Increasing Capacity for Innovation in Bureaucratic Primary Care Organizations: A Whole System Participatory Action Research Project

Abstract

PURPOSE We wanted to identify what organizational features support innovation in Primary Care Groups (PCGs).

METHODS Our study used a whole system participatory action research model. Four research teams provided complementary insights. Four case study PCGs were analyzed. Two had an intervention to help local facilitators reflect on their work. Data included 70 key informant interviews, observations of clinical governance interventions and committee meetings, analysis of written materials, surveys and telephone interviews of London Primary Care Organizations, interviews with 20 nurses, and interviews with 6 finance directors. A broad range of stakeholders reviewed data at annual conferences and formed conclusions about trustworthy principles. Sequential research phases were refocused in the light of these conclusions and in response to the changing political context.

RESULTS Five features were associated with increased organizational capacity for innovation: (1) clear structures and a vision for corporate and clinical governance; (2) multiple opportunities for people to reflect and learn at all levels of the organization, and connections between these “learning spaces”; (3) both clinicians and managers in leadership roles that encourage participation; (4) the right timing for an initiative and its adaptation to the local context; and (5) external facilitation that provides opportunities for people to make sense of their experiences. Low morale was commonly attributed to 3 features: (1) overwhelming pace of reform, (2) inadequate staff experience and supportive infrastructure, and (3) financial deficits.

CONCLUSIONS These features together may support innovation in other primary care bureaucracies. The research methodology enabled people from different backgrounds to make sense of diverse research insights.

INTRODUCTION

Clusters of general practices in the United Kingdom now relate to Primary Care Organizations (PCOs) that in England are called Primary Care Trusts (PCTs). PCOs commission services from secondary and intermediate care for populations of approximately 200,000. They have subsumed the roles of such other primary care organizations as health authorities and community trusts. Starting in 1999 an earlier phase of Primary Care Groups (PCGs) paved the way for these new organizational structures.

The PCO is intended to be a cornerstone of the National Health Service (NHS) charged with “empowering frontline staff to use their skills and knowledge to develop innovative services.”¹ Lifelong learning is considered key through enabling practitioners to “change things for the better.”² Large organizations, such as PCOs and PCGs, however, necessarily develop features of a bureaucracy to gain adequate accountability of their diverse portfolios of work. Such large-scale and mechanistic organizations^3,4 may be functional for processing a large volume of routine transactions but, conversely, can be dysfunctional in terms of rapid and radical adaptation to a changing environment.^5,6 Little is known about how to overcome this tendency or how to enhance innovation in faster moving environments within the health care sector.

The successive creation of collective organizations in UK primary care since 1990 (Multi-Funds, PCGs, and now PCOs) with a strategic management function has been accompanied by a growing recognition that theory and models from learning organizations and action research are effective at managing change in the NHS.⁷ These approaches are commonly used in discrete projects. There is a dearth of experience, however, about how complex bureaucratic organizations can embed such approaches in the long term throughout their diverse domains.⁸ This project sought to provide the evidence. It focused on clinical governance programs because these were core functions of the PCGs charged with facilitating quality improvements within local primary care. The project took place in 2000–2002, during the transition from Primary Care Group to Primary Care Trust.

METHODS

A research method that enabled a broad cross-organization research community to make sense of data generated by a set of research projects was adopted for the following reasons:

1. General practitioners were not accustomed to being part of bureaucratic organizations. Their experience of what did and did not work would be small.

2. There would be several complementary features that facilitated innovation. What worked in one place might not work in another.

3. Practitioners and managers were unfamiliar with action research and organizational learning. Some exposure to what these concepts meant in practice was needed.

Details of whole system participatory action research are set out in the Supplemental Appendix, available online at http://www.annfammed.org/ http://www.annfammed.org/cgi/content/full/3/4/312/DC1. This approach has been previously used in care pathway research.⁹ It requires regular conferences at which people from all parts of the system crystallize¹⁰ meaning from different research projects. Ongoing feedback to the wider research community enables this sense-making process.

Ethical committee approval was gained.

The set of research projects included 4 case study primary care groups and 3 research subprojects.

Case Study Primary Care Groups

Two PCGs were selected because they expressed interest at a very early stage (1 in a deprived inner-city area and the other in a more suburban area). These PCGs received an intervention of a monthly half-day course for multidisciplinary teams of facilitators to reflect on their work. This resource was not available to other PCGs. Two other case study PCGs were matched for similar demographic profiles. Data gathered from these sites included observations of PCG meetings, analysis of strategic documents, and semistructured interviews (first piloted) of 70 informants from a variety of disciplines and levels in the organizations, at the beginning of the project and 2 years later.

