Methods

The study was approved by a multicentre research ethics committee. Research governance approval was obtained at each of the hospital trusts participating in the first phase of the SPI. We recruited participants from all four hospital sites by asking hospital contacts involved in the SPI to identify people in particular roles, including the chief executive, medical director, senior nurses, senior pharmacists, and others involved at a strategic level with implementation. Interviews were also conducted with a small number of stakeholders whom we term the designer stakeholders: these were senior individuals in the Health Foundation and the Institute for Health Improvement (IHI) who had been involved in commissioning and designing the SPI and organising the delivery of learning sessions and support for participating hospitals. Materials used at these learning sessions were available to the research team.

Participants were interviewed by telephone using a semistructured prompt guide. The emphasis in the interviews was on allowing participants to describe the SPI in their own terms. We were interested to identify what participants saw as the aims of the SPI, what ToC they were working with and what barriers and facilitators they saw to change. We also explored participants' views on the acceptability and validity of the initiative. Interviews were transcribed verbatim. Analysis of the interviews was based on the constant comparative method.13 “Open” codes to describe each unit of meaning were initially generated. Through comparison across transcripts, the open codes were developed into higher-order thematic categories and subcategories to provide frameworks for coding, assisted by QSR N5 software (QSR N6, Victoria, Australia). Documents and other materials from learning sessions were analysed informally and treated as background information.

Results

The SPI began in the four participating hospitals in January 2005, and interviews were conducted over the period November 2005 to November 2006. Interviews were conducted with 60 hospital stakeholders (14 from site A, 14 from B, 19 from C and 13 from D). Fourteen participants were in very senior strategic positions, including at each hospital the chief executive, the medical director, and the nursing director, and at least one clinical director. Other participants included 36 staff in senior managerial positions at the “front line”, including consultant-grade doctors, nurses and pharmacists. The remaining six participants were in mid-level managerial positions (such as theatre managers and patient safety coordinators). Although some hospital participants had been directly involved in the competitive bidding process for their trust to participate in the SPI programme, most had not. Five designer stakeholders, all at senior level in their organisations, were also interviewed.

The ToC underlying the SPI was communicated to the hospital stakeholders through the learning sessions and other resources provided by the IHI. The “theory of change” was not formally expressed as a “theory of change” in these resources and materials, and—in part because the interviews were conducted over a relatively long period of time—to some extent the ToC emerged “in use” rather than being fully specified a priori. Nonetheless, the interviews suggested a very high degree of congruence between participants, even between those interviewed earlier and later in the process of the SPI's embedding.

Aims of the SPI

All hospital participants spoke of the initiative as aiming to address patient safety problems, particularly by reducing the number of adverse incidents and making the hospital environment safer. All participants emphasised the legitimacy and importance of this aspiration, locating its warrant in a narrative about an uncontested need to assert control over risk.

It's trying to improve patient care generally, but specifically to try to reduce the number of adverse events. I think with a bit of hard work and persistence to institute it, it could embed quite well in the NHS and it would be a healthy mechanism for maintaining standards and improving standards. (consultant, site C)

If we can't make patients safe in hospital … then we're wasting our time. … So I strongly recommend it and I think it is transformational. (head of nursing, site B)

Seven participants described SPI solely in terms of cultural change and 11 solely in terms of changes to systems, but most (42) described SPI in terms of changes both to culture (such as increasing awareness or changing attitudes) and to systems, practices and processes, particularly in increasing consistency and monitoring and implementing evidence-based practice.

it is … about us looking at how we change the culture and the way that people think so that everybody realises this isn't an extra add-on to their day-job which probably at the moment some staff do perceive, but actually this is how we do business here and safety is our first and foremost, you know, ethos. (patient safety manager, site D)

It's trying to look at the high risk areas and actually assessing whether our processes are reliable at the moment and if they're actually putting patients at increased risk, and trying to identify ways to improve those processes. (pharmacist, site C)

Fifty-three hospital participants expressed positive attitudes towards the SPI during their interview, although 19 of these also made statements that indicated a muted positive or ambivalent attitude. Less positive attitudes included expressions of scepticism (10) or feeling apprehensive, overwhelmed or confused by the initiative (17). These more negative accounts tended to be descriptions of how the interviewees had felt when the initiative had first been introduced; many of those interviewed later in the study described having developed more positive attitudes over time. Four hospital participants described feeling only negatively about the initiative. Thus, for the most part, strategic stakeholders saw the SPI as a legitimate and promising response to the problems of patient safety.

