Integrating patient movement and pathogen genomics to support hospital infection prevention with PathoPath: a method development study
Discuss this preprint
Start a discussion What are Sciety discussions?Listed in
This article is not in any list yet, why not save it to one of your lists.Abstract
Background
Healthcare-associated infections pose a major burden to neonatal health worldwide and remain difficult to track in low-resource hospitals because patient movement data and pathogen genomic data are rarely integrated into actionable transmission models. Existing approaches are often restricted to specific settings, highly structured electronic health records (EHRs), or analyses focused on either patient movements or pathogen characteristics alone. To address this gap, we developed PathoPath, an open-source integrative modelling platform, and evaluated its utility in a high burden paediatric hospital in Dhaka, Bangladesh.
Methods
PathoPath is an open-source R package that combines electronic health records with whole genome sequencing data to generate contact networks from direct and indirect contacts using minimal structured inputs. We retrospectively applied PathoPath to 373 cases of Klebsiella pneumoniae species complex (KpSC) infection identified in 2021 at the largest paediatric referral hospital in Dhaka, Bangladesh. Ward level patient movement trajectories were used to reconstruct contact networks, and genomic data from isolates from children <60 days were integrated to identify probable dissemination of bacterial clones and antimicrobial resistance plasmids.
Findings
PathoPath identified 750 direct contacts among 317 patients, forming 25 connected components, with the largest including 93 patients. KpSC infections were identified across 21 of 37 wards, with the neonatal intensive care unit accounting for 77.9% of all cases. Integration of genomic and network data distinguished sustained clustering of ST147 from multiple probable inter-clonal dissemination events involving IncFII plasmids carrying blaNDM-5 and/or blaOXA-181 within ST16. Four dominant sequence types accounted for 65.6% of sequenced isolates, and carbapenemase genes were detected in 95.8%.
Interpretation
PathoPath reconstructs hospital-wide contact networks and integrates them with pathogen genomics to map probable dissemination of pathogens and antimicrobial resistance using minimal structured clinical data. It could support more targeted infection prevention and control in hospitals where granular digital records are not available.
Funding
Gates Foundation, National Institute of Health Research, Wellcome Trust, and the David Price Evans Endowment.
Research in context
Evidence before this study
We searched PubMed ( pubmed.ncbi.nlm.nih.gov ) and Semantic Scholar ( www.semanticscholar.org ) databases in May 2025 for literature about existing methods for modelling healthcare-associated transmission of infectious diseases. We identified at least 3 Bayesian, 1 SEIR (Susceptible, Exposed, Infectious, Recovered) and 33 other models/approaches such as StEP and outbreaker2 to support these efforts. 19/37 of these methods have demonstrated their capability to reconstruct transmission or detect outbreak with varying level of accuracy depending on the context and pathogen. However, most of these tools frequently rely on dense patient metadata (e.g., Radio Frequency Identification sensors), specific types of electronic healthcare records that might not be readily available across settings, or a specific spatial structure. Also, most of these tools either focus on patient records or genomic data so there was a need for a generalised, integrative framework that can bridge the gap between clinical records with minimal digital input and pathogen genomic information.
Added value of this study
We developed PathoPath, an open-access integrative pathogen agnostic method that uses minimal electronic healthcare data and genome information to construct transmission pathways of pathogen and antimicrobial resistance within a healthcare facility. We have tested this package using 373 cases of Klebsiella pneumoniae species complex (KpSC) infection at Bangladesh Shishu Hospital and Institute (BSHI), the largest tertiary paediatric healthcare facility in Dhaka which serves as a referral centre in Bangladesh for the sick paediatric population. Using PathoPath, we successfully identified sustained K. pneumoniae ST147 clonal transmission and a few events of potential dissemination of carbapenemase-encoding plasmids among K. pneumoniae ST16, providing us a better understanding on how resistance is moving through the hospital network.
Implications of all the available evidence
PathoPath provides a scalable approach to generate high resolution data to inform infection prevention and control in settings that still lack well-defined electronic healthcare records. This could provide a greater level of understanding of transmission than that offered by standard practice and enable the implementation of targeted, cost-effective interventions to prevent further outbreaks or spread of antimicrobial resistance within hospitals.
