Hostname: page-component-7b7567c8c-z4zkx Total loading time: 0 Render date: 2026-11-16T21:58:48.109Z Has data issue: false hasContentIssue false

Clostridioides difficile surveillance: 9-year comparison between automated surveillance and conventional surveillance in acute care hospitals

Published online by Cambridge University Press:  24 February 2025

Jean Stanciu
Affiliation:
Medical Microbiology and Infectious Diseases, Centre Intégré de Santé et de Services Sociaux du Bas-Saint-Laurent, Rimouski, QC, Canada
Patrick Dolcé
Affiliation:
Medical Microbiology and Infectious Diseases, Centre Intégré de Santé et de Services Sociaux du Bas-Saint-Laurent, Rimouski, QC, Canada
Charles Frenette
Affiliation:
Medical Microbiology and Infectious Diseases, McGill University Health Centre, Montreal, QC, Canada
Marie-Claude Roy
Affiliation:
Medical Microbiology and Infectious Diseases, CHU de Québec-Université Laval, Quebec City, QC, Canada
Lina Kouider
Affiliation:
Medical Microbiology and Infectious Diseases, Centre Intégré de Santé et de Services Sociaux du Bas-Saint-Laurent, Rimouski, QC, Canada
Yves Longtin*
Affiliation:
Medical Microbiology and Infectious Diseases, Infection Prevention and Control Unit, Jewish General Hospital, Montreal, QC, Canada
*
Corresponding author: Yves Longtin; Email: Yves.longtin@mcgill.ca

Abstract

Objective:

To develop and validate an automated surveillance system for healthcare-associated Clostridioides difficile infections (HA-CDI).

Design:

Multicenter cohort study.

Setting:

16 acute care hospitals.

Patients:

Patients admitted to participating hospitals between 2013 and 2022.

Methods:

An automated surveillance system was developed with retrospective extraction from admission/discharge/transfer and laboratory databases and compared with conventional surveillance based on clinical definitions collected prospectively by infection control professionals. Comparison of HA-CDI incidence rates calculated by automated vs conventional surveillances were performed with χ2, incidence rate ratios, and linear regression. A subset of discordant cases was further investigated by reviewing medical records.

Results:

Overall, conventional surveillance reported 3,211 cases of HA-CDI for an incidence rate of 4.94 per 10,000 patient-days. Automated surveillance detected 4,708 cases, for an incidence rate of 7.24 per 10,000 patient-days (incidence rate ratio, 1.47; 95% CI, 1.40–1.53). Full concordance between both surveillance methods was observed in 62% of cases, while 34% of cases were detected only by automated surveillance, and 4% were detected by conventional surveillance only. Between 2013 and 2022, an identical declining trend in HA-CDI incidence rates of –0.54 cases per 10,000 patient-days was observed with both surveillance methods. A subset of 49 cases detected only by automated surveillance were reviewed; the main reasons for discrepancy were delayed testing (39%), colonization (24%), misclassifications (14%), and interinstitutional transfers (12%).

Conclusions:

HA-CDI incidence rates calculated by automated surveillance were higher than those of conventional surveillance, but the overestimation was consistent over time, suggesting that a correction factor could improve precision.

Information

Type
Original Article
Creative Commons
Creative Common License - CCCreative Common License - BY
This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted re-use, distribution and reproduction, provided the original article is properly cited.
Copyright
© The Author(s), 2025. Published by Cambridge University Press on behalf of The Society for Healthcare Epidemiology of America
Figure 0

Figure 1.The automated surveillance algorithm for categorization of positive C. difficile tests consists of one starting point: a recognized positive C. difficile laboratory test. Next are exclusions: for tests that are not recognized, incomplete data, tests not linked to the declaring institution, duplicates (other positive test <2 weeks), recurrences (other positive test <10 weeks), patient in a unit excluded from surveillance, patient that was not hospitalized. If a sample was collected ≥3 calendar days after admission of <4 weeks after discharge, the case was considered healthcare-associated. Remaining tests were considered community-associated.

Figure 1

Figure 2.Linear graph illustrating the annual aggregated healthcare-associated Clostridioides difficile infection (HA-CDI) incidence rates calculated by automated vs conventional surveillance methods among 16 hospitals in Quebec, Canada over a 9-year period. Linear trendlines to detect secular trends in HA-CDI incidence are also shown with dotted lines for both methods.

Figure 2

Figure 3.Incidence rate ratio (IRR) and 95% confidence intervals (CI) for healthcare-associated Clostridioides difficile infections (HA-CDI) incidence rate calculated by automated vs conventional surveillance. The overall IRR and 95% CI is illustrated by the horizontal bars. The sample size of each institution is illustrated by the size of the dot.

Figure 3

Figure 4.Sankay diagram illustrating the change in overall ranking of participating institutions when healthcare-associated Clostridioides difficile infections (HA-CDI) incidence rates are computed by automated versus conventional surveillances.

Figure 4

Figure 5.Scatter plot illustrating yearly healthcare-associated Clostridioides difficile infections (HA-CDI) incidence rates calculated by automated vs conventional surveillance methods for each group of hospitals among 16 hospitals in Quebec, Canada over a 9-year period. Linear regression yielded a linear trendline equation of y = 1.12x + 1.46, where y represents HA-CDI incidence rates calculated by automated surveillance and x represents HA-CDI incidence rates calculated by conventional surveillance. Coefficient of determination (R2) = 0.91.

Figure 5

Figure 6.Linear graph illustrating the annual aggregated healthcare-associated Clostridioides difficile infection (HA-CDI) incidence rates calculated by conventional surveillance, automated surveillance, and automated surveillance adjusted with the correction factor among 16 hospitals in Quebec, Canada over a 9-year period.