Article Text


‘Just so you know, the patient is staff’: healthcare professionals’ perceptions of caring for healthcare professional–patients
  1. Mia Svantesson,
  2. E Carlsson,
  3. M Prenkert,
  4. A Anderzén-Carlsson
  1. Faculty of Medicine and Health, University Health Care Research Center, Örebro University, Örebro, Sweden
  1. Correspondence to Dr Mia Svantesson; mia.svantesson-sandberg{at}


Objective To explore healthcare professionals’ conceptions of the care of patients who are also healthcare professionals.

Design Explorative, with a qualitative, phenomenographic approach.

Participants and setting 16 healthcare personnel within different professions (doctors, nurses, assistant nurses, physiotherapists, occupational therapists) were interviewed about the care of 32 patients who were themselves members of different healthcare professions, in one healthcare organisation in Sweden.

Results The care of patients who are healthcare professionals was conceived in five different ways, as: usual, dutiful, prioritised and secure, insecure and responsive. An initial conception was that their care was usual, just as for any other patient, and also a perceived duty to treat them and to protect their right to be a patient—as any other patient. Exploring further, informants described that these patients did receive secure and prioritised care, as the informants experienced making a greater commitment, especially doctors giving privileges to doctor–patients. A conception of insecure care infused the informants’ descriptions. This comprised of them feeling intimidated in their professional role, feeling affected by colleagues’ stressful behaviour and ambiguity whether the healthcare professional–patient could be regarded as a competent professional. The deepest way of understanding care seemed to be responsive care, such as acknowledging and respecting the patient's identity and responding to their wishes of how treatment was to be met.

Conclusions Caring for healthcare professionals seems to trigger different ethical approaches, such as deontology and ethics of care. According to ethics of care, the findings may indeed suggest that these patients should be cared for just as any other patients would be, but only if this means that they are cared for as persons, that is, they are given ‘person-centred care’. This would imply balancing between acknowledging the vulnerable patient in the colleague and acknowledging the identity of the colleague in the patient.

Statistics from

Strengths and limitations of this study

  • A major strength was the inclusion of informants from diverse healthcare professions caring for patients who were also members of diverse healthcare professions. As a uniform pattern was discovered in the results, this might reflect a shared interprofessional understanding of the care for healthcare professional–patients.

  • Another strength was the capturing of experiences connected with concrete patient situations, instead of capturing a general opinion of what could be regarded as the ‘right’ way to care for healthcare professional–patients.

  • One limitation might be the recruitment method—most informants actively communicated their interest to participate. One could speculate that these individuals had an urge to talk about these complicated relationships. However, this was the case in only half of the interviewed patient situations.


Caring for healthcare professional–patients might entail challenges other than those faced when caring for lay-patients, an observation that has foremost been reported anecdotally and almost solely in relation to doctors caring for doctor–patients. Empirical studies are needed to further investigate ‘doctoring’ doctors, but also to gain knowledge from other healthcare professions, as each profession constitutes a part of the wholeness of care. As there might be risk of emotional distress, knowledge of experiences of caring for these patients might entail awareness of one's own reactions and this might have implications for interprofessional learning in order to provide good care.

The experience of doctors treating other doctors has been an object of interest since the 1950s, but has been reported mostly in the non-empirical literature. A considerable amount of anecdotal reports have been published, such as case reports, correspondence, personal reflections and editorials,1–5 and some have even been published in high-impact journals.6–10 One older review was found, but it comprised mostly of mental health, non-empirical literature.11

However, four empirical studies were found about ‘doctoring doctors’12–16 (one reported in two papers: refs. 14 and 15). All but one focused on general practitioners’ experiences, that is, when the doctor–patients were not severely ill. The care could be experienced as rewarding,12 but provoked additional emotional responses such as anxiety, awkwardness, self-doubt, feelings of role ambiguity and difficulty in defining boundaries of the relationship.12–14 Behavioural responses of both overinvolvement and underinvolvement of the doctor–patient were experienced.13 ,14 Overinvolvement could comprise overidentification with the patient and being too sociable, as well as an urge to provide the best care.11 ,13–15 Underinvolvement13–15 implied care not being supportive enough17 and being suboptimal, experienced due to the doctor–patients’ own preferred, but inaccurate, treatment expectations.13 ,17

