Identifying Strategies to Decrease ED Boarding

Keeping patients in the emergency department (ED) after an admission decision has been made, also known as ED boarding, leads to ED overcrowding. Boarded patients continue to use ED resources and prevent new patients from being seen. Boarding can also have adverse effects, such as higher mortality rates, longer inpatient stays, increased ED length of stay (LOS), and increased waiting room times. Factors that could contribute to longer boarding times include bed availability, hospital staffing, and delays in nurse-to-nurse report or physician consultation. The purpose of this project was to identify strategies to reduce the time a patient is kept in the emergency department after an admission decision has been made. The first objective of the project was to collect and analyze data on current boarding times in the ED at a local community hospital and compare them to the state and national average boarding times. The mean boarding time over the past 3 months at the local community hospital was 56 minutes. The median boarding times captured by Medicare were 47 minutes for the local hospital, 66 minutes for similar sized hospitals in Utah, and 88 minutes for similar sized hospitals throughout the nation. The second objective was to identify barriers in the admission process that could be causing delays in moving an admitted patient to the floor from the ED. For this objective, a data list of the time it takes to complete various admission tasks was created using the electronic medical record (EMR), call logs, and staffing sheets. An admission delay log was also filled out by ED nurses. Based on this log, the top reason for delays was when an on call nurse had to be called in prior to accepting the patient. The third objective was to develop recommendations based on the research findings. To do this, the data list was analyzed using two-tailed t tests, linear regression, random forest, and Pearson R to look for trends that could be increasing the boarding time, such as the consulting physician, certain hospital floors, times of the day, days of the week, or when a nurse was called in from home. Various lists were created to help understand and calculate the data. Staffing and census on the hospital floors and in the ED had no significant effects on boarding time. The closest correlation found using Pearson R was that of Telemetry and BHU census on boarding time, which had weak correlations with R=0.22299 and R=0.15116 respectively. Waiting for the on call nurse to come in created a statistically significant increase in mean boarding time, 72 minutes compared to 55 minutes when no nurse was called in. Using a two-tailed t test, this was a significant difference, p-value=0.00215. The random forest analysis showed that boarding time was very hard to predict, but the top three variables affecting boarding time were floor of admit, floor census, and if a nurse was called in. One of the main recommendations was to make changes that streamline nurse-to-nurse report. The final objective was to present recommendations to key hospital stakeholders, and disseminate project findings to a professional organization. Based on the project findings and recommendations, a presentation was made and presented to key hospital stakeholders. An abstract of the project was sent to the American College of Emergency Physicians. This project identified strategies to decrease ED boarding time. The findings of the project could be utilized to make changes that could prevent ED overcrowding through the reduction in boarding time, total length of stay and ED waiting room times. Making these changes could improve patient outcomes and reduce mortality.

Running head: STRATEGIES TO DECREASE ED BOARDING Identifying Strategies to Decrease ED Boarding Darrin VansCoy, BSN, RN, CEN, NREMT-P University of Utah In partial fulfillment of the requirements for the Doctor of Nursing Practice 1 STRATEGIES TO DECREASE ED BOARDING 2 Executive Summary Keeping patients in the emergency department (ED) after an admission decision has been made, also known as ED boarding, leads to ED overcrowding. Boarded patients continue to use ED resources and prevent new patients from being seen. Boarding can also have adverse effects, such as higher mortality rates, longer inpatient stays, increased ED length of stay (LOS), and increased waiting room times. Factors that could contribute to longer boarding times include bed availability, hospital staffing, and delays in nurse-to-nurse report or physician consultation. The purpose of this project was to identify strategies to reduce the time a patient is kept in the emergency department after an admission decision has been made. The first objective of the project was to collect and analyze data on current boarding times in the ED at a local community hospital and compare them to the state and national average boarding times. The mean boarding time over the past 3 months at the local community hospital was 56 minutes. The median boarding times captured by Medicare were 47 minutes for the local hospital, 66 minutes for similar sized hospitals in Utah, and 88 minutes for similar sized hospitals throughout the nation. The second objective was to identify barriers in the admission process that could be causing delays in moving an admitted patient to the floor from the ED. For this objective, a data list of the time it takes to complete various admission tasks was created using the electronic medical record (EMR), call logs, and staffing sheets. An admission delay log was also filled out by ED nurses. Based on this log, the top reason for delays was when an on call nurse had to be called in prior to accepting the patient. The third objective was to develop recommendations based on the research findings. To do this, the data list was analyzed using two-tailed t tests, linear regression, random forest, and Pearson R to look for trends that could be increasing the boarding time, such as the consulting physician, certain hospital floors, times of the day, days of the week, or when a nurse was called in from home. Various lists were created to help understand and calculate the data. Staffing and census on the hospital floors and in the ED had no significant effects on boarding time. The closest correlation found using Pearson R was that of Telemetry and BHU census on boarding time, which had weak correlations with R=0.22299 and R=0.15116 respectively. Waiting for the on call nurse to come in created a statistically significant increase in mean boarding time, 72 minutes compared to 55 minutes when no nurse was called in. Using a two-tailed t test, this was a significant difference, p-value=0.00215. The random forest analysis showed that boarding time was very hard to predict, but the top three variables affecting boarding time were floor of admit, floor census, and if a nurse was called in. One of the main recommendations was to make changes that streamline nurse-to-nurse report. The final objective was to present recommendations to key hospital stakeholders, and disseminate project findings to a professional organization. Based on the project findings and recommendations, a presentation was made and presented to key hospital stakeholders. An abstract of the project was sent to the American College of Emergency Physicians. This project identified strategies to decrease ED boarding time. The findings of the project could be utilized to make changes that could prevent ED overcrowding through the reduction in boarding time, total length of stay and ED waiting room times. Making these changes could improve patient outcomes and reduce mortality. The project committee was project chair Dr. Clint Child, program director Dr. Denise Ward, Assistant Dean Dr. Pam Hardin, and content experts Dr. Les Greenwood and Dr. Andrew Wilson. STRATEGIES TO DECREASE ED BOARDING 3 Table of Contents Problem Statement ...........................................................................................................................5 Clinical Significance ........................................................................................................................5 Purpose and Objectives ....................................................................................................................7 Literature Review.............................................................................................................................7 Theoretical Framework ..................................................................................................................11 Implementation and Evaluation .....................................................................................................13 Results ............................................................................................................................................20 Recommendations ..........................................................................................................................27 Recommendations for Future Projects ...........................................................................................32 DNP Essentials...............................................................................................................................33 Conclusions ....................................................................................................................................34 References ......................................................................................................................................35 Appendix A. DMAIC Theoretical Framework ..............................................................................40 Appendix B. Project Proposal PowerPoint ....................................................................................42 Appendix C. IRB Reliance Agreement ..........................................................................................46 Appendix D. IRB Exemption.........................................................................................................51 Appendix E. Admission Delay Log ...............................................................................................54 Appendix F. Abstract Submission .................................................................................................56 Appendix G. Various Tables and Images of Results .....................................................................60 Appendix H. Hospital Stakeholders PowerPoint Presentation ......................................................79 Appendix I. Project Poster .............................................................................................................88 STRATEGIES TO DECREASE ED BOARDING 4 Identifying Strategies to Decrease ED Boarding Problem Statement The number and the acuity of people seeking care at emergency departments (ED) is increasing throughout the United States. This overcrowding of emergency departments has caused an increase in wait times and length of stays, putting many patients at risk for adverse outcomes (Sun et al., 2013). A major contributing factor to ED overcrowding and increased length of stay is a delay in moving admitted patients to hospital floors. This