Public value accounting and the use of performance measurements as a management tool in a context of various assessments

Abstract

Purpose

The purpose of this paper is to conceptualise our understanding of public value accounting (PVA) by studying the use and usefulness of performance measurements (PM) as a management tool. The authors do this from a perspective in which they address the complexity of various (sometimes conflicting) assessments of performance measurement and management (PMM) by different stakeholders.

Design/methodology/approach

An interpretative case study using qualitative methods. The paper is based on 30 interviews conducted in 2018 and 2019 with respondents working with PMM at different levels, such as politicians, officials and health-care professionals. The study context was Region Stockholm (RS) in Sweden and its health-care division.

Findings

PMs become an instrumental tool for PMM, which led to output being promoted above outcome. The authors show that there is a conceptual shortcoming in the discussion of PVA, as the effort needed to achieve outcome-based information might exceed the ability of an organisation to deliver it. The authors address the importance of studying the interaction among different stakeholders, including politicians, the public and media, in research on PVA, as well as possible power relationships among stakeholders.

Originality/value

The authors contribute to the growing research on PVA and its call for more empirical research by offering a more nuanced interpretation of PVA activities. The authors do this by studying PMM and the nature of these activities in a public sector organisation from a multiple-stakeholder perspective.

Citation

Höglund, L., Mårtensson, M. and Nylinder, P. (2023), "Public value accounting and the use of performance measurements as a management tool in a context of various assessments", Journal of Accounting & Organizational Change, Vol. 19 No. 4, pp. 603-621. https://doi.org/10.1108/JAOC-12-2021-0186

Introduction

Recently there has been enhanced interest in public value accounting (PVA), resulting in the development of PVA into a distinct research field that we need to pay attention to (Bracci et al., 2014, 2019; Höglund et al., 2021b; Steccolini, 2019). PVA takes a special interest in strategic matters of public value creation and its relationship to accounting processes of performance measurement and management (PMM) (Höglund et al., 2021b). Research on PVA has been limited so far (Steccolini, 2019) and there is a call for empirical studies (Bracci et al., 2019; Salemans and Budding, 2022). We find this surprising, as research has shown for some time the importance of strategically envisioning long-term value creation in relation to performance measurements (PM; Moore, 1994, 1995; Spano, 2009). In line with this, we intend to contribute to the discussion on PVA by addressing previous research on PMM and the use of PM as a management tool. In this work, we view PM as a management tool for “measuring of output, outcome, efficiency, effectiveness and equity at various levels of organisations” (Siverbo et al., 2019, p. 1801).

Within PVA it is important to distinguish between PMM of output and outcome (Benington and Moore, 2011). To deliver public value, organisations need to focus on strategically managing their outcomes and the long-term perspective (Talbot, 2008). Yet this is precisely what has been proven to be difficult in daily practice and, therefore, output tends to be favoured over outcome (Höglund et al., 2018; 2021b). This is often described as problematic, as the use of PM as a tool for measuring output does not provide any insight into the delivered public value (Bryson et al., 2014), and research indicates that using specific output indicators may even be counterproductive (Kallio et al., 2021). Moreover, to achieve specific outcomes it is important to organise stakeholder contact to clarify the strategic goals before they are operationalised (Moore, 1995; Höglund et al., 2021b). While strategic goals seem to express public value thinking, we do not know the extent to which stakeholders are involved in defining them (Bryson et al., 2014). In line with Salemans and Budding (2022), we argue that the information PVA delivers will be enriched by such feedback. As Curuana et al. (2023) argue, assessing the influence of PM based on different stakeholders and their decision-making process would make interesting studies, as would investigating the effects of PM in relation to how problems of public value are framed and how decisions are taken. In other words, we intend to address such important questions as how concepts of value are determined in contexts where conflicting interests and political agendas are in play (Steccolini, 2019).

Petit and Lux (2020) highlight the fact that there might be a collision between different stakeholders (e.g. management and politicians) regarding the use and usefulness of PMM. Likewise, Pollitt (2018) addresses how PMM assessment is perceived differently by different stakeholders. In other words, different stakeholders have different perceptions of which measures are relevant to assessing the desired performance. In this context, Lewis (2015) states that, when it comes to public sector organisations (PSOs), the substantial question of PMM is not technical but political. That is, who decides on PMs and how they are linked to the organisation’s structure, management and governance. This results in a PSO’s performance having to be negotiated among various stakeholders (Dobija et al., 2019). In other words, PSOs have many different stakeholders who sometimes have different ways of assessing and understanding value performance and, thus, advocate different PMs.

The purpose of this paper is to conceptualise an understanding of PVA by studying the use and usefulness of PM as a management tool from a perspective in which we address the complexity of various (sometimes conflicting) assessments of PMM by different stakeholders. To achieve this aim, we approached Region Stockholm (RS) in Sweden and its health-care division. In 2017, RS began large-scale strategic changes involving the implementation of a new PMM model. We used an interpretative case study approach, including 30 interviews conducted in 2019 with respondents working with PMM at different levels, including politicians, officials and health care professionals. In this way, we contribute to the growing research on PVA and its call for more empirical research by offering a more nuanced interpretation of PVA activities through the study of PMM and the nature of these activities in a PSO from a multiple-stakeholder perspective.

