Exploring the experience of weight loss in people with advanced cancer


  • Jane Hopkinson PhD RN,

  • David Wright PhD,

  • Jessica Corner PhD RN

Jane Hopkinson,
Macmillan Research Unit,
University of Southampton,
Southampton SO17 1BJ,
E-mail: jbh@soton.ac.uk


Aim.  This paper reports a study of the experience of and concerns about weight loss described by patients with advanced cancer, their caregivers and nurse specialists.

Background.  Weight loss is reported to be one of the commonest symptoms experienced by patients with advanced cancer. There is evidence that it can be of concern to patients and their caregivers. However, little is known about why this is the case or how people might be helped to live with the symptom.

Method.  An exploratory study with a purposive sample of 30 patients, 23 caregivers, and 14 specialist nurses from the South of England was conducted in 2003. The in-depth interviews focused on the experience of weight loss and its management. Interviews were transcribed verbatim, then analysed using an approach informed by Wolcott's framework for qualitative data analysis and Miles and Huberman's ‘mixed strategy for cross-case analysis’.

Findings.  Concern was experienced when advanced cancer became visible through weight loss. Visible weight loss symbolized proximity to death, loss of control and both physical and emotional weakness. Despite this, weight loss was not routinely assessed by palliative care nurse specialists, who, like others in the patient's social network, respected a weight loss taboo in the belief that little could be done to help people live with the symptom.

Conclusion.  Weight loss-related concern might be mitigated if clinicians adopted a systematic and proactive approach to the management of the symptom that breaks through the weight loss taboo.

What is already known about this topic

  • Weight loss is one of the most commonly experienced symptoms in people with advanced cancer.
  • Nutritional and pharmacological interventions for weight loss in people with advanced cancer are of limited value in arresting or reversing the symptom.
  • The reasons why weight loss can be of concern are unknown.

What this paper adds

  • Weight loss can be a source of concern for patients with advanced cancer when it becomes visible and leads to change in the lived experience of weight.
  • Offers a conceptual model of the experience of living with weight loss and advanced cancer.
  • Weight-related concern may be ameliorated by adopting a systematic and proactive approach to the management of the symptom.


Weight loss is one of the most common symptoms experienced by patients with advanced cancer (Addington-Hall & McCarthy 1995, Vainio et al. 1996, Potter et al. 2003). As new treatments improve survival time, it seems likely that more patients will be living not only with cancer, but also with weight loss. Yet little is known about how weight loss is experienced (Poole & Froggatt 2002a,b), or how patients might be best helped to live with the symptom.


The biomedical approach to the weight loss caused by advanced cancer has focused on working towards arresting or reversing cachexia. It is now widely agreed that neither pharmacological manipulation of metabolic processes using corticosteroids, progestational drugs, or prokinetic drugs, nor nutritional supplementation can achieve this end (Tisdale 2002, Strasser 2003). Currently, a way forward for management of cachexia is being sought by testing approaches that combine pharmaceutical agents, nutritional supplementation and energy conservation (Brown 2002, Strasser 2002).

In the literature, a number of interventions other than pharmacological, nutritional supplementation and energy conservation are proposed as having the potential to support patients with cachexia. Firstly, physical exercise may be beneficial, even for those confined to bed (Brown et al. 2003). It has been found to increase lean body mass with no adverse side effects in people with acquired immune deficiency syndrome (AIDS) (Arey & Beal 2002). Secondly, if cachexia is a chronic inflammatory condition, then certain foods with anti-inflammatory properties may be helpful in combating its impact e.g. anti-oxidants and polyphenols in fruit and vegetables (McCarthy 2003). Thirdly, psychological and behavioural interventions, such as attention to ambience at meal times and relaxation techniques, may help (Higginson & Winget 1996). Finally, written information and nutritional counselling may be important (Strasser 2003). An appreciation of patients concerns about weight loss would provide insight into the likely ability of these proposed interventions to enhance well being of people living with advanced cancer.

A systematic literature search of electronic databases (PubMed 1966–2005/03; CINAHL 1982–2005/03; EMBASE 1980–2005/03; CancerCD 1990–2003/09) identified four studies across the international literature of the impact of unintentional and non-treatment related weight loss in patients with advanced cancer (Ovesen et al. 1993, Sarna et al. 1993, O'Gorman et al. 1998, Davidson et al. 2004). Weight losing patients have reported poorer quality of life than weight stable patients in both a study of 107 pancreatic patients receiving nutritional intervention in Australia and 119 gastrointestinal patients in the United Kingdom. Likewise, in Denmark, 48 weight losing patients with breast, lung and ovarian cancer reported poorer quality of life than 56 weight stable patients, irrespective of functional status. But in the United States of America, overall symptom distress in 28 people with lung cancer was greatest in those who had lost most weight and who had greatest functional impairment. Quality of life tools were used in these studies to examine the experience of weight loss and highlight associated problems with physical functioning and general symptom distress. However, the assessments reveal little about the concern that weight loss might precipitate for patients, or healthcare professional management of concern. Several authors have proposed that intractable weight loss may have a negative impact upon a patient's body image, their sense of control, and may act as a signal of advancing disease (Ottery 1995, Higginson & Winget 1996). These claims have yet to be formally investigated in an empirical study.

