ORIGINAL ARTICLE
Focused Process Improvement Events: Sustainability of Impact on Process and Performance in an Academic Radiology Department Andrew B. Rosenkrantz, MD, MPA, Kirk Lawson, MBA, Rosina Ally, BBA, David Chen, BS, Frank Donno, Steven Rittberg, Joan Rodriguez, Michael P. Recht, MD Abstract Purpose: To evaluate sustainability of impact of rapid, focused process improvement (PI) events on process and performance within an academic radiology department. Methods: Our department conducted PI during 2011 and 2012 in CT, MRI, ultrasound, breast imaging, and research billing. PI entailed participation by all stakeholders, facilitation by the department chair, collection of baseline data, meetings during several weeks, definition of performance metrics, creation of an improvement plan, and prompt implementation. We explore common themes among PI events regarding initial impact and durability of changes. We also assess performance in each area pre-PI, immediately post-PI, and at the time of the current study. Results: All PI events achieved an immediate improvement in performance metrics, often entailing both examination volumes and ontime performance. IT-based solutions, process standardization, and redefinition of staff responsibilities were often central in these changes, and participants consistently expressed improved internal leadership and problem-solving ability. Major environmental changes commonly occurred after PI, including a natural disaster with equipment loss, a change in location or services offered, and new enterprise-wide electronic medical record system incorporating new billing and radiology informatics systems, requiring flexibility in the PI implementation plan. Only one PI team conducted regular post-PI follow-up meetings. Sustained improvement was frequently, but not universally, observed: in the long-term following initial PI, measures of examination volume showed continued progressive improvements, whereas measures of operational efficiency remained stable or occasionally declined. Conclusions: Focused PI is generally effective in achieving performance improvement, although a changing environment influences the sustainability of impact. Thus, continued process evaluation and ongoing workflow modifications are warranted. Key Words: Radiology, radiology practice, process improvement, performance metrics J Am Coll Radiol 2015;12:75-81. Copyright 2015 American College of Radiology
INTRODUCTION A radiology department is a highly complex environment [1]. A wide variety of individuals, including radiologists, technologists, nurses, licensed independent practitioners, schedulers, registrars, billers, research coordinators, and administrators work together to provide a comprehensive spectrum of high-quality imaging services to a large number of patients in a safe, timely, and reliable manner while taking advantage of numerous sophisticated medical and
Department of Radiology, NYU Langone Medical Center, New York, New York. Corresponding author and reprints: Andrew B. Rosenkrantz, Department of Radiology, Center for Biomedical Imaging, NYU Langone Medical Center, 660 First Avenue, New York, NY 10016; e-mail: Andrew. [email protected].
ª 2015 American College of Radiology 1546-1440/14/$36.00 n http://dx.doi.org/10.1016/j.jacr.2014.08.016
information technologies. These processes must be responsive to a complex and frequently changing regulatory environment in such areas as compliance, billing, and reimbursement. Furthermore, ongoing reform in health care payment systems, including declining reimbursements, creates strong pressures for radiology departments to increase their efficiency and optimize workflows. If achieved, such efforts will not only improve departmental operations, but also enhance employee engagement and patient care. Process improvement (PI) is well established in the business community as a formal approach to achieving operational efficiency and is now considered a critical component of organizational competitiveness and survival [2-4]. Key aspects include the following: representation of all constituents involved in the process undergoing change; collection and analysis of data as a basis for change; empowerment of frontline workers to devise
75
and implement changes; use of IT systems to facilitate change; and ongoing iterative evaluation of results, with continued implementation of further process refinement as warranted. PI has been applied effectively throughout various specialty areas in medicine and is documented to lead to improvements in diverse settings such as intensive care units [5], pharmacies [6], nutrition services departments [7], and billing departments [8]. PI is often performed using a broad and disruptive approach that potentially entails use of a large amount of organizational resources. For instance, within our institution, a robust Lean Six Sigma office conducts extensive re-engineering processes that involve multiple departments and constituents. Projects typically entail at least 1 week of daily 8-hour meetings, although they sometimes run for considerably longer, as well as financial costs to support the office’s operations. We believe that PI can also be effectively performed in a focused fashion at the individual departmental level, entailing far less resource utilization. In particular, we have independently conducted internal PI events that involve essentially no cost and a much lower time commitment for the entire team in comparison with institution-led re-engineering. With this approach, the improvement process is rapidly completed within a narrow time span in hopes of achieving prompt “breakthrough” change [9]. For instance, we have previously described a successful effort to apply focused PI to produce immediate performance improvements within our MRI division [3]. Although PI has demonstrated strong potential to generate early effective change, the continued sustainability of a PI event’s impact is not well understood. Such insight is important given the potential to be misled by an emphasis on the immediate impact of process improvement. One report observed that the long-term impact of PI is influenced by such factors as the handling of activities that do not add value and optimization of IT support [10], suggesting that the results of PI are not guaranteed to endure. This concern regarding lack of sustainability is of particular importance for rapid, focused PI processes, such as those we have conducted in our department, given that they have a briefer and narrower initial scope. In the time since our department completed its initial focused PI in MRI, we have conducted additional focused PI events using a similar structure in ultrasound, CT, breast imaging, and research billing. At least 1 year has elapsed since completion of these focused PI events, providing an opportunity to investigate their continued effectiveness. Thus, the current study was designed to evaluate the sustainability of the impact of focused PI events on process and performance within an academic radiology department.
quality-improvement initiatives; protected health information related to individual examinations was not used for purposes of this study, and thus, institutional review board approval was not required. Our department conducted PI over the course of 2011 and 2012 within the 5 previously noted areas. Although the PI process differed among areas, owing to unique issues and workflows, we believe that several common elements of the PI process contributed to its success. Each implementation of PI entailed participation from all stakeholders, including physicians, technologists, nurses, licensed independent practitioners, schedulers, registrars, billers, research coordinators, and IT representatives (total range of participants: 8-13), and was led and facilitated by the department chair with assistance from the senior administrator who had previously received formal training in PI methodology. An initial meeting determined which data were relevant and necessary for the PI process. These data were then collected over a period of 2-3 weeks; the content of a sample datacollection form is shown in Table 1. Approximately 1 week following the start of data collection, PI participants began meeting several times a week, analyzing and understanding the data, identifying potential areas for improvement, as well as suggesting and discussing potential changes and solutions to identified problems. A key element of the PI process was the empowerment of all individuals involved to criticize existing processes and workflows and suggest new ideas and solutions without fear of retribution from supervisory personnel. To achieve this context, the chair clearly articulated this goal at the beginning of each PI initiative and strongly advocated for a culture of change. The PI meetings were generally held over a
METHODS
Note: IV ¼ intravenous. Staff completed items for all patients imaged during a 1-day period, and the data were then used to identify bottlenecks, inefficiencies, and commonly encountered workflow problems.
