Last updated: May 10, 2023 at 7:43 pm

The goal of our project is to create a novel communication platform that allows a sepsis team in the PICU to see each step of the sepsis management process. We aim to augment a process map of the sepsis management process with live antibiotic delivery tracking in order to alert each person on the team when it is their turn to act.

• Students:
  • Sofia Posada (CS '24)
  • Rahul Swaminathan (MSE in BME ' 23)
• Mentor(s):
  • Dr. Jim Fackler (M.D., Anesthesiology and Critical Care Medicine)
  • Dr. Kimia Ghobadi (PhD, Department of Civil and Systems Engineering)

Sepsis is the body’s extreme reaction to an infection. It is a main cause of global morbidity and mortality, with peak mortality occurring in pediatric and elderly patients. Based on a systematic review with 15 studies included in the meta-analysis, they found that there are, on average, 48 sepsis cases in children per 100000 person-years. Within these cases, mortality ranged from 13 to 54%. The majority of deaths from children with severe sepsis occur during the first 24 hours after the children are referred to the pediatric intensive care unit.

Since sepsis attacks the body so quickly, timing is of utmost importance when treating children with sepsis. A child can appear healthy one moment and quickly develop sepsis and need to be put on life support, so detecting sepsis as soon as possible is critical. Once detected, delivering antibiotics to a septic child within one hour is critical, with a noted decrease in mortality by 40%. The data showed that the children who received antibiotics within one hour had, on average, a shorter hospital length of stay and shorter in-hospital mortality rate. Evidently, timing is critical when treating sepsis.

Though it is known that identifying sepsis early and treating patients within one hour is critical, there is no implementation in place to accomplish this. After speaking with Dr. Fackler and other attending physicians at the Johns Hopkins PICU, it is evident there is a communication issue when it comes to treating children with sepsis.

Our goal is to find a viable solution to this problem and be able to successfully provide IV antibiotics within an hour to children diagnosed with a serious bacterial infection. We will then test our solution in the PICU and collect new timestamps. We hope our solution will allow antibiotics to be delivered within one hour and, therefore, indirectly improve treatment outcomes for children diagnosed with sepsis.

• Minimum: (Expected by 3/17/2023 4/10/2023) (Changed on 3/26/2023)
  1. Finalized process map of sepsis management workflow (COMPLETE)
  2. Short report of gap analysis showing where a communication solution is most needed (COMPLETE) (Rahul and Sofia)

• Expected: (Expected by 4/28/2023)
  1. Physical prototype of communication solution that allows providers to visualize where the antibiotics are within the hospital (COMPLETE for Pneumatic Tube System)
     1. Track antibiotics through pneumatic tube system with Bluetooth proximity sensor (COMPLETE) (Rahul)
     2. Track antibiotics in PICU with barcode scanning (Prototype COMPLETE) (Sofia)
     3. Correlate antibiotics location with roles on process map (COMPLETE) (Sofia, Rahul)
     4. Set up alerts to automatically inform the relevant personnel when it is their turn (COMPLETE) (Rahul)

• Maximum: (Expected by 5/18/2023 5/11/2023) (Changed on 3/30/2023 to reflect the removal of one planned maximum deliverable)
  1. Implemented prototype of communication solution that uses a monitor to display the location of the antibiotics and whose turn it is within the process map
     1. Create mockup of idealy monitor display (COMPLETE) (Sofia)
     2. Set up monitor in PICU telemonitoring room to display antibiotic location and process map (INCOMPLETE)
     3. Program live process map that lights up depending on antibiotic location (INCOMPLETE)
  2. Dataset of timestamps and locations for antibiotic delivery after the solution is implemented (Sofia)

The project workflow is split into the following main components:

1. Augment the Process Map
   1. Augmenting the process map will involve conducting clinician and provider interviews with the help of Chen Wang, a Master’s student currently working with Dr. Fackler. We will start in the hospital pharmacy, where the antibiotics first originate, and interview the pharmacists there to learn more about the beginning of the workflow and any problems they face with initial delivery. We will then interview nurses and attending physicians to learn about problems they are facing and what their ideal communication solution would entail. We anticipate that different groups within the hospital may use different communication platforms, so we will also identify common means of communication and what works best for each group.