trend, commonly referred to as ED boarding, can be due to a lack of inpatient beds, lack of nursing staff, or poor admission processes. Delays in moving patients to the hospital floor increases the length of stay of other patients and increases the time other patients spend in the waiting room (Kyriacou et al., 1999; White et al., 2013). Patients with increased ED length of stays have been found to have a higher risk of mortality and longer inpatient stays (McCusker et al., 2014; Sun et al., 2013). Clinical Significance Centers for Medicare and Medicaid Services (CMS), as well as the Joint Commission, are now tracking emergency departments' times and setting standards that encourage hospital corporations to make changes to reduce ED length of stay and to admit patients in a timely manner. These times are now being reported to the public on the CMS website, allowing people to compare the performance of neighboring hospitals. Overcrowding in the ED can be due to multiple factors, many of which affect one another. If the hospital floors are not able to receive a patient that is being admitted, the patient remains in the ED and takes up a bed. Boarding admitted patients in the ED prevents another patient from being treated, and increases the time patients spend in the waiting room. Boarded patients also STRATEGIES TO DECREASE ED BOARDING 5 continue to use the limited and expensive resources in the ED. Nurses and physicians are forced to continue to care for patients they have already made dispositions on, which takes time away from caring for the new ED patients who have yet to be evaluated and diagnosed. According to White et al. (2013), boarding patients in the ED increases the length of stay of the patients who are being discharged to home. Long ED boarding times can have major impacts on patient outcomes both in the ED itself and on inpatient floors. According to McCusker et al. (2014), having an increased amount of ED beds occupied raises the mortality rate of patients being cared for in the department. Another study by Sun et al. (2013) found that patients admitted to the hospital when the ED was at full capacity, had higher mortality rates and longer inpatient hospital stays. Singer et al. (2011) also showed that a patient with an increased boarding time in the ED had increased inpatient mortality and a longer hospital stay. Overcrowding in the ED increases the time patients spend in the waiting room. Some patients become discouraged by long wait times and leave the hospital prior to being seen. Patients with true emergency conditions who leave without being seen could be at risk for adverse outcomes, and could return to the ED at a later time in far worse condition. Kyriacou et al. (1999), found that reduced ED waiting times decreased the number of patients who left without being seen. The stakeholders of this project are the patients who visit the emergency department and the health care providers who care for them. Other stakeholders include the hospital administrative staff, who are often compelled to make changes to hospital operations to decrease boarding times, ED length of stay, and overall patient flow. STRATEGIES TO DECREASE ED BOARDING 6 At a local community hospital there are many potential areas where improvements in ED flow could be made. The major area this project focused on was identifying strategies that could expedite the process of moving the patient to the hospital floor after an admission decision has been made. Project Purpose The purpose of this project was to identify strategies to reduce ED boarding times and present the findings to hospital stakeholders. Objectives • Objective 1 - Collect and analyze data on current boarding times in the ED at a local community hospital. Collect data on national average boarding times and compare to the local community hospital. • Objective 2 - Identify barriers in the admission process that could be causing delays in moving an admitted patient to the floor from the ED. • Objective 3 - Develop recommendations based on research findings. • Objective 4 - Present recommendations to key hospital stakeholders, and disseminate project findings to a professional organization. Literature Review Background With the growing problem of overcrowding of emergency departments around the country, there is becoming an increased focus on finding ways to reduce ED waiting room times and total length of stay. This literature review will discuss processes that may be increasing the time patients spend in the ED, and the affects these increased times have on patient safety and outcomes. STRATEGIES TO DECREASE ED BOARDING 7 Databases used for this literature review included PubMed and CINAHL. Search terms used were ED boarding, emergency department crowding, and ED length of stay. Other appropriate articles were found from these searches by using the search for similar articles tool within these databases. When discussing times in the emergency department, there are three time periods that are usually tracked, waiting room time, total length of stay and the time between the admission decision and when the patient is transferred to the hospital floor (boarding time). All of these times can be affected by the processes of the emergency department, however the staff on hospital floors also have a large impact on ED times, and they may not be aware of it. All of these times are interconnected and influence one another. For example, a delay in an admission leads to increased ED length of stays, and causes other patients to wait longer in the waiting room. ED Waiting Room Time Increased waiting room times can cause patients to leave the ED prior to being cared for. Hospitals should be concerned about reducing waiting times to prevent lost revenue from patients who leave without being treated. According to Pines et al. (2011), hospitals have the potential to increase their revenue by $10,000 per day by reducing the boarding time of patients by one hour. This revenue comes from the ability to see more patients who would have normally left the ED or would have been taken to a different hospital by an EMS crew. ED bed availability. Waiting room time can be affected by ED bed availability. Factors that influence ED bed availability are how quickly treatments are being completed so a disposition can be made, and STRATEGIES TO DECREASE ED BOARDING 8 the time it takes admitted patients to be transferred up to a hospital floor. Staffing is one of the key aspects that impact the transfer of patients to the inpatient floor. Staffing. Low RN staffing on the inpatient side of the hospital can have an overall negative affect on ED length of stays. Most inpatient floors have set certain nurse-to-patient ratios, which allow for the safe care of patients. When the safe nurse-to-patient ratio is already maxed out on the hospital floor, the nurses on that floor often chose not to accept another patient from the ED until another nurse is available. These staffing related factors could cause a delay in a patient being moved to the hospital floor due to nurses being called in from home, the floor nurse's refusal to take additional patients, or the unavailability to give a nurse-to-nurse report. ED Total Length of Stay Reducing the time patients spend in the emergency department not only allows for more patients to be seen, but can also improve the care and outcomes of the patients who are being cared for. In a study conducted by Sun et al. (2013), patients who were seen in crowded emergency departments had a longer inpatient stay and a 5% increased risk of mortality during that hospital stay. Factors that can affect length of stay are disposition time and delayed discharge or admission. Delay in disposition. Delays in the disposition process could come from delays in drawing labs, delays in processing labs, delays in obtaining or reading imaging studies, delays in consultations with other physicians and excess workloads of ED physicians. One major aspect of disposition delays that could extend ED boarding time is delays in reaching consulting physicians. STRATEGIES TO DECREASE ED BOARDING 9 Physician consultation. Kang et al. (2014) researched ways to improve physician consultation processes. The authors found that adding an intermediary internal medicine physician, who acted as the sole person to consult with the ED physicians, reduced the ED length of stay of patients being admitted by 35%. Length of Time Between Admission and Transfer Increased ED boarding times can cause harm to patients. Chalfin et al. (2007) discuss that ED boarding times increase a patient's mortality and their inpatient length of stay. The authors found that patients boarded in the ED for 6 hours or more had a 5% increase in mortality and a one day longer inpatient stay than those who were transferred to the hospital floor in less than 6 hours. An additional risk to patients who are boarded in the ED for an extended period of time is delays in treatments ordered by the admitting physician. These treatments could either be delayed or completely missed all together. Coil et al. (2016) found that patients boarded in the ED for over 6 hours had five more orders completely missed than the patients who were transferred to the floor in less time. Bed availability. Boarding of patients in the ED can be due to lack of availability of rooms on the hospital floors. Increasing the availability of inpatient beds has the potential to decrease ED boarding times. Khanna et al. (2016) discuss ways to improve bed availability. The authors found that most ED admissions come in the afternoon and evening. If the hospital floors made it a priority to discharge patients in the morning, more hospital beds would be available later in the day during peak admission times. According to Khanna et al. (2016), discharging patients from the STRATEGIES TO DECREASE ED BOARDING 10 hospital floors by 11:00 am increased the bed availability by 9 beds, and also reduced the average inpatient length of stay. Rabin et al. (2012) also mentioned improving the discharge processes of the hospital floors as a way to improve emergency department flow. Rabin et al. (2012) focused on other ways to prevent ED boarding. The authors recommended moving patients to inpatient hallways rather than keeping them in the ED bed, and spreading out direct admission elective surgical patients so that more inpatient beds are available. Nurse to Nurse Report. Delays in nurse to nurse report can lengthen boarding times. Baker and Esbenshade (2015) recommend preventing delays in transferring patients through streamlining nurse to nurse report processes and changing the floor staff's mindset by showing them how they influence ED times. Rabin et al. (2012) also mentioned improving admission processes as a way to decrease boarding patients in the ED. Site Specific Review Patients who wait in the emergency department after an admission decision has been made become upset with their care and have made complaints to the hospital administration staff. Possible causes of longer boarding times at the local community hospital where the project will be carried out include, delays in physician consultation, delays in nurse-to-nurse report, and delays due to nurse staffing. This project identified