1 University of Michigan Health System Program and Operations
Contents
1. Figure 4 Total length of stay for each patient and current location 12 Patient LOS by Three Intervals Figure 5 shows the patient LOS with the ED visits split into three intervals Arrival to Exam Room Exam Room to Physician and Physician to Discharge Each bar represents a patient and the height of the bar is the amount of time that the patient has spent at that interval Open Main Delete Graph Fatlent Length of Ste ar sin each Interval June 4 22 22 Aci LOS Minutes 17 Number of Fallente In sach 21 Lvcallvn Location Intervala Figure 5 Patient length of stay for selected intervals 13 Individual Graphs Figure 6 shows seven individual graphs to represent patient LOS in those seven areas of the ED It doesn t show any additional information when compared to Figure 3 but because all of the information is side by side no information is hidden from the viewer It takes longer to understand what these graphs are showing because the scales are all different Had all of the scales been set to be the same and the graph areas set to be the same a lot of information would be lost because the scales would be inappropriate for the data Patient Length of Stays in each Location Individual Graphs June 4 22 22 Humber of Patient id 1 1 3 F a in S a e m 2 E ee Se E ft fe fs ee a te ss el 13 14 Member of Patieat Figure 6 Patient length of stay for each location 14 Patient2. a A gt w bcm ers te ta a ae This will open up a message box like the one shown below 26 C toh Er A r automatically disabled diur ee ene sare MacOS Low not recommended You are not protected from pot ntially unsafe macros rab N E PA 2 Select your macro Security Level either Medium recommended or Low not recommended Checking the radio button beside Medium which is what I d recommend means Excel will prompt you for permission each time a macro tries to run Checking the radio button beside Low not something I d encourage means Excel will have free reign to run any and all macros in the spreadsheets you open You might notice the words Virus scanner s installed in the bottom left corner of that screenshot Having a good antivirus program running at all times is a great idea having a good antivirus program running when you re using Excel macros is a must Still Not Working There May Be an Excel Quirk at Work l Complete the procedure above then save and close your spreadsheet 2 Close Excel completely J Reopen Excel and your spreadsheet 4 Complete the procedure above once more 27 5 Save and close Excel again 6 When you open Excel the next time your macros should be operating normally Yeah it s odd but I find that in Excel 2003 I usually have to change the security settings save and close Excel a couple of times before
3. Count in Each Area Figure 7 shows the patient count in each area The x axis has the names of the areas and the y axis shows the number of patients in each area is saan sepia Patient Count in Each Location 30 S 15 Humber ol Patents 1g Arrival Trage Waling Reon Ezam Rawi Wilh Physician Decision Admi Decision Discharge Figure 7 Number of patients in each area 15 Box plot of LOS Figure 8 shows a box plot of the LOS s of all adult patients in the main ED This graph is important because it shows the long tails and outliers The tails show the range of patient LOS in each location The box shows the 25 and 75 percentile The x in the box shows the median Open Main Delete Graph LOS Boxplot by Location of June 4 22 22 api LOS Minutes 700 600 End Points are Max and Min Purple x are Median Box ends are 25th and 75th Percentile 500 400 300 200 Decision Admit Decision Discharge Waiting Room Exam Room With Physician Arrival Triage Location Figure 8 Box plot of patient length of stay by location 16 Average LOS by Hour Figure 9 shows the average LOS by hour of the day along with the 95 and 5 percentiles The 95 and 5 percentiles are based on the last 30 days of data The average LOS is calculated by taking the ED LOS for each patient currently in the ED and averaging them This graph allows comparison of the data and hour with the previous 30 days at