A case study

We draw upon Stake (1995), interpreting the case of the health-care division at RS to conceptualise an understanding of PVA. Flyvbjerg (2006) argues that when the aim is to generate as much information as possible on a given problem, an atypical or extreme case often reveals the best information. Based on this, we argue that the health-care division in RS is interesting to study in relation to PVA as it has been the subject of several strategic change initiatives aimed at achieving more efficient PMM using PM as a tool for management. We were also able to include three common types of stakeholders (officials, professionals and politicians) in the context PSOs.

Data collection

The empirics in this paper are built on a larger research project, held from 2017 to 2019 in RS and including a total of 52 interviews. To these interviews, we added various sources, such as meetings, strategy documents, budgets, operational plans, reports, newspapers, websites and internal PowerPoint presentations. The documents alone ran to over 2,400 pages. In this paper, the document study served as a guide to the context of RS, the health-care division and its health providers. It also allowed us to study the division’s formalised management control practices, the intents of PMM, how PMs are used for performance management purposes and what kinds of PMs were included in the new PMM model. The larger research project addressed RS as a whole – all four divisions of health care, public transport, culture and regional development. It focused on issues related to theoretical perspectives within management control, strategic management in the public sector, public value management, PVA and PMM.

In 2020, two Swedish reports on the project were published. Results from the first report showed that long-term strategic work and public value were challenged by standardised management control, a focus on quantitative PMs and detailed political governance. In the second report, the focus was on RS and the health-care division from a management-by-objectives (MBO) perspective. This showed that different stakeholders – officials, professionals and politicians – had conflicting ideas and assessments of PVA, PMM and the difficulties of implementing PM as a management tool among professionals. We found those results interesting and wanted to dig deeper. As Lewis (2015) states, when studying PMM in practice one must consider the entire decision chain including politicians, officials and representatives of the profession.

For this paper, we have included a total of 30 interviews conducted between November 2018 and November 2019, including 33 respondents working with health-care-related issues. Nine of the interviewees (politicians and officials) worked at the RS headquarters, 11 were posted at the health-care division (politicians and officials), 12 worked at the health-care service provider level and one had another health-care-related position (Table 1). At RS, the politicians are organised (separately from the officials and professionals) into different political committees at both RS headquarters and the health-care division. The interviewees categorised as “professionals” are medical professionals, such as doctors and nurses, working as managers. The “officials” include senior managers or those working with management control, strategy or the health-care budget. The interviews lasted between 43 and 114 min, were transcribed and can be described as more of a conversation between the interviewer and the interviewee. They covered discussions about public value, management control, strategy, the political governance of the region and the health-care division, the new PMM model, and the use and usefulness of PM as a management tool.

Data analysis

We began the analysis in this paper from a position of having written two reports on RS. The results of these reports were based on an iterative research strategy (similar to Modell, 2009) and the discussion on different levels of triangulation and abduction. We analysed the documents and interviews at various times and from different theoretical angles, discussing the results on several occasions with stakeholders from the headquarters, the health-care division and the operational level. This made it clear that opinions on PMM and the way measurements were assessed were quite different among the various stakeholders. For example, for some stakeholders, using PM for management purposes contributed to better performance, while others stated the opposite. Another interesting result was that all the respondents (directly or indirectly) made some form of statement on quality and public value in relation to PMM without any prompting from us. We found these two results interesting and decided to return to the 30 interviews we held about the health-care division. The results of this work are presented in this paper.

In qualitative research, there is always a trade-off between showing the rich data on which findings are based and the constraints of an academic article (Höglund et al., 2021a). Based on this, we did two levels of analysis: Firstly, we focused on what was common to the entire corpus of data. Hence, when analysing the interviews again (to understand the use and usefulness of PM as a management tool from the perspective of PVA and different stakeholders), we identified three common narratives concerning:

1. quantitative measurements;

2. compatibility for benchmarking; and

3. negotiation of performance.

Based on this, we constructed a theoretical frame of reference and presented the results of our empirical analysis in relation to these main findings. It is important to state that the notion of quality in this paper should be viewed as an empirical finding and not a theoretical concept.

Secondly, we focused on variations in the statements within the three identified narratives. This means that the interview extracts were selected to show variation in the assessment of PMM relating to the three common themes in the data corpus.

Theoretical frame of reference

Quantitative measurements

In previous research on PMM, as well as within health care, there is a tendency to highlight quantity before quality in determining performance, as there seems to be an inherent “logic of escalation” (Pollitt, 2018). Huston et al. (2022) state that, when focusing on quantified measures and their instrumental use, other important aspects and their non-instrumental use (e.g. political, rhetorical, symbolic, opportunistic, ceremonial and passive use), which are often more difficult to measure or even unquantifiable, are neglected. Perrin (1998) addresses consequences in terms of “flaws and limitations”. One such flaw is the varying interpretations of performance models, i.e. similar terms and concepts are interpreted differently. This leads to what Smith (1995) refers to as misinterpretation, i.e. that data is interpreted incorrectly and at times even deliberately so. Another flaw, according to Perrin (1998), is labelled “goal displacement,” which leads to creaming. One example of creaming is when doctors choose easily treated patients instead of those with multiple diagnoses, i.e. more difficult to treat, as a measure to fill numbers instead of improving patients’ health. This is also addressed by van Thiel and Leeuw (2002), who state that creaming might limit the focus of a given programme and not improve the actual outcome, as work is devoted to what is easy to measure and achieve. This creaming effect results in PMs of perceived numbers and volumes (Bouckaert and Balk, 1991).

