Safety cases performance measurement

A Safety Case is a narrative that literally makes the case that an adequate level of safety has been reached for an installation. It requires looking at all potential hazards which could lead to a loss of the installation, a loss of life, or a major pollution event. A risk analysis is performed on each hazard evaluating the probability of the event occurring and describing the magnitude of the consequences. A discussion is then given of the measure undertaken to lower the probability of occurrence or to mitigate the consequences and a case is made that the risk for the installation meets the ALARP safety criteria. 

Alternatively we may decide to create more specific claims, perhaps that the system performs at a level that is conducive to practical operation in a clinical environment and/or with a defined degree of availability. Whatever measure is selected it should be clear and concise and, most importantly, be capable of being demonstrated within the scope of the planned CRM assessment process. In other words it should be possible to draw lines of logical inference between each claim and the arguments set out in the body of the safety case. 

Each of the safety performance measurement methods can be characterised as a combination of the following three facets (a) data source (b) identification procedrrre and (c) data category. Typical data sources for safety performance in healthcare are patient case records, incident reports, field data (for example, obtained by observation), and self-reported data (typically elicited by the use of a questiormaire or interview). 

To overcome this weak point of self-reported data, we propose eliciting safety-related attitudes and indications of actions by asking respondents to report their likely actions with regard to fictitious adverse event cases (vignettes). This method for safety performance measurement will be mentioned later in this section, illustrated with some case studies. 

National consensus standards can be used to formulate performance levels and safety metrics. Because the standards are already developed, and most cases viewed as an industry-accepted level of performance, the safety manager may choose to adopt the standards—thus eliminating the need to create new performance measures. Because the national consensus standards have been adopted as acceptable performance within an industry, the additional burden of having to validate acceptability has also been addressed through the standards development process. 

OSHA-Based Incidence Rates one of the more common methods of measuring safety performance, including OSHA recordable injury incidence rates, recordable illness incidence rates, death rates, and lost day case injury rates. 

For a project this author had undertaken, there was a need to collect actual cases where initiatives undertaken by safety professionals to resolve injury and illness problems also resulted in improved productivity and cost efficiency. Eleven safety directors were asked for help. Surprisingly, all agreed to contribute their success stories. That was an important indicator of how their jobs had changed. They were into productivity, cost efficiency, and performance measurement (Christensen and Manuele, 1999,4). 

Now, do some wild speculation and consider whether the following example begins to have credibility. An employer has 500 employees who work I million hours a year. Do the OSHA statistics, the recordable case rate and the lost-workday case rate, for an exposure of I million hours begin to achieve a reasonable confidence level as measures of the quality of safety performance Perhaps somewhat, but additional performance measures are needed. An entity of this size would more than likely purchase workers compensation insurance. Cost trending and the trending of the company s experience modification rate serve as additional measures. 

With regard to reactive analysis it is often the case, particularly in relation to fire, that a small number of incidents or accidents may lead to complacency and a belief that all is well, when this may not necessarily be the case. Statistics themselves are likely only to demonstrate a knowledge of outcomes, i.e. twisted ankle during evacuation and not the cause such as poor flooring conditions presenting a slip and trip risk. It is therefore essential for a combined reactive and proactive approach to be taken when considering safety monitoring or performance measurement systems. 

The conclusion of the Safety justification was that a set of restrictions, dependencies and other safety measures had been identified which enabled the train operations, associated with V2.01 testing, to be performed at a level of risk which is tolerable and ALARP. This Safety justification can thus be seen a subset of the reasoning model that is the System Safety Case for VLUP. 

The Functional Safety Assessment is described in part 2 as part of Confirmation Measures and in part 4 of ISO 26262 at the end of system development after the safety validation and before the release for serial production for the product to be developed. Primarily object of the Functional Safety Assessment is the assessment of the Safety Case according to the requirements in part 2 of ISO 26262. The requirements of how the Functional Safety Assessment should be performed are given in part 4 of ISO 26262. 

In order to be measure safety performance a consistent set of terms and reporting standards is required. In the area of occupational safety, considerable standardization has already been achieved through the use of measures such as the number of first aid cases or recordable injmies. Although different organizations will apply these terms slightly differently from one another, there is sufficient consensus to allow for their use across broad swathes of industry. These data are referred to as lagging or trailing indicators. 

Teaching human factors principles, motivational evaluation of safety measures and performing selected safety measures. - The training program starts with the presentation of a common safety problem to the audience. On the basis of this case study the motivation of accident involved individuals are examined. Safe and unsafe behavior will be discussed with respect to the following topics ... 

Non-prescriptive management programs are always performance-based because the only measure of success is success. And success can only be achieved by not having incidents. But, from a theoretical point of view, such a goal is impossible to achieve. No matter how well-run a facility may be, accidents will occur risk can never be zero. Accidents can always occur. For this reason an organization that follows a non-prescriptive RMP can never be in compliance. Do whatever it takes on your facility not to have accidents. It is up to the managers, the technical experts, and the opera-tions/maintenance personnel to determine how this should be done. (Of course, detailed rules do have to be followed when they apply the safety case is not a justification or excuse for avoiding compliance.)... 

Development candidates must be measured against multiple performance criteria, including such aspects as potency, safety, and novelty. Conflict may be experienced between the criteria, in which improved performance in one criterion can only be achieved at the expense of detriment to another. In this situation—as is often the case for activity against bioavailability—a trade-off is said to exist between the objectives. A trade-off between potency and safety may also be present. 

For values obtained in the neighborhood of the MRL, quantitative accuracy and precision are of extreme importance. Thus, a laboratory should take a safety measurement zone depending upon the validation performance of the used method. Questions may also sometimes arise as to which way some numerical concenttation values (such as 1.56 ppb) would be rounded to be converted into a yes or no answer when the action level is, for example, 2 ppb. In that case, caution should be exercised with regard to the number of significant values to reflect the precision of the applied analytical method. 

---