This retrospective HIPAA-compliant study used aggregate departmental performance data collected as part of 76
Table 1. Content of sample data-collection form used by focused process improvement team in CT Date Patient name Examination scheduled Time examination scheduled Time of patient arrival in department Time of completion of registration Time of oral contrast administration Time of patient assessment by physician assistant Time of IV placement Time patient entered examination room Time examination completed Time patient left examination room Scanner on which examination scheduled Scanner on which examination completed Problems encountered
Journal of the American College of Radiology Volume 12 n Number 1 n January 2015
period of 3-4 weeks, thereby completing the formal initial PI effort within a short time. Our previous report of the application of the PI process within our MRI department described the PI process in further detail [3]. Although our PI process was loosely based on principles underlying established quality improvement systems such as Lean and the Toyota Production System, these systems were not formally applied in our initiatives [4,11,12]. Based on our approach, a broad spectrum of changes was devised, with implementation initiated immediately. The PI participants defined specific metrics to be tracked and monitored following implementation. No financial costs were incurred as a result of the focused PI events themselves, although financial costs were often associated with changes made on the basis of these events. Key participants in each PI initiative also provided qualitative feedback related to the operational challenges identified for improvement by each PI initiative and the major changes implemented. PI participants provided additional qualitative feedback regarding trends that had developed since the initial PI team meetings, including unanticipated later events, further workflow alterations, and the overall durability of the implemented changes. Common themes among these conversations were identified. In addition, key metrics tracked by each PI initiative were recorded for 3 time points: pre-PI, immediately post-PI, and the time of the current study. Since each PI was allowed to define its own metrics, including how such metrics were tracked, the methodology underlying these metrics was not uniform across PI events. The key metrics for each PI were subjectively compared among the 3 time points.
RESULTS MRI The content of the MRI PI initiative has been previously described in detail [3]. Primary objectives of the PI team included reducing delays in the scheduling and completion of examinations. Changes implemented included the following: (1) creation (in collaboration with Primordial Design, Inc., an outside vendor) of a real-time digital “grease board” that uses visual cues in presenting an enterprise-wide view of the flow of patients throughout all of the department’s MRI facilities; (2) modification of the responsibilities of technologists and other MRI staff to allow a greater number of technologists to be assigned to each scanner at a given time without requiring hiring of additional staff; (3) notification of outpatients the night before an examination via a live call rather than an automated appointment system, as had been used previously; (4) development of a more efficient process for preparing both patients and scanning rooms before examinations, including placement of intravenous lines in a dedicated room rather than within the scanning room itself; (5) collaboration with other
hospital departments to ensure arrival of each day’s first inpatient to be scanned by 7:00 AM; and (6) restructuring of the outpatient pediatric anesthesia schedule to allow for a greater number of such slots per week. Following the PI implementation, a natural disaster resulted in destruction of the department’s primary MRI facility, located within the main hospital, including 4 MRI scanners within this location. MRI services were quickly shifted to an alternate outpatient facility providing an overall capacity of 1 less scanner. Due to this geographic shift, it was no longer possible to move patients between inpatient and outpatient scanners in efforts to maintain a regular patient flow. In addition, after completion of the focused PI event, our department opened a new outpatient MRI center to which the majority of musculoskeletal MRI examinations were shifted. Then, as part of a research agreement with a major industry collaborator, time on 1 scanner became periodically reserved for purposes of research and development and was unavailable for routine clinical use. Nonetheless, additional new measures were adopted following completion of the initial focused PI, such as instituting a policy that technologists would remain in-house each night until completion of all ordered inpatient examinations. Metrics taken from the pre-PI period and that immediately post-PI have been previously published [3]; more recent data are provided in the current article. The daily inpatient backlog, defined as the mean number of inpatients not scanned each day (SD), was 4.2 4.8 pre-PI, 1.3 2.3 immediately post-PI, and at the time of current data collection, 0.2 1.1. The outpatient anesthesia backlog, defined as the mean waiting timing to schedule an outpatient pediatric anesthesia examination (SD), was 28.4 21.5 days pre-PI, 5.9 4.3 days post-PI, and at the time of current data collection, 2.5 0.4 days. The percentage of 30-minute examinations completed within their time-slot (previously reported as the percentage of all examinations completed within 30 minutes, regardless of the scheduled duration) was 56.5% pre-PI, 64.6% post-PI, and at the time of current data collection, 58.9%. Previously reported data pertaining to total daily MRI examination volumes and on-time performance are not compared across time periods as part of the current assessment, owing to the strong impact of the natural disaster, which included loss of scanner capacity, upon these particular metrics.
CT Primary objectives of the CT team included growth in volume and improved turnaround time in completing exams. Changes implemented in CT included the following: (1) creation of a digital grease board customized to CT workflow; (2) designation of separate scanners for performing contrast and noncontrast examinations, each with varied appointment durations; (3) creation of separate
Journal of the American College of Radiology Rosenkrantz, Lawson, Ally n Focused Process Improvement Events
77
registration lines for CT and walk-in radiography; (4) initiation of drinking of oral contrast promptly upon arrival within the waiting area; (5) evaluation of a triage checklist for each patient by the afternoon prior to the scheduled examination day, to identify and address factors that could create a delay, such as a missing prescription, preauthorization, blood work, or radiologist protocol; and (6) reassignment of staff responsibilities with newly defined roles for team “leads” stationed at registration and scanning areas to ensure proper triaging and movement of patients throughout the examination process. The natural disaster resulted in the loss of 1 CT scanner that was located in a separate site than the main CT facility and had reduced scanning hours compared with those at the main site. The main site absorbed this scanner’s capacity, without acquisition of a replacement scanner. On-time patient arrival, defined as the percentage of examinations in which the patient completed registration by the scheduled time, was 86.4% prePI, 94.3% post-PI, and at the time of current data collection, 95.2%. On-time starts defined as the percentage of examinations started within 30 minutes of the scheduled time, was 81.7% pre-PI, 91.5% post-PI, and 91.8% at the time of current data collection. The total monthly CT examination volume (SD), defined over a 3-month window, was 1,624 182 pre-PI, 1,728 146 post-PI, and 1,890 157 at the time of current data collection.
Ultrasound Changes implemented in ultrasound included the following: (1) creation of a separate waiting area for ultrasound patients; (2) creation of more changing rooms, including a separate changing area for pediatric patients; (3) an additional scanning room with a new ultrasound machine; (4) hiring of 1 new ultrasound technologist; and (5) instituting a more stringent policy of accommodating add-on examinations in a timely fashion. After completion of the PI initiative, the spectrum of examinations offered broadened, including the addition of more advanced vascular ultrasonography. In an effort to expand procedural volume, thyroid biopsies later began to be offered on a daily basis. On-time patient arrival, defined as the percentage of examinations in which the patient completed registration by the scheduled time, was 81.6% pre-PI, 89.1% post-PI, and 70.7% at the time of current data collection. On-time starts defined as the percentage of examinations started within 30 minutes of the scheduled time, was 83.5% pre-PI, 85.5% postPI, and 79.5% at the time of current data collection. The total monthly ultrasound examination volume (SD), defined over a 3-month window, was 1,265 156 pre-PI, 1,395 73 post-PI, and 1,514 176 at the time of current data collection. The monthly thyroid biopsy procedural volume (SD), 78
defined over a 3-month window, was 55 4 pre-PI, 63 5 post-PI, and 78 23 at the time of current data collection.
