OUP user menu

The Effect of a Lean Quality Improvement Implementation Program on Surgical Pathology Specimen Accessioning and Gross Preparation Error Frequency

Maxwell L. Smith MD, Trent Wilkerson, Dana M. Grzybicki MD, Stephen S. Raab MD
DOI: http://dx.doi.org/10.1309/AJCP3YXID2UHZPHT 367-373 First published online: 1 September 2012


Few reports have documented the effectiveness of Lean quality improvement in changing anatomic pathology patient safety. We used Lean methods of education; hoshin kanri goal setting and culture change; kaizen events; observation of work activities, hand-offs, and pathways; A3-problem solving, metric development, and measurement; and frontline work redesign in the accessioning and gross examination areas of an anatomic pathology laboratory. We compared the pre- and post-Lean implementation proportion of near-miss events and changes made in specific work processes. In the implementation phase, we documented 29 individual A3-root cause analyses. The pre- and postimplementation proportions of process- and operator-dependent near-miss events were 5.5 and 1.8 (P < .002) and 0.6 and 0.6, respectively. We conclude that through culture change and implementation of specific work process changes, Lean implementation may improve pathology patient safety.

Key Words
  • Quality improvement
  • Near-miss
  • Lean
  • Gross examination room
  • Active errors
  • Latent errors

The safety pyramid was developed in 1931 by Heinrich1 to highlight the relationship between frequency of error and degree of patient harm. In this model, a large number of errors occur at the broad base of the pyramid that do not result in harm. As the pyramid narrows toward the peak, the number of errors decreases but the degree of harm increases. In complex systems, multiple process errors may share root causes. Therefore, the study and elimination of frequent errors that may not cause direct harm have the potential to indirectly decrease both overall error rates and harm. The Institute of Medicine report in 1991 estimated the financial and humanitarian cost of preventable medical errors at between 17 and 29 billion dollars and between 44,000 and 98,000 hospital deaths.2

As a first step in the process of error reduction, errors must be identified and classified. The most common classification systems are often based on clinical impact. Sentinel events are errors that reach the patient and are associated with significant patient harm. No-harm events are errors that reach the patient but do not result in patient harm. Near-miss events are incidents, actions, and processes that are expected to result in significant patient harm if they are not corrected before they reach the patient. An example of a near-miss event is a specimen identification mismatch in the technical processing area of a pathology laboratory that results in a malignant diagnosis being rendered on the wrong patient specimen and entered into the laboratory information system. If, before releasing the report, the pathologist recognizes the error and corrects the report in the system, the original specimen identification mismatch is a near-miss event. Near-miss events have been classified as process-dependent (events occurring as part of the normal workflow) and operator-dependent (events associated with staff activities outside the standard workflow).3 Near-miss events make up the bulk of Heinrich’s safety pyramid base.

After the identification of errors, systematic methods for analyzing and decreasing errors are most effective for error reduction. Use of Lean-based quality improvement programs (LQIP) previously has been shown to improve laboratory productivity and patient safety.47 Through a direct observation method, we previously reported a high level of baseline active and latent factors in the surgical pathology system that contributed to near-miss events.3 Examples of these near-miss events include incorrect labeling of patient containers with surgical pathology accession number, incorrect pairing of one patient’s requisition with another patient’s specimen, and stacking cassettes for one patient’s specimen on to the specimen of another patient. We hypothesized that the implementation of an LQIP in the laboratory would reduce the number of near-miss events in the accessioning and biopsy gross examining areas of the laboratory. The purpose of this study was to test our hypothesis by measuring the proportion of near-miss events in these specific areas of our anatomic pathology laboratory before and after implementing an LQIP.

Materials and Methods

LQIP Setting

We evaluated the effect of an LQIP on an anatomic pathology laboratory in an urban tertiary care medical center that accessions approximately 20,000 surgical pathology specimens annually. For this study, we evaluated the post-implementation near-miss error proportion in the specimen accessioning and tissue gross examination preparation testing phases in the surgical pathology gross tissue examination room. Three full-time equivalent staff individuals and a supervisor worked in this area.

Lean may be implemented in a number of ways, ranging from a small focused intervention using specific Lean tools to a large-scale implementation involving initiation of culture change and widespread changes in work processes led by the frontline personnel. We chose the latter method of Lean implementation in the entire anatomic pathology laboratory, including the areas of specimen accessioning and gross tissue examination described here. The timeline of the implementation of this LQIP and the major events of the process are shown in Figure 1.

LQIP Components

Lean Education

The laboratory leaders and frontline staff who eventually organized and directed the LQIP activities first attended Lean courses that ranged from several hours to 4 days in duration (eg, Perfecting Patient Care course offered by the Pittsburgh Regional Healthcare Initiative, Pittsburgh, PA).

