Everyday work as the starting point
FRAM starts with how activities succeed under variable conditions. It does not assume that an unwanted outcome requires a single broken component or someone acting incorrectly; it examines how functions and adjustments combined within the system.
This approach connects with Safety-II and organisational resilience. Understanding what normally works reveals dependencies that may remain hidden when investigation starts directly with the last visible error before an event and ignores ordinary operating practices.
Describe functions and connections
Functions are characterised through inputs, outputs and other relevant aspects: preconditions, resources, controls and time. Filling every box is not the objective; the analysis needs the aspects necessary to understand the activity and its connections with other functions.
The model is not simply a flowchart. It represents potential dependencies between functions; a particular situation allows examination of which become active and how variations combine, for example in timing, precision or availability of information.
A coordination example
Imagine an intervention whose start depends on authorisation, available materials and confirmation of an area’s condition. Small adjustments within those functions may be compatible on most days yet coincide unfavourably on a particular occasion.
Analysis would examine the connections enabling that coincidence and the information needed to detect it. Findings may concern coordination or shared resources rather than merely requiring each participant to perform their individual task more carefully.
Contribution and limitations
FRAM supports retrospective analysis, future scenarios and design. It requires information about actual work and review with people who understand the system. Clear boundaries and assumptions matter more than drawing an extremely large network of functions.
Its core is qualitative. Quantitative extensions exist, but no universal probability follows automatically from a FRAM diagram. Recommendations need conversion into concrete changes and checks during operation; the model helps formulate questions and understand relationships rather than replacing evidence or independently guaranteeing safety. The team should retain the reasoning connecting each proposed change with the variability it seeks to manage.
