For an example of a man who has turned a personal tragedy into a force for good, look to Martin Bromiley.
In March 2005 Martin kissed his wife Elaine and, along with his two children aged 6 and 5, waved goodbye as she was wheeled into an operating theatre to have minor surgery on her nose to alleviate her sinusitis.
She never woke up.
At first he was told it was an unanticipated emergency and doctors had done all they can.
But Martin got the hospital's permission to conduct an investigation. And he discovered Elaine's death had been preventable, if only the surgeons had followed the protocols laid out for the problem they had faced.
Martin is an airline pilot. An industry where following safety protocols is paramount, and an industry with a deep understanding of how to eliminate human error as a potential cause for accidents.
And so Martin recognised what the surgeons had done. It's the same mistake recorded on the black boxes retrieved from some of the most infamous plane crashes in history.
The surgeons had fixated on one perceived problem, and failed to realise what else was happening; that the situation was rapidly deteriorating for other reasons.
And just as the airline industry has learnt from the human errors that lead to plane crashes, Martin Bromiley wants the Health Service to learn too.
He now runs the Clinical Human Factors Group, a coalition of experts determined to put an understanding of human error at the heart of everything the NHS does.
Checklists, openness, clear leadership without oppressive hierarchy - these are the methods safety-critical industries use, the kind of strategies Martin Bromiley wants to see in healthcare too.
Which brings me to today's news: the publication of hospital trust performance on reporting mistakes.
It's been established how many errors to expect from medical treatment, and now hospitals which are reporting fewer errors than average are being rated as poor.
One in five trusts have been identified as under-reporting in this way.
What does Martin think? It's an experiment, he tells me. Done correctly it could certainly help. But he is concerned too.
"It's a question of quality not quantity" he says.
The quality of the improvements made after a mistake, the quality of learning from errors, is far more important to Martin than simply counting incidents.
"How will it drive management behaviour?" is another of Martin's questions.
If the pressure not to be labelled as poor leads to bullying of staff, the whole exercise will be pointless.
"Where's the learning?" Martin asks. It's not a question that will be answered from looking at the statistics published from today.