what is contributing to delays in moving an admitted patient to the hospital floors so that changes can be made to improve the admission process and reduce boarding time. Theoretical Framework The framework that was used to guide this project was the DMAIC Model created by Bill Smith from Motorola. An illustration of this model can be found in Appendix A. DMAIC stands STRATEGIES TO DECREASE ED BOARDING 11 for define, measure, analyze, improve and control. DMAIC has been used by very large corporations, such as Motorola, to develop process improvements. This model has also been used as a guideline by healthcare corporations and other researchers to implement quality improvement changes (Momani et al., 2016). The researcher of this project completed the define, measure, analyze and a small section of the improve portion of this model. Due to the time constraints of the project, implementation and evaluation were not completed, which would correlate with the improve and control sections of the DMAIC Model. The researcher of this project is hopeful that the improve and control portions will be completed by the hospital stakeholders now that the findings of the project were presented to them. It is also possible that a future project could focus on completing the last two steps of the model. In the define portion of the model, processes are looked at, problems are identified, and goals for improvement are created (Thoennes, 2009). The define portion of the project was completed by observing the admission processes, and led to the idea of the project. Patient flow through the emergency department was observed and a problem of long boarding times was identified. A goal was created by the hospital to reduce the ED boarding time. This project was created based on the define process, and focused on a potential area for improvement, reducing ED boarding. In the measure section of this model, data is collected about what is currently happening (Thoennes, 2009). This portion of the project looked at the various aspects of the admission process and collected data regarding the time it took to complete the admission processes and searched for other factors that may be causing delays. Data collection included compiling a list of the times it took to complete the various admission processes. STRATEGIES TO DECREASE ED BOARDING 12 During the analyze section of this model, the data that was collected was analyzed to find possible relationships between the identified variables. This stage included analyzing the data to look for trends and identifying relationships. The researcher of the project looked for possible relationships of the multiple variables of the admission process, and found certain aspects that are linked to longer ED boarding times. During the improve section of the model, a plan for implementing changes is developed, and changes are implemented (Thoennes, 2009). Part of the improve process is the development of a list of potential improvements. The final goal of this project was to create a list of possible changes and present the findings to the hospital stakeholders. This project did not actually implement a change, and only completed the first part of the improve portion of the DMAIC Model. Recommended changes were presented to the medical facility staff, and it is hoped that they implement the changes that were recommended. In the control stage of the model, the changes are monitored for their effectiveness (Thoennes, 2009). Part of this section also includes reanalysis of the current process for any changes that will need to be made in the future, and the need to start the process over and define a new problem. The control aspect of the DMAIC model is a way to evaluate a change that has been implemented. This project did not implement a change; therefore, the researcher of the project did not complete the control portion of the model. Implementation and Evaluation The researcher of the project passed the DNP project proposal presentation. This presentation is listed in Appendix B. The researcher of the project submitted a review of the project to the University of Utah IRB. A reliance agreement between the local community hospital and the University of Utah IRB was completed, allowing the University of Utah IRB to STRATEGIES TO DECREASE ED BOARDING 13 be used as the sole IRB for the research project (Appendix C). After IRB review, the project received an IRB exemption and was granted authority to begin the research project (Appendix D). Objective 1 The first objective was to collect and analyze data on current boarding times in the ED at a local community hospital, collect data on national average boarding times, and make comparisons. The electronic medical record (EMR) at a local community hospital was used to collect data on the boarding time and length of stay of every patient admitted to the hospital from the emergency department over three months. With the help of the local hospital's IT staff and the hospital corporation staff, a report was generated using the EMR, which included a list of all patients admitted over those three months. This list included various times throughout a patient's ED visit including time of arrival, time of admission decision, and time the patient left the ED. This list was imported to a spreadsheet and the boarding time for each patient was found by calculating the difference between the time the admission decision was made and the time the patient left the ED. Over 4,000 hospitals certified by Medicare have their boarding times tracked and listed on the Hospital Compare website (Medicare, 2017a). The Medicare Hospital Compare website was utilized to find the median boarding times for the local hospital where the project took place, similar sized hospitals throughout Utah, and similar sized hospitals throughout the nation. The Medicare Hospital Compare website was accessed and the local hospital profile was found. The boarding time of the ED was found within the emergency department care section. In this section, the local hospital boarding time is listed as well as the median boarding time of similar volume hospitals in Utah and the Nation. STRATEGIES TO DECREASE ED BOARDING 14 This objective was successful by compiling a data list that included the boarding times of all admitted patients over the past 3 months, and by finding the median boarding times of similar hospitals throughout the country to compare to the local community hospital's times. Objective 2 The second objective was to identify barriers in the admission process that could be causing delays in moving an admitted patient to the floor from the ED. The researcher used the report extracted from the EMR, the ED clerk's call log, the nursing supervisor's daily staffing sheet, and the ED daily staffing sheet to compile a list of data that could be used to make comparisons and find trends. The data list included information on 928 patients. This was every patient admitted to the hospital from the ED over the months of October through December 2016. No patients were excluded. If the information was not relevant or could not be found for a certain patient that field in the data list was left blank. The data list did not include patients transferred to other hospitals for admission or patients admitted from standalone EDs. Headings created on the data list included patient name, date of admit, day of the week, time of arrival, time placed in room, waiting room time, time of admission decision, time left ED, total length of stay, time crisis worker was called, time consulting physician was called, crisis consult time, time to obtain consult, boarding time, name of admitting physician, ED physician, floor of admit, ED staff hours per patient, ED census, floor staff, floor census, and if a nurse was called in. It was necessary to keep patient names within the data list until the physician consult and crisis worker consult information was inputted, as this was how the information was located on the call log. After this information was inputted, the patient names were deleted from the data list. STRATEGIES TO DECREASE ED BOARDING 15 Other than the hospital's IT staff who generated the original list, the researcher of this project was the only person who had access to the data list information while patient identifiers were present. The patient list extracted from the EMR included the patient name, date and many times such as arrival, time placed in room, time of admission decision and time the patient left the ED. The extracted list also included name of the ED physician, the admitting physician, and the floor the patient was admitted to. A calendar was used to input the day of the week based on the date of admission. The waiting room time was calculated by finding the difference between the time the patient was placed in the room and the time of arrival. The total length of stay was calculated by finding the difference between the time the patient left the ED and the time of arrival. The boarding time was calculated by finding the difference between the time the patient left the ED and the time the admission decision was made. The ED clerk's call log was used to find the time a crisis worker was called by searching for that patient's name during the date and time the patient was seen in the ED. The total crisis consult time was calculated using the difference between the time of admission decision and the time a crisis worker was called. Only behavioral health patients being admitted to the behavioral health floor were used to find the crisis consult time. Consult times for the admitting physician to the behavioral health unit were not available as the crisis worker initiates this contact and it is not captured in the ED clerk's call log. Also, the majority of the patients admitted to the behavioral health floor did not have the time the crisis worker was called captured in the ED clerk's call log. In this case the field was left blank. The time the consulting or admitting physician was called was found on the ED clerk's call log by searching the patient's name during the date and time frame they were admitted. The total consult time was calculated by finding the difference between the time of admission decision and STRATEGIES TO DECREASE ED BOARDING 16 the time the consulting physician was called. If multiple physicians were called for a certain patient, only the call time of the admitting physician was used. The ED census and ED staff hours per patient was inputted using the daily staffing sheets from the ED. Staffing within the ED at this hospital is monitored using the total hours all ED staff worked that day divided by the number of patients seen that day. The majority of the daily staffing sheets had this calculation already listed, as this is completed by the charge nurse or the ED director. If this was not already listed, the hours per patient was calculated. The hours per patient and the census for the day were listed for each patient that was admitted that day, so that everyone admitted on the same day had the same hours per patient and ED census listed. The floor staff and floor census was inputted using the nursing supervisor's daily staffing sheets. These sheets included a list of nurses working on each floor, and