Sean L. Barnes

Preventing the Transmission of Multidrug-Resistant Organisms: Modeling the Relative Importance of Hand Hygiene and Environmental Cleaning Interventions

OBJECTIVE Hand hygiene and environmental cleaning are essential infection prevention strategies, but the relative impact of each is unknown. This information is important in assessing resource allocation. METHODS We developed an agent-based model of patient-to-patient transmission-via the hands of transiently colonized healthcare workers and incompletely terminally cleaned rooms-in a 20-patient intensive care unit. Nurses and physicians were modeled and had distinct hand hygiene compliance levels on entry and exit to patient rooms. We simulated the transmission of Acinetobacter baumannii, methicillin-resistant Staphylococcus aureus, and vancomycin-resistant enterococci for 1 year using data from the literature and observed data to inform model input parameters. RESULTS We simulated 175 parameter-based scenarios and compared the effects of hand hygiene and environmental cleaning on rates of multidrug-resistant organism acquisition. For all organisms, increases in hand hygiene compliance outperformed equal increases in thoroughness of terminal cleaning. From baseline, a 2∶1 improvement in terminal cleaning compared with hand hygiene was required to match an equal reduction in acquisition rates (eg, a 20% improvement in terminal cleaning was required to match the reduction in acquisition due to a 10% improvement in hand hygiene compliance). CONCLUSIONS Hand hygiene should remain a priority for infection control programs, but environmental cleaning can have significant benefit for hospitals or individual hospital units that have either high hand hygiene compliance levels or low terminal cleaning thoroughness.

Contribution of interfacility patient movement to overall methicillin-resistant Staphylococcus aureus prevalence levels.

OBJECTIVES The effect of patient movement between hospitals and long-term care facilities (LTCFs) on methicillin-resistant Staphylococcus aureus (MRSA) prevalence levels is unknown. We investigated these effects to identify scenarios that may lead to increased prevalence in either facility type. METHODS We used a hybrid simulation model to simulate MRSA transmission among hospitals and LTCFs. Transmission within each facility was determined by mathematical model equations. The model predicted the long-term prevalence of each facility and was used to assess the effects of facility size, patient turnover, and decolonization. RESULTS Analyses of various healthcare networks suggest that the effect of patients moving from a LTCF to a hospital is negligible unless the patients are consistently admitted to the same unit. In such cases, MRSA prevalence can increase significantly regardless of the endemic level. Hospitals can cause sustained increases in prevalence when transferring patients to LTCFs, where the population size is smaller and patient turnover is less frequent. For 1 particular scenario, the steady-state prevalence of a LTCF increased from 6.9% to 9.4% to 13.8% when the transmission rate of the hospital increased from a low to a high transmission rate. CONCLUSIONS These results suggest that the relative facility size and the patient discharge rate are 2 key factors that can lead to sustained increases in MRSA prevalence. Consequently, small facilities or those with low turnover rates are especially susceptible to sustaining increased prevalence levels, and they become more so when receiving patients from larger, high-prevalence facilities. Decolonization is an infection-control strategy that can mitigate these effects.

MRSA Transmission Reduction Using Agent-Based Modeling and Simulation

Methicillin-resistant Staphylococcus aureus (MRSA) is a significant ongoing problem in health care, posing a substantial threat to hospitals and communities as well. Its spread among patients causes many downstream effects, such as a longer length of stay for patients, higher costs for hospitals and insurance companies, and fatalities. An agent-based simulation model is developed to investigate the dynamics of MRSA transmission within a hospital. The simulation model is used to examine the effectiveness of various infection control procedures and explore more specific questions relevant to hospital administrators and policy makers. Simulation experiments are performed to examine the effects of hand-hygiene compliance and efficacy, patient screening, decolonization, patient isolation, and health-care worker-to-patient ratios on the incidence of MRSA transmission and other relevant metrics. Experiments are conducted to investigate the dynamic between the number of colonizations directly attributable to nurses and physicians, including rogue health-care workers who practice poor hygiene. We begin to explore the most likely threats to trigger an outbreak in hospitals that practice high hand-hygiene compliance and additional preventive measures.

