Exploring types of individual unlearning by local health-care managers: an original empirical approach

Article history

The research reported in this issue of the journal was funded by the HS&DR programme or one of its proceeding programmes as project number 09/1002/34. The contractual start date was in June 2010. The final report began editorial review in July 2012 and was accepted for publication in November 2012. The authors have been wholly responsible for all data collection, analysis and interpretation, and for writing up their work. The HS&DR editors and production house have tried to ensure the accuracy of the authors' report and would like to thank the reviewers for their constructive comments on the final report document. However, they do not accept liability for damages or losses arising from material published in this report.

Copyright statement

© Queen's Printer and Controller of HMSO 2013. This work was produced by Coombs et al. under the terms of a commissioning contract issued by the Secretary of State for Health. This issue may be freely reproduced for the purposes of private research and study and extracts (or indeed, the full report) may be included in professional journals provided that suitable acknowledgement is made and the reproduction is not associated with any form of advertising. Applications for commercial reproduction should be addressed to: NIHR Journals Library, National Institute for Health Research, Evaluation, Trials and Studies Coordinating Centre, Alpha House, University of Southampton Science Park, Southampton SO16 7NS, UK.

The concept of individual unlearning

The capability to unlearn is important as the inability to give up or abandon knowledge, values, beliefs and/or practices can produce a rigidity in thinking and acting, and can create a blinkered outlook, limiting a health-care manager's or organisation's adaptability. 3 This can occur when existing views are never questioned or challenged, for example through the use of ‘defensive routines’. 4 The inability to question what may have been successful organisational norms, values, practices and knowledge can create what have been referred to as ‘competency traps’,5 in which useful competencies become outdated through never being challenged, revised or abandoned. Thus, the ability of health-care managers or organisations to unlearn established knowledge, behaviours or values can be a significant catalyst to and facilitator of change.

The analysis developed in this report builds from and extends the work of Tsang and Zahra,2 who developed a conceptual analysis of organisational-level unlearning processes. Tsang and Zahra 2 concluded their article by suggesting which areas of knowledge on unlearning are still limited and which require development (p. 1454). One key area where knowledge is particularly limited is the dynamics and character of individual-level processes of unlearning, and how they connect to and interact with organisational learning and unlearning. Individual-level processes of unlearning represent a neglected topic in an under-researched field, in which analysis has been relatively limited. Our concern here is to take a first step towards addressing this gap in knowledge.

Rushmer and Davies6 argue that distinctions can be made between three separate types of individual unlearning. The first type of individual unlearning that they identify, which they label fading, refers to the slow gradual process of unlearning that can occur over time, when particular skills and capabilities are forgotten through a lack of usage. This form of unlearning is typically not experienced as being significant or challenging for people, and because of the gradual way that it occurs people may not be conscious that they are experiencing it. The second type of individual unlearning identified by Rushmer and Davies is labelled wiping. This represents a more significant, deliberate and conscious form of unlearning. Wiping is a form of unlearning focused relatively narrowly on a particular practice, or activity, in which a person consciously unlearns through making deliberate efforts to give up a particular way of thinking/acting, as a precursor to changing these beliefs/practices. Rushmer and Davies argue that this process of unlearning may be initiated by some external catalyst, such as a change initiative, which places demands on people to change their ways of working and/or thinking. The third form of individual unlearning identified by Rushmer and Davies is deep unlearning. Deep unlearning represents a radical form of unlearning that typically occurs rapidly and unexpectedly, and is experienced by people as significant as it brings into question some basic assumptions and values. Because of these characteristics this form of unlearning can produce significant emotions (fear, confusion, etc.) in the people who experience it. Rushmer and Davies argue that some external catalyst, such as an event whose characteristics or outcomes are unexpected, initiates deep unlearning. A key difference between this form of individual unlearning and wiping is the speed at which the change is made and the emotional impact associated with the experience. Deep unlearning is argued to be a fast process involving high emotional impact for the individual experiencing the unlearning.

Context of the study

The central focus of this study was on examining the extent to which health-care managers engage in processes of individual unlearning. A typical catalyst for individual unlearning is processes of change that require the adaptation of working practices. 2 The pace of change in the NHS in recent years has been significant. For example, Greener7 talked about how the ‘pace and scope of change’ was ‘intense’ (p. 206). Thus, as a result of the amount of ongoing change in the NHS, all health-care managers are likely to have experiences of individual unlearning, in which they have had to adapt their work practices and routines as a result of change. Further, individual unlearning may also be experienced by health-care managers who have undergone a significant role transition, such as would happen when a clinician moves into a managerial role. For example, a clinician moving into a clinical management position may need to adapt the way he or she uses knowledge to act and make decisions, as anecdotal evidence suggests that clinical and managerial decision-making processes are underpinned and supported by different knowledge bases. Clinicians moving into managerial roles thus arguably need to engage in processes of individual unlearning to effectively adapt how they act and make decisions in the new managerial roles they occupy. The lack of research on the topic of individual unlearning means that there is a limited understanding of how frequently health-care managers experience it or the type of events/circumstances that trigger it. Therefore, in this study we also investigated the triggers to individual unlearning with all participating health-care managers.

This research was also concerned with understanding the barriers to and enablers of individual unlearning that exist. Although individual unlearning has the potential to be an important component in the processes of learning and change, research suggests that there are many barriers to unlearning at both the individual and the team/organisational level, which means that learning from mistakes, failure and crisis often does not happen. 8 Similar observations can be made regarding enablers to individual unlearning. For example, Becker9 argues that, to enable staff to have a positive attitude to change and consequently to unlearning, managers need to demonstrate commitment, make the case for change, involve staff in planning and provide reassurance. Other examples of enablers include having secure but challenging conversations with trusted colleagues,6 and the informal support of colleagues and managers9,10 and mentors. 10 The small amount of research on unlearning6 and learning from failure that has been carried out in the health sector suggests that this is a work context in which the barriers and enablers to unlearning can be significant. However, our understanding of the key barriers to and enablers of different forms of unlearning remains sketchy.

Existing research also suggests that NHS managers with clinical and non-clinical backgrounds may have a different approach to issues such as accountability, use of guidelines and finance. It has been argued that these differences are a result of each discipline's training, beliefs and experiences. 11 Similarly, there is evidence to suggest that clinicians and managers have different priorities concerning health-care delivery: clinicians tend to prioritise patient care, whereas managers focus more on cost. 12 Research by Guven-Uslu13 suggests that these different mindsets influence how managers with a clinical background are likely to approach decision-making compared with those without clinical experience. Given the increased emphasis that is being given to encouraging clinicians into leadership and managerial roles,14 these different mindsets are likely to become increasingly more significant to the efficient running of the NHS. Therefore, the individual unlearning processes associated with significant change, such as changes in role, may also have implications for the way that health-care managers review and reflect on their decision-making processes, and ultimately how they make future decisions.

Thus, there is a need for more research to understand the nature of individual unlearning in the health-care environment. The main aims of this study were to examine whether or not health-care managers engage in processes of individual unlearning, and the individual- and/or organisational-level barriers that may inhibit effective unlearning and learning from occurring or facilitate it. It also explored the implications of individual unlearning for managerial decision-making in the health service.

This study investigates these aims through an examination and synthesis of the academic literature concerning the processes of individual unlearning, and capturing and analysing health-care managers' experiences of individual unlearning. The study utilised a case study approach, investigating experiences of unlearning in two different types of NHS trust (an acute trust and a mental health trust). A case study-based approach represents a suitable methodology for the investigation of individual unlearning because, as outlined, individual unlearning is a relatively neglected and unexplored aspect of learning and change processes, and qualitative case studies provide an effective way to conduct exploratory research, which can give rich, qualitative insights into managers' experiences of individual unlearning. 15,16

Within each organisation, the main source of data was face-to-face, one-to-one, semistructured, qualitative interviews with a range of health-care middle managers. The purposive sample participants varied in terms of clinical/non-clinical background, type of department/unit and length of time as a manager. A total of 29 health-care managers were interviewed across both sites, which captured a total of 57 episodes of unlearning and 28 episodes of fading. The participants were also invited to attend a workshop to hear about initial findings and to comment on those findings.

Contributions

Although this original exploratory study is relatively small in scale, with the data drawn from 29 health-care managers from two NHS trusts, it still provides a number of important contributions. First, we have developed the concept of individual unlearning. We argue that individual unlearning is a distinctive type of learning, involving a conscious decision to give up knowledge, values or behaviours. However, this abandoned knowledge is not permanently lost but put to one side, and it remains retrievable for future use. Using this conceptualisation of unlearning, we argue that the category of fading proposed by Rushmer and Davies6 is more akin to a process of unintentional forgetting rather than unlearning. Indeed, fading may not enter individuals' consciousnesses unless they are invited to identify lost skills or capabilities. Therefore, as unlearning requires conscious and intentional action, it is questionable whether or not fading can be conceptualised as unlearning.

Second, we conceptualise two distinct types of individual unlearning and their differentiating features and dynamics. The first type of unlearning (behavioural) is triggered by a deliberate process of change that has been externally imposed. We argue that behavioural unlearning has a bidirectional relationship with organisational change and that an individual's attitude to change may shape his or her attitude to unlearning. A significant weakness in the unlearning literature is its neglect of people's attitudes to unlearning, an area that must be considered when attempting to understand individual behavioural-type unlearning processes. The second type of individual unlearning (cognitive) is triggered by an unexpected external event that questions some basic assumptions of the individual. We argue that it is important to recognise that the cognitive unlearning process is emotionally challenging, but may not occur suddenly or produce an instant change in behaviour.

Third, we develop a new typology that distinguishes between four separate types of individual unlearning, as shown in Table 1. Previous research on unlearning6 proposes that each type of individual unlearning is linked to a different type of catalyst. Thus, using Rushmer and Davies'6 labels, change events are the catalyst for wiping (behavioural unlearning), whereas individual experiences are the catalyst for deep unlearning (cognitive unlearning). However, during the course of data collection it became apparent that this distinction between catalysts was questionable, as the catalysts for both wiping and deep unlearning included both individual experiences and change events. Thus, our analysis suggested that, although the distinction between unlearning types was useful, as was the distinction between types of catalyst, Rushmer and Davies' typology could be improved through a recognition that specific catalysts were not associated with each type of unlearning. This led us to reconceptualise their typology (see Table 1). This was carried out by combining Rushmer and Davies' two types of unlearning, which results in a typology that distinguishes between four separate types of individual unlearning.

In addition, with two exceptions,6,17 the literature on this topic defines unlearning as the conscious and intentional abandoning of skills, behaviours and values, whereas fading equates to forgetting and is unconscious and unintentional (see Chapter 2, Types of individual unlearning and the nature of individual unlearning processes, Fading). De Holan and Phillips18 take this further by distinguishing between memory loss or accidental forgetting and unlearning, which is intentional. In our view, lack of conscious action means that fading is more akin to unintentional forgetting than unlearning, hence the exclusion of fading from the typology.

Another way in which we reconceptualised Rushmer and Davies' typology6 was by relabelling wiping as behavioural unlearning and deep unlearning as cognitive unlearning. The reason for doing this was to create labels for the unlearning types that were more explicit and clear regarding what was being unlearned, as supported by our empirical data analysis. In this new typology, ‘behavioural unlearning’ refers to the individual unlearning of specific behaviours such as practices, activities or routines, which have no (or limited) impact on people's underlying knowledge, values and assumptions. As in Rushmer and Davies' model,6 behavioural unlearning, like wiping, does not have a significant affective impact. Deep unlearning, which is relabelled ‘cognitive unlearning’, is emotionally charged as it involves giving up or abandoning more deeply held knowledge, values and assumptions. In making this distinction, we further found that, although behavioural unlearning may be restricted to this domain, cognitive unlearning is likely to be accompanied by or lead to behavioural unlearning. However, it must be emphasised that the typology was developed from the data analysis and informed by a review of the relevant literature and that Rushmer and Davies' unlearning typology was the initial perspective utilised to differentiate between different types of unlearning.

Fourth, our study also considers the influence of unlearning on health-care managers' decision-making, a variable that has not been examined in other unlearning studies. Our study provides evidence to suggest that individual unlearning can impact on some health-care managers' decision-making processes. This impact involves managers moving away from an idea imposition process of decision-making to a discovery-led process of decision-making. The discovery-led process has been argued in the literature to be more successful and, therefore, the findings from this study suggest that there may be a relationship between health-care managers who engage with individual unlearning and improvements in decision-making.

Fifth, our study involved a type of research data not previously utilised in this domain – qualitative interview data – which give insights into how individual health-care managers understand the changes that their unlearning experiences have produced.

Structure of the report

In the following chapter we review in more detail the literature concerning unlearning. The chapter clarifies and develops the unlearning concept before reviewing the academic literature on unlearning. The types of individual unlearning are discussed along with the nature of the individual unlearning process. The chapter also provides a summary of recent change in the NHS, which provides the context for the study. The chapter concludes with the research questions that the study investigated.

Chapter 3 discusses the key methodological considerations for the study. The choice of research strategy, selection of case study sites and processes of data collection and analysis are explained. The chapter also provides a description of the case study sites.

Chapter 4 presents the findings of the study and reflects on the research objectives presented at the end of Chapter 2. It is divided into six main sections. Episodes of individual unlearning and fading describes the nature of unlearning and fading episodes and the ways in which health-care managers experience unlearning and fading. Barriers to individual unlearning and Enablers of individual unlearning and fading report, respectively, the barriers to and enablers of unlearning and fading that were identified by participants. The impact of individual unlearning and fading on managers' decision-making reports the influence of individual unlearning on decision-making and the changes that participants had made in relation to the different types of unlearning experienced. The influence of health-care setting and professional background on health-care managers' individual unlearning and fading reports the analysis examining whether health-care context or the professional background of participants differentiated participants' experiences of individual unlearning and fading. Finally, Workshop findings provides a summary of the key findings from the workshop conducted at each case study site.

Chapter 5 discusses the key findings of the report in relation to the study research objectives and the existing literature. It also provides a discussion of the relationship between learning, unlearning and change that became evident during the course of the study. Finally, conclusions, limitations, recommendations for future research and implications for practice are presented in Chapter 6.

Introduction

This chapter draws on a range of literature to suggest that the process of individual unlearning may have particular features. The analysis of individual unlearning presented here is based on a review of the contemporary academic literature on unlearning, but to help address themes that are relatively unexamined by this literature it also draws on a wider body of work on learning and change. After the chapter unpacks and defines the concept of unlearning, it presents the details of the literature search that was conducted. Following this is a large section that differentiates between two different types of individual unlearning, and which suggests that each type of individual unlearning may have its own distinctive features and dynamics. The final two sections summarise recent changes in the UK health-care sector and present the research questions of the study.

Clarifying and developing the unlearning concept

Thus far, ‘unlearning’ has been very broadly defined as abandoning or giving up knowledge, ideas or behaviours. This is in contrast to ‘learning’, which has been defined as increasing one's capacity to take effective action through the addition of new skills or knowledge. 19 However, to fully understand the concept of unlearning it is necessary to define it in greater detail. Although unlearning can be traced back to the 1950s, until recently it has been largely neglected in the literature on learning and knowledge management. Akgün et al. 20 trace the origins of unlearning to literature on learning and cognitive psychology published in the 1950s and 1960s. Another perspective on unlearning emerged in the 1950s, which explored ‘unlearning . . . the inherent dominative mode’ in relation to Western thinking about the ‘other’ (p. 376). 21 The ‘other’ referred to how people in the West view people or perspectives they regard as non-Western. Unlearning in this context concerned Westerners trying to think beyond their own perspective to take account of others. 21 Indeed, some educational literature22 applied Williams'21 meaning to understand how trainee teachers could unlearn their attitudes to ethnic minority and working-class young people. Similarly, Mavin et al. 23 utilise unlearning in the same sense and suggest that an unlearning process is necessary to challenge the unquestioned and unchallenged gender-blind and male-biased character of the academic business and management literature.

In the business and management literature unlearning tends to be linked to one or more of organisational learning, change and memory,2,8,9,20 knowledge management24,25 or human resource management. 26 However, some authors draw on more than one discipline, relate unlearning to individuals and/or groups and locate the concept in the disciplines of education and learning2,22,27 or psychology. 26

If unlearning involves the giving up or abandonment of knowledge, values or behaviours, it needs to be acknowledged that this can happen both unconsciously and deliberately. The unconscious or accidental giving up of something is typically referred to as forgetting, as it occurs over time through particular knowledge or behaviours becoming unused and eventually forgotten. 6,17 This process of forgetting contrasts with deliberate unlearning, which involves a process of consciously choosing to abandon or give up particular knowledge, values or behaviours. As with Tsang and Zahra,2 the assumption here is that unlearning is a conscious and intentional process and, as such, is distinct from forgetting. De Holan and Phillips18 agree about the deliberate nature of unlearning, although they do not distinguish in the same way as others between forgetting and unlearning. Indeed, for them, ‘managed unlearning’ is one of four modes of forgetting old knowledge.

