"I should have discharged him but I felt guilty": a qualitative investigation of clinicians' emotions in the context of implementing occupational therapy

Rehabilitation clinicians' delivery of good-quality care can be influenced by a wide range of factors. A recent review of frameworks about factors influencing delivery of care [1] lists 57 clusters of factors, some of which (e.g. clinicians' knowledge and skills [2]) have already received considerable research attention. Strategies to facilitate delivery of good-quality care have also been identified (e.g. educational materials [3], and audit and feedback [4]). However, a substantial proportion of the variation in clinicians' delivery of care remains unexplained [5], and the effects of known strategies for improving the quality of care are small [3],[4]. There is a continuing need to identify more effective ways to improve health care delivery, including delivery of rehabilitation care. One part of addressing this need is to investigate under-researched factors related to clinicians' practice and delivery of good-quality care. The results of such studies can be used to identify new strategies that are as yet untested [6].

One under-researched factor is clinicians' emotions concerning their practice and delivery of care [5]. For example, a search of Implementation Science in July 2014 for papers with titles including any of the terms ‘emotion’, ‘feel’, ‘feeling’ or ‘affect’ yielded no papers, and the UK National Institute for Health and Care Excellence guidance [7] on changing professionals' practice makes no reference to clinicians' emotions. Considering that positive and negative emotions are known determinants of people's actions in general [8]-[10], and that clinicians are people, it is plausible that clinicians' emotions may play a role in their practice. This hypothesis is also supported by the few studies that have included clinicians' emotions among the factors that they have investigated. A study of primary care practitioners [11] identified clinicians' feelings of embarrassment as a possible determinant of the clinicians' avoidance of discussions about sexual health with patients, and a study of mental health workers [12] identified that clinicians' feelings of fear and wariness were a possible determinant of their reluctance to provide treatment to patients. A small number of studies with medical doctors have also found that clinicians' negative emotions concerning a perceived medical error can affect their future practice, sometimes for years after the event that initially aroused the emotions [13]-[16].

While the evidence described above provides support for the general hypothesis about the role of clinicians' emotions, it provides little evidence about the nature and context of the emotions. In other words, questions about what emotions clinicians feel in implementing care (e.g. happiness, satisfaction, anger) and what situations trigger these emotions (e.g. patient encounters, team interactions) remain largely unanswered. Yet, it is this specific evidence that is required for emotions to be considered as a target for implementation interventions. Early evidence of this kind is typically obtained from open-ended, qualitative studies. However, qualitative studies of clinicians' emotions have so far largely focused on emotions as `emotional labour' or `emotional skills' [17],[18], that is, as something used to manage other people's (usually patients') emotions and behaviours. The aims of the present paper are to (i) provide examples of the nature of clinicians' emotions and (ii) identify clinical situations in which, according to clinicians, emotions emerge. This evidence about the phenomenon (i.e. what emotions clinicians feel and in what situations these emotions emerge) contributes to developing theory about clinicians' emotions in the context of implementation of good care. Developing such a theory is the recommended first step to developing effective implementation interventions [6],[19].

This was a qualitative, semi-structured, face-to-face interview study that was part of a wider mixed methods programme of research [20] about clinicians' practice in one clinical speciality (paediatric occupational therapy). The programme of work (i) identified factors related to clinicians' practice that may explain—qualitatively or quantitatively—variation in the quality of patient care and (ii) designed an intervention to target these factors. The present qualitative study was part of the identification of factors and focused on generating evidence about clinicians' views, their practice, and the context of their practice including any emerging, previously unidentified issues. The programme of work adopted a pragmatic [21] perspective on evidence; this meant identifying `what might be important and might work' [21],[22] in the context of applied clinical practice. As the first exploratory step of the programme, the present study used an inductive approach—that is, we used the data to identify themes and categories, and then compared these with existing literature to identify relevant theories and frameworks that could be used in subsequent steps (see also ‘Discussion’). The programme of research had National Health Service (NHS) Research Ethics Committee approval (07/S0801/55).

All 26 clinicians agreed to participate. Due to scheduling difficulties, only 25 were interviewed. All participants were of senior grade, had been qualified for a substantial length of time (median 12 years, interquartile range [IQR] 9-20) and had several years of experience working with children (median 8 years, IQR 6-13).

Several themes emerged from the transcripts. One of the key themes was "negative emotions" and is reported here—other themes ("beliefs about responsibilities", "aims of therapy" and "structure of the therapy process") have been previously reported elsewhere [30]. Examples of clinicians' positive emotions were actively sought from the transcripts; however, descriptions of positive events were not accompanied by clinicians' expressions of their emotions.

Within the theme of negative feelings, four clusters emerged. The first three related to the types of emotions and circumstances in which these emotions emerged: (1) guilt and anger in relation to perceived failure, (2) worry and ‘niggle’ in relation to possibility of failure and (3) frustration and inadequacy in relation to perceived failure to achieve positive outcomes. The content of these clusters was highly consistent across contexts and clinicians. The fourth cluster, (4) emotions and practice, was more tentative—it was covered in fewer interviews and the narratives between clinicians diverged. The four clusters are elaborated on below.

