Geoff (no thread creep, this is the S92 not the accident)
For extremely good reasons (the necessity to avoid the skewing of results), analysis is done on world-wide data. In one of the original texts we had a clause that incidents/trends of a local nature would attract the attention of the operator and the Authority. For some regions, this would have had some effect (if for example they were clustered around the North Sea - just like lightning strikes) but not for others. When it comes to individual States and Operators, the skewing effect of reasonably probable events is too damaging to be of much use. I'm not sure what the latest figures are but the last time I saw any world-wide figures (August 2008) for the S92 they were at about 200,000hrs (some would argue statistically insignificant).
Maintenance faults are really an anomaly; for example if someone does not correctly refit a B-Nut should that effect the reliability analysis? Yes, it would attract some rectification action (probably a memo to field) but under the current guidance, it would not be counted in the index -the same is true for FOD (although they would both figure in the analysis). When the system was first mooted, it was suggested that such failures should not even be contained in the system/report; common sense prevailed of course as one of the significant reasons for the system was to provide confidence through transparency - not just the reliability analysis and index.
Maintenance faults might be seen as a weakness in the system but they are part of the oversight regime of a State and should be addressed by the appropriate Authority. The manufacturer can assist with this by analysing procedures to ensure that they are not too complex and intervention is warranted (improves safety and does not reduce it).
If you read the regulations, you will see that, whilst there is an absolute reliability figure 1x10^-5; there is the ability to consider a worse rate if it is associate with an improving trend. If for example there were to be a couple of failures caused by a bad batch of spur gears, fleet-wide replacement of the offending gears (and monitoring of the resulting efficacy) would provide the ability to re-assess reliability.
The benefit of this system is that we can now see all occurrences of engine-failures and each State (region) can make a judgement about the resulting analysis and reliability assessment. Before this, it was just a finger-in-the-air. We have to be pleased at the introduction of this system because, as you well know, 'we cannot manage it unless we can count it'.
In some sense one of the dangers of having raw data - such a CADORS - presented to us is that it causes the stomach to knot. Just removing the duplications would have made it appear less threatening.
Jim