In summary, the literature shows that measures focused on output outcompete the discussion of outcomes (Rajala et al., 2018; Pollitt, 2018) and public value (Bryson et al., 2014). Hence, PMM practices focus on PM, measuring efficiency rather than effectiveness and output rather than outcomes (Pollanen et al., 2017), and, hence, quantity rather than quality (Pollitt, 2018). Despite this, most public managers are expected to promote outcome, while at the same time they are accountable for output towards the politicians and society (Norman, 2007).

Comparability for benchmarking

When it comes to the use of PM for improvements, it seems that a choice must be made between using the PM to develop and improve the local operations or using it for comparability on a macro-level (Johnsson et al., 2021; Zidarov et al., 2017). For example, Johnsson et al. (2021) showed that models of working with quality and continuous improvements in performance resulted in improvements on a local level, but when used for benchmarking between organisations, the results instead emphasised the lack of flexibility in the models and difficulties in using them for improvements. In line with this, Lewis (2015) concludes that PMs are mainly used at the meso-level, by managers, and not by politicians at the macro-level nor at the micro-level of operations. Mutiganda (2016) in turn states there are no direct relationships between the macro/meso-levels (where the budgets are set) and the micro-level (the actual medical practices).

Zidarov et al. (2017) address a similar problem with comparability for benchmarking. In their study, it was reported on a macro-level that the PMs for benchmarking purposes worked and were perceived as useful and valid, but became too general and difficult to use on a micro-level. This indicates that comparing measures for the same activities can be used for benchmarking purposes, but when the perspective is moved to the organisational level, measures may lose relevance for the operative level (Zidarov et al., 2017; Johnsson et al., 2021).

In summary, an ongoing discussion portrays PMM information as either output-based, i.e. information that is simple, comparable and efficient to gather (Kurunmäki et al., 2016), or outcome-based, focusing on creating and strengthening public value and how outcomes are experienced by the public (Norman, 2007). However, empirical research shows that most health-care practitioners tend to take an output focus, where comparability for benchmarking becomes an important aspect of PMM.

Negotiation of performance

PSOs are often large, complex and diverse, involving many stakeholders, such as politicians, taxpayers and citizens. Thus, it is important to identify all stakeholders to better understand each stakeholder’s wishes regarding performance. According to Freeman (2010), this means that stakeholders must be involved in both planning and management of the PMM system.

Audette-Chapdelaine (2016) argues that politicians should be included in PVA and PMM work with the public managers. This is because PMM in PSOs aims at creating public values that in turn require long-term planning. But, as politicians are elected for shorter periods, they are not always strategic (Liguori and Steccolini, 2018), which may give rise to tensions between the different timeframes. Similarly, Lewis (2015) addresses the importance of focusing on the politics and its consequences for PMM, as there are no clear lines in practice. Politicians and public managers tend to interact in relation to PMM (cf. Audette-Chapdelaine, 2016; Rajala et al., 2018), both formally and informally, and their involvement shifts depending on the question at hand. At times the politicians are more active and steering, at other times the public managers and officials are the ones pushing a specific matter forward (Brorström and Norbäck, 2020).

In summary, previous research has addressed that performance has to be negotiated between various and many different stakeholders in a context of PSOs, and must, therefore, be taken into consideration when studying PVA and PMM (Dobija et al., 2019). However, Pollitt (2018) emphasises that it is unrealistic to meet all stakeholders’ wishes, and therefore it is important to decide who the PMM systems are designed for, and that relevant PMs are selected based on this choice.

Analysis of the case

RS is one of Sweden’s largest employers, with a yearly turnover of EUR 10bn. RS is mainly financed by taxes and led by an elected political assembly, its highest decision-making body, which appoints an executive committee. The assembly decides on taxation, budget, strategies and objectives for RS. The executive committee directs, coordinates and controls the administration of RS, which means that it has the overall responsibility for the region’s performance management. A new PMM model was introduced in 2017 and is operationalised through the yearly budget document.

The Swedish model of government administration builds on three levels of governance: the national level (which has the legislative power), the regional level (21 regions) and the local authority level (290 local authorities that each have an elected political assembly). The 21 regions are self-governing regional authorities with a great deal of autonomy, including the power of taxation and the freedom to make their own decisions. All regions in Sweden have three obligatory tasks such as health care, public transport and regional development. In addition, most regions also undertake the optional responsibility for arts and culture. This means that RS is divided into four divisions where the health-care division – which is the focus of this study – is by far the largest, with approximately 41,000 employees, and is one of Europe’s largest health-care providers.

In the health-care division, operations are carried out by RS itself, by limited companies owned by RS, and by external contractors using a procurement model. The health-care division, as well as the others, report on their performance to RS based on the budget, which sets overall objectives, sub-goals and performance indicators. The health-care division differs from the other three divisions in that it also works with PMs in the contractual agreements where money is transferred from the division to the health-care service providers.