Applications of Agent-Based Modeling and Simulation to Healthcare Operations Management

Agent-based modeling (ABM) is rapidly gaining momentum in many fields, and it has added to the insights previously contributed by other modeling and simulation methods such as system dynamics and discrete event simulation. Healthcare operations management is one field that is particularly well-suited for ABM because it involves many individuals that interact in different ways. ABM is capable of explicitly modeling these individuals and the interactions among them, which facilitates the discovery of system behavior that cannot be observed using other methods. ABM has been applied successfully to several focus areas within the field of healthcare operations management, including, but not limited to: healthcare delivery, epidemiology, economics, and policy. In this chapter, we review and evaluate a selected body of research in which agent-based modeling and simulation techniques are applied to problems in healthcare. We also highlight specific areas where agent-based modeling and simulation filled a significant gap that was not addressed previously by other methods. Finally, we propose some new questions in the field which may be of interest moving forward.

Machine-Learning-Based Electronic Triage More Accurately Differentiates Patients With Respect to Clinical Outcomes Compared With the Emergency Severity Index

Study objective Standards for emergency department (ED) triage in the United States rely heavily on subjective assessment and are limited in their ability to risk‐stratify patients. This study seeks to evaluate an electronic triage system (e‐triage) based on machine learning that predicts likelihood of acute outcomes enabling improved patient differentiation. Methods A multisite, retrospective, cross‐sectional study of 172,726 ED visits from urban and community EDs was conducted. E‐triage is composed of a random forest model applied to triage data (vital signs, chief complaint, and active medical history) that predicts the need for critical care, an emergency procedure, and inpatient hospitalization in parallel and translates risk to triage level designations. Predicted outcomes and secondary outcomes of elevated troponin and lactate levels were evaluated and compared with the Emergency Severity Index (ESI). Results E‐triage predictions had an area under the curve ranging from 0.73 to 0.92 and demonstrated equivalent or improved identification of clinical patient outcomes compared with ESI at both EDs. E‐triage provided rationale for risk‐based differentiation of the more than 65% of ED visits triaged to ESI level 3. Matching the ESI patient distribution for comparisons, e‐triage identified more than 10% (14,326 patients) of ESI level 3 patients requiring up triage who had substantially increased risk of critical care or emergency procedure (1.7% ESI level 3 versus 6.2% up triaged) and hospitalization (18.9% versus 45.4%) across EDs. Conclusion E‐triage more accurately classifies ESI level 3 patients and highlights opportunities to use predictive analytics to support triage decisionmaking. Further prospective validation is needed.

A dynamic patient network model of hospital-acquired infections

We investigate the transmission of infectious diseases in hospitals using a network-centric perspective. Patients who share a health care worker (HCW) are inherently connected to each other and those connections form a network through which transmission can occur. The structure of such networks can be a strong determinant of the extent and rate of transmission. We first examine how the density of the patient network affects transmission. Our experiments demonstrate that nurses are responsible for spreading more infection because they typically visit patients more often. However, doctors also pose a serious threat because their patient networks are more highly connected, which creates more opportunity for transmission to spread to multiple cohorts in the unit. We also explore the effects of patient sharing among HCWs, which temporarily alters the structure of the patient network. Our results suggest that this practice should be done in a structured manner to minimize additional transmission.

