In 2003 the Robert Wood Johnson Foundation (RWJF) launched a multiphase initiative called Transforming Care at the Bedside (TCAB) to help hospitals engage front-line staff in change processes that would improve both the work environment and the care on medical–surgical units. Led by the Institute for Healthcare Improvement (IHI), TCAB embodies a participatory, bottom-up approach to change. Such an approach, sometimes referred to as local participatory quality improvement, relies heavily on front-line staff, rather than on administrators or staff from outside the unit, to decide what changes to make and how to make them.1
To engage front-line staff in change processes, the IHI promoted a distinctive approach that combines methods from Langley and colleagues" Model for Improvement2 and the "deep-dive" approach devised by the design firm IDEO.3 The Model for Improvement encourages use of three questions—What are we trying to accomplish? How will we know that a change is an improvement? What change can we make that will result in improvement?—and a trial-and-learning approach emphasizing use of Plan–Do–Study–Act cycles to test and refine changes. (See Testing and Implementing Changes Using the Plan–Do–Study–Act Cycle in "TCAB: The 'How' and the "What"' by Rutherford and colleagues in this supplement for a description of this process.) "Deep dive" and its smaller cousin "snorkel" are brainstorming exercises designed to generate new ideas from front-line staff.
Hospitals selected for piloting the TCAB initiative encouraged the participation of front-line staff by:
* forming unit-level improvement teams that had decision-making responsibility for change
* involving staff in brainstorming sessions ("snorkels" and "deep dives") designed to identify problems and generate ideas for change
* encouraging staff to experiment with little changes by conducting many small, rapid tests of change
* enabling staff to measure the effects of the rapid tests on care processes and outcomes
* encouraging staff to use these measures to decide whether to abandon, adapt, or adopt the change
The IHI used collaborative learning sessions, site visits, conference calls, one-on-one consultation, an extranet, and an electronic mailing list to encourage the implementation of these change processes, facilitate collaboration among participating hospitals, and provide ideas for innovations to test.
Not all efforts to transform health care use such bottom-up, staff-based change processes. A study of quality improvement at 169 Department of Veterans Affairs facilities, for example, found that only four used local teams to determine the changes.4 Many facilities use a top-down approach in which administrators, researchers, or other experts determine what changes to make on the unit and how.1 In other cases top-level leaders determine the changes to make but staff members customize the intervention to the local setting.5-7
Evidence suggests that the participation of front-line staff members in quality improvement decision making increases their support for the process and the likelihood that it will be successfully implemented and staff efforts will be sustained.1, 8 An intensive, longitudinal, comparative case study of 12 health care systems attempting to fundamentally transform their systems found that the use of "improvement initiatives that actively engage staff in meaningful problem solving" was one of five critical elements for successful transformation.9
Most previous efforts at workplace transformation included some nurses in quality improvement decision making, along with physicians and other clinical staff. However, TCAB is the first major national initiative to focus directly on nurses in promoting a participatory approach to transforming the work environment and improving the quality of hospital care.
The IHI-led TCAB initiative had three phases. During the first phase (12 consecutive months in 2003 and 2004), the IHI led three hospitals, all known innovators, in collaborative activities to test prototype processes and strategies to improve nurses" working conditions and patient care. Ten additional hospitals with good track records in quality improvement joined the collaborative in the second phase of TCAB (24 consecutive months in 2004 through 2006), and these 13 hospitals worked to identify and implement additional innovative changes.10-12 At the end of this phase, 10 hospitals continued the program for an additional two years (2006 through 2008).
This paper focuses exclusively on the experiences of the 13 hospitals in the second phase of TCAB, during which they tested more than 400 innovations.13 We investigated how the hospitals implemented the change processes in order to determine if such a participatory approach can be effectively implemented. Major research questions included the following:
* Was implementation of the change processes feasible for the TCAB hospitals?
* Did all TCAB hospitals implement the change processes uniformly?
* Did front-line nurses participate in the testing and evaluation of changes?
Two companion papers in this supplement evaluate TCAB through phase 3: Needleman and colleagues in "Overall Effect of TCAB on Initial Participating Hospitals" and Parkerton and colleagues in "Lessons from Nursing Leaders on Implementing TCAB."
Hospitals. For the TCAB pilot project, the IHI chose hospitals with recognized track records in innovation and quality improvement. The 13 hospitals that participated in phase 2 were overwhelmingly urban and nonprofit and were distributed across the United States. Eight were members of large hospital systems, seven had more than 400 beds, and nine had a hospitalist on the pilot unit. Nine of the 13 hospitals designated one nursing unit to be their pilot TCAB unit. The other four hospitals selected two pilot units. Thus, phase 2 included 17 pilot units.
Data sources included interviews with hospital staff, a survey of the organization, observation of collaborative meetings, and documentation the hospitals provided as part of the TCAB project.