Quantitative measurements

When addressing PM in terms of quantitative measurements, the interviewees commonly did so in conjunction with an assessment of quality and public value. They made several statements consistent with what previous research had acknowledged; that using PM as a management tool tends to favour quantity over quality. In other words, it favours what can be measured in terms of volume (Bouckaert and Balk, 1991) versus what is easily measured (Perrin, 1998; Smith, 1995; van Thiel and Leeuw, 2002). A professional explained:

We get volume estimates of how many in-patient and out-patient care episodes, how many are emergencies and how many are elective. […] DRG [diagnosis-related groups] is what we’re measured and monitored by. […] 10% of our pay is linked to quality. (R31)

The quality indicator is measured quantitatively. A manager stated:

Everything is regulated based on produced DRG volumes […]. The current contract consists of 3% target-related pay, linked to about 40 quality parameters, where we have deductions if we don’t achieve them. […] That’s linked to volume-based pay. (R17)

In relation to actively meeting their goals, the hospitals were punished, when it came to payment, if they did not deliver on the volume-based quality measurements. As explained by one professional:

Quality penalties, when we don’t reach those levels we lose pay. […] We lost maybe 10 million [SEK] in quality penalties last year. (R27)

What we see is a focus on measuring efficiency rather than effectiveness, output rather than outcomes (cf. Pollanen et al., 2017; Rajala et al., 2018) and, hence, quantity rather than quality (Pollitt, 2018). Thus, in health-care contracts, the economic incentive to deliver on volume prevails. A professional said:

A hospital loses resources if we don’t live up to the healthcare agreement. (R38)

Similarly, an official pointed out there will be a:

Financial sanction if we do not achieve our targets. (R32)

In other words, the health-care division depends upon the four-year contracts that set the financial framework of each health-care provider. An official noted that the politicians at the time forgot that:

The financial conditions […] are set out in the contracts. […] People think [the politicians] regulate a lot through the budget, but they don’t, because most of the financial conditions are regulated in contracts. […] politicians assess things a year at a time, but they want the hospitals to look at things with several years’ perspective, but that’s difficult to do because the politicians only make decisions with one year’s perspective. (R33)

A manager said:

We have four-year hospital contracts, […] then we get a new budget every year […] a product of politics, it’s the ideology that controls it. (R31)

In other words, the officials and professionals do not always agree with the politicians about the usefulness of having budgets with PMs that are set by politicians. This is because health-care providers are financially dependent on the contracts.

In our case, the new PMM model in the budget became something the professionals only reported to the politicians and to officials at the central administration level, and not anything they included in the daily operations. Officials at the health-care department tried to resolve some of these issues by incorporating goals and indicators from the budget into the contractual agreements that are tied to the financial resources (cf. Speklé and Verbeeten, 2014). One challenge, however, was that the contractual agreements have a time frame of four years, whereas the PMs in the budget are reported once a year. This means that both officials and professionals have to work in parallel systems to produce the performance information the politicians demand. Often, they just report the performance information to politicians in the parallel system, and then continue to work with what they consider to be important, i.e. the performance indicators followed up by the contractual agreements. This has a direct consequence for performance management, as the politicians do not get sufficient performance information (cf. Mutiganda, 2016; Jethon and Reichard, 2022; Raudla, 2022), but officials and professionals in the health-care department also become less efficient in their work, as they must produce PMs that are not relevant for the health-care division.

In line with Huston et al. (2022), we see how PMs become an instrumental tool for performance management at RS, leaving out non-instrumental uses related to such things as quality and public value. Smith (1995) refers to this as “tunnel vision”, when a PSO focuses too much on quantified measures and points out that this often comes with unintended consequences. For example, a lot of time and resources are devoted by officials and professionals to interpreting what politicians might mean when setting political visions, strategy and overall objectives in the budget. This, therefore, has to be operationalised into performance indicators that make sense to health-care providers. One of the officials gave a concrete example, referring to the goal that “no-one should have to die alone in our hospitals” and asked “how do we measure that?”:

‘Not having to’ is quite easy to understand, but what is being referred to? Do the politicians mean that no one should die [alone], is it the moment of death […] that we’re monitoring in the quality register, or the process up to death? If we were to guarantee that no-one should take their last breath alone, well then you’d need an incredibly large monitoring organisation. (R35)

The officials and professionals report on the undesired consequences of the new PMM model in terms of too many parameters, goals and performance indicators being used (cf. Bouckaert and Balk, 1991), a fact which has evolved into micromanagement by politicians (Höglund et al., 2021b). In relation to this, one official said:

We’re going to have a joint performance management system in RS […] These operations are super-complex. We have a huge number of goals from the political level. Those need to be broken down [in performance indicators] and allocated to the different divisions and their operations. (R2)

The number of measurements and assignments from the politicians is sometimes seen as overwhelming, resulting in quite heavy criticism of these metrics by health-care providers. An official who works with the metrics says:

Heaps of assignments rain down on us. I wish they [the politicians] would have a dialogue with us [officials] beforehand, because some of them can’t be done in a year, some aren’t appropriate, sometimes not even legal to do. I think it would have been valuable to check with the officials before making assignments. (R29)

Regarding quality and the difficulties of measuring it, a manager said:

The quality concept contains several things. There’s our availability – do we have queues, how long do you have to wait in A&E – to how we manage hygiene. […] There are far too many parameters. (R31)

One of the professionals stated:

We’ve never had this many targets from the council [politicians] before. This is completely unique. I counted 60 objectives […] Usually there are 3-4 of them. (R27)

Various researchers state that politicians only make limited use of PMs for decision-making. In part, this is because of the low-quality data they yield (cf. Mutiganda, 2016; Jethon and Reichard, 2022; Raudla, 2022). At RS the politicians are the ones demanding output-based results which use performance indicators.

