CUSUM: A tool for early feedback about performance?

Department of Trauma & Orthopaedics, Dumfries and Galloway Royal Infirmary, Dumfries, DG1 4AP, UK.
BMC Medical Research Methodology (Impact Factor: 2.27). 02/2006; 6(1):8. DOI: 10.1186/1471-2288-6-8
Source: PubMed


Modern day clinical practice demands evidence justifying our choice of treatment methods. Cumulative sum techniques (cusum) are amongst the simplest statistical methods known. They provide rapid analysis and identification of trends in a series of data. This study highlights use of these techniques as an early performance indicator of a clinical procedure before its implementation.
Twenty consecutive patients who underwent total hip or knee arthroplasty received a simple dressing--blue gauze and Tegaderm. Cusum charting was used to assess the dressing with regards to skin blistering. At an acceptable level of performance the curve would oscillate about the horizontal axis and the overall trend therefore said to be flat. If performance is unacceptable, the cusum slopes upward.
The cusum plot for the twenty patients did not cross the specified control limits. This showed that our simple dressing met specified standards with regards to wound blistering postoperatively.
We recommend the use of this simple, yet versatile cusum technique in the early evaluation of a clinical procedure before its implementation.

Download full-text


Available from: Winston R Chang, Apr 12, 2014
1 Follower
14 Reads
  • Source
    • "One way to improve the false positive rates of algorithms is to increase the threshold limits. However, by raising the threshold, it generally takes longer to detect an outbreak [16]. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Detection of outbreaks is an important part of disease surveillance. Although many algorithms have been designed for detecting outbreaks, few have been specifically assessed against diseases that have distinct seasonal incidence patterns, such as those caused by vector-borne pathogens. We applied five previously reported outbreak detection algorithms to Ross River virus (RRV) disease data (1991-2007) for the four local government areas (LGAs) of Brisbane, Emerald, Redland and Townsville in Queensland, Australia. The methods used were the Early Aberration Reporting System (EARS) C1, C2 and C3 methods, negative binomial cusum (NBC), historical limits method (HLM), Poisson outbreak detection (POD) method and the purely temporal SaTScan analysis. Seasonally-adjusted variants of the NBC and SaTScan methods were developed. Some of the algorithms were applied using a range of parameter values, resulting in 17 variants of the five algorithms. The 9,188 RRV disease notifications that occurred in the four selected regions over the study period showed marked seasonality, which adversely affected the performance of some of the outbreak detection algorithms. Most of the methods examined were able to detect the same major events. The exception was the seasonally-adjusted NBC methods that detected an excess of short signals. The NBC, POD and temporal SaTScan algorithms were the only methods that consistently had high true positive rates and low false positive and false negative rates across the four study areas. The timeliness of outbreak signals generated by each method was also compared but there was no consistency across outbreaks and LGAs. This study has highlighted several issues associated with applying outbreak detection algorithms to seasonal disease data. In lieu of a true gold standard, a quantitative comparison is difficult and caution should be taken when interpreting the true positives, false positives, sensitivity and specificity.
    BMC Medical Informatics and Decision Making 11/2010; 10(1):74. DOI:10.1186/1472-6947-10-74 · 1.83 Impact Factor
  • Source
    • "Manly 2001), biomedicine (e.g. Chang and McLean 2006), clinic tests, public health (e.g. Starks and Flatman 1991). "
    [Show abstract] [Hide abstract]
    ABSTRACT: Statistical Process Control (SPC) methods are extensively used in manufacturing contexts to monitor production processes. This article illustrates their potential utility for monitoring the state of marine ecosystems, using survey indicators. We exemplify the use of one SPC tool, the cumulative sum (CUSUM) control chart, to detect persistent changes in the state of a system as new observations are collected, using simulated and real data. Practical guidelines are given on how the chart parameters can be tuned to achieve an acceptable compromise between the ability to detect anomalies quickly and keeping the risk of false alarms low. The common performance measures associated with control charts depend on some key assumptions being met, and the potential impact of their violation is indicated.
    Aquatic Living Resources 04/2009; 22(2). DOI:10.1051/alr/2008058 · 1.01 Impact Factor
  • Source
    • "At an acceptable level of performance, the cusum curve runs randomly at or above a horizontal line (no slope). However, when performance is at an unacceptable level, the cusum slope changes [14]. Cusum is calculated using the formula Sn ¼ SðX 0 2X i Þ, where X i ¼ 1 for a success, X i ¼ 0 for a failure, and X 0 represents for a predetermined acceptable level of performance. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Perforation of colorectal cancer (CRC) is rare and is associated with a significantly high mortality and morbidity. The aim of the current study was to evaluate various factors influencing the outcome in these patients. A retrospective analysis of 42 patients with perforated CRC between 1999 and 2003 was performed. A number of variables including age, sex, site of perforation, presence of faecal peritonitis, grade of surgeon, presence of metastasis, stage of tumour, type of surgery, ASA grade and CR POSSUM score were analysed for their influence on the outcome in these patients using MS Excel, MS Access and Stata. Of the 42 patients 19 were female and 23 were male. The mean age of the patients was 70.5 (range 44-96yr). Thirty patients had perforation at the tumour, 10 proximal to the tumour, and one distal to the primary tumour. The perforation was localised in 25 patients. However, 17 patients had free perforation with frank faecal peritonitis. Twenty-one patients had resection and anastomosis, 18 patients had resection without restoration of bowel continuity and 3 had palliative colostomy. The in-hospital mortality (within 30d) was 40.5% (n=17) with only 15 patients being alive at the end of 2yr with an overall mortality of 64.3% (n=27). The outcome was not altered by variables such as sex, surgeon's grade, surgical procedure, Dukes' staging or the site of perforation (p>0.5). Univariate analysis showed that advanced age (p<0.01), higher ASA grade (p<0.001), higher CR POSSUM score (p<0.001) and degree of peritonitis (p<0.01) were strongly associated with adverse outcomes. However, in stepwise multivariate logistic regression analysis ASA grade (p=0.01) and CR POSSUM score (p=0.01) were the only significant predictors of in-hospital mortality. The outcome of perforated colonic cancer continues to be poor. ASA score and CR POSSUM score are good predictors of the short-term outcome.
    Surgical Oncology 09/2006; 15(2):91-6. DOI:10.1016/j.suronc.2006.09.001 · 3.27 Impact Factor
Show more