Multi-modal Approach for Predicting Infection Routes in Nursing Homes

Leveraging artificial intelligence and wearable proximity sensors to uncover hidden patterns in infection transmission, enabling targeted interventions to protect vulnerable populations in long-term care facilities.

Patients
Healthcare Workers
MDRO Carriers
Introduction

The Challenge of MDROs in Healthcare Settings

Understanding why innovative approaches are critical for infection control in nursing homes

The growing threat of Multidrug-Resistant Organisms (MDROs) in healthcare settings, particularly in nursing homes and long-term care facilities, has become a critical public health concern. The vulnerability of older adults to infections, combined with the challenges of communal living and frequent healthcare interactions, creates conditions highly favorable for the spread of MDROs.

This issue is especially pressing in Arizona, where a large elderly population resides in such facilities. Nursing homes in the state have reported higher rates of MDRO infections compared to the national average, with some facilities experiencing outbreaks that have led to significant morbidity and mortality among residents.

The urgent need for innovative solutions has led to the exploration of multi-modal approaches that leverage artificial intelligence and advanced technologies to predict and prevent infection routes. By harnessing the power of these tools, we aim to uncover hidden patterns in infection transmission that have previously eluded traditional methods of surveillance and control.

Key Innovations

  • High-resolution proximity sensors for tracking close-range interactions
  • AI-driven analysis to identify potential transmission events
  • Synthetic dataset generation for privacy-preserving research
  • Early warning system prototype for county health departments
  • Targeted interventions to reduce unnecessary antibiotic prescriptions
Goals

Research Objectives

A comprehensive approach to understanding and preventing MDRO transmission

I

Illuminate Infection Transmission Pathways

Elucidate the complex network of infection routes within nursing homes by integrating high-resolution data from proximity sensors and electronic health records, investigating associations between disease spectrum, seroconversion status, and social contact patterns.

II

Targeted Intervention Development

Utilize detailed understanding of infection transmission pathways to develop and implement targeted interventions, informed by granular sensor-based information. This will enable an early warning system prototype for county health department use.

III

Optimization of Antibiotic Prescribing

Leverage granular data to identify patterns of antibiotic overprescribing for conditions not warranting treatment, particularly non-UTIs and non-RTIs, thereby reducing the risk of antibiotic resistance development.

IV

Synthetic Dataset Creation

Generate synthetic datasets using Generative AI that accurately reflect complex infection transmission scenarios while preserving patient privacy, facilitating broader research efforts and contributing to health equity.

V

Alignment with Evolving Evidence Practices

Ensure study design and outcomes align with the current and future landscape of data collection, analysis, and evidence-based practice, incorporating the latest analytic methods and data collection technologies.

Approach

Methods & Technology

Advanced proximity sensing and data analysis for infection tracking

SocioPatterns Proximity Sensors

We employ SocioPatterns proximity sensors that use ultra-low-power radio communication in the ISM (Industrial, Scientific, Medical) band to measure close-range proximity of approximately less than 1.5 meters between individuals wearing them.

These sensors log time-resolved proximity relations in onboard non-volatile storage, from which the complete contact history can be extracted after devices are collected. The sensors have been successfully deployed in hospital settings to track close-range proximity between healthcare workers and patients.

The data collected can be used to identify potential sources of infection and track the spread of infectious diseases within the facility. Sensors are carried in a plastic case with a clip that attaches to name badges or clothing at chest height, ensuring consistent and reliable proximity detection.

Previous research using this approach has shown that healthcare workers have significantly more frequent and longer-duration contacts compared to patients, highlighting their potential role as transmission vectors—insights that are critical for developing targeted interventions.

SocioPatterns Sensor

ISM Band Radio Communication

<1.5m
Detection Range
20s
Time Resolution
ISM
Radio Band
NVM
Storage Type
Researchers

Research Team

An interdisciplinary team combining expertise in epidemiology, AI, and digital health

Principal Investigator Onicio B Leal Neto

Onicio B Leal Neto

PhD, MS

Faculty, Department of Epidemiology
College of Public Health
University of Arizona

Kate Ellingson

Kate Ellingson

PhD, MPH

Faculty, Department of Epidemiology
College of Public Health
University of Arizona

