Beth Israel-Deaconess Medical Center (BIDMC) is a large tertiary care hospital, part of the Longwood Medical District in Boston. BIDMC is the result of two separate hospitals that merged in 1996, resulting in a campus that has two specific areas (East Campus is the former Beth Israel Hospital; West Campus isnthe former Deaconess Hospital) that are two blocks apart in a dense urban area. 

Since the merger, clinical specialties have consolidated to one campus or the other, but some support services are provided on both campuses due to the distance between them. Each campus has its own central pharmacy. 

The West Campus pharmacy is located in the 1990s era clinical tower and subsequently has been renovated to incorporate an automated distribution system. In 2006, the East Campus pharmacy renovated its clean room/IV prep area to comply with 997 standards, but did not renovate the entire pharmacy. 

TROJB was retained to design the renovation of the remainder of the East Campus pharmacy, and to incorporate automated distribution technology similar to that already operating on the West Campus. 

The primary challenges of the East Campus pharmacy renovation were: 

  1. Organizing existing, oddly shaped space in four different buildings into a coherent and efficient pharmacy;
  2. Maintaining the operation of the IV/clean room throughout construction (the central pharmacy functions were relocated to a temporary pharmacy during construction);
  3. Fitting the automated carousels into the space; and
  4. Moving the research component of pharmacy services from a remote location into the central pharmacy area for proximity to the IV/clean room. 

In short, the firm had to put more function into less space and incorporate large bulky distribution units as well—a perfect Lean process improvement project. 

Support from above

BIDMC has a robust process improvement program. The office of the president established a business transformation (BT) unit, headed by Alice Lee, that is tasked with improving the quality of care by targeting operational efficiency through the implementation of Lean process improvement. Over a period of five years, BT has run numerous Lean events, involving staff at many levels. 

At the same time, BIDMC has undertaken an ambitious training program to indoctrinate all senior level management in Lean operations. 

BIDMC embraced the East Campus pharmacy renovation as the first Lean improvement project to include major construction. As a test project for how it might approach ongoing facility improvement endeavors, BIDMC put considerable resources into creating a model Lean 2P (process/preparation) event. 

The event occurred over two discrete time periods. April 12-13, 2010, about 20 people met for an introductory seminar about Lean and Lean principles. The group included representatives from pharmacy, facility planning and design, and the BT group, as well as Bob Broach of Broach Consulting, the TROJB design team, and a representative from Berry Construction.

 Several weeks later, the same group—enhanced by representatives from BIDMC’s support services, nursing units, and Cardinal Healthcare (pharmacy product supplier— convened for a five-day Lean 2P event May 10-14, 2010. 

Pre-event

TROJB held a series of programming/documentation meetings with pharmacy staff prior to the Lean initiative. The firm prepared a benchmark program and corresponding concept plan based on current operations, compiled a list of proposed pharmacy equipment with relevant model information, and analyzed the existing pharmacy space. 

A foundation density analysis of subsurface conditions also was performed in the hope that pits could be built to accommodate vertical distribution carousels, which occupy a smaller footprint and are more efficient to operate than floor mounted models. Unfortunately, the subsurface analysis indicated a high water table and challenging grade beams, so floor-mounted distribution units were used.

Introductory event

BIDMC’s approach to Lean has included broad training for all senior management in Lean principles. The director of facilities and the director of pharmacy were both well versed in Lean. As Lean events are staged within specific departments, line staff receive preparatory training in order to apply the concepts appropriately. 

The first introductory day included an overview presentation of Lean history, concepts, and tools (waste, value, flow, pull). After receiving guidance on effective workflow observation, TROJB went to the existing pharmacy to observe and document work in progress. 

From that exercise, five overarching workflows were identified: 

  1. Front line—techs receive orders, generate labels, fill prescriptions, prepare pharmacist checks;
  2. Unit-based pharmacist on clinical floors;
  3. Central pharmacy based pharmacist;
  4. Sterile product preparation; and
  5. Repackaging. 

The firm also visited the West Pharmacy during the time it does its daily restocking of floor-based Pyxis inventory from the automated carousel to observe that workflow. 

In the afternoon, the key workflows that needed mapped were determined.

These included: 

  1. Regular floor Pyxis restocking;
  2. IV/chemo processing and distribution;
  3. Narcotics processing and distribution;
  4. Special orders;
  5. Repackaging from bulk to unit dose;
  6. Receiving/stocking; and
  7. Code tray review and restock. 

TROJB developed current state maps for each of these flows. 

The second intro day, the team delved into performance failure mode effects analysis (PFMEA), an important Lean concept that determines a relative numerical value to measure the likelihood of an error occurring during a process, the probability that error will be detected prior to delivery, and the potential harm that could result from that error. 

A Starbucks process simulation also was created to understand how the chain achieves a high degree of customization in an efficient process without batching or inventory. Finally, current state maps were completed and a PFMEA analysis of the current state work processes was done. 

Interim planning

During the three weeks between the Lean introduction and the Lean 2P event, the BIDMC pharmacy gathered data on its operations. 

The East Campus pharmacy fills approximately 3 million prescriptions annually. East Campus services include oncology and NICU, therefore, the East Campus pharmacy prepares large amounts of specialty chemotherapy and infant-proportioned drugs. 

The pharmacy staff operated in the existing space with “new eyes” from their training, and several wore pedometers to measure their steps during a typical work day. 

