Www.umich.edu



Michigan MedicineOptimizing a Staffing Model for theUniversity of Michigan Medical Group Ambulatory Care UnitsFinal ReportTo:Balqis Elhaddi, Administrative Director - University of Michigan Medical GroupWhitney Walters, Manager - Continuous Improvement University of Michigan Medical GroupMark Van Oyen, Instructor - Industrial and Operations Engineering DepartmentMary G. Duck, Continuous Improvement Advisor; Quality Department - Michigan MedicineDavid Hyatt, Sr. Continuous Improvement Specialist - Michigan MedicineElaine Wisniewski, Technical Communication Advisor - Program in Technical CommunicationDistribution:Allie Mukavitz, Continuous Improvement Specialist - University of Michigan Medical GroupCarali Van Otteren, MHSA, Continuous Improvement Specialist - Clinical Design & InnovationNicki Schmidt, RN, BN, Continuous Improvement Specialist - Quality DepartmentFrom:Kate Burns, Student - Industrial and Operations EngineeringAnirudh Sharma, Student - Industrial and Operations EngineeringAsia Woods, Student - Industrial and Operations EngineeringYansheng Xiong, Student - Industrial and Operations EngineeringDate:December 14, 2020Table of Contents TOC \h \u \z Executive Summary PAGEREF _heading=h.gjdgxs \h 5Introduction PAGEREF _heading=h.30j0zll \h 5Methods PAGEREF _heading=h.1fob9te \h 5Recommendations PAGEREF _heading=h.3znysh7 \h 5Expected Impact PAGEREF _heading=h.2et92p0 \h 6Introduction PAGEREF _heading=h.plpi8lmi2iat \h 7Background and Key Issues PAGEREF _heading=h.3dy6vkm \h 7Goals, Objectives, and Expected Impact PAGEREF _heading=h.1t3h5sf \h 7Project Scope PAGEREF _heading=h.4d34og8 \h 8Design Process PAGEREF _heading=h.2s8eyo1 \h 8Engineering Challenges PAGEREF _heading=h.17dp8vu \h 8Literature Research Update PAGEREF _heading=h.3rdcrjn \h 9Design Constraints PAGEREF _heading=h.26in1rg \h 10Design Requirements PAGEREF _heading=h.lnxbz9 \h 10Design Standards PAGEREF _heading=h.35nkun2 \h 10Design Tasks PAGEREF _heading=h.1ksv4uv \h 11Alternatives Considered PAGEREF _heading=h.44sinio \h 12Criteria for Evaluation PAGEREF _heading=h.2jxsxqh \h 12Decision Matrix PAGEREF _heading=h.z337ya \h 12Data Collection and Analysis Accomplishments PAGEREF _heading=h.3j2qqm3 \h 12Business Intelligence Manager PAGEREF _heading=h.1y810tw \h 13Administrative Manager PAGEREF _heading=h.4i7ojhp \h 13Clients PAGEREF _heading=h.2xcytpi \h 13Interviews PAGEREF _heading=h.1ci93xb \h 13Literature Search & Survey PAGEREF _heading=h.3whwml4 \h 13BCSC Structure Visual PAGEREF _heading=h.2bn6wsx \h 13Cycle Time Graphs PAGEREF _heading=h.qsh70q \h 13Conversion Tool PAGEREF _heading=h.3as4poj \h 13Findings & Conclusions PAGEREF _heading=h.1pxezwc \h 14Staffing Model PAGEREF _heading=h.49x2ik5 \h 14Next Steps PAGEREF _heading=h.2p2csry \h 14Survey Results PAGEREF _heading=h.147n2zr \h 14MGMA Benchmarks PAGEREF _heading=h.3o7alnk \h 14Staffing Factors PAGEREF _heading=h.23ckvvd \h 15Unique Work PAGEREF _heading=h.ihv636 \h 17Clinic Manager Standard Work PAGEREF _heading=h.32hioqz \h 17Other Findings PAGEREF _heading=h.1hmsyys \h 17Length of Sessions PAGEREF _heading=h.41mghml \h 17Analysis of Visits Per Provider PAGEREF _heading=h.2grqrue \h 17Recommendations PAGEREF _heading=h.vx1227 \h 17References PAGEREF _heading=h.wyt5dpm42o8p \h 19Annotated Bibliography PAGEREF _heading=h.3fwokq0 \h 20Appendix PAGEREF _heading=h.1v1yuxt \h 22Appendix A: ACUs PAGEREF _heading=h.gh2r3kuxcrfh \h 23Appendix B: Constraints and Standards Matrix for Optimizing a Staffing Model for the University of Michigan Medical Group Ambulatory Care Units PAGEREF _heading=h.2u6wntf \h 24Appendix C: Preliminary Pugh Matrix for Optimizing a Staffing Model for the University of Michigan Medical Group Ambulatory Care Units PAGEREF _heading=h.19c6y18 \h 25Appendix D: Updated Gantt Chart PAGEREF _heading=h.3tbugp1 \h 26Appendix E: Screenshots of the Survey (Page 1 of 4) PAGEREF _heading=h.l460br7jjrxs \h 27Appendix E: Screenshots of the Survey (Page 2 of 4) PAGEREF _heading=h.hb8nb8fl6vhz \h 28Appendix E: Screenshots of the Survey (Page 4 of 4) PAGEREF _heading=h.d7e3yjite8u \h 30Appendix F: Extended Survey Results* (Page 1 of 18) PAGEREF _heading=h.28h4qwu \h 31Appendix F: Extended Survey Results (Page 2 of 18) PAGEREF _heading=h.1mrcu09 \h 32Appendix F: Extended Survey Results (Page 3 of 18) PAGEREF _heading=h.46r0co2 \h 33Appendix F: Extended Survey Results (Page 4 of 18) PAGEREF _heading=h.111kx3o \h 34Appendix F: Extended Survey Results (Page 5 of 18) PAGEREF _heading=h.206ipza \h 35Appendix F: Extended Survey Results (Page 6 of 18) PAGEREF _heading=h.2zbgiuw \h 36Appendix F: Extended Survey Results (Page 7 of 18) PAGEREF _heading=h.faruye5mb5bp \h 37Appendix F: Extended Survey Results (Page 8 of 18) PAGEREF _heading=h.r7vm4ddllwku \h 38Appendix F: Extended Survey Results (Page 9 of 18) PAGEREF _heading=h.28f2uemby0wo \h 39Appendix F: Extended Survey Results (Page 10 of 18) PAGEREF _heading=h.geyrjvf18sy \h 40Appendix F: Extended Survey Results (Page 11 of 18) PAGEREF _heading=h.3ygebqi \h 41Appendix F: Extended Survey Results (Page 12 of 18) PAGEREF _heading=h.xfdfrd5wx8h1 \h 42Appendix F: Extended Survey Results (Page 13 of 18) PAGEREF _heading=h.9mldxcxgor4a \h 43Appendix F: Extended Survey Results (Page 14 of 18) PAGEREF _heading=h.sqyw64 \h 44Appendix F: Extended Survey Results (Page 15 of 18) PAGEREF _heading=h.3cqmetx \h 45Appendix F: Extended Survey Results (Page 16 of 18) PAGEREF _heading=h.1rvwp1q \h 46Appendix F: Extended Survey Results (Page 17 of 18) PAGEREF _heading=h.tjek8oqaxb9p \h 47Appendix F: Extended Survey Results (Page 18 of 18) PAGEREF _heading=h.yd3q9fwfhipi \h 48Appendix G: BCSC Structure Visual (Page 1 of 3) PAGEREF _heading=h.4bvk7pj \h 49Appendix G: BCSC Structure Visual (Page 2 of 3) PAGEREF _heading=h.2r0uhxc \h 50Appendix G: BCSC Structure Visual (Page 3 of 3) PAGEREF _heading=h.1664s55 \h 54Appendix H: Cycle Time Graphs (Page 1 of 2) PAGEREF _heading=h.3q5sasy \h 54Appendix H: Cycle Time Graphs (Page 2 of 2) PAGEREF _heading=h.25b2l0r \h 55Appendix I: Model Screenshot PAGEREF _heading=h.kgcv8k \h 58List of Tables & FiguresTable 1: The percentage of staff that consider each factor when determining daily staffingTable 2: Factors staff consider when requesting to hire additional staff Table 3: The percentage of staff members that ranked each factor as the most importantTable 4: Wound CareTable 5: Retaking VitalsTable 6: TestingTable 7: DownloadsTable 8: Extra CleaningTable 9: ChaperoneTable 10: Procedure AssistTable 11: FormsTable 12: OtherFigure A: The average length of session of all departments by monthsExecutive SummaryIntroductionThe University of Michigan Medical Group (UMMG) has been forced to adapt to the various changes in protocol due to the onset of the global COVID-19 pandemic. A major challenge that has been presented is the management of staffing levels of the various Ambulatory Care Units (ACUs) under the UMMG umbrella. The Brighton Center for Specialty Care (BCSC or the “Center”) houses several of these ACUs which themselves contain a multitude of specialty clinics, or sub-ACUs. These clinics used to be responsible for their own staffing levels. However, that has grown to be rather difficult due to various changes instigated by COVID-19, such as virtual appointments, increased safety guidelines, and staffing shortages. UMMG has organized a team from the IOE 481 Senior Design Class to perform research, analyze trends, and formulate a flexible model to optimize staffing levels across all ACUs. This document presents the work performed by this group from September to December 2020.MethodsTo acquire the information needed to create an adequate model, the following methods were utilized:Literature Search: Three research papers involving formulation of staffing models were used to guide our own approachAnalysis of Previous Projects: A float-pool staffing project from the previous year was used to provide further insight. Previous models from older projects were presented by one of our coordinators as well, which became the foundation for our own final model.Discussion with Business Intelligence (BI) Manager: The BI manager shared his knowledge of the data collected by their team and how it was used, in conjunction with benchmarks set by the Medical Group Management Association (MGMA), to make a rough staffing calculator.Staff Interviews: Learned from managers and employees about the challenges of the in-scope roles. Allowed us to account for extra, unseen work within our model.Surveys: Feedback from dozens of clinic managers was given to us, highlighting that the MGMA benchmarks were not found to be useful. Data Analysis: Studies of appointment counts and cycle times over the lifetime of the Center showed how those variables changed due to the various changes related to COVID-19. Trends were found to vary across different clinics, highlighting the need for flexibility across models.The inability to conduct interviews on-site and the shortage of time largely increased the difficulty of data collection, ultimately leading to less data being collected than the initial expectation of the team. Therefore, the team decided to place the main focus on the flexibility of the staffing tool which requires more user input. RecommendationsWe recommend that the flexible staffing tool we developed be made available to all interested parties within the University of Michigan Medical Group. Our work suggests that clinic managers themselves should be provided our analysis and model as a flexible tool where they can input their own data and receive a recommended staffing level. Our guide on how to perform time-studies should also be provided, allowing the managers to receive more accurate data and in-turn a more accurate recommendation. Further research on this topic can definitely prove to be beneficial, so we also recommend UMMG has another team from the IOE 481 senior design course work to elaborate on our work.Expected ImpactImplementation of the recommendations above will allow clinic managers to have a far more personalized tool in order to determine their staffing levels. The negative sentiment towards the MGMA benchmarks, as shown in the results of our survey, shows that the rigid guidelines set by the MGMA doesn’t account for a multitude of factors. The model will also be able to adapt to the changes caused by the COVID-19 pandemic, something unaccounted for by the MGMA benchmarks. This will increase the overall satisfaction of the clinic managers, allowing them to better serve their patients and providersIntroductionThe Brighton Center for Specialty Care (BCSC or the “Center”) is one of many health facilities under the purview of the University of Michigan Medical