Case Study High-Performing Health Care Organization • December 2008 Luther Midelfort Mayo Health System: Laying Tracks for Success J ennifer E dwards , D r.P.H. H ealth M anagement A ssociates The mission of The Commonwealth Vital Signs Fund is to promote a high performance Location: Eau Claire, Wis. health care system. The Fund carries out this mandate by supporting Type: Nonprofit, physician-led integrated health system, including three rural critical access independent research on health care hospitals (nonteaching) and 220-physician multi-specialty clinic with 12 outpatient locations issues and making grants to improve Beds: 305 health care practice and policy. Distinction: Top 1 percent of hospitals in composite of 22 process-of-care quality measures among roughly 2,000 hospitals (about half of acute care hospitals in the U.S.) eligible for this analysis; also scored in top 1 percent of hospitals for prevention of surgical infections. Timeframe: Second quarter of 2006 through first quarter of 2007. To be included, hospitals must have submitted data to the Centers for Medicare and Medicaid Services for all 22 measures, with a minimum of at least 30 cases for at least one measure in each of four clinical areas. See Appendix A for full methodology.      For more information about this study, Summary please contact: Exemplary quality scores at Luther Midelfort Mayo Health System in Eau Claire, Jennifer Edwards, Dr.P.H. Health Management Associates Wis., are the result of a long-term commitment to building quality and safety into jedwards@healthmanagement.com systems of care. Since joining the Mayo Health System in 1992, Luther Midelfort has advanced a culture that supports staff in their efforts to test new ideas and improve care. Although work focused on Centers for Medicare and Medicaid Services (CMS) core measures began recently, the newer efforts follow the phi- losophy and strategy established years earlier. In a recent analysis by The Commonwealth Fund, Luther Midelfort achieved the third-highest score in the coun- To download this publication and try on a composite of 22 process-of-care measures that all hospitals report to CMS. learn about others as they become available, visit us online at www.commonwealthfund.org and register to receive Fund e-Alerts. Commonwealth Fund pub. 1194 Vol. 2 2T he  C ommonwealth F und The key components of Luther Midelfort’s Strategies for Success quality strategy are: Luther Midelfort does not use exotic tools or unusual • exposing clinical leaders to the concepts of clinical strategies. It follows the plan-do-check-act quality improvement, then giving them time to (PDCA) approach, providing rapid feedback of quality work with teams to apply their knowledge; indicators to frontline staff and benchmarking their • creating expert, interdisciplinary teams for performance against other hospitals in the system. each clinical area to determine the right con- What may make Luther Midelfort successful is the tent of care and then turning over the imple- culture of experimentation and learning that underlies mentation to a team with expertise in design- its work. ing workflow and care processes; and Learning from the Quality Leaders • rapid measurement and feedback, supported According to Borman, a driving force behind Luther by a quality resources department that coaches Midelfort’s success is its longstanding relationship teams through improvement cycles. with the Institute for Healthcare Improvement and other thought leaders in the field, including the Juran Organization Institute and Everett Rogers. Hospital leaders have Luther Midelfort Mayo Health System is a physician- participated in training and improvement activities for led, integrated health system serving west central the past 15 years and have learned the best techniques Wisconsin. Luther Hospital was founded in 1905 by a in system improvement. “Quality and safety are prop- group of Norwegian clergymen. It grew over time to erties of well-designed systems,” says Borman. “So if have 305 beds and provide a full range of services, the outcome is not what’s needed, we look first to including comprehensive cardiac, trauma, and mater- redesign systems.” nity care. The system includes three rural critical Everett Rogers, in his book Diffusion of access hospitals. Midelfort Clinic is a 220-physician Innovations, advised that creating slack would spur multi-specialty clinic with 12 outpatient locations. The new thinking. Luther Midelfort recognizes that combined Boards of Directors include seven elected improvement is work and staff need dedicated time to physicians, five community members, three physicians do it. Thus, staff members are given time apart from from Mayo Clinic Rochester, and one Mayo