Official rates of suicide are perhaps the most important and enduring measures of population mental health. They are however prone to variations in reporting usually relating to deaths where the intention was uncertain, though most official statistics circumvent this by including ‘events of undetermined intent’ (ICD10 Y10-34 and Y87.2) along with ‘intentional self-harm’ in their official statistics. It is however unclear how successful this strategy has been and whether significant sources of bias still persist.

To systematically examine the dramatic change in rates of death from suicide (and undetermined intent) in Northern Ireland, that coincided with a major overhaul and reorganisation of the Coroners Service in 2005/6, to understand the extent to which the initial investigation by the coroners’ office, legal processing, registration and coding practices can influence official suicide statistics.

In the space of one year, Northern Ireland went from having a standardised rate of suicide (incl undetermined intent) of 12.6/100,000 in 2004 to 26.6/100,000 in 2006 (a 111% increase) and in doing so went from having consistently the lowest to consistently the highest registered rate of suicide in the UK.

Initial analyses rules out changes in the police service, pathology service or registration coding practices as causative and suggest that the introduction of a coroner's liaison officer (with a formal data gathering role) was the most likely factor. The centralisation of the coroners’ service also reduced variations according to age, sex, geography and coding.

The process underpinning official statistics need to be routinely scrutinised.

The authors have not supplied their declaration of competing interest.