Software may decide how many nurses care for you tonight. Nurses are pushing back. Staffing is a patient safety issue. When a unit has too few nurses, responses slow, changes in condition go unnoticed and escalation stalls. Delays like these can become the heart of a medical malpractice case. On August 27, National Nurses United planned protests in eight cities against Palantir, a data analytics company. The union calls it its largest coordinated action against the company yet. Its demand is simple: hospitals should end their Palantir contracts.
The union represents more than 225,000 registered nurses. It says Palantir now manages data systems tied to more than one-fifth of U.S. hospital beds. In addition, those tools handle staffing and scheduling.
Who Controls Nurse Staffing?
The union’s press release points to HCA, the country’s largest for-profit hospital group. According to NNU, HCA uses a Palantir-built tool called Timpani to schedule nurses. The union says it removes local managers from staffing decisions. Instead, an anonymous office in Nashville makes them. Does profit pressure shape staffing decisions like these?
NNU also argues that executives are rushing to use untested, unregulated technology in place of clinical judgment. As a result, the union says, software now helps decide how many nurses care for patients on each shift.
However, hospitals are split. MaineHealth defends its use of a Palantir tool, while NYC Health + Hospitals is winding down a roughly $4 million contract.
Patients Left in the Dark
The union says HCA and MaineHealth have refused to disclose how far their work with Palantir goes. That includes how they protect patient data. Because no one describes the risk, patients cannot weigh it.
In fact, what do privacy notices actually tell patients about where their information goes once they share it with a provider?
Three questions deserve straight answers. First, should patients know how, and why, their data is protected? Second, should hospitals need opt-in consent before using records for research or AI training? Finally, can patients say no to AI in their own care?
Honest answers would require real informed consent conversations. However, few hospital systems will want to have them. After all, they take time, and the answers may not be popular.
Washington’s informed consent law lets a patient bring a claim when a provider leaves out an important fact about their care and the patient agrees to treatment without knowing it. The question is whether an ordinary patient would have refused treatment had they known. Whether staffing software and AI tools count as important facts is still open. However, a patient who learns about them only after harm occurs has lost the choice. Instead, understaffing can lead to exhaustion and medical errors. When this happens, the hospital may face exposure for corporate negligence. Software can build the schedule. Can it answer for the outcome?