Group (UMMG). BCSC houses several Ambulatory Care Units (ACUs), which serve as groups of outpatient clinics, covering specialties such as dermatology, gastroenterology, and psychiatry. Each clinic has handled staffing independently in previous years; however, the COVID-19 pandemic has complicated the staffing process. On top of factors such as patient cycle time, the number of rooms and providers, and patient activity, the clinics now have to account for COVID-19 safety guidelines, web-based appointments, and their patients' and staff's increased concerns. UMMG hopes to streamline staffing across the ACUs with an adaptable and standardized model. Our team was tasked with researching similar models, suggesting additional data collection, and analyzing all given information to finally create said model. This report entails the progress made on this project from September 8th through December 14th of 2020. Background and Key IssuesPhysicians and advanced practice providers (APPs), which include Physician Assistants and Nurse Practitioners, require adequate supporting staff to provide optimal care. The staffing model was designed to determine the necessary quantities and types of supporting roles, such as nurses, medical assistants (MAs), and patient service assistants (PSAs) needed in each ACU. Without universal staffing standards, ACUs are often understaffed in an effort to save money, or overstaffed, in an effort to guarantee adequate support for APPs. In order to remedy suboptimal staffing, UMMG has implemented benchmarks set by the Medical Group Management Association (MGMA). These benchmarks estimate support staff needed per number of clinical visits by using factors such as clinical specialty and physician productivity. While these benchmarks have been helpful, UMMG doesn’t consider them sufficient, and would like to build upon these benchmarks. This issue has been approached by UMMG’s business intelligence (BI) team, which has already gathered staffing data and created an unpolished “Staffing Calculator”. The data must be further analyzed and the “Calculator” must be further refined before the ACUs can accurately optimize staff. There must be enough staff to effectively assist APPs and patients, but not too many as to create unnecessary expenses. This balance has been further complicated by the COVID-19 pandemic. One major change was the shift to web-based appointments, a service that has not previously been offered by many ACUs. Online appointments make in-person contact unnecessary, but create a host of new challenges such as increased paperwork. COVID-19 regulations, such as social distancing and mandatory masks, also had an impact on the efficiency of in person appointments. These new factors were very important to account for in an accurate staffing model. Goals, Objectives, and Expected ImpactThe primary goal of the project was to create a system that determines the number of staff needed per week at six BCSC sub-Ambulatory Care Units (ACUs). As a secondary goal, the system should be accessible and relevant to the staffing of an additional 160+ University of Michigan Ambulatory Care Units. Due to the variability of roles and different needs at each ACU, the model should be adaptable. To complete this goal, the team will complete the following objectives: Understand the current methodologies used for staffing clinicsResearch similar staffing models to determine an approach that is more accurate and efficient than the current modelDetermine the different positions of the staff currently working, their daily assignments and the restrictions regarding their hours worked per weekDetermine the number of staff needed per position that will satisfy each ACU’s weekly needThis project will impact the staff of the sub-ACUSs at BCSC by changing their overall working schedule. With an adequate number of staff allocated to each ACU, staff will not have too much nor too little to do during their work shifts. Project ScopeThe scope of this project included six sub-ACUs at BCSC. This includes Adult Multispecialty, Comprehensive Cancer Center (CCC), Children Specialty, Comprehensive Musculoskeletal Center (CMC), Ophthalmology (Opthy), and Otolaryngology (Oto). The scope of this project does not include BCSC Alcohol Drug & Treatment Unit (ADTU) Center, Infusion, Medical Procedures Unit (MPU), Operating Room (OR/PACU), Therapy Services, Surgery or non-UMMG ACUs. There are several other locations this model may have staffing applications for in the future at other UMMG ACUs that can be found in Appendix A. However, only in-scope locations were taken into consideration while developing this model.Within the scope of this project are Registered Nurses (RNs), Licensed Practicing Nurses (LPNs), Medical Assistants (MAs), and Patient Services Assistants (PSAs) that work in office. The model aims to inform the recommended staffing levels of these roles. The model will not inform the recommended staffing levels of PSAs who work strictly in the call center, Nurse Practitioners (NPs), Physician Assistants (PAs), and Physicians as they are out of the scope of this project.Additionally, virtual visits, in-person visits, and time off work for staff are within the scope of this project. Individual staff schedules, and patient schedules are out of scope of this project.Design ProcessEngineering ChallengesA major challenge in constructing the model was to make sure it can scale across all clinics. The differences between the ACUs were analyzed further to make this a reality. The model must also be flexible in terms of workflow changes. The model should be usable in the post-COVID world as well, and portions of the Center’s operations are likely to slowly go back to pre-COVID protocols, while others are likely permanently changed. To account for the spectrum of possibilities, the model must include additional factors, such as the ratio of in-person appointments to web-based appointments. The model also must be user-friendly enough so it can be easily used by the staff of each ACU. The primary platform must be able to perform advanced data analysis tasks itself, or be compatible with other platforms that can perform said tasks. Challenges, such as needing to work within the limits of the platform or learning how to perform tasks in an unfamiliar platform, will be faced by our team.Literature Research UpdateThrough the comprehensive literature searches, it was clear that staffing was always a difficult problem for the hospital management. Many efforts have been devoted from the fields of mathematics, engineering and operations research to develop the most suitable staffing model for the unique hospital environments in question. In 2001, Canet and colleagues [1] studied the possibility of mathematical modeling for determining the adequate staffing in the anesthesia and post-anesthesia intensive care units and pain clinics (A-PICU-PC). The model revealed that although variations existed in the ratio of number of staff positions to the number of persons employed by an A-PICU-PC due to independent differences that existed among departments, the possibility of creating a uniform model of a common language to be used by healthcare managers was promising. In the year of 2020 with the pandemic of COVID-19, Balluck and colleagues [2] realized the importance of a safe staffing model for staff safety as well as the expected patient surge and utilized the ADKAR (Awareness, Desire, Knowledge, Ability, and Reinforcement) change model to guide a transition from primary to team nursing. We direct interested readers to the actual article for further detailed discussion of the implementation of the ADKAR change model and its effects on hospitals with variable staff populations. Similar efforts have been done by Kester and colleagues [3] to improve the predictability and accuracy of hiring using historical staffing data, quality improvement and workforce engagement, creating a prospective staffing model and implementing the AACN (American Association of Colleges of Nursing) Healthy Work Environment standards. While all of the above-mentioned work is related to staffing and utilized multiple mathematical methodologies, the team will not be able to directly utilize any of the models of mathematical foundations presented due to various reasons, explained in section Annotated Bibliography. However, insights on prospective modeling does provide directional advice on the future steps as we aim to build a prospective model ourselves. At the same time, similar past projects have been studied extensively as well. The first project analyzed was an IOE 481 project from Fall 2019 involving float pools of staffing. Float pools served as backup reserves of staff to be assigned to various ACUs, but these are no longer in effect due to COVID-19. Though the focus wasn’t identical to ours, the overall approach to the project that the group used in approaching their project was definitely helpful in our own approach. Our discussions with one of our coordinators on a few previous staffing projects provided insights on the correlations between staffing and cycle time with various types of encounter and visit types. Certain ideas such as breaking down the cycle times for each appointment type were mentioned and were further investigated.Design ConstraintsIn the design process, there were three main hard constraints that had to be met in order for the design to be successful. Firstly, the end user’s technical knowledge will not change — meaning the model had to be in a platform that was usable for all clinic managers and all levels of UMMG leadership. The clients previously mentioned that Excel was a good option for user-friendly software. Secondly, the inputs identified in the final model are easily obtained by the end-user. Lastly, the team was not permitted to conduct in-person observations or studies. However, staff were able to conduct virtual interviews with assistance from the administrative manager. Design RequirementsRequirements for a successful staffing model included:Usability for clinic managers and UMMG leadership at all levelsAdaptability across all in-scope specialtiesCapability for long-term use Accountability for variation between clinics inOverall approach staffingTypes of visits, including virtualSkill set required for each role Flexibility with individual schedules and time-off of full-time and part-time The design needed to include all of these requirements for this model to properly inform all clinic managers and all levels of UMMG leadership on appropriate future staffing levels. These requirements were measured by surveying clinic managers using a Likert scale to see the level of which each of the