Breast Imaging Changes implemented in breast imaging included the following: (1) hiring of additional radiologists; (2) expanded hours during evenings and on weekends; (3) a greater number of slots reserved for add-on examinations each day; (4) greater flexibility for patients in scheduling follow-up or next routine examinations at the time of their appointment; (5) real-time reading of screening mammograms with additional imaging, if needed, on the same day; (6) installation of monitors outside of examination rooms to improve utilization; (7) scheduling of diagnostic mammography and ultrasound 1 hour apart for improved efficiency; (8) standardized examination start and end times; and (9) faster turnaround times, including completion of second-look breast MRI within 24-48 hours and of outside consultations within 48 hours, including internally scheduled follow-up, as needed. Following completion of the PI initiative, an additional outpatient women’s imaging center opened, serving to absorb the growing volume from the original location. This new site offered bone-density examinations, which had previously not been offered. On-time starts defined as the percentage of examinations started within 30 minutes of the scheduled time, was 45.1% pre-PI, 86.4% post-PI, and 84.7% at the time of data collection. The total monthly breast imaging examination volume (SD), defined over a 5-month window, was 2,213 133 pre-PI, 2,677 151 post-PI, and 2,975 56 at the time of data collection.
Research Billing Before the PI initiative, research imaging examinations designated to be reimbursed by grants were frequently not captured and billed, or they were billed at delays of up to 1 year. In addition, substantial duplication of work occurred in the form of repeated submission of unprocessed claims. Changes implemented included the following: (1) establishment of an entirely new, formalized, intradepartmental “revenue cycle operations” pathway for ensuring consistent and timely capturing and billing of research examinations; (2) new, defined roles for billing staff within this pathway; (3) creation of new IT tools for tracking research examinations; (4) creation of new, dedicated electronic forms for billing external departments for research examination; (5) creation of a more organized system for assigning account numbers and modifiers to research trials; and (6) collaboration with other groups, including radiology scheduling and departmental research coordinators. After completion of the PI initiative, EPIC was adopted as the medical center’s electronic medical record and was Journal of the American College of Radiology Volume 12 n Number 1 n January 2015
used for all billing functions across the enterprise. Thus, the processes developed during the PI initiative were adapted to fit within EPIC’s billing workflows. The total monthly amount of research billing, defined over a 3-month window and reported as a percentage of the pre-PI dollar amount, was 100.0% pre-PI, 99.1% post-PI, and 139.1% at the time of current data collection. The delay between the month of service in which research examinations were conducted and the month in which all corresponding revenue was received ranged from 1 to 6 months during the 3 months pre-PI; was 1 month during each of the 3 months post-PI; and was 0 months (all funds received within the same monthly billing cycle) during the 3 most recent months. The performance metrics for each PI area described in this Results section are summarized in Table 2.
Follow-up to Focused PI Events After completion of the initial, focused PI event and implementation of its resulting changes, the MRI PI team continued to meet every 2 weeks, facilitated by the department chair. Through these follow-up sessions, the team modified previously established workflows in response to the changing circumstances, most notably the natural disaster mentioned earlier. In addition, the CT team held biweekly meetings for approximately 3 additional months, after which no further meetings occurred. The ultrasound, breast imaging, and research billing teams held no follow-up meetings after the initial event.
DISCUSSION We have demonstrated the potential to achieve sustained improvement within an academic radiology department through
focused PI. This result is particularly evident in the potentially opposing measures of total examination volume and on-time starts in some areas. Common themes stressed by the PI teams included the existence of chronic problems that were considered unfixable before the PI initiative; the urgency for change created by collection and sharing of data; the buy-in achieved by involving frontline workers throughout the process; and the power attained in devising innovative solutions by creating a dialogue among a broad array of stakeholders. Although in some instances the PI plan entailed hiring of additional staff, in other instances, volumes were increased, on-time performance was improved, or both, without increasing staffing. Utilization of IT services and standardization of roles and processes were additional important experiences shared by the PI teams. An important aspect of our work is the demonstration of substantial improvements through focused efforts requiring use of a relatively small amount of departmental resources. This efficiency stands in contrast to the much greater cost and resource utilization of larger, more encompassing, and disruptive systems such as formal “business process reengineering” [13]. It is hoped that the detailed exposition of how we conducted such events may provide a model for other radiology departments that wish to implement similar, focused PI initiatives. An additional, important theme, relevant to the sustainability of the PI initiative’s impact, was the apparent inevitability of major shifts in the external environment following completion of the PI team’s work. These shifts included a natural disaster, opening of a new location, offering of new examination types, and adoption of a new electronic medical record/radiology information system and billing system. Such changes, whether anticipated or not, required flexibility in the implementation plan, including modification of previously established solutions. Thus, it
Table 2. Summary of performance metrics for each PI area PI Area MRI
CT
Ultrasound
Breast imaging Research billing
Metric Daily inpatient backlog* Outpatient anesthesia backlog (days) 30-minute examinations completed within time-slot On-time patient arrival On-time start Monthly CT volume* On-time patient arrival On-time start Monthly ultrasound volume* Monthly thyroid biopsy volume* On-time start Monthly breast imaging volume* Monthly research billing, relative to pre-PI dollar amount Delay until research revenue received (months)
Pre-PI 4.2 4.8 28.4 21.5 56.5% 86.4% 81.7% 1,624 182 81.6% 83.5% 1,265 156 55 4 45.1% 2,213 133 100.0% Range, 1-6
Post-PI 1.3 2.3 5.9 4.3 64.6% 94.3% 91.5% 1,728 146 89.1% 85.5% 1,395 73 63 5 86.4% 2,677 151 99.1% 1
Current 0.2 1.1 2.5 0.4 58.9% 95.2% 91.8% 1,890 157 70.7% 79.5% 1,514 176 78 23 84.7% 2,975 56 139.1% 0
Note: PI ¼ process improvement. Data are %, unless otherwise indicated. *Data represent number of examinations.