A second area of difference in the unlearning literature relates to whether or not the knowledge or behaviours being given up are obsolete, outdated and in some way inferior to new knowledge or behaviours that are subsequently acquired. As highlighted in Table 2, a number of articles make this assumption. 24,28,29 Thus, for example, Srithika and Bhattacharyya28 define organisational unlearning as ‘the identification or removal of ineffective or obsolete knowledge’ (p. 68). However, making such a value judgement regarding the inferiority of the knowledge to be abandoned is unnecessarily restrictive and judgemental. Thus, similar to Tsang and Zahra,2 we suggest that it is more appropriate to define unlearning simply as abandoning or giving up knowledge or behaviours without making any judgement on the status of the knowledge or behaviours being unlearned.

In considering individual unlearning, an issue that is typically neglected is what happens to the knowledge or behaviours that people unlearn. It is important to acknowledge that what is unlearned is not permanently lost by people such that they are unable to think or act in the way that they had done previously. Arguably, the only ways that particular capabilities could become permanently lost are through some type of medical or neurological intervention (drugs, surgery, etc.), through developing an illness or having an accident (such as having a stroke or a car accident that results in brain injury) or through lack of use over a long time period. Thus, the type of deliberate individual unlearning considered here does not involve the permanent loss of something, but instead involves a person putting particular values, knowledge or behaviours ‘to the side’ and consciously choosing not to continue using them. Thus, individual unlearning in this sense is not necessarily permanent because, either consciously or unconsciously, people may at some point in the future begin to reuse that which they had previously abandoned or unlearned. An example of this would be when someone changed how he or she undertook a task by returning to do it in a way that had been previously abandoned.

The final issue in developing the concept of unlearning is how it is relates to learning. Tsang and Zahra2 consider that unlearning may precede learning, occur simultaneously with learning or occur independently of learning. However, the dominant perspective regarding the relationship between the sequencing of unlearning and learning is that unlearning is a unique stage and is a prerequisite to, and a precursor of, learning. 20,26,30–32 For example, Cegarra-Navarro et al. 30 define unlearning as ‘the elimination of obsolete knowledge’, which is regarded as a necessary precursor to learning, or ‘the creation and absorption of new knowledge’ (p. 901). However, an alternative way to conceptualise the relationship between unlearning and learning is to consider unlearning as a distinctive type of learning. 33 This is the perspective utilised by Argyris and Schön,34 who argued that ‘we may also speak of the particular kind of learning that consists of “unlearning”: acquiring information that leads to subtracting something (an obsolete strategy, for example) from an organisation’s existing store of knowledge’ (pp. 3–4).

In summary, in examining individual unlearning this section suggests that individual unlearning should be conceptualised as a distinctive type of learning. It involves a conscious process of choosing to give up knowledge, values or behaviours. No value judgement should be made regarding the value or status of what is abandoned, and that what is unlearned is not permanently lost to people and may be utilised again at some point in the future.

Reviewing the academic literature on unlearning

Although Tsang and Zahra2 conducted a partial review of the unlearning literature, their central focus was organisational-level unlearning rather than individual unlearning. Their review identified 34 separate pieces of work. However, they do not specify the boundaries of the search that was conducted to produce this list, with it including book chapters and journal articles published over a wide time frame.

Because of the multidisciplinary nature of the interest in the topics of learning and unlearning, we searched several management (Business Source Complete, Emerald), psychological (PsycINFO), health (MEDLINE) and education [Education Resources Information Center (ERIC)] electronic databases for English-language articles that were published between January 2000 and March 2012. We searched for articles that had ‘unlearning’ in the title, abstract or keywords. Additionally, we searched for articles on ‘abandoning behaviour or knowledge’ and ‘giving up behaviour or knowledge’. This search generated over 330 articles. From these sources we concentrated on those published in peer-reviewed scientific journals, leaving 261 articles. After removing duplicate search results we examined the abstracts of these articles and excluded studies that reported on animal-based, psychological or memory experiments. We also excluded personal viewpoint and unreferenced opinion articles, leaving approximately 100 articles. After a first round of reading of the collected articles, we selected those articles that investigated the topic of unlearning, either theoretically or empirically. This led to the exclusion of articles that used the term ‘unlearning’ in the abstract or title, but which were not fundamentally concerned with investigating it as a topic. Although the focus in this article is fundamentally on individual-level unlearning, our initial review included all articles on unlearning, whether they were focused on individual-, team- or organisational-level unlearning (see Tables 2 and 3). We also searched the reference lists of all sources collected and performed citation searches, which resulted in the addition of several relevant articles.

A total of 34 articles2,3,6,9,10,20,22–26,28–32,37–49 were identified for analysis (Table 2). An initial observation from the list is that < 35 relevant articles were identified in a time period of over 10 years, which highlights the extent to which the concept of unlearning is neglected and underdeveloped. This neglect is in stark contrast to the considerable level of interest in the topic of learning since the mid-1990s. 5,35,36

In terms of how the literature defines unlearning (see Table 2, column 2), although many authors develop their own particular form of words, what is noticeable about the way that unlearning is defined (also found to be the case by Tsang and Zahra2) is the striking degree of homogeneity that exists. What is common to these definitions is that unlearning involves ‘abandoning’, ‘eliminating’, ‘rejecting’, ‘discarding’ or ‘giving up’ something – with that something being particular values, assumptions, knowledge or behaviour at the individual level, and knowledge, assumptions or routines at the organisational level.

In terms of the type of unlearning examined, as Table 3 highlights, there has been a greater focus on organisational or group or team unlearning (21 articles2,3,20,23,24,26,28,31,37,40,42–44,46,47,50–54) than on individual-level unlearning (16 articles6,9,10,22,23,25,26,29,30,32,37–39,41,45,48). In this context, group or organisational unlearning, as with organisational learning, refers to norms, assumptions, behaviours and routines that are collectively shared and understood. 20,24 Although a few articles look at multiple levels of unlearning,37 or the inter-relationship between different levels of unlearning,38 the vast majority of articles focus on one level of unlearning alone.

Another difference in the focus of the reviewed articles was whether they examined the antecedents, process or consequences of unlearning. The largest proportion of articles (n = 183,9,10,20,22,25,26,30,39,41–47,53) examined the antecedents of unlearning, with unlearning facilitated by the organisational context,9,25,30 environmental turbulence51 and organisational size. 26 A total of 133,23,24,29,31,37,38,40,48–51 of the reviewed articles examined the consequences of unlearning, with unlearning argued to be related to a diverse range of processes and outcomes including knowledge transfer processes,49,50 processes of internationalisation,24 the non-academic impact of academic scholarship,33 organisational performance37,54 and health-care policies. 31 Finally, only 62,6,20,28,32,52 of the 34 articles examined the character and dynamics of unlearning processes.

Table 3 also reveals that of the 34 articles reviewed only two6,32 focused on individual processes of unlearning. These are the articles by Rushmer and Davies6 and MacDonald. 32 Further, of these two only MacDonald32 presents any empirical evidence, which was a reflection on personal experience. Thus, to say that there is a conceptual and empirical gap in knowledge with regard to the process of individual unlearning is an understatement.

The following utilises the work of MacDonald32 and Rushmer and Davies6, as well as some other literature on learning and change, to consider the character and dynamics of the process of individual unlearning. In so doing it is suggested that distinctions can be made between different types of individual unlearning.

Types of individual unlearning and the nature of individual unlearning processes

As has been outlined thus far, distinctions can be made between individual- and organisational-level unlearning, and between the unlearning of values/assumptions, beliefs, skills, knowledge and/or behaviours. A number of authors go beyond these distinctions to construct distinctive categories of unlearning. 6,51,54 For example, Akgün et al. 51 develop a typology that links differences in the nature of the business environment to the character of organisational unlearning. Sinkula54 on the other hand distinguishes between the unlearning of axiomatic knowledge and the unlearning of procedural knowledge. Here, axiomatic knowledge is defined as fundamental unquestioned beliefs and values, and procedural knowledge is considered to be equivalent to Argyris and Schön's34 concept of ‘theory in use’, referring to the tacit knowledge that shapes the way that people act.

However, the categorisation proposed by Rushmer and Davies6 is the most relevant to individual unlearning as they propose a useful distinction between three possible separate and distinctive types of individual unlearning: fading, wiping and deep unlearning (Table 4). Each type of unlearning is argued to differ in respect of catalyst, intentionality, speed and impact. First, fading or routine unlearning occurs gradually over time through lack of use. It is regarded as neither significant nor challenging for people. Indeed, fading may not enter individuals' consciousness unless they are invited to identify lost skills or capabilities.

Wiping is the second category of individual unlearning identified by Rushmer and Davies. 6 The catalyst for wiping is a change initiative external to the person. Wiping can be defined as a process of unlearning that results from a deliberate process of change that has been externally imposed, for example a change initiative or a change in job role. Wiping is deliberate, conscious and more significant than fading, and is typically focused on a relatively narrow practice or activity, with a person consciously making deliberate attempts to give up a particular way of thinking and acting as a precursor to changing his or her beliefs and practices.

Parallels between wiping and categories of unlearning developed by other authors can be discerned. For example, wiping is similar to ceasing a particular behaviour and making incremental change8 and to intentional forgetting of new or existing knowledge. 18 The notion of wiping is reinforced and extended by other categories such as ‘operational level unlearning’, whereby performance routines (enacted by individuals) and ostensive routines (codified systems) are discarded as a result of evolutionary, continuous, incremental change. 2 Wiping is also similar to ‘adjustive unlearning’, in which incremental changes in beliefs are accompanied by fundamental changes in routine, and to ‘operative unlearning’, which involves small-scale changes in beliefs and routines. 51

The third category of individual unlearning proposed by Rushmer and Davies6 is deep unlearning. This radical form of unlearning is argued to occur rapidly as a result of an external event whose characteristics and/or outcomes are unexpected, and which bring into question some basic assumptions. Characteristically, it has a significant impact on the individuals who experience it, leading them to question their values and beliefs, and possibly their frame of reference. As a consequence, deep unlearning may be accompanied by challenging emotions such as anxiety, fear and confusion. Thus, in contrast to wiping, deep unlearning is more likely to involve the unlearning of values and assumptions.

Like wiping, the notion of deep unlearning is echoed elsewhere. Baumard and Starbuck8 talk about challenging core beliefs and Tsang and Zahra2 characterise deep unlearning as discarding values and norms as a result of episodic, discontinuous change. Deep unlearning is also similar to reinventive unlearning and formative unlearning identified by Akgün et al. 20,51 Reinventive unlearning occurs when an organisation changes both beliefs and routines in response to a changing and highly unpredictable environment, whereas formative unlearning occurs when new beliefs structures are combined with incremental routine change. Finally, there are similarities between deep unlearning and what MacDonald32 defines as transformative unlearning (see Deep unlearning).

Finally, linking back to the idea that unlearning represents a distinctive type of learning, it can be suggested that, although wiping has parallels with single-loop learning (incremental learning or change in which basic assumptions remain unchallenged), deep unlearning can be equated more with double-loop learning (learning or change in which existing assumptions and values are questioned and reflected on).

The above definitions and descriptions provide only a brief overview of the general character of fading, wiping and deep unlearning processes. Thus, to develop a fuller understanding of the dynamics and character of the processes of fading, wiping and deep unlearning it is necessary to consider them in more detail. The definitions suggest that there are significant differences between the ways that people experience and understand fading, wiping and deep unlearning, and in the process dynamics of fading, wiping and deep unlearning. Therefore, it is useful to consider each separately, which is presented in the following three subsections.

Fading

Rushmer and Davies6 argue that unlearning could be considered as a process that will automatically occur when the factors that sustain the original learning are removed. Schein55 refers to unlearning as the forgetting curve, which suggests that some past learning will simply fade away over time. Rushmer and Davies6 propose that, with simple behavioural actions, this passive replacement can be seen as moving down the unlearning curve. They give an example of a health-care professional comprehending a new, altered, mandatory health form. The professional may first puzzle over the new layout and in completing the form may make errors through an erroneous habitual response. However, Rushmer and Davies argue that, as the health-care professional continues to complete the form, familiarity and confidence are gained with the new layout. The old way of completing the form recedes, discomfort and previous expectations fade and forgetting takes place.

However, as unlearning requires conscious and intentional action it is questionable whether or not fading should be conceptualised as unlearning. A lack of conscious action suggests that fading is more akin to a process of unintentional forgetting, rather than unlearning. Even authors who use the term ‘forgetting’ in their definitions of unlearning18,30 identify intentionality or purposefulness as defining characteristics of unlearning. We could not find any other studies that conceptualise passive unlearning, possibly because of the difficulty of investigating the concept. Without longitudinal study it would be very hard to gather data on learning that had been unintentionally forgotten, as by definition the participant could struggle to identify it. For this study we decided to continue to use the fading category for the purposes of empirical testing and fulfilling our original project proposal. This allowed us to examine the concept of fading and the value of this element in Rushmer and Davies'6 categorisation.

Wiping

In terms of the relationship between unlearning and change, the dominant perspective in the unlearning literature is that unlearning is a facilitator of change. 2,9,10,23,31,38,50 Although this may be true in relation to wiping (see Table 4), it also needs to be acknowledged that the relationship works in the opposite direction, with external change acting as the prime catalyst for wiping. Thus, in the context of wiping, the primary reason why people engage in unlearning is because it is perceived as being a necessary element of a specific organisational change initiative. This is articulated explicitly in a textbook on change,56 which, in talking about the importance of change for contemporary organisations, says that ‘people are being required to unlearn old ways and develop new competencies’ (p. 7).

Although much of the unlearning literature suggests that the relationship between unlearning and change is close, it has considered only unlearning as being a facilitator of change and has not examined the relationship between unlearning and change in any detail. Consequently, the literature provides limited insight into how change can be a facilitator of wiping, or how individuals experience the character and dynamics of the process of wiping.

The most useful article for considering the bidirectional relationship between unlearning/wiping and change is the conceptual article by Akgün et al. 20 In talking about the relationship between unlearning and change it refers explicitly to Lewin's57 three-stage model of change. This very simplified and much criticised model of change suggests that change happens through the sequential processes of unfreezing, change and refreezing (see Akgün et al.,20 pp. 800–1; Hayes,56 p. 52). Akgün et al. assume that unlearning and learning together constitute the second stage in Lewin's model. This conceptualisation of the wiping/change relationship, with unlearning being at the centre of change, highlights the bidirectional nature of the change/wiping relationship. However, the focus of Akgün et al.'s20 article is on organisational-level unlearning/wiping and, thus, it does not provide insights into the character and dynamics of individual-level unlearning/wiping.

Tsang and Zahra2 also examine the relationship between learning, unlearning and organisational change. They distinguish between different types of change (continuous and episodic) and suggest that each type of change will involve a distinctive form of unlearning. They define continuous change as change that is incremental and gradual in character. By contrast, episodic change is typically discontinuous and infrequent and is greater in scope than continuous change. Episodic change can also be linked to a process of double-loop learning in which basic assumptions are challenged. Thus, in relation to the types of unlearning considered here, continuous change can be linked more to wiping, whereas episodic change can be linked more to deep unlearning.

As wiping is so closely inter-related with processes of organisational change, it is useful to refer to some change-related concepts. In this context, if organisational change provides the catalyst for wiping/unlearning, people's attitude to unlearning is likely to be closely linked to and virtually inseparable from their attitude to the change process that precipitated it. Thus, if people do not regard the changes being undertaken as favourable they are unlikely to have a positive attitude to any unlearning that flows from the change. Equally, if the opposite is the case and people do regard change as necessary and important, they are likely to have a more positive attitude to any unlearning it precipitates. Although Tsang50 does not explicitly use the concept of resistance to change, the reluctance to unlearn and learn that he found in relation to the knowledge transfer processes that were examined can be argued to constitute resistance to change.

The concept of resistance to change is useful when considering people's attitudes to change and unlearning. The change literature suggests that, because of the uncertainty caused by change, resistance is common. A key theme in the change literature is concerned with anticipating, managing and minimising any potential resistance to change. 56 Although some of the unlearning literature touches on the topic of resistance to change,47,52 people's attitudes to unlearning are neglected. This neglect may be because of the assumption that people will embrace wiping-type unlearning relatively willingly. However, this assumption represents an important omission because people's attitude to unlearning is likely to be shaped by their attitude to change. Thus, to understand the character and dynamics of individual-level wiping-type unlearning processes it is fundamentally necessary to take account of people's attitudes to the changes that precipitated them.