Despite the lack of empirical studies, authors have made recommendations that doctor–patients should be treated just like other patients,7 ,8 ,11 ,17 ,18 as also stated by the British Medical Association.19 One article summarised: “It is crucial they're treated as patients,”17 and other authors have emphasised that those patients should be given the same information as lay-patients.7–9 ,17–19 Furthermore, authors have suggested that doctors should avoid overly close identification with the patient,7 ,18 leave plenty of time for discussions and deal openly with the patient's anxiety.7 ,18 Finally, it has also been stated that doctor–patients have the same right to confidentiality as other patients.18 ,19

Only two case studies were found about the experiences of healthcare professionals other than doctors, those being, nurses caring for a nurse-colleague.20 ,21 Similar to caring for doctors, the nurses experienced anxiety and awkwardness.20 ,21 We have not been able to find additional studies of other healthcare professionals’ experiences of caring for a colleague, yet we find it likely that they also experience some kind of emotional reactions in this care.

In summary, to enhance understanding of the phenomenon of caring for another healthcare professional, there is a need for both a broader and deeper understanding of this care, through inclusion of all healthcare professions involved in the patients’ care and capturing conceptions connected to concrete patient cases. In this study, the term ‘care’ means treating, nursing or responding to a patient and also encompasses those significant meanings that the various healthcare professionals place in their interpretation of the term. Thus, the aim was to explore healthcare professionals’ conceptions of the care of patients who are also healthcare professionals.



An explorative design with a qualitative phenomenographic approach inspired by Marton,22 and Sjöström and Dahlgren,23 was applied, in order to describe various ways of conceiving and dealing with the phenomenon, and to investigate the conceptions’ internal relationships.22 According to Marton, in a group of people, there seems to be a limited number of qualitatively different ways of understanding, making sense of and dealing with a specific phenomenon, and that insight has implications for learning.24

The present study was part of a larger project describing healthcare professionals’ experiences of being a patient or being a family member of a hospitalised patient.25 It was initiated by the board of a university hospital in a Swedish County Region, with the purpose of promoting organisational learning through generating knowledge from the experiences of various healthcare professionals.

Setting and participants

The recruitment strategy was a passive invitation to all clinical healthcare personnel employed at three hospitals in one Swedish County Region (approximately 4900 persons), through announcements on the intranet and bill posting. The inclusion criteria were: having at least 1 year of professional experience and having contributed in the care given to a patient who is also a healthcare professional. Twelve persons contacted the project team via email. Nine of these met the inclusion criteria. To be able to recruit more persons and gain variation of healthcare professions and specialty, we purposively asked selected managers (eg, for rehabilitation) to recruit informants at workplace meetings. Eight additional persons then contacted the first author. In total, 17 healthcare professionals gave their informed consent to participate (table 1). All were interviewed, but there was one drop out from the interviews—an occupational therapist—due to technical problems with the audio recording, resulting in a total of 16 interviews to analyse (table 1).

Table 1

Characteristics of the informants and the patients discussed in the interviews

Data collection

Data were collected by means of semi-structured interviews by the first author, inviting the informants to both talk about their experiences and to reflect on them through a dialogue.22 Half of the informants chose to be interviewed in their workplace and half in the office of the first author. The informants each described caring for between one and three healthcare professional–patients (figure 1).

Figure 1

Illustration of the distribution of narrations of caring for patients of different healthcare professional backgrounds, for example: from the doctors, there was one narration about one assistant nurse–patient.

The main inquiry was; ‘Please, tell me how you experienced the care for this patient with a healthcare profession’. Additional questions centred on whether and how they perceived differences in the care of the healthcare-professional-patient compared to other patients. Follow-up questions were used, such as, ‘What do you mean by …?’, and ‘Can you say more about…?’.

A pilot interview (not included in the study) was conducted to test the interview questions, resulting in minor reformulations. The digitally recorded interviews varied between 30 and 80 min (mean 43 min) and were transcribed verbatim by an experienced research secretary.

Data analysis

The phenomenographic analysis was inspired by Marton,22 and Sjöström and Dahlgren.23 The focus of the analysis was to find the various ways of how the phenomenon of being a healthcare professional caring for a healthcare professional–patient was understood and dealt with.22 First, all data were read and re-read. Second, quotes expressing ways of conceiving the phenomenon were identified and formulated as short condensations or interpretations of conceptions, facilitated by the software program NVivo 10.