the number of patients on each floor during four-hour intervals throughout the day. The number of nurses working on the floor at the time the patient was admitted was inputted for each patient. The number of patients on the specific floor the patient was being admitted to, and within the closest timeframe listed prior to the time the patient left the ED, was inputted into the floor census section. The times a nurse was called in from home were found on the nursing supervisors daily staffing sheet. On the staffing sheet, the time a nurse was called in was written next to the nurse's name that was on call for that day. If this on call time correlated with the specific floor and the time the patient left the ED, then it was counted as a time a nurse was called in due to an admission. The researcher also obtained information about the common reasons for delays in nurse to nurse reporting by creating an admission delay log that the ED nurses filled out during 2 months of the project. After 2 months of data collection, a data list was created. Headings for the data list included the reason for delay, the floor of admission, the time the patient could have STRATEGIES TO DECREASE ED BOARDING 17 been transferred up to the floor, the actual time patient left the ED, and the time of delay. This information was written down by the ED nurse and was not taken from the actual patient chart. The delay time was calculated by finding the difference between the time the patient left the ED and the time the patient could have been transferred to the floor. An example of the delay log can be found in Appendix E. This objective was successful after a data list was created that included the common reasons for delays in moving an admitted patient to the floor from an admission delay log, and a data list was compiled that included the various times throughout the patients ED stay, consult times, and information about floor and ED staffing and census. Objective 3 The third objective was to develop recommendations based on the research findings. After the data lists were compiled, the lists were analyzed to look for any trends. Parts of the data list were separated into additional spreadsheets for closer analysis of physician consult times, boarding times, and total length of stay. After the lists were created, the calculation tools within the Excel program were used to find the mean and median times and to rank the lists in order. StatPlus within the Excel program was used to calculate linear regression, Pearson R, and two-tailed t tests comparing boarding times based on floor census, floor staffing, when a nurse was called in from home, and shift change. After simple analysis to compare the previously listed variables, the researcher of the project sought to gain a better understanding of the data set as a whole. A statistician was consulted, and after reviewing the data list he determined a random forest approach would be best to analyze the data set. This was chosen due to the high correlation between the multiple variables within the data set, which would have likely skewed the results if traditional models STRATEGIES TO DECREASE ED BOARDING 18 were used. First, the data was cleaned and formatted using SAS 9.4. Case analysis was completed and all cases with missing information were removed, as less than 1% of the cases had missing information. The data was then exported and a random forest analysis was run using R version 3.3.2. Candidate predictors were found and ranked by relative importance based on the percent increased mean squared error. Using the statistical program R version 3.3.2, the generalized linear model was then used to find the partial effect plots. Based on the research findings and the information obtained from the literature review, the researcher of this project identified areas where processes could be improved and compiled a list of recommended changes. These are listed in the recommendations section below. This objective was successfully completed by creating a list of trends from the data analysis, a list of possible improvements from the literature review, and developing a list of recommendations. Objective 4 The fourth objective was to present recommendations to key hospital stakeholders, and disseminate project findings to a professional organization. The researcher of the project organized a meeting and disseminated the project findings, in the form of a face-to-face presentation, to key hospital stakeholders during a transfer committee meeting on March 13, 2017. This committee was formed soon after this research project began. The committee has a goal to find ways to improve admission processes and reduce boarding times. The committee includes the directors of the ED, admitting, hospital floors, as well as the trauma, stroke and EMS coordinators. The CNO of the hospital also attended the meeting. STRATEGIES TO DECREASE ED BOARDING 19 The project findings were also disseminated to a professional organization. An abstract was submitted to the American College of Emergency Physicians (ACEP). A copy of this abstract submission can be found in Appendix F. This objective was successfully completed once the professional organization received the abstract of the project, and the project findings were presented to the key hospital stakeholders. Results Objective 1 The Centers for Medicare and Medicaid Services collect and publish the times it takes hospitals to complete various tasks. ED boarding time is one of the times collected and published on the Medicare Hospital Compare website. According to the Medicare website the current times that were collected for this project were developed using data from April 1, 2015 through March 31, 2016 (Medicare, 2017b). This data is submitted by hospitals throughout the nation to the Centers for Medicare and Medicaid Services through the CMS Abstraction and Reporting Tool (Medicare, 2017c). The following times were collected from the Medicare Hospital Compare website. The local hospital's median boarding time was 47 minutes. Medicare also listed the boarding times of hospitals within Utah and across the nation that see similar amount of ED patients per year (between 20,000 and 39,999 patients annually). The median boarding time of similar hospitals in Utah was 66 minutes. The median boarding time of similar hospitals throughout the nation was 88 minutes. Various tables and lists of results can be found in Appendix G. STRATEGIES TO DECREASE ED BOARDING 20 Using the Excel program, the boarding time column of the data list was used to calculate the mean and median boarding times of the 928 patients. The boarding time for all patients admitted over the three months had a mean of 56 minutes and a median of 51 minutes. The median boarding time listed on the Medicare site was similar to the median time found on the data list for patients seen during the three months, 47 minutes versus 51 minutes respectively. Based on the data from Medicare, the local community hospital has a median ED boarding time that is 41 minutes less than the national average of similar hospitals and 19 minutes less than the Utah average of similar hospitals. The researcher believed a limitation of using the Medicare site would be that hospitals submitted data to Medicare that was not accurate. The similarity of the median boarding time found from the data list and from the Medicare site helped confirm the reliability of the Medicare data. An unintended consequence was finding that the local community hospital's boarding time was already below the national average. The project had a goal of reducing the boarding time, however, what the hospital is currently doing to reduce boarding time is already better than the majority of hospitals throughout the country. Objective 2 A data list was created and a list of common reasons for delays was obtained. Key facilitators in creation of the data list included the ability to obtain a report from the EMR system. Barriers that led to increased time consuming data inputting were that the day of the week, total length of stay, and total boarding time were not included in this report and had to be calculated or inputted. Barriers in inputting the ED staffing hours per patient were that some dates did not STRATEGIES TO DECREASE ED BOARDING 21 have the calculation listed on the staffing sheet. In these cases, the total hours worked per patients seen had to be calculated prior to inputting the information into the data list. Limitations included missing information from the data set. Much of the missing information came from the ED clerk's call logs. The time the admitting physician was called was captured in the call log on the majority of the patients, 766 to be exact. However, there were 162 patient admissions that did not have consult times written in the call log. This made obtaining the consult time for that patient impossible due to missing data. This was especially true for patients admitted to the behavioral health unit. In this instance, the section was left blank for those patients. The results would have been more comprehensive if all 928 admissions had the consult times captured, however the sample size of 766 was large enough to make generalizations regarding consult times. It was found that the crisis worker is the one who initiates the physician consultation for the behavioral health patients and none of these patients' consult times were captured in the ED clerk's call log. After this finding, the researcher was able to capture the time it took to consult and complete a crisis evaluation, however it was limited data as only 36 patients had the time a crisis worker was called in the ED clerk's call log. Another limitation involved the data collected with the hospital floor census. The number of patients on the floor was written on the supervisor's staffing log in four-hour intervals. The true census of the floor at the exact time of admission was not available. The number of patients listed in the time frame prior to the admission was used. An unexpected finding with consult time was that some physicians came to see the patient in the ED after being called, for example, orthopedic surgeons or GI specialists. These STRATEGIES TO DECREASE ED BOARDING 22 physicians may have called back and talked with the ED physician right away, however if they came to the ED and lengthened the admission decision time, then the consult time was increased. A limitation of the ED delay log was that there was a delay in starting the log due to the extended time it took to complete the IRB process. The researcher of the project planned to start the delay log in December, but was not able to start it until January. Two months of data were collected, from January 1, 2017 to February 28, 2017. The information from the delay log was used to create a data list within the Excel program. The data was then analyzed and placed into tables and diagrams. Some of these included common reasons for delays, how many delays each floor had, the mean time in minutes of the delay and the top reasons for the delay. The findings from the delay log showed that patients being admitted to the ICU floor had the longest boarding times with a mean of 87 min. The ICU floor also had the most number of delays, with 