Introduction

The governing of public health services in many OECD (Organisation for Economic Co-operation and Development) countries changed with the advent of New Public Management Reforms from around the 1980s onwards (Bevan and Hood, 2006; Pollitt and Bouckaert, 2011). Central public health authorities, ministry of health and regional councils responsible for the hospitals, have tried to make hospitals and their activities accountable. Accreditation is one of many ways in which central political authorities have tried to render medical staff accountable for the quality of their work. Of course this ambition to render medical services accountable is not unique. Thus, over the last few decades, most public service institutions in the OECD countries have seen an expansion in demands for various forms of accountability (Power, 1997). In particular, there has been a growth in administrative accountability measures, including value-for-money auditing, whereby politicians and top-level managers seek to hold subordinate public service agencies accountable for the quality of the services produced (Flinders, 2001). This trend has made some warn of an accountability overload (Koppell, 2005; Lewis and Triantafillou, 2012).Footnote 1

In November 2001, the Danish Ministry of Health decided that a national quality system should be developed. This system should be compulsory for all Danish public hospitals. Following a protracted political process, the so-called Danish Quality Model (DDKM), a comprehensive accreditation system, was implemented in 2009 with a view to assure the quality of hospital services. The result of this process was an accreditation model containing 82 standards and a staggering 570 indicators specifying how the standards are to be put into operation and verified. The protracted process leading to the implementation of the DDKM in 2009 was not only a result of the necessity to solve complex technical matters, but also a result of widespread scepticism and criticism from almost all parts of the medical profession. They argued that not only were there no evidence for the clinical benefits of accreditation, the costs to make the system work would be significant and could only be covered by cutting in other, necessary hospital services. Notwithstanding these – in my view – quite sober criticisms made by a powerful interest group, the political authorities headed by the Ministry of Health went ahead with a compulsory accreditation system.

Some general explanations have been offered for the growth of administrative forms of accountability, linking this largely to the New Public Management Reforms and the attempts at delegating responsibility for the ways in which public services are produced and delivered to administrators and frontline staff within broadly defined political goals (Lapsley, 1999; O’Faircheallaigh et al, 1999). There is also a substantial literature trying to disentangle the factors influencing professional groups’ adherence to clinical guidelines (for example, Baiardini et al, 2009; Gurses et al, 2010). A general insight here is that new and externally imposed guidelines must resonate with health professionals’ standards of good clinical practice (for example, Brown, 2011). A meta-review of existing studies found that the characteristics of five factors were important: the guidelines themselves, the implementation strategies, the professionals, the patients and the environment (Francke et al, 2008). With regard to the characteristics of the guidelines themselves, the reviewed studies suggest that simplicity is the most important factor for securing compliance. Only one study points to the importance of local development of guidelines to enhance adherence (Sachs, 2006). In brief, there is a need to examine how the character of the guidelines developed through DDKM may shape the ways in which medical staff uses these.

Thus, the article addresses two distinct, though interrelated research questions. First, why does hospital staff seem to accept, albeit grudgingly and partially, DDKM? To the extent that DDKM may be seen as accountability overload in the sense that its clinical effects are unproven while it clearly imposes substantial administrative burden, this is I think a rather obvious question to ask. Second, how does DDKM contribute to rendering Danish hospital activities accountable? This question is based on my hypothesis, further explained below, that the specific way in which these activities are made accountable enable a particular way of governing the medical staff. This form of governing is important for making the staff largely accept DDKM as part of their everyday practice.

It is argued, first, that Danish hospital services are rendered accountable through the employment of a wide range of administrative and clinical standards and indicators. The medical practices must be translated and in the process partially modified in order to be in lines with these standards and indicators. Second, DDKM has been difficult to resist and to some extent accepted because its procedural standards work through accountable forms of freedom. The article is structured as follows: first, I review competing explanatory frameworks and argue for the potential of the Foucauldian notion of government in addressing the kind of accountability pursued by DDKM as a particular way of rendering medical activities governable. After a brief methodological account of the case study, the political process of introducing hospital accreditation is analysed in two steps. First, the design and mechanism of DDKM is analysed. Then, I examine the way in which DDKM is brought into play at the hospitals through the enactment of organizational and clinical guidelines. The article finally discusses these findings and concludes that the accountability produced through DDKM hinges not on strict compliance with guidelines, but rather on making the staff actively participate in the forming of relevant guidelines and, not least, accounting for the ways in which they seek to adhere to these.

