We ran a piece yesterday about the apparent high death rates at St Mary’s that were highlighted in the Nationally-reported figures by Health researchers Dr Foster.
In the morning, before we published, we emailed the Head of Communications at St Mary’s, Andy Hollebon, asking him for the Isle of Wight NHS Trust’s side of the story. Waiting nearly two hours, we hadn’t heard back from him, so we published the story.
St Mary’s response
At 4pm Andy Hollebon sent VB an email pointing us to a page on the St Mary’s site (we’ve duplicated it below).
We’ve read it over a few times, but are still struggling to fully understand what it says, but we _think_ they’re saying that the way that the hospital collected the data made the figures appear higher than they actually are – and that Dr Foster were aware of this in advance of publishing their report.
Asking for further clarification
In the correspondence we had with Dr Foster yesterday, at no time did they raise the issue of data collection at St Mary’s with us. We’ve forwarded the St Mary’s statement to Dr Foster, asking for their response.
We’ve asked St Mary’s for the “evidence from Dr Foster that the figures showing a higher than average mortality ratio are explained by the previous coding practice of coding from hospital notes,” that St Mary’s refer to in the statement.
St. Mary’s statement
Dr Mark Pugh, Medical Director at St. Mary’s Hospital on the Isle of Wight said:
“We fully investigated this issue in 2010 which for St. Mary’s Hospital is mostly one of data collection. We have evidence from Dr Foster that the figures showing a higher than average mortality ratio are explained by the previous coding practice of coding from hospital notes. A recalculated hospital standardised mortality ratio (HSMR) done by Dr Foster using this years figures and reported to the Trust on 2nd November 2011 gives an HSMR of 110 based on last years base and an estimated rebased figure of 117, which although high is within the normal range.
“A high mortality ratio on its own is not an indication of poor standards of care but it is a trigger to take action. Since November last year (2010) we have been coding all deaths from patient notes rather than discharge summaries, so for example previous data did not include all co-morbidities. We will continue to monitor the situation closely, and are participating in regular mortality reviews but as we highlighted last year it will take time for these changes to come through in the nationally published data.
“Although I am confident that St. Mary’s is a safe hospital I will not be satisfied until the figure is closer to or below 100. We expect the coming months to show further reductions. To ensure that we resolve every issue we are doing further work to provide the necessary confirmation that the figure is coming down through improved data collection.
“NHS Isle of Wight supports publication of information about hospital mortality ratios. Monitoring mortality ratios is just one of the ways in which we measure the clinical effectiveness and quality of our services and ensure the safety of patients. Since June 2009, as part of our transparent and open processes we have published mortality data as part of the Quality Report at our Public Board meetings. However recognising the complex issues involved in the analysis of mortality ratios the NHS Isle of Wight Board asked for the local data to be reviewed.”
The 2010 review of the collection of mortality data for St. Mary’s Hospital showed that:
- The then Hospital Standardised Mortality Ratio (HSMR) was based on data collected by St Mary’s Hospital which was then analysed by the Dr Foster group based in London. The method aimed to correct for the different ages of patients and their underlying condition comparing the local outcome with the average national outcome. Other groups performed similar analysis on the same data often producing quite different results. The mortality ratio for St. Mary’s could also be represented as being 97, i.e. below the national average.
- The Island has a growing number of people over 65 (25% by 2012 making demographics more typical of those of England predicted in 2048) and there is a very large prison population (1.25%). Older people have more co-morbidities – the complicating factors of multiple ill-health – but these were not taken into account by Dr Foster, because of our previous coding practice.
- Palliative and end of life care were not being recorded within the data submitted to the Department of Health. A new system has been put in place to ensure this happens.
Image: davidbally under a CC BY 2.0 license