Case study data underwent 3 stages of analysis:

1. Transcripts of interviews were reviewed by hand to identify patterns of similarities and differences across the data sources.¹¹

2. Interim analysis of the cases was made prior to annual conferences.

3. Toward the end of the project, case profiles were written to a common format.

Internal validity was gained through feedback to the sites. External validity was gained through intersite comparison and reflections of the research teams and participants at pan-London annual stakeholder conferences. Reports from the annual conference were sent to the broad group of persons who had become interested in the project, inviting comment.

Research Subprojects

Nurse Study

Telephone interviews and a focus group with health visitors, practice nurses, and district nurses from the intervention sites (20 informants) were undertaken to explore their perceptions of organizational change and how it affected their ability to innovate. This project has been separately published.¹²

Study to Identify Financial Constraints to Innovation

In this study, documents and interviews of 6 finance directors and managers from the 4 case study PCGs were reviewed.

London-Wide Surveys and Telephone Interviews

In 2000, data from 30% of all London PCGs and PCTs were analyzed (63 organizations) about their organizational development strategies. In 2001 a purposive sample of these informants was interviewed about progress. In 2002 PCG and PCT chief executives and directors were asked to rate their agreement with 9 statements derived from the study (47% response rate from 32 organizations). Also in 2002, 6 purposively selected telephone interviews were conducted to gain deeper understandings about ideas expressed in the survey.

Different people participated each year in the conferences: 25 participants (mainly local practitioners) in 2000, 70 in 2001, and 41(mainly persons with strategic roles) in 2002. Only 2 persons other than the researchers attended all 3 events.

Conferences were facilitated using brainstorming, mapping, and small-group–large-group techniques that enable learning and consensus in large groups.¹³ Initially decisions were made at meetings between the intervention PCGs and the research teams. Subsequently, decisions about a new direction were made at the conferences. To minimize bias, the research teams were disallowed to discuss data except in preparation for the conferences. Table 1 ▶ shows how the research focus changed in response to the learning of the previous stage and to new political realities.

Table 1.

Emergence of the Study Design

Year	Intended Study Design for the Year	Events Within the Year Requiring Design Change
First year (2000)	Comparative case studies of clinical governance programs of 4 PCGs	Local facilitators never started in-practice work and withdrew from the project
	Intervention PCGs used local multidisciplinary teams to facilitate change within practices	Unexpected announcement of fast progress toward PCT status altered priorities of clinical governance groups
	Amalgamated near-clinical indicators from practice computers would contribute to comparative data	Near-clinical indicators impossible to gather from practice computers
Second year (2001)	Comparative aspects of the study abandoned At request of intervention sites, facilitators refocused away from the clinical governance programs to support reflection and action inside the clinical governance subcommittees themselves. Doing so focused on effective meeting behavior and how to carry into the PCT their most valuable work Studies of nurses and financial directors took place here	Emerging enthusiasm among stakeholders that ideas about learning organizations and participatory action research help to understand learning and change in the health care system Recognition among the stakeholders that facilitation of learning and innovation had less to do with formal structures and more to do with facilitative processes that free up conversations and reflections throughout the system, helping local people to make “top down” models relevant within their specific context
Third year (2002)	Refocus away from structures and roles within PCGs toward analysis of what things help and hinder people to do things for themselves	The start of PCTs and a realization among participants that their size and complexity requires rethinking about the meaning of leadership in primary care

PCG = Primary Care Groups; PCT = Primary Care Trust.

RESULTS

Five features were associated with increased organizational capacity for innovation, and 3 were associated with low morale.

Feature 1

Understandable corporate and clinical governance, high-quality leadership, and the intention to work with the ideas of a learning organization, including reflective practice, provide a good basis for a primary care organization to facilitate innovation.

There was a broadly accepted need for clinical and corporate governance plans. These plans, which were soon in place, included multidisciplinary boards, subcommittee structures, and core management teams. In all case study sites clinical leaders worked with individual practices to help them understand clinical governance.

The plans focused on general practitioners, however, and many nurses did not feel supported by the PCG: “I don’t feel I’m in the PCG” (Nurse Study¹²).

One PCG explicitly described a vision for lifelong learning and learning organizations: “… the primary care group as a learning organization, promoting adult lifelong learning, based on reflective practice to achieve the change identified through clinical governance activities” (Primary Care Investment plan PCG2, 1999).

