Article Text

Download PDFPDF

Investigating the association of alerts from a national mortality surveillance system with subsequent hospital mortality in England: an interrupted time series analysis
  1. Elizabeth Cecil1,
  2. Alex Bottle1,
  3. Aneez Esmail2,
  4. Samantha Wilkinson3,
  5. Charles Vincent4,
  6. Paul P Aylin1
  1. 1 Primary Care and Public Health, Imperial College London, London, UK
  2. 2 Health Services Research & Primary Care, University of Manchester, Manchester, UK
  3. 3 London School of Hygiene & Tropical Medicine, London, UK
  4. 4 Medical Science Division, University of Oxford, London, Oxfordshire, UK
  1. Correspondence to Dr. Elizabeth Cecil, Primary Care and Public Health, Imperial College London, London W6 8RP, UK; e.cecil{at}


Objective To investigate the association between alerts from a national hospital mortality surveillance system and subsequent trends in relative risk of mortality.

Background There is increasing interest in performance monitoring in the NHS. Since 2007, Imperial College London has generated monthly mortality alerts, based on statistical process control charts and using routinely collected hospital administrative data, for all English acute NHS hospital trusts. The impact of this system has not yet been studied.

Methods We investigated alerts sent to Acute National Health Service hospital trusts in England in 2011–2013. We examined risk-adjusted mortality (relative risk) for all monitored diagnosis and procedure groups at a hospital trust level for 12 months prior to an alert and 23 months post alert. We used an interrupted time series design with a 9-month lag to estimate a trend prior to a mortality alert and the change in trend after, using generalised estimating equations.

Results On average there was a 5% monthly increase in relative risk of mortality during the 12 months prior to an alert (95% CI 4% to 5%). Mortality risk fell, on average by 61% (95% CI 56% to 65%), during the 9-month period immediately following an alert, then levelled to a slow decline, reaching on average the level of expected mortality within 18 months of the alert.

Conclusions Our results suggest an association between an alert notification and a reduction in the risk of mortality, although with less lag time than expected. It is difficult to determine any causal association. A proportion of alerts may be triggered by random variation alone and subsequent falls could simply reflect regression to the mean. Findings could also indicate that some hospitals are monitoring their own mortality statistics or other performance information, taking action prior to alert notification.

  • health services research
  • statistical process control
  • healthcare quality improvement

This is an open access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See:

View Full Text

Statistics from


  • Contributors All authors contributed to the conception and design of this study. EC carried out the analysis. AB provided statistical advice. All authors took part in interpreting the data for this study. All authors commented on and helped to revise drafts of this paper. All authors have approved the final version.

  • Funding This study was funded by the National Institute for Health Research, Health Services and Delivery Research Programme (HS&DR - 12/178/22).

  • Competing interests All authors have completed the Unified Competing Interest form (available on request from the corresponding author), and PPA and AB declare that they are partially funded by grants from Dr Foster Intelligence, an independent healthcare information company. CV reports funding from the Health Foundation for Research and Haelo (a commercial innovation and improvement science organisation) for consultancy work. EC, SW and AE declare no support from any organisation for the submitted work; no financial relationships with any organisations that might have an interest in the submitted work in the previous 3 years; and no other relationships or activities that could appear to have influenced the submitted work.

  • Patient consent Not required.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data sharing statement No additional data available.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles