HCA nurses say Palantir-built Timpani scheduling tool may harm patients

WIRED reports that nurses at HCA Healthcare, the largest hospital chain in the US, are pushing back against Timpani, an AI scheduling tool the company co-developed with Palantir. Since 2023 HCA has rolled it out at roughly 130 of its 190 locations. Timpani was built on Palantir Foundry, a platform for storing and analyzing data from across a company's operations, and it uses algorithmic forecasts of patient volumes and staffing needs to build nursing schedules automatically.
The lead example is Amber Retzloff, a 34-year-old critical care nurse in Florida and a National Nurses United union leader with a decade of bedside experience. Over the past four months she requested 50 specific 12-hour shifts. By her account, over half the time the software assigned her different ones, often putting her on back-to-back-to-back days. She says it honors her preferences far worse than managers did when they scheduled by hand. Her hospital adopted Timpani nearly two years ago. "No one is happy with Timpani," she says, echoing five nurses at other locations who spoke to WIRED. Other purported HCA workers voice similar complaints on Reddit.
The nurses allege that Timpani has potentially harmed patients by leaving care teams understaffed, imbalanced and exhausted. They say that, with minimal oversight, it routinely schedules too few nurses or too few experienced veterans, particularly on Sundays, and frequently ignores their stated preferences on spacing shifts, working nights and weekends, or taking specific days off. Retzloff recalls recent shifts where all four of her colleagues were junior; as the only senior nurse she says she had to delay care for the sickest patients to guide the novices through milder cases. "We are overriding clinical judgment and the human-to-human piece of health care with an app," she says.
Nurses say they now spend more time trading shifts and appealing schedules than before. Previously, managers scheduled manually and generally consulted nurses well in advance when overriding requests; with Timpani, nurses report needing to make alternative arrangements on shorter notice, and more temporary reassignments to other teams. They also believe colleagues are increasingly calling out, although doing so more than a few times a year risks termination. Initially some local managers could edit Timpani's schedules. Now nurses must flag problem shifts to a centralized team at HCA headquarters in Nashville, which reviews appeals without direct communication or room for explanation, according to nurses. If an appeal is denied they try to swap shifts, but Barker and Retzloff claim the chance of success is only about 50 percent.
HCA spokesperson Harlow Sumerford says nursing leaders, not Timpani, make final scheduling decisions, and calls assertions that the tool is designed to reduce staffing at the expense of patient care misrepresentative. HCA's goal, he says, is to have the right caregivers with the right skills and experience available for patients while also considering colleagues' scheduling preferences, and the company continues to improve Timpani based on nurses' feedback. According to HCA, Timpani on average schedules nurses for 1 percent of their requested days off, known as red days. Lee Barker, an HCA nurse in Missouri, says that figure may look low, but being scheduled on a red day was previously unheard of at their facility. Less than a year after Timpani arrived, Barker says essentially every nurse has been assigned to work on a day they were supposed to be off. Barker delayed a long-planned doctor's appointment for a chronic condition after struggling to trade an assigned red day.
HCA's innovation chief Michael Schlosser said in a paper this past July that Timpani has drastically reduced the hours managers spend on scheduling each month, lowered reliance on costly contract nurses and increased retention. He wrote that the efficiency comes from AI generating stronger initial schedules, and that over 98 percent of schedules include a mix of skill and experience. At launch in 2023 he told WIRED it would save time and money and eliminate favoritism in assigning shifts. HCA's responsible AI policy stresses protecting people from physical and emotional harms, and its marketing features nurses praising the tool.
The dispute has reached labor and legal channels. National Nurses United, which represents about 10,000 of HCA's 100,000 nurses, says HCA has rejected its requests for information about the technology. Barker, a union leader at an HCA psychiatric center, claims several colleagues quit over Timpani and that workplace injuries have climbed in the first 12 months with the tool. The center's union of about 500 nurses filed a grievance arguing HCA fails to sufficiently honor scheduling preferences, and the matter is likely headed to arbitration. In July, Angelique Russell, a former HCA data science manager, sued HCA for allegedly firing her in retaliation for raising concerns about Timpani. She accused HCA of routinely deleting the data the tool uses to build schedules, which prevented auditing how well it worked and which she claimed may have violated several healthcare laws. HCA has yet to formally respond in court, and Russell declined to comment.
WIRED also points to a separate case. Rayus Radiology, a large US imaging chain, began receiving complaints about scheduling errors earlier this year after introducing a different tool also built on Palantir Foundry, according to internal documents WIRED viewed. The software populates scheduling fields from doctors' orders, but in a number of cases it linked requests to the wrong patient profiles, listed tests or scans that did not match the orders, or made up personal details about patients. Rayus schedulers suspect non-standard forms, illegible handwriting and insufficient human review contributed. Rayus spokesperson Kit Crancer says the broad characterizations do not fairly reflect its processes, safeguards or commitment to safe, accurate and timely patient care, and that Rayus uses technology to support, not replace, appropriate human judgment and oversight. An unnamed Palantir spokesperson declined to comment by name, though the company has described its software as merely presenting information more helpfully and has stressed that customers are responsible for data, policies and decisionmaking.
