I’m honoured to be regularly invited to share my safety case expertise with organisations in the form of delivering our Assurance Case Masterclass Course which runs over two-days. At the beginning of every course we ask the cohort what they hope to take away from the course (and why), and also ask them to share their biggest ‘bugbears’ with the use of safety cases. Regrettably, there are a few recurring (if not constant) themes we hear, so I thought it prudent to discuss them and offer some potential solutions.
In this article I’ll consider each of these themes and potential solutions in turn.
We don’t get the right safety case artefacts from our suppliers
There are many reasons why your supplier may not be delivering the required safety case artefacts to you at the right quality, in the required manner, at the right time. The first question to ask though, is ‘have you asked for the right things?’ If you have, and they’re not delivering, then this is a legal and / or commercial matter to resolve. But let’s consider when this is NOT the case.
In the ‘Purpose of an Assurance Case’ article , I consider the ‘who’ of a Safety Case.
The ‘who’ of a safety case are your stakeholders, and one type of stakeholders are the contributors to it. It is vital to not only identify these contributors at the outset of your project, but to also define explicitly what you require of them – when; to what quality; and to what purpose.
A supplier cannot possibly hope to deliver the correct artefact at the right time, at the right quality if they don’t understand your needs, nor understand the role their product(s) play in the causal chain to your hazards. So, make sure your stakeholders are fully informed, and know precisely how they contribute – and the role they play in the delivery of your safety case. Expectation management is key in other words.
The customer doesn’t use the Safety Cases we supply them
Safety Cases are designed to inform decisions, and help decision-makers – even if that decision-maker is you. The decisions a safety case informs are many, but include:
- the acceptance of fundamental concepts and safety strategy
- the decision whether to deploy
- whether changes are required
- whether further evidence needs to be collected
- whether and when to retire an entity, and
- the prioritisation of actions.
During the course, we are somewhat brutal in asserting that if your safety case is not informing decisions, then you are wasting your time. BUT, if you believe your safety case is shelfware, then you should invest the time in extolling the benefits of a strong safety case regime to your clients and internal stakeholders.
Our Safety Cases don’t focus on the things that matter
When the entry criteria for an item into a safety case is whether something could have any bearing on safety or the assurance of safety, just about any item, argument, or claim is admitted. This results in a rambling, incoherent safety case in which it is not possible to ascertain the most significant risks, nor even what the claims are. This is solved by the use of targeted approaches such as Risk, Confidence, Conformance (RCC) [1] in which the strands of risk, conformance with standards, and associated confidence (with both) are separated into distinct (but linked) arguments.
Proportionality in a safety case is key is achieving focus. Even if time and resource were infinite, there is still no such thing as ‘absolute safety’, and evidence and argument are – by their nature – imperfect. The use of a Confidence leg in an RCC Safety Case facilitates two key goals, therefore. Firstly it allows the safety case author to apply effort proportionate to the areas of greatest risk, and secondly it allows the author to highlight the nodes of a safety case which require more scrutiny.
This article is a quick canter through some of the issues with safety case practice – and deliberately so. Although there is no one-size-fits-all approach available to rectify the issues with safety cases, I hope these small pointers will be of value.
Reference
- Assurance Case Guidance, Challenges, Issues, and Good Practice, Version 1, The Assurance Case Working Group (ACWG), SCSC-159.
