Conversion as a form of non-radical institutional change

In its investigation of institutional conversion, this article follows James Mahoney and Kathleen Thelen's injunction to take the “next step” of investigating “the sources of institutional change: sources that are not simply exogenous shocks or environmental shifts” (Reference Mahoney and Thelen2010, 3). The NRLS has been chosen as a “critical case” of institutional conversion, allowing “for direct assessment of received wisdom on causal mechanisms or linkages” (Mahoney and Thelen, Reference Mahoney and Thelen2009: 1). Mahoney and Thelen have, usefully, promoted the merits of examining institutions within the context of distributional power conflicts. They suggest that institutions often represent compromises between different power-seeking groups, and hence should often be viewed as inherently unstable.

While this approach has much to offer, it arguably fails to accord sufficient importance to the activities of internal institutional actors. These actors may be allied to powerful groups that are rooted outside the institution but, equally, they may fail to possess such connections (or at least, it may not be possible to “read off” their actions from the preference of such power coalitions) (Woll and Jacquot, Reference Woll and Jacquot2010). The article argues that, if we are truly to understand institutional change, it may be necessary to examine the activities of internal actors (the stuff, traditionally, of organisational analyses) as well as adopt a “political coalitional approach” (Thelen, Reference Thelen2009: 476), which examines the influence of broader power coalitions and how these have impacts internal to institutions.

The figure below outlines different factors promoting institutional conversion: intra-institutional factors, the unintended consequences of institutional design, relations with external bodies and environmental turbulence.

The importance of intra-institutional factors in explaining the stability and growth of agencies has not been extensively examined in neo-institutional work, but it has been the focus of much of the “bureaucratic politics” literature. This literature suggests that bureaucrats “are politicians, and bureaucracies are organizations of political actors” (Carpenter, Reference Carpenter2001: 352; see also Carpenter, Reference Carpenter2010; Teles, Reference Teles2008). As such, the growth or contraction of agencies does not result merely from functional pressures (Skowronek, Reference Skowronek1982), nor are agencies merely “passive actors” buffeted by external processes (Maor, Reference Maor2010: 154). According to these works, agencies’ jurisdictions will expand or contract depending not only on their relationship with other agencies (ibid.) but also on their leaders’ ability to “practice a politics of legitimacy” (Carpenter, Reference Carpenter2001: 352). Success within the “politics of legitimacy” involves agencies’ relationships with external actors, in particular, their ability to network with a “diverse [supportive] coalition wrought from the multiple networks in which they are engaged” (Carpenter Reference Carpenter2001: 352). Bureaucratic politics, and in particular the assertion and maintenance of bureaucratic autonomy is, then, a matter of “coalition” (Carpenter Reference Carpenter2001: 354) and “reputation” building (Maor, Reference Maor2010). This literature also notes that the capacities of agencies are often judged by the reputations of their chief executives – despite the fact that limits exist to the impact that even the most senior actors can have on highly complex organisations (Derthick, Reference Derthick1990: 187).

The article argues that intra-institutional factors can help explain why institutional conversion occurs, as well as the other factors identified by neo-institutional authors such as Streeck and Thelen (Reference Streeck and Thelen2005) and Mahoney and Thelen (Reference Mahoney and Thelen2009). These authors suggest that three factors may promote institutional conversion: the unintended consequences of institutional design, relations with external bodies, and environmental turbulence.

The results demonstrate that all three factors contributed to the institutional conversion of the NRLS. While it is, in many ways, no surprise that a particular type of institutional design might lead to unintended consequences – as with the undertaking of any purposive social action (Merton, Reference Merton1936) – this factor is often overlooked. Both policy-makers and, indeed, academics, often fail to appreciate the possibility for gaps between the rules that constitute institutions and the behaviour that follows from and contextualises them, gaps that may offer opportunities for political contestation (Thelen, Reference Thelen2009).

