Kowinsky, Amy M. RD, LDN; Shovel, Judith MSN, RN; McLaughlin, Maribeth MPM, BSN; Vertacnik, Lisa MSN, RN; Greenhouse, Pamela K. MBA; Martin, Susan Christie MSN, RN; Minnier, Tamra E. MSN, RN, FACHE
THE high rate of interruptions in nursing work flow has been well documented in the literature1–3 and is at least partly to blame for lack of reliability of meeting patient care needs and completing patient care tasks. A 2010 study4 reported an average of 3 to 6 interruptions per nurse per hour, while another 2010 study5 found up to 10 or more interruptions per nurse per hour in inpatient settings in 2 hospitals, an academic medical center and a community-based teaching hospital. While there is a paucity of data in the literature on the rate of interruptions for assistive nursing personnel such as patient care technicians (PCTs), observations conducted at the University of Pittsburgh Medical Center (UPMC) in early 2009 showed similar rates of work interruptions for PCTs as the rates for nurses cited in the literature.
In 2009, quality improvement specialists at the UPMC held a 2-day session to discuss ways to improve the reliability and timeliness of completion of patient care tasks on the inpatient units. They had observed that frequent interruptions in work flow led to a lack of reliable completion of routine, predictable patient care tasks (eg, ambulating, turning/repositioning, feeding, routine bedside procedures), and a lack of timely attention to nonroutine, unpredictable patient care tasks, for example, answering call bells, taking patients to the bathroom, blood draws, transporting patients on and off the unit for tests or therapy, and admissions/discharges/transfers. It also had become apparent that care innovations that had been implemented over the previous few years such as hourly rounding to prevent patient falls and skin breakdown were not consistently occurring. Patient care staff was unable to effectively meet patient needs on a consistent basis.
The 2-day session began with a nominal group process to answer the question, “What are the challenges to achieving efficient work flow on the inpatient units?” Answers included the following: frequent interruptions to work flow, numerous distractions, conflicting priorities, and developing team attitude with support staff. Following the nominal group process, participants divided into teams to brainstorm alternative care models that would potentially improve work flow and the reliability of completing patient care tasks. Several alternative care models were proposed and simulated. At the end of the 2 days, a decision was made to pilot a care model that would (1) separate patient care tasks among PCTs based on predictability and (2) create role specificity to ensure reliable and timely completion of routine patient care tasks. An additional goal was to create this new care model in a cost-neutral manner, given the ongoing need to control expenses.
Results of the redesigned care model pilot would be measured at 90 days and again at 1 year by analyzing pre- and postimplementation data on timely completion of care tasks: call bell response times, blood collection time from order to draw, and assessing patients on return to the unit from an off-unit activity. In addition, patient satisfaction rates would be evaluated for improvement. If results were positive, the care model redesign would be expanded to additional inpatient units.
Factors contributing to inefficient work flow
Acute inpatient nursing requires nurses and assistive nursing personnel such as PCTs to frequently shift their attention, reevaluate decisions, and modify priorities in a dynamic environment.6 They must integrate complex thinking processes with psychomotor and affective skills7 with the goal of delivering the right care at the right time every time. During the observations of patient care tasks conducted in 2009 at UPMC, it was noted that nurses and PCTs often found themselves voicing “now, where was I?” statements after interruptions. At best, interruptions can negatively affect work flow and patient and staff satisfaction; at worst, they can be linked to serious patient safety failures such as medication errors.8,9 It is, therefore, important to identify ways to minimize interruptions to maintain the concentration and focus necessary to carrying out multiple complex activities.
On many inpatient units at UPMC, care models based on patient assignment and an unspoken rule that “everyone is responsible for answering call bells” and completing other patient care tasks have given mixed messages to staff as to who is responsible for which tasks. Each individual nurse and PCT has had some leeway in determining when patient care tasks would be completed within a “reasonable” but unspecified time frame. By not specifying whether a nurse or PCT is accountable for a particular care task and when it should be completed, tasks are sometimes accomplished late or not at all, or work-arounds and redundant layers of work are added to the care delivery process. The result is that delays occur such as long wait times for answering call bells, late blood draws, and long wait times for toileting and assistance with morning care. During times of high activity some patient requests can even be inadvertently forgotten.
Staff perception: Quality of care delivery and care task completion
A baseline staff survey on 1 medical-surgical inpatient unit at UPMC, to assess staff perception of quality of care and their ability to complete routine patient care tasks in a consistent and timely manner, showed that only 46% (16 of 35) of survey respondents (nurses and PCTs) described the quality of care on the unit as good or excellent. Furthermore, staff perceived there was room for improvement in the reliable completion of care tasks on their unit. Their answers indicated a perception that while many routine patient care tasks were “usually” completed, none were “always” completed. Walking, feeding, and Ins and Outs documentation were perceived to be completed “sometimes” and hourly rounding “rarely.” One staff member commented, “Some days we're able to [complete these tasks] if there's a low [patient] census, but when we are full or close to full the patients suffer because we don't have time to do all we need to do for each patient.”