4. ED Data Errors Within the data received there were multiple errors Some patient timestamps did not follow the correct chronological order of the ED For some situations this might reflect the actual flow of that patient through the ED For example a patient may be rushed into the exam room before being triaged This patient would not be following the chronological order of the ED However due to the large amount of data the team had to make the assumption that this was a data error Another example is when a patient departs before they are seen by a physician Also a patient should never be in the ED department for more than three days Table gives the summary of data errors Along with data not following the chronological order of patient flow there were many nulls where timestamps were supposed to be This inaccurate data entry has led to many outliers in the analysis of patient flow Table 1 Data Errors Location time differences Negative Greater than 3 days Percent of Total In Exam Room Arrival 1061 0 1 92 Seen by physician Arrival 119 82 0 36 0 Seen By Physician In 887 T 1 64 Exam Room Departure In exam room 62 0 0 11 Departure Seen by 2625 0 4 76 ohysician Exam Room Triage 9940 0 18 02 Exam Room Waiting Room Discharge Decision to Discharge 460 0 84 aE Decision to 98 20 0 2 1 Admit Data Cleansing 1695 0 3 07 2 For the program to run accurately these data errors need to be
5. Excel 2002 sp3 Visual Basic for Applications VBA was chosen This program allowed the team to analyze data as well as generate graphs using various algorithms 2002 VBA for Excel was chosen because it was the most available to the clients and it could perform the necessary tasks of the graphical display Along with choosing the proper language the team had to determine how to code in each language Graphical Analysis Graphs from the data were created to display the current situation in the ED The team created graphs to visually show the situation Multiple graphs were created for the clients to choose the best for them to look at the graph and know quickly what currently is happening The graphs the team created are in the appendix Debugging The code needed to be validated to ensure that the graphs created by the program were correct Debugging was necessary to ensure the program ran correctly The VBA program provided tools for debugging that included a watch window break points and allowed stepping through code Testing of the program through outside sources and fool proofing the code further debugged the program First outside sources were asked to evaluate the ease of use of the program Next they were asked to try to break the code to assist with the debugging Selected Graphs Based on meetings with the clients and careful consideration 9 graphs were chosen to be included in the ED Overview Program These 9 graphs are clickable
6. LOS Minutes Number of Patlents In each a Locatlon 19 ae 25 eating Fecm Tiepa i hiii Lecatlon Figure 4 3D LOS Graph Viewing the Database By viewing the data base you can see patient numbers This will help you pinpoint the patient that is causing problems in the system 1 Open the ED Overview Excel Program 2 Click Open Main 3 Adjust the list boxes shown in Figure 4 to the date and time that you are interested in viewing graphs from Adjust them in this order Month Day Hour Min Month Day Hour Min SubmitDate January 1 0 0 Figure 5 Date entry 4 Click Submit Date Wait until the occupancy has been updated in the upper right corner of the form 6 Click Database Figure 5 shows the Database form Here you can see which patient is in which area Using the Plus 15 button or changing the value in the date time box will not adjust the time in the Main Form Also you can t generate graphs from the Database form a 24 This button will hide the database form This button will return you to the main form 13805683 13807320 Arrival Exam Room Wait Room Decision to Discharge 13795720 1464 13803594 319 13797819 1378 13803601 298 13802112 1059 13803607 236 13802317 1018 13803608 255 13803615 313 13803617 244 13803781 105 13803844 25 13804454 106 13804802 10 13804724 98 13804812 23 13804858 5 13805164 5 13805567 Decision t
7. a preliminary framework for a patient flow monitoring system the team e Documented patient flow through the UM ED e Obtained validated and analyzed relevant ED data to determine which charts should be included in the display With this information the team e Developed the UM ED Overview Tool which consists of a display to allow for quick analysis of the ED status from 2008 Project Scope This project included the patient processes of the Adult Main ED at UM The project examined the location of patients and how long they are at each location The patient flow through the ED was separated into the following steps arrival Triage waiting room bed pre MD bed post MD and discharged These steps range from when a patient enters the ED and is registered to when the patient is released from the ED The project included development of a visual display to show the status of the ED which included the status of each patient Figure 1 gives a visual description of the scope of the project The parts highlighted in yellow are things that were completed to create the ED Overview Program The scope included historical data excel data cleansing and graphical development Make Graphs Real Time Data Data pulled via macra VBA Cleanse Data This project did not include any hospital operations or processes outside of the Emergency Department It did not include any inpatient or outpatient departmental processes The display was not designe