During the same period, the facilities staff prepared the Lean 2P event site, the top floor of a recently completed lab building that had not yet been fit out for tenant use. 

TROJB taped the outline of the existing pharmacy on the concrete floor, including columns and other fixed elements. The value stream maps were posted on walls. Four sets of half-inch scale equipment cutouts were prepared for table exercises, and full-scale, three-dimensional cardboard models of all major equipment, including the floor mounted carousels for the 2P simulations, also were made. This work took three interns two weeks to complete. 

Lean 2P event

Our five-day event began with 31 people representing the pharmacy, BT, facilities, business planning, intern builders, Cardinal Healthcare, as well as the consultant team. This mix represented the Lean mantra of having every perspective of senior management and line staff at the table on a first-name basis. 

The group was divided into cross-disciplinary groups charged with mapping the identified workflows. First, the current state was vetted, and then the future state was revised. Takt time analysis was applied to the future state maps and iterated the future state to develop balanced Takt times. 

The second day began with a review of the future state maps, which got tighter from the perspective of a night’s sleep. 

When the future state maps seemed right, the team performed a PFMEA analysis on the map. PFMEA is a challenging concept for healthcare providers, who are accustomed to operating under the precept that there are no acceptable errors and are hesitant to rank relative severity. 

At first, the pharmacy group considered any error within the pharmacy as life-threatening, but eventually the team was able to provide some differentiation among the severity of potential errors and future state maps were adjusted accordingly. 

In the afternoon, groups were shuffled into four teams. Each team was given the half-inch scale cutouts to lay out on a table top covered with Kraft paper. They could use the cutouts, create new ones they thought necessary, and draw connections on their paper.

The groups worked for 30 to 45 minutes, broke to review all the teams’ work, and then iterated again. By the end of the day, a concept of two discrete but related pharmacist/tech work areas with a continuous circulation path around them began to emerge. 

On the third day, the group returned to the cutout diagrams, but this time the teams worked on the actual floor plan, as opposed to a blank spread paper. 

The transition to a well-defined, long, and narrow space occurred with little frustration. Each team was able to find a way to land the identified work zones in the actual space. Within two iterations, patterns emerged. 

Receiving moved to the front of the unit. The research area migrated to the odd-shaped zone furthest from the entry. The primary pharmacist work space aligned with the existing pass-through to the IV/clean room. 

Next, a full-scale simulation was run. The teams placed the cardboard equipment in an arrangement they thought viable. When it was set, pharmacy staff acted their parts while a narrator called out work entering the pharmacy department, either by phone, fax, p-tube, or personal delivery.

The first simulation was five minutes of chaos. The process was revised to have the narrator describe a busy, but not excessive, work flow. The pharmacy data informed how many calls, faxes, codes, etc., occurred over the course of a day. This was edited to a reliable script that the narrator used to call out work.

The second iteration went much smoother. The group studied film of the event and identified changes in the layout to test further opportunities for efficiency. 

On the fifth day, the same procedure was followed. The group modified the layout, ran a narrated simulation, watched the film, suggested improvements, and changed the mockup accordingly. 

Seven iterations were done in all. In the end, a separate narcotics room was abandoned; narcotics came into the primary work area where they are under constant surveillance. 

After each simulation, the main work cell area got tighter; the circulation route for the water spider became clearer; the receiving zone work cells got better defined. By mid-afternoon, there was a plan that met everyone’s approval. 

The group outlined the 30-60-90-day implementation plan to determine the “to-do’s” necessary to realize the approved layout. The architectural team translated the simulation mockup into an actual plan while the BT team put together a PowerPoint presentation of the week’s effort. 

On the final day, BIDMC’s vice president of administration joined the group for the report. Pharmacy staff members took turns describing what was done on a day-by-day basis, accompanied by the PowerPoint BT had prepared. The architectural team pointed out the key attributes of the plan.

After the formal presentation, the entire group went to the simulation site and senior management watched the staff run through a simulation of the final result. 

Observations

The BIDMC East Campus pharmacy is currently in construction. What is being built is a very accurate reflection of what the pharmacy group built in simulation.  Because it was done thorough an as-built analysis, the simulation model was very close to actual conditions, and facility constraints were accommodated as part of the simulation.

 The BIDMC pharmacy is unique among pharmacies TROJB has designed. The large carousel machines are not set back out of the way; they occupy center stage and separate the small, focused work cells where pharmacists and tech work with few interruptions from the more open areas for drug repackaging, receiving, and supply replenishment. 

The firm has never arrived at this layout, nor would a client likely approve it, without executing the Lean 2P exercise. In the end, it was the client who championed smaller, more focused work areas. Of the four separate spaces that pharmacy occupied pre-Lean, the two most remote (20% of total area) have been turned back to the hospital for other uses. 

The pharmacy will work very well as modeled by the Lean 2P event. The real test of its efficacy will be how well it can adapt to change over time. Lean solutions have a tendency towards the specific, while changes in healthcare operations are in constant flux. 

Some growth capacity was built in terms of carousel and narcotics area capacity, but not in terms of staffing or unidentified processes. The design is predicated on a keen understanding of the envisioned operations and an attitude that the future will not be addressed through more space, but through ever-enhanced process. 

Paul Fallon RA, LEED AP, EDAC, is a senior associate with TRO Jung|Brannen.