Journal of the American College of Radiology Rosenkrantz, Lawson, Ally n Focused Process Improvement Events
79
was important for teams and workflows to remain dynamic, responding quickly to both challenges and opportunities posed by evolving circumstances. Our findings highlight a critical potential weakness of the focused PI process: The efficiency of the initial process did not naturally lend itself to structured follow-up. Only 1 of the 5 focused PI teams conducted regular follow-up sessions, which in part were in response to a subsequent natural disaster. Thus, sustained improvement from the PI initiative was not guaranteed. On one hand, continued progressive increases in examination volumes between the “post-PI” and “current” time-points in CT, ultrasound, and breast imaging, despite an absence of any follow-up PI events in these areas, suggest that changes allowing for increased volume may be more ingrained and long-lasting in nature. On the other hand, no metric related to operational efficiency (ie, on-time arrival, on-time start, or completion of examination within its time-slot) demonstrated an improvement between the two post-PI time points. In fact, in ultrasound, on-time arrivals, and on-time starts all showed an initial improvement after the focused PI, before decreasing, in the long-term, to lower than the baseline performance. Although the reason for the deterioration of some metrics is uncertain, several possible causes include changes in the external environment (eg, a natural disaster), since the initial PI, that in turn require new solutions; changes in the balance between competing factors, such as increased examination and procedure volume versus on-time starts and, perhaps most important, decreased focus on certain metrics and processes over time. Therefore, the focused PI process ought not be viewed as a onetime event capable of achieving a long-lasting ideal state. Rather, the team’s work must be periodically revisited. The MRI team has continued to meet biweekly until the writing of this article; these meetings have proven to be immensely useful in responding to the immediate impact, as well as continual institution-wide changes, associated with the major natural disaster. Given that a large number of frontline workers in various roles are involved in providing radiology services, with natural turnover in such workers over time, the recurring PI efforts have been particularly helpful as a mechanism for ensuring a continual focus by the staff on the PI goals and metrics. As noted earlier, no such structured re-evaluation occurred for the 4 other PI teams. Given our observations, we suggest that it may be possible to conduct intermittent, relatively streamlined or simplified PI re-evaluations to examine the selected metrics and assess for sustained performance. Such follow-up assessment provides a structured opportunity for the PI team to address environmental changes or any undesirable trend in the tracked metrics. Based on the results of the current study, regularly scheduled follow-up meetings have been instituted for all PI teams. Although PI is relatively common in other industries and settings, it is less commonly applied in academic radiology 80
departments. The reason may relate to several unique challenges to successful PI within academic radiology departments. First, a significant hierarchical structure that is often present in an academic department may lead to a feeling of lack of empowerment of frontline employees. We encountered this feeling in the early stages of PI in several areas. Staff members who were not considered leaders, such as technologists, schedulers, and billers, expressed concern about disagreeing with their hierarchical supervisors, such as physicians and administrators. As mentioned earlier, we believe that a key element in the success of our PI process was that the department chair served as its champion, being intimately involved in its operation and stressing the need for change. Second, it is not always obvious to staff in academic radiology departments that they have aligned incentives. Attending radiologists, house staff, technologists, nurses, schedulers, and billers may start a focused PI event with different goals and interpret events and data from differing perspectives. An important element of our PI events was to discuss and agree on common goals as well as to make clear to all participants the benefits of each goal. Third, in many academic radiology departments, staff members report to different leaders. For example, radiologists and house staff members may report to the department chair, whereas technologists and schedulers may report to the hospital administrator or vice president in charge of ancillary services. In such arrangements, the leaders themselves often have competing incentives and goals, which in turn make it difficult to successfully implement PI. We are fortunate to have a single reporting structure that eliminated this potential challenge. A past review of the application of “lean” methodology in radiology departments suggested that this approach is best applied in a gradual, continuous, and comprehensive fashion and is less effective when applied for targeted improvement of specific problems. Although we acknowledge the importance of continuous follow-up, in this article we take a somewhat different view that targeted PI in individual areas within a radiology department plays an important role in breaking down initial barriers and accelerating early change [14]. One final important theme was the impact of the PI process on the culture and atmosphere within the respective areas, an effect not necessarily reflected by metrics alone. Participants reflected on the PI process as leading to a more empowered work environment with development of internal leaders among the staff members in each area and a greater sense of teamwork. These effects resulted in staff being able to apply greater problem-solving skills and face challenges as they arose, possibly accounting for the sustained, if not continual, improvement in some areas, despite no formal PI activities after the initial PI event. Furthermore, participants expressed a sense of being more aware of their work environment, being more productive with their time, and Journal of the American College of Radiology Volume 12 n Number 1 n January 2015
dealing with less waste. Examples of this sense of empowerment include continual small experiments directly conceived and attempted by the MRI staff to further improve workflow, as well as spontaneously developed new business initiatives proposed by staff, intended to expand services and enhance the level of patient care. From the perspective of staff development, team learning, and fostering of leadership potential, all of which help workers in being more prepared to address new problems as they arise, this cultural change may be the most important effect of the PI process. A number of limitations of this study warrant mention. First, although we examined sustained impact of PI, we were not able to assess true long-term results beyond, at most, several years. In addition, although many of the changes were intended to improve the patient experience and subsequent patient satisfaction, we do not have available longitudinal patient satisfaction data to document the success of such efforts. The exact reasons for the variable degrees of success in the various tracked metrics are not known with certainty based on our retrospective assessment; a prospective evaluation may be better equipped to identify features of PI most likely to achieve a long-lasting positive effect.
n n
n
n
n
All PI events achieved an immediate rapid improvement in at least 1, if not multiple, performance metrics. An important ancillary benefit in conducting PI is the empowerment of the staff most directly involved in the processes to more effectively problem-solve and sustain success. Major external environmental changes, whether planned or unplanned, commonly occurred after completion of a PI initiative, warranting continued process evaluation and ongoing workflow modifications. Sustained improvement in the performance metrics was frequently, but not universally, observed among the PI events. One limitation of focused PI was the general lack of structured follow-up to the initial PI event, which may have contributed to lack of progressive improvement in measures of operational efficiency. Continued process evaluation and ongoing workflow modifications may be warranted to ensure sustainability of the PI event’s initial impact.
REFERENCES CONCLUSION We have observed focused PI to be effective in achieving an immediate positive performance improvement in numerous areas within an academic radiology department. Such focused PI events have the potential for achieving sustained positive impact, although this outcome is not guaranteed. An observed limitation of focused PI was a general lack of structured follow-up to such events, which may have contributed to a lack of continued progressive improvement, if not an actual decline in performance, following the initial PI event, in measures related to operational efficiency. Whether or not anticipated, substantial changes in the external environment frequently occur, requiring continual modification of the PI implementation plan, if not a formal revisit of the workflow by the PI team in some situations; thus, focused PI events ideally will formally incorporate such follow-up efforts from the outset. These observations and suggestions may be useful for other academic radiology departments in guiding their own focused PI initiatives.
TAKE-HOME POINTS n
An academic radiology department conducted rapid, focused PI in ultrasound, CT, MRI, breast imaging, and research billing, entailing less resource utilization than traditional broader re-engineering methodologies.
1. Ondategui-Parra S, Gill IE, Bhagwat JG, et al. Clinical operations management in radiology. J Am Coll Radiol 2004;1:632-40. 2. Chang PJ. Challenges and opportunities for radiology in the next millennium: re-engineering the radiology practice in an electronic world. Radiographics 2001;21:1013-4. 3. Recht M, Macari M, Lawson K, et al. Impacting key performance indicators in an academic MR imaging department through process improvement. J Am Coll Radiol 2013;10:202-6. 4. Clark DM, Silvester K, Knowles S. Lean management systems: creating a culture of continuous quality improvement. J Clin Pathol 2013;66:638-43. 5. Lilly CM, McLaughlin JM, Zhao H, et al. A multicenter study of ICU telemedicine reengineering of adult critical care. Chest 2014;145: 500-7. 6. Chou YC, Chen BY, Tang YY, et al. Prescription-filling process reengineering of an outpatient pharmacy. J Med Syst 2012;36:893-902. 7. Miller MA, Kinsel K. Patient-focused care and its implications for nutrition practice. J Am Dietet Assoc 1998;98:177-81. 8. Kaplan C. Transforming the back office with a single keystroke. Healthc Financ Manage 2011;65:118-20. 22, 24. 9. Espinosa JA, Treiber PM, Kosnik L. A reengineering success story: process improvement in emergency department x-ray cycle time, leading to breakthrough performance in the ED ambulatory care (Fast Track) process. Ambul Outreach 1997. Winter:24-7. 10. Caccia-Bava MC, Guimaraes VC, Guimaraes T. Important factors for success in hospital BPR project phases. Int J Healthc Qual Assur 2013;26:729-45. 11. Toussaint JS, Berry LL. The promise of Lean in health care. Mayo Clinic Proc 2013;88:74-82. 12. Kissoon N. The Toyota way. or not?—new lessons for health care. Physician Exec 2010;36:40-2. 13. Hammer M, Champy J. Reengineering the corporation: a manifesto for business revolution. New York: HarperBusiness; 2006. 14. Kruskal JB, Reedy A, Pascal L, Rosen MP, Boiselle PM. Quality initiatives: lean approach to improving performance and efficiency in a radiology department. Radiographics 2012;32:573-87.
