System safety engineering is embraced at NASA from the beginning of the program/project life cycle to the end. Historically, an assurance model has been the paradigm, expressed at each life cycle stage via oversight or insight into requirements development and compliance.
Assumptions are made to identify critical areas of risk so that advanced analytical tools such as Probabilistic Risk Assessment (PRA) can be reasonably and efficiently applied. This has proven to be a successful technical approach, except when the assumptions themselves miss scenarios driven more by complex social interactions.
We can learn from a sentinel 1988 event in the petroleum industry: the loss of 167 personnel and $3.4 billion damage following fire and explosions on the Piper Alpha offshore oil platform. Design flaws hindering communications, emergency procedures and evacuation conspired with an unfortunate configuration change and deficient work permit process to doom workers. The North Sea oil drilling industry changed dramatically as a result, with new regulations calling for a "safety case"--a compelling set of documents that could prove a drilling system was safe to an acceptable degree. Ever since the safety case concept was developed, the entering assumptions for safe system development and operation could be covered completely and systematically.
The NASA System Safety Handbook, Volume 1 is your source to discover how the NASA safety case, called a Risk-Informed Safety Case (RISC) should be constructed.