Potential causal effect of contact precautions and isolation on Clostridioides difficile infection in the hyperendemic setting: Interrupted time-series analyses before and after implementation

Abstract

Background: Recent studies disputed the effectiveness of efforts to comply with contact precautions and isolation (CPI) considering relatively low intra-hospital transmission rate of healthcare facility-associated Clostridioides difficile infection (HCFA-CDI). We evaluated the potential causal effect of CPI on HCFA-CDI occurrence by comparing the incidence rate (IR) for different time periods with and without CPI implementation.Methods: Long-term observational time-series data were separated into three periods (pre -CPI: January 2012 -March 2016, CPI: April 2016 -April 2021, post-CPI: May 2021 -December 2022). CPI was suspended owing to the restriction of isolation rooms during the COVID-19 pandemic. We inferred potential causal outcomes by comparing predicted and observed IRs of HCFA-CDI using interrupted time-series analyses, including the Bayesian structural time -series or autoregressive integrated moving average (ARIMA) model in the R-language or SAS software.Results: The monthly observed IR (44.9/100,000 inpatient-days) during the CPI period was significantly lower than the predicted IR (90.8) (-50.6% relative effect, P Z 0.001). However, the observed IR (52.3) during the post-CPI period was significantly higher than the predicted IR (39.1) (33.6%, P Z 0.001). The HCFA-CDI IR decreased during CPI (-14.3, P < 0.001) and increased post-CPI (5.4, P < 0.001) in the multivariable ARIMA model, which controlled for antibiotic usage, handwashing with soap and water, and number of toxin tests.Conclusions: Various time-series models revealed that CPI implementation had a potential causal effect on the reduction of HCFA-CDI incidence.Copyright (c) 2023, Taiwan Society of Microbiology. Published by Elsevier Taiwan LLC. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).