Journal of the American College of Radiology Rosenkrantz, Lawson, Ally n Focused Process Improvement Events
81
Focused Process Improvement Events: Sustainability of Impact on Process and Performance in an Academic Radiology Department Andrew B. Rosenkrantz, MD, MPA, Kirk Lawson, MBA, Rosina Ally, BBA, David Chen, BS, Frank Donno, Steven Rittberg, Joan Rodriguez, Michael P. Recht, MD Abstract Purpose: To evaluate sustainability of impact of rapid, focused process improvement (PI) events on process and performance within an academic radiology department. Methods: Our department conducted PI during 2011 and 2012 in CT, MRI, ultrasound, breast imaging, and research billing. PI entailed participation by all stakeholders, facilitation by the department chair, collection of baseline data, meetings during several weeks, definition of performance metrics, creation of an improvement plan, and prompt implementation. We explore common themes among PI events regarding initial impact and durability of changes. We also assess performance in each area pre-PI, immediately post-PI, and at the time of the current study. Results: All PI events achieved an immediate improvement in performance metrics, often entailing both examination volumes and ontime performance. IT-based solutions, process standardization, and redefinition of staff responsibilities were often central in these changes, and participants consistently expressed improved internal leadership and problem-solving ability. Major environmental changes commonly occurred after PI, including a natural disaster with equipment loss, a change in location or services offered, and new enterprise-wide electronic medical record system incorporating new billing and radiology informatics systems, requiring flexibility in the PI implementation plan. Only one PI team conducted regular post-PI follow-up meetings. Sustained improvement was frequently, but not universally, observed: in the long-term following initial PI, measures of examination volume showed continued progressive improvements, whereas measures of operational efficiency remained stable or occasionally declined. Conclusions: Focused PI is generally effective in achieving performance improvement, although a changing environment influences the sustainability of impact. Thus, continued process evaluation and ongoing workflow modifications are warranted. Key Words: Radiology, radiology practice, process improvement, performance metrics J Am Coll Radiol 2015;12:75-81. Copyright 2015 American College of Radiology
INTRODUCTION A radiology department is a highly complex environment [1]. A wide variety of individuals, including radiologists, technologists, nurses, licensed independent practitioners, schedulers, registrars, billers, research coordinators, and administrators work together to provide a comprehensive spectrum of high-quality imaging services to a large number of patients in a safe, timely, and reliable manner while taking advantage of numerous sophisticated medical and
Department of Radiology, NYU Langone Medical Center, New York, New York. Corresponding author and reprints: Andrew B. Rosenkrantz, Department of Radiology, Center for Biomedical Imaging, NYU Langone Medical Center, 660 First Avenue, New York, NY 10016; e-mail: Andrew. [email protected].
ª 2015 American College of Radiology 1546-1440/14/$36.00 n http://dx.doi.org/10.1016/j.jacr.2014.08.016
information technologies. These processes must be responsive to a complex and frequently changing regulatory environment in such areas as compliance, billing, and reimbursement. Furthermore, ongoing reform in health care payment systems, including declining reimbursements, creates strong pressures for radiology departments to increase their efficiency and optimize workflows. If achieved, such efforts will not only improve departmental operations, but also enhance employee engagement and patient care. Process improvement (PI) is well established in the business community as a formal approach to achieving operational efficiency and is now considered a critical component of organizational competitiveness and survival [2-4]. Key aspects include the following: representation of all constituents involved in the process undergoing change; collection and analysis of data as a basis for change; empowerment of frontline workers to devise
75
and implement changes; use of IT systems to facilitate change; and ongoing iterative evaluation of results, with continued implementation of further process refinement as warranted. PI has been applied effectively throughout various specialty areas in medicine and is documented to lead to improvements in diverse settings such as intensive care units [5], pharmacies [6], nutrition services departments [7], and billing departments [8]. PI is often performed using a broad and disruptive approach that potentially entails use of a large amount of organizational resources. For instance, within our institution, a robust Lean Six Sigma office conducts extensive re-engineering processes that involve multiple departments and constituents. Projects typically entail at least 1 week of daily 8-hour meetings, although they sometimes run for considerably longer, as well as financial costs to support the office’s operations. We believe that PI can also be effectively performed in a focused fashion at the individual departmental level, entailing far less resource utilization. In particular, we have independently conducted internal PI events that involve essentially no cost and a much lower time commitment for the entire team in comparison with institution-led re-engineering. With this approach, the improvement process is rapidly completed within a narrow time span in hopes of achieving prompt “breakthrough” change [9]. For instance, we have previously described a successful effort to apply focused PI to produce immediate performance improvements within our MRI division [3]. Although PI has demonstrated strong potential to generate early effective change, the continued sustainability of a PI event’s impact is not well understood. Such insight is important given the potential to be misled by an emphasis on the immediate impact of process improvement. One report observed that the long-term impact of PI is influenced by such factors as the handling of activities that do not add value and optimization of IT support [10], suggesting that the results of PI are not guaranteed to endure. This concern regarding lack of sustainability is of particular importance for rapid, focused PI processes, such as those we have conducted in our department, given that they have a briefer and narrower initial scope. In the time since our department completed its initial focused PI in MRI, we have conducted additional focused PI events using a similar structure in ultrasound, CT, breast imaging, and research billing. At least 1 year has elapsed since completion of these focused PI events, providing an opportunity to investigate their continued effectiveness. Thus, the current study was designed to evaluate the sustainability of the impact of focused PI events on process and performance within an academic radiology department.
quality-improvement initiatives; protected health information related to individual examinations was not used for purposes of this study, and thus, institutional review board approval was not required. Our department conducted PI over the course of 2011 and 2012 within the 5 previously noted areas. Although the PI process differed among areas, owing to unique issues and workflows, we believe that several common elements of the PI process contributed to its success. Each implementation of PI entailed participation from all stakeholders, including physicians, technologists, nurses, licensed independent practitioners, schedulers, registrars, billers, research coordinators, and IT representatives (total range of participants: 8-13), and was led and facilitated by the department chair with assistance from the senior administrator who had previously received formal training in PI methodology. An initial meeting determined which data were relevant and necessary for the PI process. These data were then collected over a period of 2-3 weeks; the content of a sample datacollection form is shown in Table 1. Approximately 1 week following the start of data collection, PI participants began meeting several times a week, analyzing and understanding the data, identifying potential areas for improvement, as well as suggesting and discussing potential changes and solutions to identified problems. A key element of the PI process was the empowerment of all individuals involved to criticize existing processes and workflows and suggest new ideas and solutions without fear of retribution from supervisory personnel. To achieve this context, the chair clearly articulated this goal at the beginning of each PI initiative and strongly advocated for a culture of change. The PI meetings were generally held over a
METHODS
Note: IV ¼ intravenous. Staff completed items for all patients imaged during a 1-day period, and the data were then used to identify bottlenecks, inefficiencies, and commonly encountered workflow problems.