The study


The aim of the study was to explore the experience of and concern about weight loss in people with advanced cancer, their caregivers and nurse specialists. It was hoped that insights emerging from the study could provide a foundation for developing interventions to help patients live with weight loss.


The findings reported in this paper are from the interview data of a larger mixed methods exploratory study of two palliative care teams delivering home care to patients and their families. The teams served a population of just over one million people in the South of England in 2003.

Interview data collection and analysis were informed by Paterson and Zderad's (1976) Humanistic Nursing Theory. The theory sets out a philosophy for nursing, a way of practicing as a nurse, and a methodology for research. It is based on thinking drawn from existentialism and hermeneutic phenomenology (Hopkinson 1998). Nursing is conceptualized as a response to a human situation that includes a lived dialogue, which has the purpose of enhancing the well being of the patient through revealing and supporting choice. Research can adopt the same approach for knowledge generation, but the theory does not prescribe method, making it possible to incorporate ways of working with information that have been articulated by other methodologists.


Each patient under the care of two community palliative homecare services was approached and invited to take part in the larger study. Of the 233 people approached, 199 (response rate 85%) agreed to complete a short questionnaire that assessed concern about self-perceived weight loss and 143 (74·9%) gave consent (on the questionnaire) to be interviewed. The questionnaire included statements such as, ‘I am concerned about my weight’, to which the patient gave a yes/no response.

Initial contact with potential interview participants was made by telephone. If they agreed, the patient was mailed a letter of introduction to the study and an information sheet explaining the nature and purpose of the project. They were given a minimum of 24 hours to consider their decision. Of the 32 patients approached, one decided not to take part and a second died before the agreed interview date.

These interview participants were sampled purposively. The intention was to recruit as diverse a group of patients, across age, gender and diagnosis as possible, in order to strengthen the rigour of the work in terms of the likely representativeness of people with advanced cancer in general. Interviews were conducted with 30 patients, of whom 29 reported losing weight on the questionnaire. Of the group sampled, ages ranged from 43 to 85 years (median 70 years) and 14 were women. The primary site of cancer fell into the following groups: breast cancer n = 5; lung cancer n = 4; prostate cancer n = 4; head and neck n = 4; gastrointestinal n = 6; others n = 7. None were receiving active treatment or artificial feeding at the time of interview. Within 7 months of the interview 19 participants had died (median time to death for this subgroup = 2 months).

Additional interviews were conducted with 23 caregivers (14 in the presence of the patient) and 14 nurse specialists. All were nominated by patient participants and none refused to participate.


Findings of the literature review (Poole & Froggatt 2002a,b) and the expert opinion of the project steering group members informed the development of the interview schedules. All interviews were semi-structured, lasted 20–60 minutes and were tape recorded with consent. Each interview began by the researcher stating her particular interest in weight loss and then invited the participant to talk about their illness. The pattern of the interview followed the particular concerns or aspects of the illness experience relating to weight that were of importance to the patient. At the end of the interview patients were asked what had helped them live with any described changes and concern. The caregiver and healthcare professional interviews were conducted using a similar approach.

Ethical considerations

The study was approved by two Local Research Ethics Committees and conformed to local healthcare organisations’ research governance requirements. Study participants were approached as described above, and gave written consent to participation.

Data analysis

The analysis and interpretation was informed by Wolcott's framework for qualitative data analysis (Wolcott 1994) and conforms to Miles and Huberman (1994)‘mixed strategy for cross-case analysis’. Familiarity with the data was achieved through verbatim transcription, reading of the transcriptions and through cycles of coding. Each cycle of coding was driven by a study question:

  • Cycle 1. What are the manifestations of weight loss?
  • Cycle 2. What is the current management of, or response to, weight loss?
  • Cycle 3. What is the meaning of weight loss?
  • Cycle 4. What contextual factors influence the experience of weight loss?