The professionals reported they have different ways of handling the large number of goals in the budget when they have to manage their operations. Some handled the large number of goals by not following the recommendation of breaking them down into indicators for their operations. Others selected indicators relevant to their own operations and what generates quality for the patient, while ignoring others. A professional said:

There winds up being too many indicators. It’s hard to communicate; […] we have to pick what’s important. We have a responsibility to look at those that have major significance to the patients. (R18)

In general, the professionals criticised the new PMM model and its usefulness as contributing to micromanagement, where goals and indicators are too detailed. One professional said:

I think they [the politicians] micromanage quite a bit […] How many assignments do we have, 130? […] All these assignments from the council, I think they’re quite detailed. (R20)

One of the politicians confirmed and explained:

We have more indicators in the budget than we’ve had before […] the politicians are present on a regional level […] that’s the answer to why narrow or micro issues become highly visible at a high level in our governance. (R39)

Another politician acknowledged the problem in the following way:

From a regional perspective, it would be like running down to your local clinic and telling them how to do things. And you know you shouldn’t do that and it won’t be optimal if you do. But sometimes you just get incredibly frustrated, and I think that not all politicians are capable of stopping themselves or reflecting on their actions. (R28)

However, others believe that the administration officials have (too) much power when it comes to the goals and metrics. Regarding this, a professional said:

The more micromanaged a system you have, the more power the managers have. But they’ll never admit it if you ask them. No, it’s the politicians who have set these objectives. But take this extremely complicated payment system; of course not a single politician has the skill or the energy to get into the details of them, so it’s very much a product of the managers’ powers. Then they say it’s the politicians who decided – sure, in the formal sense, but who’s behind it? (R36)

Comparability for benchmarking

It was decided at the political level that RS should work with a new PMM model, aiming at more standardised work procedures. The main idea is that the goals and indicators should support health-care in the region with a more long-term direction supporting public values (cf. Bryson et al., 2014). However, the PMM information tends to be output-based, i.e. information that is simple, comparable and efficient to gather (Kurunmäki et al., 2016). Like Mutiganda (2016), our results show no direct relationships between the macro-level (where budgets are set) and the micro-level (actual medical practices). There has been much criticism by professionals, saying that the RS level is too far removed from the business operations and that the objectives are drafted too generally to fit the health-care division (cf. Zidarov et al., 2017). Not only must the performance indicators be precise and clear, meeting their obligations regarding the overall budget framework of RS; they are also to be used as a metric to compare activities between units and departments. The health-care service providers must be able to steer their activities in a direction that leads to goal fulfilment and they must be able to compare their performance indicators with each other through benchmarking activities (Johnsson et al., 2021). A politician said:

You should also be able to compare, so that you can see each other’s [performance]. […] there are more hospitals working with the same things. What are their goals? What is their goal fulfilment? And how come they succeed […] in something we cannot handle? In this way we can build a learning organisation by using the performance management system. (R25)

This quote highlights the importance of comparability between hospitals as an attempt to enhance quality by connecting performance management of quality to standardised performance indicators in the RS budget. For example, the budget has one goal concerning safe care that is to be operationalised by the following indicators:

• The quantity of nosocomial (health care-associated) infections.

• Readmission to inpatient care within seven days (for those aged 80+).

• Antibiotic prescriptions (number per 1,000 inhabitants).

The new PMM model set general objectives for the whole region. These are addressed to all political committees in the form of main objectives, sub-goals, indicators and target values. There are also specific objectives aimed directly at the operational health-care organisations, in other words, individual hospitals. Examples of objectives aimed directly at hospitals are getting health care at the right time. One indicator of this is the requirement to ensure that patients stay no longer than 4 h in the emergency department (there is no target value). Another indicator is that the patient should be given an initial appointment with a specialist doctor within 30 days (there is no target value). It is noteworthy that, even in 2019, most indicators had no set target value.

Using standardised indicators, the politicians argue, will enhance the quality and public value of health care by using the results from the performance indicators to improve “low-performance” hospitals. However, for performance indicators to be used for improvements, previous research has shown that a choice must be made either between using PMs to develop and improve one’s own local operations or using them for comparability purposes (Johnsson et al., 2021; Zidarov et al., 2017). In our case, they tended to choose comparability. For example, performance is tied to financing, which is mainly based on the quantity – volume – of patient contacts. In such cases, where there are internal improvements that decrease the number of patient contacts, the hospital loses its financial compensation. A professional stated that having contractual agreements tied to financial results makes them bring the patient back for further appointments. This increases their volume, yielding better finances and better performance against other hospitals:

We have a requirement to take in patients at clinics, […] but we don’t have the beds. So we can’t take extra patients at clinics. It’s unreasonable because we won’t be able to manage them. […] the easiest way to handle the situation is to have unnecessary check-ups. (R38)