8. creates a chronological flow to accurately mimic the patient s flow through the ED If a patient did not have a discharge time the patient s information was removed Programming A programming language had to be chosen to graphically display patient flow Microsoft Excel 2002 sp3 Visual Basic for Applications VBA was chosen This program allowed the team to analyze data as well as generate graphs using various algorithms 2002 VBA for Excel was chosen because it was the most available to the clients and it could perform the necessary tasks of the graphical display Along with choosing the proper language the team had to determine how to code in each language Debugging The code needed to be validated to ensure that the graphs created by the program were correct Debugging was necessary to ensure the program ran correctly The VBA program provided tools for debugging that included a watch window break points and allowed stepping through code Testing of the program through outside sources and fool proofing the code further debugged the program Interface selection and developing visuals Iterative development led to the selection of the 9 graphs that were selected to properly reflect the patient flow The graphs needed to contain useful information while being easy to read and understand The Main Form interface can be seen in Figure 1 and it includes snapshots of the 9 graphs chosen Main Form I 25 Occupancy 60 3D Three Locatio
9. the changes take effect It s annoying but that s life right Enabling Macros in Excel 2007 1 In Excel click the Office button in the upper left corner of the screen Pagelayout Formulas Data Rewiew View Click here to open save or print and to see everthing else you can do with your document Alignm The Office Button is the fancy button with the colorful Office logo Click it to open the menu shown in the next step 2 Click the Excel Options button in the lower right 28 Spittheras Amazon Book Links als Mileage Tracker 2007 215 PUBL CAPACITIES x1s Nissan PreApproval_ Warksheel_vid alt Maint Puss Check _Registersort tilargeprirtas 48 Customerialt Cat family eels Whi Page Update Logri Freedom Account 2006 05 Plan 2007 OL ats New Mimggott Ottice Excel Worksheet star Spreadsheet Buyers Checkboolcals d Togli Foe bot 9 Business Contat Manager Check RegisterSart_whd xiT A E Close isi KA on H RA io ji 2 pen Click the Trust Center button on the left Then at the bottom right select Trust Center Settings as shown below oe ee ee ee Lae ee FE ea eee ee eee re eee re Be i ET A EA Sh Gh HH de ed egy hey LRP ma BE ee ee eds ee ec te re bboy hs poe Cees Phin oT pp Bi DA Pe ed pe Cd i imuar Wr i romm mma d FFA pia da ail i Fiar Era iribar The image above can be enlarged by clicking 4 In the next win
10. University of Michigan Health System Program and Operations Analysis Creating a Multi Dimensional Tracking Chart for the University of Michigan Emergency Department Project Final Report To Jeff Desmond MD Service Chief Adult Emergency Medicine John Younger MD Associate Chair for Research Steven Kronik MD Associate Service Chief Emergency Department CC Sam Clark IE Lean Coach Program and Operations Analysis Mark Van Oyen Industrial and Operational Engineering Faculty From IOE 481 Project 3 Team Program and Operations Analysis Chris Chang Josh Hiemstra Robert Kahn Nicole Roels David Ruebenson Date April 27 2009 Table of Contents Executive SUMMA y oria A OEA E 3 Miodu ionns ee torte cuentas eeu epee eons antes 6 Dackoround eee nen ame Ree ee rant a one aie ret ea er a mPa ee OO 6 FS E aha serra ee eect re ee te ce omer A a setae EA 6 SOGIS AIO DIC CIV CS ch sont tea oteceias A em eectias yeas lesauaanece reels iee a saueeet tea slab iaaabeh enon ea 7 PROICCE SCOP C ace cits bsszona tosceeiis E nate anaes tetas eae S 7 e YG UN 0 Sc womeneee tery neater ere Te mecenp Mera Wey Ry erates TY RECN warn Wey SY cere Teer re we Tyce ee Teen ene ere 8 SSC CLC LAPS arme see shed lad A E tatsten dalle ndsecasin celal 10 Limita ONS ioe facades iuahatere icin E E O i 20 FES Ce Ta E e arrestee E EE A E EEA aed E E EO I T E E AE 20 Recommendations for Real Time Data Teamen E E E R 20 APDE AN E ES 22 ADDO Bree E A ste ncoctanabaan