requirements was satisfied. More information regarding the survey results can be found in the section labeled Survey Results. Design StandardsSome design standards that were to be followed include the Health Insurance Portability and Accountability Act (HIPAA) and Private Health Information (PHI) for all patient data while analyzing the type and length of appointment of patients. The design also closely considered MGMA staffing benchmarks. These benchmarks were the main tool managers use to determine staffing levels. Additionally, the design had to closely consider acts and union contracts that highlight how much time and under what circumstances staff can take off of work, such as Michigan Medicine Human Resources information about Employee Time Away From Work, the Family Medical Leave Act (FMLA), the Expanded Family Medical Leave Act (EFLMA), and the University of Michigan Professional Nurse Council (UMPNC) contract. These acts and laws are especially important during the COVID-19 pandemic because many people on staff are taking this time off.A Constraints and Standards Matrix can be found in Appendix B. The criterions for the Pugh Matrix are included in Appendix C.Design TasksTo obtain an adequate amount of information to create an effective model, the following design tasks were determined. Literature Research. Literature research was conducted on existing work done related to staffing modeling in the healthcare field and other industries. As a team, we hoped while enhancing our own knowledge of staffing optimization modeling methods, we would also obtain a more comprehensive view of the staffing levels, including different roles within health care, and also gain insights that lead to the final solution of the project. Detailed updates of Literature Research and pertinent discussions can be found in the aforementioned section labeled Literature Research Update. Discussion with Business Intelligence Manager. The business intelligence manager not only is familiar with all aspects of the data associated with the project but was also able to provide us with more understanding of the project environment and hard constraints and requirements. As a team, we met with the business intelligence manager to have a more profound understanding of the current developing process of the staffing calculator for the staffing modeling across all clinics as well as some knowledge of the data such as certain variables and parameters involved in the project. Staff Interviews. To help with the project and to facilitate the data collection process, input opinions from the current working staff in the clinics are undeniably valuable. The administrative manager was involved in helping arrange interviews between the team and the staff members. The communications with the current staff helped us with problem identification and provided us with key insights to focus on throughout the design of the engineering solution to the project. Several meetings with the administrative manager have occurred and provided the team with valuable feedback.Surveys. Surveys were produced and distributed to the staff members and were collected to understand the benefits and drawbacks of the current staffing method. The information was integrated systemically with the data collected from the business intelligence manager and interviews to provide the team with more holistic preparation for analysis and to possibly yield more accurate analysis results. Data Analysis and Problem Identification. Data was obtained from further suggested studies and was grouped by data provided by the BI department. The combined data was centralized and further organized in an Excel spreadsheet. Data visualization tools within Excel were used to identify possible trends and areas of concern within the data. New factors were created in the spreadsheet, such as the percent of online appointments, to make the model flexible in a post-COVID environment. This data was then refined, and the significance of the factors were analyzed using Excel. The analysis provided a rough model, which was enhanced by forecasting in Excel, and any other analysis methods learned about through the literature search. Throughout this process, the data was compared to MGMA benchmarks to build upon them. Cross-validation methods were used to make sure the process properly scales across the ACUs. From the analysis result, the team identified the major issues in the current processes. Feedback Construction and Model Creation. With the insights gained from research and knowledge of the field, constructive feedback is provided on the existing processes for continuous improvement and created a new staffing model to resolve the key project issues which can be found in the Recommendations section. Validation and Testing. The new staffing model has been validated with a fold of the original data not used for the formulation of the model and was tested in simulations to confirm the achievement of the objective. Alternatives ConsideredThroughout the semester, several models with varying specificity have been theorized and/or developed. A model that used the different types of encounters for each department as inputs was considered. This model would determine adequate staffing levels by evaluating the distribution of time taken per encounter. Another model was theorized to determine adequate staffing levels based on the different subgroups within the clinic. As we collected more data and began building our model, we determined several adjustments that needed to be made. We learned that our tool may be used by nursing leaders in addition to clinic managers, and UMMG leadership. As a result, the audience of the survey increased to include both clinic managers and nursing leads. The scope of the institution research was modified to include past staffing projects within UMMG instead of reaching out to other academic medical institutions. We decided not to use simulation tools like Pro Model to examine the data. Creating a simulation would take a substantial amount of time and effort that was better allocated to further analysis of the given data. Additionally, we expected to follow the Gantt Chart in Appendix D. However due to scheduling obstacles with virtual interviews for staff interviews, the proposed timeline was adjusted. Criteria for EvaluationOur criteria for evaluation is very similar to our project's requirements. Each requirement was important inorder to create a successful model. Below are the criteria for evaluation.Incorporates time-off allocationsIncorporate variation in staff hourly designationsIncorporates skill set and scope of practice for each roleIncorporates future growth capabilitiesFlexibility for variety of visit typesFlexibility for full time and part time employeesLong-term use capabilitiesUsability for all specialities in scopeUsability for all intended usersDecision MatrixWe used the criteria mentioned above to compare two versions of a model. The first option was the original model formulation we created. It is based on a variety of visit types. Each visit type has an associated amount of time that a staff role spends doing work for this appointment type. Additionally, there is a count of the amount of this type of appointment based on prior data. It is difficult to add visit types to this model, The second option was the second model formulation we created. It is based on the tasks that each role does for a visit. The task is associated with a percentage of appointments that require this task based on the total number of appointments. Both models are specific to departments. Below is the decision matrix comparing the two versions of the model with our criteria.The second model formulation based on tasks won while assessing the criteria compared to the current standard for staffing models, which is the MGMA Benchmarks. The second model incorporates variation in staff hourly designation while the first model only looks at total time for an appointment. The second model also incorporates future growth capabilities by allowing tasks to be added and the inputs to be changed, while the second model only includes past appointment types. Additionally, the first model type does not have much usability for intended users because it is extremely difficult to estimate the associated time a staff role spends on a specific appointment. It is easier for the end user to obtain data associated with the time it takes staff roles to complete specific tasks. In conclusion, the second model based on tasks won the decision matrix.Data Collection and Analysis AccomplishmentsInterviewsThe adult multi specialty manager got our team in contact with various staff members in Brighton. Through conversations with six different staff members in the BCSC Adult-Multispecialty unit we were able to gain insight into the challenges that each department and role faces. We also identified the main driving factors for different role’s work. PSA’s work is driven by the number of appointments and number of in-basket messages. The MA’s work is mostly driven by the number appointments, and number of appointments associated with nurse assist work. The LPN’s work is mostly driven by the number of patients. If a patient comes into the clinic the LPN is responsible for checking in with patients, and refilling medication. The RN’s work is driven by the number of procedures, and education visits. These insights were some of our largest accomplishments of this project. The specific tasks based on department and role are highlighted in the ‘Unique Work’ section of this report.Literature Search & SurveyThrough our conversations with the business intelligence manager, administrative manager, and clients and our literature searches, we obtained some helpful information. The survey used to obtain more information and satisfaction benchmarks was completed and approved by the client to be sent to all clinic managers, and nursing leaderships. Screenshots of the survey can be found in Appendix E while detailed information regarding the results can be found in Appendix F.BCSC Structure VisualOur team developed a visual to analyze the types of departments, encounters, visits, and resources in each Sub-ACU. A screenshot of the visual is included in Appendix G. The specific details of clinics within each ACU and the tasks they performed were previously very difficult to keep track of. This visual was a very useful visual reference and helped our team to understand these complexities. It was not a deliverable for our client; however, our coordinators greatly valued this tool and have requested to use it for their team’s future work.Cycle Time GraphsWe were provided access to data from the MiChart