How will it work?

When asked to describe how the SPI would work (ie, the programme ToC), five hospital participants explicitly stated that they were unable to give an account of any specific element of the SPI, and a further two appeared to have very little knowledge of the initiative. However, most of the accounts from hospital stakeholders appeared to demonstrate a shared understanding of the main elements of the initiative. In many cases, hospital stakeholders' accounts demonstrated a high degree of symmetry with designer stakeholders, as these extracts illustrate in their emphasis on the need for standardisation of practice as a way of improving safety:

[we] believe very strongly that by the application of our model of the rapid cycle and the reliability principles that we're going to have less variability and more reliable processes to ensure that the care is delivered the way that we would want it to be delivered. (IHI “designer”)

It's about standardisation. A lot of these things, over the years consultants have developed their own particular practices and of course that's a problem if you've got juniors going from one consultant to another, they get a bit confused in the end and so it doesn't happen. So it's about bringing everyone together and making an agreement about what the process is than then making sure it happens. (pharmacist, site D)

An example of the shared understanding among strategic stakeholders of the ToC could be found in hospital stakeholders' descriptions of the Plan-Do-Study-Act (PDSA) cycle (see box 2), which was described as a critical element of the SPI in the designer accounts and in the learning sessions. Most (41) hospital participants offered a reasonable description of PDSA, with 22 explicitly using the term PDSA, although accounts varied considerably in their precision. For many, a key element of their ToC relating to PDSA could be summarised as “seeing is believing”, and an emphasis on how involving staff could lead to ownership and hence to effective changes in practice. Again, there was considerable symmetry between designer stakeholders and hospital stakeholders in their accounts.

it's based on the IHI's own kind of interpretation of … approaches to quality improvement which are based around what they call these PDSA cycles which is a [way] of getting people into making relatively small tests of change—doing the measurements, introducing the improvements, reviewing and then moving on and building that up over time to give everybody involved a sense of purpose and confidence. (Health Foundation, “designer”)

so there's a kind of planning stage on how it's gonna work and then they start doing it in a small group of patients first one patient as far as I know and then looked at how it's going and then roll it out to three patients, five patients and then everybody. (health professional in haematology, site C)

In addition to identifying the principal elements of the design of the SPI programme, participants often identified very specific strands or components of the initiative. These included leadership (37), interventions relating to general ward (35), intensive/critical care (34), perioperative/theatres (40), medication/drug errors (45) and hospital-acquired infections (33). Participants also spoke about how the initiative worked in terms of practices that have been adopted, including s-bar (15), early warning and critical response system (29), safety huddles (15) and multidisciplinary teams (12).

Securing implementation

Participants' accounts highlighted factors that they believed would facilitate the implementation of the initiative. For hospital stakeholders, these included good leadership in the trust (43), motivation and commitment of staff (41), the existing culture or structure of the trust (42), resources (18), taking a bottom-up approach and ensuring local ownership (15), good information and communication (22), education and training (9), sharing of experiences with others involved in SPI (9), and evidence that the initiative works in practice (14). Many (42) talked of factors specific to their organisation that were already in place as providing “fertile ground” for SPI to embed. Such factors included people being receptive to change, a good track record with clinical and risk management, good communication, high levels of awareness and an open culture.