11. ays Open Main Delete Graph Census of June 422 22 and 30 day 96th amp 6th Percentiles by hour e 95 Parcantila m 5 FParceantila k Todays Data Consus Me of Patiente 12 3 4 6 amp FF amp 8 11 2 1 4 16 16 177 8 10 OA 2 Bo A Hour of the Day Figure 11 Current census compared to 5 and 95 percentile Assumptions for Percentile Graphs A sample of 30 data points was used to create the 5 and 95 percentiles in figures 9 10 and 11 The small sample size of 30 data points allows for more outliers to be found above and below the 5 and 95 percentile Because 30 data points are needed to create the 5 and 95 percentiles this graph cannot be created until after February 15 The small sample size was used because a larger sample size would need more data points and prevent development of these graphs until further into the year With the limited number of only 2008 data 30 days was the most appropriate Program The final product is an ED Overview program The program allows users to chose a date and time from 2008 and click on any of 9 buttons to view different graphs Each graph shows a different aspect of the ED at that time see the attached User s Guide The program is robust and other ED data could be added to the file 19 Limitations This project can be very useful to the ED but there are still many limitations to its capabilities The program that was chosen Excel is limited in the capabi
12. cleansed e Ifa patient didn t have a time stamp for the most important parts of patient flow then they were deleted due to inaccurate data entry These included all of the patients who were missing one or more of the following time stamps arrival in exam room seen by a physician and departure e All of the patients which were left before exam LBE were deleted e Ifa patient s in exam room time with physician time or departure time was before their arrival time then this patient was deleted e Ifa patient s with physician or departure time was before their in exam room time then this patient was deleted from the data set e Ifa patient s departure time was before their with physician time this patient was deleted e For triage time and waiting room time if there was no time stamp then the arrival time was entered e If the triage time or the waiting room time was greater than the in exam room time then the arrival time was used for this time stamp e If the decision to admit or decision to discharge were after the discharge time then the patient s discharge time was entered for either decision to admit or decision to discharge e Ifa patient had been in the ED for a total of 3 days they were deleted from the data set These cleansing steps deleted the errors and made each patient follow the system chronologically Programming Language A programming language had to be chosen to graphically display patient flow Microsoft
13. d to be used with any real time data It was developed using Screen Saver to Display data from 2008 Implementation of the UM ED monitoring tool was not included Validation of current data was included in the project but the team did not collect any raw data Save Graphs Graphics Figure 1 Project Scope The scope did not include e Staff workflow patterns e Staff workflow analysis Findings The team s findings include data collection data errors data cleansing programming graphical analysis and debugging Data Collection The data that the team received from the coordinator was calendar year 2008 Centricity Adult ED Main data and was presented in an Excel format It gave data for each patient that entered the Main ED This data was received by time stamps When the patient entered one of the new states a time stamp of this time was recorded There was a total of 55 146 data points received Through discussion and observation the important factors to include in the UM ED Overview Tool were determined and are shown in Figure 2 These factors were the time each patient entered the ED time they were triaged time they entered the waiting room time they were in an exam room pre physician time they were seen by a physician time the decision was made to admit to the hospital or discharge and time they were discharged TRENES Decision Admit Discharge Discharge from ED Figure 2 Tracking states in the