Health their daily responsibilities to test new ideas. Some System administrator. The Boards of Directors set the physicians spend 10 to 30 percent of their time on mission and vision of the organization, determine pol- safety or quality improvement activities. Borman icy, and direct the management to implement policy. believes this allowance of time, paired with the tech- The objective of Luther Midelfort’s work on the niques of process redesign, quick cycle improvement, core measures is to provide the right care 100 percent and benchmarking, contribute to the hospital’s success. of the time. According to Terrance Borman, M.D., the medical director, following the care processes is “not Staff need time away from their rocket science.” Still, delivering recommended care for everyday work to test new ideas and innovate, every patient requires constant attention. Luther according to hospital leadership. Midelfort’s strategy is to “lay a track for all trains to run down”—getting the processes right so that care is delivered according to plan. Borman and his col- Specialized Teams leagues involve the staff members who will implement Many hospitals use teams to study, redesign, and mon- the care processes in their design. If a certain process itor quality improvement efforts. Luther Midelfort rec- does not provide the desired outcome, they keep work- ognized that staff members who have a particular clin- ing at it until they have made it reliable. ical expertise are not always closest to the delivery of L uther M idelfort : L aying Tracks for S uccess 3 patient care. Therefore, each quality improvement Rapid Measurement and Feedback effort is supported by two teams. The first team The third strategy Luther Midelfort has found to be defines the clinical standard—what should be done for critical to its success is rapid measurement of quality each patient presenting with a particular condition. The indicators and timely feedback to frontline staff and team is typically led by a physician and includes phy- the implementation team. Early in a project, process sicians, nurses, pharmacists, and others with expertise measures are collected weekly, even if only for a small in a particular clinical area. These teams report to one number of patients. Once the care processes repeatedly of the assistant medical directors. Appendix B illus- produce the right results, monitoring becomes trates the care processes for pneumonia treatment biweekly. Regular feedback help keeps priority areas developed by the pneumonia team. in the minds of staff. Subsequently, local data are The standard of care developed by the clinical joined with reports from across all Mayo sites to team is then handed off to a Hospital Implementation inspire competition, which Borman believes has pro- Team (HIT), which has experience in process redesign moted better performance. and can determine the best way to translate that stan- Luther Midelfort’s Quality Resources dard into a highly reliable system. The HIT serves the Department offers resources including advice on qual- entire hospital and focuses on how work is being done ity improvement techniques such as PDCA cycles and and any impediments to achieving the best outcome manual data checking systems. The Quality Resources each time. It charts progress on a weekly basis and Department also uses case managers to monitor continues tweaking care processes until the process of achievement of standards in real time. If they discover care is carried out consistently in the manner expected. a deficiency, they can alert medical leadership, who It may consult with the expert team at various stages can contact a physician while a patient is still in the of the project. hospital and address it. The HIT has members from all care delivery Soon, Luther Midelfort hopes to leverage its sites, so that implementation strategies are tailored to new electronic medical record system to support local needs, rather than a “one-size-fits-all” solution. reporting on core measures. Once the new processes are implemented, the Hospital Improvement Team signs off on the project. The expert Results team takes back responsibility for longer-term moni- Despite the fact that Luther Midelfort has a strong toring of results and quality control. Appendix C is a foundation in quality improvement, Borman says it tracking sheet for the pneumonia care process, illus- took one to two years to find the right way to improve trating how measures and process are connected care in the four clinical areas assessed through the core throughout the patient’s stay. measures (heart attack, heart failure, pneumonia, and surgical infections). Once the hospital worked out its Teams specialize: either they have care processes, it achieved