database involving counts and cycle times of various monthly appointments across departments and appointment types. Our analysis shows that the cycles times have remained constant across the life of BCSC. Additionally, we have looked at the difference in counts and cycle times of new patient (NP) and returning visit (RV) encounters and their virtual counterparts. There was a drop in counts of appointments during the peaks of lockdown, but they recovered afterwards. We also noticed that cycle times for virtual appointments seem to be lower on average compared to in-person appointments . These graphs can be found in Appendix H.Conversion ToolAdditionally, we have made a rough draft of a potential tool that allows for user input in percentage of virtual appointments and change in number of providers. The tool has data from Mi Chart of types of appointment from each department, count of each type of appointment, and percentage of the type of appointment compared to the total appointments in the department. The tool also has columns for the time associated with each type of appointment for each staff role. This input will be received through interviews with staff. Additionally, it has two columns for adjustments of percentage of appointments, and counts of appointments based on the input that the user enters. Finally, on the following sheet in the Excel book tool the FTE needed for that department is calculated by summing all time needed per staff type for all appointments. Screenshots of this tool can be seen in Appendix I. Our survey and interviews acted as a benchmark for satisfaction along with informing also with other factors we may have not considered including patient acuity, and hours of operation. Findings & ConclusionsIn this section there is information regarding the staffing model we created, survey results, unique work done by clinic roles per department, and additional findings we came across during this project that are technically out-of-scope. The aforementioned visual structure of BCSC can be found in Appendix G.Staffing ModelThe current staffing model we have created has a sheet for each type of provider. This model is incredibly flexible. Screenshots of this model can be found in Apendix I. All light-blue boxes are cells that should be edited. Column B a list of tasks associated with each role. These tasks can be added and deleted. Column C has a percentage. This percentage is associated with what percent of the total appointments have the associated task. For example, if MAs must take vitals for 70% of the appointments in that department then put 70% in the box next to the task “measure vitals” Column D calculates the count of appointments that have that associated task. For this example, it means that if there are 100 appointments in total then there will be 70 appointments that measure vitals. Column E as the time associated with that task. The time associated with a task and the number of appointments that involve that task are then multiplied. All tasks are multiplied and added to get the total time in minutes needed. From there the total time is multiplied by the productivity level. This productivity should account for time off and human capability to complete tasks and keep focused for a whole day. The level should be changed. The total minutes multiplied by the productivity level is then divided by 10,080. 10,080 is the number of working minutes in a month for a FTE (21 days * 8 hours * 60 minutes). Finally the model outputs the FTE for that staff role for that department. Rows 2 and 3 of each page of the model hold the inputs that may change frequently, especially while looking at growth potential. The inputs for the PSA model are the number of appointments and the number of inbasket messages. The inputs for the MA, LPN, and RN model are the number of sessions per month, and the department. This means one provider working one AM session in the Allergy department would count as 1 session. Our team has identified the conversion factor for 20 departments from number of sessions to number of appointments. Those factors can be found in the tab #SessionsTo#Appt. This conversion will only work if the department inputted exactly matches the name in the #SessionTo#Appt tab. If the desired department is not in this tab then the number of appointments should be used as the input. The staffing model can be found in the file StaffingTool.xlsx.Survey ResultsManagers across UMMG were asked to complete a survey regarding their current staffing process. As of December 3, 2020, at 1:19 PM EST, the following data was collected from the survey pictured in Appendix 6. One hundred staff members started the survey, however, only seventy-three completed it in full. Staff Members from nineteen identified centers within UMMG including BCSC responded. The identified centers are listed in the Appendix. Out of the surveyed managers:Fifty-five people managed MAsSixty-seven managed PSAsTwenty-six managed LPNsTwenty-six managed RNsFifty-eight managed other employees. MGMA BenchmarksThe survey helped to confirm the assumption that MGMA benchmarks are not an accurate way to determine staffing for the in-scope roles (RNs, LPN, MAs, & PSAs). Of seventy-three respondents, twenty-five percent found them helpful when determining staffing levels. Fourteen percent of respondents found the benchmarks were relevant for determining services while eleven percent found them to be relevant for roles. Additionally, of the eighty-four respondents, forty-one percent used a different model or system other than MGMA benchmarks when determining staffing levels. Staffing FactorsOf the eighty respondents, fifty-eight percent determined the current system used for staffing was accurate for their clinic. The survey asked respondents to determine the factors used when creating daily schedules. Of the eighty-seven that responded, the most common factor identified was the number of physicians and APPs in the clinic. The percentage of staff that consider each factor can be found below in Table 1. Table 1: The percentage of staff that consider each factor when determining daily staffingDaily Staffing Factors% of staffNumber of total physicians and APPs in clinic64.37%Special services that require extra staff55.17%Complexity of visit types48.28%A staffing model37.93%Input or concerns from staff (RNs, LPNs, MAs & PSAs)37.93%MGMA Benchmarks34.48%Other factors28.74%Input or concerns from physicians and APPs26.44%FTE of physicians and APPs in clinic25.29%Additionally, when the eighty staff members were asked to identify the factors considered when requesting new staff sixty-eight percent considered MGMA Benchmarks while fifty-one percent considered special services that require extra staff. A complete list of the considered factors can be found in Table 2. Table 2: Factors staff consider when requesting to hire additional staffFactors Considered When Hiring% of StaffMGMA Benchmarks67.82%Special services that require extra staff50.57%Complexity of visit types48.28%Number of total physicians and APPs in clinic48.28%A staffing model35.63%FTE of physicians and APPs in clinic35.63%Input or concerns from staff (RNs, LPNs, MAs & PSAs)34.48%Input or concerns from physicians and APPs32.18%Other factors21.84%Staff were asked to identify and rank the factor they deemed the most important when considering staffing needs. Of the eighty managers, forty-nine percent consider the skill set and scope of practice for each staff role to be the most important. Below in Table 3 the percentage of staff that ranked each factor as the most important is detailed. Table 3: The percentage of staff members that ranked each factor as the most importantMost Important Factor for Staffing Needs% of staffThe skill set and scope of practice for each staff role48.75%Time-off allocations18.75%Full-time and part-time considerations15%A variety of Visit types12.50%Variation in staff hourly designations5%Virtual Visit0%Additionally, staff listed in basket volume or call volume, patient count and provider count, the volume of studies and testing, the number of procedures or surgeries, the number of staff visits without a provider, the hours of operation, and overtime, delay in care, and unacceptable turnaround times to be important factors for staffing. Input and concerns from staff, the program growth or referral volume and the acuity of patients were also mentioned to be important factors; however, these are harder to measure.Unique WorkOur team interviewed six different staff members. These interviews lead us to determine unique tasks associated with each position for each department. These tasks do not include all departments in scope and may need additional interviewing to be fully comprehensive. A more detailed description of the tasks and challenges of each department can be found in the file ACUChallenges.xlsx. In this file there is a sheet for MA challenges, LPN and RN challenges, COVID-19 challenges and ratios for specific departments that were mentioned in interviews for the number of providers to number of MAs.We consolidated the MA tasks mentioned in the interviews into main groups. MA’s additional tasks other than typical intake of patients include wound care, reatking vitals, testing (urine, EKG, walking, blood, etc.), downloads, extra cleaning, chaperoning, procedure assisting, filling out forms with patients and other. We do not have accurate time and frequency of these tasks, however we did collect some estimated times and frequencies which were mentioned in the interviews. Below are tables that help explain some of the additional tasks.Table 4: Wound CareTable 5: Retaking VitalsTable 6: TestingTesting 7: DownloadsTable 8: Extra CleaningTable 9: ChaperoneTable 10: Procedure AssistsTable 11: FormsTable 12: OtherThe above tables highlight the additional tasks associated for MA’s more details can be found in the file ACUChallenges.xlsx. Additional, LPN and RN Challenges, along with COVID challenges can be found in this file as well.Other FindingsAnalysis of Visits Per ProviderIn order to find the number of appointments per session our team had to assess HR data. The number of appointments per session varies greatly depending on the department. We analyzed 20 departments and associated an average number of appointments per session, and the average number of appointments per hour of a session. Specific department data can be found in the file “10.28 Session-Provider Breakdown Analyzed.”Length of SessionsOur client requested additional research into the average length of sessions over time. This work is out of our project scope, however it is valuable to our client. We analyzed the same HR data that was mentioned in the