At a Glance
- Hospitals can use a DRG-based tracer approach to analyze and correct inefficiencies in the flow of care delivery.
- The approach works best when the analysis is focused on a high-volume DRG.
- In the case of Mount Sinai Hospital in Chicago, this approach was instrumental in improving efficiencies of care processes associated with DRG 143: chest pain.
Using a detailed analysis of care processes related to a single, high-volume DRG (DRG 143: chest pain), Chicago's Mount Sinai Hospital achieved significant improvements in care resulting in reduced length of stay.
Shrinking profit margins, payer reimbursement that does not always cover cost of care, and hospital closures are just a few of the factors driving healthcare institutions to pursue cost efficiencies relentlessly. Yet the complexities of a hospital's operations, patient care, and regulatory requirements make it a daunting challenge for hospital managers to successfully eliminate those inefficiencies that undermine quality of care and result in poor use of resources.
A performance improvement team at Mount Sinai Hospital (MSH) in Chicago found that a DRG-based tracer approach to analyzing and correcting inefficiencies in the flow of care delivery offers a particularly effective way to meet this challenge. This approach analyzes the operational components impacting various aspects of patient care as defined by a specific DRG.
MSH was facing the compounded challenges of a Medicaid- and uninsured-dominated payer mix, inadequate and delayed government reimbursement, and facilities whose efficiencies were sometimes adversely affected by the age of the plant. MSH's performance improvement team perceived that one of the best strategies for meeting these challenges would be to focus initially on identifying opportunities to reduce length of stay (LOS).
The team also understood that this analytical approach works best when the analysis is focused on a high-volume DRG. To maximize the benefits of process flow improvement, therefore, the team decided to focus on identifying opportunities to reduce LOS associated with DRG 143: chest pain, due to its high volume and relative high cost of care. This application of the DRG-based tracer approach illustrates how it can be used to identify not only the opportunities to reduce LOS, but also the actions required to achieve those reductions.
In calendar year 2007 (CY07), MSH admitted 547 patients with this DRG, with a weighted average LOS (ALOS) of 2.36 days. At this ALOS, the average cost to the hospital per patient was $5,144. As a result of the performance improvement effort, the weighted ALOS for chest pain patients decreased to 1.88 days in CY08. The reduced LOS translated to an estimated savings of $535 per patient and an actual total savings of nearly $317,255 for the hospital.
The DRG 143 effort began with key stakeholder meetings. Attendees included clinical and administrative leaders from the emergency medicine, telemetry unit, cardiology, environmental services, transportation, and utilization review departments. As a basis for the analysis, the team created a detailed patient care process flow diagram, including clinical and support departments impacting patient flow, which could be used to identify opportunities for efficiency improvements. (See Web Extra: Tracer Analysis: Patient Care Process Flow, DRG 143.)
Although the team observed that several factors were influencing LOS, it decided to focus only on those factors that represented real, actionable opportunities for reducing LOS and/or improving patient care improvement. The team identified six such factors:
- Bed availability
- Bed turnaround (housekeeping)
- Use of the emergency department (ED) chest pain order set upon inpatient admission
- Physician practice variation
- Transportation delays
- Cardiology department's role on the telemetry unit
The 45-bed telemetry unit is the busiest medical-surgical nursing unit at MSH, with an average of 16 bed turnovers each day. Admission criteria are followed stringently as patients are assessed for placement on this unit. The ED staff was concerned that delays in the telemetry unit's acceptance of new patients were impeding patient throughput. The telemetry unit's nursing staff attributed these delays to multiple admissions to the unit at the same time, occasional short staffing on the unit, and beds listed as "unavailable" for various reasons. Blocked admissions due to "unavailable beds" included those resulting from delayed entry of discharge orders, delayed patient transport, and/or delayed bed making.
The situation was exacerbated by nursing bed coordinators not using the available communication tools, such as the electronic bed management system, which would highlight the bottlenecks. The bed coordinators instead would rely on information provided to them by nursing and admitting staff or by physically walking the floors to ascertain bed availability.
Solution: The project team determined that admission delays in the telemetry unit could be reduced by placing the electronic bed tracking screens in all units and by instituting housewide training and implementation of the tracking system. This solution created hospitalwide transparency regarding available beds, because all units could now easily view each unit's actual census, as well as the numbers of patients waiting to be seen in the ED or waiting for a bed assignment. Consequently, bed placement discussions are more effective and actual placements are executed more expeditiously.
