Methicillin resistant Staphylococcus aureus transmission reduction using AgentBased

Methicillin resistant Staphylococcus aureus transmission reduction using Agent-Based Modeling and Simulation Sean Barnes Ph. D Student, Applied Mathematics & Scientific Computation University of Maryland, College Park sbarnes@math. umd. edu Dr. Bruce Golden Professor, Robert H. Smith School of Business University of Maryland, College Park bgolden@rhsmith. umd. edu MRSA

Agenda Recap Python and Sim. Py Implementation Validation and Testing Future Tasks Questions 2/27/2021 MRSA 2

Recap • The goal of this project is to model the transmission dynamics of methicillin resistant Staphylococcus aureus (MRSA) within a hospital, primarily through simulating the direct interactions between health care workers and patients • The software is an agent-based simulation package, developed in Python, which will be used to evaluate the effectiveness of infection control procedures in reducing the incidence of hospital-acquired infections 2/27/2021 MRSA 3

Interactions and State Transitions Visitors Patients Not Colonized Susceptible Decolonization Colonized Infected Treatment HCWs Contamination Susceptible Hand wash Colonized No environmental effects 2/27/2021 MRSA 4

Python and Sim. Py • Processes: Executed during the course of the simulation, proceeding through active and passive states Process Execution Method (PEM) • Resources – Resource: Finite (or infinite) capacity resource where processes can request a single unit – Store: Finite capacity reserve where processes can request and return single or multiple resources, which can be processes • Scheduler – Using the PEMs of all processes, schedules and processes discrete events until some specified time 2/27/2021 MRSA 5

Starting and Stopping Processes • Yield statements – hold: Process holds for a fixed amount of time – request/release: Process holds until it acquires/finishes resource unit – put/get: Process holds until it can put/get store unit(s) • Interruptions – Processes can interrupt other active processes • Events - Sim. Event – waitevent: All processes that wait for an event are restarted when the event is signaled – queueevent: Processes enter a queue for the event and are restarted in a FIFO manner 2/27/2021 MRSA 6

Software Modules • HAgent – Agent and object definitions • HSim – Simulation Model • HAux – Auxiliary functions 2/27/2021 MRSA 7

HAgent • Hospital class • Patient class • Health Care Worker (HCW) class – Nurse class – Physician class • Visitor class • Auxiliary classes – – Source class: Generates agents during simulation Global class: Stores global data and parameters Time class: Marks fixed time period HStat class: Stores infection metrics and other useful data Bold classes are modeled as processes 2/27/2021 MRSA 8

Hospital Class • Characteristics: – Population statistics – Rooms: Single/Double capacity resources – Staffs: Nurse/Physician stores – Infection Control Policy • PEM: Hospital opens, allowing patients to enter waiting room each day and request beds 2/27/2021 MRSA 9

Patient Class • Characteristics: – State and risk level – Length of stay and # of HCW visits per day – Transmission probabilities – Infection control parameters • PEM: Patient enters hospital, requests bed, receives care, and is discharged 2/27/2021 MRSA 10

HCW Class • Characteristics: – State – Hand hygiene parameters (Low/high risk) – Transmission probabilities • PEM: HCWs are requested by patients, provide care, implement infection control procedures, and become available for the next patient 2/27/2021 MRSA 11

Visitor Class • Characteristics: – State – Transmission probabilities • PEM: Visitor visits a patient in hospital at random and transmits MRSA to patient with some probability and then departs hospital 2/27/2021 MRSA 12

Source Class • Used to generate agents: – Patients: Generated each day until hospital “waiting room” is filled to capacity – HCWs: Fixed number generated at the beginning of the simulation, specified by parameters – Visitors: Fixed number generated each day, specified by a parameter 2/27/2021 MRSA 13

Infection Control Programs • Hand hygiene compliance • Patient screening – On admission – With some frequency during stay • • Decolonization (Treatment) Patient isolation Patient cohorting Nurse cohorting 2/27/2021 MRSA 14

Infection Metrics • Successful introduction rate: No. of secondary cases • Ward prevalence: Percentage of days on which at least one colonized patient was present • Colonized patient days: Percentage of total days spent as a colonized or infected patient • Attack rate: Ratio of MRSA transmissions to uncolonized patient days • Basic reproduction number, R 0: Mean number of secondary MRSA cases as the result of a single primary case 2/27/2021 MRSA 15

HSim • Import required modules: HAgent, HAux, Sim. Py, Num. Py, random number generator package (Sci. Py) • Define simulation parameters: Hospital size, number of nurses and physicians, simulation time, replications • Initialize simulation: initialize() • Instantiate and activate processes: Hospital, patients, and sources • Instantiate resources: Nurse and physician staffs (stores) • Execute simulation: simulate(until=stop) • Accumulate statistics • Output results 2/27/2021 MRSA 16

General Operation • Initialization – Hospital is defined with single/double rooms and infection control policy – Nurse and physician staffs are created in the hospital • Simulation – – – Patients enter hospital and request a bed Patients are admitted once a bed becomes available Patients are visited each day by nurses, physicians, and sometimes visitors During each visit, MRSA can be transmitted to/from the patient For each patient acquiring an infection, their stay is extended, otherwise they are released at the end of their original stay – Any infection control procedures are implemented by HCWs • Output of Results – Simulation parameters – Infection control policy – Simulation results, including infection metrics 2/27/2021 MRSA 17

Sample Output 2/27/2021 MRSA 18

Computing Genome Cluster: 16 processors, 64 GB RAM • Experiment (Small problem): • Experiment (Large problem): – 100 days – Hospital with 5 single and 5 double rooms – Staff with 5 nurses and 3 physicians – Patient screening and decolonization policies – 5 replications • Run times: – 1000 days – Hospital with 100 single and 200 double rooms – Staff with 50 nurses and 20 physicians – Patient screening and decolonization policies – 50 replications • Run times: – PC: 13. 1 seconds – Genome: 9. 2 seconds 2/27/2021 – PC: 13. 2 hours – Genome: 9. 2 hours MRSA 19

Validation 2/27/2021 MRSA 20

Future Work • • • Implement remaining infection control procedures Improve simulation run time Design parameter input system Further validation and testing Time permitting: – Parallel Implementation: Parallel Python (PP) – Graphical User Interface (GUI) – Visualization (MATLAB, Animation) 2/27/2021 MRSA 21

Questions? Acknowledgements Dr. Harold Standiford, UMMC Carter Price, University of Maryland Dr. Edward Wasil, American University Dr. Catherine Dibble, Aiki Labs MRSA

References 1. Cooper BS, Medley GF, Scott GM, 1999. Preliminary analysis of the transmission dynamics of nosocomial infections: stochastic and management effects. Journal of Hospital Infection, Vol. 43, pp. 131 -147. 2. Raboud J, Saskin R, Simor A, Loeb M, Green K, Low D, Mc. Geer A, 2003. Modeling Transmission of Methicillin-Resistant Staphylococcus Aureus Among Patients Admitted To A Hospital. Infection Control and Hospital Epidemiology, Vol. 26 (7), pp. 607614. 3. Mc. Bryde ES, Pettitt AN, Mc. Elwain DLS, 2007. A stochastic mathematical model of methicillin resistant Staphylococcus aureus transmission in an intensive care unit: Predicting the impact of interventions. Journal of Theoretical Biology, Vol. 245, pp. 470481. 4. Beggs CB, Shepherd SJ, Kerr KG, 2008. Increasing the frequency of hand washing by healthcare workers does not lead to commensurate reductions in staphylococcal infection in a hospital 2/27/2021 MRSA 23 ward