outstanding results. Each of clinical expertise or process redesign expertise. the 22 measures is close to 100 percent reliable, and Both types of teams are needed. has been for most of the past two years (Table, page 5). During a recent month, there was a dip in one measure. Borman explored the potential causes with Another role of the HIT is to coordinate the the clinical expert committee. The care processes were multiple improvement efforts that occur on a hospital so well defined that it took little time to discover that unit. Making sure the efforts are staged, coordinated, one group was confused about the timing of the first and streamlined can prevent staff overload and burn- antibiotic administration for pneumonia patients. A out. In this role, the team becomes a critical link in clarification was communicated to staff and perfor- communication up and down the organization and mance levels on this measure bounced back. among staff members. 4T he  C ommonwealth F und Physician buy-in has been extremely high, Physician champions have been extremely help- which Borman attributes to the fact that the core mea- ful in bringing attention to this work. Their commit- sures are not that controversial. In addition, the care ment, paired with accountability that extends up to the processes are so well designed that they leave little Boards of Directors, leaves no room for doubt among room for argument. For example, order sheets are staff about the hospital’s priorities. To keep resources designed so that the right way to provide the care is focused on improvement, Luther Midelfort engages in also the easiest way to order it. strategic planning every 180 days to identify what will be done in the next six months. It collects data and Lessons Learned monitors progress. Since CMS instituted them, core Above all, Borman says Luther Midelfort has learned measures have been a focus of improvement and there- to focus on the process. “You can’t achieve better fore at the forefront of the organization’s plans and results just by encouragement,” he says. The hospital’s resources. strategy of putting together the right workgroups to “lay the tracks for all trains to run down” has been key For More Information to its success. For more information about Luther Midelfort’s quality Luther Midelfort also believes in giving people improvement strategies, contact: Terrance Borman, the tools they need to improve. They have mostly M.D., medical director, Luther Midelfort, or encouraged use of the PDCA approach to quality borman.terrance@mayo.edu or (818) 907-4540. improvement. They have also turned to Six Sigma and Also see www.luthermidelfort.org. Lean methodology when appropriate. L uther M idelfort : L aying Tracks for S uccess 5 Table. Luther Midelfort’s Scores on 22 CMS Core Measures Compared with State and National Averages National Minnesota Luther Midelfort Indicator Average Average Hospital Heart Failure Percent of heart failure patients given discharge instructions 69% 77% 92% of 163 Percent of heart failure patients given an evaluation of LVS function 87 89 99% of 213 Percent of heart failure patients given ACE inhibitor or ARB for LVS dysfunction 87 89 96% of 55 Percent of heart failure patients given smoking cessation advice/counseling 89 89 100% of 15 Pneumonia Percent of pneumonia patients given oxygenation assessment 99 100 100% of 191 Percent of pneumonia assessment patients assessed and given 98% of 198 78 84 pneumococcal vaccination Percent of pneumonia patients whose initial emergency room blood culture was 95% of 133 90 94 performed prior to the administration of the first hospital dose of antibiotics Percent of pneumonia patients given smoking cessation advice/ counseling 85 87 98% of 57 Percent of pneumonia patients given initial antibiotics within six hours after arrival 93 97 98% of 101 Percent of pneumonia patients given the most appropriate initial antibiotic(s) 87 90 94% of 104 Percent of pneumonia patients assessed and given influenza vaccination 75 75 96% of 56 Heart Attack Percent of heart attack patients given aspirin at arrival 94 97 99% of 118 Percent of heart attack patients given aspirin at discharge 91 95 100% of 241 Percent of heart attack patients given ACE inhibitor or ARB for LVS dysfunction 88 88 96% of 52 Percent of heart attack patients given smoking cessation advice/counseling 92 90 100% of 77 Percent of heart attack patients given beta blocker at discharge 92 94 99% of 264 Percent of heart attack patients given beta blocker at arrival 89 90 99% of 94 Percent of heart attack patients given fibrinolytic medication within 30 minutes no patients met 40 45 of arrival inclusion criteria Percent of heart attack patients given PCI within 90 minutes of arrival 67 81 74% of 23 