Projects

Inferring HAI characteristics

Using fine-grained spatiotemporal data, we aim to improve our understanding of characteristics of HAIs such as C.diff. Our recent work studies the impact of exposure to family members with C. diff infection and other risk factors on the likelihood of acquiring C.diff infection [1, 2, 3], the significance of spatiotemporal interactions on C.diff infections within a hospital [4], and the impact of hospital transfers on C.diff infection rates in hospitals [5].

  1. A. C. Miller, A. M. Segre, S. V. Pemmaraju, D. K. Sewell, and P. M. Polgreen, “Association of Household Exposure to Primary Clostridioides difficile Infection with Secondary Infection in Family Members,” JAMA Network Open, vol. 3, iss. 6, 2020. [PDF]

  2. A. C. Miller, L. A. Polgreen, J. E. Cavanaugh, and P. M. Polgreen, “Hospital Clostridium difficile infection (CDI) incidence as a risk factor for hospital-associated CDI,” American Journal of Infection Control, vol. 44, iss. 7, pp. 825-829, 2016. [PDF]

  3. A. C. Miller, L. A. Polgreen, J. E. Cavanaugh, and P. M. Polgreen, “Hospital Clostridium difficile Infection Rates and Prediction of Length of Stay in Patients Without C. difficile Infection,” Infection Control and Hospital Epidemiology, vol. 37, iss. 4, pp. 404-410, 2016. [PDF]

  4. S. Pai, P. M. Polgreen, A. M. Segre, D. K. Sewell, and S. V. Pemmaraju, “Spatiotemporal Clustering of In-Hospital Clostridioides difficile Infection (CDI),” Infection Control and Hospital Epidemiology, vol. 41, iss. 4, pp. 418-424, 2020. [PDF]

  5. D. K. Sewell, J. E. Simmering, S. Justice, S. V. Pemmaraju, A. M. Segre, and P. M. Polgreen, “Estimating the Attributable Disease Burden and Effects of Inter-Hospital Patient Sharing on Clostridium difficile Infections,” Infection Control and Hospital Epidemiology, vol. 40, pp. 656-661, 2019. [PDF]

Inferring agent-behavior in healthcare settings

Using data, sometimes gathered using novel technology, we aim to infer behavior of healthcare personnel and patients in hospital settings. We use electronic medical records, sensor network instrumentation, kinect cameras, etc. to estimate contact networks of healthcare personnel and patients [1], hand hygiene behavior of healthcare personnel [2], and duration of close-contacts between healthcare personnel and patients in hospital-rooms [3].

  1. D. E. Curtis, C. S. Hlady, G. Kanade, S. V. Pemmaraju, P. M. Polgreen, and A. M. Segre, “Healthcare Worker Contact Networks and the Prevention of Hospital-Acquired Infections,” PLOS One, 2013. [PDF]

  2. V. Galluzzi, T. Herman, D. J. Shumaker, D. R. Macinga, J. W. Arbogast, E. M. Segre, A. M. Segre, and P. M. Polgreen, “Electronic Recognition of Hand-Hygiene Technique and Duration,” Infection Control and Hospital Epidemiology, vol. 35, iss. 10, pp. 1298-1300, 2014. [PDF]

  3. R. Butler, M. N. Monsalve, G. W. Thomas, T. Herman, A. M. Segre, P. M. Polgreen, and M. Suneja, “Estimating Time Physicians and Other Healthcare Workers Spend with Patients in an Intensive Care Unit Using a Sensor Network,” The American Journal of Medicine, 2018. [PDF]

HAI risk prediction

We build machine learning prediction models using detailed electronic medical record data overlaid with hospital architectural layout for predicting patient risk. In recent work [2] we use a 2-stage prediction model to identify latent C.diff infections (e.g., asymptomatic C.diff carriers). In [1], we predict the daily risk of a patient acquiring a C.diff infection by taking the temporal ordering of events into account as features.

  1. M. N. Monsalve, S. V. Pemmaraju, S. Johnson, and P. M. Polgreen, “Improving Risk Prediction of Clostridium Difficile Infection Using Temporal Event-Pairs,” in 2015 International Conference on Healthcare Informatics, 2015, pp. 140-149. [PDF]

  2. “epiDAMIK Workshop.” Accessed August 24, 2023. [PDF]

Disease-surveillance

We study the use of social media — Twitter [1], Wikipedia [2], and Craig’s list [3] — in helping with disease surveillance. We model the geographic placement of surveillance sites as an optimization problem [4] and propose methods for computing optimal screening rates in [5]. We also build apps for individual-level surveillance [6].

  1. A. Signorini, A. M. Segre, and P. M. Polgreen, “The Use of Twitter to Track Levels of Disease Activity and Public Concern in the US During the Influenza A H1N1 Pandemic,” PLOS One, 2011. [PDF]

  2. G. C. Fairchild, S. DelValle, L. DeSilva, and A. M. Segre, “Eliciting Disease Data From Wikipedia Articles,” in ICWSM Workshop on Wikipedia Research Challenges and Opportunities, Stanford, CA, 2015.

  3. J. A. Fries, P. M. Polgreen, and A. M. Segre, “Mining the Demographics of Craigslist Casual Sex Ads to Inform Public Health Policy,” in IEEE International Conference on Healthcare Informatics, Verona, Italy, 2014.

  4. G. C. Fairchild, P. M. Polgreen, E. Foster, G. Rushton, and A. M. Segre, “How Many Suffice? A Computational Framework for Sizing Sentinel Surveillance Networks,” International Journal of Health Geographics, vol. 12, iss. 56, 2013. [PDF]

  5. A. C. Miller, L. A. Polgreen, and P. M. Polgreen, “Optimal Screening Strategies for Healthcare Associated Infections in a Multi-Institutional Setting,” PLOS Computational Biology, vol. 10, iss. 1, pp. 1-11, 2014. [PDF]

  6. A. C. Miller, I. Singh, E. Koehler, and P. M. Polgreen, “A Smartphone-Driven Thermometer Application for Real-time Population- and Individual-Level Influenza Surveillance,” Clinical Infectious Diseases, vol. 67, iss. 3, pp. 388-397, 2018. [PDF]