The creation, development and conversion of the NRLS highlights the influence of predispositions towards what could be described as quantitative rationalism (Porter, Reference Porter1995; Power, Reference Power2004) and negativity bias (Soroka, Reference Soroka2006) amongst some UK politicians and media. Here, quantitative rationalism refers to the tendency of actors to privilege quantitative over qualitative sources of information and to assume that quantitative information is in some way objective. Negativity bias refers to the phenomenon whereby “negative information more strongly influences people's evaluations than comparably extreme positive information” (Ito et al., Reference Ito, Larsen, Smith and Cacioppo1998: 887).

Both quantitative rationalism and negativity bias appear to have undermined the NRLS’ original, formative, role and were not, it appears, anticipated when the NRLS was created. For example, many commentators appeared determined to view incident reports lodged within the NRLS as a measure of patient safety, hence engaging in quantitative rationalism. These commentators presented incident reports as “objective” measures indicative of safety “levels”, and presented increasing numbers of reports as “evidence” of increasingly unsafe care. In contrast, the NPSA and a small number of other expert commentators were keen to emphasise that the NRLS database offered only a very partial indication of patient safety incident rates, given that only a small proportion of incidents tended to be reported to it (Vincent et al., Reference Vincent, Aylin, Franklin, Holmes, Iskander, Jacklin and Moorthy2008); and that rising incident rates were likely to be indicative of greater proclivity to report, rather than of increasingly unsafe care. The greater political and media resonance of quantitative rationalism and negative news compared with the expert position does not appear to have been anticipated by institutional founders, yet in the long-run led to the NRLS being undermined.

Streeck and Thelen (Reference Streeck and Thelen2005) and Thelen and Mahoney (Reference Mahoney and Thelen2009) have also suggested that the rules and norms constituting institutions often represent both initial and ongoing compromise between different power coalitions. As such, institutions may be shaped over time by changing power balances amongst cognate and competing institutions and, hence, by their relationship with external bodies. The NRLS (and its home agency, the NPSA) was faced with a highly dynamic institutional context where a variety of institutions arguably shared its focus on promoting patient safety. While the agency was able to cooperate with some of these and the processes they set in train, this arguably confused the NPSA's initial, formative, role to aid NHS organisations to learn from patient safety incidents (as opposed to providing, or helping to provide, summative measures of “safety”). As Mahoney and Thelen suggest (Reference Mahoney and Thelen2009: 22), “short-run behaviours” can be distinguished from “long-run strategies”, where “immediate rule-conforming behaviour” can conflict with “the overall goal of institutional maintenance”.

Finally, Streeck and Thelen (Reference Streeck and Thelen2005) and Thelen and Mahoney (Reference Mahoney and Thelen2009) have suggested that institutional change rarely results solely from external pressures (apart from under certain rather specific circumstances). Nonetheless, they do suggest that the changing balance of power between different groups can affect internal institutional dynamics, by offering more (or less) succour to particular intra-institutional elements. Hence, environmental turbulence can promote institutional conversion. While a change in the partisan composition of the UK government towards a more market-friendly coalition certainly appeared to encourage the use of NRLS data to drive resource allocation decisions, many of the processes that led to this were already in train under the previous government.

Mahoney and Thelen do consider how the “over-time distributional effects of institutions” can “trigger divisions among institutional power holders”, and how “subordinate groups” can organise and attempt to effect change after being undermined by existing institutional arrangements (2009: 9–10). However, such change processes tend to focus on different groups as part of power coalitions that have a life beyond the institution in question, rather than internal institutional dynamics. Yet, the case of the NRLS suggests that, in addition to the extra-institutional factors specified above, intra-institutional factors were important in explaining institutional conversion: in particular, the lack of a clear institutional rationale and the lack of internal stability. Hence, for example, the NPSA's leadership frequently appeared to alter the agency's approach to patient safety and failed to articulate a clear, consistent and resonant rationale for the agency's continued existence, which differentiated it from other institutional actors. This appeared to be less a result of external power coalitions than of internal issues such as successive changes in leadership.