Paulina Colombo

Paulina Colombo

MPH

Graduate Research Associate
Department of Epidemiology
University of Arizona

Ciro Cattuto

Ciro Cattuto

PhD

Scientific Director
ISI Foundation
Turin, Italy

Rodrigo Paiva

Rodrigo Paiva

PhD

AI/ML Researcher

Citations

References

Scientific literature supporting this research

1 Dumyati, G., Stone, N. D., Nace, D. A., Crnich, C. J., & Jump, R. L. P. (2017). Challenges and strategies for prevention of multidrug-resistant organism transmission in nursing homes. Current Infectious Disease Reports, 19(4). https://doi.org/10.1007/s11908-017-0576-7
2 Chris Dall, M. A. (n.d.). US researchers to investigate multidrug-resistant organisms in nursing homes. CIDRAP. https://www.cidrap.umn.edu/antimicrobial-stewardship/us-researchers-investigate-multidrug-resistant-organisms-nursing-homes
3 Huebner, C., Roggelin, M., & Flessa, S. (2016). Economic burden of multidrug-resistant bacteria in nursing homes in Germany: a cost analysis based on empirical data. BMJ Open, 6(2), e008458. https://doi.org/10.1136/bmjopen-2015-008458
4 McKinnell, J. A., et al. (2020). High Prevalence of Multidrug-Resistant Organism Colonization in 28 Nursing Homes: An "Iceberg Effect". Journal of the American Medical Directors Association, 21(12), 1937–1943.e2. https://doi.org/10.1016/j.jamda.2020.04.007
5 Russakoff, B., et al. (2023). A Quantitative Assessment of Staphylococcus aureus Community Carriage in Yuma, Arizona. The Journal of Infectious Diseases, 227(9), 1031–1041. https://doi.org/10.1093/infdis/jiac438
6 Scott, S. E. (2020). Notes from the Field: Carbapenemase-Producing Klebsiella pneumoniae in a Ventilator-Capable Skilled Nursing Facility—Maricopa County, Arizona, July–November 2018. MMWR. Morbidity and Mortality Weekly Report, 69.
7 Almulhim, A. S., & Alamer, A. (2020). The prevalence of resistant Gram-negative bacteraemia among hospitalized patients in Tucson, Arizona over a 12-month period. The Journal of International Medical Research, 48(1). https://doi.org/10.1177/0300060519829987
8 Ellingson, K. D., et al. (2022). Interfacility transfer communication of multidrug-resistant organism colonization or infection status. Infection Control & Hospital Epidemiology, 43(4), 448–453. https://doi.org/10.1017/ice.2021.131
9 Hawkes, B. A., et al. (2023). Healthcare System Distrust and Non-Prescription Antibiotic Use. Antibiotics, 12(1), 79. https://doi.org/10.3390/antibiotics12010079
10 Vanhems, P., et al. (2013). Estimating potential infection transmission routes in hospital wards using wearable proximity sensors. PLoS One, 8(9), e73970.
11 ETH Zurich. (2024). Finding and blocking infection routes in hospitals. https://ethz.ch/en/news-and-events/eth-news/news/2024/02/finding-and-blocking-infection-routes-in-hospitals.html
12 Isella, L., et al. (2011). Close encounters in a pediatric ward: measuring face-to-face proximity and mixing patterns with wearable sensors. PLoS One, 6(2), e17144.
13 Barrat, A., et al. (2014). Measuring contact patterns with wearable sensors: methods, data characteristics and applications to data-driven simulations of infectious diseases. Clinical Microbiology and Infection, 20(1), 10-16.
14 Fragkouli, S.-C., et al. (2024). Synthetic data: How could it be used for infectious disease research? arXiv. http://arxiv.org/abs/2407.06211
15 Bioglio, L., et al. (2016). Recalibrating disease parameters for increasing realism in modeling epidemics in closed settings. BMC Infectious Diseases, 16(1). https://doi.org/10.1186/s12879-016-2003-3
16 Seibert, K., et al. (2021). Application Scenarios for Artificial Intelligence in Nursing Care: Rapid Review. Journal of Medical Internet Research, 23(11), e26522. https://doi.org/10.2196/26522
17 Danchin, A. (2024). Artificial intelligence-based prediction of pathogen emergence and evolution in the world of synthetic biology. Microbial Biotechnology, 17(10), e70014. https://doi.org/10.1111/1751-7915.70014
18 Dumyati, G., et al. (2017). Challenges and Strategies for Prevention of Multidrug-Resistant Organism Transmission in Nursing Homes. Curr Infect Dis Rep, 19, 18. https://doi.org/10.1007/s11908-017-0576-7
19 Crnich, C. J., et al. (2015). Optimizing Antibiotic Stewardship in Nursing Homes: A Narrative Review and Recommendations for Improvement. Drugs Aging, 32, 699–716. https://doi.org/10.1007/s40266-015-0292-7
20 Taylor, L. N., & Crnich, C. J. (2022). Accelerating the Growth of Antibiotic Stewardship in Nursing Homes. JAMA Network Open, 5(2), e220211. https://doi.org/10.1001/jamanetworkopen.2022.0211
21 Eikelenboom-Boskamp, A., et al. (2023). A practice guide on antimicrobial stewardship in nursing homes. Antimicrobial Resistance & Infection Control, 12, 120. https://doi.org/10.1186/s13756-023-01321-0
22 Neto, O. L., Haenni, S., Phuka, J., Ozella, L., Paolotti, D., et al. (Additional reference from uploaded document)