Deep unlearning

In examining how individuals experience deep unlearning and the process through which it unfolds, few of the unlearning articles reviewed are relevant. Of the six articles2,6,20,28,32,52 that focus on the process of unlearning (see Table 3), only two examined processes of deep unlearning,6,32 with the other four concerned with individual and organisational unlearning, which is more equivalent to wiping. In tentatively outlining a model for the dynamics of the process of deep unlearning this section draws on MacDonald's32 empirical and conceptual work and links it with some wider, relevant literature on learning.

MacDonald32 suggests that the process of what she labels ‘transformative unlearning’, which has much in common with deep unlearning, has three distinctive but overlapping steps. The key features of transformative unlearning that resonate with deep unlearning are that it involves questioning, reflecting on and giving up some core values, assumptions, knowledge and practices, and also that this process is deeply emotional and challenging for people to undertake. Similar to the mainstream perspective in the unlearning literature, MacDonald conceptualises unlearning as a necessary precursor to learning, and that both together are interlinked components of change. Finally, the catalyst for transformative unlearning is a process of change that brings a person's pre-existing values, assumptions knowledge and practices into question.

The first stage in MacDonald's model is receptiveness, in which a person accepts the possibility that there are perspectives and viewpoints that challenge his or her assumptions and that he or she is prepared to consider these perspectives. Following this is the second stage of recognition, which is the process through which a person acknowledges the veracity of these alternative viewpoints, and the limitations that exist in his or her own perspectives. Finally is the process of grieving, which she suggests is the emotional core of transformative unlearning, whereby a person comes to terms with ‘the loss of prior ways of seeing – the loss of fundamental assumptions which until now had brought certainty and security’ (p. 174). It is only after these three stages of the transformative unlearning process have been undertaken that a person is able to effectively change and learn new assumptions, knowledge and practices.

MacDonald's32 model can be illustrated by summarising the example she uses, which involves her own experiences as a practising nurse in relation to changes in the recommended sleeping position for infants following research on cot death, or sudden infant death syndrome (SIDS). In the 1970s and 1980s, the generally accepted and unquestioned practice was that babies were put to sleep in the prone position (lying on their front with the head to the side). This was a practice that MacDonald not only recommended to young mothers but also used with her own children. However, research in the early 1990s into the causes of SIDS suggested that putting infants to sleep in a supine position (lying on the back or side) reduced the risk of SIDS. Being aware of and reading this research represented the process of receptivity for MacDonald. However, this process was not quick or easy as it challenged ideas and practices she had taken for granted her whole working life. As more research emerged in the early to mid-1990s on the causes of SIDS, and as MacDonald read more of it, she entered the recognition phase in which she started to accept the validity of the new perspective on infant sleeping positions and began to accept the need to change her own assumptions and practices. This process was facilitated by dialogue with other nurses who were also going through the same process. However, before she was able to comfortably and fully accept the need for this change, MacDonald suggests that it was necessary to go through a process of grieving in which she came to terms with the fact that current research and advice suggested that her previous knowledge and practices had their limitations. This was argued to be the most emotional aspect of the unlearning process as it ‘touched the emotional core of my identity as a nurse’ (p. 174). As with the process of recognition, this process of grieving was facilitated by dialogue and communication with others.

Although this model of the process of transformative unlearning has resonances with Rushmer and Davies'6 concept of deep unlearning, it also differs from it in two key respects. First, the catalyst for transformative unlearning was external change rather than some specific incident or experience. Second, a more significant difference was the timescale over which transformative unlearning occurred. Although Rushmer and Davies suggest that deep unlearning involves a sudden and rapid realisation of the need for change, and an equally sudden process of change, in MacDonald's model transformative unlearning was a slow process that occurred over a period of years. As both transformative and deep unlearning involve the emotionally challenging process of unlearning fundamental assumptions and values, they represent comparable forms of unlearning. However, what MacDonald's perspective highlights is that not all deep unlearning occurs through the sort of sudden Archimedean epiphany which produces an instantaneous change in behaviour. Although particular events may lead people to question their values and assumptions, it may take more time for behavioural change to follow.

Two other process models of learning from the learning literature also have potential relevance to understanding the character of deep unlearning processes. First is Garud et al.'s58 narrative model for learning from what they call ‘unusual experiences’. Unusual experiences are defined as ‘situations that bear little or no resemblance to the types of experiences that have occurred in the past’ (p. 587). Although Garud et al. talk of learning rather than unlearning, there is the potential that, in making sense of unusual experiences, people's assumptions, values and practices may be brought into question and a process of individual unlearning may be undertaken, in which certain values and/or behaviours are abandoned and changed. However, although the narrative model that they develop for how people make sense of unusual experiences may have some relevance to understanding the dynamics of individual unlearning processes, a major limitation of their framework is the lack of consideration given to emotional issues. Despite acknowledging that making sense of unusual experiences involves reflecting on basic assumptions, and that dialogue with others in doing so may result in contrasting perspectives, issues of emotion are not considered. For this reason, Garud et al.'s model is not considered appropriate for understanding the dynamics of the process of deep unlearning.

A final very general model that is relevant for understanding the process of deep unlearning, which has some synergy with MacDonald's model, is the process of organisational inquiry outlined by Argyris and Schön,34 which takes inspiration from the work of the pragmatist philosopher John Dewey. Similar to deep unlearning, and Garud et al.'s58 unusual experiences, the catalyst for a process of inquiry is the experience of doubt, which is defined as the experience of a ‘problematic situation’, triggered by a mismatch between the expected results of action and the results actually achieved. Such experiences, they argue, inhibit continued action and encourage a process of reflection/inquiry aimed at resolving the doubt. Thus, the start of this process of inquiry can be considered equivalent to the initial stage of receptiveness in MacDonald's32 model, with the objective of resolving the sense of doubt that has been experienced, providing the primary catalyst to this receptiveness. The process of inquiry outlined by Argyris and Schön34 is relatively generic and lacking in detail but it involves combining reflection and action as well as an active process of dialogue with others. The ultimate aim of this process of inquiry is doubt resolution, whereby the mismatch that was experienced is made sense of. Arguably, this process of inquiry is equivalent to both the receptiveness and the grieving stages of MacDonald's model. Once the process of inquiry has resulted in the doubt that was experienced being resolved, this makes it possible for learning and change to occur, whereby people adapt their knowledge and actions to take account of the recently experienced mismatch.

Overall, therefore, MacDonald's three-stage model of dealing with transformative unlearning provides a useful template for understanding how people experience and make sense of what is referred to here as deep unlearning. However, as the only empirical illustration of this model is MacDonald's reflections on her own experience, further research is necessary to empirically test and evaluate this model before its validity can be established. As has been revealed by reviewing the literature on individual unlearning types, change can feature as the catalyst to individual unlearning and unlearning can trigger further change. Therefore, the following section briefly reviews the literature concerning individual-level transitions in change theory.

Reviewing the literature concerning individual-level transitions in change theory

When examining individual attitudes to and experiences of change there are two main literatures of relevance: resistance to change and emotion/attitude to change. The following reviews each of these literatures and considers how they may relate to individual unlearning.

Organisational change has been studied from many perspectives with the use of many methods. 59 In their literature review, Armenakis and Bedeian60 identified five themes in the change literature: content issues, which deal with the substance and nature of a particular change; contextual issues, which deal with forces and issues that exist in an organisation's internal and external environment; process issues, which examine the actions taken to implement change; criterion issues, which focus on outcomes from organisational change efforts; and affective and behavioural reactions to change. It is the last category that is of most interest to this study.

Resistance to change is a concept that is frequently discussed in conjunction with the unfreezing stage of Lewin's57 change model. It is often presented as a negative barrier to change, defined as a restraining force acting to maintain the current state of equilibrium. 61 Traditional approaches to resistance to change frame compliant behaviours as non-resistant and non-desirable behaviours as resistant. The early solution presented to overcome resistance was to increase employee participation. 62 However, Lawrence63 argued that this was an oversimplified solution to resistance and that change had both technical and social characteristics. It was a lack of attention to the latter that often led to resistance to change. Lawrence's view was that attitudes were important in how managers think about resistance to change and that expecting resistance to change from employees often delivered resistance.

In his review of the literature Foster59 observes that other scholars have conceptualised resistance to change as an emotional reaction rather than a purely behavioural one. For example, Argyris and Schön64 discussed resistance in terms of defensive routines and frustration, and Kanter65 refers to feelings of uncertainty and loss of control in understanding individual responses to change. Klarner et al. 66 argue that the literature concerning emotions during change can be divided into two parts. First, some studies in change management focus on employee behaviour as a result of emotions evoked during change. Second, other studies have used psychological stage models to explain the relationship between emotions and employee behaviour during change. Both of these parts are considered briefly below.

It has been argued that emotion drives the adaptive behaviour of employees in response to change. 67 This adaptive behaviour involves learning new strategies and being resilient to setbacks. 68 Several studies have considered the behavioural outcomes of employees who show either positive or negative emotions. It has been argued that failure to emotionally adapt to change leads to resistance among employees. 69 For example, resistance can lead to withholding of participation during change events. By contrast, Avey et al. 68 found that positive emotions could help individuals cope and could create support for the change. Positive emotions were found to increase the level of commitment and emotional engagement to the organisation, even in times of disruption. 70 However, Piderit71 has argued that dichotomising individual responses to change as either altogether negative or altogether positive is an unhelpful oversimplification. Piderit71 proposes that conceptualising resistance to change as multidimensional would help account for the possibility of ambivalence or mixed feelings towards change, and presents a tripartite attitudinal model of resistance to change that includes three components: emotional (affective), intentional (behavioural) and cognitive. She concludes that the term ‘resistance to change’ is inaccurate and unhelpful, and does not provide the best framework for understanding organisational change implementation. Ford et al. 72 add that, by assuming that resistance is necessarily bad, change agents have missed potential contributions that could help build awareness and momentum and eliminate unnecessary or counterproductive elements, increasing the likelihood of successful implementation of change. Both Piderit71 and Foster59 conclude that studies at the individual level would be better served using the term ‘response to change’. Further, Klarner et al. 66 argue that studies on emotions during change treat change as a snapshot event but would be better served by characterising change as a process, with the result that emotions could evolve during change. Klarner et al. 66 add that studies often focus on single change but neglect to consider the repetitive nature of change that occurs in many organisations and that emotion should also be examined as a continual construct.

Studies in human resource management and organisational behaviour have also attempted to use psychological stage models to explain the relationship between triggers for employee emotions, coping behaviours and change outcomes. 66 It has been argued that the trigger for the change process can produce strong emotions as well as the change itself. 73 Other studies have looked at the relationship between employee emotions and coping behaviour during change. Smith and Lazarus74 advocate psychological appraisal theory to explain this relationship, with the employee engaging with an appraisal process that considers the implications of the change for his or her personal goals and well-being. The appraisal process shapes the individual emotional response, which is translated into specific coping behaviour. These coping strategies can influence the organisational change outcome. 75 Positive employee emotions towards change can increase engagement, affecting the emotions of other team members. 76 By contrast, resistance to change is demonstrated by disengagement, which can hamper effective change. 68 Klarner et al. 66 contend that these stage models suffer from the same deficiencies as those of change management studies that have examined emotion. The studies tend to consider emotions as either positive or negative rather than considering multiple possible emotions during change. They treat change as a snapshot event and fail to consider how emotions may evolve during a change process, with different events during a change process triggering different emotions. Consequently, they call for more studies that consider the evolving character of emotions during a change process, as well as studies that consider employees' emotions during repeated change.

Reflecting on the literature presented in this section suggests synergies between individual unlearning and individual transitions from change. Several studies have proposed that a catalyst to unlearning can be an organisational change event. 2,6 Therefore, individual unlearning may be part of the adaptive behaviour process that employees experience as a result of change. Employees may be required to give up or abandon some previous strategies or knowledge as well as learning new ones, and this abandonment may trigger strong emotions. MacDonald32 argues that the unlearning process can involve giving up some core values, assumptions, knowledge and practices, which can be challenging and deeply emotional for people to undertake. Similarly, studies of individual transitions during change have found that employees report feelings of uncertainty, loss of control and frustration and enact defensive routines. 64,65 Positive emotions have been found to increase commitment and engagement with change, but may also demonstrate an increased willingness to engage with individual unlearning, seeing it as an opportunity to refresh old thinking and knowledge. Therefore, it is possible that employee emotions during change may also influence their engagement with unlearning processes, which may explain how employees respond and adapt to change.

Emotion has also been linked to decision-making processes, the main outcome of interest for this study, and the following section reviews the decision-making literature in relation to individual unlearning.

Reviewing the academic literature on decision-making

In reviewing the literature concerning individual decision-making and decision-making processes we examined the learning, knowledge management and organisational behaviour fields. By searching for literature published from 2005 we identified 29 articles that considered learning and decision-making in the article title, but we found none of these articles to be directly relevant to this study. Similarly, since 2005, 11 articles had been published that included knowledge management and decision-making in their titles, but these also did not provide any further insights for this study. Neither field provided any new material regarding unlearning and decision-making. However, searches in the organisational behaviour field were more successful, especially concerning work psychology, which is most relevant to this study because of its focus on the individual. Consequently, the following discussion draws heavily from this discipline.

Much of the work psychology literature on decision-making has considered how individuals deviate from strictly ‘rational’ processes, with ‘rationality’ defined as choosing the option that has the highest expected value among those options available to the individual. 77 However, this definition assumes that all other things are equal, and Tetlock78 has argued that in different circumstances individuals may have different motives and goals for a decision and that the attractiveness and utility of certain decisions is subjective to individuals' perceptions of these possible outcomes. Arnold77 states that there are many different strategies for decision-making, such as evaluating options using a very limited range of important criteria, or searching for an option that an individual considers to be good enough. This latter approach has been termed ‘satisficing’ and can be contrasted to ‘maximising’ in which a thorough search is made for the best possible option. Arnold77 argues that a satisficing strategy may be suitable if a person does not have time and/or the ability to use a more thorough approach. Frisch and Clemen79 argue that, although attention to the perceived attractiveness and utility of a decision outcome is useful, attention should also be given to the early stages of the decision-making process. In particular, they argue that to understand decision-making well it is important to consider consequentialism (the expected personal consequences of making a decision), thorough structuring (considering more than one option and the likely consequences) and compensation (the trade-offs between one benefit and another from making different decisions). These three influences on decision-making highlight the key role of information processing in the decision-making process.

There are several different aspects to information processing in decision-making such as attitudes to uncertainty and risk, gains compared with losses and framing of information. Although these aspects are relevant to decision-making we identified four areas that may have particular significance for individual unlearning and process management decision-making. The first of these areas was the role of emotion, action and inaction. Isen80 provides evidence to suggest that individuals who are experiencing positive emotion are able to be more creative and are more able to process information than those who are not. However, it appears that positive emotion can also lead individuals to be unrealistically optimistic about future events and outcomes. 81 Further, individuals seem likely to consider future emotions when making decisions and, in particular, the emotion of regret. It is argued that individuals regret actions more than inactions, probably because the consequences from actions are more tangible. In line with these results, Baron82 found that worries concerning future regrets encouraged inaction over action. This may suggest that managers could choose not to abandon certain behaviours or knowledge even if they feel that they are outdated or not well suited to the current environment if the alternative is a step into the unknown.

The second area was escalation of commitment. This phenomenon is now well established77 and emphasises that, if a decision is made freely and explicitly, the individual making it is committed to it and retrospectively seeks to find reasons to justify his or her rationale for making it. McCarthy et al. 83 report evidence of this escalation of commitment in some managerial decisions. This phenomenon is important for managerial decision-making because laboratory experiments have shown that, when decisions are reported to individuals as unsuccessful, they will feel compelled to continue to support their decision in the belief that they will ultimately prove that their wisdom was correct. Simonson and Staw84 investigated how to counter this compulsion and found that decision-makers who set advance minimum outcome levels were less likely to pursue the course of action if the minimum outcome had not been reached than decision-makers who had not specified a minimum outcome level. Escalation of commitment was also reduced if decision-makers were told that it was the quality of the decision-making process that would be evaluated, not the outcomes achieved (which in any case may be beyond the control of the decision-maker). Such strategies may have value to managers when assessing the quality and effectiveness of their decision-making by providing indicators of when existing decision-making practices should be reviewed and potentially abandoned.