Well I think it is more likely to work here. Because I think we have a culture of openness, freely admitting mistakes, good communication. We're regarded I suspect as a can-do organisation. … I think we're an organisation that readily accepts change. (medical director, site A)

I think that it's a very positive and supportive organisation … so it's not seen as something that's being brought in from the top. It's hopefully being seen as a genuine culture change. (clinical director, site C)

However, most (57) hospital stakeholders also described barriers to the initiative's successful implementation. For these stakeholders, success was seen as depending to a large extent on how far other stakeholders in the organisation—particularly those at the sharp end—could be mobilised around the initiative. Participants cited difficulties in changing attitudes and culture, and particularly in getting front-line staff to realise that their current practice was not necessarily safe practice; reconciling (new) standardised practices with clinical autonomy; problems in getting people to do things properly (due to lack of knowledge, education, engagement or time); lack of ownership; lack of leadership; difficulties communicating within multidisciplinary teams; and people not believing the evidence.

you do have that element where you know individuals would think that they work safely. So you know kind of like, “Well I always make sure my patient is safe”. (senior sister, site D)

Accounts stressed the need for the changes in practices and attitudes encouraged by the SPI to become “taken for granted”, and reproduced routinely in organisational settings without contestation or resistance, but many strategic stakeholders feared that the SPI would not enjoy the legitimacy it needed among front-line practitioners. The most common view (raised by 52 hospital stakeholders) was that “people” issues were hugely important, and that although some people in the organisation were enthusiastic, other sharp-end and often powerful stakeholders (especially doctors) were “cynical”, “wary”, “too busy” or “resistant”, although only seven participants felt that people in their organisation primarily felt negatively or sceptical towards the initiative.

I think the reaction has been mixed. I think there's been a lot of people who have felt it's nothing to do with them and therefore haven't shown any interest. I think there's been a lot of people who certainly at the outset felt it was just yet another thing that the management had signed us up for. I think there were some people who thought, “Yeah actually this has got some useful things in it but some of it's not for us”, and I think there were one or two zealots who thought it was the greatest thing that they'd ever heard. (consultant, site C)

Really just the staff are upset about having more work to do. I would say it would be the general feeling on the on the ground. (senior nurse, site B)

Specific barriers to implementation identified in strategic stakeholders' accounts included increased work and bureaucracy, lack of communication about the initiative, trying to take on too many changes at once, lack of preparation, unrealistic time scales and concerns about maintaining momentum over longer periods of time. The “complainability” of these issues was seen as being a threat to legitimacy, in part because of the costs they imposed.

When you start to implement change and bring maybe an additional form that they have to fill in, they automatically, you get negative feedback … “We're filling in enough forms and we've enough to do as it is. We haven't time for this” (pharmacist, site B)

These doubts about the organisational feasibility of implementing the initiative were joined by doubts about the scientific legitimacy of some aspects of the programme in many (32) hospital stakeholder accounts. Scientific doubts arose in relation to difficulties measuring change and reaching targets in that relate to changes in behaviour, lack of evidence for some of the interventions, changes that did not fit easily with current practice, problems with definitions leading to difficulty in measuring outcomes (particularly in relation to infections), the abstract nature of the initiative, and problems with translating and applying an programme developed in the USA to the UK context. These accounts often appeared to point to the existence of professional subcultures that did not share the same view of specific issues as the SPI.

[There has been] a lack of knowledge or a lack of belief if you like in the evidence, and there have been a couple of points in the surgical site infection bundle where people have questioned the evidence that has been given forward for doing something, and in fact in one of those instances they, with regards to beta blockers there is such widespread disbelief in the evidence that the Safer Patient Initiative has watered that down now. (medical director, site B)

…we're not sure what an infection is you know and I feel strongly … maybe it's from a scientific background as opposed to a management background but if we're counting apples and oranges we need to know what an apple is. And not just a round fruit. (clinical director, site A)

Structural barriers related to the organisation/trust were raised by 50 participants and included the large size of the organisation, lack of systems already in place, lack of resources (money or staff) and the mobility of medical staff.

