Evidence for patterns of colonization risk in the nationwide hospitalized patient network in France
Abstract
Patient movement within and between hospitals creates a complex hospital network that facilitates the spread of antimicrobial-resistant bacteria and contributes to healthcare-associated infections. Yet most national-scale studies infer spread from aggregated hospital-to-hospital patient sharing, leaving patient-to-patient contact pathways and the role of routine patient mobility in shaping colonization risk poorly characterized. Using Frances 2023-2024 national hospital-discharge data, we reconstructed a temporal patient-based network of 10.4 million patients across 1,408 acute-care hospitals, comprising 326.6 million overnight ward co-location interactions. We simulated colonization over 365 days on the observed day-by-day temporal network using a stochastic susceptible-colonized-susceptible model, with a mean colonization duration of 180 days. We used national carbapenemase-producing Enterobacterales (CPE) surveillance data to calibrate differences in transmission risk across clinical settings. Across 2,000 simulations initiated on randomly selected dates in 2023, we calculated patient-level colonization probabilities and summarized risk across patients residential areas, wards, and hospitals. We further assessed hospital-level consistency between simulated risk and recorded CPE occurrence. Simulated risk was highly heterogeneous across patients and care settings. When mapped by place of residence, hospital-associated colonization risk showed strong spatial clustering in both urban and rural areas. Multivariable regression analyses identified patient- and hospital-level characteristics associated with heterogeneity in simulated colonization risk. Sensitivity analyses varying initial seeding and colonization duration showed that relative risk patterns remained highly stable. This nationwide framework shows how patient mobility and ward co-location patterns can reveal patterns of hospital-associated colonization risk and could support pathogen-specific surveillance and intervention evaluation.
Significance StatementAsymptomatic carriage allows antimicrobial-resistant bacteria to spread through healthcare systems while colonized patients may remain unrecognized in the absence of active screening. We reconstructed a nationwide, time-resolved network of ward co-location among hospitalized patients in France. We accounted for differences in transmission risk across clinical settings using national surveillance data on carbapenemase-producing Enterobacterales. Using a transmission model, we characterized colonization-risk patterns and identified patient- and hospital-level features associated with variation in risk. This framework offers health authorities a way to examine where patient movement and hospital care may contribute to concentrations of colonization risk, although pathogen- and context-specific adaptation and evaluation remain necessary to guide surveillance or interventions.
Author contributionsS.B., L.C., L.T., A.K., T.D., V.C., and P.C. designed research; S.B. and P.C. performed research; S.B. and P.C. analyzed data; S.B. and P.C. wrote the paper; all authors reviewed and edited the final manuscript.
Related
- Epidemiology of hospitalized strongyloidiasis in Japan: A nationwide administrative database study
- Automated compartment-based quantification reveals distinct patterns of cutaneous innervation with aging and microbial colonization
- Hospitalized Infection Risk and Organ-specific Infection Patterns Across Advanced Therapies in Ulcerative Colitis: A Japanese Claims Analysis
- Glucocorticoid in giant cell arteritis: real-world treatment patterns over time and associated burden in a nationwide hospital-based study within the French health insurance database
- Socioeconomic deprivation and post-discharge mortality in heart failure: evidence from a nationwide universal health coverage system
BibTeX
@misc{bhuria2026evidence,
title = {{Evidence for patterns of colonization risk in the nationwide hospitalized patient network in France}},
author = {Shrichand Bhuria and Lucille Calmon and Madhav Chaturvedi and Laura Temime and Andre Karch and Tjibbe Donker and Vittoria Colizza and Pascal Crepey},
howpublished = {HAL},
year = {2026},
month = sep,
doi = {10.64898/2026.09.17.26363228},
url = {https://hal.science/hal-05781294v1},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Evidence for patterns of colonization risk in the nationwide hospitalized patient network in France
AU - Shrichand Bhuria
AU - Lucille Calmon
AU - Madhav Chaturvedi
AU - Laura Temime
AU - Andre Karch
AU - Tjibbe Donker
AU - Vittoria Colizza
AU - Pascal Crepey
PY - 2026
DA - 2026/09/17
PB - HAL
DO - 10.64898/2026.09.17.26363228
UR - https://hal.science/hal-05781294v1
AB - Patient movement within and between hospitals creates a complex hospital network that facilitates the spread of antimicrobial-resistant bacteria and contributes to healthcare-associated infections. Yet most national-scale studies infer spread from aggregated hospital-to-hospital patient sharing, leaving patient-to-patient contact pathways and the role of routine patient mobility in shaping colonization risk poorly characterized. Using Frances 2023-2024 national hospital-discharge data, we reconstructed a temporal patient-based network of 10.4 million patients across 1,408 acute-care hospitals, comprising 326.6 million overnight ward co-location interactions. We simulated colonization over 365 days on the observed day-by-day temporal network using a stochastic susceptible-colonized-susceptible model, with a mean colonization duration of 180 days. We used national carbapenemase-producing Enterobacterales (CPE) surveillance data to calibrate differences in transmission risk across clinical settings. Across 2,000 simulations initiated on randomly selected dates in 2023, we calculated patient-level colonization probabilities and summarized risk across patients residential areas, wards, and hospitals. We further assessed hospital-level consistency between simulated risk and recorded CPE occurrence. Simulated risk was highly heterogeneous across patients and care settings. When mapped by place of residence, hospital-associated colonization risk showed strong spatial clustering in both urban and rural areas. Multivariable regression analyses identified patient- and hospital-level characteristics associated with heterogeneity in simulated colonization risk. Sensitivity analyses varying initial seeding and colonization duration showed that relative risk patterns remained highly stable. This nationwide framework shows how patient mobility and ward co-location patterns can reveal patterns of hospital-associated colonization risk and could support pathogen-specific surveillance and intervention evaluation. Significance StatementAsymptomatic carriage allows antimicrobial-resistant bacteria to spread through healthcare systems while colonized patients may remain unrecognized in the absence of active screening. We reconstructed a nationwide, time-resolved network of ward co-location among hospitalized patients in France. We accounted for differences in transmission risk across clinical settings using national surveillance data on carbapenemase-producing Enterobacterales. Using a transmission model, we characterized colonization-risk patterns and identified patient- and hospital-level features associated with variation in risk. This framework offers health authorities a way to examine where patient movement and hospital care may contribute to concentrations of colonization risk, although pathogen- and context-specific adaptation and evaluation remain necessary to guide surveillance or interventions. Author contributionsS.B., L.C., L.T., A.K., T.D., V.C., and P.C. designed research; S.B. and P.C. performed research; S.B. and P.C. analyzed data; S.B. and P.C. wrote the paper; all authors reviewed and edited the final manuscript.
ER -