Third, similar conceptions were grouped together and reformulated, and preliminary categories of description across the interviews were developed. Effort was made to make sense of each particular conception in relation to the collective. All data that related to general perceptions which were not grounded in the experiences of the specific patient situations and the informant's interpretation of the patients’ experiences were omitted from the analysis.

The process of analysis was iterative, moving from the whole of the transcripts to the condensed, to the preliminary descriptions and back again to make sure that informants’ conceptions were correctly apprehended, and that all the ways of understanding the phenomenon, not only the dominant ones, were identified. Preliminary ‘categories of description’ emerged, which were discussed in an iterative process with the coauthors. After these explorations, the first author went back to the data, recategorised, reformulated or formulated new descriptive categories until final agreement was reached.

The categorisation resulted in the generation of a preliminary ‘outcome space’ (figure 2), where the internal relationships between various ways of understanding the care were described. This included the interpretation of a hierarchal relationship of levels of understanding; from surface to deep understanding.22 This implied recategorisations in the light of the whole and trying to grasp the relationships between the categories until a final agreement was reached among the authors. Also, the significance of the categories was assessed by determining the ‘frequency’ of quotes and their ‘pregnancy’, elucidating which aspects the informants had emphasised explicitly as being more important22 ,23 (table 3).

Table 2

Outcome space of the hierarchal relationship between various ways of understanding the care of healthcare professional–patients, formulated as categories of descriptions and conceptions

Table 3

The significance of frequency and pregnancy of the categories of description23 in the informants’ narratives: ‘+’ represents that the content of the category was mentioned only briefly and ‘++’ that the content emerged repeatedly and/or was explicitly described (n=number of quotes for each category)

Figure 2

Outcome space of the hierarchal relationship between various ways of understanding the care of healthcare professional–patients, formulated as categories of descriptions and conceptions.

Ethical considerations

An advisory statement specifying no objections to the study was provided by the Uppsala Regional Ethical review board (Dnr 2011/4589). A confidentiality issue identified was the fear of revealing sensitive matters related to a colleague and the risk of recognition, as the setting was only one healthcare organisation. Therefore, the informants were invited to read the manuscript and were asked about any evidence of this risk, which was refuted by the eight who responded.


Five ways of understanding the care of healthcare professional–patients emerged in the data; usual care, dutiful care, secure and prioritised care, insecure care and responsive care. All the conceptions were described by almost all informants, but insecure care was described most frequently (table 3). Usual care seemed to be the most spontaneous and unreflected way of understanding care, that is, on a surface level, and responsive care seemed to be the deepest way to understand, comprising all the other ways of understanding (figure 2). Illustrative quotes are presented in figure 3.

Table 4

Quotes for the categories of descriptions

Figure 3

Quotes for the categories of descriptions.

Usual care

In the initial stage of the interviews, informants mostly understood care as treating and responding to the healthcare professional–patient as they would treat and respond to any other patient. A conviction was described of the caring not being any different to that of lay-patients or that they had not reflected over whether it was different. Informants described recognising that the patient was a healthcare professional, but expressing not being influenced by this (figure 3, Q1). This seemed to be the surface (most shallow) way to understand the care.

Regarding patients who were recognised as colleagues in the hospital, the interaction was described as more personal, but informants felt that they could still make a distinction between the professional and the private role (Q2). Also, there was an experience of having less difficulty providing usual care to healthcare professional–patients who seemed to see themselves solely as patients, not as healthcare professionals.

Dutiful care

Providing equal care for healthcare professional–patients could also be understood as an obligation; a duty to treat these patients as any other patient. The focus was on the informant's own role. A major duty seemed to be to adhere to the ward routines and ‘the rule book’. This strategy could imply stepping into an even more professional role in order to prevent mistakes.

Professionalism was defined as being much stricter, being more objective in assessments, and being direct and clear, without being afraid of harming the patient's professional pride (Q3). There was a conception that it would be a relief if the patients removed themselves from their professional role and that the patients would get better care if staff realised that they were as vulnerable as other patients. Other reasons conveyed for treating the patient as they would treat other patients were related to patient safety and justice. This often entailed giving the same information to them as given to other patients.

Informants also felt a duty to seek support to be able to treat the patient as they would any other patient, and also a duty to influence and support colleagues to treat the patient as they would other patients (Q4).