section above. The typical provider session length is thought of as being four hours long. There are typically two sessions a day. We measured session length as the difference in time between the start of the first appointment and the end of the last appointment in a session.Figure AFigure A shows the average length of session of all departments by months. We see that the average session time stays between 2.5 and 3 hours with the exception of April and May 2020. April and May 2020 was during the peak of COVID-19 shutdowns. There is a slight upward trend in the months leading up to COVID-19, however it does not approach 4 hours.CheckoutThe number of patients not checking out on the second floor of BCSC was a notable finding from our interview. Approximately 25% of patients do not check out. They either do not checkout because they were in a virtual appointment which has no check out process, or they exited the building without checking out. This causes extra work for PSAs. Throughout the day PSAs must identify patients in their system that have been there for hours longer than expected, and manually check them out. Additionally, all patients that have a follow up appointment and didn’t check out must be called until they confirm a new appointment time. This confirmation must be verbal, so PSAs must call patients until they get a hold of them.Potential improvements could be made to lessen this work load for PSAs. A more robust check out system in clinic could be implaced to prevent people from exited without checking out. The use of signs, and the placement of the checkout counter could be considered. Additionally, a virtual appointment check out could be created. The patient could be directed to a PSA after their virtual visit to ensure the follow up appointments are scheduled right after the appointment with the provider. This finding and recommendations was out of scope, however we thought it was valuable to mention.RecommendationsRead our analysis so you can use the work we have done for any future work in other areas especially interview work pertaining to workload in different specialties.Have this project be a 2nd semester 481 project. Leave them with the model framework of the model, and they can find more true data to put into the model with a time study, further analysis of data provided, and so on.Distribute model to appropriate managers, with the message that this model is meant to be flexible, they know their clinic better than us, so feel free to alter inputs and the model will compute the math for them, also send them instructions on how to do a time study so if they wanted, they could find more accurate data to input into their altered model.References[1]J. Canet, V. Moral, A. Villalonga, D. Pelegri, C. Gomar, and A. Montero, "[Model to predict staffing for anesthesiology and post-anesthesia intensive care units and pain clinics]," Rev Esp Anestesiol Reanim, vol. 48, no. 6, pp. 279-84, Jun-Jul 2001. [Online]. Available: . Modelo de cálculo de plantillas de los servicios de anestesiología, reanimación y terapéutica del dolor.[2]J. Balluck, E. Asturi, and V. Brockman, "Use of the ADKAR(R) and CLARC (R) Change Models to Navigate staffing model changes during the COVID-19 pandemic," Nurse Lead, Aug 20 2020, doi: 10.1016/j.mnl.2020.08.006.[3]K. M. Kester, M. Lindsay, and B. Granger, "Development and evaluation of a prospective staffing model to improve retention," J Nurs Manag, vol. 28, no. 2, pp. 425-432, Mar 2020, doi: 10.1111/jonm.12945.[4]Robinson J, Porter M, Montalvo Y, et al. Losing the wait: improving patient cycle time in primary care. BMJ Open Quality 2020;9:e000910. doi:10.1136/ bmjoq-2019-000910 Annotated Bibliography[1]J. Canet, V. Moral, A. Villalonga, D. Pelegri, C. Gomar, and A. Montero, "[Model to predict staffing for anesthesiology and post-anesthesia intensive care units and pain clinics]," Rev Esp Anestesiol Reanim, vol. 48, no. 6, pp. 279-84, Jun-Jul 2001. [Online]. Available: . Modelo de cálculo de plantillas de los servicios de anestesiología, reanimación y terapéutica del dolor.In their work of Canet and colleagues, it was illustrated that Adequate staffing is a key factor in providing for both effective care and professional staff development. Changes in professional responsibilities have rendered obsolete the concept of one anesthesiologist per operating room. Duties must be analyzed objectively to facilitate understanding between hospital administrators and A-PICU-PC chiefs of service when assigning human resources. A model was created for for estimating requirements for A-PICU-PC staffing based on three factors: 1) Definition of staff positions that must be filled and criteria for assigning human resources; 2) Estimation of non-care-related time required by the department for training, teaching, research and internal management, and 3) Estimation of staff required to cover absences from work for vacations, personal leave or illness. The model revealed that the ratio of number of staff positions to number of persons employed by an A-PICU-PC is approximately 1.3. Differences in the nature of services managed by an A-PICU-PC or the type of hospital might change the ratio slightly. The model can be applied universally, independently of differences that might exist among departments.Similarly, to what Canet and colleagues did and discovered in 2001, the team is also investigating in utilizing mathematical models to optimize staffing levels. However, while the general topics do agree, the details presented in their work could not be transported to our project directly and the lack of detail in the mathematics involved with the model formulation in the paper prevented aid from the paper. [2]J. Balluck, E. Asturi, and V. Brockman, "Use of the ADKAR(R) and CLARC (R) Change Models to Navigate staffing model changes during the COVID-19 pandemic," Nurse Lead, Aug 20 2020, doi: 10.1016/j.mnl.2020.08.006.Balluck and colleagues discussed that in the early 2020, hospitals faced unprecedented patient volumes resulting from the COVID-19 outbreak. Nurse executives at a faith-based, not for profit healthcare system, quickly responded to ensure safe staffing, conservation of personal protective equipment (PPE) and implementation of infection prevention strategies. A significant challenge was safe staffing for the expected patient surge. To address this, a team of nurse executives utilized the ADKAR change model to guide a transition from primary to team nursing. The processes varied between hospitals, but core principles and implementation strategies were the same. Their article briefly discussed a case study at one health care system. The approach proposed in the work of Balluck seemed applicable to our staffing project. However, the ADKAR change model is not covered in the IOE (Industrial and Operations Engineering) curriculum and it would be time consuming to learn such a method from scratch. [3]K. M. Kester, M. Lindsay, and B. Granger, "Development and evaluation of a prospective staffing model to improve retention," J Nurs Manag, vol. 28, no. 2, pp. 425-432, Mar 2020, doi: 10.1111/jonm.12945.The work presented by Kester and colleagues aimed to improve predictability and accuracy of hiring using historical staffing data, quality improvement and workforce engagement. They used retrospective, secondary data analysis to develop a Prospective Staffing Model and conduct a five-year longitudinal evaluation of the implementation of the model in a convenience sample at a quaternary academic Cardiothoracic Intensive Care Unit. They used a team-based, quality improvement approach to restructure recruitment and hiring strategies, standardize new graduate nurse orientation and implement AACN Healthy Work Environment standards. It was resulted that retention increased (n = 286 days) and turnover decreased (17.6%) between 2014 and 2018; hence improvements in workforce stability were sustained. The work done by Kester was very similar to the topic of our project in staffing. While they used a prospective staffing model to improve nurse retention and to decrease turnovers, we will adopt a similar prospective approach in developing our model in staffing. [4]Robinson J, Porter M, Montalvo Y, et al. Losing the wait: improving patient cycle time in primary care. BMJ Open Quality 2020; 9:e000910. doi:10.1136/ bmjoq-2019-000910 This article focuses on cycle times of primary care clinics in particular. The primary goal was to determine a strategy to reduce cycle time in the family medicine clinic. The group examined inefficiencies such as lengthy computer searches, redundancy and lagging of insurance related work and a steady increase in check-in time throughout the day. The group was able to reduce cycle time by adjusting the MA schedule, streamlining the check-in process, and investing in additional staff in the front-office. The cycle time was reduced by 12% with these changes.The work done by Robinson was able to provide an idea of how various factors may affect patient cycle time. They made charts breaking down cycle time across the timeline of their project as well. The design of the charts inspired a few charts created by our group that showed variation in cycle time across clinics. These charts showed that while cycle time across all clinics didn’t seem to vary much across the time frame of our data, certain clinics were very much affected by the COVID-19 pandemic. These factors would be further analyzed in our final model AppendixAppendix A: ACUsAppendix B: Constraints and Standards Matrix for Optimizing a Staffing Model for the University of Michigan Medical Group Ambulatory Care UnitsAppendix C: Preliminary Pugh Matrix for Optimizing a Staffing Model for the University of Michigan Medical Group Ambulatory Care UnitsAppendix D: Updated Gantt ChartAppendix E: Screenshots of the Survey Appendix F: Extended Survey ResultsAppendix G: BCSC Structure VisualAppendix H: Cycle Time GraphsAppendix I: Model ScreenshotAppendix A: ACUsA Alfred Taubman Health Care CenterBriarwood Center for Reproductive MedicineBriarwood Center for Women, Children and Young AdultsBriarwood Health AssociatesBurlington BuildingDomino’s FarmsEast Ann Arbor Health and Geriatrics CenterFrankel Cardiovascular CenterKellogg Center - Ann Arbor - Westbloomfield - Huron River DriveWest Ann Arbor Health CenterLivonia Center for Specialty CareNorthville Health CenterSaline Health CenterYpsilanti Health CenterAppendix B: Constraints and Standards Matrix for Optimizing a Staffing Model for the University of Michigan Medical Group Ambulatory Care UnitsAppendix C: Preliminary Pugh Matrix for Optimizing a Staffing Model for the University of Michigan Medical Group Ambulatory Care UnitsAppendix