20 delays reported over the period of two months. Some of the top reasons for delays in moving the patient to the floor were calling in a nurse, waiting for physician orders, not having an available bed, having to transfer a patient off the floor first and having to clean a room. Another barrier with the ED delay log was that it relied on the ED nurses to fill it out when they felt a delay happened. Although information about the ED log was provided at the staff meetings and emailed to the nurses, when discussing it with nurses face-to-face, they often didn't know where to find the log and had not been filling it out. A limitation was that there were likely many more delays that went uncaptured due to non-compliance in filling out the log. Objective 3 The sort and filter settings within Excel were used to create various lists to help analyze and understand the data. Calculation tools were then used within the Excel program to obtain median and mean times. The lists created included: STRATEGIES TO DECREASE ED BOARDING • • • • • • • • 23 The mean and median boarding time of all hospital floors The mean boarding times based on the days of the week The mean and median consult times of admitting physicians The mean and median consult times of ED physicians The mean and median boarding times of admitting physicians The mean and median boarding times of ED physicians The mean and median length of stay of ED physicians The mean consult and boarding time during shift change These findings showed that some hospital floors have longer boarding times than others. For example, the ICU floor had the longest mean boarding times based on the data list at 67 min. As discussed previously, the delay log findings provided further confirmation that the ICU floor had the most delays and the longest boarding times when compared to other hospital floors. The day of the week had very little effect on boarding time. Consult and boarding times varied based on the ED physician and the admitting physician. For example, mean consult times of admitting physicians ranged from 6 to 58 minutes, showing that physicians could play a major role in reducing consult and boarding times. Using the StatPlus feature within the Excel program, a two-tailed t test was used to compare the mean boarding time during shift change compared to not during shift change. Shift change was considered 4:30 am - 6:30 am and 4:30 pm - 6:30 pm. The findings were a mean of 59 minutes during shift change compared to a mean of 55 minutes during times other than shift change. This finding was not significant with a P value of 0.17639. A two-tailed t test was also used to compare consult time during shift change compared to not during shift change. The findings were a mean consult time of 21 minutes during shift change and a mean of 19 minutes not during shift change. This finding was not significant with a P value of 0.41353. STRATEGIES TO DECREASE ED BOARDING 24 Using the StatPlus feature within the Excel program, the data was analyzed with linear regression and Pearson R to gain an understanding of the relationship between boarding time and the variables. These variables included: • • • • • • • • • • • ED staff effect on boarding time. No correlation, Pearson R 0.06239 ED census effect on boarding time. No correlation, Pearson R -0.02702 ICU staff effect on boarding time of patients being admitted to that floor. No correlation, Pearson R 0.06846 ICU census effect on boarding time of patients being admitted to that floor. No correlation, Pearson R 0.06944 Medical staff effect on boarding time of patients being admitted to that floor. No correlation, Pearson R 0.00847 Medical census effect on boarding time of patients being admitted to that floor. No correlation, Pearson R -0.01144 Surgical/Pediatrics staff effect on boarding time of patients being admitted to that floor. No correlation, Pearson R 0.02488 Surgical/Pediatrics census effect on boarding time of patients being admitted to that floor. No correlation, Pearson R 0.06507 Telemetry staff effect on boarding time of patients being admitted to that floor. Weak positive correlation, Pearson R 0.11484 Telemetry census effect on boarding time of patients being admitted to that floor. Weak positive correlation, Pearson R 0.22299 Behavioral Health Unit census effect on boarding time of patients being admitted to that floor. Weak positive correlation, Pearson R 0.15116 The findings included weak positive correlations between boarding times and telemetry staff, telemetry census, and behavioral health unit census. As the patient census on telemetry and behavioral health floors increased, boarding time also increased. However, these were weak correlations, as a perfect correlation would have a Pearson R nearing 1.0. As listed above the highest Pearson R found was that of Telemetry census at R = 0.22299. Overall, the patient census and amount of nursing staff in the ED and on the hospital floors had very little effect on boarding times. STRATEGIES TO DECREASE ED BOARDING 25 Again, using the StatPlus feature within Excel. Using a two-tailed t test there was a significant difference when comparing the mean boarding time when a nurse was not called in (55 minutes) versus the mean boarding time when a nurse was called in (72 minutes), with a P value of 0.00215. Waiting for the on call nurse to arrival prior to accepting report and transferring the patient to the hospital floor causes significant delays and lengthens boarding time. With the help of a statistician, the data list was analyzed using random forest analysis with the statistical program R version 3.3.2. The three most important candidate predictors of boarding time were floor of admit, floor census, and if a nurse was called in. Once the three main predictors were found, a generalized linear model was used to find the partial effects. Although the floor of admit, floor census and nurse called in were statistically significant, these were not likely predictable. The partial effects could only explain 9.5% of the variation of boarding times. Random forest analysis found that boarding time is extremely hard to predict, and there is much variation in boarding times that is left unexplained by the variables within the data set. Capturing other reasons that could be contributing to increased boarding times, such as those found in the admission delay log, could help explain the other variables that affect boarding time. To improve the results of the random forest analysis, the information in the data set could have been recategorized into more condensed groups. For example, the statistical program was unable to analyze the long list of physicians. Possibly recategorizing the physicians into groups, such as hospitalist, intensivist, orthopedics, etc. would have shortened the list of physicians and allowed for the program to include physicians as one of the variables within the calculation. Using information obtained from the literature review and the findings from the analysis of the data, a list of recommendations was successfully created. These recommendations are discussed in detail in the recommendations section below. STRATEGIES TO DECREASE ED BOARDING 26 An unintended positive consequence of the data analysis included the finding of long delays in crisis consultation time. Objective 4 The recommendations were successfully presented to key hospital stakeholders in the form of a face-to-face presentation. A copy of this presentation can be found in Appendix H. An abstract of the project was successfully submitted to the American College of Emergency Physicians for possible dissemination at their scientific assembly from October 29 - November 1, 2017 in Washington DC. A project poster was created and a poster presentation was presented at the University of Utah College of Nursing on March 31, 2017. A copy of this project poster can be found in appendix I. A barrier to submission of an abstract was that many of the professional organizations had abstract submission deadlines of January or February. The researcher intended to submit an abstract to the Emergency Nurses Association for the possibility of presenting a poster at the upcoming conference. Although the conference is not until September 2017, the abstract deadline was in January. Recommendations The community hospital where this project took place has a median boarding time that is less than the state and national average. A reasonable option would be to continue with the current practices without making any changes. If improvements are desired, the below recommendations could be utilized to further reduce boarding time. Based on the literature review and the research project findings, recommendations are to create changes that reduce the length of stay of behavioral health patients, improve physician consultation, implement changes to streamline nurse-to-nurse reporting, improve bed availability STRATEGIES TO DECREASE ED BOARDING 27 by prioritizing discharges and transfers to happen in the morning, and implement further changes to prevent delays in transferring a patient to the hospital floor. Implement Strategies to Reduce Length of Stay of Behavioral Health Patients Length of stays and boarding times of behavioral health patients are longer than other patients due to the need for crisis evaluation and the difficulty in finding inpatient beds. This is especially true when the patient is being transferred to another facility. The data obtained during this project was regarding behavioral patients who were admitted to the same hospital, to an available behavioral health bed. The average ED length of stay for a behavioral health patient being admitted to the hospital was 5 hours and 31 minutes. The national average ED length of stay for a behavioral health patient being admitted is 15 hours (Zeller, 2015). Although the hospital the project took place already has a substantially shorter total length of stay for behavioral health patients than the national average, the time between calling a crisis worker and disposition of the patients could be reduced. The average time from calling the crisis worker to an admission decision was 3 hours and 55 minutes. This time was longer than the average length of stay of all other patients in the ED other than behavioral health patients, which was 3 hours and 39 minutes. Crisis workers are often at other facilities completing crisis evaluations or have multiple patients who require crisis evaluations in the ED at the same time. Goals should be set to find ways to reduce the time it takes for the crisis worker to arrive and complete a crisis evaluation. A recommendation would be to increase the availability of the crisis worker in the ED. This could be achieved by hiring additional crisis workers that could be located in the ED 24 hours a day or at least during peak times. Hiring more staff would have additional costs. The facility would need to analyze whether these additional costs would be acceptable for the possible improvements in patient care and safety. Improving the time to obtain a crisis evaluation STRATEGIES TO DECREASE ED BOARDING 28 could reduce ED overcrowding and increase ED bed availability through reduction in length of stay. Preventing ED overcrowding should be a