Secure and prioritised care

The focus was partly on one's own role, partly on colleagues’ roles and also on the patient's role. There seemed to be an understanding that quality of care and security for these patients was influenced by the fact that the patient was a healthcare professional. This was particularly expressed when the patient was a doctor. Informants expressed being more inclined to be committed to healthcare professional–patients, that it was right that these patients received additional benefits. The efforts made for them could comprise shortcuts to investigations and treatments, letting the patient choose where to be cared for, and by whom and for how long. Informants described making these efforts in order to avoid mistakes, but also with the hope of treatment succeeding (Q5).

Colleagues were perceived to prioritise these patients as they kept reminding the informants that these patients were special, saying things such as, ‘just so you know, she is a doctor’, or writing ‘nurse’ in the box for diagnosis in the operation-chart. For the doctor–patients, privileges mostly comprised of more thorough investigations and treatment options, and doctors having more contact with the patient and his/her family on the ward. Nurses perceived that they received substantially more support from the doctors in this care (Q6). Junior doctor informants described a personal feeling of security when they involved other senior doctor-colleagues in the treatment of these patients, but they could also feel excluded when senior doctors themselves suddenly took over the responsibility for the care of the doctor–patients.

There was also a perception that these patients themselves contributed to providing both secure and safe care, in that their knowledge could be trusted, such as knowledge of self-care and needing less information and explanations. Experience of sharing professional knowledge with a patient, particularly when the patient had the same profession or was a workmate, was also expressed as contributing to a feeling of security (Q7).

Insecure care

The focus of this category was on the informants’ own feelings of insecurity when giving care, but also on their colleagues’ behaviour and partly on the patient's role. This category dominated the informants’ descriptions (table 3).

Informants could feel intimidated when learning about the patient's healthcare profession. Intimidation due to the patient's superior professional knowledge or status was perceived as fear of exposing ignorance, failing in front of the patient and an unpleasant experience of a swayed power balance (Q8, figure 3). For some, it felt as though they were forcing themselves to be natural, which in turn induced performance anxiety. Feelings of being too intimate with a colleague-patient were perceived as an emotional strain and there was a conscious endeavour to maintain professionalism. The feelings of insecurity comprised fear, embarrassment, concern and compassion when identifying themselves with the patients.

Many informants described feeling affected by their colleagues’ stressful behaviour, such as creating a stressful atmosphere around the patient and a loss of professional distance. They noticed how their colleagues’ change in attitude affected their own role. These situations could comprise feelings of being persuaded by others to sidestep routines, resulting in undertreatment (Q9).

Whether the healthcare professional–patients could be regarded as competent professionals and their knowledge could be relied on was perceived to generate feelings of ambiguity. Informants expected the patient to understand better than lay-patients and placed more responsibility on them (Q10). Surprise was described when these patients did not comply with the prescribed treatment. For instance, patients rejected or overdosed medicine or could not understand advice, such as why driving was hazardous after stroke. This was related to any healthcare professional–patient, but was reported especially in older patients, those affected by a deteriorating medical condition, or patients working within specialities other than the informant's own specialty. For instance, this could involve a patient on a psychiatric ward who was employed as a surgical nurse. Informants described raised insecurity and being shocked by the experience that trusting the patient's professional knowledge and overestimating the patient's understanding of their illness could be risky (Q11).

Responsive care

The deepest way to understand the care of patients with a healthcare professional background seemed to include those conceptions that paid equal attention to the particular individual's needs and one's own role in providing that care. This descriptive category focused more on the patient's role and well-being than on the informants’ own emotions and behaviours. Acknowledging and respecting the patient's identity encompassed mostly being sensitive to the patient's level of knowledge, adapting the medical language, as well as trying to connect to his/her knowledge. Informants were aware that it could be a risk to trust the patient's knowledge and that the patients therefore should be treated as lay-patients, but they expressed that their most important mission was to acknowledge the patient as a person (Q12). This centred on being respectful, but not necessarily following the patient's medical preferences; it was mostly about respect for the person. This involved respecting the preferred identity that the patient wished to reveal and included respectful tactics such as giving the same information as that given to other patients but presenting it as ‘this you probably already know’. Another respectful strategy that the informants adopted was to let the patient ask questions or to give them the opportunity to demonstrate their knowledge by asking them questions.

