Key takeaways
- Accredited hospitals must run a proactive risk assessment on at least one high-risk process every 18 months; FMEA is the best-known method.
- Map the process as it really runs, with the people who do each step, before listing failure modes.
- Treat scores as a ranking aid; review catastrophic-severity failure modes even when their total score is modest.
- Give every mitigation an owner, a due date and a measure, and re-check after go-live.
What FMEA is
Failure mode and effects analysis (FMEA) is a proactive, team-based method for finding how a process could fail before it harms anyone. The team maps the process step by step, lists the ways each step could fail (failure modes), their likely causes and effects, scores how serious each one is, and plans actions for the highest-risk failure modes.
FMEA began in US military reliability engineering, where a 1949 procedure (MIL-P-1629) set out how to analyze the effects of equipment failures, and it was later used widely in aerospace and manufacturing. In 2002, the VA National Center for Patient Safety published Healthcare FMEA (HFMEA), which combines ideas from FMEA, the food-safety hazard analysis method HACCP and the VA's root cause analysis tools.
There are two common ways to score failure modes. Traditional FMEA, used in IHI's tool, rates the likelihood of occurrence, the likelihood that the failure would not be detected and the severity of the effect, each from 1 to 10, and multiplies them into a risk priority number (RPN) from 1 to 1,000.
HFMEA rates severity and probability on four-point scales to give a hazard score from 1 to 16; failure modes scoring 8 or more, or that are a single point of weakness, go through a decision tree that asks whether existing controls or detection are enough.
Under The Joint Commission's National Performance Goals effective January 2026, accredited hospitals must select at least one high-risk process every 18 months and conduct a proactive risk assessment. FMEA is the best-known method, but the standard does not mandate a particular tool.
Sources
- Institute for Healthcare Improvement, "Failure Modes and Effects Analysis (FMEA) Tool" (opens in a new tab)
- DeRosier J, Stalhandske E, Bagian JP, Nudell T, "Using Health Care Failure Mode and Effect Analysis: The VA National Center for Patient Safety's Prospective Risk Analysis System," Joint Commission Journal on Quality Improvement 28(5), 2002 (opens in a new tab)
- The Joint Commission, National Performance Goals, effective January 2026 for the Hospital Program (NPG.02.03.01) (opens in a new tab)
When to use FMEA
Use FMEA before a new process, piece of equipment or technology goes live, or when redesigning a high-risk process: before the first adverse event, not after it. It is especially worthwhile for processes with many handoffs, processes where a failure could cause serious or irreversible harm, and high-volume processes where even a rare failure happens often in absolute terms.
FMEA looks forward. To understand an event that has already happened, use root cause analysis tools such as fishbone, 5 Whys or HFACS, and use what you learn to update the FMEA.
ImprovementFlow
FMEA and Proactive Risk Assessment
ImprovementFlow runs FMEA alongside the other analysis tools teams use, and turns the mitigations it produces into plan steps with an owner and a due date, with linked measures and a follow-up review to confirm the change held.
How to run an FMEA
- 1Choose a high-risk process and define its scope: where it starts, where it ends and which units are included.
- 2Assemble a multidisciplinary team that includes people who do each step, plus an advisor who knows the method.
- 3Map the process as it actually runs, numbering each step and sub-step.
- 4For each step, list the failure modes, their possible causes and their effects on the patient or the process.
- 5Score each failure mode with your chosen scale (RPN or HFMEA hazard score) and agree which ones need action.
- 6Plan actions with an owner, a due date and a measure, favoring changes that remove or design out the failure over reliance on vigilance.
- 7After go-live, check the measures and the event reports, and re-score the failure modes you acted on.
Example: automated dispensing cabinets before go-live
This is an illustrative example. A pharmacy-led team runs an FMEA before installing new automated dispensing cabinets on three inpatient units, walking through retrieval, restocking and returns step by step.
The highest-risk failure modes cluster in three places: selecting the wrong patient profile when retrieving a medication, loading a medication into the wrong pocket during restocking, and expired stock remaining in the cabinet. Each is serious, plausible on a busy unit and hard to detect before the dose reaches the patient.
Before go-live the team adds barcode scanning when cabinets are restocked, a second-person check for high-alert medication pockets and an expiration-date report for pharmacy technicians, each with an owner and a due date. A follow-up review three months later checks event reports and audit results against the failure modes the team identified.
How ImprovementFlow supports this
- FMEA is one of the analysis tools teams can use within Improvement Projects, alongside 5 Whys, fishbone diagrams and process mapping.
- Mitigations become plan steps with an owner, a team and a due date, so actions agreed in the FMEA meeting are tracked to completion before go-live.
- Linked measures and follow-up reviews show whether the redesigned process is performing as intended after go-live.
