philipat;
Just to be certain, my comment regarding a configuration error as statistically "not significant" - it is indeed a serious, though highly infrequent operational issue.
In our flight data analysis program, we examine about 60,000+ departures per year of one fleet type. While we see other events of significance, we do not see takeoff configuration events. I further submit, again, without diminishing the seriousness of even one such departure!, that six accidents in 30 years (and many millions of departures) attributed to configuration error is not as statistically significant as, say, CFIT or mid-air/ground collision accidents, from which the industry has learned a great deal and dealt with in terms of technical and procedural response.
What should be done, in your opinion, to better, or more relevantly, address the issue of the three killers?
Well, everything and nothing. Human error is inevitable. What is not inevitable is individual response to error. Error can be engineered out of complex systems to a certain extent but no engineer and no one pilot can suggest ways which might eliminate all potential for error in all cases. Many times a beautifully-designed and even elegant system, like software itself, can perform for years and decades, only to be placed in singularly unique circumstances which, unforseen by all, defeats all efforts at prevention. Sometimes "who woulda thought that..." is the only response. Sometimes hindsight is genuinely chagrined and sometimes even learning takes place. Recall the many comments about money. In aviation, ever-moreso today, it is extremely difficult to get aviation managers and especially those who don't fly, to comprehend the risk that is so clearly perceived by some and never seen by others. Justifying expense, which in aviation is always very high, is tougher than ever. Sometimes it takes quite a lot of work and focus to ignore what is very plain to others in terms of risk but by and large the industry is very responsive to demonstrated risk.
As such, I personally would have nothing further to say with respect to "killer items" as this thread covers them sufficiently, for those with eyes, anyway. They're well known, well understood and from Hour One, are part of a pilot's toolkit for staying alive and keeping his/her passengers similarly alive. Certainly, nothing I say will add measurably to the helpful and professional dialogue that has already been extended on the subject.
For a very good read on these matters and human factors, I highly recommend that you read Sidney Dekker's, "
Just Culture: Balancing Safety and Accountability" or "
The Field Guide to Understanding Human Error" both available through Amazon and elsewhere.