Problem investigated: The activation and implementation of the Medical Emergency Team (MET) system. Procedures followed: The ability of the objective activation criteria to accurately identify patients at risk of three serious adverse events (cardiac arrest, unexpected death and unplanned intensive care admission) was assessed using a nested, matched case-control study. Sensitivity, specificity and Receiver Operating Characteristic curve (ROC) analyses were performed. The MET implementation process was studied using two convenience sample surveys of the nursing staff from the general wards of twelve intervention hospitals. These surveys measured the awareness and understanding of the MET system, level of attendance at MET education sessions, knowledge of the activation criteria, level of intention to call the MET and overall attitude to the MET system, and the hospital level of support for change, hospital capability and hospital culture. The association of these measures with the intention to call the MET and the level of MET utilisation was assessed using nonparametric correlation. Results obtained: The respiratory rate was missing in 20% of subjects. Using listwise deletion, the set of objective activation criteria investigated predicted an adverse event within 24 hours with a sensitivity of 55.4% (50.6-60.0%) and specificity of 93.7% (91.2-95.6%). An analysis approach that assumed the missing values would not have resulted in MET activation provided a sensitivity of 50.4% (45.7- 55.2%) and specificity of 93.3% (90.8-95.3%). Alternative models with modified cut-off values provided different results. The MET system was implemented with variable success during the MERIT study. Knowledge and understanding of the system, hospital readiness, and a positive attitude were all significantly positively associated with MET system utilisation, while defensive hospital cultures were negatively associated with the level of MET system utilisation. Major conclusions: The objective activation criteria studied have acceptable accuracy, but modification of the criteria may be considered. A satisfactory trade-off between the identification of patients at risk and workload requirements may be difficult to achieve. Measures of effectiveness of the implementation process may be associated with the level of MET system utilisation. Trials of the MET system should ensure good knowledge and understanding of the system, particularly amongst nursing staff.

This thesis explores triage code as a predictor of direct nursing care time, thus its potential usefulness in a model for calculating and allocating nurse requirements in emergency departments. A framework for nursing work is proposed. This framework is based on the works of O'Brien-Pallas, Irvine, Peereboom, and Murray (1997) and Houser (2003). It suggests that the structures of environmental complexity, nursing characteristics, patient nursing complexity, and patient medical condition and severity, impact on the processes of direct and indirect nursing care to affect patient outcomes. A prospective, non-experimental study was undertaken to examine the relationship between direct nursing care time and triage code. Six potential confounding variables were selected for this study: length of stay, age, ethnicity, sex, complaint type, and discharge category. Data were collected for 261 visits over a three day period in one New Zealand emergency department. Patient visits averaged 200 minutes. The mean direct nursing care time per visit was 49 minutes. On average, patients with more urgent triage codes, longer length of stay, or who were not discharged, received more direct nursing care. The model developed predicted 49% of variation in direct nursing care time (p < .05) related to triage code (16%), length of stay (31%) and disposition category (2%).Further exploration of the proposed framework has potential to develop a model allowing managers to identify nurse staffing required for optimal nursing care in emergency departments.

Problem investigated: The activation and implementation of the Medical Emergency Team (MET) system. Procedures followed: The ability of the objective activation criteria to accurately identify patients at risk of three serious adverse events (cardiac arrest, unexpected death and unplanned intensive care admission) was assessed using a nested, matched case-control study. Sensitivity, specificity and Receiver Operating Characteristic curve (ROC) analyses were performed. The MET implementation process was studied using two convenience sample surveys of the nursing staff from the general wards of twelve intervention hospitals. These surveys measured the awareness and understanding of the MET system, level of attendance at MET education sessions, knowledge of the activation criteria, level of intention to call the MET and overall attitude to the MET system, and the hospital level of support for change, hospital capability and hospital culture. The association of these measures with the intention to call the MET and the level of MET utilisation was assessed using nonparametric correlation. Results obtained: The respiratory rate was missing in 20% of subjects. Using listwise deletion, the set of objective activation criteria investigated predicted an adverse event within 24 hours with a sensitivity of 55.4% (50.6-60.0%) and specificity of 93.7% (91.2-95.6%). An analysis approach that assumed the missing values would not have resulted in MET activation provided a sensitivity of 50.4% (45.7- 55.2%) and specificity of 93.3% (90.8-95.3%). Alternative models with modified cut-off values provided different results. The MET system was implemented with variable success during the MERIT study. Knowledge and understanding of the system, hospital readiness, and a positive attitude were all significantly positively associated with MET system utilisation, while defensive hospital cultures were negatively associated with the level of MET system utilisation. Major conclusions: The objective activation criteria studied have acceptable accuracy, but modification of the criteria may be considered. A satisfactory trade-off between the identification of patients at risk and workload requirements may be difficult to achieve. Measures of effectiveness of the implementation process may be associated with the level of MET system utilisation. Trials of the MET system should ensure good knowledge and understanding of the system, particularly amongst nursing staff.

Objective: To explore the decision making approaches used by nurses who chose to work in critical care and factors influencing the process of decision making. Design and methods Qualitative descriptive methods using semi-structured interviews with nurses who had chosen to work in critical care within the past year. Results Ten nurses weighed staying in their current positions with moving to critical care. Two nurses considered two or more specialty units. The nurses used rational-intuitive and satisficing decision making approaches in making their employment decision. Limited knowledge and unrealistic expectations of work life in critical care, pressure from others, and inadequate professional support made nurses’ employment decisions more difficult. Social support, personal values for growth and learning, and paid educational incentives within employment offers facilitated their employment decisions. Conclusions Several factors negatively influenced nurses’ decision making approaches to making an employment decision. Findings revealed the need for decision support interventions focused on making employment choices for nurses.