The third area was heuristics. Heuristics are rules of thumb that people use to simplify information processing and decision-making. 77 In their seminal work, Tversky and Kahneman85 identified three common heuristics. The ‘representativeness’ heuristic refers to when an individual judges something purely according to how representative it appears of a particular category, regardless of the probability of it actually belonging to that category. For example, if an individual encounters three medical students who are all aggressive and demanding, the individual may conclude that all medical students are aggressive and demanding, regardless of the probability of this actually being the case. The ‘anchoring bias’ heuristic refers to a person's failure to change his or her views as much as he or she should in the light of new information, for example the difficulty an interviewer may have in dislodging the first impression of a candidate over the course of an interview. The ‘availability’ heuristic concerns a person's tendency to consider an event more probable if it can be easily imagined than if it cannot. For example, after seeing news reports about redundancies in local government, a manager in the NHS may believe that he or she is in danger of being made redundant, regardless of whether or not there is any genuine risk in their sector. Such rules can therefore skew individual decision-making processes and risk poorly founded decisions being implemented. The process of unlearning may provide opportunities for managers to become aware of the influence of heuristics on their decision-making processes, abandoning old rules in favour of more structured information-gathering procedures, enabling more objective decision-making.

Finally, the fourth and final area in which information processing can affect decision-making was evidence-based management. Evidence-based management is defined by Briner et al. 86 as follows:

Evidence based-management is about making decisions through the conscientious, explicit and judicious use of four sources of information: practitioner expertise and judgement, evidence from the local context, a critical evaluation of the best available research evidence, and the perspectives of those people who might be affected by the decision.

p. 19

They argue that evidence-based management is a relatively new concept, although the use of research evidence to help make managerial decisions is well established. Reay et al. 87 observe that many of the calls for the increased use of evidence-based management reference the more mature discussions in the medical and health-care literatures, which promote the use of the ‘best available evidence’ in medical decision-making.

There is still considerable debate concerning the value of evidence-based management for improved decision-making. In the review of the literature by Reay et al.,87 it was found that much of the existing research lacks strong evidence from which to draw conclusions. Reay et al. 87 also concluded that research had so far failed to provide clear evidence to determine whether or not employing evidence-based management improved organisational performance. There is also debate concerning what constitutes suitable ‘evidence’ on which to base management decisions. Evidence-based medicine draws heavily on evidence gathered using rigorous methods such as systematic reviews and randomised controlled trials. However, Briner et al. 86 caution against following a blind adoption of a big science perspective that privileges randomised controlled trials and meta-analyses above all other kinds of research evidence. This is partly because of a lack of systematic reviews in management or organisational studies for managers to draw on, but also, as Walshe and Rundall88 note, because the constrained, contested and political nature of many managerial decisions may make the application of research evidence difficult for managers, even when it is available.

A further dimension that has attracted growing attention in evidence-based medicine research has been investigating how to effectively translate research findings to practice settings. Reay et al. 87 argue that this area of research, known as knowledge translation, has potential as a model for evidence-based management. Knowledge translation research in health care focuses on understanding how complex interventions, based on research evidence and involving multiple providers, recipients and organisations, can be implemented. What has become clear from knowledge translation research is that the existence of evidence is critical but not sufficient to guide action. Despite the existence of evidence to support a particular behaviour change, Kitson et al. 89 found that changing practitioner behaviour remained difficult. Reay et al. 87 believe that innovative and creative methods are required to bring evidence to managers to overcome barriers to change. Consequently, it is possible that the inability of a manager to engage with processes of unlearning may help to explain why some managers resist changing or abandoning established behaviours and practices, despite the availability of evidence to support the change.

Furthermore, the shortage of systematic management research evidence compared with medical research evidence, and the increased political and constrained aspects of management decision-making, may mean that clinicians moving into managerial roles are not able to translate established evidence-based medicine practices into effective strategies for management decision-making. Walshe and Rundall90 highlight the divide between the biomedical and the social science evidence bases that clinicians and managers are likely to draw on. One is experimental, replicable and generalisable and the other is non-replicable and contextual. As a result, clinicians who become health-care managers may feel that they have to abandon or give up their established views of what constitutes reliable evidence and learn to use new forms of evidence for their decision-making. This type of change is illustrated by recent work investigating the use of evidence in NHS commissioning processes. 90 Swan et al. 90 concluded that commissioners needed to view evidence as something that is coproduced rather than being purely external and independent. For example, they argue that knowledge of the local context may be just as relevant as the latest research-based findings or policy guidance. They also argue that commissioners needed to recognise the limits of medical and scientific models for effective management decision-making and that evidence-based medicine templates could not be simply transferred to management decision-making processes.

The decision-making process

Several authors have proposed various conceptualisations of decision-making processes. For example, Briner et al. 86 suggest that an evidence-based management process could comprise five steps. First is the manager's attempt to define or state the problem or issue. Second, having stated the problem, internal organisational evidence is collected to check the problem relevance and validity, leading to a reformulation of the problem as necessary. Third, external evidence from published research about the problem is identified and critically appraised. Fourth, the views of stakeholders and those likely to be affected by the decision are considered along with ethical considerations. Fifth, having critically appraised all of these sources of information, a decision is made that attempts to take into account these four sources of information. Alternative conceptualisations of a decision-making process have been presented by several authors although, as Nutt91 observes, these conceptualisations differ significantly, which makes integration of the findings difficult. For example, Dean and Sharfman92 classify the decision-making process through procedural features of rationality, political behaviour and flexibility. Others focus on coalition formation and decision-maker attributes such as tolerance for ambiguity, uncertainty or risk aversion. 93 However, Nutt91 argues that, although these studies are helpful, they do not explain how decisions were made.

To address this weakness, Nutt91 compared four different types of decision processes to investigate their adoption, value and duration. First, based on interpretive studies, Nutt argues that decision-makers cater to the interests of powerful stakeholders by looking for ready-made solutions that they can support. These solutions provide a clear course of action leading to an outcome that can easily be visualised, a process called ‘idea imposition’. Second, by contrast, the literature calls for decision-makers to gather intelligence about their needs, discover and evaluate options according to their benefits, and implement the chosen decision by working with relevant interest groups. This process is called ‘discovery’ because it calls for learning about possibilities. Nutt also investigates two hybrid processes to decision-making. If an idea presented in an idea imposition process seems ill advised then the decision-maker might change to a ‘redevelopment’ process to find a replacement. If a discovery process throws up an ‘emergent opportunity’ then the discovery process may be abandoned if the opportunity seems beneficial.

Nutt91 found compelling evidence to support the discovery process model of decision-making with little empirical support for the idea imposition process. He also found that the discovery process was more successful than either of the hybrid approaches. Nutt concludes that the sequencing and steps in decision-making matter. The most important steps include to investigate needs, conduct a sensitive stakeholder implementation and set a direction. Further, these steps should be conducted early in the decision-making process. There was an emphasis on logical and political rationality with expectations needing to be set and politics managed early in the process. Decision-makers who preferred idea imposition seem to be keen for action, the motivation for a quick solution being motivated by pragmatics and fear. However, Nutt argues that such quick fixes can create a trap that limits search and discourages knowledgeable people from offering ideas. In addition, the opportunity can be seen as a personal idea, raising questions about the decision-makers' motives and requiring the rapid action to be delayed while the decision-maker formulates a defence.

The attractiveness of remedy-driven solutions that seem compelling may be due, in part, to a decision-maker's desire to remove ambiguity. By following this path, sources of ambiguity are swept away, giving temporary relief but stopping the search for intelligence and alternatives. The approach also distracts decision-makers away from contemplating their needs, and draws decision-makers to document their ideas and the stakeholder response. Nutt91 concludes that accepting ambiguity is a necessary motivator to effectively address difficult and troubling issues. This transition may well be an important aspect for managers who have moved from clinical roles with established evidence-based decision-making practices into managerial roles that have a less well-defined evidence base and involve higher levels of ambiguity. Managers in these roles may initially adopt idea imposition as a process of decision-making to remove the ambiguity associated with their new positions. If, as Nutt's research suggests, this method of decision-making proves unsuccessful, this experience may act as a catalyst for unlearning this approach and adopting a more discovery-orientated approach. As a key antecedent of individual unlearning has been found to be change at an individual, an organisational or a policy level, the following section reviews recent major changes in the NHS to provide a contextual background to the study.

Recent changes in the National Health Service

Over the last 40 years, significant organisational changes have taken place in the NHS. Major reforms in, for example, general practitioner (GP) fund holding, NHS organisations moving to trust status, governance management and patient involvement have meant that staff frequently experience major structural and organisational transformations. The introduction of ‘formal’ management positions into the NHS followed the Griffith Report in 1983. 94 In addition to recommending the introduction of general managers into health authorities, the report also focused on specific management training for NHS staff, particularly for those with a medical background. 95 Since the Griffith Report, changes in the health service have been accompanied by the introduction of many formal management positions. Managers from both clinical and non-clinical backgrounds now occupy many of the crucial senior roles within the health service, bringing about local organisational change and reform.

The pace of change in the NHS has continued to be significant in recent years. For example, Greener7 talked about how the ‘pace and scope of change’ was ‘intense’ (p. 206). Most recently, the introduction of the health and social care reforms for the NHS in England in 2012–13 will require many NHS staff to further adapt in their roles and responsibilities. In summary, the reforms have meant that almost every GP practice across England has joined with neighbouring surgeries to form clinical commissioning groups (CCGs). The 152 local NHS primary care trusts (PCTs) that managed most of the NHS budget in England have merged at a senior managerial level to form 51 clusters. Some staff will stay on after PCTs are abolished in 2013 in organisations that will continue to advise the new GP-led groups on how to plan and buy health care. A new NHS Commissioning Board that was established in shadow form in October 2011 will undertake specialised care that needs to be planned at a national level. Below the board will be four regional hubs and then beyond that 50 local offices. From 1 April 2013, strategic health authorities and PCTs will be abolished, and the CCGs and National Board will take on their full functions. 96,97

Thus, it is anticipated that, because of the amount of ongoing change in the NHS, all health-care managers are likely to have experiences of unlearning processes such as Rushmer and Davies'6 categories of wiping and fading in which they have had to adapt their work practices and routines as a result of change. Rushmer and Davies' category of deep unlearning on the other hand may be experienced by health-care managers who have undergone a significant role transition, such as would happen when a clinician moved into a managerial role. Thus, it is possible that processes of deep unlearning may be more evident in managers who have undergone such transitions. To examine these aspects in more detail we developed the following research questions for the study.

Research questions

The overall research aim of this project was to empirically investigate the extent to which health-care managers engage in processes of individual unlearning and the barriers that exist to them doing so. This is important because it will improve our understanding of how managers make decisions and will help provide relevant support and training for those involved in running the health service. An addition was made to the second research question from the outset of the project to investigate enablers as well as barriers to unlearning. This change was made to ensure that the study captured a full understanding of the factors influencing health-care managers' unlearning processes.

The specific research questions investigated were:

1. To what extent do health-care managers engage with the process of unlearning and what impact does the engagement with this process have on health-care managers' decision-making?

2. What are the barriers and enablers that influence the engagement of health-care managers with the process of unlearning?

3. To what extent does the health-care setting affect the level of engagement with the process of unlearning by health-care managers?

Having provided a comprehensive review of the literature and provided a brief summary of the key changes in the NHS that were being implemented at the time of data collection, in Chapter 3 we explain the methods adopted for the study and the process of data collection and analysis.

Organisation of the study

Coombs took overall responsibility for the management of the study, with support from Hislop and Holland. Coombs had experience in undertaking health-care management research projects using multisite case studies and managing research associates. Hislop provided the project with subject area expertise and project management support for Coombs and managed the data collection and analysis. Holland managed the liaison with NHS trusts, contributed to the data analysis and facilitated the workshops. A research associate, Bosley, was appointed and commenced work in January 2011. In late 2011, Bosley was appointed to a lectureship at the School of Business and Economics, Loughborough University, Loughborough, UK. A new research associate, Manful, was appointed on a secondment from the Loughborough University Centre of Research into Social Policy and commenced work in January 2012. It was agreed that Bosley would remain part of the research team to ensure continuity for the project and because unlearning was a research interest that Bosley wished to follow. The project team had monthly meetings to monitor progress and additional regular supervisory meetings with the research associate.

Choice of research strategy

Unlearning is a relatively neglected and unexplored aspect of learning and change processes. A case study-based approach was identified as a suitable method because qualitative case studies provide an effective way to conduct exploratory research that can generate rich, qualitative insights into managers' experiences of unlearning. 15,16 Given the limited literature on and knowledge of the concept of unlearning, the case study design enabled an inductive approach to theory-building rather than seeking to test theoretical hypotheses. The intention was to investigate the experiences of unlearning in three different types of NHS trust (an acute trust, a PCT and a mental health trust) to explore the extent to which the organisational context affects unlearning. The reason for choosing these three types of context is that they represent some of the most important and common organisational types that the managers being investigated work in.

Unlike some case study designs, the research does not draw on a range of different data sources but focuses on individual interviewees' experiences of episodes of unlearning. In other words, the unit of analysis is the episode of unlearning experienced by the individual manager rather than the individual manager or case study site, both of which provide a context for comparison. For example, it was possible to compare experiences of episodes of unlearning between managers with clinical backgrounds and those with non-clinical backgrounds, and also between managers working in acute health-care environments and those working in mental health-care environments. Our main source of data was the individuals themselves, who provided descriptions of their job roles, immediate work context and the barriers and facilitators of unlearning, some of which related to the department, NHS trust or wider policy context. We justified this on the grounds that our interest was individual rather than organisational unlearning, which can be illuminated only through one-to-one interviews.

Selection of case study sites

The intention was to recruit an acute trust, a mental health trust and a PCT. Given that participation involved 14 managers being released for 1 hour to be interviewed, a key criterion for inclusion was securing senior management support and willingness to commit some resource to assist in recruitment. Holland used her NHS contacts to identify interested NHS trusts and to secure initial written agreement to take part. The acute and mental health trusts were identified relatively quickly using this approach, but the PCT that initially agreed to participate later withdrew. The engagement of another PCT was secured in March 2011 but, as described in Recruitment at site C: primary care trust/clinical commissioning group, recruitment at this site failed to generate any participants. Consequently, the results of the study are based on two rather than three case study sites.

Ethical approval

Loughborough University gave ethical approval for the conduct of the exploratory interviews on 29 March 2011. The main study did not require approval from the University Ethical Advisory Committee because it was subject to NHS ethical clearance. Application for NHS ethical approval and NHS R&D clearance was made through the Integrated Research Application System. Notification of clearance by the National Research Ethics Service Committee [. . .] Proportionate Review Sub-Committee was received on 10 May 2011. (The name of committee was not included to ensure anonymity.) The following documents were provided to the committee along with the completed NHS ethics application form:

• summary CVs for the chief investigators

• letter of invitation to participants, participant information sheet (PIS), consent form, interview schedule, pre-interview questionnaire, examples of unlearning

• transcriber's confidentiality agreement

• letters from the sponsor and funder, professional indemnity and employers' liability certificates

• the protocol, summary flow chart of the protocol, referees' assessment form and comments on the research proposal.

Application for NHS R&D clearance was handled through the Co-ordinated System for Gaining NHS Permission (CSP). Documents submitted for this stage were the research protocol, summary flow chart of the protocol, letter from the funder, summary interview schedule and pre-interview questionnaire. Permissions were secured from case site B on 9 June 2011, from case site A on 15 June 2011 and from case site C on 24 November 2011.

Data collection

Data were collected through one-to-one semistructured interviews with health-care managers. A semistructured approach enabled the interviewees to respond to the researcher's questions using their own words and to introduce and discuss new topics, issues and experiences that they deemed appropriate. 98

Interviews were framed by a topic guide that was informed by a typology of unlearning proposed by Rushmer and Davies6 and used critical incident technique. 99 Interviewees were invited to identify specific incidents of unlearning and fading that they regarded as significant and could clearly recall, along with identifiable antecedents and consequences. They were invited to identify one incident each of fading (the gradual forgetting of skills/capabilities no longer required), wiping (deliberate efforts to unlearn specific beliefs and/or behaviours) and deep unlearning (sudden and unexpected unlearning that challenges deeply held beliefs and assumptions). Using Kemppainen's100 three-component approach, questions were asked about the circumstances and context of the incident being discussed, the specific details of how they experienced and remembered the incident being examined and their understanding of the consequences or outcomes resulting from the incident.

Pilot interviews

To test the topic guide, nine formative, exploratory interviews were conducted with health-care managers who were or had recently been studying part-time on Masters programmes in Medical Management and Healthcare Governance in the School of Business and Economics at Loughborough University. These interviews took place between the end of March and early May 2011. All participants were sent an information sheet, an interview schedule and examples of the different types of unlearning in advance of the interview. All were invited to complete a pre-interview questionnaire providing background about their previous work and qualifications, although not all did this. All gave either written consent or oral consent, which was digitally recorded before beginning the interview.