This retrospective HIPAA-compliant study used aggregate departmental performance data collected as part of 76
Table 1. Content of sample data-collection form used by focused process improvement team in CT Date Patient name Examination scheduled Time examination scheduled Time of patient arrival in department Time of completion of registration Time of oral contrast administration Time of patient assessment by physician assistant Time of IV placement Time patient entered examination room Time examination completed Time patient left examination room Scanner on which examination scheduled Scanner on which examination completed Problems encountered
Journal of the American College of Radiology Volume 12 n Number 1 n January 2015
period of 3-4 weeks, thereby completing the formal initial PI effort within a short time. Our previous report of the application of the PI process within our MRI department described the PI process in further detail [3]. Although our PI process was loosely based on principles underlying established quality improvement systems such as Lean and the Toyota Production System, these systems were not formally applied in our initiatives [4,11,12]. Based on our approach, a broad spectrum of changes was devised, with implementation initiated immediately. The PI participants defined specific metrics to be tracked and monitored following implementation. No financial costs were incurred as a result of the focused PI events themselves, although financial costs were often associated with changes made on the basis of these events. Key participants in each PI initiative also provided qualitative feedback related to the operational challenges identified for improvement by each PI initiative and the major changes implemented. PI participants provided additional qualitative feedback regarding trends that had developed since the initial PI team meetings, including unanticipated later events, further workflow alterations, and the overall durability of the implemented changes. Common themes among these conversations were identified. In addition, key metrics tracked by each PI initiative were recorded for 3 time points: pre-PI, immediately post-PI, and the time of the current study. Since each PI was allowed to define its own metrics, including how such metrics were tracked, the methodology underlying these metrics was not uniform across PI events. The key metrics for each PI were subjectively compared among the 3 time points.
RESULTS MRI The content of the MRI PI initiative has been previously described in detail [3]. Primary objectives of the PI team included reducing delays in the scheduling and completion of examinations. Changes implemented included the following: (1) creation (in collaboration with Primordial Design, Inc., an outside vendor) of a real-time digital “grease board” that uses visual cues in presenting an enterprise-wide view of the flow of patients throughout all of the department’s MRI facilities; (2) modification of the responsibilities of technologists and other MRI staff to allow a greater number of technologists to be assigned to each scanner at a given time without requiring hiring of additional staff; (3) notification of outpatients the night before an examination via a live call rather than an automated appointment system, as had been used previously; (4) development of a more efficient process for preparing both patients and scanning rooms before examinations, including placement of intravenous lines in a dedicated room rather than within the scanning room itself; (5) collaboration with other
hospital departments to ensure arrival of each day’s first inpatient to be scanned by 7:00 AM; and (6) restructuring of the outpatient pediatric anesthesia schedule to allow for a greater number of such slots per week. Following the PI implementation, a natural disaster resulted in destruction of the department’s primary MRI facility, located within the main hospital, including 4 MRI scanners within this location. MRI services were quickly shifted to an alternate outpatient facility providing an overall capacity of 1 less scanner. Due to this geographic shift, it was no longer possible to move patients between inpatient and outpatient scanners in efforts to maintain a regular patient flow. In addition, after completion of the focused PI event, our department opened a new outpatient MRI center to which the majority of musculoskeletal MRI examinations were shifted. Then, as part of a research agreement with a major industry collaborator, time on 1 scanner became periodically reserved for purposes of research and development and was unavailable for routine clinical use. Nonetheless, additional new measures were adopted following completion of the initial focused PI, such as instituting a policy that technologists would remain in-house each night until completion of all ordered inpatient examinations. Metrics taken from the pre-PI period and that immediately post-PI have been previously published [3]; more recent data are provided in the current article. The daily inpatient backlog, defined as the mean number of inpatients not scanned each day (SD), was 4.2 4.8 pre-PI, 1.3 2.3 immediately post-PI, and at the time of current data collection, 0.2 1.1. The outpatient anesthesia backlog, defined as the mean waiting timing to schedule an outpatient pediatric anesthesia examination (SD), was 28.4 21.5 days pre-PI, 5.9 4.3 days post-PI, and at the time of current data collection, 2.5 0.4 days. The percentage of 30-minute examinations completed within their time-slot (previously reported as the percentage of all examinations completed within 30 minutes, regardless of the scheduled duration) was 56.5% pre-PI, 64.6% post-PI, and at the time of current data collection, 58.9%. Previously reported data pertaining to total daily MRI examination volumes and on-time performance are not compared across time periods as part of the current assessment, owing to the strong impact of the natural disaster, which included loss of scanner capacity, upon these particular metrics.
CT Primary objectives of the CT team included growth in volume and improved turnaround time in completing exams. Changes implemented in CT included the following: (1) creation of a digital grease board customized to CT workflow; (2) designation of separate scanners for performing contrast and noncontrast examinations, each with varied appointment durations; (3) creation of separate
Journal of the American College of Radiology Rosenkrantz, Lawson, Ally n Focused Process Improvement Events
77
registration lines for CT and walk-in radiography; (4) initiation of drinking of oral contrast promptly upon arrival within the waiting area; (5) evaluation of a triage checklist for each patient by the afternoon prior to the scheduled examination day, to identify and address factors that could create a delay, such as a missing prescription, preauthorization, blood work, or radiologist protocol; and (6) reassignment of staff responsibilities with newly defined roles for team “leads” stationed at registration and scanning areas to ensure proper triaging and movement of patients throughout the examination process. The natural disaster resulted in the loss of 1 CT scanner that was located in a separate site than the main CT facility and had reduced scanning hours compared with those at the main site. The main site absorbed this scanner’s capacity, without acquisition of a replacement scanner. On-time patient arrival, defined as the percentage of examinations in which the patient completed registration by the scheduled time, was 86.4% prePI, 94.3% post-PI, and at the time of current data collection, 95.2%. On-time starts defined as the percentage of examinations started within 30 minutes of the scheduled time, was 81.7% pre-PI, 91.5% post-PI, and 91.8% at the time of current data collection. The total monthly CT examination volume (SD), defined over a 3-month window, was 1,624 182 pre-PI, 1,728 146 post-PI, and 1,890 157 at the time of current data collection.
Ultrasound Changes implemented in ultrasound included the following: (1) creation of a separate waiting area for ultrasound patients; (2) creation of more changing rooms, including a separate changing area for pediatric patients; (3) an additional scanning room with a new ultrasound machine; (4) hiring of 1 new ultrasound technologist; and (5) instituting a more stringent policy of accommodating add-on examinations in a timely fashion. After completion of the PI initiative, the spectrum of examinations offered broadened, including the addition of more advanced vascular ultrasonography. In an effort to expand procedural volume, thyroid biopsies later began to be offered on a daily basis. On-time patient arrival, defined as the percentage of examinations in which the patient completed registration by the scheduled time, was 81.6% pre-PI, 89.1% post-PI, and 70.7% at the time of current data collection. On-time starts defined as the percentage of examinations started within 30 minutes of the scheduled time, was 83.5% pre-PI, 85.5% postPI, and 79.5% at the time of current data collection. The total monthly ultrasound examination volume (SD), defined over a 3-month window, was 1,265 156 pre-PI, 1,395 73 post-PI, and 1,514 176 at the time of current data collection. The monthly thyroid biopsy procedural volume (SD), 78
defined over a 3-month window, was 55 4 pre-PI, 63 5 post-PI, and 78 23 at the time of current data collection.
