BID indicates twice daily; DC, discharge; PT, physical therapy; and SW, social work.
The shaded area represents the project implementation period. The mean length of stay (LOS) (after adjusting for covariates) for the preproject, project, and sustainment periods are displayed above the chart (P < .001). CL indicates center line; LCL, lower control limit; and UCL, upper control limit.
The non–Veterans Affairs care costs for total hip arthroplasty and total knee arthroplasty are shown for fiscal years 2008 through 2012.
The number of cases of total hip arthroplasty and total knee arthroplasty in Veterans Affairs (VA) and non-VA care are shown for fiscal years 2008 through 2012. The proportion of patients using non-VA care showed a statistically significant decrease each year (P < .05).
Gayed B, Black S, Daggy J, Munshi IA. Redesigning a Joint Replacement Program Using Lean Six Sigma in a Veterans Affairs Hospital. JAMA Surg. 2013;148(11):1050-1056. doi:10.1001/jamasurg.2013.3598
In April 2009, an analysis of joint replacement surgical procedures at the Richard L. Roudebush Veterans Affairs Medical Center, Indianapolis, Indiana, revealed that total hip and knee replacements incurred $1.4 million in non–Veterans Affairs (VA) care costs with an average length of stay of 6.1 days during fiscal year 2008. The Joint Replacement Program system redesign project was initiated following the Vision-Analysis-Team-Aim-Map-Measure-Change-Sustain (VA-TAMMCS) model to increase efficiency, decrease length of stay, and reduce non-VA care costs.
To determine the effectiveness of Lean Six Sigma process improvement methods applied in a VA hospital.
Design, Setting, and Participants
Perioperative processes for patients undergoing total joint replacement were redesigned following the VA-TAMMCS model—the VA’s official, branded method of Lean Six Sigma process improvement. A multidisciplinary team including the orthopedic surgeons, frontline staff, and executive management identified waste in the current processes and initiated changes to reduce waste and increase efficiency. Data collection included a 1-year baseline period and a 20-month sustainment period.
Main Outcomes and Measures
The primary endpoint was length of stay; a secondary analysis considered non-VA care cost reductions.
Length of stay decreased 36% overall, decreasing from 5.3 days during the preproject period to 3.4 days during the 20-month sustainment period (P < .001). Non-VA care was completely eliminated for patients undergoing total hip and knee replacement at the Richard L. Roudebush Veterans Affairs Medical Center, producing an estimated return on investment of $1 million annually when compared with baseline cost and volumes. In addition, the volume of total joint replacements at this center increased during the data collection period.
Conclusions and Relevance
The success of the Joint Replacement Program demonstrates that VA-TAMMCS is an effective tool for Lean and Six Sigma process improvement initiatives in a surgical practice, producing a 36% sustained reduction in length of stay and completely eliminating non-VA care for total hip and knee replacements while increasing total joint replacement volume at this medical center.
The Veterans Health Administration (VHA) provides care to more than 8 million patients annually at a cost of nearly $50 billion.1 Although the population of veterans is declining, total expenditures and medical expenditures for this population are increasing. The rising costs of medical care suggest that use of the Veterans Affairs (VA) health care system is likely to increase. The VA system has acknowledged its limitations in resources and infrastructure by allowing individual facilities to contract with non-VA health care providers or private-sector facilities (non-VA care) to provide services when VA medical facilities are not “feasibly available.”2,3 However, non-VA care costs have doubled in recent years, increasing from $2.2 billion in fiscal year 2007 to $4.5 billion in fiscal year 2011.4 The excessive contribution of non-VA care to the overall cost of providing care for veterans has made this a target for cost-reduction efforts.
The Richard L. Roudebush VA Medical Center (RLR VAMC) is located in Indianapolis, Indiana. The RLR VAMC is part of the Veterans in Partnership Healthcare Network (VISN 11) regional network, 1 of 21 regional networks in the VA system. A VA central office request to reduce non-VA care costs within VISN 11 prompted a review of administrative data, which identified orthopedic care as a significant contributor to VISN 11 non-VA care costs. This led to an internal data review at the RLR VAMC in Indianapolis where the executive management team identified total hip arthroplasty (THA) and total knee arthroplasty (TKA) as the orthopedic surgical procedures requiring the highest non-VA care costs, totaling $1.42 million for this 1 facility in fiscal year 2008. As a direct result of this investigation, THA and TKA became target areas for increased efficiency to reduced non-VA care.
Key stakeholders within the joint replacement process at the RLR VAMC used Lean Six Sigma (LSS) methods to increase efficiency by reducing waste in THA and TKA perioperative care pathways (throughout this article, the terms LSS and Vision-Analysis-Team-Aim-Map-Measure-Change-Sustain [VA-TAMMCS] may be used interchangeably). Team members applied LSS techniques including the identification of value, value stream mapping, waste reduction, and establishing single-piece flow to influence process changes. This project was named the Joint Replacement Program (JRP); results and insights from this redesign initiative are discussed herein.