Grasping the Governing of Medical Conduct by Standards

A number of well-trodden analytical approaches seem relevant for trying to explain the way in which professional groups respond to demands of managerial accountability. First, rational choice in the sense of agents driven by maximizing self-interest seems a probable explanation for the quality managers at the hospitals because they may make a career out of being in charge of quality management schemes (Dunleavy, 1985). This may include part of the medical staff that is appointed to be in charge of quality management schemes. Indeed, the present case study does indicate that some nurses regard the local implementation of the DDKM as a career opportunity. However, rational choice seems a rather improbable explanation for actions of the bulk of medical staff whose career depends less on how they abide by managerial accountability standards, and more on their ability to produce clinical results and having these results published in prestigious medical journals.

Second, isomorphic pressure, in particular regulative pressures (DiMaggio and Powell, 1983), may have played a role in the spread of a new logic of appropriateness guiding local hospital conduct (March and Olsen, 1989). As some hospitals start complying with new guidelines, colleagues in other institutions may feel compelled to join in because of peer pressure and, perhaps, because some may fear losing their job in cases of overt non-compliance. While it does not seem very likely that hospital management would go as far as firing medical staff blatantly resisting accreditation, we may expect that medical staff would at least pay lip service to the potential benefits of accreditation. That is, one could expect that the medical staff would employ more subtle resistance strategies by thwarting or neglecting the standards and indicators in the everyday making of administrative accountability.

Such ‘subterraneous’ resistance strategies lead us to a third explanatory thesis, namely de-coupling. As argued by Meyer and Rowan some time ago, professions may formally accept externally imposed reform schemes and provide an image of implementation to the outside world, while the everyday technical practices of the organization remain largely unchanged (Meyer and Rowan, 1977). In the present case, this would entail that the hospital quality management department produces impeccable accreditation reports and instigates accounting rituals that really do not change the core of medical practices. While it seems reasonable to assume that large and complex public organizations such as hospitals may be able to fool some superior political-administrative authorities some of the time, it seems unlikely that they are able to fool all of them all the time (Scott, 2004). Also, local hospital quality managers will risk their career if they let the de-coupling strategy advance too far.

The analytical frameworks discussed so far do provide useful suggestions, but they remain partial at best. In particular, they seem to overlook the ways in which medical services are rendered accountable and the ways in which accreditation hinges on the active participation of the medical staff. Accordingly, we need a framework that is better able to illuminate the ways in which accreditation seeks to shape the conduct of medical practitioners not by thwarting their will and freedom, but by mobilizing and shaping them in distinct ways. Accreditation is here regarded as a mode of accountability, which in turn is understood as the practice by which a person or group (accountee) accounts for her or his action to another person or group. This understanding is close but not identical to Mark Bovens’ definition: ‘the relationship between an actor and a forum in which the actor has an obligation to explain and justify his or her conduct, the forum can pose questions and pass judgement, and the actor may face consequences’ (Bovens, 2007). Both understandings emphasize the relationship between the producer and the consumer of accounts. However, Bovens’ definition does not give much attention to the practice of producing accounts. This is important for the present purposes inasmuch as the very way in which accounts are produced enables particular ways of governing the said activity – by the account producer and the account recipient.

It has been argued, in my view convincingly, that things, people and activities are not accountable per se, but have to be made accountable (Power, 1996). In the process of making things, people and activities accountable, they are changed (Hoskin, 1996). The ways in which activities, such as public health services, are rendered accountable may be changed in several ways. In the context of accreditation, two distinct but closely interrelated changes are relevant, namely visibility and governability. All forms of accountability entail that the activities are made visible in new ways. To account for an activity is to make it visible to others in ways that differ from the grid of visibility employed by the practitioner in his or her everyday work. If the grid was the same, there would really be no reason to provide (additional) accounts. In other words, the activities have to be made visible to others than those conducting the activities. Those others rarely have the resources or time to observe the practices made accountable, nor do they have the expertise to understand them. Thus, visualization entails that the complex activities need to be conveyed in a manner that simplifies these (Porter, 1994; Munro, 2004; Dahler-Larsen, 2014).