Responding to the patient's wishes and how these were to be met could mean being sensitive to changing desires, such as shifting between being professional and private. The most dominant adaptation related to providing information. Informants were sensitive to the variable wishes of the patients of either wanting or not wanting to speak in medical language (Q13). From their experiences, they could conclude that these patients should be treated individually, which implied responding to them as healthcare professionals, patients, or both, and being sensitive to their fluctuating wishes.


A broader perspective, including the conceptions of a variety of healthcare professionals, offers a more complex picture of caring for a healthcare professional–patient than has previously been reported.7 ,8 ,18 ,19 The findings are, however, in line with previous studies reporting evoked anxiety in this care,12–14 ,20 ,21 which, in the present study, was understood as insecure care. This was particularly apparent in the care of doctors, which further supports the notion that care of a doctor is special.

A salient finding was how awareness emerged during the interviews—from stating that these patients are and should be treated just like other patients, to an understanding that they do receive prioritised care. It appears healthcare professional–patients can be subjected to either secure or insecure care; secure care compared to the lay-patient regarding resources, but less secure care due to the professionals’ insecurity. Awareness emerged that responsive care is important for these patients, with a focus on individual needs. This could mean acknowledging either the patient in the colleague or the colleague in the patient. Throughout the findings, ethical issues arose around how these patients should be cared for and therefore the findings will be interpreted from an ethical perspective.

There was a notion among the informants of the principle of a duty to not discriminate against these patients and it seemed to emanate from a conviction that usual care was the best care for healthcare professional–patients. However, justice-related reasons were also given to prioritise them. It could be argued, given that the experiences of insecure care drastically outweighed secure care, that healthcare professional–patients might have different emotional needs.26 These needs might be because of more input due to them having too much knowledge, which might cause an increase in their apprehension.27 To not acknowledge colleague patients might lead to worse care, which in turn would imply injustice.

The conception of prioritised care was judged as either right or wrong by the informants; wrong when relating how colleagues positively discriminated healthcare professional–patients. However, the conception also entailed aspects of secure care, because while informants perceived that such discrimination was wrong, they also described that positive discrimination contributed to secure care. The judgements concerned mostly ‘doctoring’ doctors and might support the notion of overinvolvement and underinvolvement, as previously reported.11 ,13–15 The most complicated relationship seemed to be between a junior doctor and a senior doctor–patient, a finding that is consistent with those of other studies.11 ,17 Teng et al12 showed that, with increasing experience, anxiety decreased, so that might be a reason to appoint senior doctors to care for doctor–patients. However, such practice might imply unjust positive discrimination.

There was ambivalence in relation to whether or not the patients’ professional competence could be trusted. Informants were mostly judging the patient's trustworthiness, but this can also be interpreted as an analogy of Aristotle's golden mean; that this virtue exists between two extremes.28 Being too trustful of the patient's knowledge might lead to unsafe care, a finding that has also been previously reported.13–15 Being too mistrustful might lead to degrading of the patient's integrity. This might be described as an ethical conflict between sacrificing the value of control and sacrificing the value of the patient's autonomy in shared decision-making.3 In a recent study, general practitioners’ strategies for doctor–patients were either to ignore or acknowledge their background, but there is no evidence about which strategy is most beneficial.16

Responsive care seems to be the deepest way of understanding the care of these patients, as it may encompass all the other ways of understanding the care, but it also adds taking a non-judgemental perspective of the patient. The other conceptions in this study were more focused on deontological right or wrong, while this conception was more in line with the theory of ethics of care, expressed by Nortvedt et al29 as: “with the right intention and attitude, and respecting persons for who they are, not merely because they are representatives of a particular role (eg, being a patient) is important”.