Surgical Care Improvement/Surgical Infection Prevention Percent of surgery patients who received preventive antibiotics one hour 97% of 911 84 90 before incision Percent of surgery patients who received the appropriate preventive antibiotics for 100% of 924 91 95 their surgery Percent of surgery patients whose preventive antibiotics are stopped within 24 hours 98% of 813 82 88 after surgery Percent of surgery patients whose doctors ordered treatments to prevent blood clots 94% of 891 80 87 (venous thromboembolism) for certain types of surgeries Percent of surgery patients who received treatment to prevent blood clots within 24 91% of 891 77 84 hours before or after selected surgeries Note: ACE = angiotensin-converting enzyme; ARB = angiotensin receptor blockers; LVS = left ventricular systolic; PCI = percutaneous coronary intervention. Source: www.hospitalcompare.hhs.gov. Accessed on 10/24/08. Data are from CY2007. 6T he  C ommonwealth F und Appendix A. Selection Methodology Selection of high-performing hospitals in process-of-care measures for this series of case studies is based on data submitted by hospitals to the Centers for Medicare and Medicaid Services (CMS). We use 22 measures that are publicly available on the U.S. Department of Health and Human Services’ Web site, Hospital Compare (www.hospitalcompare.hhs.gov). The 22 measures, developed by the Hospital Quality Alliance (HQA), relate to practices in four clinical areas: heart attack, heart failure, pneumonia, and surgical infections. Heart Attack Process-of-Care Measures Percent of Heart Attack Patients Given ACE Inhibitor or ARB for Left Ventricular Systolic Dysfunction (LVSD) Percent of Heart Attack Patients Given Aspirin at Arrival Percent of Heart Attack Patients Given Aspirin at Discharge Percent of Heart Attack Patients Given Beta Blocker at Arrival Percent of Heart Attack Patients Given Beta Blocker at Discharge Percent of Heart Attack Patients Given Fibrinolytic Medication Within 30 Minutes of Arrival Percent of Heart Attack Patients Given PCI Within 90 Minutes of Arrival Percent of Heart Attack Patients Given Smoking Cessation Advice/Counseling Heart Failure Process-of-Care Measures Percent of Heart Failure Patients Given ACE Inhibitor or ARB for Left Ventricular Systolic Dysfunction (LVSD) Percent of Heart Failure Patients Given an Evaluation of Left Ventricular Systolic (LVS) Function Percent of Heart Failure Patients Given Discharge Instructions Percent of Heart Failure Patients Given Smoking Cessation Advice/Counseling Pneumonia Process of Care Measures Percent of Pneumonia Patients Assessed and Given Influenza Vaccination Percent of Pneumonia Patients Assessed and Given Pneumococcal Vaccination Percent of Pneumonia Patients Given Initial Antibiotic(s) Within 4 Hours After Arrival Percent of Pneumonia Patients Given Oxygenation Assessment Percent of Pneumonia Patients Given Smoking Cessation Advice/Counseling Percent of Pneumonia Patients Given the Most Appropriate Initial Antibiotic(s) Percent of Pneumonia Patients Whose Initial Emergency Room Blood Culture Was Performed Prior to the Administration of the First Hospital Dose of Antibiotics Surgical Care Improvement/Surgical Infection Prevention Process-of-Care Measures Percent of Surgery Patients Who Received Preventative Antibiotic(s) One Hour Before Incision Percent of Surgery Patients Who Received the Appropriate Preventative Antibiotic(s) for Their Surgery Percent of Surgery Patients Whose Preventative Antibiotic(s) Are Stopped Within 24 hours After Surgery The analysis uses all-payer data from the second quarter of 2006 through the first quarter 2007. To be included, a hospital must have submitted data for all 22 measures (even if data submitted were based on zero cases), with a minimum of 30 cases for at least one measure in each of the four clinical areas. Approximately 80 percent of U.S. acute care hospitals submitted data on the 22 measures. Approximately 2,000 facilities—about half of acute care hospitals—were eligible for the analysis. No explicit weighting was incorporated, but higher-occurring cases give weight to that measure in the average. Since these are process measures (versus outcome measures), no risk adjustment was applied. Exclusion criteria and other specifications are available at http://www.qualitynet.org/dcs/ContentServer?cid=1141662756099&pagename= QnetPublic%2FPage%2FQnetTier2&c=Page)  L uther M idelfort : L aying Tracks for S uccess 7 Appendix B. Luther Midelfort’s Care Processes for Pneumonia Patients Appendix B: Luther Midelfort's Care Processes for Pneumonia Patients Patient Self- Care Education Information Given Smoking Cessation Process of Educating Patient Accesses Luther Hospital Sources of Care Nursing Homes LH Emergency Dept