How ECRI made the list

ECRI and the Institute for Safe Medication Practices (ISMP) reviewed millions of reported patient safety events and surveyed their members to nominate patient and workforce safety concerns for evaluation. Nominators supported their proposals with information from scientific literature, event reports, causal analyses, and more.

A cross-disciplinary team of ECRI and ISMP experts then evaluated each nominated topic using the following criteria:

• Severity: How seriously would patients be harmed if this safety issue occurred?
• Frequency: How likely is this safety issue to occur?
• Breadth: How many patients would be affected if this safety issue occurred?
• Insidiousness: Is this issue difficult to recognize or challenging to address once it occurs?
• Profile: Would this safety issue put significant pressure on an organization?

Based on this, the team chose and ranked the top 10 patient safety concerns for 2025.

The top 10 patient safety concerns for 2025

ECRI and ISMP concluded that the top 10 patient safety concerns for 2025 are:

1. Risks of dismissing patient, family, and caregiver concerns
2. Insufficient governance of artificial intelligence (AI) in healthcare
3. The wide availability of and viral spread of medical misinformation
4. Medical error and delay in care resulting from cybersecurity breaches
5. Unique healthcare challenges in caring for veterans
6. The growing threat of substandard and falsified drugs
7. Diagnostic errors related to cancers, major vascular events, and infections
8. Persistence of healthcare-associated infections in long-term care facilities
9. Inadequate communication and coordination during discharge
10. Deteriorating community pharmacy working conditions

In addition, as more health systems explore the clinical uses of AI, they need to track which devices use AI, understand when and how developers update software, and work more closely with vendors, ECRI said. Without the right guardrails, AI could skew results based on data biases and lead to unnecessary care.

"AI has a lot of potential to improve outcomes, increase access, create efficiencies and reduce disparities. But AI has the potential to cause harm if it is not governed properly," said Scott Lucas, VP of device safety at ECRI. "Providers and payers need to be transparent on AI tools used in treatment, and if there are concerns, they should be considered and discussed with patients."

ECRI said solutions to its list of problems require a holistic approach that considers how every aspect of a health system can promote safety.

"Patient safety events are not isolated incidents. They are often products of the system that clinicians and patients operate within, and how that system supports the people it serves," said Shannon Davila, executive director of total systems safety at ECRI. "Tackling threats to patient care requires rejecting the current fragmented approach and designing systems that promote a true culture of safety."