The RLR VAMC is a 180-bed, level 1A facility (the highest level of patient complexity in the VA system) providing care to more than 60 000 veterans annually. It is 1 of 8 medical centers in VISN 11. The JRP was conducted at the RLR VAMC from November 1, 2009, through July 31, 2010, following the VA-TAMMCS model of system redesign—a process improvement framework developed within the VA from Lean and Six Sigma principles. A summary timeline of significant events and changes associated with the JRP is shown in Figure 1. This study received institutional review board approval from the Indiana University Office of Research Administration and concomitant approval from the RLR VAMC Research and Development Department. The requirement for informed consent was waived as the study was determined to present low risk to study participants.
A multidisciplinary team formed prior to project initiation included orthopedic surgeons, an orthopedic nurse practitioner, a registered nurse, physical therapists, an occupational therapist, a patient safety advocate, and a member of the executive management team. The project scope was focused on processes outside the operating room (OR); staff who work exclusively with intraoperative processes (OR personnel, central processing) were not included in the project team. Members of the JRP team conducted a literature review that identified a non-VAMC with significant prior experience redesigning joint replacement perioperative processes.5 A site visit to this facility to observe perioperative processes helped to establish benchmark practices and provided considerable baseline information to inform interventions.
The JRP team met weekly for a 9-month period from November 1, 2009, through July 31, 2010. The initial meetings focused on mapping perioperative processes prior to implementing changes (the current state). After mapping the current state, the project team designed a future state with single-piece flow for the patient from the preoperative clinic through hospitalization. Single-piece flow is a Lean principle that refers to connecting process steps. In the current state, work processes for the preoperative clinic, hospital admission, and postoperative care were largely disconnected from one another. The team identified barriers preventing implementation of the future state and conducted measurements to further characterize the degree of waste associated with each barrier. Finally, the team designed changes to address identified barriers and performed small tests of change to evaluate the effectiveness of proposed changes for reducing waste. The team identified metrics to capture changes in efficiency and gathered data to objectively support these small tests of change. Successful tests of change led to broader process change implementation. After the conclusion of the project period, the JRP team met weekly to review length of stay (LOS) data for 20 months during the sustainment period (August 1, 2010, to March 31, 2012). These data were also displayed prominently on the orthopedic recovery ward.
Following the sustainment period, a medical record review identified patients undergoing joint replacement prior to, during, and after the project period (n = 638) to determine the effect on LOS. Patient demographic characteristics were compared for each period (Table 1). Continuous variables were compared between the 3 groups using analysis of variance, and categorical variables were compared using Fisher exact test.
A general linear model was fit to log-transformed LOS including the categorical variable of group (before, during, and after process implementation) and covariates of patient age, sex, type of surgery (TKA or THA), and an indicator for revision surgery. A smearing estimate6 was used to estimate mean LOS for each period and to calculate the difference in LOS between patients after implementation and before implementation. The analysis was completed using SAS version 9.3 statistical software (SAS Institute, Inc).
In a secondary analysis, the team collected and analyzed data on cost and volume of non-VA care for total joint replacements from October 1, 2008, to September 30, 2012. This analysis period provided a wider scope than the LOS data collection period, allowing a comparison of 5 complete VA fiscal years. The proportion of patients using non-VA care for each fiscal year was estimated along with 95% confidence intervals. Fisher exact test was used to compare fiscal years. Cost savings were calculated as the reductions in non-VA care costs. Non-VA care cost data were derived from 2 sources. Fully outsourced care costs (care provided in non-VA facilities) were extracted from the non-VA care inpatient episode of care cube, an administrative database of payments made by the VA for non-VA care. Remaining non-VA care resulted from care provided within the Indianapolis VAMC by non-VA physicians under a scarce medical care contract. Scarce medical care contract costs were gathered directly by referencing archived contracts. Neither method explicitly accounted for OR time, although the data cube necessarily included OR charges as part of outsourced surgical care. Descriptive statistics were only estimated for cost information owing to the complexity of the derived costs. Finally, subtracting direct project costs from the cost-savings estimate produced an estimated return on investment (ROI).
Significant events and changes around the time of the JRP are summarized in Figure 1. Project team members made changes to address barriers and waste identified through application of the VA-TAMMCS model. Barriers and waste identified from the perspective of patients and various health care providers are listed in Table 2. Changes implemented as part of the JRP are shown in Table 2, also from the perspective of patients and various health care providers.