The officials and professionals all described the usefulness of indicators in terms of the side effects of the PMs and financial system on patients and whether they were getting the right level of health care. The professionals said they had to select the more easily treated patients (“creaming” cf. Perrin, 1998; van Thiel and Leeuw, 2002). This resulted in PMs relating to perceived numbers and volumes (Bouckaert and Balk, 1991). One of the officials, a medical professional, said:

When you calculate the price in Swedish kronor [SEK] per unit of time, patients who have hour-long appointments are not very profitable. This means that there is an incentive to turn away patients you don’t want, i.e. refer them somewhere else. (R37)

The politicians and some officials at the RS headquarters reported that they had high expectations that the new PMM model would make RS more efficient and enhance the quality of health care, while at the same time admitting that it is hard to formulate goals and relevant indicators for measuring these goals. However, so far the quality of the work in the hospitals have not been improved. As one of the officials said:

There haven’t been any improvements in quality like we thought, and we haven’t changed the content as we thought either. Instead, all of our hospitals have fairly significant deficits, and they are not even producing 95% of what they promised in total. (R26)

In general, officials and professionals stated that the PMM model lacks incentives for quality improvement and delivering public value. Some professionals were self-critical and understood that politicians wanted hospitals to do something about the quality of care, given their repeated failure to improve. As one of them explained:

Year after year it’s bad. We don’t wash our hands […] we have nosocomial infections […] I understand that people eventually get tired of us in the profession […] we have a huge amount of unnecessary mortality that wouldn’t be accepted anywhere else […] So much is just measure, measure, measure, measure, measure. We’ve been doing it for an awful lot of years and we can’t get rid of [the problems]. […] but if we don’t feed back [the results of the measurements] […] in a discussion of what that means, you don’t get any effect […] why do we still have 8% of patients getting infections in our department […], but we can’t just look at them, we also have to […] do something about it. (R27)

At the same time, several of the professionals are also critical of the fact that objectives, sub-goals and indicators are too general, as they are supposed to be comparable and used in all units providing health-care service, regardless of the characteristics of the specific operation. There are, of course, major differences between an emergency hospital and a health-care centre, but also between different departments of the same hospital (cf. Kurunmäki et al., 2003). Despite this, all health-care service operations are supposed to strive towards fulfilling the same goals and indicators. In primary care, for example, both the professionals and officials described the problem of having the same level for an indicator, regardless of the type of health care provided. One such example is the indicator prevalence of health care-associated infections. This indicator has a target value of ≤7.9% and one respondent reflected on the consequences:

It’s supposed to be less than 10 % […] For us, 10 % is through the roof, incredibly high. But if you imagine a hospital with an intensive care unit […] it’s fairly reasonable […] We think it should be less than 3 [per cent]. If we were to go out and say less than 10 %, they’ll think, ‘Aha, they want it to be dirtier here’. (R24)

The quotation shows an unintended consequence of the new PMM model where hospitals’ units are compared through performance indicators. The indicator does not support any quality enhancement of the primary care unit at all; rather, it gives them “permission” to have more infections and still be considered a good performer.

The objectives, sub-goals and indicators are described by most professionals as political products mostly used as tools for monitoring the health-care service providers’ operations. This means that several goals and indicators are used only for reporting purposes and their relationship to the contractual agreements are weak and at times even non-existent. However, as the contractual agreements are related to the financial resources, which the goals in the budget are not, the professionals often question the goals formulated via the budget and seldom regard them as relevant, as they do not add to the quality of their work with their patients. One such case related to the education and training of specialist surgeons. On this subject, a professional said that they could not, or would not, fully adapt to the overall assignments, the new PMM model or the budget, due to there being a greater responsibility for long-term public value:

I know exactly what our assignment is, what we are and are not to operate on […]. But we have a major responsibility to train surgeons, because there’s no one else doing it. […] We’re training surgeons for future needs for the entire country. I’ve been trying to keep certain types of operations here […] because they are the basis of manual training for surgeons. (R16)

This quotation illustrates how outcome-based performance actually prevents output (cf. Pollanen et al., 2017) that is less commonly practised.

Among the officials the measurements are used, but mainly to see that you have the right level of care in specialist units. An official said:

They [the indicators] can be used to compare hospitals and to ensure that patients are at the right level of care. A dermatologist should not be examining moles and things; that should be done in primary care. (R35)

Lastly, when it comes to unintended consequences of benchmarking, one of the professionals said:

Our expenses are increasing too much, the number of doctors is increasing […], the volumes per doctor are decreasing all the time, no one operates as little as Swedish surgeons, and that’s a bit worrying. (R36)

Negotiation of performance

The empirical analysis shows how the content of performance management is negotiated not only among politicians, but also between officials and politicians (cf. Audette-Chapdelaine, 2016; Rajala et al., 2018; Brorström and Norbäck, 2020). This means that goals and performance indicators are set after discussions among the officials at the central administrative level, politicians at the regional level and politicians in the health-care committee. However, some politicians raised a concern about the negotiation process, about how they ended up caught between two sets of responsibilities: one that takes an external perspective and one focusing more on the internal processes. A politician is appointed to implement a certain policy; at the same time, there are requirements from officials and health-care professionals that they must take into consideration, as they know the operations. If politicians do not negotiate with the officials and the professionals, it could backfire on them internally, with dissatisfied employees having a direct effect on performance management and the quality of the operations. At the same time, it can backfire on them externally, as politicians are elected and have obligations towards their citizens, those who evaluate the policy and ultimately the politicians when they vote. One of the politicians said:

One of the problems I’ve experienced […] is a democratic duality. We have the politicians, who are put in place through democratic elections, but then very often we have to pursue democratic work internally, too, because many employees feel that they should have an influence, which of course they should have as employees. (R28)

However, the officials and professionals also stated that some politicians have their own political agendas. This agenda becomes visible when goals and performance indicators are set by the politicians in what officials and professionals have described as “weird goals”, as they are disconnected from actual health-care operations. An official said:

It isn’t always the decision support that’s incorrect. […] the politicians don’t always follow my recommendation. (R29)

In this context, a politician explained that the officials’ statements are rarely enough to make a decision and that they need to find other ways of getting sufficient information:

Officials’ statements, but that is rarely enough. I try to be active, listen, get out there in real life. Visit different clinics, meet with patient associations, participate in meetings […] Of course it’s anecdotal, but it gives me a bit more to go on and a bit more explanation of what I read in the officials’ statements, which are often quite brief. (R39)

Moreover, some of the goals and performance indicators are affected by media coverage; i.e. politicians tend to react to what the media are reporting. A professional gave an example:

They’ve [the politicians] made promises [in the media] that there will be one midwife per woman in labour. Of course we can’t manage that. It’s an unreasonable medical priority. (R38)

A politician said:

Sometimes it’s media-controlled and sometimes it comes from other places and the media piles on. I think that’s part of how social debates develop right now. It’s a bit too simplistic and a bit too short-term. (R28)

Part of this ad hoc and short-term reaction to the media is to construct performance indicators that can gain legitimacy not only with RS, but also with society at large. This means that the PMs are selected not necessarily to contribute to efficiency or enhanced quality, but rather to satisfy a political goal, the media and society at large. The media are, thus, reported as being an important actor to manage in the “political game” as the politicians tend to prioritise specific matters over others if they give them a good political position (cf. Höglund et al., 2021a). In relation to this, several officials stated that the media had the power of silencing the politicians. A professional explained:

The logic of politics is the logic of the media, because they [the politicians] are very media-driven. They are significantly more media-controlled now than before. (R26)

An example given by a professional regarding media and politics involved salary-setting for specific professionals:

Right now there’s competition for pay in some professional categories. […] certain groups make themselves heard in the media quite a bit and tend to get much, much more. (R38)

Another professional said:

In addition, politicians go in and say that [a hospital with a specific mission] has to survive, to get votes. Then it turns out that [that hospital] did ⅓ or ¼ of what the nearby departments did, and they had much less patient safety because there was no doctor linked to them. We noticed that at the surgical department because we had to take care of emergencies from there. When they closed, the quality of care at [in this case] the women’s clinic improved immensely. (R16)

This quote shows the power of the media in relation to the politicians, which is interesting, but our empirical analysis also shows that the professionals at times use the media to strengthen their position in discussions with politicians and officials. For example, one official reported on doctors who threatened to use the media by reporting on quality deficiencies that would be of interest to the public in an attempt to get more funding. A professional in turn stated that:

[The Public Healthcare Services have] a political structure, but there are many managers there who are almost as political as the politicians themselves and pursue their own issues. (R33)

Our results show that the politicians and some of the officials not only use PMs to follow up and report performance internally, but also externally to the media and the public. In this way, PMs are used to provide legitimacy in relation to politicians at other levels, and to the media and the public, as they are important stakeholders when reporting on performance management (cf. Bryson et al., 2014). This puts a lot of pressure on the organisation to use the “right” measurements. If not, the system might backfire and create a bad media reputation and in the long run create distrust among the citizens. Thus, the media have a direct influence (cf. Strömbäck, 2011) on PMs, as well as the conduct of performance management, which in turn influences how quality is constructed (cf. Kallio et al., 2020). Moreover, some of the professionals also use the media as a resource to “force” the politicians to take actions that benefit specific professionals or health-care units. They succeed in this because they threaten to go to the media to expose a lack of quality in the health-care system.

Concluding discussion

This paper contributes by conceptualising our understanding of PVA. We do this by studying the use and usefulness of PMs as a management tool in PSOs from a perspective in which we addressed the complexity of various (sometimes conflicting) assessments of PMM by different stakeholders. There have been several calls to develop a more nuanced empirical understanding of PVA that can contribute to further conceptualising the research area (see e.g. Bracci et al., 2014, 2019; Höglund et al., 2021b; Salemans and Budding, 2022; Steccolini, 2019). We will elaborate on our contributions below.