Role redesign and simulation phase
UPMC was fortunate to have an empty 10-bed inpatient unit to serve as a simulation setting in which to create and test the foundations of a new care model. With reliability of care task completion as the focus, the care model redesign process began with separating patient care tasks based on their predictability as a method to address reliability. In a half-day session, the improvement specialist team defined predictable and unpredictable work as follows and then separated nursing care tasks by category (Table 1).
* Predictable work: work that happens repetitively and reliably, and that can be scheduled;
* Unpredictable work: nursing care tasks that are not scheduled and that occur randomly over the course of 24 hours.
With these definitions in mind, a new patient care model was created featuring a reliable rounder (RR) PCT for predictable work done at predictable times and a variable rounder (VR) PCT for unpredictable work at unpredictable times and for predictable work at unpredictable times. By separating out these 2 roles, the RR PCT would follow a specific schedule for routine patient care tasks and would be able to stay on task because interruptions and variable work were routed to the VR PCT. The VR PCT role was designed to be interruptible, with the VR handling multiple unscheduled tasks. The role redesign was based on the nature of the work rather than on patient assignment.
In this simulated environment, patient scenarios, scripts, and observers were used to role-play patient and staff experiences on a typical medical-surgical inpatient unit, first using manikins and later using people as patient actors. The scenarios and scripts were created from actual observations on busy nursing units to accurately represent real-time patient care events. Participants role played using the scenarios and scripts that specified times and patient care tasks needing to be completed over the course of several hours. The simulation enabled participants to create and fine-tune the care model to ensure separation of predictable and unpredictable work and specific role delineation such that routine patient care tasks would be reliably completed and nonroutine patient care tasks would be completed in a more efficient and timely fashion.
In a separate half-day session, front line nurses from the system-wide Professional Practice Committee and nursing leaders from the majority of UPMC hospitals were invited to the simulation unit to observe and participate in the simulation themselves as both “patients” and observers. After this simulated experience, they provided feedback on the new care model design to fine-tune it further.
The new care model features highly specified work and work distribution that is not based on patient assignment but rather on the work's predictability. Prior to the care model redesign, a PCT who was bathing a patient might need to interrupt the bathing process to quickly check a patient returning to the unit from an off-unit activity. This same PCT might be interrupted again before returning to the patient being bathed to answer a call bell (or wait to answer the call bell while finishing bathing the first patient). In the care model redesign, the RR PCT (Table 2) bathes the patient and the VR PCT receives the patient returning to the unit from an off-unit activity and answers call bells. The VR PCT carries a pocket phone and can be summoned by the unit clerk, a nurse or another PCT, or even a patient. The VR PCT's pocket phone number is provided on the white boards in patient rooms.
While simulation on an empty patient unit can provide an environment in which to try new roles and models of care without burdening staff and patients on an actual inpatient unit, only testing in a real clinical environment provides enough information to determine whether a care model redesign that worked well on the simulation unit will work in practice. Following the simulation sessions, a UPMC hospital volunteered to be the first “live” test site for the new model.
A 36-bed medical telemetry unit at Magee Women's Hospital of UPMC was selected as the first unit on which to trial the new care model that was designed and tested on the simulation unit. Factors leading to the selection of this patient unit were its track record of success with innovation as a former Transforming Care at the Bedside unit10–12 and the leadership commitment to the care redesign effort. Baseline data on the 36-bed pilot unit showed that call bell response times ranged from 1 to 14 minutes, the time from blood work order to lab draw ranged from 10 minutes to 2½ hours, and patients were seen on average 9 minutes after returning to the unit from an off-unit activity. Twenty percent of patients on the unit responded “always” to the patient-satisfaction survey question, “I received toileting help as soon as I wanted.” Observation showed 37 interruptions in one 12-hour PCT work shift.
The pilot unit was budgeted for 3 PCTs; 2 were reassigned to the RR role and 1 to the VR role, making the role redesign budget neutral. The new care model design was initiated on the pilot unit in January 2010 and began with 12 patients for 4 hours. At the end of the first test, staff provided positive feedback about the design. Over the next several days, the number of patients and the amount of time included in the trial were increased incrementally until all 36 beds were included in the role redesign over the 12-hour daylight shift; the 2 RRs and 1 VR were able to cover the entire unit.