Method

The article is based on the analysis of content, produced between 2001 and 2010, from parliamentary reports (from the UK's Hansard), institutional reports from the NPSA (including minutes of meetings, press releases and rapid response alerts) and press reports (across the broadsheet, tabloid and specialist press). First, relevant parliamentary discussions and press reports were collected by conducting a keyword search for “patient safety” and “NRLS” and/or “National Reporting and Learning System” and/or “National Reporting and Learning Service” (a misnomer frequently used in discussions of the NRLS), using Hansard (the UK database of parliamentary debates) and the relevant full-text websites for high-circulation UK newspapers. The discussions and press reports were then examined in detail to ascertain the manner in which data from the NRLS was used by authors and/or speakers. This process of inductive content analysis led the authors to identify the various key themes discussed below.

The findings presented here have been extensively discussed with a number of key actors from this policy area, including staff from the NPSA (who remain anonymous), to check our interpretation of events. These discussions have been undertaken rather than formal interviews given the low likelihood of staff being willing to go on the record whilst they were in the process of attempting to salvage their employment within the health service. The discussions corroborated the key findings presented here. However, any errors or omissions are the unique responsibility of the authors.

The National Reporting and Learning System

After the publication of an Institute of Medicine report in the US in 1999 on patient safety, many countries began to discuss plans for establishing national patient safety reporting systems in the framework of quality of care, risk management and clinical governance (Runciman et al., Reference Runciman, Sellen, Webb, Williamson, Currie, Morgan and Russell1993; Scally and Donaldson, Reference Scally and Donaldson1998; World Health Organization, 2005). In the UK, the NPSA was created in 2001 as a special health authority following the Kennedy inquiry into the Bristol Royal Infirmary scandal, where abnormally high death rates in babies receiving heart surgery were detected (Lyall, Reference Lyall2006). The NRLS enables UK health care staff (doctors, nurses, surgeons, etc.) to report information about patient safety incidents, including both “near-misses” and instances where a patient was directly harmed, into a national database. In July 2001, a pilot for an NRLS started in 28 NHS organisations, running until June 2002. This followed a number of NHS hospitals in England and Wales developing their own incident reporting (IR) systems from 1995 after the publication of national risk management standards promoting this (Sari et al., Reference Sari, Sheldon, Cracknell and Turnbull2007).

This was followed by development work with 39 NHS organisations (Parker, Reference Parker2004), with the NRLS being officially launched across the UK in February 2004. The NRLS was only the second national-level incident reporting system (British Medical Journal, 2004). By just April 2002, the pilot NRLS was “the biggest patient safety database in the world” (Greenwood, Reference Greenwood2003). The NPSA was initially described as analysing data from the NRLS to “develop national solutions that will make it easier to do things right and more difficult to do things wrong”. (Health Services Journal, 2003b).

The voluntary and confidential nature of the NRLS was initially presented as promoting more reporting of incidents by health care staff, and thus leading to greater effectiveness. The NPSA drew lessons from the data collected in the NRLS and, on the basis of this information, issued safety alerts to UK hospitals and other NHS organisations (as well as undertaking a wide variety of other safety-related activities).

However, the UK's new Conservative-Liberal Democrat coalition government announced in July 2010 that it would abolish the NPSA, following a review of quangos (non-departmental public bodies) (Smith, Reference Smith2010). This review concluded that the functions of the NPSA “do not of themselves need to be performed at arm's length of the Department [of Health]” (Department of Health, 2010: 26). The NRLS is to be shifted to a Patient Safety Subcommittee of a new NHS Commissioning Board. The new NHS Commissioning Board is to provide oversight of the new quasi-market within the NHS, whereby general practitioners (GPs, family doctors) will be able to commission services for their patients, in conjunction with other health professionals. In addition, the NHS Commissioning Board will itself commission some of the services that currently report patient safety incidents into the NRLS, including primary care and related services, and national and regionally specialised services.

The proposed Patient Safety Subcommittee” will cover “the whole function from getting evidence to working up evidence-based safe services” and thus, it is claimed, provide “an opportunity to preserve the synergy between learning and operational practice that already exists in the system” (ibid.), whilst other parts of the NPSA will be shifted elsewhere. The new NHS Commissioning Board is due to assume all its responsibilities by April 2013. The responsibilities of some other UK quangos with quasi-regulatory or regulatory roles (such as the National Information Governance Board for Health and Social Care) will be passed to the Care Quality Commission, an existing “super-regulator” (Department of Health, 2010).