Ethics of care may lead us to the notion of ‘person-centred care’, as an ethical idea to treat the patient as a person.30 In connection with responding to the healthcare professional–patient, either as a patient or as a colleague, the aspect of understanding the whole of the person30–32 and respect for “who they think they individually are”31 ,33 seemed to be essential. There are, however, different definitions of person-centred care,31 and some question as to whether it implies respecting the patient's autonomy and preferences.30 ,33 In the present study, respect for patients’ medical preferences did not seem to be included. Responsive care centred more on respecting the identity that the patient wished to reveal, as also reported by others.30 Thus, the definition of person-centred care that is most in line with the present findings seems to be “comprehensive care that meets each patient's physical, psychological, and social needs” (p.8).31 For a patient whose profession is an important part of their identity, there might be a risk of infringing on the person's integrity and dignity if they are treated the same as any patient. The reverse is also true—their integrity and dignity might be compromised by revealing sensitive psychosocial facts about the patient beyond those that are medically motivated.32

A strategy used by the informants in this study, in their interactions with healthcare professional–patients, was to ask questions to acknowledge the patient's knowledge. This strategy was also identified in Fox et al's study of former doctor–patients.13–15 Grounded in their own previous experience of disempowerment as patients, the doctors had been more motivated to ask questions instead of only informing, in order to empower the patients.13–15 Person-centred care may not contradict the previous recommendation of treating the healthcare professional–patient in the same manner as other patients, as long as other patients are also acknowledged for their own preferred identity, such as being an engineer, mother, dog-breeder and suchlike.

Finally, information has previously been reported as an important issue,7–9 ,17–19 and recurred in most of the informants’ narrations in this study. They had different perceptions about whether or not the information should be adapted to suit the patient who is a healthcare professional. When practicing responsive care, informants expressed sensitively adapting the information to the perceived identity and perceived capacity of the patient while conforming to relational-oriented ethics.

Strengths and limitations

It could be argued that the heterogeneous sample of a variety of healthcare professions might be considered as a weakness, as the professions have different roles in caring. However, despite this and also other variations, in age, specialty and profession of the patients, a uniform pattern was identified in the informants’ conceptions (table 3), which we consider to be a major strength in the study. From the perspective of the phenomenographic method, the variations seemed to exist within the individual healthcare workers and were connected to their relationships with different patients. The uniform pattern might reflect that healthcare professionals from different professions share an understanding of the care that should be provided to healthcare professional–patients.

Another strength was that capturing the experiences connected to concrete patient situations rather than simply generally describing how informants dealt with these patients contributed to a rich understanding of the phenomenon. An argument to be made is that there is a stipulated difference between descriptions of how you have dealt with a concrete situation compared to how you perceive that you generally act. There were general reflections made in the interviews, but these were omitted from the analysis, according to phenomenographic method.

One threat to the trustworthiness of the study might be the recruitment method—informants actively communicated their interest to participate. One could speculate that these persons had an urge to narrate about complex relationships. This was the case in half of the interviewed patient situations, but some informants did describe both, complicated and uncomplicated relationships.

The first author constantly probed whether the informants’ perceptions were connected to the fact that the patient had a healthcare profession. One might argue that this might have provoked the informants to focus more on the differences between this care and that provided to lay-patients. However, if this follow-up question had not been used, it would have been impossible to know whether the conception was connected to the phenomenon or not.

Conclusion and relevance to clinical practice

Caring for healthcare professionals seems to trigger different ethical approaches, such as deontology or ethics of care. According to ethics of care, the findings may indeed suggest that these patients should be cared for just as any other patients would be, but only if this means that they are cared for as persons, that is, they are given ‘person-centred care’. This would imply balancing between acknowledging the vulnerable patient in the colleague and acknowledging the identity of the colleague in the patient. To improve care for healthcare professional–patients, we suggest a qualitative meta-analysis of studies of the experiences of both, healthcare professional–-patients and healthcare professionals, in order to connect their perspectives.


The authors would like to thanks the informants, who sacrificed their time to tell us their rich stories and also to Aileen Ireland for valuable proof reading.


View Abstract


  • Contributors MS, AA-C, EC and MP designed the study. MP was responsible for the literature search. MS conducted all the interviews, except one, which AA-C conducted. MS was responsible for the analysis, but all the coauthors (AA-C, EC and MP) participated in the co-assessment of the different steps of analysis and the formulation of categories of descriptions. All the authors (MS, AA-C, EC and MP) were involved in drafting the work or revising it critically for important intellectual content and final approval of the version to be published.

  • Funding This work was supported by Orebro County Region Research Committee for funding, (grant number OLL-393461 and OLL-238541, grant number 6930 GBP).

  • Competing interests None declared.

  • Ethics approval Uppsala Regional Ethical review board, Sweden (Dnr 2011/4589).

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data sharing statement No additional data are available.

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.