Upon reviewing the typical patient discharge times on the telemetry unit, the team determined that afternoon housekeeping staffing was inadequate to meet the demand for rapid bed turnover. Analysis of inpatient bed turnover and ED admissions made it clear that this was a highly active time for the hospital.
Solution: To match bed turnover demands with resources, the team reviewed the environmental services department's staffing pattern to ensure that it was appropriately staffing to demand. Based on an analysis of activity patterns by time of day and by location, the environmental services staffing pattern was reorganized, deploying additional housekeepers to the telemetry unit during peak activity times.
Chest Pain Order Sets
Prior to the analysis, low-risk patients with chest pain in the ED had three sets of enzymes drawn, were admitted for observation, and then had a stress test ordered by the inpatient service.
The project team found that inpatient service staff were often failing to refer to patients' ED orders in a timely manner after patients were admitted, and in some cases, they were discontinuing the ED orders. This failure to reference ED records would sometimes cause patients to undergo duplicate troponin tests, setting back patient throughput by eight to 12 hours.
Other issues were that stress tests were available only during limited hours of operation and that patients with orders to avoid mouth feeding were sometimes receiving inadvertent feedings.
Solutions: By requiring that the ED order set be used with earlier diagnostic testing, MSH was able to reduce the need for admissions. As the project progressed, the team explored the idea of placing a treadmill in the ED to expedite stress testing. Patients with chest pain would have been able to complete stress tests in the ED before being admitted to an observation or inpatient location, thereby potentially avoiding admission, or at the very least, reducing LOS. However, the facility's configuration would not allow for this treadmill placement.
On further consideration of the problem, the team determined that the duplicate test could be avoided by requiring that all ED orders always be placed at the front of the patient chart and that attending physicians and nurses always address the chest pain order set during "hand offs" to ensure that everyone was aware of the patient's status within the care pathway (see Exhibit 2).
MSH used ongoing education and reinforcement of the process to ensure that nurses and physicians in the ED and on the telemetry unit would embrace these changes. Under the new standard handoff protocol, when patients are admitted to the telemetry unit, the ED nurse provides a verbal report to the telemetry nurse over the phone, following a specific "Situation-Background-Assessment-Recommendation" format. If the patient is admitted to the intensive care unit (ICU) or is on a medication drip, the ED nurse accompanies the patient on a cardiac monitor to the ICU or floor and provides the verbal bedside "hand off" in person. Regardless of the inpatient admission location, the original ED chest pain order set is now always attached to the ED chart and is sent with the patient to the inpatient location.
The treadmill relocation discussion led to a subsequent recommendation to revise the ED orders and require that ED personnel either order a stress test for patients at the time inpatient beds become available and or facilitate the performance of stress tests if they are to be performed while the patients are still in the ED.
In the end, the order set was revised for three reasons: to avoid duplication of diagnostic ancillary and laboratory tests, to improve timeliness for ordering stress tests, and to ensure coordinated handoff between emergency medicine and the inpatient service. Because the revised order set prescribes that stress tests be performed earlier in the protocol, MSH quickly saw a 3 percent decline in low-risk patients with chest pain being admitted for observation and a corresponding increase in the percentage of these patients being discharged from the ED after testing negative. And as indicated previously, overall LOS for patients with chest pain dropped from 2.36 days in 2007 to 1.88 days in 2008.
Physician Practice Variation
The team also analyzed the impact of the physicians' role in the control of care management, and identified a wide range of practice variation among physicians, with many physicians significantly exceeding the benchmark average LOS for chest pain patients.
Solutions: MSH undertook an educational initiative to raise the medical staff's awareness of the impact of their care management on LOS. To alert physicians of their performance against MSH's LOS goal of 2.0 days (i.e., Medicare's recommended average LOS) and the performance of their respective physician peer groups, each physician's average LOS weighted by the volume of patients managed was shown using quadrant-style scattergrams that would serve as dashboards.
Public recognition was given to the top performers at various quality improvement meetings as well as in monthly patient throughput newsletters. The dashboards also were used as a foundation for dialogue between senior physician leaders and physicians whose LOS was higher than the benchmark average LOS for chest pain patients.
This strategy proved to be a primary driver of the project's sustained success, and the change in physician practice as a result of using these dashboards was significant. To illustrate, the first scattergram (see Exhibit 3) shows that only one physician managing more than 20 patients in a year reached the benchmark LOS goal of 2.0 days. However, in the second scattergram (see Exhibit 4), three physicians managing more than 20 patients not only met the benchmark goal, but also accounted for extremely high volume of cases.
