Use of optimised dual statistical process control charts for early detection of surgical site infection outbreaks

Early recognition and response to increases in surgical site infections using optimised statistical process control charts—The early 2RIS trial: A multicentre stepped wedge cluster randomised controlled trial

“…Our findings confirm the results of the retrospective analyses we performed in the same network of hospitals that led to the selection of optimised SPC methods used in this trial. 22 , 26 Our prior retrospective analysis estimated that optimised SPC would have 90% sensitivity and 56% positive predictive value in detecting clinically important increases in SSI rates for hospital and procedure combinations. 26 Prospective evaluation in this trial yielded similar performance characteristics.…”

“…Both charts utilised rolling baseline windows, control limits of ±1 standard deviation, and other chart characteristics as previously described. 22 SSI data for each procedure under surveillance were plotted with monthly resolution. Any data point above the rolling upper control limit on either chart indicated detection of a potentially important SSI rate increase and was considered an “SPC signal,” which required evaluation (appendix p 17).…”

“…SPC signals that occurred within 12 months before or after true positive traditional signals were considered to have detected the same important SSI rate increase. 22 Procedure months of surveillance were calculated by multiplying the sum of the number of procedure types under surveillance at each hospital by the duration of surveillance in months. For calculations of rates of SSI investigations, procedure months were subtracted for hospital/procedure combinations that were receiving active investigation and not capable of generating new SSI investigations.…”

“… 21 Therefore, infection prevention personnel often initiate SSI outbreak investigations months after SSI rates first increase. 22 …”

“… 21 , 26 When applied retrospectively to 30 previously investigated SSI outbreaks, this chart combination identified each outbreak, typically several months earlier than traditional surveillance. 22 We therefore hypothesised that prospective use of the same optimised SPC approach in addition to traditional SSI surveillance would decrease rates of SSI compared to use of traditional surveillance methods alone.…”