D: Updated Gantt ChartFigure above is a detailed project plan and updated current progress indicated in Gantt Chart. It can be seen from the updated Gantt Chart that current progress indicates 100% completion. Appendix E: Screenshots of the Survey (Page 1 of 4)Appendix E: Screenshots of the Survey (Page 2 of 4)Appendix E: Screenshots of the Survey (Page 3 of 4)Appendix E: Screenshots of the Survey (Page 4 of 4)Appendix F: Extended Survey Results* (Page 1 of 18)*All data can be found in the attached ACU Clinic Manager Staffing Survey_December 3, 2020_11.16.xlsxLocation58 total unidentified clinicsIdentified ClinicsA. Alfred Taubman Health Care CenterBriarwood Center for Reproductive MedicineBriarwood Health AssociatesBriarwood Unsure whichBriarwood RadiologyBrighton Center for Specialty CareBurlington BuildingCanton Health CenterDomino's FarmsEast Ann Arbor Health and Geriatrics CenterFrankel Cardiovascular CenterKellogg Eye Center - Ann ArborLivonia Center for Specialty Care - Specialty ClinicsLivonia Health CenterRachel Upjohn BuildingSaline Health CenterUniversity of Michigan Rogel Cancer CenterWest Ann Arbor Health Center - Parkland PlazaYpsilanti Health CenterRolesPositionTotal Number of Managers that Manage each positionMA55PSA67LPN26RN26Other58Appendix F: Extended Survey Results (Page 2 of 18)List of Roles ManagedMSW, RDDental AssistantCall Center & PSIUS TechRadiologic TechnologistsVascular SonographersSurgery SchedulersAdministrative Assistant, Health Records AnalystsSpeech Language PathologyOphthy Techs and MPU tech'stherapiesAdministrative Assistants, Inventory Clerks, Call Center RepresentativesSurgery SchedulersAdmin AssistOphthalmic TechnicianRN manager is off site and we work together to manage this group.PA, NP, DietitiansPharmacists, Techs, Intake, DriversCall Center RepScribe, Ortho TechCall Center, clinic leadsPT, OTFertility NavigatorRTCall Center Reps, Athletic Trainers, radiology techsAdmin AsstCall Centerpartner with nursingDental AssistantsRadiation TherapistsPharm Techs and InternsAdministrative AssistantsExercise PhysiologistsI do not manage staffing directlyPanel ManagersAdministrationSpeech Language PathologistClinic LeadPT, PTA, ATC, Rehab TechsAudiologyNurse PractitionerCall Center RepresentativesAppendix F: Extended Survey Results (Page 3 of 18)Daily Staffing FactorsDaily Staffing FactorsTotal Number of Managers that use each factorPercentageNumber of total physicians and APPs in clinic5664.37%Special services that require extra staff4855.17%Complexity of visit types4248.28%A staffing model3337.93%Input or concerns from staff (RNs, LPNs, MAs & PSAs)3337.93%MGMA Benchmarks3034.48%Other factors2528.74%Input or concerns from physicians and APPs2326.44%FTE of physicians and APPs in clinic2225.29%Additional Factors Used when Determining Daily Staffing ScheduleAccess; next available appointmentStudy/testing volumes, inpatient volumes, surgeriesThis PSA's at this clinic schedule and check in/out for physicians and therapistsThe MGMA benchmark is what I held to when planning RN and LPN staffing for the activation of BCSC.You need to consider nurse assist work.Visit volumes and hoursProvider/Patient NumbersInterpreter Services neededCurrently only staffing COVID Hotline/Drive thru testing sitesVisit volume anticipatednumber of physicians and APP's virtualnumber of patient visitsCapacity adjustments as guided by MM safety protocolInput and concerns from therapists andProceduresNoneteam based careIn-Basket volumesIn basket volume, call volume, number of nurse visits, testing volume(echo/stress)Appendix F: Extended Survey Results (Page 4 of 18)Requesting Additional StaffFactorsPercentCountMGMA Benchmarks67.82%59Special services that require extra staff50.57%44Complexity of visit types48.28%42Number of total physicians and APPs in clinic48.28%42A staffing model35.63%31FTE of physicians and APPs in clinic35.63%31Input or concerns from staff (RNs, LPNs, MAs & PSAs)34.48%30Input or concerns from physicians and APPs32.18%28Other factors21.84%19Additional Factors Considered when Requesting Hiring Need to consider Nurse assist workHours of operationTesting site needs, clinic needs, under leadership directivesProgram growth and opportunities, competitionVolume MetricsAcuity of patients. MGMA does not provide benchmarking for wound clinics or hyperbaric clinicsCall center model, volumesservice levels (in-baskets/phone)Home developed acuity toolCompare direct patient care hours and estimated MA duty (e.g. in-basket, prior auths, etc.) to total MA hours available.New Patient Averages/new orders to access/ and total fill ratesnonein Basket volumes, referral volume, call volumeData: patient panel size, productivity data, workload volume MiChart data, Overtime, patient safety concerns, noted delays in care or unacceptable turnaround timesreview of data: in-basket messaging, aspect phone reports, overtime trigger reports.Appendix F: Extended Survey Results (Page 5 of 18)Alternative Models to MGMA BenchmarksCurrently use a different model than MGMA benchmarks for staffingCountPercentYes3440.48%No5059.52%Non-MGMA Staffing Model UsedNo modelWe created our own staffing models.MGMA does not provide an accurate basis for an academic REI clinic.Radiology staffing model-based on rooms, previous data regarding patient flow, and modalities of service.There really isn't a staffing model for testing, we just need to plan to have enough staff to cover exams rooms and inpatient volume across all of our sites.The MGMA staffing model is not totally accurate for a specialty clinic with physicians and therapists.MGMA looks at number of patients however you need to also consider the work that MA are doing who are in the nurse assist assignments.staffing metric with MGMAProvider presently working at site and patients they will seeIn addition to the MGMA benchmarks also look at the new patient growth over the years and the payroll expenses on the monthly financial reports.I determine staffing based off of clinic volume to technician work-up ratio/ length of work up time.While in GI and Hepatology, we developed a staffing model based on contact hours with patients, behind the scenes work to support patient care, load of work from MiChart, time to develop staff.Internally developed.Long-term staffing.I use MGMA as a reference/starting pointDepending on Provider coverage needed as well as limiting how many staff can be off on a given day. We also consider seasonal patient volumes.Number of provider sessions scheduled, type of specialty, acuity and ease of patient (podiatry patient with foot wound versus spinal cord injury with multiple pressure ulcers)I don't have a benchmark that is relevant to us for the most part. I look at the Erlang calculator for call center, the other call center metrics from aspect, hours of operation. For PSAs an MAs it is the volume, the overtime (historically), # of in-baskets, hours we are open, etc. For insurance authorization - work queues, current status of needs by insurance - e.g. require in person or can we use portal to submit authorization, etc.Identified support needs based on acuity and workflows required in an oncology settingBesides MGMA - I also determine the MA staffing by the number of providers in clinic as well as the volume of patients per provider. We try to be MA to provider. PSA is also determined by the volume of provider/patients.Appendix F: Extended Survey Results (Page 6 of 18)Non-MGMA Staffing Model Used (Continued)it is difficult to compare MGMA benchmarks to transplant. UNOS benchmarking is a "better" benchmark but still does not prove effective. I have worked with my teams to develop an acuity tool that we use in part to help with staffing determinations.Review of benchmarks combined with acuity and special needs of particular services. sometimes the model is different depending on the day and services / procedures that will be provided on particular days in particular services.Asst provider ratiosApex accreditationI use the Erlang Calculator for Call Agents, and patient volume for referral staffing.While MGMA is a good base being a academic medical center with current workflow & role expectations there can be variability based on physician practice to ensure patients are safely cared for.I also use the Erlang Calculator for call center staffing - it estimates how many call center agents are needed based on the volume and call length of incoming phone calls.While the model my staffing levels were based off was MGMA. I feel that it does not capture the complexity of our operations and the non-patient facing demand/responsibilities that attribute to our belief of needing more patient services staff to adequately support the volume of work in a timely manner.Erlang calculator, which we are working build and propose a different model that factors in outbound call work and other off the phone work.MGMA benchmarks do not apply well in ophthalmology who utilize ophthalmic techs primarily we utilize a staffing model based on productivity, volume of patients and specialized skill neededThere is no clear Ambulatory model measure for staffing.It depends on the ACU, I mostly use MGMA however, for Peds Oto there is a hearing aid business that requires staff not usually included in the MGMA surgical staffing models. This clinic still falls within MGMA benchmarking but the clinic runs very lean and looks average compared to MGMA.The MGMA benchmarks do not accurately allow us to compare apples to apples. As we continue to create standard work and shift the right work to the right people, I think MGMA could become a valuable tool. I have used AGMA for Advanced practice.Appendix F: Extended Survey Results (Page 7 of 18)Current System Staffing Level AccuracyThe current system accurately determines staffing levelsTotal Number of ManagersStrongly Disagree12Disagree0Somewhat Disagree10Neither agree nor disagree12Somewhat Agree31.00Agree0Strongly Agree15Percent DisagreePercent Agree27.50%57.50%Important Staffing Factors FactorAverage ImportanceTotal of managers that ranked each factor as the most importantPercent of Most Important FactorsTime-off allocations3.11251518.75%Virtual Visit4.262500.00%A variety of Visit types3.46251012.50%Full-time and part-time considerations3.7251215.00%Variation in staff hourly designations4.162545.00%The skill set and scope of practice for each staff role2.2753948.75%Appendix F: Extended Survey Results (Page 8 of 18)Important Staffing Factors - MAMATotal MA managers49FactorsTotal number of managers that ranked the factor as the most importantPercent ranked firstTime-off allocations918.37%Virtual Visit00.00%A variety of Visit types816.33%Full-time and part-time considerations714.29%Variation in staff hourly