BCAR, a problem in a 'very safe' industry is if incidents really do preceded accidents (not just our view with hindsight), but if so, how often is 'often', particularly if there are fewer related incidents.
Your questions are very pertinent;
Facts, few; incidents, few and questionable if they relate to accidents.
Management; from self reporting, with hindsight, data driven, counting numbers, and highly questionable relevance to the issues in accident reports.
How safe is the industry; depends on the definition of safe. If safety is considered as 'what is done' vs 'what we have', then there is always room for improvement.
We need to move away from the headline fatalities, but not forgetting this is more likely to influence public opinion.
Do not use body count as it is unrelated to accident rate (
onboard fatalities page 16).
We should place greater emphasis on non fatal accidents where the risk of fatality might be the same or greater than in the fatal accidents; (
we need to look and think beyond numbers).
Also, look beyond accident reports and consider those unreported issues because they were not 'factual'; human factors, influences on human behaviour, SOPs and management pressure.
We need to question our assumptions about safety.