14. dediateceenaisenteees 26 RETTEN C Sn moet taaakiwneleiwleade actrees Raw alee ee ee a ace 32 Executive Summary This project is a detailed account of the analysis of the patient flow at the University of Michigan s Emergency Department ED and the proposed framework for real time snapshot display of patient flow Jeff Desmond Service Chief of Adult Emergency Medicine and John Younger Associate Chair for Research asked the team to provide data analysis and to develop a preliminary framework of a multidimensional graphical display to help with the development of an automated programming tool for patient flow in the Adult Emergency Department The overcrowding in the ED has been reported as an ongoing issue which has resulted in many patients not receiving care in a timely fashion A real time graphical chart of patient information can assist in the discovery of bottlenecks in the process flow of the ED The patient information necessary to create this display would need to include the patient s status in the ED and the time spent in each designated location The purpose of this project was to analyze the previous year s patient flow data while researching different display techniques to develop a multidimensional graphical display Background The ED at the University of Michigan currently does not have a way to display patient flow The implementation of a graphical display system could improve the flow of patients through the ED The dis
15. dow select Macro Settings then select the radio button for Disable all macros with notification Trusted Publishers hi Geit Fat magas in documents polin a trusted location fdd irs O Digable ap macros without motication Trusted Locations C Disable all macros except digitally signed macros O Enable all macros not recommended potentially dangerous code can run Message Bar External Content pila ia CI Trust access to the YEA project object model Pracy Optrans 5 To close the Trust Center window click the lower right OK button 6 Save and close Excel completely 29 7 Reopen Excel and your spreadsheet one with macros anyway Now there should be aSECURITY WARNING notification beneath the Office ribbon as shown in the next step 8 In the SECURITY WARNING banner click the Options button Ca Copy S ee Past J Format Painter iu le Ja A a Esa Macros amp Activex Maras and one or more Actives coniras have been disabled The actve content might contain viruses or other security hazards Do not enable this content unless you trust the source of this fie Warning It is not possible to determine that this content came from a trustworthy source You should leave this content disabled unless the content provides critical functionality and you trust its source Fie Path C e9tep Home Documents Check Register Check _RegeterSert_v40 5LT C Hep protect me from unknown content recommended G Enable t
16. erify accuracy of the data as much as possible This would allow faculty and staff at the hospital to reduce overcrowding and fix bottlenecks in patient flow Introduction Emergency department ED overcrowding has been reported as an ongoing issue at the University of Michigan UM EDs treat patients with injuries illnesses or severe pains that require urgent attention However overcrowding causes some patients to not get care in a timely fashion The first step to avoiding overcrowding in the ED is to understand and document the patient situation in the ED however there are currently no visual displays to provide a snapshot of the current patient situation of the ED A real time graphical chart of patient information including the patient s status in the ED and time spent there can assist in the discovery of bottlenecks in the process flow of the ED An effective chart will divide the patient visits into the following steps arrival triage waiting room bed pre Medical Doctor bed post MD and discharged and waiting Using previous year patient flow data the team developed multidimensional graphs for the ED display to show snapshots of changes in patient status This report presents the design of this monitoring system Background A hospital ED is one of the most vital components of any health care institute providing treatment to a broad spectrum of illness or injuries that may require immediate attention Overcrowding in the ED ha
17. his content After you click OK and the window closes your macros should function normally Because you selected Disable all macros with notification in Step 4 above you will need to follow Steps 8 and 9 each time you open a spreadsheet that contains macros if you want the macros to run that is Alternately in Step 4 above you could select Enable all macros This would allow any and all macros to run freely at all times But I would not suggest it Don t do this unless you are ABSOLUTELY SURE you won t be downloading or using unsafe spreadsheets and by 30 extension macros from ANY source Malicious macros and Active X controls can do serious damage to your computer and or network 31 References Enabling Macros in Excel Spreadsheets Moneyspot org It s Your Money Personal Finance Debt and Budgeting Steps and Spreadsheets Sonata Creations LLC 26 Apr 2009 lt http www mdmproofing com 1ym macros php gt Visual Basic sample code Functions for calculating the Min Max and StdDev Source code examples Happy Codings Visual Basic Programming Sample Code Source Code Search Engine Code Examples Happy Codings 26 Apr 2009 lt http www visualbasic happycodings com Applications VB A code8 html gt 32