priority of every ED, as many sources report higher mortality rates of patients who are seen in an overcrowded ED (McKusker et al., 2014, Sun et al., 2013, Singer et al., 2011). Implement Policies That Streamline Physician Consultation One of the main reasons for delays in consult time is waiting for the admitting physician to call back. Consult times could be decreased if the ED physician is able to immediately talk with the admitting physician. A recommendation would be to require admitting physicians to use cell phones rather than pagers to help prevent the waiting for a call back. The findings showed that certain physicians had longer consult times than others. A recommendation would be to talk with specific physicians who have longer consult times and set a standard call back time that they are expected to meet. Consult times were longer when the orthopedic specialist came to see the patient in the ED. If possible, have orthopedic physicians make admit decisions over the phone rather than waiting for them to assess the patient in the ED. Implement Policies Which Reduce Excuses and Streamline Nurse-to-Nurse Reporting The admission delay log showed that most of the delays in nurse reporting were due to floor staff refusing to take report. Educate nursing staff on the hospital floors about the goal of reducing ED boarding times and the important role they play in reducing this time. Nursing staff on the hospital floors may not understand how refusing to take report on one patient affects the other patients in the ED. Nurses on the hospital floors should be educated about increased boarding times and how it plays a roll in increased patient mortality, reduction of patient satisfaction, increased ED waiting room times, and increased number of patients who leave STRATEGIES TO DECREASE ED BOARDING 29 without being seen. The floor nurses must have a mindset of wanting to accept and take care of their newly admitted patient as soon as possible, rather than trying to avoid taking a new admission. One of the common reasons for delays from the delay log was that nurses on the behavioral health unit and obstetrics floors would not accept report until they had received orders from the admitting physician. A recommendation would be to allow patients to be admitted without waiting for admitting physician orders. This could be done by creating a basic protocol of orders for nurses to complete for every behavioral health and obstetrics admit so that the patient can be moved to the floor prior to the RN receiving orders from the admitting physician. The number one reason for delay from the delay log was waiting for a nurse to be called in from home. This was also found to cause a statistically significant increase in boarding time. A solution to this problem could be to have the current floor staff care for the new patient while the on call nurse is en route, or to staff an additional nurse to help coordinate admissions. This admissions nurse could go to the ED when there is an admission and help initiate the transfer to the floor, help settle the patient in their new room, and start the admitting physician's orders while the on call nurse is coming in from home. Even when a nurse would not need to be called in, this position would help smooth the transition from ED to the floor by helping floor staff transfer the patient and get them settled in their new room. This position could also be utilized as a lunch break nurse in the down time between admissions. This extra position would increase the amount of money spent on nursing hours, and the facility would need to analyze whether these increased costs are worth the potential benefits. Increasing nursing staff in hospitals has been found to improve patient safety, reduce mortality, reduce medical errors, improve nurse retention, improve patient satisfaction and reduce hospital readmissions (American Nurses Association, STRATEGIES TO DECREASE ED BOARDING 30 2014). If the improvements in patient safety, satisfaction, and outcomes are not enough incentive, the potential saving of improving nurse retention should be explored. Nursing Solutions Incorporated conducted a survey of RN staffing retention for the year of 2015. They found that "the average turnover rate for bedside nurses was 17.2% and the cost to turnover one nurse was between $37,700 to $58,400" (Nursing Solutions Incorporated, 2016, p. 1). Another common delay was that a room needed to be cleaned before the patient could be admitted. A recommendation would be to always have a clean room on every floor. Rather than cleaning rooms just prior to an admission, rooms should be cleaned right after a patient is discharged. Many of the delays in nurse-to-nurse reporting are possible because the floor nursing staff can refuse to take report. A recommendation would be to set a policy that eliminates the ability to refuse to accept report. Implement a one call report policy. After a bed on the floor has been assigned, the ED nurse attempts to call the floor to give report one time. After this, the patient is taken to the floor and bedside report can be given. Increase Bed Availability The admission delay log showed that some patients had delays in being admitted due to beds not being available, or the need to transfer a patient from one floor to another. To prevent these delays and increase bed availability, prioritize patient discharges and patient transfers to other floors to be completed first thing in the morning. The ED is busiest in the afternoon and evening, with most admits happening at that time. Over the three months, 582 admits occurred between noon and midnight compared to 346 admits between midnight and noon. If discharges and transfers were completed in the morning, there would be more bed availability for admissions in the afternoon and evenings. STRATEGIES TO DECREASE ED BOARDING 31 Implement Further Changes That Will Prevent Transfer Delays Another common reason for delay in transferring the patient to the floor was that the staff in the ED was too busy to transfer the patient. If hiring an admissions nurse isn't a feasible option, then hiring a patient transfer technician to transfer patients from the ED to the hospital floors should be considered. When not transferring a patient to the floors, this technician could be utilized in triage to help transfer patients back to the ED rooms and clean ED rooms. Not only will this help reduce boarding times, it could also reduce waiting room times, the number of patients who leave without being seen, and door to provider time. Once again, the facility would need to analyze if the cost of hiring an additional staff member is worth the potential improved patient satisfaction, patient outcomes and the possible reduction of patients who leave without being seen. For example, Pines et al. (2011), estimates $1,096 in revenue is lost from one patient who leaves without being seen. The median average wage for emergency medical technicians listed by the Bureau of Labor Statistics (2015) is $15.38. This would have a daily cost $369.12 to staff a transfer technician 24 hours a day. While analyzing the call logs, it was found that some patients were being kept in the emergency department unnecessarily. If a patient is going to have a surgery and will be admitted after the surgery, have the patient transferred to the floor to wait for the surgeon and surgical team to come in. Surgeons often request to have the patient wait in the ED. This patient continues to take up an ED bed, use ED resources and could be lengthening the stay of other ED patients. Recommendations for Future Projects The hospital's newly created transfer committee could take over the project from here. Recommendations regarding the future of this project are to make the recommended changes and STRATEGIES TO DECREASE ED BOARDING 32 reevaluate the boarding time and any other data of interest after the changes have been implemented. To evaluate if the changes were successful, repeat this same project, and compare the findings to the original project. For further information, evaluate if the changes made to reduce boarding time affected other times within the ED stay. For example, did the changes reduce total length of stay, waiting room time, and the number of patients who leave without being seen? This project could be applied to other areas, such as identifying ways to reduce ED total length of stay. A project could be developed to analyze the time it takes to complete tasks within the ED such as triage, blood draws, lab processing, imaging completion, and radiology reads. After obtaining the baseline times and analyzing the processes, implement changes and reevaluate. This project could also be repeated in emergency departments in different medical facilities throughout the country. The recommended changes could be implemented in other hospitals if they find their facility has some of the same barriers. An additional research project could be to implement a patient transfer technician in the ED that can work in triage and transfer admitted patients to the floor as needed. Look at total cost of EMT coverage vs the total amount of lost revenue from patients who leave without being seen. Obtain average times for waiting room, door to physician, total length of stay, boarding time and the number of patient who leave without being seen before and after implementation and compare. Another research project could be to implement an admissions or transfer RN to help coordinate admissions, and prevent the delay in waiting for the on call nurse to arrive prior to transferring the patient to the floor. Comparisons that could be made before and after STRATEGIES TO DECREASE ED BOARDING 33 implementation could include patient and nurse satisfaction, nurse retention and turnover rate, and patient complaint rates. It would also be beneficial to compute the annual extra cost of the additional nursing staff and compare that to the annual savings from improved nurse retention. DNP Essentials Two of the DNP essentials that this project addressed were Essential II and VI. Essential II is organizational and systems leadership for quality improvement and systems thinking (AACN, 2016). One of the main focuses under this essential is making quality improvements to improve patient safety (AACN, 2016). This project developed recommendations that have the potential to decrease ED boarding time. As discussed in detail in the literature review section, reducing boarding times helps improve patient safety by decreasing adverse outcomes, mortality, and inpatient stays. Essential VI is interprofessional collaboration for improving patient and population health outcomes (AACN, 2016). This essential has a focus on working as a healthcare team to improve the care a patient receives (AACN, 2016). This project helped promote interprofessional collaboration between ED staff, floor staff nurses, inpatient physicians, and hospital administration. Implementing the recommendations will improve communication and teamwork between hospital floors and the ED. Conclusion In summary, the ED times that are usually considered are waiting time, length of stay and boarding time. All of these times affect one another and the processes on the hospital floor and in the