The shifting of the NRLS’ functions to a new NHS Commissioning Board, following a move to compulsory reporting of serious untoward incidents, and the abolition of the NPSA suggest that the grounding principles of the NRLS have been entirely abandoned. The move suggests that, henceforth, the NRLS will be mainly used not as a source of information for the NHS to learn from itself but as a means for commissioners and patients to assess services’ safety as part of a newly emerging quasi-market for health services. In sum, the NRLS, rather than being a tool to support professional improvement, is becoming a mechanism to facilitate resource allocation.

The change in the location and function of the NRLS signals a quite remarkable institutional conversion. Initially, the NRLS’ role was entirely separate from resource allocation, with patient safety being seen as best promoted by a voluntary reporting scheme. Gradually, patient safety data such as that on “serious untoward events” (including deaths and surgery on the wrong side of the body, for example) was incorporated into funding decisions until, in July 2010, it was proposed that the entire NRLS should be subordinated to the commissioning system.

The following sections consider the role of a variety of factors promoting institutional conversion. Although downplayed in many accounts, we suggest that, first, intra-institutional factors were important in explaining the conversion of the NRLS. This was alongside factors that are more commonly identified such as the unintended consequences of institutional design, relations with external bodies and environmental turbulence.

Intra-institutional factors within the NRLS: instability and lack of clarity

The NRLS lacked a clear institutional rationale and was extremely unstable in its internal organisation over time, for reasons not reducible to external pressures. The analysis presented below focuses on two intra-institutional factors: the lack of a clear institutional rationale for the NRLS, and the lack of stability within the NPSA. Both of these internal factors increased the vulnerability of the NRLS to institutional conversion and, inter alia, its ability to build a strong reputation amongst external audiences.

The unintended consequences of institutional design

The view that incident reporting can drive up safety has informed the creation of incident reporting systems in industries including military, nuclear energy, oil, and rail and aviation (Macrae, Reference Macrae2008; Smith and Mahajan, Reference Smith and Mahajan2009). Experience from these sectors has informed the view that “for all practical purposes, a safe culture [can] be equated to an informed culture”, which itself results from a “reporting culture” whereby incidents are recorded and learned from (Reason, Reference Reason1998: 302). Given estimates that around 10 per cent of hospital admissions lead to the patient suffering from an adverse incident of some kind (Department of Health, 2000: 11), it was anticipated that analysis of these incidents would then enable the NHS to learn from its mistakes and make its systems, processes, equipment and practices safer. However, despite strong initial support from government and academic commentators, in practice the NRLS suffered severely from unintended consequences resulting from its institutional design, particularly those arising from what is described here as quantitative rationalism and negativity bias.

Quantitative rationalism and the NRLS

Data drawn from the NRLS were frequently presented as representing appropriate estimates of incident rates, as part of a narrative that health care was unsafe. For example, the Daily Mail newspaper suggested in 2005 that the NRLS had “estimated (sic.) that around 840 deaths could be linked to safety incidents” (Daily Mail, 2005) whilst The Times newspaper talked of “[t]en thousand safety alerts (sic.) over medication given to children […] being issued annually in the NHS” (Lister, Reference Lister2009). The NRLS was described, erroneously, as a “monitoring programme” (Lister, Reference Lister2006) instituted to “collect trustworthy data” (The Times, 2006). This was paralleled in discussions of particular types of adverse incident. Hence, a rapid response report on drug errors was described (wrongly) as an “official warning” by the Observer newspaper (Campbell, Reference Campbell2008). The Sunday Times newspaper spoke of more “than 200 errors with epidurals in 12 months” having “been detected” by the NRLS, even though these errors were revealed through voluntarily given information (Oakeshott, Reference Oakeshott2006). Such misinterpretation also, albeit less frequently, occurred in the UK specialist health press. Hence, for example, the British Medical Journal reported in 2005 that more “than half a million errors are made in patient care in acute NHS hospitals each year, according to the National Patient Safety Agency”, apparently ignoring the fact that within the same article, an NPSA spokesman had stated that “it was difficult to be precise on numbers” (Coombes, Reference Coombes2005).