The backdrop is Palantir's growth: its sales have nearly doubled recently to almost $2 billion per quarter, and health care providers, squeezed by costs and staffing shortages, are eager customers. Cleveland Clinic credited Palantir software with boosting patient capacity by forecasting ER discharges and replacing manual coordination. Palantir has also drawn pushback from civil rights activists over its defense and immigration work and the stated political views of cofounders Peter Thiel and Alex Karp. In the UK, nurses and activists have demanded cancellation of its $440 million National Health Service contract, which aims to cut appointment wait times, over privacy and security concerns. At HCA, some nurses want Palantir gone. "Everyone checks immediately and every month you hear, 'Man, why did it schedule me like that?'" Barker says.
Key facts
- HCA Healthcare has rolled out Timpani, an AI scheduling tool co-developed with Palantir, at roughly 130 of its 190 locations since 2023.
- Nurses allege it has potentially harmed patients by understaffing teams and ignoring shift preferences; Amber Retzloff says over half of her 50 requested 12-hour shifts over four months were assigned differently.
- HCA says nursing leaders, not Timpani, make final scheduling decisions, and that on average Timpani schedules nurses for 1 percent of their requested days off.
- A union grievance at an HCA psychiatric center (about 500 nurses) is likely headed to arbitration, and former HCA data science manager Angelique Russell sued in July alleging retaliatory firing over Timpani concerns.
- WIRED also reports scheduling errors at Rayus Radiology with a separate tool built on Palantir Foundry.
Why it matters
Hospital staffing sits close to patient safety, and Timpani is a large-scale case of an algorithm replacing manual scheduling in that setting. HCA is the largest hospital chain in the US, and the tool covers roughly 130 of its 190 locations. Nurses say the software leaves too few nurses or too few experienced veterans on shifts, and in one account left a single senior nurse supervising four junior colleagues while sick patients waited. The story also shows how a promised efficiency gain, which HCA's Michael Schlosser describes in terms of fewer manager hours, less reliance on contract nurses and better retention, can look very different from the bedside. Palantir's health care push, with sales near $2 billion per quarter and a $440 million UK NHS contract, gives the dispute wider weight.
Who it affects
Mainly HCA nurses, about 100,000 in total, of whom National Nurses United represents about 10,000. They bear the back-to-back shifts, shorter notice for changes and extra time spent appealing and swapping. Patients are affected if the nurses' staffing and experience-mix allegations hold. Managers lost some control: initially some local managers could edit schedules, but appeals now go to a centralized team in Nashville. Other Palantir health care customers, such as Rayus Radiology and Cleveland Clinic, and Palantir itself, are part of the picture.
How to use it
Timpani is an internal HCA tool, and the source gives no way for outsiders to use it. The practical points are for those working with it. Nurses flag problem shifts to the centralized team at HCA headquarters in Nashville; if an appeal is denied they try to swap shifts, where Barker and Retzloff put the chance of success at about 50 percent. At one HCA psychiatric center, union nurses have filed a grievance. For other health systems weighing similar software, the story's cautions are that Palantir says customers are responsible for the data, policies and decisionmaking, and that Rayus's schedulers suspect insufficient human review contributed to its errors.
How solid is it
It rests on a single WIRED report. Its nurse evidence comes from Retzloff, five nurses at other locations, Lee Barker, and an anonymous non-union nurse in Texas who feared retaliation; the figures on missed requests and 50 percent swap success are the nurses' own counts and claims. The 1 percent red-day figure and the 98 percent mixed-skill figure come from HCA, in Schlosser's case from a paper he wrote. No independent audit or verified data confirming patient harm is given; the harm is the nurses' allegation, worded as 'potentially harmed'. Russell's claims about deleted data are allegations in a lawsuit that HCA has yet to formally respond to. The Rayus errors are documented in internal documents WIRED viewed, but involve a different tool.
Risks and caveats
HCA disputes the framing: it says nursing leaders make final decisions and that it keeps improving Timpani from nurses' feedback. The claims that colleagues are calling out more, quitting, or suffering more workplace injuries come from nurses' beliefs and Barker's claims; the source gives no counts for how many quit or how much injuries rose. The Rayus case is separate; it shares Palantir Foundry with Timpani but is not Timpani and is not an HCA tool. The arbitration has no stated date or outcome, and the source mentions no regulator or government action against HCA over Timpani. Studies show AI-based scheduling holds promise, the source notes, so the dispute concerns this deployment rather than the idea in general.
“We are overriding clinical judgment and the human-to-human piece of health care with an app.”
— Amber Retzloff, critical care nurse in Florida and National Nurses United union leader