The pilot interviews showed that the research documentation worked satisfactorily but also drew our attention to the need for participants to have an opportunity to discuss the concept of unlearning and to read the documentation in advance of the interview. The interviewer also gained from the experience by developing skills and supplementary questions to tease out unlearning experiences from the interviewees.

Recruitment of interviewees for the main study

The research team identified and worked with a key contact and human resources manager at each site. The key contacts were sent a briefing sheet about the project (see Appendix 3) and then the principal investigator and research associate arranged to meet with the contact and a human resources manager. At this meeting the aims and time frame of the project were explained, any questions and issues were addressed and the most appropriate method or methods of recruitment were identified. During the revisions to the original project proposal it was identified by the Department of Health through the Health Services and Delivery Research (HS&DR) programme that the employer of NHS staff who participated in the study should be compensated for the loss of staff resources. Consequently, a payment of £141 per 1-hour interview was offered to each case study site. This figure was based on the rate used in clinical trials to calculate a doctor's time. The payment was not passed directly to participants. The key contact and human resources manager helped the research team identify the target group and the most effective method of recruitment. For this reason, different recruitment strategies were used at different sites.

Inclusion criteria

We envisaged that health-care managers who may have had to adapt their work practices and routines as a result of policy changes, responded to new research on clinical effectiveness or undergone a significant role transition, such as moving from a clinical to a managerial role, were most likely to be found in middle management roles. Therefore, for the main study, the aim was to identify a purposive sample of 14 middle managers at each case study site. Given the definitional issues surrounding the term ‘middle manager’ and different usage in different organisational structures, the research team acted on the advice of the steering group and sought managers working at Agenda for Change bands seven and eight. Thus, participants needed to (1) be working at Agenda for Change bands seven or eight or be GPs or consultants; (2) have some managerial responsibilities that involved making decisions about the use and allocation of resources, such as staff and/or finances, data collection and management or service delivery; (3) be able to identify and be willing to talk about their experiences of unlearning; and (4) be willing to complete the pre-interview questionnaire, participate in an interview of approximately 1 hour in length and give informed written consent to the conditions of participation.

Workshops

The final part of the project was to conduct feedback workshops with interviewees from the participating sites. The aim of the workshops was to provide an opportunity for additional data gathering. The workshops had three main objectives: to feed back preliminary results from the study; to gather participants' views on the preliminary findings; and to discuss the practical implications of the findings with participants in terms of how the information could be used within the sites. Each workshop was attended by the principal investigator (Coombs), workshop facilitator (Holland) and Bosley (site A) or Hislop (site B). Hislop and Bosley captured the key points from each workshop on flip charts that were transcribed for analysis.

The feedback workshops were held on the 15 March 2012 and 12 March 2012 in sites A and B respectively. All interviewees were invited to attend the session at their site. In addition, after discussion with our key contacts at both sites, a number of additional staff who were interested in the study but who had not participated in the interviews were invited to attend the workshops. These staff included middle managers, research managers and, at site B, senior trust management. Nine health-care managers, including three health-care managers who did not participate in the interviews, attended the workshop at site A. At site B there were 20 participants, including 11 health-care managers who did not participate in the interviews.

Data analysis

Data analysis followed the three concurrent activities identified by Miles and Huberman101 of data reduction, data display and conclusion drawing/verification (p. 10). Both deductive and inductive approaches were used. A deductive approach was used initially to reduce the data. A coding frame was developed based on the research questions and concepts being examined and applied to the data collected in the exploratory interviews. The research team subsequently tested the coding frame with each member independently coding two transcripts selected from the corpus of the data and comparing and discussing their coding decisions with the whole team. This led to a further iteration and additional testing of the coding frame.

To promote consistency the research associate coded all transcripts. The research associate used an inductive approach to code findings from the interview data. Thus, she added further codes and split and spliced codes as she worked on new transcripts. Team members reviewed coding and conducted periodic queries in NVivo 10 (QSR International, Southport, UK) to test and validate coding decisions.

Using the software package facilitated the coding process, enabling the research associate to compare and revisit codes during the coding process and thus ensure consistent application and identify codes that could be split or spliced. NVivo also enabled quick retrieval and comparison of data segments allocated the same code and the construction of conceptually ordered displays (data display). Such displays or matrices enabled the researchers to study the variables in more depth and generate more explanatory power. The matrices were created through a series of iterative steps, with initial drafts created by the research associate with the initial and subsequent drafts reviewed and discussed by the project team. This approach ensured consistency and validity when categorising data such as the nature of unlearning episodes, barriers, enablers and impact on decision-making processes. As the draft matrices were often large and covered several pages, an example of part of a draft matrix is shown in Table 6.

The data from the workshops were analysed using content analysis to derive the key themes that emerged from the meetings. Because of the method of data collection, the level of analysis of the workshop data was largely descriptive, but it does provide some interesting additional perspectives for the study.

Description of case study sites

Data about the case study sites were gathered from NHS trust websites, annual reports and informal discussions with site contacts. Data from these sources were used to construct the case study descriptions. As the case study sites (like other NHS trusts) were subject to considerable change during the study period, it was difficult to construct accounts that accurately represented them. The site descriptions are intended to provide the reader with sufficient detail to make sense of the findings while protecting the anonymity of the participating NHS trusts.

Episodes of individual unlearning and fading

The first research question for this study was to investigate the extent to which health-care managers engaged with the process of individual unlearning. This section reports how individual experiences or change events resulted in participants experiencing behavioural and cognitive unlearning. Also in this section we present examples of responses relating to episodes of fading. From the interview data we identified 85 episodes of individual unlearning and fading: 29 behavioural unlearning episodes, 28 cognitive unlearning episodes and 28 fading episodes. Based on the new unlearning typology, Table 8 provides a summary of the types of knowledge that had been unlearnt by participants in the study. The table shows that the types of knowledge reported as being unlearnt by participants were varied, with the most common knowledge unlearnt being concerned with procedures and practices for both behavioural and cognitive unlearning. These episodes are reported in more detail in the following sections before considering episodes of fading.

Barriers to individual unlearning

The second research question of this study focused on investigating the barriers and enablers that influenced health-care managers' level of engagement with the process of unlearning. This section addresses the first part of that question. Overall, participants mentioned 36 examples of barriers encountered during episodes of behavioural unlearning, but mentioned fewer examples of barriers encountered during cognitive unlearning episodes (n = 19). Some participants mentioned more than one barrier to their unlearning or fading episode, which is why the number of barriers identified is greater than the number of unlearning episodes described in the previous section. Table 9 shows that the most common forms of barrier identified were staff resistance to changes in practice, personal behaviour, lack of management support, high workloads and organisational culture. Other barriers that were mentioned included financial constraints and job insecurity. In addition, we also identified 13 examples of barriers to fading, the most common of which was the ability to recall past knowledge and experience. The following explains the nature of the barriers identified by participants and the types of unlearning they were associated with, before considering barriers to fading.

Enablers of individual unlearning and fading

This section examines the factors that managers perceived enabled them to engage with the process of unlearning or fading. With the exception of one factor, all of the enablers mentioned by participants could facilitate either behavioural or cognitive unlearning. As was the case when discussing barriers, participants often identified multiple enablers to particular unlearning episodes, resulting in a higher number of enablers than individual unlearning episodes. As shown in Table 10, a total of 48 and 49 examples of enablers of behavioural and cognitive unlearning were identified respectively.

By contrast, participants identified a greater variety of enablers facilitating fading episodes. Participants identified 32 examples of enablers for fading, the most common being the nature of the current job role, introduction of new practices and knowledge, development in technology and infrequent use of skills.

The impact of individual unlearning and fading on managers' decision-making

In this section we illustrate how the process of individual unlearning can impact on health-care managers' decision-making. Table 11 provides a summary of the types of impact on decision-making that participants identified. From the data analysis 30 examples of changes in decision-making were identified of which 19 were related to individual unlearning triggered by change events and 11 were related to unlearning episodes triggered by individual experiences. Six participants did not give any examples of how the individual unlearning process had impacted on their decision-making. From the data we identify that the process that leads to each unlearning type results in different changes in decision-making. Behavioural unlearning resulted in more collective and decisive decision-making, adaptation of management style and more effective reporting plans. Engagement with cognitive unlearning resulted in more rational decision-making, a more negotiative management style, changes in attitudes towards work and changes in procedures.

None of the participants was able to provide an example of fading having an impact on their decision-making. This is likely to be because of the difficult nature of investigating the fading concept as it requires participants to recall examples that they have, by definition, forgotten and then attempt to recall whether or not this forgotten knowledge or behaviour altered their decision-making processes.

The following sections present examples of decision-making changes related to each individual unlearning type according to the trigger (individual experience or change event) associated with the unlearning.

The influence of health-care setting and professional background on health-care managers' individual unlearning and fading

The third research question of this study was to investigate whether or not the health-care setting of managers influenced their level of engagement with the process of unlearning. Therefore, the interview data were analysed to assess whether or not there were any significant differences between data collected from different case study sites in terms of unlearning episode type, enablers and barriers. In addition, we were also interested to examine whether or not there was any empirical evidence of health-care managers' professional background influencing their level of engagement with unlearning. The following sections present the results of our analysis regarding each of these questions.

Workshop findings

An important aspect of our original project plan was to undertake a feedback workshop at each of our case study sites. The workshop objectives were to feed back preliminary results from the study, to gather participants' views on the preliminary findings and to discuss the practical implications of the results with participants in terms of how the information could be used within the sites. Details of the administration of the workshops and participants are provided in Chapter 3.

Summary of findings

This chapter has presented the findings from our analysis of our interviews with health-care managers at two case study sites and the findings from workshops conducted at both sites. The previous sections have shown that health-care managers who participated in this study engaged with both forms of individual unlearning and fading. The engagement with individual unlearning was triggered through either an individual experience or a change event and in some cases these triggers were similar. For example, a change in role could be initiated by an individual or by a wider organisational change. Participants also referred to the greater emotional impact often associated with cognitive unlearning than with behavioural unlearning, and this was a strong theme that also came from the workshops. There was little evidence to indicate that health-care setting or professional background had a strong influence over whether or not health-care managers participating in this study engaged with individual unlearning. However, the small sample size and similar characteristics of the two case study organisations must be kept in mind when considering this finding.

Participants identified a variety of barriers to individual unlearning and there was some evidence of a possible relationship between the trigger that initiated the individual unlearning and managing the associated barrier. For example, if the trigger was an individual learning episode, managers may need support in building confidence, the challenging of their need to be liked, relinquishing control and breaking habits. On the other hand, if unlearning was triggered by a change event, some action may be needed at an organisational level, for example to encourage senior management support and address concerns about workload and job insecurity.

With one exception the same factors enabled behavioural and cognitive unlearning. Overall, personal skills, attitudes and relationships were more commonly cited than organisational enablers such as policies, procedures and work circumstances. More enablers were identified in relation to individual unlearning triggered by individual experiences than in relation to unlearning triggered by change events.

The findings also suggest that engaging with individual unlearning did seem to have some impact on the decision-making of health-care managers who participated in this study, although the nature of the impact seems to vary depending on an individual's circumstances. However, a positive finding was the associated improvement that managers inferred having engaged with individual unlearning. In several cases participants felt that the effect of individual unlearning on their decision-making was that they were able to work in a more effective way, for them personally and/or in their organisational role.

Fading episodes were often triggered by a change in role and resulted in the forgetting of particular skills or knowledge. However, many participants felt that they could still recall some core knowledge associated with tasks and it was specific details that had most commonly been forgotten. In addition, no participants indicated that they had made a conscious action to forget particular skills or knowledge; these specific details had simply faded from memory. This was reflected in the barriers and enablers that were identified, with reference being made to the ability to recall (barrier) and ingrained experience (barrier) or external change (enabler) and infrequent use (enabler) when discussing episodes of fading. We found no evidence of fading having an impact on participating health-care managers' decision-making. However, this may also be partly a reflection of the challenge of capturing this possible change for participants, as it requires them to remember skills that they had forgotten. Similarly, for both types of unlearning, we did not find strong evidence to suggest that health-care setting or the professional background of participants could differentiate experiences of fading, although this finding is also subject to the same limitations. The following chapter discusses these findings in more detail and in reference to the existing literature regarding unlearning.

The extent to which individual unlearning and fading occurred and the nature of individual unlearning experiences

This section focuses on the first part of the first research question of the project, the extent to which health-care managers engage in unlearning. It is divided into two parts with the first part analysing the data related to behavioural and cognitive unlearning and the second part examining fading.

The barriers to and enablers of individual unlearning and fading

As noted in Chapter 2, the literature focuses on organisational unlearning and the antecedents of unlearning, identifying unlearning as facilitated by organisational context,9 environmental turbulence3,51 and organisational size. 26 Only one article32 explores the process of individual unlearning and in so doing identifies receptiveness and recognition as necessary conditions for unlearning to take place. Antecedents – pre-existing conditions that enable or prevent the unlearning from beginning – may be the same as or distinct from the enablers of or barriers to the unlearning process. Less attention has been given to the latter, although some studies offer relevant insights from empirical studies. 9,10 As discussed in Chapter 2, when antecedents equate to ‘triggers’ (individual experiences or change events that initiate individual unlearning) they are integral to the definitions of the different types of unlearning that we have identified. However, the discussion that follows considers enablers and barriers in relation to both initiating and engaging in the process of individual unlearning and fading.

The impact of individual unlearning and fading on managerial decision-making

An important aspect of our study was investigating the extent to which managers' experiences of unlearning impacted on their decision-making. In general terms, as outlined in Chapter 4 (see The impact of individual unlearning and fading on managers' decision-making), we did find some evidence that health-care managers' individual unlearning experiences impacted on the way that they made decisions, although there was no evidence of fading having an impact. However, it should be noted that our participants often had difficulty relating their unlearning and fading experiences to specific changes in their decision-making processes. Out of the 85 episodes of unlearning and fading, we were able to identify only 30 examples of changes in decision-making, none of which was linked to fading episodes. Consequently, we can offer only relatively tentative insights concerning the relationship between unlearning and health-care managers' decision-making. Before reflecting on these findings, this section begins by looking back at the literature review on unlearning that was presented in Chapter 2 to highlight the nature of the contribution to this literature that our empirical data make.

As outlined in Table 3, of the 34 articles on unlearning that were reviewed, 13 examined the consequences of unlearning. Of these 13, five were conceptual articles and only eight presented original empirical data on the outcomes or consequences of unlearning. Furthermore, of these eight articles, only three examined individual unlearning. 37,38,48 In all three articles the empirical data presented are taken from survey-based studies and no qualitative data are used; thus, there are very few empirical data on the consequences of individual-level unlearning.

Therefore, our exploratory empirical findings on the impact of individual unlearning on managers' decision-making processes make a contribution to knowledge in two respects. First, our study involves a type of research data not previously utilised in this domain – qualitative interview data that give insights into how individual managers understand the changes that their unlearning experiences have produced. Second, it also looks at a variable – managerial decision-making – that has also not been examined in other unlearning studies.

The analysis of our research data found that the likelihood of people's unlearning experiences producing changes in the way that they made decisions was typically the same for all types of unlearning examined. Thus, for both types of behavioural unlearning and for cognitive unlearning initiated by a change event, changes in managers' decision-making behaviours were identified in approximately 50% of cases. The exception to this pattern was cognitive unlearning initiated by an individual experience, for which changes in health-care managers' decision-making were identified in just under one-third of episodes. A possible reason for this is that change is less likely with this type of unlearning because of the combined effect of it being a challenging type of unlearning (cognitive) and the individualised nature of it. The individual aspect means that people typically experience it in isolation from others and may have limited opportunities to share their experiences with others and learn from them.

Changing focus to look at the type of changes in decision-making that were identified, a number of observations can be made. First, although the decision-making changes that people reported undertaking were described as being positive, in general terms the scale of the changes undertaken was relatively small and incremental. For none of the unlearning episodes examined could it be said that they produced a radical transformation in the way that people thought or acted. Thus, a typical example of the scale of change in decision-making that people's unlearning experiences produced was when one manager reported taking more account of the fact that there were limitations to his ability to make decisions because of the constraints placed on him by more senior colleagues. This is in line with the observations of Walshe and Rundall88 that managerial decision-making is often more politicised and constrained than decision-making in medical practice.

In broad terms, the types of changes in decision-making that were identified can be grouped into two primary categories. The first concerns changes in how people consulted and communicated with colleagues regarding the implementation of change (such as adaptations of participants' management styles aimed at getting more involvement and commitment to change, and consulting with a wider range of people before making decisions). The second concerns the processes of planning and information seeking that are undertaken before decisions are made (labelled as more rational decision-making and better reporting plans).