One case study of this bias towards quantitative rationalism comes from the area of the physical safety of mental health patients. The NPSA commissioned a detailed analysis of the reports of rapes, sexual assaults and incidences of sexual harassment of mental health patients that had been reported to the NRLS, following the incorporation of mental health organisations into the reporting system. This research was part of a programme of work that allegedly constituted “the first time in the world that an agency has used a national incident reporting system to look at the safety of mental health patients” (Kmietowicz, Reference Kmietowicz2006).

However, two high-circulation UK daily newspapers, The Times and the Daily Mail, reported in July 2006 that the report had been held up for a number of months by the Department of Health due to its sensitivity (Craven and Merrick, Reference Craven and Merrick2006; The Times, 2006). The UK Department of Health, for its part, maintained that “some information remains unclear and work is underway to clarify this and ensure the report reflects the most complete picture” (Craven and Merrick, Reference Craven and Merrick2006). This was criticised by the political Opposition for the length of time it had taken (eight months) to verify a small number of cases (ibid.). The report was eventually published on the 17 July 2006, with the caveat that an additional enquiry was being carried out into those most serious incident reports (Lister, Reference Lister2006).

The NPSA was criticised by some commentators for failing to challenge the delay in publication, given that the report suggested some very serious failures to protect vulnerable women, including in particular the suspected occurrence of 19 rapes in mental health wards (The Times, 2006). However, it later transpired that of those 19 reported rapes, 13 allegations were subject to “serious doubt” (Hawkes, Reference Hawkes2006). As such, the delay in publication was arguably justified; yet this went unacknowledged by the press. This case illustrates the apparent tendency of both UK politicians and the press to try to present data from the NRLS as offering a picture of patient safety within the NHS, even when the quality of this data was highly questionable.

Figure 1

Factors promoting institutional conversion

The NPSA continually attempted, unsuccessfully, to emphasise publicly the partiality of incident reports as a measure of patient safety. As one of many examples, the NPSA's nursing lead for patient safety stressed publicly that “variation in reporting of patient safety incidents varied vastly, with higher reporting from community nurses and mental health staff than clinicians in general practice settings” (Carlowe, Reference Carlowe2009). This cautious and nuanced approach was also reflected in the NPSA's own presentation of data from the NRLS. Rather than presenting absolute numbers of patient safety incidents as some kind of measure of safety, the NPSA developed and promoted more sophisticated approaches as means of measuring patient safety. The most detailed of these was the NPSA's development of what it described as a Harm Susceptibility Model (Pham et al., Reference Pham, Colantuoni, Dominici, Shore, Macrae, Scobie, Fletcher, Cleary, Goeschel and Pronovost2010). This used the relative number of reports falling into certain categories in 20 high-reporting hospitals (known as trusts) in order to develop a model that could be used for inter- and intra-trust comparison. The model suggested that most variability in “harm”, as indicated by the NRLS, “occurred among trusts (55%), with differences in odds of harm across work areas within a trust accounting for 30%” (ibid., 2). The Harm Susceptibility Model related “the probability of near-miss events to the probability of harmful events”, and in doing so offered “a simple quantifiable measure of a hospital of work area's ability to catch and defend against errors” (ibid., 4).

Even the Harm Susceptibility Model was subject to a range of caveats identified by its creators. Most significantly, it was acknowledged that the ratio of harmful to non-harmful errors reported in different units might vary dependent on reporting biases, differences in coding errors as harmful or non-harmful, and a range of other related factors (ibid., 5). In addition, by viewing “non-harmful” incidents as related to resilience, the authors acknowledged that they discounted the potential importance of “mere chance or luck on the day” (implying completely degraded risk resiliency; the “as bad as it gets but somehow we didn't hurt anyone” situation) (ibid.).