designations12.04%The skill set and scope of practice for each staff role2448.98%Important Staffing Factors - PSAPSATotal PSA Managers59FactorsTotal number of managers that ranked the factor as the most importantPercent ranked firstTime-off allocations1016.95%Virtual Visit00.00%A variety of Visit types813.56%Appendix F: Extended Survey Results (Page 9 of 18)Important Staffing Factors - MA (continued)FactorsTotal number of managers that ranked the factor as the most importantPercent ranked firstFull-time and part-time considerations1016.95%Variation in staff hourly designations46.78%The skill set and scope of practice for each staff role2745.76%Important Staffing Factors - LPNLPNTotal LPN Managers22FactorsTotal number of managers that ranked the factor as the most importantPercent ranked firstTime-off allocations418.18%Virtual Visit00.00%A variety of Visit types313.64%Full-time and part-time considerations313.64%Variation in staff hourly designations00.00%The skill set and scope of practice for each staff role1254.55%Appendix F: Extended Survey Results (Page 10 of 18)Important Staffing Factors - RNRNTotal RN Managers21FactorsTotal number of managers that ranked the factor as the most importantPercent ranked firstTime-off allocations419.05%Virtual Visit00.00%A variety of Visit types314.29%Full-time and part-time considerations29.52%Variation in staff hourly designations00.00%The skill set and scope of practice for each staff role1257.14%Additional Staffing FactorsAdditional factors that are important when considering staffing needsaccess -- next available appointment for current staff, and whether or not productivity goals have been met.we do not have an RN therefore all of this work in done by an APP or MAI do not have any part time staff.We are a 7 day a week operation therefore we must make sure we are adequately staffed throughout the week, based on IVF demands but also have adequate staff available on the weekends with little to no plexity of visitDoes time off include FMLA & EFMLA? Who is covering our nurse assists? How many patients are in clinic?It would be nice to have one tool that all are expected and accountable to, I am sure this is the intent. it also would be helpful to recognize opportunities to share staff across oversight groups. Ancillary service volume need to be accounted for when considering the support needed (IE: PSA's, include ancillary visits to help identify increase need for ancillary services).Appendix F: Extended Survey Results (Page 11 of 18)Additional factors that are important when considering staffing needs (continued)We match staffing needs to provider hours/building hours. Staffing is needed for extended hours which are also times where there are less providers to staff.FTE providersUnderstanding the critical staffing needs and complexity of the clinic you are staffing.ExperienceOur MA's are also required to process Prior Authorizations, these are very time consuming and it is very difficult to work on them in between patients. This work is better suited for staff with clinical experience and knowledge and we do not have any extra staffing to assign this work to exclusivelyIn clinic Provider needs are determined first.The above vary by role. We change shifts as needed.Some clinics have providers that place orders and reply to portal messages; in the clinics where they don't, we need more RN's and LPN'sStaff competency for specific areasPatient care responsibilitiesPhysical Layout of ACUComplexity of Care Required; Current Staff FeedbackAttempts to meet UMMG metrics and Patient satisfactionChildcare/COVID needsChaperone needs for sensitive visits. Clerical functions that require more than clerical knowledge.face to face pt visitsclinic volumes, chaperone needs, number of procedures being doneNumber of Calls/Call length/average duration/hold time/target to answer/wait timesMedical leaves, orientation, FMLA,WorkloadAnticipating leaves of absences.Daily Patient Schedulehaving the adequate number of staff to match the volume of work associated with each work group, while having the needed contingency to support time off (planned and unplanned).Staffing needs in the clinic are determined by the number of providers mainly. This question is difficult to answer as it is unclear if time-off allocations refers to providers or support staff.the acuity of the patient in clinic, volume of basket needs, education visits, number of MD/APP FTW;s in clinic.The work being done in addition to visits (in basket messages, pump downloads, etc.)Type of work needed,Hours of clinic, overbooks, providers who tend to do more complex procedures, etcAppendix F: Extended Survey Results (Page 12 of 18)Helpfulness of MGMA Benchmarks in determining staffing levelsMGMA Benchmarks are helpful for determining the staffing level of my clinicsTotal Number of ManagersStrongly Disagree20Disagree0Somewhat Disagree22Neither agree nor disagree13Somewhat Agree16Agree0Strongly Agree2Helpfulness of MGMA Benchmarks in determining staffing levels (Continued)Percent DisagreePercent Agree57.53%24.66%MGMA Benchmarks are NOT relevant for certain servicesMGMA Benchmarks are NOT relevant for certain servicesTotal Number of ManagersStrongly Disagree0Disagree0Somewhat Disagree10Neither agree nor disagree14Somewhat Agree24Agree0Strongly Agree25Percent HelpfulPercent Not Helpful13.70%67.12%Appendix F: Extended Survey Results (Page 13 of 18)MGMA Benchmarks are NOT relevant for certain rolesMGMA Benchmarks are not relevant for certain rolesTotal ManagersStrongly Disagree0Disagree0Somewhat Disagree8Neither agree nor disagree17Somewhat Agree22AgreeStrongly Agree26Percent HelpfulPercent Not Helpful10.96%65.75%Services and staffing staffing roles the MGMA benchmarks are not relevantRDs and MSWs don't have MGMA benchmarks.Benchmarks are a number determined by visits and faculty. each practice has independent needs which cannot be determined by benchmarksPsychometrist roles, and services related to psychotherapy and neuropsych testing.All of REIMedical Assistants - Orthopaedic surgery utilizes MA's for casting, bracing, procedure assisting, etc. which I'm not sure is factored into this benchmarking.Vascular sonographers...from what I've found and understand there are no benchmarks for ancillary testing services.There are admin roles in the clinic that are not reflected in MGMA. In addition, our clinic is 80% therapy 20% physician. The 80% therapy is not reflected in the MGMA. Nor are the roles of PSA's in a therapy clinic. They do much more than check in/out patients.There is not enough information for Cancer Service line to determine information.The staffing benchmarks don't allow the nursing profession to fully execute on the role of the ambulatory care nurse as described by our professional organization. Furthermore, the current staffing doesn't allow the RN to offload re-visits for chronically ill patients. Offloading these visits from providers allows providers to see more new patients which positions us well to achieve our mission for covering more lives.It is the position of the American Academy of Ambulatory Care Nursing that:1. RNs enhance patient safety and the quality and effectiveness of care delivery and are thus essential and irreplaceable in the provision of patient care services in the ambulatory setting.2. RNs are responsible for the design, administration, and evaluation of professional nursing services within an organization in accordance with the framework established by state nurse practice acts, nursing scope of practice, and organizational standards of care.3. RNs provide the leadership necessary for collaboration and coordination of services, which includes defining the appropriate skill mix and delegation of tasks among licensed and unlicensed health care workers.4. RNs are fully accountable in all ambulatory care settings for all nursing services and associated patient outcomes provided under their direction.Appendix F: Extended Survey Results (Page 14 of 18)Services and staffing staffing roles the MGMA benchmarks are not relevant (Continued)The benchmarks are too broad to be accurate for each particular situation. We are a stand-alone urology clinic in Muskegon with unique services in our local community. We do not physically work within the UM institution, so our workflows differ considerably when compared to those in the Urology Department in Ann Arbor.MA's who work in nurse assist.In the Cancer Center our MGMA benchmarks are based on primary care. Our services and patient population are much more acute and we have found that staffing levels for our benchmarks do not reflect our needs.We have transitioned to a model of care that involves higher volume of procedures in clinic and are offering appointments to patients who may not need a procedure to other sites. As a result the volume has stayed level but the amount of MA support per case has increased.PSA's may be supporting patients that are seen both in this clinic and other Oto Clinics. Its is difficult to prorate the FTE accurately skewing the staffing level to some degree.n/aMGMA seems to be a good fit for RN, MD's, When if comes to therapies the model quickly requires modifications and estimations based on what we think would be comparable.There is no benchmarking for Infusion services.For Call Center agents you can measure in-bound work but not outbound work.Inventory Supply personnel have no benchmarking.Other roles, such as Administrative Assistants are utilized very differently across the organization. Some are used to schedule for physicians while other are utilized for a wide range of services as they support a large health center. MGMA benchmarking does not account for the variety of tasks within this title.All Clinics are not created the same. its very hard to treat them alike.If you have an MA who is working in a clinic that only rooms a patient and takes vitals you should not need as many MAs as if you have a very busy clinic that requires a lot of skill and work from the MA position. If an MA is called in to do injections, assist on procedures, run bench labs, do EKG's, A1C's etc... this takes a bunch of time to do before the next patient may be roomed.My clinic has Gen Med & Peds and I don't believe MGMA supports the appropriate number of MA's needed to do the entire rooming and follow up work required by our MA's support providers in both of these services.Data is not available for subspeciality services like hepatology, IBD, IBS.Our business unit does not operate like a ACS clinic so there is not a direct comparison that allows use of MGMA.Appendix F: Extended Survey Results (Page 15 of 18)Services and staffing staffing roles the MGMA benchmarks are not relevant (Continued)MA and PSA - the benchmarks are low and do not encompass the details of the work in specialty clinics, as more staffing is needed for these clinics. The benchmarks seem to lean towards primary care, not specialty care.Call Center representatives based on the erlang calculator we need 5.5 staff- we have never had that. If we dip below 4 our calls back up and management/leads find themselves helping cover phones during peak times. If even one staff member calls in or people have time off it negatively affects operations.Medical assistants- in family medicine the MA's perform numerous tasks and are very high functioning, because they are not only rooming patients, but managing in-basket messages providers paperwork, prior authorizations, pre-visit planning, med and supply ordering stocking rooms and running PC Quality controls MA's need to be staffed at the very least 1:1 with providers. 