Breast Imaging Changes implemented in breast imaging included the following: (1) hiring of additional radiologists; (2) expanded hours during evenings and on weekends; (3) a greater number of slots reserved for add-on examinations each day; (4) greater flexibility for patients in scheduling follow-up or next routine examinations at the time of their appointment; (5) real-time reading of screening mammograms with additional imaging, if needed, on the same day; (6) installation of monitors outside of examination rooms to improve utilization; (7) scheduling of diagnostic mammography and ultrasound 1 hour apart for improved efficiency; (8) standardized examination start and end times; and (9) faster turnaround times, including completion of second-look breast MRI within 24-48 hours and of outside consultations within 48 hours, including internally scheduled follow-up, as needed. Following completion of the PI initiative, an additional outpatient women’s imaging center opened, serving to absorb the growing volume from the original location. This new site offered bone-density examinations, which had previously not been offered. On-time starts defined as the percentage of examinations started within 30 minutes of the scheduled time, was 45.1% pre-PI, 86.4% post-PI, and 84.7% at the time of data collection. The total monthly breast imaging examination volume (SD), defined over a 5-month window, was 2,213 133 pre-PI, 2,677 151 post-PI, and 2,975 56 at the time of data collection.
Research Billing Before the PI initiative, research imaging examinations designated to be reimbursed by grants were frequently not captured and billed, or they were billed at delays of up to 1 year. In addition, substantial duplication of work occurred in the form of repeated submission of unprocessed claims. Changes implemented included the following: (1) establishment of an entirely new, formalized, intradepartmental “revenue cycle operations” pathway for ensuring consistent and timely capturing and billing of research examinations; (2) new, defined roles for billing staff within this pathway; (3) creation of new IT tools for tracking research examinations; (4) creation of new, dedicated electronic forms for billing external departments for research examination; (5) creation of a more organized system for assigning account numbers and modifiers to research trials; and (6) collaboration with other groups, including radiology scheduling and departmental research coordinators. After completion of the PI initiative, EPIC was adopted as the medical center’s electronic medical record and was Journal of the American College of Radiology Volume 12 n Number 1 n January 2015
used for all billing functions across the enterprise. Thus, the processes developed during the PI initiative were adapted to fit within EPIC’s billing workflows. The total monthly amount of research billing, defined over a 3-month window and reported as a percentage of the pre-PI dollar amount, was 100.0% pre-PI, 99.1% post-PI, and 139.1% at the time of current data collection. The delay between the month of service in which research examinations were conducted and the month in which all corresponding revenue was received ranged from 1 to 6 months during the 3 months pre-PI; was 1 month during each of the 3 months post-PI; and was 0 months (all funds received within the same monthly billing cycle) during the 3 most recent months. The performance metrics for each PI area described in this Results section are summarized in Table 2.
Follow-up to Focused PI Events After completion of the initial, focused PI event and implementation of its resulting changes, the MRI PI team continued to meet every 2 weeks, facilitated by the department chair. Through these follow-up sessions, the team modified previously established workflows in response to the changing circumstances, most notably the natural disaster mentioned earlier. In addition, the CT team held biweekly meetings for approximately 3 additional months, after which no further meetings occurred. The ultrasound, breast imaging, and research billing teams held no follow-up meetings after the initial event.
DISCUSSION We have demonstrated the potential to achieve sustained improvement within an academic radiology department through
focused PI. This result is particularly evident in the potentially opposing measures of total examination volume and on-time starts in some areas. Common themes stressed by the PI teams included the existence of chronic problems that were considered unfixable before the PI initiative; the urgency for change created by collection and sharing of data; the buy-in achieved by involving frontline workers throughout the process; and the power attained in devising innovative solutions by creating a dialogue among a broad array of stakeholders. Although in some instances the PI plan entailed hiring of additional staff, in other instances, volumes were increased, on-time performance was improved, or both, without increasing staffing. Utilization of IT services and standardization of roles and processes were additional important experiences shared by the PI teams. An important aspect of our work is the demonstration of substantial improvements through focused efforts requiring use of a relatively small amount of departmental resources. This efficiency stands in contrast to the much greater cost and resource utilization of larger, more encompassing, and disruptive systems such as formal “business process reengineering” [13]. It is hoped that the detailed exposition of how we conducted such events may provide a model for other radiology departments that wish to implement similar, focused PI initiatives. An additional, important theme, relevant to the sustainability of the PI initiative’s impact, was the apparent inevitability of major shifts in the external environment following completion of the PI team’s work. These shifts included a natural disaster, opening of a new location, offering of new examination types, and adoption of a new electronic medical record/radiology information system and billing system. Such changes, whether anticipated or not, required flexibility in the implementation plan, including modification of previously established solutions. Thus, it
Table 2. Summary of performance metrics for each PI area PI Area MRI
CT
Ultrasound
Breast imaging Research billing
Metric Daily inpatient backlog* Outpatient anesthesia backlog (days) 30-minute examinations completed within time-slot On-time patient arrival On-time start Monthly CT volume* On-time patient arrival On-time start Monthly ultrasound volume* Monthly thyroid biopsy volume* On-time start Monthly breast imaging volume* Monthly research billing, relative to pre-PI dollar amount Delay until research revenue received (months)
Pre-PI 4.2 4.8 28.4 21.5 56.5% 86.4% 81.7% 1,624 182 81.6% 83.5% 1,265 156 55 4 45.1% 2,213 133 100.0% Range, 1-6
Post-PI 1.3 2.3 5.9 4.3 64.6% 94.3% 91.5% 1,728 146 89.1% 85.5% 1,395 73 63 5 86.4% 2,677 151 99.1% 1
Current 0.2 1.1 2.5 0.4 58.9% 95.2% 91.8% 1,890 157 70.7% 79.5% 1,514 176 78 23 84.7% 2,975 56 139.1% 0
Note: PI ¼ process improvement. Data are %, unless otherwise indicated. *Data represent number of examinations.