1. Collect Data and Perform Gap Analysis
   1. We plan to re-purpose the existing Versus Information System (VIS) tracking technology in the hospital. VIS is a real-time locating system that was initially introduced to hospitals as a way of tracking the locations of nurses and patients within a hospital. VIS uses infrared and radiofrequency based tracking to identify room-level locations rather than approximate locations. Rather than using it to track nurses, we plan to collaborate with the hospital pharmacy to place VIS tags in the bags containing the antibiotics in order to track their location throughout the delivery process. This will not only give us an idea of how the antibiotics go from the pharmacy to the patient, but also where and when delays happen, and whether there are any specific problem areas within the hospital.
   2. Change (3/26/2023): After additional research and pharmacist interviews, we determined that this approach is not feasible because the VIS tracking technology cannot utilize its infrared sensors within the pneumatic tube system until the antibiotics are taken out of the opaque protective capsule. We will instead conduct our analysis using a limited dataset provided by Dr. Fackler
   3. Change (4/12/2023): After meeting with Anatoly Gimburg from JHH Facilities Management, we have learned about a publicly available dataset of pneumatic tube delivery data, so we will also be using that for our gap analysis.

1. Develop a Physical Prototype of the Communication Solution
   1. We intend to export the tracked antibiotic locations and correlate them to specific roles or team members that are in charge of transporting or administering the antibiotics at those locations. We plan to use this information in combination with either VIS’s internal alert system or Epic’s secure chat to contact the relevant personnel when it is their turn to take over antibiotic delivery. Our eventual goal is to attempt to implement a program in which we can display the augmented process map on a monitor in the hospital with certain areas highlighted to show where the antibiotic is currently located. This will allow all team members to see where they currently are in the sepsis management process as well allow the relevant team members at each location to receive timely alerts.
   2. Change (4/4/2023): Since VIS tags cannot be used in the pneumatic tube system, we intend to use it just within the PICU and instead use a barcode scanning tracking system at other steps. Nurses and pharmacy techs will scan a barcode on the antibiotics bag at every step of the delivery process, similar to package tracking done by companies like UPS and FedEx.
   3. Change (4/15/2023): We were not able to acquire VIS tags, so we will be using barcode scanning to track the antibiotics within the PICU and Bluetooth to track/detect the antibiotics through the pneumatic tube system.

1. Test Efficiency of Solution (if time permits)
   1. We will re-do the VIS tracking experiment from step 2 in order to compare the antibiotic delivery time before and after implementation of our solution, ideally reducing delivery time to be consistently below one hour.
   2. Change (3/30/2023): We have determined that we likely will not have time to test our solution post-implementation, especially since the tracking experiment from step 2 is infeasible. If we do have time, this efficiency testing will likely be based on just a few samples.

As shown in the table of dependencies, green signifies that the dependency was successfully resolved, yellow signifies that the dependency was partially resolved, and red signifies that the dependency was never met, so the contingency plan was followed.

1. Milestone name: Finalized process map
   • Planned Date: 3/7/2023
   • Expected Date: 3/7/2023
   • Status: Completed on 3/9/2023
2. Milestone name: Complete antibiotic delivery tracking dataset * Planned Date: 3/9/2023 * Expected Date: 3/9/2023 * Status: Have some data as of 3/7 but not all
3. Milestone name: Gap analysis report
   • Planned Date: 3/16/2023
   • Expected Date: 4/10/2023
   • Status: Completed on 4/25/2023 due to further delays in receiving access to data
4. Milestone name: Role-location mapping
   • Planned Date: 4/20/2023
   • Expected Date: 4/20/2023
   • Status: Completed on 4/20/2023
5. Milestone name: Physical prototype
   • Planned Date: 4/27/2023
   • Expected Date: 4/27/2023
   • Status: Completed first round of testing with Bluetooth proximity sensor prototype on 4/26/2023
6. Milestone name: Implemented prototype
   • Planned Date: 5/11/2023
   • Expected Date: 5/11/2023
   • Status: Planning to test Bluetooth sensor prototype with full workflow in mid-May
7. Milestone name: Post-implementation dataset * Planned Date: 5/18/2023 * Expected Date: 5/18/2023 * Status: Cannot begin until previous milestone is complete.