Assisted Living Clinic Home Health/Hospice Regional Hospitals Home Assessment Differential Diagnosis History Exam Radiology Studies Oxygenation Assessment Laboratory Blood Cultures Timing Sputum Cultures Smoking Cessation Outpatient Treatment Counseling Inpatient Respiratory Therapy Medications Antibiotic Selection Antibiotic Timing Follow-up & Planned follow-up Maintenance evaluation/visit Vaccinations Source: Luther Midelfort Hospital, 2008 Source: Luther Midelfort Hospital 2008. H:\Excel\Chronic Care\Pneumonia Expert Team\Pneumonia Conceptual Model.xls rev 0505 jrw Appendix C: Inpatient Pneumonia Process and Measures - Example Acceptable Key Target Measure In Control Performance Notes/Remarks Patient Presents P1 100% Oxygenation Assessment Y Y 100% 90th P2 percentile Blood Culture Performed Y Y 100% (CCU/NICU pts) P 90th P3 Blood Culture Prior to First Assess Patient P3 percentile Antibiotic Dose N (+) Y 96% (ED pts) P Initial Antibiotic Selection for 90th Immunocompetent Patients P4 percentile (ICU & NonICU) Y Y 100% Placed on Pneumonia Initial Antibiotic Selection for 90th Immunocompetent Patients Standing Orders P5 percentile (ICU) Y Y 100% P1 P15 Initial Antibiotic Selection for 90th Immunocompetent Patients P6 percentile (NonICU) Y Y 100% P4 P6 P7 90th Treat Patient P7 percentile Antibiotic Timing (Mean) Y ? 107.2 median P5 P8 P9 90th First Antibiotic Dose Within 8 P8 percentile Hours of Hospital Arrival Y Y 100% 90th First Antibiotic Dose Within 4 Audit Chart P9 percentile Hours of Hospital Arrival N (+) Y 96% 90th Tobacco Cessation P1 P10 percentile Advice/Counseling N (+) Y 100% 90th Pneumococcal Screening 94.5% Decision to remove afebrile Discharge Patient P1 P11 percentile and/or Vaccination N (+) N criteria from order set P1 93% (quarter ending Feb 2006) P13 90th Influenza Screening and/or Decision to remove afebrile criteria P12 percentile Vaccination N (+) N from order set P13 All-or-None N (+) N 90% P14 No protocol on chart (ED) N (+) N 5.50% P15 No protocol on chart (NonED) new measure N 28.60% Appendix C. Inpatient Pneumonia Process and Measures – Example At each point (P1 through P13) in the process of care flow diagram, there is an opportunity to measure compliance with the care plan. In addition, charts origi- nating in the emergency department (ED) or not in the ED, are reviewed for the presence or absence of the protocol document (measures P14 and P15). Source: Luther Midelfort Hospital 2008. 8T he  C ommonwealth F und L uther M idelfort : L aying Tracks for S uccess 9 A bout the A uthor Jennifer Edwards, Dr.P.H., M.H.S., is a principal with Health Management Associates’ New York City office. Jennifer has worked for 20 years as a researcher and policy analyst at the state and national levels to design, evaluate, and improve health care coverage programs for vulnerable populations. She worked for four years as senior program officer at The Commonwealth Fund, directing the State Innovations program and the Health in New York City program. She has also worked in quality and patient safety at Memorial Sloan-Kettering Cancer Center, where she was instrumental in launching the hospital’s Patient Safety program. Jennifer earned a Doctor of Public Health degree at the University of Michigan and a Master of Health Science degree at Johns Hopkins University. A cknowledgments We wish to thank Dr. Terrance Borman for generously sharing his time, knowledge, information, and materials with us. Editorial support was provided by Martha Hostetter. This study was based on publicly available information and self-reported data provided by the case study institution(s). The Commonwealth Fund is not an accreditor of health care organizations or systems, and the inclusion of an institution in the Fund’s case studies series is not an endorsement by the Fund for receipt of health care from the institution. The aim of Commonwealth Fund–sponsored case studies of this type is to identify institutions that have achieved results indicating high performance in a particular area of interest, have undertaken innovations designed to reach higher performance, or exemplify attributes that can foster high performance. The studies are intended to enable other institutions to draw lessons from the studied institutions’ experience that will be helpful in their own efforts to become high performers. It is important to note, however, that even the best-performing organizations may fall short in some areas; doing well in one dimension of quality does not necessarily mean that the same level of quality will be achieved in other dimensions. Similarly, performance may vary from one year to the next. Thus, it is critical to adopt systematic approaches for improving quality and preventing harm to patients and staff.