Balancing Collective Exploration and Exploitation in Multi-Agent and Multi-Robot Systems: A Review.

Multi-agent systems and multi-robot systems have been recognized as unique solutions to complex dynamic tasks distributed in space. Their effectiveness in accomplishing these tasks rests upon the design of cooperative control strategies, which is acknowledged to be challenging and nontrivial. In particular, the effectiveness of these strategies has been shown to be related to the so-called exploration-exploitation dilemma: i.e., the existence of a distinct balance between exploitative actions and exploratory ones while the system is operating. Recent results point to the need for a dynamic exploration-exploitation balance to unlock high levels of flexibility, adaptivity, and swarm intelligence. This important point is especially apparent when dealing with fast-changing environments. Problems involving dynamic environments have been dealt with by different scientific communities using theory, simulations, as well as large-scale experiments. Such results spread across a range of disciplines can hinder one's ability to understand and manage the intricacies of the exploration-exploitation challenge. In this review, we summarize and categorize the methods used to control the level of exploration and exploitation carried out by an multi-agent systems. Lastly, we discuss the critical need for suitable metrics and benchmark problems to quantitatively assess and compare the levels of exploration and exploitation, as well as the overall performance of a system with a given cooperative control algorithm.

Reporting of Perioperative Adverse Events by Pediatric Anesthesiologists at a Tertiary Children's Hospital: Targeted Interventions to Increase the Rate of Reporting.

BACKGROUND: Incident reporting systems (IRSs) are important patient safety tools for identifying risks and opportunities for improvement. A major IRS limitation is underreporting of incidents. Perioperative anesthesia IRSs have been established at multiple pediatric institutions and a national pediatric anesthesia IRS for perioperative serious adverse events (SAEs) is maintained by Wake Up Safe (WUS), a patient safety organization dedicated to pediatric anesthesia quality improvement. A confidential, electronic, perioperative IRS was instituted at our tertiary children's hospital, which is a WUS member. The primary study aim was to increase the rate of incident reporting by anesthesiologists at our institution through a series of interventions. The secondary aim was to characterize our reporting behavior relative to national practice by referencing SAE data from WUS. METHODS: Perioperative adverse events reported over a 71-month period (November 2010 to September 2016) were categorized and the monthly reporting rates determined. Effects of 6 interventions targeted to increase the reporting rate were analyzed using control charts. Intervention 5 involved interviewing pediatric anesthesiologists to ascertain incident reporting barriers and motivators. A key driver diagram was developed and used to guide an improvement initiative. Incidents that fulfilled WUS criteria for SAEs were identified and categorized. SAE reporting rates over a 27-month period for 12 WUS member institutions were determined. RESULTS: 2689 perioperative adverse events were noted in 1980 of 72,384 anesthetics. Mean monthly adverse event case rate was 273 (95% confidence interval, 250-297) per 10,000 anesthetics. A subgroup involving 54,469 cases had 529 SAEs in 440 anesthetics; a mean monthly SAE case rate of 80 (95% confidence interval, 69-91) per 10,000 anesthetics. Cardiac, respiratory, and airway events predominated. Relative to WUS peer members, our institution is a high-reporting outlier. The rate of incident reporting per 10,000 anesthetics was sustainably increased from 149 ± 35 to 387 ± 73 (mean ± SD) after implementing mandatory IRS data entry and Intervention 5 quality improvement initiative. Barriers to reporting included concern for punitive repercussions, feelings of incompetence, poor education about what constitutes an event, lack of feedback, and the perception that reporting had no value. These were addressed by IRS education, cultivation of a culture of safety where reporting is encouraged, reporter feedback, and better inclusion of anesthesiologists in patient safety work. CONCLUSIONS: Electronic mandatory IRS data entry and an initiative to understand and address reporting barriers and motivators were associated with sustained increases in the adverse event reporting rate. These strategies to minimize underreporting enhance IRS value for learning and may be generalizable.

Implementation of a standardized postanesthesia care handoff increases information transfer without increasing handoff duration.