To render public service activities accountable is not only about producing simplifying accounts of these practices, but accounts that seek to render these practices governable to the outside world, notably political authorities. To account for a set of activities also implies rendering these governable in particular ways. This should not be very surprising inasmuch as one of the key ambitions of holding someone to account is to be able to govern their conduct – at least in the case when the result of the account giving is not satisfactory. I am using the term government here to denote the ways in which activities are rendered amenable to more or less systematic attempts to orchestrate and direct these in a particular direction (Foucault, 2008). This is less a matter of controlling or determining the ways in which these activities are taking place, and more one of propagating a form of conduct conducive to wider political goals. More precisely, the notion of government is defined as the conduct of conduct and entails a certain reflexive dimension. That is, government hinges on the self-governing capacities of those subjected to government. Thus, we may distinguish between the governing of others (power) and the governing of the self – by the self (freedom). The former implies that medical activities are rendered governable to politicians, public managers and quality managers. The latter implies that the medical activities are rendered governable to the medical staff itself and its peers (collegial associations and so on).

A number of scholars have used the term neoliberal or advanced liberal governmentality to designate the political rationalities informing the more or less systematic attempts to link the governing of others (power) with the governing of the self by the self (freedom) (for example, Rose, 1999).Footnote 2 Thus, we may find that the exercise of power over a certain group is not limited by the latter’s exercise of freedom. On the contrary, power depends upon freedom and may be exercised even more effectively by working through rather than against the self-governing practices of those whose conduct is to be modified. This may sound as nonsensical inasmuch as power is usually seen as the very anti-thesis of freedom: when there is power there cannot be freedom and vice versa. In contrast to this common understanding of the relationship between power and freedom, the Foucauldian understanding allows not only power and freedom to coexist in a given situation, but goes on to allow for the former to depend upon the stimulation and guidance of the latter (Foucault, 1994). Accordingly, freedom is understood here not as the set of actions completely unconstrained by power, but as the reflections and actions chosen by a person or a group within a space often saturated with power relations. As long as these power relations allow for a certain degree of liberty allowing the self to reflect upon and choose a course of action, there is a possibility for the exercise of freedom.

In order to translate this understanding of government into an analysis of the working of hospital accreditation in Denmark, we must address its technical mechanisms (Dean, 1996), in casu: the standards and indicators, and the ways in which these are informing and inciting action at the hospitals. To grasp the kinds of governing these standards and indicators enable, we need to distinguish between standards that set out to guide or even determine a particular kind of clinical practice on the one hand, and the standards that urge the medical staff to formulate and implement their own guidelines on the other. Both elements are found in the Danish accreditation model, but the latter standards are the most common.

Method

The Danish case of hospital accreditation is to some extent unique. While most EU countries have some kind of hospital accreditation system, the majority of these are voluntary (Shaw, 2006; De Walque et al, 2008; Shaw et al, 2010). EU countries that do have compulsory hospital accreditation systems tend to have hospital systems that differ importantly from the Danish one.Footnote 3 The Danish hospital system is similar to most Nordic countries, that is, a largely tax-financed system in which the hospitals are run not by the state, but by the regions. However, neither Norwegian nor Swedish hospitals are subjected to accreditation, though other voluntary modes of quality management systems are in place there. In brief, the Danish case of hospital accreditation differs from most other existing ones. On the one hand, this may limit the potential for generalization. On the other hand, the Danish case of hospital may be a least likely case. At least, the imposition of compulsory accreditation in a system in Denmark was not only facing resistance by a very strong professional interest group, it also took place in a system in which the hospitals are run not by the state but by the regions with a very large degree of autonomy. If it is possible to implement mandatory accreditation here, then it seems fair to assume that it may be possible to implement in other – more conducive – political settings as well.

This article focuses on the contemporary political processes evolving around the making and implementation of DDKM. Here it should be noted that the current study deals with the accreditation of hospitals only, not general practitioners and pharmacies who were subsequently subjected to the same scheme. The analysis of the way in which accreditation has been employed in Danish hospitals is based both on documents and on interviews. More than 200 documents covering the period between 1990 and 2013 were collected. These include laws, regulations, agreements between the Ministry of Health and the counties, reports from the Danish medical authorities, reports from the Danish medical professions, articles in the leading Danish medical journals, newspaper articles and press statements. Laws and regulations were identified via the Danish legal database www.retsinformation.dk. The homepage of the Ministry of Health and the Danish Regions were used to identify relevant reports and programmes. Medical journal and newspaper articles were identified by using the national electronic search base www.Infomedia.dk. Moreover, references in reports and articles were carefully traced and reiterated until no additional relevant documents could be identified.