18. hs The ED Overview Program allows you to view ED data from a chosen date and time in 9 different graphs 1 Open the ED Overview Excel Program 2 Click Open Main or press ctrl shift R on the keyboard on Adjust the list boxes shown in Figure 1 to the date and time that you are interested in viewing graphs from Adjust them in this order Month Day Hour Min Month Day Hour Min Figure 2 Date entry January 4 Click Submit Date 5 Wait until the occupancy has been updated in the upper right corner of the form 22 DA Main Form Database Month Day Hour Submit Date January 1 0 0 3D Total ED LOS Average LOS Control chart Total LOS Control Chart aa an ae a ut a pel 8 ag a s we a ot a i My i J TEEN es oP n oe od A kt zE if a ue ma aler l i I Occupancy Bargraph Los Box Plok Individual LOS Graphs Figure 3 Main Form Click one of the 9 graph images Wait until the Main Form disappears and the appropriate graph is generated There are two buttons visible on the graph as shown in Figure 3 Click Open Main Opens up the main form without deleting the current graph You will be able to pick another date if you want and develop another graph Click Delete Graph Opens up the main form and deletes the graph You will be asked again if you would like to delete the graph The main form will open up after you have deleted the graph 23 Tq
19. in the program so that the 10 user can choose which graph they would like to view The following 9 graphs were chosen and included in the program Patient LOS in each area Figure 3 shows the patient LOS in each area Each bar represents a person and the height of that bar is the number of minutes that the patient has spent in that specific area Because this graph 1s 3D it is possible for some information to be hidden behind other bars This graph is important to give an overview of what patients are where Open Main Delete Graph Patlant Length of Stays In sach Location June 4 22 22 qE o TOO l J B00 Bw Eng 430 1 LOS Minutes 4 300 200 100 Number of Failent In sach Locatlon nip Location Figure 3 Patient length of stay in each location 11 Patient Cumulative LOS in Each Area Figure 4 shows the cumulative LOS for each patient Each bar represents a person and the height of the bar is the total amount of time that the patient has spent in the ED The previous graph Figure 4 showed the amount of time in the area while this one shows the amount of time spent in the ED From this graph once can also see the locations that the patients are in This graph is important in helping pinpoint which patients have been in the ED for a very long time Open Main Delete Graph Patient Total Length In ED June 4 22 22 ee 14 00 1200 1000 600 LOS Minutes 4 A 400 200 D ge Location
20. lities and options that are available Excel only has a limited number of useful graphs that could be chosen from its arsenal If another program was chosen more dynamic and useful graphs could have been chosen Excel also has many different versions that run differently The differences in Excel versions cause it to further limit Excel s functionality Most of the code and macros are alike across versions but some of it has to be changed so that the program can run properly The data used in this project is another limitation The data that was given to the team was dirty data Dirty data refers to the fact that there were many nulls and errors in the spreadsheets The errors and nulls caused the team to interpret the data which might not properly reflect the actual situation being shown in the data By cleaning the data so that it could be used in creating a display board useful data points could have gotten lost The cleansing was up to the team s discretion A cleaner and more accurate set of data would eliminate the need for interpretation and would more properly reflect the patient flow through the ED The last major limitation to this project is the problem with actual implementation of the final product As discussed above there are real time data problems that would have to be corrected as the data is entered into the database New code or a new programming language would have to be used to fix real time errors Another limitation to the po