Figure 1

Quality improvement program timeline.

Hoshin Kanri

The laboratory leadership and frontline personnel facilitated a 2-day hoshin kanri goal-setting and cultural change event to set overarching aims for the LQIP. One of the 3 anatomic pathology goals was the overall improvement of clinical service quality, the goal from which the accessioning and tissue gross examination kaizens (see below) emerged.


The core activities of the LQIP involved the implementation of kaizens, or rapid improvement events (RIEs). In the Lean model of work, less than optimal quality is attributed to failures in work activities, connections, and/or pathways. During the RIEs, work activities and their associated connections and pathways were directly observed by 1 or more nonparticipant observers, who identified near-miss errors in real time. The Lean A3 method was used during the kaizens to document error/current condition, root cause analysis, ideal state, and action plan. In addition to the Lean method of third-party observation, process mapping and timing evaluation were used to document the current condition. One or more participant observers were also recruited to report near-miss events, because the frontline staff usually has a greater familiarity and recognition of these events. Also, a blame-free atmosphere was promoted to encourage the staff to report near-miss events that may have been overlooked by the observation process.

Quality Improvement Changes

Methods to lower the frequency of process-dependent failure generally involve redesign of work pathways and the physical components of work connections or hand-offs. Methods to lower the frequency of operator failure generally involve targeting breakdown points in work activities; a combination of workflow redesign and education or training may lower the frequency of failures in skill and knowledge-based tasks.

After A3 action plans were devised during the kaizen, changes that could be made immediately were implemented. Those that could not be immediately implemented were initiated, with follow-up and completion over the ensuing several weeks.

Post-LQIP Near-Miss Measurement

Two of the authors (M.L.S. and T.W.) used a nonparticipant direct observation method to measure the frequency of near-miss errors in the accessioning area and gross examination room 15 months after the hoshin kanri. The near-miss events were classified as operator- or process-dependent. In addition, the A3 documents generated during the kaizens were reviewed to determine the root causes of error that were targeted by the subsequent quality improvement changes in the accessioning and gross examination room areas. Each A3 was classified as targeting a specific component(s) of work (ie, activity, pathway, and/or connection) to determine the relationship between error type (ie, operator or process) and improvement after implementation.

Data Analysis

The pre-LQIP implementation error proportion was previously reported.3 The pre- and post-LQIP implementation near-miss event proportions were calculated as follows: (number of near-miss events/hours of observation)/(total number of specimens accessioned/hours of observation). Thus, the near-miss event proportions represented number of near-miss events per specimen. Pre- and post-LQIP near-miss proportions were compared using the differences-between-proportions test. Statistical significance was assumed at a P value of less than or equal to .05.


The LQIP produced 164 individual A3 documents, with 29 devoted to the accessioning and gross examining area of the laboratory. Examples of some of the individual A3 documents and the components of work affected are shown in Table 1.

The A3 root cause analyses showed that specimen flow and batching were major contributors to process-dependent near-miss events and were a primary target for improvement. One A3 problem solution involved major alterations to specimen flow to move from batching to a single-piece-flow model that incorporated aspects of a pull system and leveled out the work load.

Figure 2 and Figure 3 show a process map of specimen flow pre- and post-LQIP, respectively. A schematic representation of the pull system we developed is shown in Figure 4.

Table 2 shows the number of pre- and post-LQIP process-dependent, operator-dependent, and total near-miss events per specimen. All specimens during the post-LQIP observation and measurement had at least 1 near-miss event associated with it. Statistically significant differences in pre- and post-LQIP proportions were found for total near-miss events (P < .002) and process-dependent near-miss events (P < .001), but not for operator-dependent near-miss events.


Our LQIP was successful in dramatically decreasing the frequency of process-dependent near-miss events in the accessioning and gross examining area of the laboratory, though the frequency of operator-dependent near-miss events did not significantly improve.

View this table:
Table 1
Figure 2

Current condition of accessioning, specimen set-up, and gross examination. Plan view of the work area. Batches of specimens are dropped off through a window or door (1). Specimens and requisition forms are separated from each other and stacked on the counter (2). Requisitions are taken into a clean area and accessioned into the laboratory information system (3). Requisitions are then re-paired with the specimens on the counter (4). Cassettes are made for each case on a separate counter (5). Cassettes are placed on top of the paperwork for each case (6). Tray of case ready for gross examination is delivered to biopsy (7) or big specimen (8) gross examination area. The biopsy specimen is examined, data are dictated, and the specimen is submitted into a cassette (9).

