Article Text

Download PDFPDF

Comparison of control charts for monitoring clinical performance using binary data
  1. Jenny Neuburger1,2,
  2. Kate Walker2,3,
  3. Chris Sherlaw-Johnson1,
  4. Jan van der Meulen2,3,
  5. David A Cromwell2,3
  1. 1 The Nuffield Trust, London, UK
  2. 2 Department of Health Services Research and Policy, London School of Hygiene and Tropical Medicine, London, Greater London, UK
  3. 3 Clinical Effectiveness Unit, The Royal College of Surgeons of England, London, UK
  1. Correspondence to Dr Kate Walker, Department of Health Services Research and Policy, London School of Hygiene and Tropical Medicine, 15-17 Tavistock Place, London WC1H 9SH, UK; kate.walker{at}lshtm.ac.uk

Abstract

Background Time series charts are increasingly used by clinical teams to monitor their performance, but statistical control charts are not widely used, partly due to uncertainty about which chart to use. Although there is a large literature on methods, there are few systematic comparisons of charts for detecting changes in rates of binary clinical performance data.

Methods We compared four control charts for binary data: the Shewhart p-chart; the exponentially weighted moving average (EWMA) chart; the cumulative sum (CUSUM) chart; and the g-chart. Charts were set up to have the same long-term false signal rate. Chart performance was then judged according to the expected number of patients treated until a change in rate was detected.

Results For large absolute increases in rates (>10%), the Shewhart p-chart and EWMA both had good performance, although not quite as good as the CUSUM. For small absolute increases (<10%), the CUSUM detected changes more rapidly. The g-chart is designed to efficiently detect decreases in low event rates, but it again had less good performance than the CUSUM.

Implications The Shewhart p-chart is the simplest chart to implement and interpret, and performs well for detecting large changes, which may be useful for monitoring processes of care. The g-chart is a useful complement for determining the success of initiatives to reduce low-event rates (eg, adverse events). The CUSUM may be particularly useful for faster detection of problems with patient safety leading to increases in adverse event rates.  

  • control charts
  • run charts
  • statistical process control
  • quality measurement

This is an Open Access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/

Statistics from Altmetric.com

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.

Footnotes

  • Contributors JN and KW are joint first authors. They conceived the idea for the study, conducted the literature search, and synthesised the findings. JN, KW, JvdM and DAC developed the idea for the study. DAC advised on the literature search. JN and KW carried out the analyses. JN and KW drafted and revised the paper. All authors contributed to revising the paper and all reviewed and approved the final draft.

  • Funding JN was funded by a National Institute for Health Research Postdoctoral Fellowship (PDF-2013-06-078). The views expressed in this publication are those of the authors and not necessarily those of the NHS, the National Institute for Health Research or the Department of Health.

  • Competing interests None declared.

  • Patient consent The study involves secondary analysis of anonymised data.

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