ED both have an influence on the times. STRATEGIES TO DECREASE ED BOARDING 34 Much of the research has a focus on boarding patients in the ED. The literature review supports the fact that increased boarding time leads to ED overcrowding and increases the risk for adverse patient outcomes. The project found many things that could be contributing to delays in moving the admitted patient to the floor. Some common reasons for delays were, delays in nurse-to-nurse reporting, no bed availability, and increased physician consult times. Waiting to give report while a nurse was called in from home created a significant delay in moving the admitted patient to the floor. Possible ways to reduce ED boarding times are to implement changes to reduce the length of stay of behavioral health patients, improve physician consultation, implement changes to streamline nurse-to-nurse reporting, and improve bed availability by prioritizing discharges and transfers to happen in the morning. This project identified strategies to decrease ED boarding times through the improvement of admission processes. A project which promotes the reduction of ED boarding times should also reduce waiting room times and total length of stay. This creates the potential to improve the safety, outcomes and satisfaction of patients who seek care in the emergency department. STRATEGIES TO DECREASE ED BOARDING 35 References American Nurses Association. (2014). Safe staffing literature review. Retrieved from www.nursingworld.org/2014-NurseStaffing-UpdatedLiteratureReview Baker, S. J., & Esbenshade, A. (2015). Partnering effectively with inpatient leaders for improved emergency department throughput. Advanced Emergency Nursing Journal, 37(1), 65-71. doi:10.1097/TME.0000000000000050 Bernstein, S., Aronsky, D., Duseja, R., Epstein, S., Handel, D., Hwang, U., & ... Asplin, B. (2009). The effect of emergency department crowding on clinically oriented outcomes. Academic Emergency Medicine, 16(1), 1-10. doi:10.1111/j.1553-2712.2008.00295.x Bureau of Labor Statistics. (2015). Occupational employment statistics. Retrieved from https://www.bls.gov/oes/current/oes292041.htm#st Chan, T., Killeen, J., Kelly, D., & Guss, D. (2005). Impact of rapid entry and accelerated care at triage on reducing emergency department patient wait times, lengths of stay, and rate of left without being seen. Annals Of Emergency Medicine, 46(6), 491-497. Chalfin, D., Trzeciak, S., Likourezos, A., Baumann, B., & Dellinger, R. (2007). Impact of delayed transfer of critically ill patients from the emergency department to the intensive care unit. Critical Care Medicine, 35(6), 1477-1483. Coil, C. J., Flood, J. D., Belyeu, B. M., Young, P., Kaji, A. H., & Lewis, R. J. (2016). The effect of emergency department boarding on order completion. Annals Of Emergency Medicine, 67(6), 730-736. doi:10.1016/j.annemergmed.2015.09.018 Falvo, T., Grove, L., Stachura, R., Vega, D., Stike, R., Schlenker, M., & Zirkin, W. (2007). The opportunity loss of boarding admitted patients in the emergency department. Academic Emergency Medicine, 14(4), 332-337. STRATEGIES TO DECREASE ED BOARDING 36 Joint Commission. (2012). Patient flow through the emergency department. R3 Report Requirement, Rationale, Reference. Issue 4. Kang, H., Nembhard, H. B., Rafferty, C., & DeFlitch, C. J. (2014). Patient flow in the emergency department: a classification and analysis of admission process policies. Annals Of Emergency Medicine, 64(4), 335-342. doi:10.1016/j.annemergmed.2014.04.011 Khanna, S., Sier, D., Boyle, J., & Zeitz, K. (2016). Discharge timeliness and its impact on hospital crowding and emergency department flow performance. Emergency Medicine Australasia, 28(2), 164-170. doi:10.1111/1742-6723.12543 Khare, R., Powell, E., Reinhardt, G., & Lucenti, M. (2009). Adding more beds to the emergency department or reducing admitted patient boarding times: which has a more significant influence on emergency department congestion?. Annals Of Emergency Medicine, 53(5), 575-585. doi:10.1016/j.annemergmed.2008.07.009 Kyriacou, D., Ricketts, V., Dyne, P., McCollough, M., & Talan, D. (1999). A 5-year time study analysis of emergency department patient care efficiency. Annals Of Emergency Medicine, 34(3), 326-335. McCusker, J., Vadeboncoeur, A., Lévesque, J., Ciampi, A., & Belzile, E. (2014). Increases in emergency department occupancy are associated with adverse 30-day outcomes. Academic Emergency Medicine, 21(10), 1092-1100. doi:10.1111/acem.12480 Medicare. (2016). Compare hospitals. Retrieved from https://www.medicare.gov/hospitalcompare/compare.html#cmprTab=2&cmprID=460041 %2C460003%2C460051&cmprDist=3.1%2C26.2%2C40.7&dist=50&lat=41.0602216&l ng=-111.9710529&loc=LAYTON%2C%20UT STRATEGIES TO DECREASE ED BOARDING 37 Medicare. (2017a). What is hospital compare. Retrieved from https://www.medicare.gov/hospitalcompare/About/What-Is-HOS.html Medicare. (2017b). Measures and current data collection periods. Retrieved from https://www.medicare.gov/hospitalcompare/Data/Data-Updated.html# Medicare. (2017c). Data sources. Retrieved from https://www.medicare.gov/hospitalcompare/Data/Data-Sources.html Momani, A., Hirzallah, M., & Mumani, A. (2016). Improving employees' safety awareness in healthcare organizations using the DMAIC quality improvement approach. Journal of Healthcare Quality. doi: 10.1097/JHQ.0000000000000049 Northwestern Memorial Hospital. (2008). A description of the infrastructure, the organizational committees, and the decision making bodies specifically designed to oversee the quality of patient care. Retrieved from http://ww2.nmh.org/oweb/MagnetDoc/01_oo_organizational_overview/2128/oo25_narrative.htm Nursing Solutions Incorporated. (2016). 2016 national healthcare retention & RN staffing report. Retrieved from www.nsinursingsolutions.com/.../retention.../NationalHealthcareRNRetentionReport2016 Patel, P., & Vinson, D. (2005). Team assignment system: expediting emergency department care. Annals Of Emergency Medicine, 46(6), 499-506. Pines, J., Batt, R., Hilton, J., & Terwiesch, C. (2011). The financial consequences of lost demand and reducing boarding in hospital emergency departments. Annals Of Emergency Medicine, 58(4), 331-340. STRATEGIES TO DECREASE ED BOARDING 38 Rabin, E., Kocher, K., McClelland, M., Pines, J., Hwang, U., Rathlev, N., & ... Weber, E. (2012). Solutions to emergency department 'boarding' and crowding are underused and may need to be legislated. Health Affairs, 31(8), 1757-1766. doi:10.1377/hlthaff.2011.0786 Retezar, R., Bessman, E., Ding, R., Zeger, S., & McCarthy, M. (2011). The effect of triage diagnostic standing orders on emergency department treatment time. Annals Of Emergency Medicine, 57(2), 89-99. doi:10.1016/j.annemergmed.2010.07.026 Singer, A. J., Thode Jr, H. C., Viccellio, P., & Pines, J. M. (2011). The association between length of emergency department boarding and mortality. Academic Emergency Medicine, 18(12), 1324-1329. doi:10.1111/j.1553-2712.2011.01236.x Stauber, M. A. (2013). Advanced Nursing Interventions and Length of Stay in the Emergency Department. JEN: Journal Of Emergency Nursing, 39(3), 221-225. doi:10.1016/j.jen.2012.02.015 Sun, B. C., Hsia, R. Y., Weiss, R. E., Zingmond, D., Liang, L., Han, W., & ... Asch, S. M. (2013). Effect of emergency department crowding on outcomes of admitted patients. Annals Of Emergency Medicine, 61(6), 605-611. doi:10.1016/j.annemergmed.2012.10.026 Thoennes, L. (2009). Qualipedia: DMAIC. Retrieved from http://www.qualitydigest.com/inside/six-sigma-news/qualipedia-dmaic.html# White, B. A., Biddinger, P. D., Chang, Y., Grabowski, B., Carignan, S. & Brown, D. F. (2013). Boarding inpatients in the emergency department increases discharged patient length of stay. Journal of Emergency Medicine, 44(1), 230-235. doi: 10.1016/j.jemermed.2012.05.007 STRATEGIES TO DECREASE ED BOARDING 39 Wiler, J., Gentle, C., Halfpenny, J., Heins, A., Mehrotra, A., Mikhail, M., & Fite, D. (2010). Optimizing emergency department front-end operations. Annals Of Emergency Medicine, 55(2), 142-160. doi:10.1016/j.annemergmed.2009.05.021 Zeller, Scott. (2015). Psychiatric patient boarding problems in the emergency department. Retrieved from ihssf.org/PDF/foundationbhpatientboarding.pdf STRATEGIES TO DECREASE ED BOARDING Appendix A DMAIC Theoretical Framework 40 STRATEGIES TO DECREASE ED BOARDING 41 (Northwestern Memorial Hospital, 2009). STRATEGIES TO DECREASE ED BOARDING Appendix B Project Proposal PowerPoint 42 STRATEGIES TO DECREASE ED BOARDING 43 STRATEGIES TO DECREASE ED BOARDING 44 STRATEGIES TO DECREASE ED BOARDING 45 STRATEGIES TO DECREASE ED BOARDING Appendix C IRB Reliance Agreement 46 STRATEGIES TO DECREASE ED BOARDING 47 STRATEGIES TO DECREASE ED BOARDING 48 STRATEGIES TO DECREASE ED BOARDING 49 STRATEGIES TO DECREASE ED BOARDING 50 STRATEGIES TO DECREASE ED BOARDING Appendix D IRB Exemption 51 STRATEGIES TO DECREASE ED BOARDING 52 STRATEGIES TO DECREASE ED BOARDING \ 53 STRATEGIES TO DECREASE ED BOARDING 54 Appendix E Admission Delay Log Date Time patient could have left Time patient actually left the Floor the patient was Reason for admission STRATEGIES TO DECREASE ED BOARDING the ED ED 55 being admitted to delay STRATEGIES TO DECREASE ED BOARDING Appendix F Abstract Submission 56 STRATEGIES TO DECREASE ED BOARDING 57 Control/Tracking Number: 17-AB-16-ACEP Activity: Research Forum Submission Current Date/Time: 3/16/2017 1:17:40 AM Identifying strategies to reduce emergency department boarding Author Block: Darrin VansCoy. University of Utah, Kaysville, UT Abstract: Study Objectives: Compare state and national boarding times to the local hospital's times. Identify barriers in admission processes that could be causing delays in moving an admitted patient to the hospital floor. Develop recommendations based on the findings. Disseminate the project findings. Methods: Obtained the median boarding times of similar volume hospitals from the Medicare hospital compare website. Created a data list by extracting a report from the electronic medical record, using the emergency department (ED) clerk's call log, the nursing supervisor's daily staffing sheet, and the ED daily staffing sheets. The data list included every patient admitted to the hospital from the ED over the months of Oct-Dec 2016, 928 total patients. No patients were excluded. The boarding time for each patient was found by finding the difference between the time of admission decision and the time the patient left the ED. The floor staff and floor census at the time of admit as well as the times that nurses were called in that correlated with an admission were found on the nursing supervisors daily staffing sheet. The time the admitting physician was called was found on the ED clerks call log and the difference between that time and the time of admission decision was used as consult time. Common reasons for delays in RNto-RN reporting were captured in an admission delay log that the ED nurses filled out during 2 STRATEGIES TO DECREASE ED BOARDING 58 months of the project. Data was analyzed using linear regression, Pearson R, t-tests, and random forest. Results: Delay log findings showed top reasons for delays in moving an admitted patient to the floor as calling in a nurse, not having orders from the admitting physician and not