Participants at the 2000 conference agreed on the importance of working with the principles of organizational learning; however, participants at the 2001 conference recognized that they lacked the skills to do so.

Feature 2

External facilitation is useful when it provides an opportunity for participants to make sense of their experience and to agree both long-term vision and short-term steps toward this vision. All 4 PCG boards valued externally facilitated ”away-days.”

The plan in the intervention PCGs to use the external facilitators to support teams of local facilitator teams was abandoned after an uncomfortable 6 months of misunderstandings. Both intervention PCGs asked the facilitators to instead help the clinical governance subcommittees think through their transition to Primary Care Trusts. PCG1 invited the facilitators to observe their 13 subcommittees to give feedback about effective meeting behavior. Eleven subcommittees agreed to participate. Feedback from the exercise was presented to a cross-PCG group of more than 30 people. This feedback was greeted with much interest and led to the PCG’s decision to pay attention to developing and supporting the chairs of the subcommittees and to creating cross-committee representation to better connect learning throughout the PCG.

Facilitated reflection on data at all 3 annual conferences resulted in some special moments of energy and motivation, when participants from different backgrounds recognized that they were sensing the same insight from different perspectives. The potential of external facilitation to empower participants is indicated in these comments:

“I think [the intervention] … has educated us all to be reflective and not to say, well, this is a loss, let’s drop it [but] to see why it’s lost and go onto something else” (Clinical Governance Lead PCG1, January 2002).

“It was an important intervention in terms of enabling us to deal with the conflicts that had arisen, and I do think any organization, probably any group, needs to build in that sort of time out” (CEO PCG1, January 2002).

The potential of external facilitation to make sense of experience was apparent:

“Facilitators provide an environment where people can reflect, make sense of and act for change” (Conference, 2001).

Feature 3

Multiple opportunities for reflection and learning are needed at all levels in the organization. Such learning spaces must connect throughout the whole system if innovative thinking in one part of the system is to be built on elsewhere.

Subcommittees in both intervention PCGs became important places for practitioners and managers to reflect together on their different insights. PCG1 devised a system of subcommittee cross-representation that helped share learning. The usual situation, however, was for potentially valuable insights to remain in the place of origin: “You have a voice within your own workplace but I don’t see that it goes much further than that” (Nurse Study).¹²

Feature 4

Both clinicians and managers are needed in leadership roles. A facilitative leadership style that encourages participation is needed from both clinicians and managers.

All sites and the nurse study recognized a need for a leadership role that appeals to different constituencies and reaches different parts of the system: “… we do have an effective manager who is helping communication between PCG and locality” (Nurse Study).¹²

In PCG3 the PCG chair and chief executive were seen as a powerful partnership. They were perceived to combine good local knowledge with an ability to keep abreast of political developments, while encouraging others to take part. Informants were not sure what good leadership entailed, but it had something to do with nurturing people, helping them to make sense of it all, and giving them permission to unlearn certain attitudes:

“What those practices needed (was) somebody to go in and really take them through it and hold their hands” (Clinical Governance Lead, PCG4, January 2002).

“…nurturing leadership and the management of change [is needed], within a learning—not blaming—culture” (Primary Care Investment plan, PCG2, 1999).

“Leaders for PCTs, especially middle-management, must facilitate sense-making—not merely managing up and managing down but helping people from the whole system to make sense of their place in the whole system” (Conference, 2002).

“... you get those who are just soaked to the skin in NHS management traditions … it’s a big, big block to the learning process” (Lay member PCG2 Board, December 2001).

Feature 5

The right timing for an initiative is an important determinant of success and established interventions for learning and change need to be adapted to the local context. In PCG3 there was a locality where there was no history of collaboration between practices. They decided to first build trust and receptiveness to change by facilitating change in prescribing practice. PCG4 used locality groups to facilitate local relationships, and later adopted a program of practice visits once understanding of clinical governance had increased. PCG1 abandoned its clinical governance program that revolved around community-oriented primary care. They came to recognize that the timing was wrong for this approach:

“I had different people and lay people working together [on COPC] … and then it turned out you can’t get people and professionals working together… they can work in their own sphere and then … at the time there was a great sense of loss and then I thought ‘Oh, this is the end of it, we’re no good’… then we reflected and we came out with something different” (Clinical Governance lead, PCG1, Jan 2002).

The need to tailor interventions that had been successful elsewhere was expressed:

“Should avoid ‘cloning’ change models although interventions may have been successful elsewhere they need to be tailored for local use” (Conference, 2001).