BACKGROUND: In the transition of a patient from the operating room (OR) to the postanesthesia care unit (PACU), it was hypothesized that (1) standardizing the members of sending and receiving teams and (2) requiring a structured handoff process would increase the overall amount of patient information transferred in the OR-to-PACU handoff process. METHODS: A prospective cohort study was conducted at a 311-bed freestanding academic pediatric hospital in Northern California. The intervention, which was conducted in February-March 2013, consisted of (1) requiring the sending team to include a surgeon, an anesthesiologist, and a circulating nurse, and the receiving team to include the PACU nurse; (2) standardizing the content of the handoff on the basis of literature-guided recommendations; and (3) presenting the handoff verbally in the I-PASS format. Data included amount of patient information transferred, duration of handoff, provider presence, and nurse satisfaction. RESULTS: Forty-one audits during the preimplementation phase and 45 audits during the postimplementation phase were analyzed. Overall information transfer scores increased significantly from a mean score of 49% to 83% (p < .0001). Twenty-two PACU nurse satisfaction surveys were completed after the preimplementation phase and 14 surveys were completed in the postimplementation phase. Paired mean total satisfaction scores increased from 36 to 44 (p =. 004). The duration of the handoffs trended downward from 4.1 min to 3.5 min (p = 0.10). CONCLUSION: A standardized, team-based approach to OR-to-PACU handoffs increased the quantity of patient information transferred, increased PACU nurse satisfaction, and did not increase the handoff duration.

Variability of characteristics and outcomes following cardiopulmonary resuscitation events in diverse ICU settings in a single, tertiary care children's hospital*.

OBJECTIVE: The primary objective of this study was to compare and contrast the characteristics and survival outcomes of cardiopulmonary resuscitation for "monitored" events in pediatric patients treated with chest compressions more than or equal to 1 minute in varied ICU settings. DESIGN: Retrospective observational study. SETTING: Three different specialized ICUs in a single, tertiary care, academic children's hospital. PATIENTS: We collected demographic information, preexisting conditions, preevent characteristics, event characteristics, and outcome data. The primary outcome measure was survival to hospital discharge. Secondary outcome measures included return of spontaneous circulation, 24-hour survival, and survival with good neurologic outcome. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Four hundred eleven patients treated with chest compressions for more than or equal to 1 minute were included in the analysis: 170 patients were located in the cardiovascular ICU, 157 patients in the neonatal ICU, and 84 patients in the PICU. Arrest durations were longer in the cardiovascular ICU than other ICUs. Use of extracorporeal cardiopulmonary resuscitation was more prevalent in the cardiovascular ICU (cardiovascular ICU, 17%; neonatal ICU, 3%; PICU, 4%). Return of spontaneous circulation, 24-hour survival, survival to hospital discharge, and good neurologic outcome were highest among neonatal ICU patients (survival to discharge, 53%) followed by cardiovascular ICU patients (survival to discharge, 46%) and PICU patients (survival to discharge, 36%). In a multivariable model controlling for patient and event characteristics, using cardiovascular ICU as reference, adjusted odds of survival in PICU were 0.33 (95% CI, 0.14-0.76; p = 0.009) and odds of survival in neonatal ICU were 0.80 (95% CI, 0.31-2.11; p = 0.65). CONCLUSIONS: Comparative analysis of pediatric patients undergoing cardiopulmonary resuscitation in three different ICU settings demonstrated a significant variation in baseline, preevent, and event characteristics. Although outcomes vary significantly among the three different ICUs, it was difficult to ascertain if this difference was due to variation in the disease process or variation in the location of the patient.

Improving code team performance and survival outcomes: implementation of pediatric resuscitation team training.