21. n LOS Average LOS Control Chart Total LOS Control Chart ae 7 es dti Tig a Eg in 7 k j rl ral _ i re Bg i HE E a a 2 ao i E vt me at va ji a i Hi r zy ua a fe H C a P An z H aF Ps i mahi o 1 r ThE rice i 1 1 T r r i i Occupancy Bargraph LOS Box Plot Individual LOS Graphs Figure 1 Main Form interface Limitations The final product is very useful but has limitations to its capabilities Excel only has a limited number of graphs that can be created to display the patient flow information Excel also has many versions making compatibility an issue e Dirty data led to interpretations of the data which was cleansed and converted manually This process cannot be performed in real time with the current program e Implementation of the graphical display into the ED is not possible with the final product The final product is a desktop application and not a screen saver Also this instrument only has Adult Main ED data loaded in Recommendations The final product an ED Overview program shows that a display tool is possible to create This program provides the preliminary framework for an automated monitoring system and has the ability to display non normalized data The team would recommend to a real time data team that the program be reworked to function with real time data The new program should store additional hourly data on real time implement real time data cleansing and v
22. o Admit Admitted Physician Figure 6 Database Form 7 To return to the main form click Main Form Show Occupancy Plus 15 This button will display the total number of patients in the ED This button will add 15 minutes to the current time and update the database This shows the current date and time in minutes since January 2007 at 12am 25 Appendix B Enabling Macros In Excel Spreadsheets Macros where Excel is concerned are simply sets of instructions that tell Excel to automate some process or processes In my case I love to use macros to sort and or clear out spreadsheet data with a single click and to perform a variety of similar mundane duties Macros can be huge time and effort savers but because they could conceivably be programmed to do pretty much anything to the computer they re running on say wipe out the hard drive or even something worse we have to be very careful with the Excel macros we run You always want to verify a spreadsheet macro s source and make sure you trust the soul who created the thing In any case when you re ready to enable macros in your Excel program you ll find that it s a snap Enable macros in Excel 2003 and older e Enable macros in Excel 2007 Enabling Macros in Excel 2003 and Older 1 In your Excel menubar click TOOLS gt MACRO gt SECURITY as shown below E G me git vew Insert Fomat fT D e Uenga aT
23. play would assist in finding bottlenecks in the current system and would provide the current patient flow status of the ED Methodology Many different methods were used during this project All of the team s activities and their explanations are listed below e Collected data on patient flow Sam Clark project coordinator provided calendar year 2008 Centricity Adult Main ED data The data provided time stamps of the patient flow through the ED which was separated into the following steps arrival triage waiting room in exam room with physician decision discharge or admit and discharge from ED e Data Errors The data had quality problems which varied between invalid entries to blanks in the data set The following were different situations found in the data set o The data did not follow proper chronological order with an example being that the time stamp of a patient s discharge was before that patient was seen by a physician o Empty cells in spreadsheets o Inaccurate data entry which led to outliers in the patient flow e Data Cleansing The 2008 data provided contained errors and nulls that needed to be eliminated Patients who never saw a physician LBE s were removed from the system All data was rearranged to follow the chronological order of the patient flow In the case of a null value the patient s data was analyzed based on a set of assumptions to determine the accurate timestamp to fill that null The time stamp
24. real time program more easily The program will need to be reworked to receive and function with real time data To assist with development of a real time tool a user s manual is 20 included in Appendix A and the programming code has been commented To run the program in Excel Macros must be enabled and the steps to do that are included in Appendix B Data Cleansing When cleansing the data the team made many assumptions previously mentioned in the findings When the program is running in real time many of these assumptions cannot be made Store Additional Data To calculate the 95 and 5 percentiles for the graphs the values had to be calculated ahead of time and stored in a separate column of Excel However if the values could be continuously updated as the data is being collected no additional time would be added in developing graphs Verify Accuracy of Data Because the program will be running in real time data will need to be collected accurately It will be very important that the time stamps are correct because people will be using the program as an assistant in recognizing bottlenecks If the data is incorrect false assumptions will be made about how the ED is running 21 Appendix A To Operate the ED Overview Program You must enable macros in Excel for the program to work If you do not know how to enable macros please follow the directions in the attached PDF called Enabling Macros in Excel Developing Grap