There were three rounds of semistructured interviews with hospital staff who were knowledgeable about the site's TCAB activities. The first interviews occurred after one year of phase 2. During a visit to each site in spring and summer 2005 we interviewed 150 people, including chief nursing officers (CNOs) and other hospital leaders, nursing unit managers, front-line nursing staff, quality improvement and performance measurement personnel, and patient representatives from the 17 TCAB units. Two of us conducted each interview using a semistructured interview guide. Interviews lasted 40 to 90 minutes. A second round of 27 telephone interviews occurred in spring 2006. For these interviews, we selected unit managers and quality improvement staff or persons responsible for data collection. A third round of 29 telephone interviews took place in summer 2006, at the end of phase 2, with the CNOs at each of the 13 hospitals and managers of 16 of the 17 units. The telephone interviewers also used semistructured interview guides, and the interviews were approximately one hour long. Our employers" institutional review boards reviewed and approved the protocols for all three rounds of interviews.
After each interview, one of the interviewers coded the data according to categories suggested by the semistructured interview guides and defined by the team as a whole. The other interviewer reviewed the coding, and then they discussed and resolved any discrepancies. If discrepancies occurred across data sources, the whole evaluation team analyzed and resolved them.
Another data source was a questionnaire on the organizational characteristics of the participating hospitals and nursing units. The RWJF distributed this questionnaire to hospital staff in June 2004, before phase 2 of the TCAB pilot began.
At least one evaluation team member was present at the seven multifacility learning and innovation community meetings that the IHI set up for TCAB team members and hospital leadership. We also observed TCAB site visits and conference calls. We reviewed and abstracted data from project documents distributed at these sessions and from information posted on an extranet that the IHI hosted for TCAB participants.
Data analysis. As is appropriate for implementation and dissemination research,14 we used a mixed qualitative and quantitative approach to analyze implementation of the TCAB intervention at the nursing unit level. We first looked at the nursing units" fidelity to the TCAB approach by examining their adherence to the IHI-advocated change processes (using unit-level teams; having staff generate ideas; experimenting with small, rapid tests of change; and using measurement data to make change decisions). Where possible, we created dichotomous or categorical variables to indicate the unit's adherence to key aspects of the TCAB process. Adherence to the IHI recommendation to use unit-level teams for decision making, for example, was defined in terms of team formation and frequency of team meetings. We created three dichotomous variables of team use: unit team formed in year 1, weekly team meetings in year 1, and weekly team meetings in year 2.
We also assessed the overall intensity of each hospital's implementation of TCAB. Consistent with the TCAB bottom-up model, the intensity measure included indicators of unit-level team participation, staff brainstorming, staff performance of tests of change, and use of measurement data for assessing tests of change. Because sustained performance is an important dimension of intensity, the analysis looked at indicators of performance in the second year of TCAB (except for data use, which was not collected separately by year). The TCAB intensity measure combined four dichotomous variables: weekly team meetings held in year 2, at least one unit brainstorming session in year 2, unit team or staff serving as primary decision makers in year 2 for what change ideas to test, and team use of measurement data to refine tests of change in year 1 or 2.
This analysis uses descriptive statistics to examine pilot units' adherence to the TCAB approach and the intensity of their implementation. Given the small sample size, the findings are exploratory in nature and can only suggest hypotheses for use in subsequent analyses of larger-scale TCAB interventions.
Table 1 shows the percentage of nursing units that implemented change processes the IHI promoted through the TCAB collaborative. Many units implemented the basic organizational processes required for TCAB. For example, 88% initially created a unit-level team, including staff nurses, to direct TCAB activities, and 88% held an initial brainstorming session (a "snorkel" or "deep dive") to solicit ideas for change from front-line staff. All units conducted rapid-cycle tests of change in the first year of phase 2, and 88% did so in the second year. Approximately 70% of pilot units successfully encouraged the use of data in TCAB decision making, in particular to refine their tests of change.
The support of senior leadership and increased autonomy of nursing units made this high degree of change implementation and decision making by front-line nurses possible. On 62% of units the unit manager played an active role in forming the TCAB team and leading the TCAB effort. In 59% of cases the nursing unit itself, rather than the hospital's quality improvement department, led TCAB activities in the first year. In 71% of the units staff were given dedicated time to work on TCAB.
Implementation of some TCAB processes continued at high levels into the second year, but tapered off for others. By the second year on 63% of the units, the TCAB teams or front-line staff, rather than the unit managers or more senior leadership, were deciding which of the many proposed changes to test. Only 47% of the units continued to conduct TCAB team meetings on a weekly basis in the second year, however, and only 18% held an additional brainstorming session where staff could suggest problems and ideas for change. Staff members in 44% of the units consistently were allowed dedicated time to work on TCAB in the second year.