Firstly, the paper shows how PMs become an instrumental tool for performance management, leaving out non-instrumental use related to such things as quality (cf. Huston et al., 2022) and public value (Bryson et al., 2014). Overall, our analysis paints a picture of PMM and the use of PMs as a tool formulated in a more process- and primarily efficiency-oriented way, which resembles New Public Management thinking. This results in a focus on measuring efficiency rather than effectiveness, output rather than outcome (cf. Pollanen et al., 2017; Rajala et al., 2018) and, hence, quantity rather than quality (Pollitt, 2018). In line with Salemans and Budding (2022), we argue that this way of working with PMM becomes an obstacle for steering towards public value creation. Moreover, both previous research (Johnsson et al., 2021) and our results imply that PMs must be developed to better fit the specific context of the operation. In other words, benchmarking activities tend to stimulate general and easily produced metrics, which, in our case, had a direct effect on quality and public value creation.

Secondly, we show the importance of studying the interaction among different stakeholders in PVA research (cf. Audette-Chapdelaine, 2016; Modell, 2022; Rajala et al., 2018). Our results exposed a significant influence, not only from the politicians but also from the public and the media (cf. Strömbäck, 2011), as to which PMs were selected. This highlights the importance of including the media as well as the public as actors with significant constitutive influence (Höglund et al., 2021a) on the choice of PMs for management purposes and the creation of public value (Höglund et al., 2021b).

Thirdly, we argue that there is a conceptual shortcoming in the discussion of PVA, because the effort needed to achieve outcome-based information might exceed the ability of an organisation to deliver it (Lowe, 2013). In particular, the officials were displeased with the politicians regarding the inability to work with more outcome-related performance of long-term public value. Politicians are not always strategic, which may give rise to tensions between different time frames (Liguori and Steccolini, 2018). Hence, short-term political pressures affect most PSOs, resulting in difficulties in developing useful and meaningful PMs, especially in relation to public value (Höglund et al., 2021b). Most officials working with management are expected to promote outcome, whilst simultaneously being accountable to politicians and society for the output (Norman, 2007). Politicians wanted the new PMM model to focus on output, not on outcomes, as the latter do not help them in political struggles and deals (Jethon and Reichard, 2022).

Fourthly, the results indicate that to achieve specific outcomes it is important to organise stakeholder contacts among different stakeholders to clarify strategic goals before they are operationalised (cf. Bryson et al., 2014). In our case, the new PMM model became something the professionals only reported on, but did not include in their daily operations (cf. Petit and Lux, 2020). This had a direct consequence for performance management, as the politicians did not get sufficient performance information (cf. Mutiganda, 2016; Jethon and Reichard, 2022; Raudla, 2022). Officials and professionals in the health-care department also became less efficient in their work, as they had to produce PMs that were not relevant to the health-care division.

Lastly, we addressed what kind of PMs are valued and determined in contexts where conflicting interests and political agendas are in play (Steccolini, 2019). In relation to this, there is a need to consider possible power relationships associated with the various conceptions of performance and public value. How we frame performance in relation to PVA is motivated by the notion of public value; i.e. broader or narrower conceptions of performance are likely to evolve and this will, in turn, influence what kind of stakeholders are represented in processes of public value creation. In line with Modell (2022), our results suggest the need to further enhance our understanding of which stakeholders gain or lose power through such processes. Scholars could take research on PVA forward while, at the same time, infusing research on PMM with more explicit, critical intent focusing on power struggles.

Table 1.

An overview of interviewed respondents

Code	Position	Org. level	Situated	Year-month	Min
R1	Official	Health-care division	Management office	2018–11	81
R2	Official	RS headquarter	Management office	2018–11	88
R3	Official	RS headquarter	Management office	2018–11	82
R7	Official	Health-care division	Management office	2018–11	53
R8	Official	Health-care division	Management office	2018–11	78
R13	Professional (manager)	Health provider	Hospital	2019–02	56
R14	Professional (manager)	Health provider	Primary health care	2019–02	63
R15	Manager	Health-care company		2019–02	53
R16	Professional (manager)	Health provider	Hospital	2010–02	70
R17	Manager	Health provider	Hospital	2019–03	56
R18	Professional (manager)	Health provider	Hospital	2019–03	68
R19	Official	RS headquarter	Management office	2019–03	56
R20	Official	Health-care division	Management office	2019–03	54
	Official	Health-care division	Management office
R21	Professional (manager)	Health provider	hospital	2019–03	100
R22	Professional (manager)	Health provider	Primary health care	2019–03	54
R23	Manager	Health provider	Hospital	2019–02	87
R24	Professional (manager)	Health-care division	Management office, primary health care	2019–04	75
R25	Politician	RS headquarter		2019–06	63
	Official	RS headquarter
R26	Official	RS headquarter	Management office	2019–06	93
	Official	RS headquarter	Management office
R27	Professional (manager)	Health provider	hospital	2019–06	73
R28	Politician	RS headquarter		2019–06	66
R29	Official	Health-care division	Management office	2019–06	56
R31	Professional (manager)	Health provider	hospital	2019–10	88
R32	Official	Health-care division	Management office, primary health care	2019–10	57
R33	Official	RS headquarter	Management office	2019–10	88
R35	Official	Health-care division	Management office	2019–10	79
R36	Professional (manager)	Health provider	hospital	2019–10	65
R37	Official (professional)	Health-care division	Management office, primary health care	2019–10	114
R38	Professional (manager)	Health provider	hospital	2019–10	58
R39	Politician	Health-care division	Political committee	2019–11	76
				In total: 36 h (20,150 min)

Source: Authors’ own creation