Aaaah! "Maintenance induced failure". This is the first cousin of "infant mortality" - premature failre by stuffing up replacement with new parts.
I had a classic at Ansett many years ago. The MTBF of F27 gyros was only a few thousand hours over many years. Sunfish analysed the data - and found a classic infant mortality curve graphing failure vs time in service. What happend was that the delicate gyro bearings were being stuffed (brinelling) by ham fisted installers.
Sunfish solution: stop changing them so often, TBO increased by ten thousand hours. Problem solved