The intensity of the units" implementation of TCAB change processes also varied considerably (see Figure 1). At one extreme, three units carried out none of the four processes included in the intensity measure; that is, in the second year the unit TCAB team did not hold a brainstorming session, meet weekly, decide which tests of change to initiate, or use data to refine tests of change. At the other extreme, three units performed all four of these processes. Five units implemented one of these processes, two units engaged in two, and the remaining four units accomplished three.
Clear patterns emerged in the nursing units" implementation of these processes. Most units that implemented only one of the processes used data to refine tests of change. In all nine units that implemented two or more processes, unit staff made decisions on tests of change. Most of the units that held frequent meetings also used data to refine tests and had staff or team decision-making structures. The three units that continued to hold staff brainstorming sessions in the second year ranked the highest in intensity, as they implemented the other three processes as well.
Units succeeded in engaging a substantial portion of their nursing staffs in process improvement. By year two, half of the units had at least 80% of the front-line nurses participating in tests of change. Units with a high degree of TCAB implementation were more likely to achieve this level of participation. Six of the seven units that performed three or four of the processes included in the intensity measure reported 80% or higher participation, whereas seven of the nine units with scores of less than 3 failed to reach this level of participation.
The 17 nursing units participating in phase 2 of TCAB implemented many of the change processes at relatively high levels. Almost all created unit-level improvement teams, solicited staff ideas for change in an initial brainstorming session, and conducted rapid-cycle tests of change on the unit. Most units had active teams that met weekly in the first year, made decisions about which change ideas to test, and experimented with using measurement data to assess and refine tests of change. Senior leadership at most of these hospitals supported unit-level responsibility and authority for TCAB activities and provided some help to facilitate staff participation.
This evaluation of the second phase of TCAB also suggests that these mechanisms do engage front-line nurses in change—and not just a select few nurses. Half of the hospitals reported that 80% or more of their front-line nurses were participating in TCAB tests of change by the second year, and most of these units sustained high levels of intensity by engaging in multiple TCAB processes.
Although engaging front-line nurses in change processes is feasible, this evaluation shows that implementing the front-line participatory approach is not necessarily easy. The unit variation in performance testifies to this, as does the drop-off in implementation in the second year. Lack of familiarity with a work climate that values experimentation and ongoing data assessment, demands on busy units, other changes in the work environment, and a shift in attention to newer projects are possible reasons for the decreased performance of TCAB processes in some units over time.
The intensity analysis demonstrates that performing some of the TCAB change processes doesn't necessarily mean others will be performed. Only 18% of the units engaged in all four processes in the intensity measure. The processes least likely to be continued were brainstorming and frequent team meetings in the second year. The process most frequently continued was using measurement data to refine tests of change.
Use of measurement data in assessing change is generally considered critical in most quality improvement methods,15 and it is impressive that 71% of the units reported doing so. However, the criterion for this variable was that the team had used at least one measure to inform a test of change, which gives no indication of how extensive data use was beyond a single test of change.
About two-thirds of the units reported that front-line staff—usually the TCAB team—continued to decide which changes to test in the second year. However, the extent to which nurses continued to generate ideas for change is not clear, given that only 18% of the units continued staff brainstorming sessions in the second year. It may be that units tapped the pool of untested ideas from the first deep dive. Perhaps once change activities began, staff members continued to suggest new ideas without formal brainstorming sessions or weekly meetings. Less frequent team meetings may have provided adequate opportunities for nurses to continue participating in the change process. A question for further research is whether the decreases in brainstorming sessions and the frequency of team meetings reflect efficient adaptation of the TCAB approach or missed opportunities for further engaging nurses in change processes.
In interviews staff nurses expressed enthusiasm about becoming agents of change. Several nurses said they found the TCAB team activities exciting. As one nurse elaborated, "I see what we've done, the changes that we've actually implemented on the floor, and how that helps us." A front-line staff member at another hospital described the TCAB experience as offering a "shared feeling of inclusiveness. It is more empowering. Everyone feels they can make a difference; they can make change. TCAB offers us the chance to come up with ideas rather than follow suit with the way things have always been done. TCAB has given us access to change."
These hospitals were selected to participate in the TCAB pilot because of their previous experience with innovation and quality improvement. The results they achieved therefore might not be generalizable to a broader set of hospitals. Nonetheless, this evaluation demonstrates that it is possible for at least some hospitals to successfully implement TCAB mechanisms and engage nurses in a bottom-up approach to change, although sustaining these processes is challenging. Whether the TCAB approach can be implemented as effectively with other hospitals that have less quality improvement experience or different organizational characteristics is being addressed in the fourth phase of the TCAB initiative, in which 67 hospitals are working with the American Organization of Nurse Executives to implement the same change mechanisms.16