In summary, whilst the NPSA continually refrained from describing NRLS data as indicative of safety levels without further qualification, this was effectively ignored by external commentators. Whilst the superior expertise and information-gathering capacity of the NPSA might have been expected to have strengthened its autonomy in relation to external actors (Carpenter, Reference Carpenter2001: 364), it was unable to control the use made of such information once it was released into the public domain.

Relations between the NRLS and external bodies

The NPSA operated within an almost suffocatingly complex and competitive institutional context, which changed considerably over time. Lacking formal powers to enforce the safety alerts it produced from NRLS data, the NPSA was dependent on the support of outside institutions for the success of its activities (Lyall, Reference Lyall2006). However, both the institutional ecology and organisational priorities of these bodies altered almost continually. This made it more difficult for the agency to articulate a clear identity that was separate from, but complementary to, such other bodies.

First of all, the creation of the NPSA coincided with the intensification of regulatory mechanisms covering the NHS. These included, for example, the development of new regulatory regimes covering medical staff, such as the revalidation scheme run by the General Medical Council (the body that registers doctors to practice in the UK), which required doctors to be able to indicate a capacity to learn from significant events. This paralleled the development and increasing importance of inspection-based organisations such as the (then) Commission for Healthcare Audit and Inspection (now an element of the Care Quality Commission super-regulator). In order, apparently, to demonstrate its utility in the regulatory context, the NPSA developed a standard for safety that was incorporated into this commission's inspections process (Katikireddi, Reference Katikireddi2004). However, as discussed above, the leadership of the NPSA never precisely articulated the degree to which the agency's activities could or should be viewed as summative as opposed to formative.

It is also arguable that some of the NPSA's role came to be undertaken by a new Patient Safety First Campaign, which focused on interventions relating to the highest-risk areas as identified by the NRLS. This campaign was formally supported by the NPSA in addition to other institutions, and ran from June 2008 (Creamer, Reference Creamer2008). The English campaign was described as based on the “principles of social movements”, was run by NHS staff and occurred in tandem with the creation of similar campaigns in Wales, Scotland and a number of other nations (Woodward, Reference Woodward2009). Its approach was similar to that of the US Institute for Healthcare Improvement, which differed fundamentally from that of the NPSA given the latter's focus on incident reporting. While the NPSA formally supported the Patient Safety First Campaign, the campaign formed yet another (competing) element of an already crowded institutional context.

Another potentially competing body was the NHS Information Centre, which was established as a “special health authority” in April 2005. The NHS Information Centre did not suggest that incident data itself could be used to measure safety. However, the centre developed a different measure of safety based on the NRLS, within its set of voluntary Indicators for Quality Improvement, based on the degree to which health care organisations had made reports to the NRLS within the preceding period.

The NPSA's development was also concurrent with an increasing focus on safety and quality within governmental health care funding decisions. In some, but not all, cases, the NPSA supported decision-making in this area, despite the fact that this involved legitimising the view that safety could be measured in some way through the collection of incident data. As mentioned, in some cases, the NPSA was not deeply involved in the creation and operation of these funding processes. For example, in order to gain funds through the Quality and Outcomes Framework within primary care, following the concluding of a new GP (family doctor) contract between GPs and the UK government in April 2004, general practices (groupings of family doctors) were required to carry out 12 “significant event audits” over a three-year period and submit reports of these (Gould, Reference Gould2008). Significant event audits differ from investigations of adverse incidents, since “significant events” can have positive as well as negative outcomes. “Significant events” therefore were not generally reported to the NRLS, nor was data from the NRLS used for the purpose of significant event audits. The development of the significant event audit system, apparently at one remove from the NPSA, was perhaps surprising given the fact that the agency was in theory meant to cover the entire NHS, and yet very few incidents within primary care appeared to be reported to it (only 1 in 200 reports to the NRLS in 2008 were made from primary care, despite this sector delivering the bulk of health care services for the general population) (Hoffman et al., Reference Hoffman, Beyer, Rohe, Genishen and Gerlach2008).