The Impact of Reducing Antibiotics on the Transmission of Multidrug-Resistant Organisms

OBJECTIVE Antibiotic resistance is a major threat to public health. Resistance is largely driven by antibiotic usage, which in many cases is unnecessary and can be improved. The impact of decreasing overall antibiotic usage on resistance is unknown and difficult to assess using standard study designs. The objective of this study was to explore the potential impact of reducing antibiotic usage on the transmission of multidrug-resistant organisms (MDROs). DESIGN We used agent-based modeling to simulate interactions between patients and healthcare workers (HCWs) using model inputs informed by the literature. We modeled the effect of antibiotic usage as (1) a microbiome effect, for which antibiotic usage decreases competing bacteria and increases the MDRO transmission probability between patients and HCWs and (2) a mutation effect that designates a proportion of patients who receive antibiotics to subsequently develop a MDRO via genetic mutation. SETTING Intensive care unit INTERVENTIONS Absolute reduction in overall antibiotic usage by experimental values of 10% and 25% RESULTS Reducing antibiotic usage absolutely by 10% (from 75% to 65%) and 25% (from 75% to 50%) reduced acquisition rates of high-prevalence MDROs by 11.2% (P<.001) and 28.3% (P<.001), respectively. We observed similar effect sizes for low-prevalence MDROs. CONCLUSIONS In a critical care setting, where up to 50% of antibiotic courses may be inappropriate, even a moderate reduction in antibiotic usage can reduce MDRO transmission. Infect Control Hosp Epidemiol 2017;38:663-669.

Exploring the effects of network structure and healthcare worker behavior on the transmission of hospital-acquired infections

We investigate the transmission of infectious diseases in hospitals using a network-centric perspective. Patients who share a healthcare worker are inherently connected to each other and those connections form a network through which transmission can occur. The structure of this network can be a strong determinant of the extent and rate of transmission. We explore the effects of healthcare worker behavior, including sharing patients and incorporating the ability for healthcare workers to infect each other. Finally, we examine how patient turnover can affect transmission dynamics in a patient network under the influence of other effects. Our results show that all of these factors can affect transmission significantly, and that this model can be used to provide additional insight to hospital administrators who are looking to improve their ability to control infections.

Deconstructing the relative benefits of a universal glove and gown intervention on MRSA acquisition

BACKGROUND The 20-site Benefits of Universal Glove and Gown (BUGG) study found that wearing gloves and gowns for all patient contacts in the intensive care unit (ICU) reduced acquisition rates of meticillin-resistant Staphylococcus aureus (MRSA). The relative importance of gloves and gowns as a barrier, improved hand hygiene, and reduced healthcare worker (HCW)-patient contact rates is unknown. AIM To determine what proportion of the reduction in acquisition rates observed in the BUGG study was due to improved hand hygiene, reduced contact rates, and universal glove and gown use using agent-based simulation modelling. METHODS An existing agent-based model to simulate MRSA transmission dynamics in an ICU was modified, and the model was calibrated using site-specific data. Model validation was completed using data collected in the BUGG study. A full 2k factorial design was conducted to quantify the relative benefits of improving each of the aforementioned factors with respect to MRSA acquisition rates. FINDINGS Across 40 simulated replications for each factorial design point and intervention site, approximately 44% of the decrease in MRSA acquisition rates was due to universal glove and gown use, 38.1% of the decrease was due to improvement in hand hygiene compliance on exiting patient rooms, and 14.5% of the decrease was due to the reduction in HCW-patient contact rates. CONCLUSION Using mathematical modelling, the decrease in MRSA acquisition in the BUGG study was found to be due primarily to the barrier effects of gowns and gloves, followed by improved hand hygiene and lower HCW-patient contact rates.

Early detection of bioterrorism: monitoring disease using an agent-based model

We propose an agent-based model to capture the transmission patterns of diseases caused by bioterrorism attacks or epidemic outbreaks and to differentiate between these two scenarios. Focusing on a region of three cities, we want to detect a bioterrorism attack before a sizeable proportion of the population is infected. Our results indicate that the aggregated infection and death curves in the region can serve as indicators in distinguishing between the two disease scenarios: the slope of the epidemic infection curve will increase initially and decrease afterwards, whereas the slope of the bioterrorism infection curve will strictly decrease. We also conclude that for a bioterrorism outbreak, the bioterrorism source city becomes more dominant as the local working probability pL increases. In contrast, the behavior of individual cities for the epidemic model presents a “time-lag” pattern, especially when pL is large. As pL decreases, the individual citys dynamic curves converge as time progresses.