• Project Plan
  • Project Plan Presentation
  • Project Plan Proposal
• Project Background Reading
  • Paper A
  • Paper B
• Paper Seminar Report
  • Seminar Report A
  • Seminar Report B
• Paper Seminar Presentations
  • Seminar Presentation
• Design Specifications
  • Sepsis Communication Solution
  • Sepsis Antibiotic Tracking
• Gap Analysis
  • Gap Analysis Report
  • Patient Data Analysis Report
• Project Checkpoint
  • Project Checkpoint Presentation
• Project Final Checkpoint
  • Final Checkpoint Presentation PDF
• Project Final Presentation
  • Final Poster PDF
• Project Final Report
  • Final Report
  • Mentor Feedback Form

1. Goldstein, B. , Giroir, B. Randolph, A. (2005). International pediatric sepsis consensus conference: Definitions for sepsis and organ dysfunction in pediatrics*. Pediatric Critical Care Medicine, 6 (1), 2-8. doi: 10.1097/01.PCC.0000149131.72248.E6.
2. Fleischmann-Struzek, C., Goldfarb, D. M., Schlattmann, P., Schlapbach, L. J., Reinhart, K., Kissoon, N. (2018). The global burden of paediatric and neonatal sepsis: a systematic review. The Lancet. Respiratory medicine, 6(3), 223–230. https://doi.org/10.1016/S2-213-2600-18-30063-8
3. Cvetkovic, M., Lutman, D., Ramnarayan, P., Pathan, N., Inwald, D. P., Peters, M. J. (2015). Timing of death in children referred for intensive care with severe sepsis: implications for interventional studies. Pediatric critical care medicine : a journal of the Society of Critical Care Medicine and the World Federation of Pediatric Intensive and Critical Care Societies, 16(5), 410–417. https://doi.org/10.1097/PCC.0000000000000385
4. Qian, J.1; Wang, Y.I.N.G.1; Zhang, Y.2; Rong, Q.2; Zhu, X.3. ABSTRACT 696: AN EVALUATION OF SIMULATION ON THE FIRST HOUR BASIC CARE TASKS OF THE SEPTIC SHOCK IN PEDIATRIC PATIENTS. Pediatric Critical Care Medicine 15(4_suppl):p 157, May 2014. | DOI: 10.1097/01.pcc.0000449422.48869.74
5. Cowart, M. C., Heath, T. S., Shipman, A. (2022). The Effect of Rapid Initiation Versus Delayed Initiation of Antibiotics in Pediatric Patients With Sepsis. The journal of pediatric pharmacology and therapeutics : JPPT : the official journal of PPAG, 27(1), 45–50. https://doi.org/10.5863/1551-6776-27.1.45
6. Mari, D. , Rush, T. Biswas, A. (2023). 730: Antibiotic Timing in Neonates and Children with Sepsis or Septic Shock Does Not Affect PICU Stay. Critical Care Medicine, 51 (1), 355-355. doi: 10.1097/01.ccm.0000908652.60631.11.
7. Im, Y., Kang, D., Ko, RE. et al. Time-to-antibiotics and clinical outcomes in patients with sepsis and septic shock: a prospective nationwide multicenter cohort study. Crit Care 26, 19 (2022). https://doi.org.proxy1.library.jhu.edu/10.1186/s13054-021-03883-0
8. Chua W, Ooi S, Chan G, Lau T, Liaw S The Effect of a Sepsis Interprofessional Education Using Virtual Patient Telesimulation on Sepsis Team Care in Clinical Practice: Mixed Methods Study J Med Internet Res 2022;24(4):e-35058 URL: https://www.jmir.org/2022/-4/e35058 DOI: 10.2196/35058
9. Nomura O, Ihara T, Morikawa Y, Sakakibara H, Horikoshi Y, Inoue N. Predictor of Early Administration of Antibiotics and a Volume Resuscitation for Young Infants with Septic Shock. Antibiotics. 2021; 10(11):1414. https://doi.org/10.3390/antibiotics10111414

To adhere to HIPAA guidelines, our important project files are stored on a SAFE Desktop.

Code to test the Bluetooth proximity sensor can be found here: scan_for_ble_tag.pdf