2 Additional to act as support/float is critical in maintaining operations when there are call ins or vacationsMA is understated for a 1:1 model on a consistent basis for Primary care.RN & LPN is clearOther roles are not defined to know if an accurate comparison is being met.MGMA does not provide benchmarks for wound or hyperbaric clinics.Seems like the do not make sense given social work and other roles in the clinic - I don't know. I would have to hire more than 10 supportive staff to match the benchmark.RN's and LPN's - MGMA bases workload off clinic visits. During COVID clinic shutdowns we had no visits in some clinics and still the same, if not more amount of nursing encounters. Visits and nursing work are NOT directly proportionate. Additionally many of our providers do not practice the way they do at other Health Systems I have experience with. Many do not place orders, use MiChart, or reply directly to patients.The overall "primary care/some specialty" benchmark doesn't incorporate the significant multidisciplinary roles needed to manage a comprehensive cancer program.MA staffing since some do other roles in the clinic that are not direct patient care such as RX team, forms, Leads.LPN support can be widely ranged in terms of the responsibilities that are performed. I believe that the opportunity for work for LPNs per site is largely affected by the presence of residents. The use of virtual care has also increased the need for LPN support.RN FTE. The MGMA benchmarks don't capture the level of nursing care needed in the individual clinic areas.I think the MGMA benchmarks are a good source of information. Our clinic has worked for years below the 75% of the MGMA benchmarks, but we are still asked to trim and eliminate positions to the point that I'm not sure why we are using these benchmarks.The models don't seem to factor all of the work that is done or needed especially for the MA role.Audiologists and Speech Therapists are not consideredAppendix F: Extended Survey Results (Page 16 of 18)Services and staffing staffing roles the MGMA benchmarks are not relevant (Continued)again sharing that there are not MGMA benchmarks for transplant.. UNOS has the "most" transplant sensitive data, but again not comparable between centers. Homegrown acuity tool has helped staff to understand work loads but it is not validated.While it's good to have a benchmark to compare to, there are too many variables [e.g. staffing continuity clinics, number of APPs in the practice, the additional responsibilities we ask MAs to do, leaves of absence, etc.] to take a one size fits all approach.PSW because & Dental assts because out clinic has to do additional tasks others ido not have to as we work within two completely separate clinics and we have to do work based on dental needs. UMMG basis their decisions on medical needs and does not understand the needs as related to a dental clinic and or dental proceduresOur surgical multi specialty clinic doesn't even exist in the benchmarking data for "like ACU's" The staffing doesn't match our needs and although we are at or below the 75th percentile for MGMA benchmarks in PSA and MA when compared to the nearest clinic type, our nursing numbers far exceed the benchmarks and it is impossible to really use the benchmarks to look at the entire clinic.MGMA does not offer benchmarks in collaboration with radiation oncology services, this is a very different clinic with different factors for staffing, as well as visit types.I do not have a lot of experience with the MGMA benchmarks, but have not found them relevant for the call agent/Insurance/billing and PSA ACU staffing for therapy areas. I could need more information on utilizing these benchmarks for these staffing areas.RNs, LPNs. Without consistent role delineation practice and acceptance it is difficult to use MGMA as an accurate measurement. MGMA primarily measures practices that have full time clinicians who function differently than many of our clinicians.The Specialty clinics are difficult to align with MGMA benchmarks due to the complexity of the responsibilities.They seem to be based on private practices and primary care or specialty care, but we have multiple subspecialties in addition to specialty care and we use clerical staff to support a broad range of work that does not need to be clinical. The benchmarks for clerical staff do not seem accurate for large, quaternary care clinics. Especially when you consider things beyond scheduling such as medical records retrieval, obtaining prior authorizations, tracking outpatient testing such as blood draws and echocardiograms, etc.there is a ton of outbound work that is being done in the clinic. It's hard to put a number on this but with all of the inbasket work, referrals, virtual platform and the work that goes behind this effort, the immense fax server and the time it takes to get faxes to the right place...all of these pieces of work is not easy to benchmark.Appendix F: Extended Survey Results (Page 17 of 18)Services and staffing staffing roles the MGMA benchmarks are not relevant (Continued)My understanding is that our ACU is being held to the assumption that we operate under one specialty, when in fact there are subspecialty nuances that generate legitimate variability that workgroups have to navigate. It would be more relevant to be benchmarked to multi-specialty standards to capture complexity. Benchmark also seems to only capture the number of staff in the context of patient-facing activity, when in reality workgroups such as checkin/out are expected to support client-facing as well as clerical responsibilities.Our ACU is considered all primary care. However, we have OB/GYN, which is more a specialty clinic in actuality.I don't feel the benchmarks take extra tasks into consideration. We have a very large number of calls and in basket messages that affect all of our roles (nursing, MAs and PSAs).Staffing is very dependent upon what types of procedures are being performed in the clinic and the complexity of the patient population being seen at certain ACU's. The MGMA benchmark does not account for this level of detail and varies greatly from one clinic to the next.In terms of Urology at Taubman, we generally have patients who are of higher complexity. I am not sure MGMA benchmarks capture the level of patient that we see in TC.PSA- as mentioned earlier, We use PSAs in Peds Oto as Hearing Aid technicians as a much lower cost staff than a Audiology Assistant. This makes Peds Oto look like it has a high level of PSA positions when it actually runs very lean.Additional comments about staffingWe determine staffing needs based on patient populations that should be supported by ancillary teams, productivity levels, and access.We operate 365 days a year which makes it challenging.NAYou might consider two staffing models? One for MA's that are in patient flow and one for those who work in the nurse assist groups.We utilize MA scribes in clinic 5 days per week. This service is important to our faculty and if staffing is short we have to do serious changes to flowThank you for asking about this.Staffing off visits does not reflect nursing work completed. It would make more sense to staff off of numbers of completed nursing encounters.happy to discuss the tool that was homegrown.Staffing considerations should include 'current state' dialogue at the local level.The important work of staffing our clinics during a pandemic according to MGMA benchmarks should also take into account the current fiscal year goal for role delineation.I have found that staffing based on clinic hours and volume spread throughout the day and then factoring in chaperone needs and procedures works the best. We huddle every morning to assign rooming and roles for the MA team. the front desk staggers hours to accommodate lunches and early openings and late closings.We believe that it would be more beneficial if the MAs and LPNs were combined in relation to providers. That is a more accurate reflection of provider support.Appendix F: Extended Survey Results (Page 18 of 18)Additional comments about staffingI always pull benchmarks using criteria for my specific specialty (Metabolism/Endo) and looking at "FTE per 10,000 encounters," but I'm not sure that is the correct criteria. There are many options and filters that can be applied, so it would be helpful to have clarification on how to pull the correct numbers.Based on the MGMA benchmarks for my clinic I believe the benchmark for MA's is too high and too low for the PSA work. There are a variety of different roles that a PSA can do and it should break it down by the clerical needs and the medical records and referral coordination needs, they should not be combined.Appendix G: BCSC Structure Visual (Page 1 of 3)Appendix G: BCSC Structure Visual (Page 2 of 3)Appendix G: BCSC Structure Visual (Page 3 of 3)Appendix H: Cycle Time Graphs (Page 1 of 2)Figure 1: Monthly total appointment counts over life of BCSCFigure 2: Cycle times by percentile across all appointments at BCSC Figure 3: Min, Mean & Max Cycle Time over BCSC’s Lifetime - Dermatology Appendix H: Cycle Time Graphs (Page 2 of 2)275590223520Figure 4: Min, Mean & Max Cycle Time over BCSC’s Lifetime - Orthopedic Surgery Figure 5: Min, Mean & Max Cycle Time over BCSC’s Lifetime - Rheumatology Figure 6: Average Cycle times of RV & NP appointments, and RV & NP video visits over timeAppendix I: Model ScreenshotFigure 1: Flexible tool for PSA. Inputs are total visits per month and total in basket messages.Figure 2: Flexible tool for MA. Inputs are department name, and sessions per month.Figure 3: Flexible tool for LPN. Inputs are department name, and sessions per month.Figure 4: Flexible tool for RN. Inputs are department name, and sessions per month. ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download