The interviews were made in two rounds. In order to provide a general insight into the main ways in which accreditation measures and procedures were translated into practice, eight persons from three hospitals were interviewed in October 2012: two quality managers, three doctors and three nurses (Myhr et al, 2013). In order to get a more detailed picture of the ways in which DDKM is actually translated into concrete procedures and routines of action, a second round of interviews was conducted in April 2014 with six other persons from two other hospitals: a head of hospital quality management, two doctors and three nurses (Erichsen and Lund, 2014). All interviews were audio recorded and transcribed. In order to illuminate how local interests and conditions impinged on the implementation of DDKM, a most similar systems design was adopted, namely two intensive care units operating under identical regional guidelines (in Region Zealand). All interviewees were promised anonymity in order to make them talk as freely as possible about the ways in which DDKM was actually adopted. Of course, this does not exclude the possibility that the staff was omitting issues or events that they found undesirable to disclose. In particular, it cannot be excluded that the medical staff may be downplaying the phenomenon of de-coupling, that is, more or less deliberate ignoring of the guidelines emanating from DDKM or its local spin-offs. Yet, the staff did seem rather frank about this. Moreover, the motivation for downplaying this phenomenon seems limited because DDKM allows staff to ignore some of those procedures they find irrelevant and also allow for procedures to be revised. More about this below.

Making the Hospitals Accountable for the Quality of Their Services

On the face of it, it seems improbable that the DDKM’s 82 standards and the 570 indicators pertaining to these standards leave little if any room for professional discretion. Of the 82 standards, 39 are organizational ones, 40 general patient process standards and three are specific sickness standards. The organizational standards are regarded by IKAS as indispensable preconditions for securing the good patient process, that is, high clinical quality treatment of the patient and a high level of patient satisfaction (Institut for Kvalitet og Akkreditering i Sundhedsvæsenet, 2012, p. 212). The general patient process standards include reception, assessment, diagnosis, invasive treatment, medication and observation. The overall aim of these standards is to ensure a coherent patient process not only within the hospital (between its many departments) but also between the hospital and other relevant institutions before and after hospital admission. Finally, hospitals are supposed to select three specific sicknesses of particular relevance to their hospitals and apply the standards specified in the model and develop their own clinically proven guidelines.

Yet, the DDKM model focuses not really on how the hospitals seek to abide by its standards, but rather if the hospitals adopt procedures that account for the ways in which they will deal with the standards and indicators. While DDKM implies that all hospitals are subjected to a mandatory external assessment every three years, IKAS emphasizes that the purpose of this assessment ‘is not that externals [assessors] assess the concrete quality of every form of treatment. DDKM must instead evaluate the hospital’s preconditions to deliver good quality’ (IKAS, 2012, p. 4). Significantly, hospitals are allowed a wide space of discretion not only in their implementation of standards pertaining to organizational matters, but also in clinical matters. While all hospitals must elect three diseases they find are in particular need to ensure quality, it is up to the individual hospital to decide which diseases to focus on. The only requirements are that the diseases occur frequently and involve complicated treatment and rehabilitation challenges (IKAS, 2013, p. 174). Even in vital areas such as patient security and risk management, hospitals must find their own way of implementing these (Institut for Kvalitet og Akkreditering i Sundhedsvæsenet, 2013, pp. 2 and 38). Tellingly, Rigshospitalet, the largest hospital in Denmark, explicates that it does not expect its medical staff to even know the DDKM standards. Instead, they should comply with the guidelines issues that the hospital has developed in order to accommodate the DDKM standards (Rigshospitalet, 2011, app. 8).