At the end of 90 days, results for the PCT care model redesign were positive. Call bell response times decreased from a range of 1 to 14 minutes to a range of 0 to 2 minutes. Blood collection time, from order to draw, decreased from a range of 10 minutes to 2.5 hours to a range of 8 to 28 minutes. On return to the unit from an off-unit activity, patients were seen faster—the time decreased from an average of 9 minutes postreturn to consistently 2 minutes or less. Patient satisfaction, as measured by the Hospital Care Quality Information from the Consumer Perspective survey, also showed improvement, as patients responding “always” to the question about received help with toileting when wanted increased from 20% to 40%.
At 1-year postcare model redesign, the results continued to be positive. Call bell response times averaged 2 minutes, response time for other patient requests averaged 1.3 minutes, blood collection times from order to draw showed 100% compliance for turnaround times for both “stat” (30 minutes) and routine (within 2 hours) blood draws, and overall Hospital Care Quality Information from the Consumer Perspective scores for the unit increased slightly at year 1 postimplementation.
Anecdotally, there have been many comments supporting the care model redesign:
* Patient care technicians have stated that it is now easier for them to remember everything they need to do and they notice the decrease in interruptions;
* Nurses have commented that care tasks such as ambulation are now completed earlier in the day and time-sensitive tasks, such as feeding patients, are completed at appropriate times during the shift;
* Unit directors have indicated that having patient care tasks scheduled makes it easier to determine which tasks are getting done and whether there are any that need additional focus.
Unit directors are also hearing fewer complaints on patient rounds about patients needing to wait too long for call bells to be answered or to be taken to the bathroom.
CHALLENGES AND NEXT STEPS
During the course of the care delivery model redesign at UPMC, a number of challenges arose. Patient care technicians struggled with not having easy access to information about specific patient care activities that were needed by individual patients, for example, which patients needed to have vital signs taken once or twice per shift? Which patients were at full activity level, needed to be turned, or were waiting to be fed? The PCTs were spending a great deal of time (up to 30 minutes per technician per shift) manually locating this information. There was variability in how nurses supervised this activity and how they contributed to PCT shift report, resulting in inconsistency in the amount, detail, and accuracy of information provided in the shift report process. Having PCTs spend valuable time manually locating information was at odds with the goals of the care redesign model, as is any deterioration in information flow between caregivers. Nurse-PCT huddles have been implemented to supplement the information obtained from the record review and address the nurses' responsibility to oversee the PCTs. A logical next step involves leveraging the electronic health record in use at UPMC to support both easy access to patient task requirements and the accurate and up-to-date flow of information between patient care staff.
Comfort level and trust between PCT and nursing staff was an adjustment, as staff needed to be accountable for sharing information and keeping each other updated. It was difficult for staff to let go of the need to do everything for their assigned patients and simply trust that the new care model would work.
The need for leaders to develop a system to monitor that the new care model is operating as designed also has been a challenge. Currently, unit leadership is monitoring completion of patient care tasks approximately 3 times per week by direct observation and chart audit and providing feedback in staff meetings and in one-to-one communications with staff members.
During the implementation phase of the new care model, staff members reported that it was most difficult to implement on days with high patient acuity and patient census. There also was PCT staff turnover on the unit as the changes were made, which led to slow-downs in implementing the new care model while new PCT staff were hired and trained. Eventually, the care model redesign was fully incorporated on the unit and showed the positive outcomes reported. The pilot unit is budgeted for 3 PCTs; when there is a PCT vacancy, 2 PCTs are unable to support the care model as redesigned. A PCT council has been formed to address the recruitment and retention issues. The staffing challenges have made it difficult to assess patient satisfaction postimplementation, particularly as many other variables have an effect on patient satisfaction.
Although there continue to be challenges, the positive results at 90 days and at 1-year postimplementation support a decision to begin to spread the redesign to additional inpatient units. The reliable-variable rounder model is currently being spread to 2 additional units at Magee Women's Hospital of UPMC. Both the RR schedule and delineation of responsibilities are undergoing customization for the new units. Additionally, plans are underway for the model to be spread to a pilot unit at each of the 12 regional UPMC hospitals.
We set out to design a nursing care model that would improve the ability of patient care staff to reliably complete tasks in a more consistent and timely manner. The care model we have created is built on the premise that the current care delivery model does not allow caregivers to be successful in reliably meeting patient needs. The reliable-variable rounder care model separates the work of PCTs based on task predictability rather than on patient assignment and creates role specificity, the lack of which is thought to be a contributing factor to nursing work fragmentation and inconsistent care task completion.
Results of this care model redesign are promising. Call bell response times, blood collection time from order to draw, the elapsed time between patients returning to the unit from an off-unit activity until they were seen by patient care staff, and patient satisfaction indicators related to timeliness of patient care tasks have improved. The results have been positive enough that the care model redesign is expanding to additional inpatient units.
© 2012 Lippincott Williams & Wilkins, Inc.