‘Online’ monitoring and retrospective analysis of hospital outcomes based on a scan statistic

Abstract

Statistical tests based on the scan statistic are introduced for detecting possible increases in the occurrence of hospital events. The tests use a moving window and the theoretical aspects of the tests are investigated using Markov chain theory. The main objective of this study is to provide a statistical technique to assist hospital staff in deciding whether the variation they observe is greater than usually expected under random variation. In this paper we develop the test for Poisson data and apply the theory to monitor the occurrence of orthopaedic wound infection and Methicillin-resistant Staphylo- coccus aureus colonization. We find that this method is sensitive in detecting the change in the process parameter which may not be detected by standard control chart methods. Both online and retrospective analyses are considered.

DeCon: A tool to detect emotional concordance in multivariate time series data of emotional responding

“…To this end, we build on the key idea of the Bernoulli scan statistic in change-point analysis (Ismail, Pettitt, & Webster, 2003) by computing a moving sum of the entries of the outlier vector. When this moving sum exceeds a threshold, evidence for change has been found.…”

On the use and Evaluation of Prospective Scan Methods for Health-Related Surveillance

“…Ismail et al. (2003) and Naus and Wallenstein (2006) proposed scan statistic methods for monitoring in the temporal–Bernoulli and temporal–exponential cases.…”

Modeling emergency department visit patterns for infectious disease complaints: results and application to disease surveillance

“…If any single-day forecast error exceeded a threshold, then the simulated outbreak was said to be detected. Sequential tests are ideal for multi-day outbreaks so the performance (the false negative rate for a fixed false positive rate) of Page's statistic or a moving window such as in [28] or the scan statistic such as in [29] would be better than the performance of one-at-a-time tests in the case where all simulated outbreaks lasted 7 days. Reis et al [28] applied several sliding detection windows, each of at most 7 days to ED visit daily counts in which simulated outbreaks (in the form of additional ED visits) lasting 3, 7, and 14 days were added to the real data in a simulation study.…”

“…In summary, we concur with [30] regarding the robustness and simplicity of Page's test. Alternatively, there are occasions when using a modest number of specific tests is effective as was done in [29]. …”