It seems then that in spite of DDKM’s overwhelming number of standards and indicators, they do allow the hospitals wide room of manoeuvre. More importantly, DDKM not only allows hospitals to implement standards differently, they urge – if not force – them do so. This encouragement to make hospitals and their staff act is both indirect and direct. It is indirect in the sense that many standards and their respective indicators are defined quite broadly, leaving a quite wide room for interpretation. For example, several of the organizational standards, such as the documentation of quality measures, allow for rather differential practices. It is direct in the sense that some standards explicitly demand that hospital staff develop and apply their own guidelines for the everyday medical service work. More precisely, a set of regional guidelines are developed covering each and every medical specialty found at the hospitals in Region Zealand. These regional guidelines are produced by a large number of quality councils made up by hospital doctors, nurses, physiotherapists and other medical staff. While each hospital is supposed to abide by the regional guidelines, they are also supposed to develop local guidelines reflecting local hospital conditions and needs. This requirement that the hospital develop its own guidelines is particularly clear in the patient process and sickness specific standards, but it is also found to some extent within many of the organizational standards.

In sum, DDKM seeks to govern hospitals and their staff not by substantive codes of conduct determining clinical or organizational practices, but by strongly urging – if not forcing – hospitals and their staff to develop their own guidelines for clinical conduct and procedures, ensuring that the staff adhere to these guidelines. Thus, it is a form of government that is conditioned upon the active self-government of hospitals and their staff. As this leaves a wide local room for manoeuvre at hospitals and their departments, the next section examines how such self-government may unfold.

Local Accountability Practices

Whereas DDKM allows IKAS to govern the hospitals and their staff through the 82 standards and associated indicators, the hospitals and their staff govern themselves through the ongoing development of regional and local guidelines for clinical conduct. Thus, as I will try to demonstrate in the following, the local accountability practices instigated by the DDKM standards hinge on the constant reflections and active participation of hospital department quality managers, doctors and nurses in developing and enacting regional and local guidelines for organizational and not least clinical conduct.

Local accountability practices revolve above all the everyday documentation of compliance (or non-compliance) with existing guidelines. The medical staff interviewed at the two intensive care units all asserted that the major effect of DDKM was that they now had to document how they abided by the regional and local guidelines. In contrast, clinical practice per se, they argued, had changed little if at all. When pressed on the issue of clinical practice, some staff did come up with a few examples of change. Moreover, anecdotal evidence from other Danish hospitals suggests that locally developed guidelines – in the area of gastrointestinal surgery – may bring about significant changes in clinical practice (Kompetencecenter for Epidemiologi og Biostatistik Nord, 2014). However, the most important point here may not be whether or not clinical conduct per se is changed, but that the DDKM standards force the hospitals to document the procedures by which they ensure that the medical staff adhere to contemporary clinical guidelines for intensive care. Thus, while there is a wide room for interpretation and even de-coupling when it comes to clinical practice changes per se, the hospital department may risk not having a positive accreditation if they fail to document their procedures for how they – seemingly – adhere to the clinical guidelines. In brief, the exercise of discretion in the interpretation of clinical guidelines is an option; failing to account for local procedures seeking to ensure compliance with such guidelines is not.

In order to further illustrate how DDKM seeks to spur local accountability practices, we may turn to patient recording. During the 1970s and 1980s, patient records were digitized and further standardized. Notwithstanding these efforts to improve the quality of the patient processing before, during and after hospital admission, DDKM has tried to reduce the significant variations in the mode of registration and the procedures for disseminating patient record information from hospitals to the patient’s general practitioner. For example, one new standard specified the making of the records upon admission of a new patient. Such records are not new, but the new standard specifies a number of information that the hospital has to retrieve from the patient (or his or her general practitioner). Another new standard specifies the making and processing of discharge letters based on the hospital patient records. This is particular pertinent in order to ensure proper follow-up monitoring and treatment by the patient’s general practitioner upon discharge. More generally, the medical staff is urged to use their professional judgement in filling out some fields in the template on patient flow while ignoring others that they find irrelevant.

