PSIF stands for potential serious injuries and fatalities. The term comes out of the safety profession’s recognition that managing safety by counting what went wrong is managing too late: by the time an injury is recorded, the event has already happened. PSIF shifts the focus to the events that carried the potential for a serious outcome, whether or not the outcome occurred. The word potential is the whole idea, and this post unpacks it.
PSIF vs SIF
The two terms are close and frequently used interchangeably, which is a mistake that costs clarity. SIF, serious injuries and fatalities, describes the outcome class: the events that produce death, amputation, or permanent disability. PSIF describes the precursor class: the events that could have produced a SIF but did not, often by luck. A pedestrian steps into an aisle and the operator stops inches short. No injury occurred, so the event is not a SIF. The potential was there, so it is a PSIF event. The SIF prevention post on this site covers the outcome side; this post covers the potential side and why it is the actionable one.
Why potential is the actionable measure
Serious injuries are rare, which makes them statistically useless for day-to-day management. A fleet can go years between serious events, and the absence of outcomes tells the safety team nothing about whether the exposure is growing. PSIF events are common: the near misses, the close calls, and the events that connect a mechanism with a person without the expected consequence. Each one is a sample of the exposure, and the sample rate is high enough to manage. The near-miss reporting post on this site is the operating side of the same idea: the log of potential events is the signal, and the injury count is the noise.
What counts as a PSIF event
The classification uses the same mechanism logic as the severe injury data on this site. A PSIF event is any event that involved a mechanism capable of producing a serious outcome, with the outcome avoided or minor. In forklift operations the mechanisms are struck-by, caught-between, falls from height, and crushing under a falling load. The classification question is not “what happened?” but “what could have happened?” A load that shifts and settles without hitting anyone is a PSIF event if the fall would have crushed a worker. A truck that passes within centimeters of a pedestrian is a PSIF event. The unsafe conditions post on this site covers the condition side that creates these events.
Running a PSIF program
A PSIF program is the SIF framework applied to the precursor stream: log every event with serious-injury potential, tag it by mechanism, review it with a named owner, and close the gap between the event and the control that should have prevented it. The measure is the PSIF rate, the number of high-potential events per shift, and the program’s success is the rate falling over time. The Risk Scorer on this site feeds the review by ranking the operation’s exposure, and the review feeds the fixes.
The meaning in one sentence
The PSIF rate in practice
The PSIF rate is the operating measure, and its practical definition is worth nailing down. The rate is the number of high-potential events per shift, counted from the near-miss log with the severity filter applied. The filter is the question from the SIF post on this site: could this event have killed or permanently disabled someone? The count is reviewed weekly, and the trend is the program’s leading indicator. A rising PSIF rate with a rising report volume is the signature of a trusting culture that is finding the exposure; a falling rate with a stable volume is the signature of the exposure being removed. The Risk Scorer feeds the review with the scored exposure, and the zero incident process sets the target. The rate is the number that the PSIF framework is managed by.
PSIF exists because safety management learned that the outcome is the last thing to find out about an event. The potential is visible earlier, it is measurable, and it is actionable. An operation that manages its PSIF rate is managing the serious event before it happens, which is the only place it can be managed.
