'When I was a regional director of public health I would see cases that were uncommon - but tragic. I became more and more of the opinion that history keeps repeating itself.
'Even when organisations had gone though a great deal of soulsearching the same things were happening again - a year or two later. ' Chief medical officer Professor Liam Donaldson (pictured, left) brings an almost evangelical air to the business of stamping out medical errors.
During his time as director of public health for Northern and Yorkshire region, he introduced a database of serious adverse incidents and near-misses, striding ahead of colleagues in other regions.
In his current job he has the power to do still more. Last week Professor Donaldson announced the creation of the National Patient Safety Agency - a 'mandatory reporting system for logging all failures, mistakes, errors and near-misses across the health service'. It also promises to provide a streamlined approach which will 'ensure that lessons are learnt and spread throughout the health service'.
The announcement came in response to publication of Organisations with a Memory - the report of his expert group on learning from adverse events in the NHS.
In an exclusive interview Professor Donaldson told HSJ: 'It is a very exciting time tinged with a certain amount of sadness. '
He was referring to the inquest last week into the case of leukaemia patient Wayne Jowett, who died after he was given an injection into the spine rather than a vein, at Queen's Medical Centre, Nottingham. Such tragedies appear to spur on Professor Donaldson: 'We want to eliminate those sorts of errors, and create a system where they can't happen. '
The announcement of the new agency has met enthusiasm in the service. But several HSJ sources shared the view of a Department of Health insider who warned that 'having a database is fine. It is a good idea. But how do you make sure lessons are learned?
'There is currently the problem that the data available is often poor. But even when it is good, little use is made of it. '
Professor Donaldson agrees that logging incidents is easy compared to the more complex task of ensuring that repeat performances are eliminated. But he is optimistic about the power of some of his report's ideas - simulation exercises, safety briefings before highrisk procedures and the chance to 'design safety into the system' - to hammer home the message.
And he takes issue with those who say the creation of another new outfit runs the risk of adding to confusion around reporting mechanisms within clinical governance.
Critics ask what the agency can do that the Commission for Health Improvement, National Clinical Assessment Authority, National Institute for Clinical Excellence, Modernisation Agency (including the clinical governance support team), national confidential enquiries and regional offices can't.
For Professor Donaldson, the agency's unique selling point is its degree of distance from services on the ground: 'All of the evidence from the airline industry is that we need to make sure there is a degree of independence and neutrality so that people feel comfortable giving their report. '
Asked how the agency will fit in with the other new organisations dedicated to clinical governance and performance, Professor Donaldson gives some examples.
'I think some of the things which come out of it will have national or even international relevance. If we saw 12 separate reports which criticised the packaging of a particular drug, that would be something which had implications all round. '
Evidence of local problems could mean a need for training within an organisation. Evidence ofwhole-systems failures would be passed on to CHI, while suggestions of individual doctors' failings could mean a referral to the NCAA.
Despite Professor Donaldson's own background at Northern and Yorkshire, he sees little role for regional offices in filtering information from trusts and HAs to the national database: 'I think it will be a more direct link. '
He is modest - or dismissive - about the potential of the regional system he created as the basis of a national model: 'It was good for its time but it was nothing compared to the depth and comprehensiveness we want the national database to bring. '