Categories and themes were developed inductively from the codes allocated during the coding cycles. Data summaries were then devised in the form of maps of overarching themes that captured all coding for the symptom of weight loss for each of the participant groups; patients, caregivers, and professional carers. These three maps formed the basis of the interpretive process of recognizing relationships between the overarching themes, which were then represented by a ‘Model of the experience of weight loss’. They also revealed that there were examples of concerns categorized under each of the themes, exposing multiple sources of concern. The data were analysed using NUD*IST6 QSR.


The model was checked and refined through a search for disconfirming evidence (van Manen 1994, Miles & Huberman 1994) by returning to the original interview transcripts, patient records and field notes collected during 150 hours of contact with study participants. It was also tested for credibility through discussion with academics and clinicians.


All interview data relating to weight loss were captured by themes. The themes represented components of the experience of weight loss; ‘visible weight loss’; ‘measuring proximity to death’; ‘personal responsibility for weight’; ‘physical weakness’; ‘emotional weakness’; ‘helplessness’; and ‘the weight loss taboo’. Each theme is described below, with examples of how it can precipitate self-perceived problems or troubling thoughts, in other words, concern from the perspective of patients, caregivers and nurse specialists. To draw attention to ways in which others can shape patient experience, helplessness and the weight loss taboo are supported by data from caregivers and nurses rather than patients. All participants were given pseudonyms during the transcription process to protect their anonymity. The themes and relationships between them are then presented as a model of the experience of weight loss.

Visible weight loss

The patients interviewed were aware that they had lost weight because their clothes were no longer a good fit. Sid knew that he had lost weight because of his belt:

My waist…you can tell by your belt the notches in the belt.

Some of the patients had not expected to lose weight. Their stories included accounts of feeling shocked by either looking different or the rapidity of the change. People who did not think of their cancer as an eating related disease expressed surprise. For example, Rita had breast cancer, and so found it difficult to understand why she was losing weight:

I always thought cancer patients lost weight because they had it in the stomach and they couldn't eat.

Concern about weight was expressed when its visibility spoke to others of illness. Visible weight loss was troublesome because of the ideas and beliefs that it evoked around death, personal responsibility for weight and both physical and emotional weakness.

Of the 15 patients who spoke of concerns about their weight, all but one also spoke of the visibility of their illness through weight loss and all appeared wasted to the researcher: they had ill-fitting clothes or protruding bones. The primary reasons given for concern were; thoughts of prognosis (n = 5); changed appearance (n = 2); uncertainty about how to maintain or gain weight (n = 4); weakness (n = 4). For example, Frank talked about the visibility of his weight loss symbolizing approaching death:

You don't necessarily have to get on the scales, you see the bones begin to protrude and feel the end is nigh.

All but one of the caregivers of patients who remarked on the visibility of their weight loss mentioned the change in the patient's appearance. A further five relatives spoke of the patient's wasted appearance when it was not mentioned as a concern by the patient. The patients who were visibly wasted but were unconcerned, explained that they had expected to lose weight, that their weight was now stable and that they would only be concerned if they lost more weight. Lack of concern was expressed through comments on the health benefits of losing weight, improvement in appearance, or not having troubled to think about it:

I do feel better for losing weight…It does help me move a bit better. (Beatrice)

Proximity to death

Patient participants had questions about the prognostic implications of their weight change. Visible weight loss signified a close proximity to death. Thelma expressed this clearly:

Interviewer: What sort of things does losing weight make you think about?
Thelma: I know I am going to die soon; there is nothing anybody can do to help me.

Weight loss could prompt concern about approaching death, but it could also be helpful in estimating time to death. Paul recognised that weight loss told something of prognosis. He was troubled that no one would tell him its meaning in terms of how long he was likely to live:

It would be wonderful if someone let me know ‘look you have lost 5 kgs so therefore the end is not far away. Perhaps with a little bit of luck in 3 weeks time if you have lost another 5 kgs that will be the end, because your body will not be able to sustain such a waste of weight’.

Personal responsibility for weight

All but three of the patients spontaneously estimated their weight loss during the interview, 21 giving precise measures in stones or kilograms of both their weight pre-diagnosis and on the day of the interview. Monitoring weight on a regular basis was, for most, no more than the continuation of a pre-illness habit. Doris weighed herself daily, which she had done for many years, regulating her eating habits on a daily basis according to her weight:

I weigh myself. This is something that I do (every day), I always keep myself weighed…I have done for years.

So, weight was viewed as something under personal control in health. For some people this personal responsibility for their weight carried through into their experience of advanced cancer. Paula felt personally responsible for ensuring that she took enough calories to maintain her weight, implying it is a failing not to take control in this way:

I am concerned that weight just seems to be falling off and I am not getting enough calories. So I try to make sure that I have at least one drink of Complan® or Build-up® (proprietary nutritionally reinforced drinks) every day as well as anything else that I can take just to give me the calories and just to give me the vitamins that I feel are necessary.