Specimen flow and pathway redesign from a batch, push-based system to a one-by-one, pull-based system with leveled workload was a major contributor to the decrease in near-miss events. There are many benefits of the single-piece-flow model, including built-in quality, flexibility, increased floor space, and increased safety.8 From a laboratory perspective, one of the most important is the single-piece-flow model’s systems engineering process that allowed staff to work on only a single patient specimen at a time. If only a single patient specimen is handled at a time there is no chance for a patient mix-up, an event that has the potential for catastrophic results. Pull systems and work-leveling are required for the single-piece-flow model to function properly. A pull-system is one in which the downstream customer signals an upstream supplier when they are ready for more product. Without this component, product piles up in batches. Heijunka, or leveling out the workload, attempts to make each set of activities take the same amount of time to complete. Without heijunka, some workers are overworked and burn out while others may feel underutilized and demoralized.

Figure 3

Specimen flow following workflow redesign. Batches of specimens are dropped off through a window or door (1). Specimens are placed on a counter in a queue for entering the work line (2). Each specimen is taken one at a time and is accessioned (3) and cassettes are made (3). Specimen is put in queue for gross examination (4). Biopsy specimen is examined grossly (5). A separate queue cart is positioned for big specimens that are placed on a designated shelf (6).

Figure 4

Single-piece-flow model with a pull system. Staff 1 takes the specimen (A) closest and performs gross examination and accessioning activities. The box holding specimen A is then replaced at the front of the inventory line (dashed box) and specimens B and C move down the line. This empty box is the queue for staff 2 to perform the task of accessioning and cassette generation for the next case. There is inventory of work for both staff 2 (specimens being dropped off in the laboratory) and staff 1 (Case B and C in the queue).

View this table:
Table 2

Although we found a significant decrease in the frequency of process-dependent errors, we were unable to reduce the frequency to zero, which was our goal. Despite changes in workflow, laboratory personnel still worked in batches on some aspects of the process. One of the difficulties was our inability to affect the work processes of suppliers (eg, involving the assembly and transport of specimens to the laboratory) outside the laboratory domain. The laboratory still received large batches of work periodically throughout the day. Setting a goal of zero defects was a point of discussion in our laboratory. Although we recognize reduction of defects to zero in a process involving humans is unrealistic, it is also unacceptable as a group to tolerate any defects when patient safety is at stake. A principle of Lean is the idea of continuous quality improvement. One approach is to have a goal of minimal defects with incremental goals of 50% reduction in defects annually.

The LQIP had no effect on the near-miss frequency of operator-dependent events despite multiple Lean educational experiences of and involvement by frontline staff. Several factors likely contributed to this: (1) high staff turnover in the area (2 of 3 personnel left their positions), (2) lack of buy-in from frontline staff, and (3) a lack of training opportunities to improve skill or knowledge-based skills.

Buy-in from frontline staff requires a unified leadership group with universal support for the change initiative. Kotter9 described 8 sequential steps in leading successful change efforts: establishing a sense of urgency, forming a powerful guiding coalition, creating a vision, communicating the vision, empowering others to act, planning for short-term success, consolidation of efforts, and institutionalization of new approaches. The lack of buy-in from some of the frontline staff suggested failures in creating a vision and empowerment. For example, many histology section personnel (customers of gross examination room personnel) and some high-level leaders in the department denigrated the LQIP and continued to blame gross examination room personnel for failures. With the absence of unified support, some frontline personnel experienced a lack of empowerment and ability to change work activities.

D’Angelo and Zarbo10 studied defects in surgical pathology and found a 27.9% frequency of defects. This is a much lower frequency than we identified and likely is related to the definition of defect and the mode of defect identification. D’Angelo and Zarbo defined a defect as a flaw, imperfection, or deficiency in specimen processing that required a delay or work stoppage to return work to the sender. This definition encompasses both operator- and process-dependent defects, but not process-related near-miss events that end up being corrected before moving down the line of work. Our operator-dependent defect rate is similar (just over 1 in 2 specimens) compared with the data reported by D’Angelo and Zarbo (just under 1 in 3 specimens). D’Angelo and Zarbo relied on the self-reporting of defects as opposed to our direct observational method which required more time and effort by a third-party observer.

Using a similar LQIP, we previously showed a marked increase in productivity and decrease in turnaround time in the histology unit of an anatomic pathology laboratory. Efficiency and timeliness are 2 other quality metrics (in addition to patient safety measured in our current study), indicating the complexity of metric development and tracking inherent in measuring change. Many pathology departments lack the resources and laboratory information systems necessary to determine the effectiveness of LQIP efforts. Laboratory personnel and leaders may use the lack of data to derail change efforts.

In conclusion, change in conditions that contribute to patient safety may be directly linked to targeting defective anatomic pathology workflow processes. The underlying factors that lead to the change or failure to change are often poorly understood and need to be studied in greater detail for dissemination of the methods of change, rather than the individual work elements that are altered.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
View Abstract