Before the intervention, MSH's transportation process was rife with inefficiencies: At time of discharge, a unit secretary called a third-party dispatcher, who in turn would pursue an available transporter. When found, the transporter would have to comb the floors in search of available equipment and eventually arrive to pick up the patient. The transporter would leave a note in the unit secretary's logbook to document that the patient was picked up. Eventually, the nurse or unit secretary electronically discharged the patient, triggering the required housekeeping. The environmental services supervisor in turn would contact staff to assign the bed to be cleaned. This process was very passive and required multiple layers of communication.
Solutions: The process was redesigned to eliminate the lag time caused by multiple handoffs. The need for transporters to search for equipment not only delayed a patient pickup, but also posed a hazard and created clutter in hospital's hallways. A common central parking location near a service elevator was designated for all equipment not in use.
New software was integrated into the transportation process, automating the transportation operation. Unit secretaries and nurses now place a call into a computerized system and follow the preset prompts to customize their transportation request. Certain destination codes designate beds as dirty, simultaneously paging the next available housekeeper. After the request is registered, the system automatically dispatches the next available transporter by pager. The transporter acknowledges the request by calling the system, registering the job from "pending" to "dispatched." The transporter calls the system once again when the patient is picked up and one final time to register the request as "complete" when the patient finally arrives at his or her destination.
The new process not only reduces idle times, but also eliminates the guesswork related to transporter's availability and location, as caregivers have access to tracking tools that monitor the progress of their requests from start to finish. The application also provides valuable data for ongoing performance improvement by continually computing turnaround averages that help create expectations for real-time start-to-finish time, which department leaders can use as a management tool. Further, the application can assist transportation managers by automatically assigning lunch breaks through real-time recognition of appropriate low-volume periods.
A preintervention sampling of transportation data shows an average turnaround time of about 45 minutes from request to completion. Since the implementation, the recorded average transportation turnaround time has been reduced to 32 minutes, or about 29 percent (see Exhibit 5).
Cardiology Role on Telemetry Unit
Stress test readings are first completed by a cardiology fellow and then dictated and signed off by a cardiology attending physician. The dictation then is transcribed and the test is considered final. Before the project was initiated, the stress test turnaround process was loosely monitored. Although final test results are not typically required to discharge a patient, occasional discharge delays would occur because several physicians strongly preferred to wait for an attending physician's final read of the test before discharging patients.
Solutions: The cardiology director has implemented tighter controls and now closely manages all intervals of the stress test turnaround process. Benchmark expectations were developed for turnaround time from first to final reading.
Voice recognition as opposed to transcription services greatly reduced both absolute time and consistency of time to completion. The average turnaround time for completing stress tests was held under a six-hour goal for five consecutive months (see Exhibit 6).
Other practices at MSH also contributed to the success of the LOS reduction initiative, including daily rounding on the telemetry unit by two cardiology hospitalists, who are also responsible for ensuring that patients meet criteria for ongoing monitoring. These physicians further help by identifying patients who are ready for discharge, expediting the discharge process. Although generally charged with overseeing that all telemetry patients meet telemetry criteria, the cardiology hospitalists were engaged to focus extra attention on chest pain patients leading to an expedited discharge from the unit if not the hospital.
Impact on Patient Satisfaction
MSH has contracted with a consultant to measure patient satisfaction based on a statistically significant number of completed telephone surveys. The project leaders decided to utilize the patient satisfaction scores from the ED as a proxy for patient satisfaction regarding their care associated only with chest pain, given that the preponderance of chest-pain-related care occurs in this department. From 2006 to the present, the percentage of patients rating the overall quality of care as "excellent" in the ED increased by 11 percent.
Impact on Cooperation
By creating transparency and using positive reinforcement techniques, SMH gained the cooperation of various constituents involved with chest pain care. The installation of electronic "big boards" on each nursing unit helped physicians and nurses not only better manage the flow on their units, but also visualize patient flow and bottlenecks on all of the other units, including the ED. A monthly newsletter, Chest Pain Express, was developed to publicly recognize the "chest pain heroes" who contributed to reducing the average LOS. Finally, the physician scattergram dashboards served as visual aids for physicians to compare their ALOS performance against that of a group of their peers, ultimately sparking competition and driving performance enhancement.