having a hospital bed available. The ICU floor had the most delays with a mean delay of 87 min. The Medicare hospital compare website listed the median boarding times as 47 min for the hospital, 66 min for Utah and 88 min for the nation. The day of the week had very little effect on mean boarding times. Using linear regression there was no significant findings that correlated boarding times to nursing staff. Telemetry and behavioral health floor census had weak correlations with boarding time, Pearson R 0.22299 and Pearson R 0.15116 respectfully. Mean consult times ranged from 6 to 58 minutes. The overall mean consult and boarding time was not significantly affected by shift change. Using a two-tailed test, the mean boarding time was significantly higher when a nurse was called in compared to when a nurse wasn't called in, 72 min and 55 min respectfully. P=0.00215. Conclusion: Recommendations to decrease boarding time are to Implement strategies that streamline nurse to nurse report, streamline physician consultation, increase bed availability, and reduce the length of stay of behavioral health patients. These recommendations include the following. One attempt call to give report, then give bedside report. Increase floor staffing to always allow for an admission without calling in a nurse. Add an additional nurse who can help with admissions and transfers. Allow nurses to take report and accept the patient without admitting physician orders. Set protocols for nurses to follow on every admitted patient prior to receiving physician orders. STRATEGIES TO DECREASE ED BOARDING 59 Require physicians to use cell phones instead of pagers. Make admission decisions over the phone rather than in person. Increase bed availability by prioritizing discharges and transfers between floors to happen first thing in the morning. Clean rooms just after discharge rather than just before an admission. Hire full time in house crisis workers. Author Disclosure Information: D. VansCoy: ; Iasis Healthcare Corporation. Category (Complete): Quality and Patient Safety Awards (Complete): EMF Grant Funded: No I would like to be considered for the Young Investigator Award: Yes I would like to be considered for the Best Resident Paper Award: No I would like to be considered for the Best Medical Student Paper Award: No Attached Files: Signature (MS-WORD, 26163 bytes) Status: Complete ***To log out, simply close your browser window. All information will be saved if you have hit the Continue button after each step. American College of Emergency Physicians 4950 W. Royal Lane Irving, TX 75063 Technical Support: OASIS Helpdesk or call 217-398-1792 Powered by cOASIS, The Online Abstract Submission and Invitation System SM © 1996 - 2017 CTI Meeting Technology All rights reserved. STRATEGIES TO DECREASE ED BOARDING Appendix G Various Tables and Images of Results 60 STRATEGIES TO DECREASE ED BOARDING Hospital, National, and State Boarding Times Median Boarding Time Nation 88 min EMR boarding time 51 min Utah past 3 months Hospital boarding time from Medicare 66 min 47 min 61 STRATEGIES TO DECREASE ED BOARDING 62 Admission Delay Log, Reasons for Delays Reasons For Delays Pt. had MRI 1st RN in meeting MD shift change Pt/RN ratio maxed, no RN to call Busy with procedures RN on lunch break RN needs to chart Admitting Dr. in ED with patient Waiting for surgeon Just refuse to take report Waiting for bed assignment Shift change RN giving meds ED staff too busy Just had another admit Have to clean room Transferring pt. off floor 1st RN with patient No available bed No orders from MD Calling in RN 0 2 4 6 8 10 12 14 STRATEGIES TO DECREASE ED BOARDING Admission Delay Log, Hospital Floor Differences Floor ICU/IMC Number of Delays in RN Report 20 Mean Delay in Minutes 87 min Top Reasons for Delay 1. Transferring a pt. out of the icu prior to admit 2. Calling in a nurse and refuse report until they arrive Medical 12 51 min 1. Nurse with patient BHU 11 55 min 1. No orders from physician, refuse to take report. Telemetry 10 90 min 1. Nurse with patient 2. Calling in a nurse and refuse to take report 2. Have to wait because we just had another admit 2. Not enough nursing staff, Pt. to RN ratio maxed out, no nurse to call in Surgical 6 69 min 1. Nurse with patient 2. Not enough nursing staff, Pt. to RN ratio maxed out, no nurse to call in ED Pediatrics OB/GYN 3 3 1 60 min 1. Too busy, no staff to transfer pt. 100 min 1. Calling in a nurse and refuse to take report 60 min 1. No orders from physician yet, refuse to take 2. Nurse with patient report. 63 STRATEGIES TO DECREASE ED BOARDING 64 Mean and Median Boarding Time of Hospital Floors Floor Mean boarding time All floors when RN called in Obstetrics ICU/IMC GYN Surg All floors Medical Telemetry Surgical Pediatrics BHU GI Lab Cath Lab L&D 72 min 67 min 66 min 64 min 56 min 56 min 56 min 55 min 52 min 52 min 35 min 25 min 25 min Mean Boarding Time by Day of Week Day of the Week Mean Boarding Time Tuesday 58 min Friday Wednesday Saturday Monday Sunday Thursday 60 min 58 min 55 min 53 min 52 min 50 min Floor Median boarding time All floors when RN called in GYN Surg ICU/IMC Medical All floors Telemetry BHU Surgical Obstetrics Pediatrics GI Lab L&D Cath Lab 72 min 60 min 57 min 53 min 51 min 51 min 50 min 50 min 47 min 37 min 28 min 21 min 18 min STRATEGIES TO DECREASE ED BOARDING Admitting Physician Consult Times Admitting Physician F P S T DD Z N FF C I K GG L O Q X B V A E J M U Y EE H W CC AA D Mean Consult 58 38 37 31 28 23 20 20 19 19 19 19 18 18 17 17 16 16 14 14 11 10 10 10 10 9 9 9 6.5 6 Admitting Physician Median Consult F 49 Z 27 S 25 P 24 T 24 DD 23 C 19 N 16 FF 14 B 13 X 12 E 12 K 11 I 10 O 10 A 10 AA 8.5 GG 8 V 8 J 8 EE 8 W 8 L 7 M 7 U 7 H 7 Q 6 Y 6 CC 6 D 6 65 STRATEGIES TO DECREASE ED BOARDING 66 ED Physician Consult Times ED Physician Q R T K M S E L CC DD A B W N U AA V Y Z C F G D BB J I P X H O Mean Consult 35 34 33 32 29 29 28 27 26 26 22 20 19 18 17 17 16 16 16 15 15 15 14 14 13 10 10 10 9 8 ED Physician M Q E S L T DD K A W R CC N AA B U Y G D V C J Z F P BB X I O H Median Consult 27 22 21 19 19 18 17 16 14 14 13 13 13 13 12 12 11 11 11 10 10 10 9 8 7.5 7 7 6 5 4 STRATEGIES TO DECREASE ED BOARDING 67 Admitting Physician Boarding Times Admitting Physician I DD C G W D A B V Q CC T M U F O H S N E AA R FF GG BB Z J K L Y P X EE Mean boarding 125 90 77 74 68 66 64 61 61 60 60 58 58 57 54 53 53 52 52 52 52 52 50 50 50 48 44 42 40 39 35 30 30 Admitting Physician I C DD V G A Q D CC B M U O T S N AA R BB Z W GG E FF H J Y F K P L EE X Median boarding 100 82 77 62 60 58 57 56 56 55 55 54 53 52 51 51 51 51 51 51 50 49 46 46 42 42 41 37 36 36 35 34 17 STRATEGIES TO DECREASE ED BOARDING 68 ED Physician Boarding Times ED Physician P O D A Q DD J BB X N I M S T U B F H C G V Z W R Y E K AA CC L Mean Boarding 71 69 67 64 63 62 62 61 60 59 59 57 57 57 57 54 54 54 53 51 51 50 49 49 49 48 47 47 46 42 ED Physician J A DD BB P U S Q D T X I E N H F V W O M Z G B CC C AA R Y L K Median Boarding 72 63 62 62 61 60 58 56 55 54 53 53 53 52 52 50 50 49 48 48 48 47 45 45 44 43 41 41 41 39 STRATEGIES TO DECREASE ED BOARDING 69 ED Physician Length of Stay ED Physician M K J S T W R Q F DD E D CC A U G Z N P X BB B AA C Y L I V O H Mean Length of Stay 312 293 282 280 278 274 264 260 241 238 227 226 223 222 221 219 218 213 208 208 200 196 196 189 189 189 188 182 177 170 ED Physician M S W T Q F K R DD G A J D E P Z X U B BB N CC C Y I H AA L V O Median Length of Stay 298 268 264 263 254 243 238 238 234 226 221 218 214 213 209 207 200 198 196 190 189 182 177 176 171 171 170 169 162 152 STRATEGIES TO DECREASE ED BOARDING 70 Boarding and Consult Time During Shift Change Mean Consult Time Mean Boarding Time During Shift Change 21 min 59 min Not During Shift Change 19 min 55 min Linear Regression ED Staff Total boarding time Scatter Diagram (Predicted Y, Total boarding time vs. ed staff hrs/pt. ) 500 400 300 200 100 0 1.2 1.4 1.6 Linear Regression ED Census 1.8 2 2.2 ed staff hrs/pt. 2.4 2.6 2.8 Pearson R 0.06239 Total boarding time Scatter Diagram (Predicted Y, Total boarding time vs. ed census ) 500 400 300 200 100 0 50 60 70 80 90 ed census 100 Pearson R -0.02702 110 120 130 STRATEGIES TO DECREASE ED BOARDING 71 Linear Regression ICU Staff Boarding Time Scatter Diagram (Predicted Y, Boarding Time vs. ICU/IMC Floor Staff Level ) 500 400 300 200 100 0 1 2 3 4 5 ICU/IMC Floor Staff Level 6 7 Pearson R 0.06846 Linear Regression ICU Census Boarding Time Scatter Diagram (Predicted Y, Boarding Time vs. ICU/IMC Census ) 500 400 300 200 100 0 2 4 6 8 10 ICU/IMC Census Pearson R 0.06944 12 14 16 STRATEGIES TO DECREASE ED BOARDING 72 Linear Regression Medical Staff Boarding Time Scatter Diagram (Predicted Y, Boarding Time vs. Medical Floor Staff Level ) 150 100 50 0 0.5 1 1.5 2 2.5 3 Medical Floor Staff Level 3.5 4 4.5 Pearson R 0.00847 Linear Regression Medical Census Boarding Time Scatter Diagram (Predicted Y, Boarding Time vs. Medical Floor Census ) 150 100 50 0 2 4 6 8 10 12 Medical Floor Census Pearson R -0.01144 14 16 18 STRATEGIES TO DECREASE ED BOARDING 73 Linear Regression Surgical/Pediatrics Staff Boarding Time Scatter Diagram (Predicted Y, Boarding Time vs. Surgical/Peds Staff Level ) 250 200 150 100 50 0 0.5 1 1.5 2 2.5 3 Surgical/Peds Staff Level 3.5 4 4.5 Pearson R 0.02488 Linear Regression Surgical/Pediatrics Census Boarding Time Scatter Diagram (Predicted Y, Boarding Time vs. Surgical/Peds Floor Census ) 250 200 150 100 50 0 0 5 10 Surgical/Peds Floor Census Pearson R 0.06507 15 20 STRATEGIES TO DECREASE ED BOARDING 74 Linear Regression Telemetry Staff Boarding Time Scatter Diagram (Predicted Y, Boarding Time vs. Telemetry Floor Staff Level ) 150 100 50 0 0.5 1 1.5 2 2.5 3 Telemetry Floor Staff Level 3.5 4 4.5 Pearson R 0.11484 Linear Regression Telemetry Census Boarding Time Scatter Diagram (Predicted Y, Boarding Time vs. Telemetry Floor Census ) 150 100 50 0 0 2 4 6 8 10 Telemetry Floor Census Pearson R 0.22299 12 14 16 STRATEGIES TO DECREASE ED BOARDING 75 Linear Regression Behavioral Health Unit Census Boarding Time Scatter Diagram (Predicted Y, Boarding Time vs. BHU Census ) 160 140 120 100 80 60 40 20 0 2 4 6 8 10 BHU Census 12 14 16 18 Pearson R 0.15116 Calling in a Nurse Effect on Boarding Time Calling in a Nurse No 56 Yes 72 0 10 20 30 40 50 Mean Boarding Time in Minutes 60 70 80 STRATEGIES TO DECREASE ED BOARDING Random Forest Variable Importance Partial Effect Plots 76 STRATEGIES TO DECREASE ED BOARDING 77 STRATEGIES TO DECREASE ED BOARDING Behavioral Health Patient Times National Average LOS Behavioral Health Patients Mean Time 15 hrs LOS Behavioral Health 5 hrs 31 min From calling crisis worker 3 hrs 55 min LOS all patients excluding 3 hrs 28 min Patients to admission decision Behavioral Health 78 STRATEGIES TO DECREASE ED BOARDING Appendix H Hospital Stakeholders PowerPoint Presentation 79 Running head: STRATEGIES TO DECREASE ED BOARDING 1 Running head: STRATEGIES TO DECREASE ED BOARDING 1 STRATEGIES TO DECREASE ED BOARDING 82 STRATEGIES TO DECREASE ED BOARDING 83 STRATEGIES TO DECREASE ED BOARDING 84 STRATEGIES TO DECREASE ED BOARDING 85 STRATEGIES TO DECREASE ED BOARDING 86 STRATEGIES TO DECREASE ED BOARDING 87 STRATEGIES TO DECREASE ED BOARDING Appendix I Project Poster 88 STRATEGIES TO DECREASE ED BOARDING 89