These experiences suggest that the changes that managers were making in their decision-making practices as a result of individual unlearning experiences, such as collecting evidence from a variety of sources, critically appraising that information and considering the views of stakeholders, are in line with the steps prescribed by Briner et al. 86 for adopting evidence-based management. However, there was little evidence from our participants that external published research was being regularly integrated into these decision processes, the information tending to be gathered from within the organisation. Furthermore, both categories suggest that participants' unlearning experiences had caused them to reflect on their decision-making processes and move away from practices that were akin to idea imposition to more discovery-orientated approaches. 91 After the unlearning experience, some participants appear inclined to adopt a more systematic gathering of information to learn about possibilities and to understand stakeholder views and perceptions.

However, there was one exception to this trend, with one manager reporting that she had moved away from a heavily consultative discovery approach to a more decisive idea imposition-based approach (see quote on p. 57, paragraph 5). This was in reference to a situation in which the change was imposed from an external policy request that was required to be implemented. In this case, it appears that the manager thought that she had been consulting excessively with stakeholders regarding change that was non-negotiable. The participant's comments suggest that this change in style had been effective, although it is not clear whether this was purely due to the specific circumstances of the change required or that the consultation aspect of the discovery approach can be over applied.

However, overall, the broad message from our exploratory findings is that the experience of unlearning seems to encourage the adoption of more organised approaches to searching for and collecting information and stakeholder opinions and critically evaluating this information for decision-making. This suggests a move towards the discovery process of decision-making, which Nutt91 has demonstrated to be more successful than other decision-making processes.

The extent to which the health-care setting affects health-care managers' engagement with individual unlearning and fading

As the findings indicate, managers at the acute and mental health NHS trust sites engaged in both types of individual unlearning and reported experiencing fading. Managers at both sites reported unlearning policies and procedures most commonly. Overall, there was little difference between what was unlearnt at the different sites, suggesting that the organisational characteristics of the two case study NHS trusts did not affect either the level of engagement in individual unlearning or the nature of individual unlearning for our participants.

Barriers to individual unlearning were largely the same at both sites, although financial constraints were reported only by managers at the mental health NHS trust and lack of management support was identified by a larger number of managers at this site than at the acute NHS trust. Enablers were also much the same at both sites and the number of times that particular enablers were cited was similar. However, only managers at the mental health NHS trust reported changes in policies and procedure as enabling individual unlearning. New practices or knowledge was mentioned as an enabler of fading by more managers at the acute NHS trust than at the mental health NHS trust.

Our findings suggest that lack of management support rather than the type of NHS trust may explain the limited differences in barriers to and enablers of individual unlearning experienced by our participants. Alternatively, the differences may be due to the different lengths of time since the NHS trusts acquired foundation trust status. At the time of the interviews the acute NHS trust had had foundation trust status for more than 6 years, whereas the mental health trust had had foundation trust status for less than 2 years. As a result, managers at the latter site may have been subject to a greater degree of change in a short time frame than those at the acute site. Having said that, at the time of the interviews both sites had recently acquired new community services and the acute NHS trust was undergoing an internal restructuring. However, it should also be acknowledged that these two types of NHS trust compared with other health-care organisations in the NHS are relatively similar in terms of their structure and manner in which they deliver care (e.g. ward based). These similarities may explain the relatively consistent findings in this study when assessing the influence of organisational setting. This issue will be revisited in the conclusions of the report and the research agenda.

The relationship between learning, individual unlearning and change

This was not a topic that the study explicitly aimed to examine. However, it is a topic that emerged as being of interest at various points throughout the study, largely because of some of the challenges of data collection and data analysis that occurred, which encouraged the research team to reflect on this topic. There is scope to make a potentially significant contribution to knowledge regarding the inter-relationship both between learning and unlearning and between unlearning and change for two reasons. First, our analysis challenges the dominant way in which the relationship between these topics is conceptualised in the unlearning literature. Second, we were able to collect rich data on people's experiences of unlearning, which gave good empirical insights into the nature of individual unlearning processes (which, as outlined previously in Chapter 2, Reviewing the academic literature on unlearning, is a neglected topic in the unlearning literature) and which helped shed light on the relationship between learning, unlearning and change.

One of the key challenges experienced both during data collection, in conducting the interviews, and during data analysis, when research data were being coded in NVivo, was the close relationship between learning, unlearning and change. These concepts were so inter-related that it was difficult to tease them apart and separate them. For example, in terms of data collection, when we asked managers to describe their unlearning experiences, they often started by talking about their experiences of change, and only on detailed probing revealed information on unlearning. We argue that these challenges do not relate to any limitations in our data collection or analysis skills but instead say something about the intrinsic nature of the relationship between individual unlearning, learning and change. In considering these issues, this section is divided into two parts, with the first considering the relationship between learning and individual unlearning, and the second considering the relationship between individual unlearning and change. Further, these discussions link back to Chapter 2 (Reviewing the academic literature on unlearning and Types of individual unlearning and the nature of individual unlearning processes) where discussion on these topics has already been undertaken.

To what extent do health-care managers engage with the process of unlearning and what impact does the engagement with this process have on health-care managers' decision-making?

The findings demonstrate that the health-care managers who participated in this study engaged with unlearning processes although the form and circumstances of the unlearning varied. The findings of the study were conceptualised into two distinct types of individual unlearning experienced by participating health-care managers. The first was behavioural unlearning, which was triggered by a deliberate process of change that had been externally imposed. The second was cognitive unlearning, which was triggered by an unexpected external event that questioned some basic assumptions of the individual. We have added further refinement to this conceptualisation by developing a new typology, which distinguishes between four separate types of individual unlearning.

The findings indicate that 57 episodes of individual unlearning were identified by the 29 health-care managers who participated in the study. All participating health-care managers were able to identify at least one unlearning episode. The most common individual unlearning type was behavioural unlearning triggered by a change event (23 episodes), followed by cognitive unlearning triggered by an individual experience (17 episodes). The least common individual unlearning type was behavioural unlearning triggered by an individual experience (six episodes) (see Table 8). These findings give some indication of the extent to which our participants engaged with processes of unlearning. However, we are not able to generalise from these findings to health-care managers working in other settings in the NHS, although we also have no reason to believe that our findings are atypical.

The findings suggest that engaging in processes of individual unlearning did impact on decision-making for some participating health-care managers, with 30 examples of decision-making changes identified from 57 unlearning episodes. However, some managers found it difficult to clearly identify examples of changes to decision-making that were a result of individual unlearning experiences and six participants did not give any examples of impact on decision-making. The difficulty that some participants encountered is probably due to a lack of awareness at the time of unlearning and not easily associating changes in decision-making as an obvious outcome. Acknowledging this limitation, the findings show that after engaging with individual unlearning several managers moved away from idea imposition approaches to decision-making in favour of more discovery-focused approaches. Particular changes involved more systematic and formalised procedures for collecting and documenting information, and greater consideration of stakeholder views and perceptions when formulating and implementing decisions. Research suggests that discovery-orientated approaches to decision-making processes are more successful than idea-imposed processes. 91 Consequently, it is likely that these managers have improved the effectiveness of their decision-making by engaging with unlearning processes.

What are the barriers and enablers that influence the engagement of health-care managers with the process of unlearning?

Participants mentioned 36 examples of barriers encountered during episodes of behavioural unlearning, but mentioned fewer examples of barriers encountered during cognitive unlearning episodes (n = 19). The most common forms of barrier identified were staff resistance to changes in practice, personal behaviour, lack of management support, high workload and organisational culture. Other barriers that were mentioned included financial constraints and job insecurity. In addition, we also identified 13 examples of barriers to fading, the most common being the ability to recall past knowledge and experience. Our findings show that barriers to individual unlearning could be individual or organisational. These barriers could influence either the individual unlearning process or the implementation of the new ways of working or behaving that the unlearning process required for full implementation by the individual. For example, an individual's personal behaviour may act as a barrier to adopting a better work–life balance because of his or her desire to maintain control and resist delegating tasks. However, a lack of support from managers to devolve decision-making to more junior members of a team, or team members not wishing to take responsibility for decision-making, may also prevent an individual from abandoning his or her old ways of working. Of course, in the latter case, it is possible that the person has engaged with unlearning but has been unable to implement the outcome of this unlearning experience. Further research is required to fully unpick the relationship between individual unlearning and barriers to unlearning processes.

A further aspect of unlearning that is worth noting in relation to barriers to individual unlearning is the variation in emotional impact between the types of unlearning reported by study participants. In particular, the cognitive unlearning experience often involved high emotional strain and appeared to be associated with feelings of sadness and grieving. Although the catalyst for the cognitive unlearning may be a relatively sudden change or event, the associated emotional adjustment and process of unlearning may take a considerable amount of time. In addition, our workshop findings suggest that this emotional impact does not reduce, even if the individual has experience of previous cognitive unlearning episodes. Studies of individual transitions during change have shown that employees often find change to be a deeply emotional process that can engender a sense of uncertainty, frustration and defensive routines. 63,64 These responses are often interpreted as ‘resistance to change’; however, our findings suggest that they may also be associated with some types of individual unlearning. Consequently, it is important that change agents do not immediately interpret these types of emotional response to change as non-desirable58 but recognise that they are likely to be the early signs of employees giving up old practices in order to adapt to new knowledge or behaviours.

Support networks, usually in the form of particular individuals, seemed to be important in helping individual managers when experiencing this change, with less reference to organisational support mechanisms. This reflects the individual and personal nature of the cognitive unlearning experience, but may also indicate a lack of sufficiently individualised organisational support mechanisms. Participants from our workshops also suggested that more continuous support may be required after change events, both to ensure that the unlearning is successfully embedded and to help the emotional transition for the staff concerned. Emotion has been identified as an important aspect influencing information processing for decision-making. Individuals who are experiencing positive emotion have been found to be more able to process information and be creative than those who are not. 80 Therefore, it is possible that a health-care manager experiencing cognitive unlearning that triggers negative emotion could find his or her decision-making processes impaired.

Regarding the enabling factors for individual unlearning, a total of 48 and 49 examples of enablers of behavioural and cognitive unlearning were identified respectively. With the exception of one factor (changes in policies and procedures), all of the enablers mentioned by participants could facilitate either behavioural or cognitive unlearning. The most common enablers were personal skills and attitudes, leadership and networks and support from specific individuals, all of which are supported in general terms by the literature. Becker9 argues that, to enable staff to have a positive attitude to change and consequently to unlearning, managers need to demonstrate commitment, make the case for change, involve staff in planning and provide reassurance. Overall, the main skill that our participants identified as supporting their unlearning was communication, a viewpoint supported by the literature, which talks about the value of questioning45 and listening6,45 along with the ability to discern what needs to be unlearnt6 and to come to terms with the loss of previous ways of seeing and of fundamental assumptions. 32 It seems likely that the positive attributes that enabled individuals to unlearn also assisted them in building supportive relationships that further supported their unlearning. In other words, the different sets of enablers are likely to interact and reinforce one another. These recommendations echo some of the key aspects advocated by Briner et al. 86 for evidence-based management, so the promotion of these activities by senior managers may help in providing role models for effective decision-making practices coupled with a supportive and more receptive environment for individual unlearning by health-care managers.

To what extent does the health-care setting affect the level of engagement with the process of unlearning by health-care managers?

This study also attempted to investigate the influence of health-care setting on individual unlearning experiences of health-care managers. The study had planned to address this question by comparing individual unlearning experiences of managers across three different case study sites (acute, mental health and primary care trusts). However, the extent to which this question could be fully addressed by the findings was limited by not being able to recruit any participants from the primary care site. Consequently, the comparison of experiences could be made only between two case study sites. Further, as our study was concerned with individual experiences of unlearning rather than organisational unlearning, our findings are limited in the degree to which they can be generalised to other health-care managers within each case study, as well as to other health-care settings. Despite these limitations the study did find some interesting results.

The findings showed that participants at both sites reported very similar levels of experience of each unlearning type and also of the types of knowledge that were unlearnt. This was also the case for barriers to and enablers of unlearning. Similarly, consistent themes were identified for fading. These similarities suggest that the type of NHS trust did not influence the level of engagement in unlearning or the nature of what was unlearnt. However, it could also be argued that the relative similarity of these two trusts in terms of their organisational structures may also explain the consistency of findings. Consequently, as noted in the research agenda later in this chapter, further research is required to investigate individual unlearning in a wider variety of health-care settings to establish whether or not there are context-specific aspects to health-care managers' unlearning experiences.

Limitations of the study

Although this study makes some important contributions, it is subject to a number of limitations that should be kept in mind when considering the findings and conclusions. First, the exploratory nature of the study means that, although the findings provide a useful foundation for future research in specific aspects and impacts of unlearning, the study is limited in the depth of insight that it can provide in terms of specific aspects of unlearning processes and the extent to which they occur among health-care managers across different health-care settings.

Second, it is acknowledged that this is a relatively small-scale study involving a small number of participants drawn from two NHS trusts. The original plan for the study involved data collection at a third site in a primary care setting, but, unfortunately, recruitment of participants for the study was unsuccessful at this site (see Chapter 3, Recruitment at site C: primary care trust/clinical commissioning group). Consequently, the loss of the primary care site limited the degree of contextual influence that could be examined in the study and the findings are therefore limited with respect to their generalisability to health-care managers working in other health-care settings.

Third, we found the process of eliciting the details of the unlearning experience from participants to be challenging. During the interviews, participants would often begin by and return to talking about change processes rather than the specific types of knowledge and behaviours that they had given up or abandoned as a result of experiencing a change event or individual experience. We dealt with this challenge through careful probing and follow-up questions during the interviews and also by two or more members of the project team reviewing potentially ambiguous examples during the data analysis to achieve an agreed interpretation. However, it is possible that the distinction between change event comments and references to unlearning could have been interpreted differently.

Fourth, because of the relatively low numbers of participants in our study, the findings provide limited insights regarding the extent to which managers engage with processes of unlearning and the impact on their decision-making. This exploratory study does show the extent to which our study participants engaged with unlearning processes and suggests that further research would be valuable to assess wider levels of engagement by health-care managers in other health-care settings. Similarly, this study's findings are helpful as they indicate that there may well be a relationship between unlearning and health-care managers' decision-making processes; however, this relationship would also benefit from further research to fully understand the precise nature of the relationship. Acknowledging the limitations of this original exploratory study highlights a number of specific areas where further research would be valuable.

Fifth, the study was able to provide only limited insights into the nature of fading, although the empirical data do provide evidence that fading should not be conceptualised as a type of individual unlearning. The study participants had difficulty recalling examples of skills and knowledge that they had forgotten. A longitudinal data collection strategy may be more effective at capturing the process of forgetting, which was outside the scope of this study.

Research agenda

1. The investigation of a greater number of health-care managers' experiences of unlearning in a wider range of settings is necessary to establish the significance of health-care setting for individual unlearning experiences and provide greater generalisability to the findings of this study. In particular, extending this study to investigate the role of unlearning in the primary care sector in 2013–14 would be of value to explore health-care managers' unlearning experiences in relation to the significant structural and organisational change they will have recently experienced.

2. The findings of this study indicate that there is a relationship between individual unlearning and health-care managers' decision-making. However, the precise nature of this relationship requires further investigation. In addition, several of our participants found it challenging to identify changes in decision-making as a result of unlearning, probably because they did not consider this relationship at the time. A new study that asks health-care managers to identify and record episodes of unlearning over a 6-month period and after each episode to identify resultant changes in decision-making may be more effective for investigating this relationship.

3. The data suggest that change of one type or another (individual-level role change and organisational-level structural change) is a catalyst for unlearning, and also that (further) organisational-level change may result from people unlearning and learning. However, further research on both aspects of this relationship is necessary to more fully understand the dynamic interaction between change and unlearning. For example, one (untested) suggestion by Tsang and Zahra2 was that different types of organisational change event might result in different types of unlearning. Thus, in further investigating this topic it would be useful to be sensitive to both the specific type of change that precedes the unlearning episode and the type of unlearning that results from such changes.

4. For practitioners, further understanding of the character and dynamics of individual unlearning processes would be valuable. This understanding may help to identify likely causes of resistance to change and organisational structures and employers' actions that may (inadvertently) present barriers to unlearning and therefore barriers to achieving desired change. Recognising employees' attitudes to change will provide an indication of how receptive they are likely to be to new change events that require them to abandon previous knowledge or ways of working. Furthermore, a better understanding of the process of individual unlearning will also enable practitioners to have more realistic expectations regarding the impact of a change process, and the likely time required for new practices and procedures to become embedded.