MSH's performance management team points to three important lessons learned, based on their success with using the DRG-based tracer approach with DRG 143.
Use a focused, rather than global, approach. Before settling on a DRG-specific approach to reducing LOS, the hospital had attempted to identify organizationwide barriers that were common to all diagnoses. Each process identified as an "LOS barrier" was assigned to a small group of caregivers charged with developing and implementing solutions to that discrete piece of the process. There was no one tracking mechanism and no team leader assigned to ensure that process changes were implemented and effective.
Switching to a narrowly defined DRG approach enabled project leaders to focus on LOS-reduction challenges that would produce the greatest impact. Focusing on a single DRG resulted in clarity on benchmarks and best practice expectations that a global approach could not produce.
Take the time to develop a thorough flowchart of the process. Developing a detailed flowchart of the patient care process for presentation of chest pain was one of the initiative's first activities. Although it may be tempting to skip this part of the process, given that the interviews with each caregiver involved are time-consuming, the precise detail obtainable through this effort is critically important. Only with such detail can an organization develop an instructive tool for identifying where significant, sustainable modifications can be made to a care plan.
Address the importance of overcoming a culture of ingrained habits and processes. An organization's culture can be one of the biggest barriers to achieving positive interventions. In many organizations, people fall into the trap of accepting the status quo simply because they are accustomed to following familiar, yet poorly designed, routines. To combat this tendency, MSH's project leaders used a careful approach balancing the introduction of innovative ideas with positive reinforcement and public recognition of success.
Using process improvement techniques in the context of the DRG-based tracer approach, MSH was able to reduce LOS for patients with chest pain by nearly 0.5 days, achieving approximately $317,000 in savings annually. MSH achieved this result by eliminating inefficiencies; reducing variation through the implementation of standard order sets; using technology to manage assignments of functional departments such as environmental services, transportation, and bed management; and creating transparency around LOS data at the physician level. Exhibit 7 illustrates both MSH's average LOS reduction trend and the hospital's improved control over the patient flow process, as evidenced by the bandwidth reduction between the upper and lower control limits.
This project's success also could be attributed to the ongoing support of senior hospital leaders, who ensured that recommendations made by the project team were adopted and executed. To ensure the long-term sustainability of the project, the team will continue to monitor key operational metrics. The lessons learned from this initiative will help drive continued success as MSH identifies other high-cost/high-opportunity DRGs and replicates the effort in those areas. As a result, MSH will be better able to deliver high-quality, cost-effective health care to Chicago's west side, with a balance sheet that reflects the hospital's commitment to maintaining financial solvency to serve its mission.
Jack Gottlieb is manager, material and clinical services, Sinai Health System, Chicago.
Aalia Khawaja is manager, executive analytic service, clinical data and information, University Heath System Consortium Oak Brook, Ill. (authorship occurred during her tenure as director of service lines, Mount Sinai Hospital, Chicago).
Karen Teitelbaum is executive vice president/chief operating officer, Sinai Health System, Chicago.
Alan Channing is president and CEO, Sinai Health System, Chicago.
Sidebar: About Mount Sinai Hospital
Located on Chicago's west side, Mount Sinai Hospital (MSH) is a 431-bed teaching, research and tertiary care facility, founded 90 years ago in response to the health needs of an immigrant community. MSH is a member of Sinai Health System, which includes Schwab Rehabilitation Hospital, Sinai Community Institute, Sinai Urban Health Institute, Sinai Medical Group and Sinai Children's Hospital. The hospital serves a predominantly low-income population, and its dependence on Medicaid is a reflection of the community it serves.
Exhibit: Mount Sinai Payer Mix 2007
MSH provides specialty care for a broad community of 750,000 people within its service area. Diagnostic and treatment capabilities include a Level I trauma center; an open heart surgery program; digital mammography; a child-abuse diagnosis, treatment, and prevention unit; and a high-risk maternity service with 4,000 annual deliveries and a Level III perinatal center. In the past 10 years, MSH opened a state-of-the-art maternity care center and a completely renovated and expanded emergency and trauma facility. Cardiovascular care is a core competency, with plans to add a second cardiac catheterization laboratory in the next year.
As a major teaching hospital, MSH trains more than 300 health professionals each year in undergraduate and graduate medical education, allied health, and continuing education for physicians, nurses, and therapists. Research activities at MSH include clinical trials and community interventions.
Publication Date: Tuesday, June 01, 2010