HCA nurses say Palantir’s Timpani is failing the schedule test

Nurses at HCA Healthcare say Timpani, a scheduling system built on Palantir’s Foundry platform, is leaving shifts short-staffed, disrupting rest days and pushing more work onto the people it was meant to help. HCA says nursing leaders—not the software—make final scheduling decisions, and that the system is designed to match staffing with patient needs. The dispute is about more than whether an algorithm can draft a roster: it is about who can challenge that roster when the people working it say it is unsafe.

What the schedule gets wrong

Five HCA nurses told WIRED they had seen Timpani assign too few nurses, or too few experienced nurses, to a shift—especially on Sundays. They said the system often overlooks requests for breaks between shifts, limits on nights or weekends, and specific days off.

The consequences are not confined to the roster. Nurses described spending more time disputing assignments and trying to trade shifts than they did before Timpani. Some colleagues, they said, increasingly take paid or unpaid leave to avoid assigned shifts. Taking leave more than a few times a year can put a nurse at risk of being fired.

One nurse, 34-year-old National Nurses United member Rätzloff, has worked at her hospital for nearly two years since Timpani was introduced. She recalled shifts where all four of her colleagues were novice nurses. As the only experienced nurse, she had to delay care for the most seriously ill patients while teaching newer staff to handle less complex cases.

Rätzloff said HCA “lets AI dictate decisions,” putting an app ahead of clinical judgment and human interaction. Nurses also described late requests to find replacements and temporary transfers to other teams—changes they said managers used to discuss with them in advance.

HCA spokesperson Harlow Summerford said nursing leaders make the final scheduling decisions, not Timpani. He rejected the claim that the system is intended to reduce staffing at the expense of patient care, saying HCA wants to match patients with staff who have the right skills and experience while taking nurses’ preferences into account. The company, he said, continues to improve Timpani based on nurses’ feedback.

The one-percent defense

HCA says Timpani assigns nurses to an average of 1 percent of days they asked to keep free. The company calls those dates “red” days. Missouri nurse Lee Barker says the figure sounds small, but before the system arrived, nurses at her facility were never scheduled on days they had marked off.

Less than a year after Timpani was introduced, Barker said, almost every nurse had been assigned work on a day they expected to have off. She had to reschedule a long-planned doctor’s appointment for a chronic condition because she could not trade her shift. The experience, she said, felt like nurses were being “pushed when they’re already down.”

HCA made small changes after complaints, but nurses in several states say the problems remain. National Nurses United leaders say HCA has refused to share information about the technology. The union represents about 10,000 of HCA’s 100,000 nurses.

Some details emerged in July, when former HCA data science leader Angelica Russell sued the company. She alleges she was fired in retaliation for raising concerns about Timpani, and says HCA regularly deleted data the system used to build schedules. Russell argued that the deletions made it impossible to assess the system’s performance and may have violated several health-care laws. HCA has not formally responded to the lawsuit; Russell declined to comment.

Timpani uses patient-flow and staffing forecasts to generate schedules automatically. When HCA launched it in 2023, chief innovation officer Michael Schlosser told WIRED it would save time and money and eliminate favoritism in shift assignments. In a July article, Schlosser said the system had sharply reduced the hours managers spent scheduling each month, lowered reliance on expensive contract nurses and improved retention. He said more than 98 percent of schedules paired employees with different skills and experience.

Source: wired.com

The missing measure

The gap between those claims and the nurses’ accounts is not just a disagreement over whether the software performs well. It is also about who gets to inspect the process and correct it. Some local managers could initially edit Timpani schedules. Nurses now have to send complaints about problematic shifts to a centralized HCA team in Nashville. Nurses say the team reviews complaints without speaking to them or hearing their circumstances. When a complaint is rejected, Barker and Rätzloff say shift swaps work only about half the time.

Barker, a union leader at an HCA psychiatric facility, says several colleagues quit because of Timpani and workplace injuries increased in the system’s first 12 months. The facility’s union, which includes about 500 nurses, has filed a complaint alleging HCA does not adequately account for staff preferences when making schedules. The dispute is likely to go to arbitration.

My read is that HCA’s 1 percent figure answers a narrower question than the nurses are asking. A small share of marked-off days can still matter if the assignments land on days nurses rely on, and the number says nothing about whether staffing levels or experience are right on a given shift. The more important test is whether nurses can understand and contest a decision before it affects care.

That question matters beyond HCA. Palantir’s quarterly sales have nearly doubled recently, approaching $2 billion, as organizations turn to the company for data analysis and automation. Health systems facing rising costs and staff shortages are among the interested buyers. Cleveland Clinic has linked Palantir software to increased hospital capacity by using it to predict emergency-department discharges and replace manual coordination.

But Palantir’s work in health care also draws scrutiny. In the UK, nurses and activists have called for the cancellation of the company’s $440 million National Health Service contract, citing concerns about privacy and data security. At Rayus Radiology, a large US imaging-center network, internal documents reviewed by WIRED showed complaints this year about scheduling errors after a program built on Foundry was introduced. Staff said it sometimes linked requests to the wrong patient profiles, entered exams that did not match doctors’ orders, or invented patient details. Rayus spokesperson Keith Cranser said broad descriptions of the problems did not properly reflect the company’s processes, safeguards or commitment to safe, accurate and timely care. He said its technology supports rather than replaces human judgment and oversight.

Palantir declined to comment on the record. The company has described its software as a way to present information more usefully, while saying customers remain responsible for the data, rules and decisions.

HCA’s responsible-AI policy promises protection from physical and emotional harm, and company leaders have said nurses were consulted during Timpani’s development. Some nurses featured in HCA marketing praise the system’s ease of use and time savings. Others say they feel ignored and disrespected. Hannah Drummond, an incoming board member of a National Nurses United affiliate, said she and her colleagues at an HCA hospital in North Carolina would oppose any rollout of Timpani.

Schlosser once acknowledged that health-care workers may be slow to adopt new technology when human lives depend on it. He imagined a future in which frontline nurses would ask for more tools like Timpani. “Then it’ll be fun,” he said. Three years after launch, the conflict is over who gets to decide whether the system has earned that trust.