Both PCG1 and PCG2 successfully adopted the RCGP Quality Team Development model. Leaders of the program explained that its local appeal seemed largely to do with the ease with which they could mould it to their specific context. This flexibility was contrasted with less well received initiatives, such as guidelines and protocols.

This research also identified 3 features of the contemporary environment of change. It was not clear whether these features assisted or detracted from innovation; however, low morale was commonly attributed to these factors.

Feature 6

The pace of policy reform and change can be overwhelming and this can limit the capacity for reflective practice:

“… and I think my feeling is that that’s where the government has completely lost it really, that actually yes there should be a continual improvement within the core provision but that’s slightly different to constantly changing the way in which you provide the whole service…” (CEO PCG1, March 2000).

“It’s difficult, they’ve been so many changes” (HV in Nurse Study).¹²

Feature 7

Staff experience and infrastructure are presently inadequate to manage the scale of change from isolated and fragmented general practice to coordinated coalitions of primary care organizations:

“… we’re understaffed, we’re under everything, we’ve been told that we’re understaffed and certain things for development (can’t be supported)… (Chair PCG1, January 2002).

“What we did have a problem with and I think the PCT always did, was the quality of the Board. I think there were only a small number of the Board who actually realized what the agenda was and what it meant and what the commitment would be and a lot of them really couldn’t cope with it...” (CEO PCG4, September 2001).

“… a lot more is expected of primary care teams … it won’t be delivered unless there is more investment and more staff” (Nurse Study).¹²

Feature 8

PCOs may have inherited financial deficits, which could dominate strategic thinking or have a limiting effect on organizational development. Interviews of finance directors revealed financial deficits that had a major impact on plans:

“The recent development money, a lot of that, a very significant proportion, has gone to supporting an Acute Trust” (Chair PCG3, September 2000).

“… we have this naive idea that we can reduce services but over the months being on the Board, I think we end up more worrying about how to contain the overspends and that also removes the focus of what you’re trying to do” (Clinical Governance Lead PCG3, September 2000).

Nurses noted greater access to training because of clinical governance, but inadequate funds to implement change (Nurse Study).¹²

DISCUSSION

This study of fledgling UK primary care organizations analyzed the partial and nuanced impact of some interesting recent management ideas and models designed to help them move away from traditional bureaucratic management styles and promote a greater organizational propensity to innovate, change, and learn.

There is a balance to be struck here. PCGs, as should all bureaucracies, must produce rules and protocols to ensure safety and provide consistent information about best practice. Such top-down processes, however, inhibit local innovation and change. Resistance to change is worsened when senior leaders have insufficient experience to facilitate locally owned innovation. Inexperienced leaders either fail to filter or magnify anxieties, thus missing opportunities in fast-changing situations.

The PCGs attempted to overcome this bureaucratizing tendency through bottom-up facilitation, helping practitioners to reflect on their work and make sense of their experiences. This approach had the desired effect of helping local practitioners to learn, but it rarely produced innovation. Most PCG innovations were existing models adapted to be locally relevant.

PCG1 provided the only example of true innovation in its whole-organization feedback of meeting behavior and cross-committee learning. No previous model was copied, and no one had the idea at the outset. It emerged as an idea in the moment and took advantage of a resource (the external facilitators) intended for another purpose. It was devised by the same PCG that had 2 previous failed plans, but not the PCG with the expressed intention to operate as a learning organization. Leadership in this PCG gave its officers confidence to test new ideas, make mistakes, and start again. Their innovation provided a potentially sustainable infrastructure of connected learning spaces to help staff to think in terms of whole systems rather than in narrow functions. There was no opportunity to observe longer term consequences of innovation.

The whole system participatory action research method enabled those from different backgrounds to develop innovative interpretations of diverse research insights, which changed the research focus. This process showed how “organizational design (can be) a verb rather than a noun”¹⁴ and how “incremental experiential learning … is a form of organizational intelligence.”¹⁵

This study therefore cautiously supports the contention that the principles of organizational learning and whole system participatory action research can increase the capacity of bureaucracies to learn and innovate, at least in receptive primary care organizations.¹⁶ But these concepts remain unfamiliar and are not free of problems. Further research could help identify what proportion of primary care organizations are receptive to more innovative organizational styles.

There were elements of this study that could have been done differently. Disallowing the research teams to discuss data in between conferences inhibited the generation of creative ideas and led to a sense of isolation. A better strategy would have been to have a series of workshops to explore emerging insights. Such workshops might have led to better use of data and made it easier to keep the overall aim of the project in view. A clearer timeline of when the direction of the research was intended to alter and how data were to be handled might have helped participants to see better the overall project rather than merely their part.