Enhanced statistical anlysis to improve air safety; How to be proactive

Abstract: LFV is Sweden‟s major provider of air navigation service. The organisation continuously strives towards providing better and safer air traffic control, and to minimize its part in incidents and accidents occurring. One part of the safety work is analysis of statistics, concerning different types of safety performance indicators; in other words, statistical analysis of occurrences. The numbers of occurrences in the investigated and monitored safety performance indicators are rising, and LFV therefore wants to find a way to use statistical analysis in a proactive way, instead of reactive as the use is today. We have identified the way LFV uses statistical analysis today, and compared it to alternative approaches. We will give suggestions to LFV on different approaches of statistical analysis, which can be used in a more proactive way compared with today. The information is collected through investigations of regulations, and of procedures at the organisations concerned by the Swedish occurrence reporting system. Experts from LFV have also been interviewed. We have, after discussions with LFV‟s experts, come up with an alternate approach to perform statistical analysis. The new approach is named 24 Hour Approach, and it sets the average statistics from a specific safety performance indicator for a specified period, in relation to the average amount of traffic for the same period. The occurrences are divided between the hours they took place during the day. We have also investigated the Alarm Model, an approach used by Transportstyrelsen. We discuss the Human Factors Analysing and Classification System (HFACS) as a proposal to future development. It is an approach towards the reports for the occurrences. By classifying the reports with the human factors involved with the occurrences, statistical analyses can be made on the causal factors involved. LFV have started to introduce human factors in LFV‟s occurrence reporting system, which will give the possibly to combine different occurrences (runway incursion, airspace infringement, clearance bust, etc.) within the same safety performance indicator, thus performing statistical analyses on causal factors instead of occurrences. We conclude and recommend LFV to use the provided 24 Hour Approach on its statistics. LFV is also recommended to keep up the on-going work with integrating human factors into the Synergi system, and to adapt the Risk Analysis Tool provided by EUROCONTROL.