At a large, academic medical center, as cardiac surgical volume increased, meaningful delays were noted in transferring post-operative patients from the operating room (OR) to the intensive care unit (ICU). We suspected that these delays might be associated with worse patient outcomes, longer lengths of stay, and a higher cost of care, and these suspicions were borne out by analysis. Staff were frustrated, but change was difficult to come by.
To make a change, a surgeon and the team of care providers came together to understand the system, root out problems, and develop solutions that leveraged the natural strengths of the staff. This principle is central to Dynamic Work Design (DWD), an emerging paradigm in management science that the medical center used to guide systemic changes to reduce delays, improve outcomes, and create sustained reductions in cost of care.
Attention to seemingly small and apparently inconsequential errors, like patient transfer delayed more than 15 minutes, can unearth greater inefficiencies in large complex health systems. These can be diagnosed and remedied for great payoff.
Changes made according to Dynamic Work Design — an emerging framework for human-centered workflow design that has been used successfully in many other industries — can improve process flow and efficacy, leading to improved outcomes for patients, at lower cost to the system.
When operating room staff prepare to move a patient to the intensive care unit after cardiac surgery, they may receive an unwanted reply from the ICU: “There is no room available. The patient needs to wait.” The pipeline is clogged as the ICU scrambles to transfer patients to the floor and the floor scrambles to discharge patients home. Unanticipated events, such as a patient going into atrial fibrillation, disrupt the best laid plans. The post-op patient is still waiting, anesthetized on the table. The next operation is supposed to start in 30 minutes, and that pre-operative patient is now waiting as well. After a few hours, the first patient finally gets a room in the ICU.
The staff is frustrated. The operating room is frustrated. Staff now have to work overtime to get through the day’s schedule. While everyone agrees that all would benefit from some change in how things are done, it is hard to change patterns that involve several departments and heavily matrixed organizations that often lack clear reporting relationships.
This scenario played out repeatedly at our hospital until a study initiated by a cardiothoracic surgeon investigated the underlying causes of the delays and yielded striking results. The authors retrospectively analyzed data on 1,136 patients undergoing cardiac surgery during the preceding year and found that these delays caused more than just staff frustration: They also posed serious risks to patients.
Patients delayed more than 15 minutes between the OR and the ICU were 34% more likely to experience 30-day mortality than patients who were not delayed, and were 10% more likely to be readmitted within 30 days. They experienced 16% more blood loss after surgery and stayed 13% longer in the ICU. (We also estimated that these compounding delays caused more than $2 million a year in additional direct costs, though the costs paled in comparison to the increased patient risk.)
These real dangers to patients motivated doctors and staff to find ways to decrease delays.
A team was assembled to discuss baseline data, delineate solutions, and guide interventions to improve process efficacy. The team included all stakeholders and motivated representatives — physicians, nurses, bed managers, administrators, and transport staff. Involvement of all parties was crucial, as investment and consensus on changes were vital for implementation and sustainability. The team used information from the authors who did the original analysis revealing the quality risks from patient transfer delays.
- The team considered baseline flow data, contributed further insights to understanding inefficiencies germane to their specific areas, and learned how these inefficiencies affected the system as a whole.
- The authors of the formal study visualized and studied relevant baseline processes regarding the direction and timing of patient flow and information flow. They followed a small, representative subset of 78 patients over 10 days and informally interviewed faculty and staff in order to make a literal map of how streams of patients and information normally flowed. Timestamp data was used to pinpoint bottlenecks.
- Varied solutions were deployed and simultaneously monitored for efficacy using predetermined endpoints. The team established a regular meeting where all parties convened to discuss and brainstorm in a judgment-free zone. Special care was taken to avoid over-burdening people, and leadership was turned over to small groups that were empowered to manage newly employed solutions. The study authors created an electronic dashboard that provided rapid feedback over the course of each day and aggregate metrics for weekly review by all.
The team devised relatively simple solutions to speed patient flow. For example, a hospital-wide effort had been in place to discharge patients from the floor at 11:00 AM. The group recommended discharging patients at 9:00 AM and instituting a second designated discharge time at 6:00 PM, which in turn created capacity to transfer ICU patients to the floor at night and reduced ICU congestion before operations the next morning.