The requirements of constantly producing and giving accounts, such as patient record maintenance, receive mixed reactions. On the one hand, many of the doctors and nurses interviewed also explained that the new guidelines and the templates that have to be filled out with patient flow are a relief in that they often simplify their work. They do not have to invent anew what information to convey every time they make a new letter. On the other hand, most of the interviewed doctors and some of the nurses lamented that a number of the guidelines emanating indirectly from DDKM did have clinical relevance, and were not conducive to enhancing the clinical quality of their services. Accordingly, they complained that they often had to spend time on what they find are absurd measures. For example, they must assess the risk of each and every patient of falling and hurting herself, a problem usually only seen among quite old patients. A chief nurse at intensive department complains over a guideline requiring that all patients are checked for the risk of contracting pressure ulcers. The nurse readily admits that when emergency patients in a critical condition are received this guideline is usually ignored. A chief doctor in the same department laments a guideline requiring her to provide data to the region for a disease (ventilator-associated pneumonia) for which there is no clear diagnosis. These and several other stories of irrelevant or inoperable guidelines were mentioned by all interviewed medical staff. More than half of these explained that they had ignored such guidelines on several occasions. Notwithstanding these common instances of non-compliance, both doctors and nurses claimed that they followed most of the guidelines on patient flow recording most of the time.

Apart from the guidelines aimed to ensure better patient flows by systematic recording procedures, the DDKM standards oblige medical staff to use ICT systems with updated information on organizational and clinical guidelines for treating various diseases. Several doctors explained that despite the presence of ICT systems with updated information on clinical guidelines, they often use handwritten papers and personal diaries when in doubt about clinical guidelines. The DDKM standards explicitly seek to eradicate this practice because of the risk of adhering to obsolete guidelines. An electronic document management system, D4, has been adopted by all the hospitals in Region Zealand in order to provide standardized and updated information on organizational and clinical guidelines for specific diseases and medical events. The D4 system has received mixed reactions. On the one hand, the interviewees all complain that the D4 system in several areas is not logically structured. Accordingly, some staff tend to stick to the use of printouts from the D4 system or on handwritten guidelines based on information from meetings or emails disseminated by colleagues. On other hand, the very same doctors also report that the system is widely used in everyday work to ensure that the medical services conform to the most recent guidelines. Some doctors even explain that D4 has made staff more aware about the utility of guidelines and has started developing new local guidelines in areas where these are seen to be lacking.

It seems then that the local accountability practices propagated by DDKM are not primarily about ensuring compliance with particular standards or guidelines, though this does play a role, but more about making local hospital assess their use of such guidelines and develop these when needed. In both the intensive care units studied, the suitability of the guidelines was discussed regularly (every month in one unit, and twice a year in the other). Regional guidelines that had been consistently criticized by local medical staff were reviewed and in many cases revised, that is, new, more suitable local guidelines were developed by the medical staff itself. Moreover, both units, though with different levels of rigour, run their own control processes to ensure that the new local guidelines were actually met in everyday practice.

Interestingly, three of the interviewed doctors and nurses seemed unaware that the guidelines they must comply with are developed not by IKAS, but by either their own hospital (in the case of local guidelines) or by colleagues in the region (in the case of regional guidelines). This ignorance may induce hostility against DDKM and increase non-compliance with the guidelines. Accordingly, IKAS does not spare any occasion to make the hospital staff aware that IKAS is responsible for developing national standards only and has no influence on the local clinical guidelines. One of the interviewed intensive care chief doctors was infuriated by a particular guideline on patient nutrition and wrote a complaint to IKAS. They responded that it was allowed to deviate from the guidelines if sound arguments could be provided, and added that he should address the regional quality council, made up by his colleagues, as they were in charge of developing and updating the regional guidelines on nutrition. Of course, this did not make the doctor find the guidelines any better, nor could they be changed immediately. However, once he understood that the guidelines were made by his colleagues in the region, rather than some bureaucrats in IKAS (his terms), his critique was redirected from what he saw as ‘the system’ (IKAS and DDKM) to his own profession. Thus, IKAS quite deliberately seeks to govern at a distance – by way of the DDKM standards – and lets hospitals and staff govern themselves – by way of making them develop their own guidelines for organizational and clinical practices.

Discussion

In the absence of coercive means of power, what makes the doctors and nurses utilize – albeit grudgingly and at times partially – the guidelines emanating indirectly from DDKM? We may start by noting that the nurses generally seemed much more positive towards the systematic use of clinical guidelines than the doctors did. In line with the rational choice perspective, this could be attributed to nurses using the guidelines to enhance their prospects of career progress. There are several indications rendering such an explanation plausible: many of the interviewed staff responsible for the local implementation of DDKM-induced guidelines were actually nurses, the nurses spend more time attending meetings to discuss how the standards may be translated into local guidelines in the most practical manner, and one of the (doctoral) department heads explained that some of the guidelines are deliberately used by the nurses to address doctors and make sure their conduct adhere to the guidelines. It seems then that DDKM is used – deliberately or not – by many nurses to install a more equitable relationship between themselves and the doctors. More generally, by taking up the responsibility of ensuring that the standards are adopted in everyday health-care service, the nurses may be carving out a niche of expertise and authority that may secure the status of their profession (see also Timmermans and Berg, 2003).

