Margaret Squires (Letters, 22 June) asks about the availability of death rate statistics in the NHS. These are not available at ward level, because patients are often moved between wards, and ward of death is not a standard data item. However, since 2013, what is now NHS Digital has published hospital trust-level mortality statistics in the summary hospital-level mortality indicator. This uses a well-designed and transparent (to those with enough methodological background) technique derived in part from statistical process control, and presents the data in easily understood charts. I suggest the funnel-plot presentations as easiest to interpret. The data is issued quarterly as rolling annual figures, and can be most quickly found by a web search for SHMI. I reported regularly on this to my director of public health, and the PCT and later CCT of our area. The data was closely monitored by the top management.
Data at GP or GP practice level is less easily accessed. In the aftermath of Shipman, the NHS and Ordnance Survey Vital Statistics Branch combined death certificate data with patients’ GP registration data, to produce the Primary Care Mortality Database. Extracts for patients registered with a GP in, who die in, or who were resident in a local authority are supplied to the director of public health or nominated deputy. These data are not publicly available because they include personally identifiable data, and the confidentiality restrictions on who may see the data and what purposes the data may be used for are very stringent. The intention, never I think explicitly stated, was that PH departments would use it to monitor their own GP mortality rates and give early warning of another Shipman. Anonymised reporting from the database would be permitted.
However, this is raw data, and to extract valid and useful conclusions requires a fair degree of statistical sophistication, as well as other data relating to size and characteristics of GPs’ register lists. This is a background task that might not be an obvious priority to the organisation without a steer from on high. I did analyse these data for my local PCT, but I don’t know how many others did similarly. When I presented what I’d done to a London-wide PH intelligence analysts’ meeting it was clearly a new idea to almost everyone present.
Lewes, East Sussex
• Join the debate – email firstname.lastname@example.org
• Read more Guardian letters – click here to visit gu.com/letters