Surgeon-specific barriers included inadequate OR time, unclear expectations for discharge, and the extra work required for patients with delayed discharges. The OR time barrier was addressed by adding block time on Tuesdays. The OR case load was shifted earlier in the week owing to the unavailability of social work support on the weekends. A 1.0 full-time-equivalent surgeon was hired out of the scarce medical care contract group with goals of eliminating the scarce medical care contract and reducing non-VA care for THA and TKA. This recruitment was anticipated prior to project initiation and not a direct result of the project. Finally, administrative support was clarified supporting discharge as soon as patients were medically cleared with a safe and appropriate discharge destination arranged.
Lack of standardization was a common theme in the identified barriers without standard care pathways for patients or standard work for caregivers. Many changes focused on standardizing work to account for the needs of patients undergoing total joint replacement. For example, standard work for physical therapy specified exercises focused on developing functional independence, and a rehabilitation room was set up on the inpatient orthopedic surgery ward to decrease transportation time and permit group therapy sessions. A social worker was dedicated to the THA and TKA patient population, and a new social work encounter form was designed to address specific needs of this patient population.
The initiation of a mandatory preoperative class for all patients undergoing THA and TKA (and a caregiver) allowed staff to educate patients on the process and set expectations for postoperative rehabilitation. From the physical therapy and social work perspectives, the class permitted baseline functional status assessments; anticipating discharge needs expedited the postoperative discharge preparation process.
Table 1 shows a 36% reduction in mean LOS after adjusting for covariates—from 5.3 days during the preproject period to 3.4 days during the sustainment period (P < .001). The control chart in Figure 2 demonstrates change in monthly LOS average over time.
Trends in non-VA care cost and volume per fiscal year are shown in Figure 3 and Figure 4. One hundred twenty-two TKA and THA cases used non-VA care at the RLR VAMC in fiscal year 2008. The percentage of patients using non-VA care services continued to significantly decrease across fiscal years from 65.8% (95% CI, 61.2%-75.3%) in fiscal year 2008 to only 1.2% (95% CI, 0.3%-3.5%) in fiscal year 2012. All non-VA care for total hip and knee replacements at the RLR VAMC was eliminated early in fiscal year 2012, representing a 100% reduction in non-VA care for THA and TKA.
Cost reductions were calculated from the elimination of non-VA care, which cost $1.4 million in the baseline year (fiscal year 2008). Care provided by other facilities contributed $900 000 to baseline non-VA care cost. The remaining $517 000 resulted from provision of care in the VA facility by non-VA surgeons (scarce medical care contract care). A 100% reduction of non-VA care for THA and TKA was achieved in fiscal year 2012. An additional orthopedic surgeon was hired in the middle of fiscal year 2010 (out of the scarce medical care contract group). The difference between the cost of the scarce medical care contract and the salary of the new surgeon was a net savings of $117 000. The net overall cost savings was $1.02 million using fiscal year 2008 expenses and case volume as a baseline.
The ROI was calculated as the difference between the cost savings and the cost of implementing the JRP. Implementation cost was estimated at $25 000 and took into account the opportunity cost for training and direct travel costs for the site visit. Aside from the hire of an additional orthopedic surgeon (accounted for earlier), there were no costs related to new equipment, staff, or other expenses. The estimated ROI for this project was $1 million annually.
The LSS methods have been adapted for use in health care with increasing frequency during the past 10 years. Many organizations have branded their own LSS methods to accommodate organizational structures, languages, and cultures. The VA system has also developed its own improvement model—the VA-TAMMCS model.
The JRP used VA-TAMMCS to improve efficiency within perioperative joint replacement processes at the RLR VAMC. Single-piece flow was established for patients undergoing THA and TKA through the preoperative and postoperative care processes. Changes reduced wasteful steps in the process and ultimately produced a 36% reduction in LOS (P < .001) while eliminating non-VA care costs and increasing THA and TKA volume. These outcomes were sustained and even showed continued improvement through the sustainment phase. Published data reporting LOS in non-VA facilities currently show an average LOS of 3.8 days following TKA7,8 and 3.5 to 4.0 days following THA.9- 11 In comparison, the JRP achieved a sustained LOS average of 3.4 days in the RLR VAMC for both procedures during the sustainment period.
To understand the systemwide impact of this project, data were gathered reflecting the function of the OR suite and related departments before and after the project. Overall OR volume and orthopedic surgical volume each increased annually from fiscal years 2008 to 2012, with each achieving 16% cumulative growth. Figure 4 demonstrates a 41% increase in total joint replacement volume during this same period. Delay and cancellation data revealed that cancellation rates were reduced by 50% during the data collection period, while delays remained unchanged. Operating room overtime did not increase, and there was no noted decrease in OR function. There were no limitations related to the availability of equipment or prosthetics with increasing joint replacement volume. High rates of flash sterilization for total joint replacement instrument sets were noted during the JRP process observations, prompting additional LSS projects to eliminate flash sterilization for joint replacement cases, reducing costs and improving safety.