The present paper reports on (i) the nature of clinicians' negative emotions about practice, including feelings of guilt, worry, inadequacy, frustration and anger, and (ii) three typical practice situations in which, according to clinicians, the negative emotions emerge: the clinician perceives s/he has failed a norm, value or standard they hold as important; the clinician perceives s/he is at risk of failing such a norm, value or standard in the future and the clinician feels unable to achieve positive outcomes for the patient.

The results expand the findings from previous studies and contribute to the field in several ways. One key contribution is to the way clinicians' emotions are theorised and subsequently operationalised for investigations and for implementation interventions. So far, literature in implementation science has identified, at a general level, that emotions may explain practice [11],[12]. In this literature, emotions are conceptualised as evaluations of affective consequences or outcomes of a behaviour, e.g. `discussing sexual behaviours with a patient [the practice action] makes me feel awkward/embarrassed [emotional consequence/outcome]' [11]. This reflects the social cognitive view of affect and behaviour [31],[32]. Social cognitive models are a common approach to investigating clinicians' practice [5]; however, there are significant limitations to their usefulness and a need to explore other approaches [33]. The evidence from the present study indicates that evaluation of consequences and outcomes covers only one aspect of clinicians' lived experiences of emotions. From the present results, a revised conceptualisation of emotions can be proposed: clinicians' emotions emerge from a combination of (i) clinicians' commitment to certain outcomes, standards or norms and (ii) a perception that they are failing to achieve these outcomes, standards or norms. Specifying clinicians' emotions in this way mirrors the literature about theories of self-regulation of human emotions and actions [34], including notions that (i) desired standards, norms and outcomes have an important role in the emergence of emotions; (ii) a person's evaluation of the rate of their performance against the desired standards, norms or outcomes produces positive or negative emotions and (iii) emotions `feed forward' to shape future actions [34],[35]. The revised specification has direct consequences for how emotions should be considered for elicitation, observation and measurement in future research. The specification indicates that research into clinicians' emotions should incorporate an investigation of the standards, norms and outcomes they hold as important, where they view themselves in relation to these standards, and how this relates to their feelings about the practices they are expected to do. The specification also indicates a need to expand the theoretical basis informing investigations of clinicians' emotions to include self-regulation as well as social cognitive theories.

Another key contribution of the present study is the identification of specific emotions likely to be relevant to investigations of clinicians' practice and implementation of care. The results from the present study provide an evidence-based description of a range of emotions likely to be relevant to clinicians' practice and a description of the situations where these are most likely to be observable. The results specifically indicate that future studies should pay particular consideration to how clinicians manage their own negative emotions in situations where they are required to consider their practices in the context of (patient or other) outcomes, norms and standards.

The results also highlight challenges for future research investigating clinicians' emotions and implementation practices. First, a key characteristic of the situations in which negative emotions emerged was that the clinician perceived a failure. (Whether or not there was also an objectively observed failure is left open.) This is an important finding for future research as it indicates that accurate identification of emotion-inducing situations relies on the clinician's willingness to disclose their perceived failures or perceived anticipated failures to the researcher. Second, the relationships between clinicians' emotions and practices were often implied rather than directly discussed. It is possible that this was because clinicians were hesitant to discuss these issues openly; however, it is equally likely that the relationships between emotions and practice are not easily accessible to clinicians and not easy to articulate'this is a known feature of self-regulatory processes [34]. Together, the first and the second point indicate that while some evidence about the nature, content and context of clinicians' emotions can be elicited by interviews, other modes of data collection may also be required to explore the relationships between emotions and practice. A solution could be, for example, the use of longitudinal, mixed methods data collection across individual clinicians' care delivery situations for a period of time. The results of the present study are of particular importance for planning such investigations as they provide a starting point for specifying the target and context of observations, i.e. the emotions and care delivery situations.

In terms of strengths and limitations, a major strength of the present study was that participants discussed their negative emotions about practice unprompted—this lends support for the credibility and confirmability of the results [26]. Another strength was that the participants were recruited across organisations, supporting relevance of the results across organisations. The main limitations were that the participants came from a single discipline and that the study relied on a single mode of data collection (interviews). The present study does not provide evidence about the direction or nature of relationships between emotions and practice or that clinicians' emotions cause undesirable practice. However, the results contribute to the development of the evidence and theory about the phenomenon of clinicians' emotions in the context of implementation of health care. This is the essential first step for investigating causal processes between clinicians' emotions and practice [6],[19].

Clinicians experience a range of negative emotions about practice. These are particularly likely to emerge in situations where clinicians' perceive that their actions and practice fall short of the standards, norms or outcomes that they hold as important. The results inform the specification of emotions and emotion-triggering situations for future investigations of health care implementation.

NK designed the study and conducted all data collection, conducted the initial data analysis, and wrote the first draft of the manuscript. NK and JM discussed the data and the first draft of the manuscript, and JM made substantial comments. NK and JM revised the manuscript, both authors providing substantial intellectual contributions on all drafts. Both authors approved the final version of the manuscript and take responsibility for its content.

NK is a Senior Research Fellow and a practicing Senior Occupational Therapist (Honorary). NK has PhD in Health Services Research; she has led the programme of work within which the described study resides since 2006. JM is Clinical and Professional Lead, and Service Manager, in a children's community occupational therapy. JM has MSc in Professional Health and Care Studies and is an active collaborator on a number of allied health care studies.