Journal of the American College of Radiology Rosenkrantz, Lawson, Ally n Focused Process Improvement Events
79
was important for teams and workflows to remain dynamic, responding quickly to both challenges and opportunities posed by evolving circumstances. Our findings highlight a critical potential weakness of the focused PI process: The efficiency of the initial process did not naturally lend itself to structured follow-up. Only 1 of the 5 focused PI teams conducted regular follow-up sessions, which in part were in response to a subsequent natural disaster. Thus, sustained improvement from the PI initiative was not guaranteed. On one hand, continued progressive increases in examination volumes between the “post-PI” and “current” time-points in CT, ultrasound, and breast imaging, despite an absence of any follow-up PI events in these areas, suggest that changes allowing for increased volume may be more ingrained and long-lasting in nature. On the other hand, no metric related to operational efficiency (ie, on-time arrival, on-time start, or completion of examination within its time-slot) demonstrated an improvement between the two post-PI time points. In fact, in ultrasound, on-time arrivals, and on-time starts all showed an initial improvement after the focused PI, before decreasing, in the long-term, to lower than the baseline performance. Although the reason for the deterioration of some metrics is uncertain, several possible causes include changes in the external environment (eg, a natural disaster), since the initial PI, that in turn require new solutions; changes in the balance between competing factors, such as increased examination and procedure volume versus on-time starts and, perhaps most important, decreased focus on certain metrics and processes over time. Therefore, the focused PI process ought not be viewed as a onetime event capable of achieving a long-lasting ideal state. Rather, the team’s work must be periodically revisited. The MRI team has continued to meet biweekly until the writing of this article; these meetings have proven to be immensely useful in responding to the immediate impact, as well as continual institution-wide changes, associated with the major natural disaster. Given that a large number of frontline workers in various roles are involved in providing radiology services, with natural turnover in such workers over time, the recurring PI efforts have been particularly helpful as a mechanism for ensuring a continual focus by the staff on the PI goals and metrics. As noted earlier, no such structured re-evaluation occurred for the 4 other PI teams. Given our observations, we suggest that it may be possible to conduct intermittent, relatively streamlined or simplified PI re-evaluations to examine the selected metrics and assess for sustained performance. Such follow-up assessment provides a structured opportunity for the PI team to address environmental changes or any undesirable trend in the tracked metrics. Based on the results of the current study, regularly scheduled follow-up meetings have been instituted for all PI teams. Although PI is relatively common in other industries and settings, it is less commonly applied in academic radiology 80
departments. The reason may relate to several unique challenges to successful PI within academic radiology departments. First, a significant hierarchical structure that is often present in an academic department may lead to a feeling of lack of empowerment of frontline employees. We encountered this feeling in the early stages of PI in several areas. Staff members who were not considered leaders, such as technologists, schedulers, and billers, expressed concern about disagreeing with their hierarchical supervisors, such as physicians and administrators. As mentioned earlier, we believe that a key element in the success of our PI process was that the department chair served as its champion, being intimately involved in its operation and stressing the need for change. Second, it is not always obvious to staff in academic radiology departments that they have aligned incentives. Attending radiologists, house staff, technologists, nurses, schedulers, and billers may start a focused PI event with different goals and interpret events and data from differing perspectives. An important element of our PI events was to discuss and agree on common goals as well as to make clear to all participants the benefits of each goal. Third, in many academic radiology departments, staff members report to different leaders. For example, radiologists and house staff members may report to the department chair, whereas technologists and schedulers may report to the hospital administrator or vice president in charge of ancillary services. In such arrangements, the leaders themselves often have competing incentives and goals, which in turn make it difficult to successfully implement PI. We are fortunate to have a single reporting structure that eliminated this potential challenge. A past review of the application of “lean” methodology in radiology departments suggested that this approach is best applied in a gradual, continuous, and comprehensive fashion and is less effective when applied for targeted improvement of specific problems. Although we acknowledge the importance of continuous follow-up, in this article we take a somewhat different view that targeted PI in individual areas within a radiology department plays an important role in breaking down initial barriers and accelerating early change [14]. One final important theme was the impact of the PI process on the culture and atmosphere within the respective areas, an effect not necessarily reflected by metrics alone. Participants reflected on the PI process as leading to a more empowered work environment with development of internal leaders among the staff members in each area and a greater sense of teamwork. These effects resulted in staff being able to apply greater problem-solving skills and face challenges as they arose, possibly accounting for the sustained, if not continual, improvement in some areas, despite no formal PI activities after the initial PI event. Furthermore, participants expressed a sense of being more aware of their work environment, being more productive with their time, and Journal of the American College of Radiology Volume 12 n Number 1 n January 2015
dealing with less waste. Examples of this sense of empowerment include continual small experiments directly conceived and attempted by the MRI staff to further improve workflow, as well as spontaneously developed new business initiatives proposed by staff, intended to expand services and enhance the level of patient care. From the perspective of staff development, team learning, and fostering of leadership potential, all of which help workers in being more prepared to address new problems as they arise, this cultural change may be the most important effect of the PI process. A number of limitations of this study warrant mention. First, although we examined sustained impact of PI, we were not able to assess true long-term results beyond, at most, several years. In addition, although many of the changes were intended to improve the patient experience and subsequent patient satisfaction, we do not have available longitudinal patient satisfaction data to document the success of such efforts. The exact reasons for the variable degrees of success in the various tracked metrics are not known with certainty based on our retrospective assessment; a prospective evaluation may be better equipped to identify features of PI most likely to achieve a long-lasting positive effect.
n n
n
n
n
All PI events achieved an immediate rapid improvement in at least 1, if not multiple, performance metrics. An important ancillary benefit in conducting PI is the empowerment of the staff most directly involved in the processes to more effectively problem-solve and sustain success. Major external environmental changes, whether planned or unplanned, commonly occurred after completion of a PI initiative, warranting continued process evaluation and ongoing workflow modifications. Sustained improvement in the performance metrics was frequently, but not universally, observed among the PI events. One limitation of focused PI was the general lack of structured follow-up to the initial PI event, which may have contributed to lack of progressive improvement in measures of operational efficiency. Continued process evaluation and ongoing workflow modifications may be warranted to ensure sustainability of the PI event’s initial impact.
REFERENCES CONCLUSION We have observed focused PI to be effective in achieving an immediate positive performance improvement in numerous areas within an academic radiology department. Such focused PI events have the potential for achieving sustained positive impact, although this outcome is not guaranteed. An observed limitation of focused PI was a general lack of structured follow-up to such events, which may have contributed to a lack of continued progressive improvement, if not an actual decline in performance, following the initial PI event, in measures related to operational efficiency. Whether or not anticipated, substantial changes in the external environment frequently occur, requiring continual modification of the PI implementation plan, if not a formal revisit of the workflow by the PI team in some situations; thus, focused PI events ideally will formally incorporate such follow-up efforts from the outset. These observations and suggestions may be useful for other academic radiology departments in guiding their own focused PI initiatives.
TAKE-HOME POINTS n
An academic radiology department conducted rapid, focused PI in ultrasound, CT, MRI, breast imaging, and research billing, entailing less resource utilization than traditional broader re-engineering methodologies.
1. Ondategui-Parra S, Gill IE, Bhagwat JG, et al. Clinical operations management in radiology. J Am Coll Radiol 2004;1:632-40. 2. Chang PJ. Challenges and opportunities for radiology in the next millennium: re-engineering the radiology practice in an electronic world. Radiographics 2001;21:1013-4. 3. Recht M, Macari M, Lawson K, et al. Impacting key performance indicators in an academic MR imaging department through process improvement. J Am Coll Radiol 2013;10:202-6. 4. Clark DM, Silvester K, Knowles S. Lean management systems: creating a culture of continuous quality improvement. J Clin Pathol 2013;66:638-43. 5. Lilly CM, McLaughlin JM, Zhao H, et al. A multicenter study of ICU telemedicine reengineering of adult critical care. Chest 2014;145: 500-7. 6. Chou YC, Chen BY, Tang YY, et al. Prescription-filling process reengineering of an outpatient pharmacy. J Med Syst 2012;36:893-902. 7. Miller MA, Kinsel K. Patient-focused care and its implications for nutrition practice. J Am Dietet Assoc 1998;98:177-81. 8. Kaplan C. Transforming the back office with a single keystroke. Healthc Financ Manage 2011;65:118-20. 22, 24. 9. Espinosa JA, Treiber PM, Kosnik L. A reengineering success story: process improvement in emergency department x-ray cycle time, leading to breakthrough performance in the ED ambulatory care (Fast Track) process. Ambul Outreach 1997. Winter:24-7. 10. Caccia-Bava MC, Guimaraes VC, Guimaraes T. Important factors for success in hospital BPR project phases. Int J Healthc Qual Assur 2013;26:729-45. 11. Toussaint JS, Berry LL. The promise of Lean in health care. Mayo Clinic Proc 2013;88:74-82. 12. Kissoon N. The Toyota way. or not?—new lessons for health care. Physician Exec 2010;36:40-2. 13. Hammer M, Champy J. Reengineering the corporation: a manifesto for business revolution. New York: HarperBusiness; 2006. 14. Kruskal JB, Reedy A, Pascal L, Rosen MP, Boiselle PM. Quality initiatives: lean approach to improving performance and efficiency in a radiology department. Radiographics 2012;32:573-87.
Journal of the American College of Radiology Rosenkrantz, Lawson, Ally n Focused Process Improvement Events
81