In-depth qualitative research would be the most suitable to explore this complex phenomenon, in a range of contexts and environments. In particular, as the time between the catalyst for individual unlearning and changes in attitudes or behaviour may vary in duration, longitudinal studies that involve participants recording changes in their behaviour and attitudes in research diaries could be of particular value. This approach would also enable the study of inhibitors of and facilitators to individual unlearning and their relationship with the different stages of the unlearning process.

Implications for practice

As a result of our study we are able to offer three tentative implications for practice. However, when considering these suggestions the small data set and exploratory nature of this study should be kept in mind.

1. The individual unlearning experiences of our participants (in their view) led to improvements in some participants' decision-making processes. These improvements involved greater and more formalised information searching and gathering, critical appraisal of this information and a consideration of stakeholder views and perceptions of possible options. These changes are all in line with what is prescribed in the literature for adopting evidence-based management decision-making processes.

2. People may find it hard to distinguish between unlearning (abandoning/giving up established practices or ideas) and learning (the development of new practices or knowledge) as these processes typically occur virtually simultaneously. Training programmes designed to support change may need to actively acknowledge and address participants' abandonment of existing knowledge and behaviours as well as concentrating on explaining and delivering new knowledge and desired behaviours.

3. Individual unlearning can be triggered by change events or individual experiences and unlearning may lead to further change. The period of transition to new ways of working or thinking may stretch beyond initial change-related training. During this period managers could benefit from ongoing informal feedback, support, advice and role modelling from colleagues, managers and mentors to support both the individual's well-being and embed the new knowledge and/or behaviours. Individuals who experience unlearning may also generate new ideas and opportunities for further changes and possible improvements in their own ways of working or the functioning of their team or department.

Contribution of authors

Dr Crispin Coombs (Senior Lecturer in Information Systems) led the study, participated in the fieldwork and analysis, contributed to the writing of the literature review, methods and conclusions and provided overall editorial support for the entire report.

Dr Donald Hislop (Senior Lecturer in Human Resource Management and Organisational Behaviour) participated in the fieldwork, led the writing of the literature review and discussion sections and provided input to the analysis.

Dr Julie Holland (Director of the Glendonbrook Centre for Enterprise Education) facilitated access to the case study sites and designed and delivered the workshops.

Dr Sara Bosley (Lecturer in Human Resource Management and Organisational Behaviour) was appointed as the project's research associate for the first 12 months of the study. Dr Bosley contributed to a number of areas including the literature review and methodology and discussion sections of the report. Dr Bosley also conducted the majority of the research interviews and developed the qualitative coding scheme for the analysis.

Dr Esmeranda Manful (Research Associate) was appointed as the project's research associate for the final 6 months of the study, following Dr Bosley's change in role. Dr Manful undertook and wrote the analysis sections of the report.

Disclaimer

This report presents independent research funded by the National Institute for Health Research (NIHR). The views and opinions expressed by authors in this publication are those of the authors and do not necessarily reflect those of the NHS, the NIHR, NETSCC, the HS&DR programme or the Department of Health. If there are verbatim quotations included in this publication the views and opinions expressed by the interviewees are those of the interviewees and do not necessarily reflect those of the authors, those of the NHS, the NIHR, NETSCC, the HS&DR programme or the Department of Health.

Full title of project: The Role of Unlearning in Healthcare Managers' Decision Making

Proposed duration (in months): 18

Proposed start date: 01 Jan 2011

Aims and objectives

The overall research aim of this project is to empirically investigate the extent to which healthcare managers engage in processes of unlearning, and the barriers that exist to them doing so. This is important because it will improve our understanding of how managers make decisions and will help provide relevant support and training for those involved in running the health service.

In addressing this aim the following specific research questions will be investigated:

1. To what extent do healthcare managers engage with the process of unlearning and what impact does the engagement with this process have on healthcare manager decision making?

2. What are the barriers that prevent healthcare managers engaging with the process of unlearning?

3. To what extent does the healthcare setting (acute care, mental health care or primary care) affect the level of engagement with the process of unlearning by healthcare managers?

Link to SDO call on research utilisation and knowledge mobilisation by healthcare managers

This research proposal responds directly to the main primary research topic area ‘the contribution of “unlearning” in managers' use of research and other forms of knowledge’ in the SDO call on research utilisation and knowledge mobilisation by healthcare managers. It addresses the increasing interest in understanding the mechanisms that encourage or reduce the adoption of new interventions such as clinical research or management practice. Unlearning may impact on the extent to which knowledge is used or transferred by managers. Old knowledge may need to be unlearnt before new knowledge can be learnt and applied to management decision making. If healthcare managers fail to unlearn existing practice, this may act as a barrier to knowledge flow and implementation. This study will explore managers' use of research and existing knowledge and how this relates to managers' own decision making through analysis of the process of unlearning.

Background

The concept of unlearning (a deliberate process of reflection which results in ‘giving up’ some existing knowledge which is deemed to be outdated) is closely linked to processes of learning and change. Fundamentally, for change to occur unlearning must happen, as people need to acknowledge the limitations of existing knowledge/practice, be prepared to ‘give up’ such knowledge and adapt their knowledge/practice by actively learning and developing new ways of working/thinking. Thus, unlearning is arguably a necessary precursor to processes of learning (Akgünet al2007). Despite this, the concept of unlearning is under-utilised and poorly understood both in the general literature on learning and knowledge management (Hislop 2009), and also in the literature on learning in the health sector (Crillyet al2009, Rushmer & Davies 2004). This project will thus contribute to developing a greater understanding of this aspect of knowledge management and learning processes.

Rushmer & Davies (2004) argue that distinctions can be made between three separate types of unlearning which vary in a number of ways, most importantly their degree of significance. One type of unlearning they identify, which they label fading, refers to the slow gradual process of unlearning that can occur over time, where particular skills and capabilities are forgotten through a lack of usage. This form of unlearning is typically not experienced as being significant or challenging for people, and due to the gradual way it occurs, people may not be conscious they are experiencing it.

The second type of unlearning identified by Rushmer & Davies (2004) is labelled wiping. This represents a more significant, deliberate and conscious form of unlearning. Wiping is a form of unlearning focussed relatively narrowly on a particular practice, or activity, where a person consciously unlearns through making deliberate efforts to give up a particular way of thinking/acting, as a pre-cursor to changing these beliefs/practices. This process of unlearning may be initiated by some external catalyst, such as a change initiative, which places demands on people to change their ways of working and/or thinking. An example of wiping in a clinical-managerial context would be someone deliberately changing a work procedure for deciding how and when patients are to be discharged from hospital.

The third form of unlearning identified by Rushmer & Davies is deep unlearning. Deep unlearning represents a radical form of unlearning which typically occurs rapidly and unexpectedly, and is experienced by people as significant as it brings into question some basic assumptions and values. Due to these characteristics this form of unlearning can produce significant emotions (fear, confusion etc.) in the people who experience it. Typically, deep unlearning is initiated by some external catalyst, such as an event whose characteristics or outcomes are unexpected. An example of deep unlearning, gained from conversations with healthcare managers studying at Loughborough was a sudden change a manager made in the way they communicated with their staff when consulting about important ward management decisions. The person concerned switched from consultation by e-mail, to making a significantly greater use of face-to-face interactions when consulting people, after it was found that there were problems and discontent among clinical staff when they were consulted only by e-mail. A key difference between this example and the wiping example, was the speed at which the change was made, and the emotional impact associated with the experience, the manager describing the experience as being ‘burned’ due to the discontent that arose among her staff and the difficult atmosphere that ensued.

Thus, Rushmer & Davies' three types of unlearning can be distinguished from each other in a number of ways including their:

• Level of significance, with fading having a low level of significance to people, and deep unlearning a high level of significance.

• Scope, or breadth, with wiping and fading typically having a narrow scope, being limited to particular, specific skills or practices, whereas deep unlearning may have a broader scope, due to the questioning of basic assumptions and values it involves.

• Speed, with fading happening relatively slowly and gradually, while deep unlearning typically happens rapidly.

• Catalyst, with wiping being a deliberate and conscious process of unlearning (typically in response to some external change), whereas both fading and deep unlearning are instead initiated by unexpected external events or the lack of use of particular skills and capabilities.

• Emotional impact/consequences, with fading being experiences as having a limited emotional impact, while deep unlearning can have a significant emotional impact for people.

One of the main aims of this research is to examine the extent to which healthcare managers undertake all three forms of unlearning (research question 1). A typical catalyst for all forms of unlearning are processes of change that require the adaptation of working practices (Tsang & Zahra 2008). The pace of change in the NHS in recent years has been significant. For example, Greener, (2008, p. 206), talked about how the ‘pace and scope of change’ was, ‘intense’. Tsang & Zahra (2008) suggest that different types of unlearning may be associated with particular forms of change, distinguishing between continuous and episodic change. With continuous change, which is small-scale, evolutionary and occurs on a day-to-day basis, unlearning is likely to take the form of wiping and fading. For episodic change (which is large-scale rapid change) unlearning is likely to take the form of deep unlearning. Thus, it is anticipated that due to the amount of ongoing change in the NHS, all healthcare managers are likely to have experiences of processes of wiping and fading, where they have had to adapt their work practices and routines as a result of change. Deep unlearning on the other hand may be experienced by healthcare managers who have undergone a significant role transition, such as would happen when a clinician moved into a managerial role. Thus, it is possible that processes of deep unlearning may be more evident in managers who have undergone such transitions.

A possible example of this process of deep unlearning generated by a significant role transition is the experience of change for a clinician moving into a clinical-management position, due to the challenges clinicians who have moved into management roles are likely to face. One challenge is the need to adapt the way they use knowledge to act and make decisions, as somewhat anecdotal evidence suggests that clinical and managerial decision-making processes are underpinned and supported by different knowledge bases. Thus, Crillyet al(2009, p. 145) argue that, ‘the professional and scientific culture of doctors stands in contrast with the world of managers. Doctors are aligned with the quantitative and positivistic nature of biomedical research where clinical decisions may be codified. Managers' decisions depend on pragmatism and subjectivity rather than research which is more often qualitative and in any case is perceived as contingent and not readily generalisable.’

Clinicians moving into managerial roles thus arguably need to engage in processes of unlearning to effectively adapt how they act and make decisions in the new managerial roles they occupy. Thus, it is anticipated that processes of deep unlearning are more likely to be experienced by managers who have undergone any type of significant role transition.

While we anticipate that clinical managers may be more likely than the non-clinical managers to experience deep unlearning, we will also look for evidence of deep unlearning in the interviews with all healthcare managers. As deep unlearning is triggered by unexpected external events it may be that healthcare managers have experienced it sometime in carrying out their work, despite not having made the type of disciplinary transition that clinical-managers have. The lack of research on the topic of unlearning means that there is a limited understanding of how frequently people experience it, or the type of events/circumstances that trigger it. Thus, examining all forms of unlearning with every interviewee allows comparisons to be made between clinical and non-clinical managers, and will also give insights into the extent to which both experience unlearning (research question 1).

This research is also concerned with understanding the barriers to unlearning that exist. While unlearning has the potential to be an important component in the processes of learning and change, research suggests that there are many barriers to unlearning, which means that learning from mistakes, failure and crisis often does not happen (Baumard & Starbuck 2005). Barriers to unlearning exist at both individual and team/organisational levels. The small amount of research on unlearning (Rushmer & Davies 2004) and learning from failure that has been done in the health sector suggests that this is a work context where the barriers to unlearning can be significant. Further, this research suggests that in such contexts both individual level barriers (concern about negative consequences/sanctions from admitting mistakes, and a preference for local/contextual knowledge) and organisational level barriers (a culture where admitting to mistakes isn't encouraged) can be causes of unlearning (Dean 2002) (research question 2). Through interviewing healthcare managers in three distinct organisational contexts (acute care, mental health care and primary care) we will investigate the extent to which organisational level barriers to unlearning are shaped by the work context (Research question 3).

The central focus of this project is on examining the extent to which the managers interviewed engage in processes of unlearning, and the individual and/or organisational level barriers which may be inhibiting effective unlearning and learning from occurring. Thus the focus of this project is distinctive from that of other SDO funded projects, which are more focussed on inter-personal processes (such as Swan's study of the ‘co-production’ of knowledge – SDO/244/2008, and Dopson's study on the motivation of health managers to use knowledge – SDO/242/2008).

This research will make a number of practical contributions to policy and practice in the NHS. Most fundamentally, it will advance knowledge on a neglected aspect of the processes of change and learning. By examining unlearning, and the barriers to achieving it that exist, the research can help develop a greater understanding of what may be an important barrier to change. In practical terms, this knowledge can be used to help facilitate processes of change and learning through addressing the barriers to unlearning that are identified. A second contribution to policy and practice in the NHS is that the research will shed light on some of the challenges to learning and decision-making faced by healthcare managers. This knowledge can thus be used to develop ways to better support healthcare managers. In terms of healthcare managers' decision making the findings of this study can be used to address the barriers to learning from mistakes and failure, thereby improving the learning process which should enable better decisions to be made in the future.

Need

This project meets and contributes to the needs of the NHS. Improving our understanding of how managers make decisions will help us provide relevant support and training for those involved in running the health service. Improving management performance will in turn lead to better delivery of services and ultimately improved patient care. The study has the potential to generate new knowledge in the area unlearning and will be of direct value to trust boards and those who provide training and support for NHS managers (e.g. HR departments, Institute of Innovation & Improvement and external training providers such as universities).

Existing research suggests that NHS managers with clinical and non-clinical backgrounds may have a different approach to issues such as accountability, use of guidelines and finance. It has been argued that these differences are a result of each discipline's training, beliefs and experiences (Degelinget al., 2003). Similarly, there is evidence to suggest that clinicians and managers have different priorities concerning healthcare delivery: clinicians tend to prioritise patient care whereas managers focus more on cost (Nicoliniet al., 2008). Research by Guven-Uslu (2006) suggests that these different mindsets influence how managers with a clinical background are likely to approach decision making compared to those without clinical experience. Given the increased emphasis that is being given to encouraging clinicians into leadership and managerial roles (Darzi, 2008) these different mindsets are likely to become increasingly more significant to the efficient running of the NHS. Clearly, both cost and high quality patient care are key considerations for healthcare delivery and getting the best balance between these two issues is important for the NHS especially in the current and likely future economic conditions. However, achieving this balance may be difficult for NHS trusts when their managers come from different backgrounds and may be influenced by their existing knowledge, training, beliefs and experiences on which they rely when making decisions.

For example, due to their structured evidence based clinical training, managers with a clinical background may find it difficult to challenge management practice because there is often no evidence based knowledge supported by hard data and publications to support their management decision making. Anecdotal evidence from our own medical management programme indicates that clinicians find management challenging because there is often no clearly defined management protocol to follow, compared to clinical practice which has a comprehensive evidence base. Discarding obsolete knowledge (unlearning) may be an alien concept to many clinicians who may consider hoarding explicit clinical knowledge as essential when dealing with the diagnosis and treatment of disease. The concept of doing things differently, particularly in a clinical setting, may be unfamiliar or frowned upon because of the unsafe consequences a poor medical decision can have. When this approach is applied to a management situation, organisational unlearning may be prevented.

Similarly, non-clinical managers may find the unlearning processes challenging when faced with, for example, resource and cost perspective issues. They may be instinctively resource led in their decision making and have a strong desire to improve the quality of data, or gather comparative data on which to make numerically based decisions. Without a direct involvement in patient care, this group may find it harder to develop a more ‘needs led’ approach to decision making.

This study will investigate the role of unlearning in healthcare managers' decision making. It will also explore the implications of unlearning for managerial decision making in the health service.

Methods

Collective research effort – contribution to research utilisation

It is envisaged that the findings from this study will provide a number of important contributions. It will identify the importance of the process of unlearning for healthcare managers and the barriers to this process. These findings should provide insights and recommendations in terms of new issues to address for managers in terms of their awareness of their current decision making processes and how these could be improved. It should also provide useful information on the importance of unlearning that could be used to enhance and refine future management training courses for managers. Unlearning is likely to be a key precursor that needs to be successfully completed before managers are able to improve their use of knowledge and may influence their motivations to use knowledge and evidence in decision-making. Therefore the findings of this study will also complement and enhance the understanding of management decision making in the NHS that is being developed from current SDO projects (e.g. Swan's study of the ‘co-production’ of knowledge – SDO/244/2008, and Dopson's study of the factors which influence the motivation of health managers to use knowledge – SDO/242/2008).

After reporting to the SDO, initial dissemination of findings must be to the people best placed to use them. Firstly, we plan to share our findings with participants, senior managers and the boards of each participating NHS Trust. This will take the form of two stages. Once data analysis is complete, we will share our finding in a workshop format allowing participants to feedback on our findings and discuss their practical applications and implications. Once this stage is complete, a summary report will be made available to the participating NHS trusts.