25. s become an increasingly significant problem in hospitals throughout the U S The implementation of a monitoring system that could help improve the flow of patients through the ED in the long term provides the current status of the ED The users of this chart will be the doctors and nurses They will be able to get a sense of where bottlenecks may be occurring in the ED With the knowledge of where the bottlenecks are occurring the doctors and nurses can take action to try to eliminate the bottleneck The steps in the patient process that will need to be monitored are e Arrival Begins when a patient is finished registered at the front desk and ends when a patient enters Triage e Triage Begins when a patient enters triage Severity assessment and ends when a patient enters the waiting room e Waiting room Begins when a patient enters the waiting room and ends when a patient is given a bed e Bed pre MD Begins when a patient arrives to a bed and ends when the MD arrives e Bed post MD Begins when the MD leaves and ends when the patient is discharged e Discharged Begins when a patient is discharged and ends when o The patient is admitted to the hospital o The patient is leaves the ED Key Issues The following factors are key issues that drove the need for this project e There is no real time visual display tool used at the UM ED e Itis difficult to know where the bottlenecks occur in the ED Goals and Objectives To develop
26. ssibility of implementation is that the display created is a desktop application and not a screen saver Therefore this display could not be streaming real time throughout the ED The code was created to only analyze historical data and provide a proof of concept for an actual screen saver used in the ED The team was only given 2008 Adult Main ED data but the program has the ability to function with additional data that could include other areas of the ED The data included in the project did not include MEDPATH or any other subsidiary department of the ED which could cause changes in the Adult Main ED Expected Impact The team provided a program with charts to display the patient flow status in the Adult Main ED The program provides a foundation for development of a real time display that will be able to retrieve data from other systems and automatically update so a current snapshot of the ED will always be available This display will have the capabilities of showing the non normalized data of the ED This snapshot will help reduce bottlenecks in the patient flow in the ED by making them easily recognizable to staff The discovery and reduction of bottlenecks will aid in the process of the improvement of overcrowding Recommendations for Real Time Data Team The scope of this project didn t include any real time data so it is necessary to make recommendations from the findings These recommendations will help the real time data team develop a
27. the same hour Open Main Delete Grapn Average LOG of June 4 22 22 and 30 day 96th amp 6th Percentiles by hour ggg aa AN Ed ZON 95 Percentile a 5 Parcantila jana Average LOS inuiesi 200 wee 100 12 23 4 6 amp F 8 0 11 12 13 A 1 te TF 18 Tt 20 21 2 as a Figure 9 Average length of stay compared to 5 and 95 percentile 17 Total LOS by Hour Figure 10 shows the total LOS by hour of the day along with the 95 and 5 percentiles The 95 and 5 percentiles are based on the last 30 days of data The total LOS is calculated by taking the ED LOS of all patients currently in the ED and adding them up This graph allows comparison of the data and hour with the previous 30 days at the same hour Open Main Delete Graph Total LOS of June 4 22 22 and 30 day 96th amp 6th Percentiles by hour 20000 16000 16000 14000 12000 e 215 Farcentila TOC m 8 Percentile a Today s Data Total LOS Minutes 1 Z 3 4 amp amp TS 9 1 ZH 1 16 16 T 1 1 OA eae Se M Figure 10 Total length of stay compared to 5 and 95 percentile 18 Number of Patients by Hour Figure 11 shows the number of patients in the ED by hour of the day compared to the 95 and oa percentiles as calculated from the last 30 days and plotted by the hour This graph allows the viewer to see how the ED census compares with the percentiles from the last 30 d
Download Pdf Manuals
Related Search