If rational choice theory may go some way in grasping why many nurses support the guidelines emanating indirectly from DDKM, this theory seems of little if any use in helping us to grasp why doctors should take them on. As already indicated above, many doctors are quite sceptical of the clinical utility of many of the guidelines, in particular the organizational ones. Also, the doctors willingly admit that ignoring standards is not an uncommon phenomenon. Thus de-coupling is taking place, though the extent of non-compliance with the guidelines is difficult to assess. Nevertheless, the general picture is that doctors do actively use most of the guidelines most of the time in their everyday work. At least two reasons may be suggested for this apparent widespread utilization of the guidelines. First, some doctors have gradually changed their mind about the utility of non-clinical standards. They seem increasingly to acknowledge that rigid procedures of hygiene and proper patient record making may have an important positive effect on clinical outcomes. This could be interpreted as the result of a new institutional logic of appropriateness gradually making its way into hospitals. However, a second reason for the doctors’ general adoption of the guidelines may have less to do with a change of mind about the appropriateness of the (organizational) guidelines and more to do with a key political-cum-technical feature of the guidelines. With important exceptions, the guidelines are developed either by the local hospital department (that is, by the staff itself) or by colleagues in the regional quality council. Accordingly, the hospital staff itself is actively contributing to their own governing in the sense that they are forced to participate actively in the development of regional and local guidelines, which they find suit their needs.

In brief, both rational choice and institutional theories are to some extent useful in grasping local accountability practices and the staff’s reaction to DDKM. However, it seems that the notion of government brings us a step further. Most of the procedural standards in DDKM amount to a form of power that works by latching on to and facilitate a particular form of freedom of those over whom or, more precisely, through which that power is exercised. This procedural form of power is akin to what Mitchell Dean and others have dubbed the governmentalization of government (Dean, 1999, pp. 193–194; Triantafillou, 2012). Doctors and nurses are not told how to carry out their medical work, but urged to act and take up the role of active and responsible civil servants constantly reflecting on how their work can be improved. Of course the freedom that doctors and nurses are encouraged to exercise is not just any freedom. On the one hand, it is a freedom that must be exercised in line with the professional norms of medical conduct largely established by the medical profession itself. On the other, it is a freedom that must make itself visible to the outside world, notably to the public health authorities, through extensive documentation of the ways in which the guidelines developed and chosen by the hospital itself are adhered to.

Conclusion

This article has tried to shed some light on two questions. First, why does the hospital staff seem to accept, albeit grudgingly and partially, DDKM? Second, how does DDKM contribute to rendering Danish hospital activities accountable? The article has tried to show, first, that DDKM contributes to render Danish hospital activities accountable through a vast set of standards and indicators. Significantly, the bulk of these standards are not strictly clinical ones, but mainly organizational standards. Moreover, the standards used are mainly meta-procedural standards that do not really specify procedures, but require hospitals to formulate their own procedures. This leaves a wide room for local discretion. However, what this room for discretion does not allow is that medical staff abstains from developing and implementing their own guidelines. Thus, they are not only obliged to be active and develop their own guidelines, they must also document their everyday clinical work in a way that allows both (internal) colleagues and (external) quality assessors from IKAS to check that they actually adhere to their own guidelines.

Second, it has been argued that the hospital staff’s acceptance of DDKM, though grudging and partial, may be explained – at least partly – by the meta-procedural character of the standards. As argued above, they not only leave substantial room for local discretion, they also urge medical staff to act in ways the staff itself finds suitable as long as they develop, explicate and monitor guidelines shaping these actions. Analytically speaking, DDKM may be characterized as a technology of what Foucault terms government, that is, a form of power that more or less systematically works through the facilitated and structured freedom of the subjects of power. This kind of freedom has clearly not fully eliminated criticism from the medical profession, but the procedural standards constituting the core of this freedom do seem to have tamed the resistance against DDKM.