Background: The neurological examination in critically ill patients is limited due to decreased level of consciousness and sedating medications. Electroencephalography (EEG) can be used to monitor brain injury; however, availability is limited. Methods: To determine the perceived need for EEG monitoring in the ICU and its current availability, we used rigorous methodology to develop and disseminate a survey to 199 Canadian critical care physicians. Results: Of 103 (52%) respondents (77% academic practice; 83% adult focus), 75% stated EEG monitoring should be a standard of care; yet, 75.5% were unable to obtain an EEG in an optimal timeframe. Technology under-use was exacerbated during non-standard working hours and greater in adult institutions. Perceived barriers to optimal care were lack of EEG technicians, physicians to interpret EEG and finances. Conclusion: Sub-optimal availability of EEG represents an important gap in the care of neurologically injured patients. Specific barriers represent targets for quality improvement.

Background: The neurological examination in critically ill patients is limited due to decreased level of consciousness and sedating medications. Electroencephalography (EEG) can be used to monitor brain injury; however, availability is limited. Methods: To determine the perceived need for EEG monitoring in the ICU and its current availability, we used rigorous methodology to develop and disseminate a survey to 199 Canadian critical care physicians. Results: Of 103 (52%) respondents (77% academic practice; 83% adult focus), 75% stated EEG monitoring should be a standard of care; yet, 75.5% were unable to obtain an EEG in an optimal timeframe. Technology under-use was exacerbated during non-standard working hours and greater in adult institutions. Perceived barriers to optimal care were lack of EEG technicians, physicians to interpret EEG and finances. Conclusion: Sub-optimal availability of EEG represents an important gap in the care of neurologically injured patients. Specific barriers represent targets for quality improvement.

Background: The neurological examination in critically ill patients is limited due to decreased level of consciousness and sedating medications. Electroencephalography (EEG) can be used to monitor brain injury; however, availability is limited. Methods: To determine the perceived need for EEG monitoring in the ICU and its current availability, we used rigorous methodology to develop and disseminate a survey to 199 Canadian critical care physicians. Results: Of 103 (52%) respondents (77% academic practice; 83% adult focus), 75% stated EEG monitoring should be a standard of care; yet, 75.5% were unable to obtain an EEG in an optimal timeframe. Technology under-use was exacerbated during non-standard working hours and greater in adult institutions. Perceived barriers to optimal care were lack of EEG technicians, physicians to interpret EEG and finances. Conclusion: Sub-optimal availability of EEG represents an important gap in the care of neurologically injured patients. Specific barriers represent targets for quality improvement.

Background: The neurological examination in critically ill patients is limited due to decreased level of consciousness and sedating medications. Electroencephalography (EEG) can be used to monitor brain injury; however, availability is limited. Methods: To determine the perceived need for EEG monitoring in the ICU and its current availability, we used rigorous methodology to develop and disseminate a survey to 199 Canadian critical care physicians. Results: Of 103 (52%) respondents (77% academic practice; 83% adult focus), 75% stated EEG monitoring should be a standard of care; yet, 75.5% were unable to obtain an EEG in an optimal timeframe. Technology under-use was exacerbated during non-standard working hours and greater in adult institutions. Perceived barriers to optimal care were lack of EEG technicians, physicians to interpret EEG and finances. Conclusion: Sub-optimal availability of EEG represents an important gap in the care of neurologically injured patients. Specific barriers represent targets for quality improvement.

Objective: To explore the decision making approaches used by nurses who chose to work in critical care and factors influencing the process of decision making. Design and methods Qualitative descriptive methods using semi-structured interviews with nurses who had chosen to work in critical care within the past year. Results Ten nurses weighed staying in their current positions with moving to critical care. Two nurses considered two or more specialty units. The nurses used rational-intuitive and satisficing decision making approaches in making their employment decision. Limited knowledge and unrealistic expectations of work life in critical care, pressure from others, and inadequate professional support made nurses’ employment decisions more difficult. Social support, personal values for growth and learning, and paid educational incentives within employment offers facilitated their employment decisions. Conclusions Several factors negatively influenced nurses’ decision making approaches to making an employment decision. Findings revealed the need for decision support interventions focused on making employment choices for nurses.

Introduction: A protocol for management of sedation and pain for mechanically ventilated patients at Flagstaff Medical Center (FMC) was implemented in August 2013. It was unknown whether the protocol is being adhered to or whether it has had an impact on patient outcomes. Objectives: To develop an audit and feedback mechanism to monitor adherence to sedation protocol at FMC and determine whether the protocol has impacted patient outcomes. Methods: A retrospective manual chart review was conducted including all mechanically ventilated adult patients for four, one-month periods: 1) pre-protocol; and 2) one month, 3) six months, and 4) 12 months post-protocol implementation. Results: 132 total patients were included (32 pre; 100 post-protocol). Mean weighted adherence score for post-protocol study groups were 5.0±0.6, 5.0±0.7, and 5.2±0.7 (p=0.926) out of ten. Time of mechanical ventilation (p=0.003) and hospital length of stay (LOS) (p=0.023) were reduced post (56±58h; 9.8±7.9days) vs. pre-protocol (90±67h; 13±7days). The adherence score was weakly correlated with hospital LOS but not time of mechanical ventilation. Conclusion: This project demonstrates improvements in patient outcomes from utilization of a sedation protocol. However, this project also highlights several challenges associated with the monitoring of protocol adherence. A lack of audit and feedback may be a factor in the observed unchanged adherence over time. Both research and monitoring activities are impaired by EHR systems that do not allow for the easy extraction of data. Ensuring that adequate audit and feedback strategies are designed and available prior to implementation of new protocols is an essential step in planning the implementation of a new protocol.