Prevention of Surgical Site Infections Using Statistical Process Control Charts
Surgical site infections (SSIs) are the most common and costly healthcare-associated infections (HAI) in the US. Therefore, innovative strategies to prevent SSI are greatly needed. Traditional or “standard” surveillance monitors rates of SSI following common surgical procedures, often calculated on a quarterly or biannual basis. Therefore, changes in SSI rates often are detected several months after the rate first changed, if at all.
Statistical process control (SPC) is an analytic approach that combines time series analysis methods with graphical presentation of data to determine whether a process or rate exhibits “common cause” natural variation or “special cause” unnatural variation due to circumstances that have not previously been inherent in the process. To date, however, SPC methods are not commonly utilized in a rigorous manner to provide real-time surveillance of HAIs such as SSIs. Similarly, the best type and design of SPC methods that would most quickly identify changes in rates of SSI while minimizing false positive alerts is unknown. The overall objective of this proposal is to determine the clinical effectiveness of SPC methods to prevent SSI.
Specific Aims :
Aim 1: To determine which SPC methods best identify increases in rates of SSI compared to standard surveillance methods.
Aim 2: To measure the effectiveness of surveillance using optimized SPC methods and feedback on rates of SSI compared to standard surveillance and feedback.
Duke Infection Control Outreach Network (DICON): The Duke Infection Control Outreach Network (DICON) has been helping community hospitals and surgery centers address these issues for over 10 years. DICON provides sophisticated data analysis and metrics, access to experts in infection control, opportunities to share successful programs, and extensive educational initiatives related to infection prevention.
This research aims to evaluate the effectiveness of numerous, rigorous statistical methods, mainly statistical process control as well as SCAN, change point, and minimal spanning tree methods, to provide real-time surveillance and detect increases in surgical site infection rates compared to standard surveillance methods.
Results to Date:
The research team analyzed 12 years of data on 13 surgical procedures from a network of 58 community hospitals and evaluated the performance of numerous statistical surveillance methods in terms of earlier detection of increases in SSIs compared to traditional surveillance. Statistically significant increases in SSI rates (signals) were identified using over 50 different SPC chart variations and several SCAN, change point, and minimal spanning tree analyses. Epidemiologists evaluated the clinical significance of the signals generated and rated them based on clinical severity and level of action needed. These ratings were used to identify which SPC, SCAN, change point, and MST approaches maximized sensitivity and specificity. Optimized approaches were implemented and evaluated within dicon hospitals using a randomized control trial (RCT) approach. The results of the RCT show the effectiveness of using optimal SPC chart for real-time surveillance in terms of prevented HAIs, hospitalizations, mortalities, and potential cost savings. These approaches were also applied to other hospital-acquired infections such as c. difficile infections and central-line associated blood stream infections.
Baker AW, Haridy S, Salem J, et al. Performance of statistical process control methods for regional surgical site infection surveillance: a 10-year multicentre pilot study. BMJ Qual Saf 2018;27:600–10.
Ilieş I, Anderson DJ, Salem J, et al. Large-scale empirical optimisation of statistical control charts to detect clinically relevant increases in surgical site infection rates. BMJ Qual Saf 2020;29:472–481. doi:10.1136/bmjqs-2018-008976.
Benneyan JC. Statistical quality control methods in infection control and hospital epidemiology, part I: introduction and basic theory. Infect Control Hosp Epidemiol 1998;19:194–214
Benneyan JC. Statistical quality control methods in infection control and hospital epidemiology, part II: chart use, statistical properties, and research issues. Infect Control Hosp Epidemiol 1998;19:265–83.
Ilieş, I., Benneyan, J., Jabur, T., Baker, A., & Anderson, D. (2020). Impact of molecular testing on reported Clostridoides difficile infection rates. Infection Control & Hospital Epidemiology, 41(3), 306-312. doi:10.1017/ice.2019.327.
Nehls N, Holler N, Ilies I, Anderson D, Jacobsen M, Baker A, Benneyan J. Investigating Statistical Process Control for Surveillance of C. difficile Infections. Poster Presented at HSPI Conference; February 21-23, 2018; Atlanta, GA. https://www.iise.org/HSPI/
Nehls N, Holler N, Ilies I, Anderson D, Jacobson M, Baker A, Benneyan J. Performance of Statistical Process Control Charts for Detecting Clinically-Significant Adverse Event Outbreak Patterns. Poster Presented at IHI National Forum; December 10-13, 2017. Orlando, FL.
Nehls N, Ilies I, Benneyan J, Baker A, Anderson D. Potential Health and Cost Outcomes of Optimized Statistical Process Control Use for Surgical Site Infection Surveillance. Poster presented at IDWeek; October 2-6, 2019; Washington, D.C. http://www.idweek.org
Bilantuono A, Etherton C, Nehls N, Ilies I, Benneyan J, Baker A, Anderson D. Performance of Statistical Process Control Charts for Detecting Clinically-Significant Increases in Clostridium difficile Infection Rates. Poster presented at IDWeek; October 2-6, 2019; Washington, D.C. http://www.idweek.org
Baker A, Nehls N, Ilies I, Benneyan J, Anderson D. Use of Dual Statistical Process Control Charts for Early Detection of Surgical Site Infection Outbreaks at a Community Hospital Network. Talk presented at IDWeek; October 2-6, 2019; Washington, D.C. http://www.idweek.org