]. Less than a year after Timpani arrived, Barker said, essentially every nurse there had been assigned one ^[14]. The two figures count different things. HCA's is a share of days and Barker's is a share of people, so a 1 percent rate can still reach nearly every nurse on a unit if the misses are spread across the staff ^[1]. Barker put off a long-planned doctor's appointment for a chronic condition after struggling to trade an assigned red day ^[15]. "It feels like we're being kicked while we're already down," Barker said ^[16].

The evidence so far is nurses' accounts on one side and company statements on the other. Retzloff's comparison with hand scheduling rests on her own count of her requests ^[2]. HCA says nursing leaders, not Timpani, make the final scheduling decisions ^[11].

A team rolling out a scheduler can sort it on two axes. The first is whether the tool honors individual requests at least as often as the hand-built schedule did, counted per person and not per day. The second is whether a manager reviews each override before the schedule publishes, with the old notice period intact. Pass both and the tool saves manager time. Honor requests but publish unreviewed, and the bad weeks reach staff with nobody checking first. Miss requests but review them, and the work goes back to managers, where at least it gets counted. Miss and publish unreviewed, and staff do the repairs. HCA places Timpani in the reviewed column ^[11]. Nurses describe overrides arriving on shorter notice than when managers consulted them well in advance, plus more temporary moves to other teams ^[10]. Neither axis can be scored without a per-person baseline taken from the hand-made schedule before the switch. In this record, the closest thing to one is Retzloff's count of her own requests ^[2].