Compared with private health care facilities, VA health care facilities present a unique environment for implementing LSS process improvement initiatives. Local, regional, and national administrative layers exist between frontline staff and the functional equivalent of a board of directors, the US Congress. This lengthy hierarchy introduces more sources for top-down, policy-driven initiatives, which can lead to organizational variability and challenge application and sustainability of process improvement. Another notable feature of this system is the absence of profit motive (or revenue generation, more accurately), creating a culture in which incentives for productivity are necessarily something other than revenue.12 As a result, VA staff may have more latitude to absorb the short-term loss in productivity from dedicating employee time to improvement initiatives. (Individual employees in the VA system are talented and committed to providing excellent care to veterans. The federal government’s exemption from a balanced budget requirement may indirectly influence the culture of the system by allowing more latitude for decreased productivity, which is actually a benefit for LSS implementation as a short-term loss of productivity is anticipated to produce a worthwhile ROI [monetary or nonmonetary], the ultimate goal of any LSS initiative.) These factors suggest that in the VA health care system, LSS initiatives may be easier to implement but more difficult to sustain.
The RLR VAMC in Indianapolis differs from other VA facilities in several respects. It is the first VA facility to implement process improvement work using a Lean Enterprise Deployment strategy. This strategy applies a large-system transformation approach, defined as a “coordinated, system-wide change affecting multiple organizations and care providers, with the goals of significant improvements in the efficiency of health care delivery, quality of patient care and population-level patient outcomes,”13 to the LSS program deployment within the facility. This approach may have contributed to an organizational culture that is more conducive to supporting the JRP initiative (Heather Woodward-Hagg, MS, Jamie Workman-Germann, MS, Deanna Suskovich, Edward Miech, EdD, George Ponte, RRT, Brian Preston, MBA, Isa Bar-On, PhD, Sharon Johnson, PhD, and Diane Strong, PhD, unpublished data, April 24, 2013).
We would like to share a few insights from the JRP to identify critical factors for success. In the planning and sustainment phases, reviewing data on LOS motivated individual staff members to pursue efficiency in their own work areas. Although changes associated with the JRP did result in significant improvements in LOS and cost savings, these improvements were not apparent in the midst of the project. “Soft data” such as teamwork, trust, and staff satisfaction were more reliable indicators of improved efficiency in the project period. Overall, the VA-TAMMCS method of process improvement is systems focused as opposed to focusing on individual inefficiencies, improving buy-in and sustainability. Finally, pulling individuals away from daily responsibilities is a barrier to process improvement in any institution. Planning for protected time for process improvement activities would potentially alleviate this stress and improve ROI by creating a space for team members to focus more fully on process improvement work and would foster better alignment of work with hospital staffing resources by accounting for scheduled process improvement events.
Limitations of this study are related to design. The outcomes were captured through a retrospective review of patient medical records and administrative databases. Adding an orthopedic surgeon at the transition between the project period and the sustainment period confounds the non-VA care cost data, even though the surgeon had been working in the VA under a scarce medical care contract prior to recruitment. In addition, the scarce medical care contract group was not exclusively dedicated to total joint replacement. Finally, health status was assumed to be comparable in the preproject and postproject study populations.
The JRP demonstrates that LSS process methods can be used to improve patient care efficiency, substantially enhancing outcomes while reducing costs in a VA facility’s orthopedic surgery department. Process changes reduced waste and were sustained, directly contributing to the complete elimination of non-VA care costs for patients undergoing THA and TKA, a 36% reduction in LOS, and a $1 million ROI while increasing total joint replacement volume.
Corresponding Author: Benjamin Gayed, MD, Department of Surgery, Indiana University School of Medicine, 545 Barnhill Dr, Emerson Hall, Room 202, Indianapolis, IN 46202 (firstname.lastname@example.org).
Accepted for Publication: June 4, 2013.
Published Online: September 11, 2013. doi:10.1001/jamasurg.2013.3598.
Author Contributions: Gayed had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Gayed, Black.
Acquisition of data: Gayed, Black.
Analysis and interpretation of data: All authors.
Drafting of the manuscript: Gayed.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Gayed, Black, Daggy.
Administrative, technical, or material support: Munshi.
Study supervision: Gayed, Munshi.
Conflict of Interest Disclosures: None reported.
Previous Presentation: This study was presented at the 2013 Annual Meeting of the Association of VA Surgeons; April 22, 2013; Milwaukee, Wisconsin.
Additional Contributions: We thank Verde Valley Orthopedic Associates, Edward Miech, EdD, for proofreading and providing detailed guidance on revising an earlier version of the manuscript, and Heather Woodward-Hagg, MS, for providing technical direction and proofreading.