Leecaster M, Veterans Affairs Salt Lake City Health Care System, Salt Lake City, UT and Department of Internal Medicine, University of Utah School of Medicine; Khader K, Veterans Affairs Salt Lake City Health Care System, Salt Lake City, UT and Department of Internal Medicine, University of Utah School of Medicine; Nelson RE, Veterans Affairs Salt Lake City Health Care System, Salt Lake City, UT and Department of Internal Medicine, University of Utah School of Medicine; Ray W, Veterans Affairs Salt Lake City Health Care System, Salt Lake City, UT and Department of Internal Medicine, University of Utah School of Medicine; Stevens VW, Veterans Affairs Salt Lake City Health Care System, Salt Lake City, UT and Department of Pharmacotherapy, University of Utah College of Pharmacy; Toth D, Veterans Affairs Salt Lake City Health Care System, Salt Lake City, UT and Department of Internal Medicine, University of Utah School of Medicine; Rubin MA, Veterans Affairs Salt Lake City Health Care System, Salt Lake City, UT and Department of Internal Medicine, University of Utah School of Medicine;
Objectives:
Hand hygiene (HH) is an important control measure for Clostridium difficile (CD), a common nosocomial infection in the United States. We sought to assess the effect of varying hand hygiene adherence (HHA) on Clostridium difficile acquisition (CDA) in an acute care setting.
Methods:
We designed an agent-based simulation of nosocomial CD transmission with static and dynamic components including: patients, healthcare workers (HCW), and rooms; patient admission, discharge, and transfer; contact between HCW and patients; contamination of rooms by patients shedding CD; HCW hand carriage and removal via HH or prevention via personal protective equipment; and acquisition of CD by patients following contact with contaminated rooms or HCW. Model parameters were derived from local data, literature, and expert opinion. The model was calibrated against local data and validated internally and externally. Five scenarios with varying HHA were simulated 400 times each over a one-year period. The base-case scenario had HHA levels from literature. Two scenarios varied the probability of HHA by provider (nurse/doctor) between 0 and 1, while two scenarios varied the probability of nurse HHA by entry and exit between 0 and 1. CDA rate was calculated as the number of nosocomial CDA per 1000 patient days (/1000PD).
Results:
The effect of varying HHA on CDA was greater for nurses than doctors. CDA at the base-case was 12.6/1000PD. Nurse HHA below base-case increased CDA up to 90/1000PD while variation in doctor HHA produced minimal change in CDA. The effect of varying nurse HHA was almost completely attributed to room exit. Nurse HHA below base-case on exit resulted in CDA up to 90/1000PD while on entry resulted in CDA up to 20/1000PD. The effect of nurse HHA above the base-case for both entry and exit reduced CDA to 6/1000PD.
Implications:
Our results suggest that HH on room exit may be most important for preventing CD transmission, particularly among nurses. Furthermore, improvements in already high HHA have smaller relative impact on CDA than improvements in low HHA.
Impacts:
This novel use of agent-based simulation for decision-making in healthcare can provide key guidance on efforts to reduce transmission of CD through interventions that increase HHA among HCW.