I think you know staffing issues is a big thing, you plan to staff a ward, you get last minute sickness, you get bank and agency nurses, and that I think is the difference really if you're getting agency nurses you haven't always got that ownership … You know it's the usual things within the NHS isn't it? It's time, staffing levels, and I think they're probably the biggest factors around why things work and things don't. (senior sister, site D)

Discussion

This paper has identified that an initiative aimed at improving patient safety across the entire hospital was broadly welcomed by strategic stakeholders in four participating hospitals. The ToC was communicated through learning sessions and other resources and also emerged in use, and could be characterised as seeking to identify risks within organisations, to make aspects of practice that contribute to those risks visible and to make orderly modifications to practices to reduce risks. Overall, there generally appeared among study participants to be a broadly shared vision and commitment to the principles and practices associated with the ToC. The ToC underlying the initiative was generally understood, although 7 of the 60 hospital stakeholders were unable to describe it. The aims of the initiative were broadly seen by hospital stakeholders as legitimate and sound. Most stakeholders expressed positive views about the initiative, with many showing considerable enthusiasm about the potential for the SPI to effect change. In Suchman's12 terms, the initiative enjoyed both “pragmatic legitimacy”, where it was seen to be of benefit to participants' organisations, and “moral legitimacy”, where it was seen as the right thing to do. However, participants also had specific doubts, particularly relating to feasibility of implementation and scientific legitimacy of some elements of the initiative.

It is important to access the views of strategic stakeholders in the early stages of a programme because, as organisational analyses have emphasised, the entrepreneur role is central to institutional processes. It could be argued that it is unsurprising that those at a strategic level in organisations who had chosen to take part in the programme were broadly supportive, but it was by no means self-evident that even early adopters and advocates for the initiative would share the same ToC, or views on legitimacy, across four different organisations and in widely differing settings within those organisations. Our study provides insight into those with organisation-wide responsibilities as well as into those with responsibilities in particular clinical areas, which had differing levels and types of strategic overview and varied in distance from the sharp end. Although overall broadly supportive, their views were not uniformly positive, either regarding the legitimacy of specific elements of the intervention or whether it would succeed in its objectives. Many stakeholders perceived that the initiative might encounter resistance at the sharp end, where many of the changes needed to occur. The blunt end is the source of demands that sharp-end practitioners must meet, and this can often lead to conflict.14 Threats to the practices and principles of the SPI becoming institutionalised were perceived to lie in possible contestation about the legitimacy and relevance of the initiative among front-line staff. Strategic stakeholders feared that, even when using the PDSA cycle to persuade and engage individuals, introducing normative rules about professional behaviour would invite trouble associated with ruling “against the grain”, and that individuals would incline towards existing norms within their individual locations15 rather than committing to new ways of doing things. Resource-related issues were also seen to have the potential to undermine the initiative because of their practical implications and because of their capacity for further destabilising legitimacy.

A limitation of this study was that the number of strategic stakeholders at each site was relatively small—on average 15 per hospital—and the assessment of the congruity between the accounts of designer stakeholders and hospital stakeholders was limited by the small number of designers interviewed. It is also possible that those who agreed to be interviewed were to some extent self-selected enthusiasts. It is also important to recognise that, to some extent, these interviews are publicly acceptable accounts of participants' views, offered, like most accounts, to present an identity that is competent, moral and responsible,16 and therefore offering uncertain access to any “inner” feelings or views. Nonetheless, this study does offer insights that can transfer outside their context of origin. First, it suggests that it is possible to get strategic-level individuals across geographically spread and organisationally diverse settings, and interviewed over a period of time, to understand and agree upon a ToC that can be used in their organisations. Second, it suggests that people at this strategic level are able to recognise the competing interests within their organisations that forms the context of negotiation in which the programme of change is enacted. In particular, they are able to acknowledge the complexities of how to mobilise people, technologies, organisational structures, resources and “culture” around a new effort.17

These findings point to the importance of identifying what ToC people are working with at the strategic end of patient safety initiatives. They also underline the need to identify how far that ToC is seen as being strategically legitimate. Further investigation is needed to explore implementation of the strategic vision: it is possible the actual change model used at a local site to achieve the programme goals may vary from that defined by the ToC intended by the designers, regardless of how well strategic stakeholders learned this theory. Thus, the programme as designed (including process, goals, tools and methods) must be distinguished from the programme as implemented, perceived or experienced locally. This in-use theory may well, of course, have advantages that might not be evident at the design phase and would benefit from systematic evaluation. Such work is underway as part of our programme of evaluation of the SPI.