In contrast, the NPSA was instrumental in developing the concept of “never events”, which was adopted as part of commissioning arrangements by Primary Care Trusts (PCTs)Footnote 2 following the New Labour Minister Lord Darzi's recommendation to do so as part of his report, High Quality Care for All. The concept of a never event was popularised in the US, where it is linked to insurance systems. Hence, Medicare, and some private health care insurers in the US, will not now pay out to hospitals for the costs of dealing with “never events” (nor for hospital-acquired infections such as MRSA) where these occurred (or were contracted) as part of an inpatient stay.

Eight “core” never events were selected to form a Never Events Framework 2009/10, ranging from wrong site surgery to inpatient suicide using non-collapsible rails. The NPSA helped promote this framework, to work as a pilot initially, to later be developed for use by commissioning PCTs (Fletcher and Huehns, Reference Fletcher and Huehns2009). Initially, PCTs were urged to simply discuss never event rates with providers from whom they commissioned services; in return, PCTs were to be informed of the occurrence of never events by their providers. PCTs were then to be encouraged to report, publicly, on the occurrence of never events, meaning that their role extended from that of commissioning to publicising safety information. It was anticipated that rates of never events would then, at some point in the future, be used to inform commissioning decisions by PCTs. The Never Events Framework clearly prioritised the use of incident reports as summative measures of safety that could be used to drive funding decisions, in departure from the NPSA's initial formative use of the incident database.

In summary, the NRLS (and its home agency, the NPSA) was faced with a highly dynamic institutional context where a variety of institutions arguably shared its focus on promoting patient safety. While the agency was able to cooperate with some of these and the processes they set in train, this arguably confused the NPSA's initial, formative, role to aid NHS organisations to learn from patient safety incidents (as opposed to providing, or helping to provide, summative measures of safety). While cooperation with regulators and the commissioning system may have had the short-term impact of ensuring that the NRLS was seen as a relevant and useful information resource, it may have undermined over the longer term the case for maintaining the NRLS as a formative (as opposed to summative) data resource.

The NRLS’ external conditions – constantly changing over time

The coming to power of a new core executive in the UK, which strongly supports radical (as opposed to partial) marketisation of the UK NHS, driven by patient choice and GP commissioning, could be seen as an example of a particularly significant change in the power resources of external coalitions. Whilst the NPSA possessed a patient experience and public involvement directorate and promoted the reporting of incidents by patients and their involvement in resolving safety issues (Health Services Journal, 2003a), it had not linked providing information on patient safety (via the NRLS) to facilitating patient choice in the manner proposed by the new UK government.

Nonetheless, it is arguable that external pressure even before this point had led to information from the NRLS being used in a summative manner, against the NPSA's original rationale of formatively aiding NHS organisations to learn from patient safety incidents. Hence, in 2006, a report from a committee within the UK Parliament, the Public Accounts Committee, suggested that the Department of Health should consider the publication of safety data for different NHS trusts (hospitals, collections of GP surgeries, community services, etc.) “as part of the NHS choice agenda”, given that the NRLS was not currently “tooled up” to do so (Clews, Reference Clews2007). It was also suggested that “a collapse in public confidence” was starting to “force politicians to take a more draconian approach”, and that this had allegedly been the case with “repeated errors and associated fatalities in the administration of intravenous/intrathecal chemotherapy”, over which the Department of Health had “resorted to threats of high-level retribution” (Burns, Reference Burns2007).

These factors appear to have motivated a major change in the use of NRLS data in April 2010. After this point, reporting of serious untoward incidents became mandatory. Incidents had to be reported to the NRLS, which would then pass on the data to the Care Quality Commission, which would then use the data for its registration process (National Patient Safety Agency, 2010). The decision to require the NPSA to be involved in mandatory reporting in this way appears to have been made by government with minimal involvement from the NPSA itself.

While the external context may, therefore, have changed considerably with the election of a new government, the previous UK government also appeared to have moved towards the view that NRLS data could, and should, be used for summative as well as formative purposes. It is, therefore, arguable that the shifting of the NRLS to the new NHS Commissioning Board represents the end of a long process of institutional conversion rather than an abrupt and dramatic change.