The retrospective data analysis identified a tipping point where hospital costs (and presumably patient outcomes) were negatively affected by increases in case volume. Specifically, when OR volume exceeded 22 cases per week, cost per case increased 20% above baseline. Delays in transfer amplified these increases: When average transfer time exceeded 23 minutes with more than 22 cases per week, costs increased by 27%. This tipping point suggests that there is an optimum case volume for this service and care setting, and there are clear inefficiencies of scale to increasing volume above that level.
Checklist admission and discharge criteria were created for the OR, catheterization laboratories, ICU, and floors, allowing nurse practitioners and physician assistants to discharge patients without waiting for a physician’s approval, thereby decreasing the likelihood of patients being held unnecessarily while waiting for a physician to come by.
Other solutions addressed problems in information flow. Although a bed availability meeting was held daily at 9:30 AM, operations started at 7:30 AM, often with no information regarding bed availability for that day. This meeting was moved to 5:30 AM to get information on demand for beds into the hands of all invested parties, so that discharges and transfers could start earlier and thereby open up beds sooner. In addition, if the demand for beds far outstripped supply, procedures could be rescheduled. The meeting was opened to a teleconference so that participants who weren’t at the hospital yet could call in.
Once the staff had been trained on these new policies, and the policies were implemented, we prospectively studied 539 patients to determine the impact of these changes on delays, outcomes, and cost. With no significant change in procedural volume, 16% fewer patients were delayed more than 15 minutes, and the time required to transfer a patient from the OR to the ICU fell from a median of 19 minutes to a median of 15 minutes.
In fact, the likelihood that a patient would experience a delay was decreased by almost 50%. Length of stay in the ICU fell by 19%. There were no significant changes in 30-day mortality, 30-day readmission rates, or unplanned return to the OR or ICU; however, we speculate that if the size of this prospective cohort were larger, we would see a similar impact to that seen on study of the retrospective cohort. The average total cost per case across the heart and vascular center fell by 19%.
In addition to the striking improvements in these metrics, the system operated with far less chaos, and providers experienced less restriction in their ability to deliver patient care. There were no more heated debates about which patient gets the bed, giving everybody more time to focus on patients.
The Role of Dynamic Work Design
The solutions responsible for these improvements were designed according to principles of Dynamic Work Design. Like other process improvement methodologies, such as Lean Processes and Six Sigma, DWD generates context-specific solutions, tests them, and conducts follow-up monitoring. However, DWD’s specific goal is not only to make improvements, but also to make work more engaging and rewarding for the people doing it.
Processes designed using DWD are characterized by these four principles:
- Continuously reconcile activity and intent.
- Connect the human chain.
- Leverage structured problem-solving.
- Manage challenge optimally.
We can “reconcile activity with intent” in the workplace by making it easy for people to know why they do what they do, and to understand how their activities affect given outcomes. Our preliminary research, demonstrating how patient flow problems were leading to poorer outcomes, reconciled activity with intent in this project.
“Connecting the human chain” refers to embedding face-to-face communication in workflow. In-person interaction is highly effective for transferring subtle and complex information. We increased opportunities for face-to-face communication as we revised our procedures.
To “leverage structured problem-solving,” we educated our workers on how to assess errors and inefficiencies using the clarifying method of Structured Problem-Solving (SPS), a worksheet tool with subsections that organize identification of process issues and clarification of solutions. Experience with SPS can allow people doing work at all levels to recognize and translate problems into actionable changes.
For example, when members of the team contended that delays occurred because there were not enough ICU beds, the authors could walk them through a data-based analysis using our own historical data, showing that there were enough beds to accommodate 99% of cardiac surgical patients in a timely manner, if transport and patient flow problems were addressed effectively.
Finally, “optimal management of challenge” in the workplace occurs when level of stress is controlled to be stimulating but not overwhelming. Each leader of a subgroup was given only three tasks to focus on. If the group lagged behind, its action items were rapidly re-prioritized to address the highest-impact items.
Ultimately, our experience makes a case for the efficacy of Dynamic Work Design in health care. Designing work to fit the humans who do the work, to capitalize on our strengths and offset our weaknesses, offers the promise of a new era in medicine, one in which all participants are deeply engaged in delivering care that is both effective and efficient.