Once the research findings have been shared with the participating NHS trusts we will then disseminate the findings in two domains. Firstly, the findings will be disseminated to a practitioner-oriented audience, both through presenting articles at relevant conferences (such as the annual Institute of Healthcare Managers conference), and also publishing articles and papers in practitioner-oriented journals (such as the Health Service Journal).

Secondly, we will disseminate the findings of our research to the academic community through presenting articles at conferences and also publishing articles in a range of relevant journals. The extent to which the concept of unlearning is neglected and under-utilised means that there is scope to make a potentially significant contribution to knowledge. The focus of our research means that the findings are relevant to a number of different academic domains, including healthcare, knowledge management, and business/management. Thus, we will present research data, and publish articles in all three domains. Examples of relevant conferences we will present our findings at include the British Association of Medical Managers Conference, the Organisational Learning, Knowledge and Capabilities conference, and the British Academy of Management conference. Examples of relevant journals we will target publications at include the Journal of Health Services Research and Policy, Health Services Management Research and the Journal of Knowledge Management.

Following the Darzi Report (2008), shifting managerial power and control of resource allocation back to clinical professionals is being encouraged in the health service. As a result, management training programmes run by the NHS and a growing number of universities must accommodate the training needs of healthcare managers. Changing the decision making processes of NHS managers will be challenging and will be best addressed by management training, coaching and mentoring. If the findings of our study show that unlearning is an important precursor to learning, we will incorporate this insight into our healthcare management training programmes at Loughborough University. This will enable us to have an immediate direct impact on NHS managers' decision making in the NHS. Understanding how healthcare managers unlearn could influence significantly the future style and content of management training programmes. We will therefore disseminate our findings to internal NHS training providers (e.g. Institute of Innovation & Improvement) and relevant universities.

Plan of investigation and timeline

Table 2 provides a detailed summary of the project plan of investigation and also indicates when the progress reports of the project will be available. It is anticipated that the start date of the project should give sufficient time to advertise and recruit a research associate so that they will be in post for the start of the project. In addition to the details provided below, there will be monthly project team meetings and more frequent supervisory meeting with the research associate.

Approval by Ethics Committees

We will gain informed consent from our research subjects. They will be guaranteed individual anonymity in terms of not being individually identified in written material and confidentiality in that interview transcripts will only be viewed by the research team and will be held in a secure and anonymised form. More generally, data will be held in accordance with the Data Protection Act. It will be necessary to apply for NHS Research Ethical Committee (REC) approval and subsequently to obtain Honorary Contracts to enable researchers to interview in the case study sites. Following consultation with RECs associated with participating trusts, we have confirmed that approval is needed from one REC only. We will therefore submit our proposal to Trent Research Ethics Committee (Appointing Authority – East Midlands Strategic Health Authority).

Project management

Coombs will take overall responsibility for the management of study, with support from Hislop and Holland. Coombs has experience in undertaking healthcare management research projects using multi-site case studies and managing research associates. Hislop will provide the project with subject area expertise, project management support for Coombs and will manage the data collection and analysis. Holland will manage the liaison with NHS trusts, contribute to the data analysis and facilitate the workshops. In addition a research associate will be appointed for the project term who will be responsible for the data collection, analysis and drafting the final report. Coombs, Hislop and Holland will be involved in the pilot interviews and at least the first two interviews of the main study with each case study trust. The presence of two interviewers for these interviews will ensure the validity of the interview schedule and rigour in data collection. All members of the research team will contribute to the overall dissemination of the project's findings as described in Section 0. The project team will have meetings every month to monitor progress and will have additional regular supervisory meetings with the research associate.

Service users/public involvement

As this study is specifically focused on the role of unlearning in NHS managers' decision making, and having reviewed the INVOLVE documentation, we do not anticipate that service users or the public will be able to usefully contribute to the study's research objectives. Consequently, we do not propose to include service users or the public in this study.

Expertise and justification of support required

The members of the team that will undertake the study bring a complementary range of skills and expertise to the project. Coombs and Hislop are both experienced researchers having conducted several qualitative research studies. Coombs has an established track record of healthcare research having held Department of Health, DTI, ESRC and Society of Radiographers grants and previously worked for the NHS Modernisation Agency Clinical Governance Support Team as a researcher. Hislop has been researching and writing on the topic of knowledge management for the last decade and has published academic work on unlearning. Hislop has strong experience of research project management having held grants from the British Academy, EPSRC, ESRC and White Rose Trust. Holland, is an expert in work-based learning, knowledge management and management education for clinical staff, and has recently been instrumental in developing and delivering the University's MSc in Medical Management which is aimed at clinicians working as clinical managers within the health service. Holland has previously studied learning and its effect on job control and decision making and the role of daily learning for problem solving in SMEs and has held grants from the East Midlands Development Agency. Previously, Holland was Director of the East Midlands NHS Innovation Hub, the knowledge transfer management and commercialisation office for primary and secondary care NHS trusts and therefore also provides expert knowledge of practical healthcare management issues and key NHS contacts to the project.

FIGURE 1.

Flow diagram for project.

Recent relevant publications for Coombs, Chief Investigator

References

Appendix 1 Outline interview schedule

The Role of Unlearning in Healthcare Managers' Decision Making Steering Group Terms of Reference

Author: Dr Sara Bosley, Research Associate, The School of Business and Economics, Loughborough University.

Date: 14 June 2011

Version number: Version 2 agreed

Purpose

This document describes the terms of reference, membership and methods of communication for the Steering Group for the research project, ‘The Role of Unlearning in Healthcare Managers' Decision making’.

Background

The aim of the research project is to empirically investigate the extent to which healthcare managers engage in processes of unlearning, the barriers that prevent and the enablers that promote their engagement in unlearning.

The concept of unlearning (a deliberate process of reflection which results in ‘giving up’ some existing knowledge) is closely linked to processes of learning and change. For change to occur unlearning must happen, as people need to acknowledge the limitations of existing knowledge/practice, be prepared to ‘give up’ such knowledge and adapt their knowledge/practice by actively learning and developing new ways of working/thinking.

This topic is important because it will improve our understanding of how managers make decisions and will help provide relevant support and training for those involved in running the health service.

The project will address the following specific research questions:

1. To what extent do healthcare managers engage with the process of unlearning and what impact does the engagement with this process have on healthcare manager decision making?

2. What are the barriers that prevent healthcare managers engaging with the process of unlearning?

3. To what extent does the healthcare setting (acute care, mental health care or primary care) affect the level of engagement with the process of unlearning by healthcare managers?

Terms of Reference

• To comment and advise on the progress of the project.

• To comment and advise on the research process, early findings and draft reports.

• To ensure that the research team adheres to the research protocol.

• To consider any new information of relevance to the research question.

• To suggest relevant literature.

• To provide detailed comment on any extension requests or substantial changes to protocol.

Membership

Chair: John Loan-Clarke (Senior Lecturer, School of Business and Economics, Loughborough University)

Meetings

Three meetings will be held over the course of the project. These will be scheduled to proceed the timing of reports submitted to the NIHR SDO i.e. in June 2011, December 2011 and June 2012. The precise dates and timing of the meetings will be agreed between the research team and Steering Group members. Meetings will take place in the School of Business and Economics at Loughborough University.

Communications

Communications between meetings will be conducted by e-mail. To reduce the burdens on Steering Group members, such communications will be kept to a minimum and mainly comprise documentation and reports sent to members prior to Steering Group meetings. On occasion, additional advice on specific issues will be sought from members between meetings.

The Role of Unlearning in Healthcare Managers' Decision Making

This project runs for 18 months and is funded by the National Institute for Health Research Service and Delivery Organisation. The Chief Investigator on the project is Dr Crispin Coombs, and Co-Investigators are Dr Donald Hislop and Dr Julie Holland, and the Research Associate is Dr Sara Bosley. The project is based in Loughborough University's School of Business and Economics.

Aims and objectives

The overall research aim of this project is to empirically investigate the extent to which healthcare managers engage in processes of unlearning, the barriers that prevent and the enablers that promote their engagement in unlearning. This topic is important because it will improve our understanding of how managers make decisions and will help provide relevant support and training for those involved in running the health service.

In addressing this aim the following specific research questions will be investigated:

1. To what extent do healthcare managers engage with the process of unlearning and what impact does the engagement with this process have on healthcare manager decision making?

2. What are the barriers that prevent healthcare managers engaging with the process of unlearning?

3. What are the enablers that promote healthcare managers' engagement with unlearning?

4. To what extent does the healthcare setting (acute care, mental health care or primary care) affect the level of engagement with the process of unlearning by healthcare managers?

Background

The concept of unlearning (a deliberate process of reflection which results in ‘giving up’ some existing knowledge deemed to be out-dated) is closely linked to processes of learning and change. Fundamentally, for change to occur unlearning must happen, as people need to acknowledge the limitations of existing knowledge/practice, be prepared to ‘give up’ such knowledge and adapt their knowledge/practice by actively learning and developing new ways of working/thinking. Thus, unlearning is arguably a necessary precursor to processes of learning (Akgünet al2007). Despite this, the concept of unlearning is under-utilised and poorly understood, both in the general literature on learning and knowledge management (Hislop, 2009), and also in the literature on learning in the health sector (Crillyet al2009; Rushmer & Davies 2004). This project will thus contribute to developing a greater understanding of this aspect of knowledge management and learning processes.

This research is also concerned with understanding the barriers to unlearning that exist. While unlearning has the potential to be an important component in the processes of learning and change, research suggests that there are many barriers to unlearning, which means that learning from mistakes, failure and crisis often does not happen (Baumard & Starbuck 2005). Barriers to unlearning exist at both individual and team/organisational levels. The small amount of research on unlearning (Rushmer & Davies 2004) and learning from failure that has been done in the health sector suggests that this is a work context where the barriers to unlearning can be significant. Further, this research suggests that in such contexts both individual level barriers (concern about negative consequences/sanctions from admitting mistakes, and a preference for local/contextual knowledge) and organisational level barriers (a culture where admitting to mistakes isn't encouraged) can be causes of unlearning (Dean, 2002).

However, the heath-care literature suggests that individual factors and contextual features may enable unlearning. To some extent enablers mirror barriers. For example, having an open culture, identifying root causes of errors and establishing systems that provide feedback to those involved in errors (Dean, 2002). More specifically an organisation may need to develop an ‘unlearning context’ (Cegarra-Navarroet al2010) which enables individuals to examine their perception and interpretation of situations, to change routines and to consolidate emergent understanding. In order to do this, individuals need the capacity to reflect on new ways of acting, to tolerate embarrassment, risk and threats to status (Rushmer & Davies, 2004).

Research Method

To empirically investigate the unlearning undertaken by healthcare managers this project will utilise a multiple case study approach, investigating experiences of unlearning in three different types of NHS trust (an acute trust, a primary care trust/Clinical Commissioning Group and a mental health trust). Conducting research in these different organisational contexts allows analysis of the extent to which the organisational contexts affects unlearning (research question 4). The reason for choosing these three types of context is that they represent some of the most important and common organisational types that the managers being investigated work in. A case study based approach represents a suitable methodology for the investigation of unlearning because, as outlined, unlearning is a relatively neglected and unexplored aspect of learning and change processes, and qualitative case studies provide an effective way to conduct exploratory research which can give rich, qualitative insights into managers' experiences of unlearning (Berg 2007; Hartley 2004).

Within each organisation, the main source of data will be face-to-face, one-to-one semi-structured, qualitative interviews with a range of healthcare managers. In each of the 3 case study Trusts 14 interviews will be conducted with healthcare managers. The research is fundamentally concerned with how the managers being investigated experience and make sense of the need to unlearn and adapt their approaches to decision making. Qualitative, semi-structured interviews provide an appropriate method for the collection of data on this topic, as they allow interviewees to articulate their experiences in their own words, and provide a mechanism via which interviewees can respond to issues raised by researchers, as well as introducing and discussing new topics, issues and experiences that they deem appropriate (King & Horrocks, 2009).

Case studies

At each case study site, the research team will recruit a sample of middle managers who vary in terms of clinical/background, type of department/unit and length of time as a manager.

To be included participants need to:

1. be middle managers. We expect the sample to comprise mainly Agenda for Change grade 7/8 staff, along GPs involved in the shadow Clinical Commissioning Group;

2. have some managerial responsibilities that involve making decisions about the use and allocation of resources, including staff and/or finances, data collection and management, service delivery etc.

3. be able to identify and be willing to talk about their experiences of unlearning; and

4. be willing to complete the pre-interview questionnaire, participate in an interview of approximately one hour in length and to give informed, written consent.

The research team will work with Human Resource Managers (HRMs) to identify the most appropriate group of managers to invite to participate in the project and the most effective way of engaging and recruiting them. The proposed recruitment process is as follows, but is open to discussion with HRMs:

• The researcher, Sara Bosley, will provide HRMs with an invitation and a Participant Information Sheet to be sent by e-mail to potential participants.

• If they are happy to do so, HR staff will send the e-mail and Participant Information Sheet to potential participants. They may also be requested to send a reminder if the initial e-mail generates limited interest.

• Those managers who are interested will be invited to reply to Sara Bosley by e-mail or phone.

• Sara will provide them with more information about the project and respond to any questions.

• Managers who volunteer to take part will be invited to complete a pre-interview questionnaire giving their current job title, employing organisation, clinical background and previous job titles. They will be invited to return the questionnaire as an e-mail attachment (password protected if they wish) before the interview and to give e-mail consent about the use of the data they provide in the questionnaire. They may return their completed questionnaire by post or at the interview, should they wish to do so.

• An interview date, time and venue will be agreed between the researcher and participant. We anticipate conducting the interviews at the participants' place of work unless they choose an alternative venue.

• The researcher will send to participants the main interview questions, examples of unlearning and a consent form before the interview.

• Before the interview starts, participants will have an opportunity to ask more questions and will be invited to sign the consent form.

After the interview, participants will be offered a copy of the interview transcript. They will be invited to attend a workshop in the spring of 2012 to hear about initial findings and to comment on those findings. However, taking part in the interview does not commit them to attending a workshop. Participants and case study contacts will be sent a summary of the final report and a web link to the full report.

References

Project title: Healthcare managers' decision making

Dear Healthcare Manager

I am writing to invite you to take part in a research study exploring healthcare managers' experiences of abandoning established ways of thinking and acting (unlearning), and how this affects their decision making. The study will help us better understand the processes of change, learning and decision making. We will use this knowledge to make suggestions about how healthcare managers can be better supported.

We are approaching you because of your position as a manager within one of the Trusts that has agreed to be a case site study. Taking part will involve completing a short pre-interview questionnaire and being interviewed for about one hour. In the interview we will ask you about:

1. Your work context

2. Your job role and responsibilities

3. Occasions when you abandoned some aspect or aspects of the way you act and think

4. Skills and knowledge that you once used but no longer use

Please read the attached Participant Information Sheet and then contact the researcher, Sara Bosley, if you are interested. Expressing an interest does not commit you to taking part. You can reach Sara by email on s.l.c.bosley@lboro.ac.uk or by phone on 01509 223177. Sara will tell you more about the study and answer any questions you may have. You can then decide if you would like to be considered for inclusion in the study.

Best wishes

Dr Crispin Coombs

Chief Investigator

Healthcare Managers' Decision Making

We would like to invite you to take part in our research study. Before you decide, we would like you to understand why the research is being done and what taking part would mean for you. One of our team will go through the information sheet with you and answer any questions you have.

Interview Schedule Healthcare managers' decision making

Instructions and prompts for the interviewer are in italics.

Introduction: check participant has received and read project information. Go through the information sheet and answer questions. Talk through consent form, and gain permission to record. Ask participant to sign the consent form. If the interviewee has completed a pre-interview questionnaire, refer to this to prompt responses to Q2.

Examples of abandoning established ways of acting and thinking

The Role of Unlearning in Healthcare managers' decision making

Part one: Name and contact information

Part Two: Academic and Professional Qualifications

Part Three: Current Job Role and Career

4) What are your main responsibilities in your current job?

5) What other jobs have you done?

Please record the titles and approximate dates of any substantial jobs have had in the table overleaf. You do not have to list all the jobs you have done or give a full CV.

Please save this with your name as part of the document title and return it to s.l.c.bosley@lboro.ac.uk or by post to Sara Bosley, Research Associate, School of Business and Economics, Loughborough University, Loughborough, Leicestershire LE11 3TU. If you wish to password-protect your questionnaire before emailing it, please let Sara know your password in a separate email or by phone (01509 223176). If you prefer not to send the questionnaire in advance of the interview, please give it to the researcher on day of the interview. Thank you for your assistance.