OBJECTIVE: To determine whether implementation of Composite Resuscitation Team Training is associated with improvement in survival to discharge and code team performance after pediatric in-hospital cardiopulmonary arrest. DESIGN, SETTING, AND SUBJECTS: We conducted a prospective observational study with historical controls at a 302-bed, quaternary care, academic children's hospital. Inpatients who experienced cardiopulmonary arrest between January 1, 2006, and December 31, 2009, were included in the control group (123 patients experienced 183 cardiopulmonary arrests) and between July 1, 2010, and June 30, 2011, were included in the intervention group (46 patients experienced 65 cardiopulmonary arrests). INTERVENTION: Code team members were introduced to Composite Resuscitation Team Training and continued training throughout the intervention period (January 1, 2010-June 30, 2011). Training was integrated via in situ code blue simulations (n = 16). Simulations were videotaped and participants were debriefed for education and process improvement. Primary outcome was survival to discharge after cardiopulmonary arrest. Secondary outcome measures were 1) change in neurologic morbidity from admission to discharge, measured by Pediatric Cerebral Performance Category, and 2) code team adherence to resuscitation Standard Operating Performance variables. MEASUREMENTS AND MAIN RESULTS: The intervention group was more likely to survive than the control group (60.9% vs 40.3%) (unadjusted odds ratio, 2.3 [95% CI, 1.15-4.60]) and had no significant change in neurologic morbidity (mean change in Pediatric Cerebral Performance Category 0.11 vs 0.27; p = 0.37). Code teams exposed to Composite Resuscitation Team Training were more likely than control group to adhere to resuscitation Standard Operating Performance (35.9% vs 20.8%) (unadjusted odds ratio, 2.14 [95% CI, 1.15-3.99]). After adjusting for adherence to Standard Operating Performance, survival remained improved in the intervention period (odds ratio, 2.13 [95% CI, 1.06-4.36]). CONCLUSION: With implementation of Composite Resuscitation Team Training, survival to discharge after pediatric cardiopulmonary arrest improved, as did code team performance. Demonstration of improved survival after adjusting for code team adherence to resuscitation standards suggests that this may be a valuable resuscitation training program. Further studies are needed to determine causality and generalizability.

Nursing mortality and morbidity and journal club cycles: paving the way for nursing autonomy, patient safety, and evidence-based practice.

The ability to perform professional scrutiny is a required component of autonomous practice, and the ability to use science to guide care delivery is a hallmark of professional practice. The cardiovascular intensive care unit staff initiated a forum for peer review to encourage the use of the best available evidence to guide system and practice changes. A focus group was formed, resulting in development and implementation of a specific process for regular cycles of nursing mortality and morbidity (M & M) conferences, a journal club, and interdisciplinary educational sessions in staff meetings. After performing several cycles, the team observed interdisciplinary practice changes that improved patient care delivery. Thus, cycles of nursing M & M conferences, a journal club, and educational sessions can help support patient safety, enhance professional autonomy, and foster evidence-based interdisciplinary practice in the cardiovascular intensive care setting.

Effect of a rapid response team on hospital-wide mortality and code rates outside the ICU in a Children's Hospital.

CONTEXT: Introduction of a rapid response team (RRT) has been shown to decrease mortality and cardiopulmonary arrests outside of the intensive care unit (ICU) in adult inpatients. No published studies to date show significant reductions in mortality or cardiopulmonary arrests in pediatric inpatients. OBJECTIVE: To determine the effect on hospital-wide mortality rates and code rates outside of the ICU setting after RRT implementation at an academic children's hospital. DESIGN, SETTING, AND PARTICIPANTS: A cohort study design with historical controls at a 264-bed, free-standing, quaternary care academic children's hospital. Pediatric inpatients who spent at least 1 day on a medical or surgical ward between January 1, 2001, and March 31, 2007, were included. A total of 22,037 patient admissions and 102,537 patient-days were evaluated preintervention (before September 1, 2005), and 7257 patient admissions and 34,420 patient-days were evaluated postintervention (on or after September 1, 2005). INTERVENTION: The RRT included a pediatric ICU-trained fellow or attending physician, ICU nurse, ICU respiratory therapist, and nursing supervisor. This team was activated using standard criteria and was available at all times to assess, treat, and triage decompensating pediatric inpatients. MAIN OUTCOME MEASURES: Hospital-wide mortality rates and code (respiratory and cardiopulmonary arrests) rates outside of the ICU setting. All outcomes were adjusted for case mix index values. RESULTS: After RRT implementation, the mean monthly mortality rate decreased by 18% (1.01 to 0.83 deaths per 100 discharges; 95% confidence interval [CI], 5%-30%; P = .007), the mean monthly code rate per 1000 admissions decreased by 71.7% (2.45 to 0.69 codes per 1000 admissions), and the mean monthly code rate per 1000 patient-days decreased by 71.2% (0.52 to 0.15 codes per 1000 patient-days). The estimated code rate per 1000 admissions for the postintervention group was 0.29 times that for the preintervention group (95% likelihood ratio CI, 0.10-0.65; P = .008), and the estimated code rate per 1000 patient-days for the postintervention group was 0.28 times that for the preintervention group (95% likelihood ratio CI, 0.10-0.64; P = .007). CONCLUSION: Implementation of an RRT was associated with a statistically significant reduction in hospital-wide mortality rate and code rate outside of the pediatric ICU setting.