Physical weakness

Weight loss was also troubling for patients because it was accompanied by weakness, leading to loss of independence and restricted activity. Victor was troubled because weight loss had led to weakness, which meant that he was only able to go out of the house in a wheel chair, which his wife had found she was unable to push. She explained:

I can't make out why he's losing weight every day. Because he can't walk around now, he can't walk around without a chair. He can't go out. (…)
He ain't got no strength at all.

When patients became physically weak, this led to them becoming different in other ways. They became unable to continue with social activities. Leaving the house became increasingly difficult and driving became unsafe leading to progressive social isolation.

Emotional weakness

When visible, weight loss symbolised illness but also spoke of undesirable personal qualities. It was interpreted as evidence of emotional weakness in the face of cancer. Metaphors were used to illustrate the impact of the cancer on weight, which implied emotional, as well as physical loss.

Steven explains how he felt about his cancer becoming visible through his weight loss. He found it distressing because it symbolized the cancer taking control of the body and told the world of both physical and emotional failing:

I know the cancer is eating away at me…I always felt I was in control. You know you are not going to be in control but subconsciously you don't believe it, then you get the physical evidence. You can see you are losing weight fast…Your clothes aren't fitting you anymore. You know it's taking over. It may be inside I'm giving up. It might be that inside I've had enough…

For Steven, weight loss symbolized the limitations of personal autonomy to control cancer.


Weight loss was something that was not discussed, either with family or healthcare professionals. Only two patients could recall conversations with their palliative care nurse about their weight. One had been advised that their weight might continue to decline and the other had been offered a financial grant to buy new clothes.

The nurse specialists spoke of weight loss as a common symptom that had the potential to cause much distress. Yet, they generally had a negative outlook on what could be done to help people live with the symptom. Claire described feeling helpless to do anything:

It's difficult to know what to give as a suggestion…There's not a lot I think you can do for weight loss.

Similarly, Jenny explained how she felt when unable to offer practical help:

Interviewer: How's that make you feel?
Jenny: Quite uncomfortable, because we are very much fixers. We like to try and fix everything and be able to offer something.

Caregivers also described the recognition of a loss of control over weight leading to a sense of helplessness, as described by Vera's husband:

But the trouble is weight always goes one way, it goes down. There's never a time when it goes up and it stays up…It's like trying to walk up an escalator the wrong way. You go faster and faster and never make no progress.

When patients were aware of the sense of helplessness engendered by their weight loss in others, then they could feel abandoned and become troubled by a feeling that nobody cared. This sense of abandonment is summed up by Paul:

Nothing can be done. These are the words they tell me.

Weight loss taboo

Because of the belief that little could be done to help patients with weight loss, nurses were careful not to initiate discussion about the symptom. They were concerned not to generate problems for patients and their families, as Angie (nurse) explained:

We generally tend to play weight loss down. If your disease is advanced you are not going to maintain your weight, how much of an issue should you make of that?

Caregivers expressed a similar concern not to problematise weight loss for the patient:

I didn't say anything, because I didn't want Mike to worry about it (Mike's wife)

Nurses suggested that it was best to manage weight loss using a ‘patient-led’ approach, meaning that it should only be addressed if raised as an issue by patients:

We don't want to make (weight loss) a bigger problem by prompting them to think, ‘Well, it is a problem if (my nurse is) asking me about it.’ (Anna)

However, Jean, a very experienced palliative care nurse was more confident of the benefit of initiating conversation about weight loss. She pointed out that if assessment of weight was routine from first contact, as with pain, it would not generate problems. It would be seen to be normal practice and something addressed with every patient.

If we treat nutrition, anorexia, cachexia in with everything else that we do and we treat it as a symptom if we can…Not to channel it, to compartmentalise it, if we could do that it would seem the normal thing to talk about.

Nurse specialists suggested that the appropriate weight loss intervention for patients and their families was to provide education and information through dialogue. Importance was placed on facilitating acceptance of the symptom for patients judged to be approaching the end of life. However, there was limited evidence from the patient interviews and records that this occurred in practice. Eleven of the patients had not had their weight assessed at any time from point of entry into the palliative care service. Of the 18 people who had been assessed, 15 had been assessed on one occasion only, two of the assessments had not involved any dialogue, and four assessments had taken place when the patient had sought help with concerns about their weight. One of the patients and two caregivers were troubled that information about weight assessments had not been shared with them:

Nobody's ever said anything about his weight. (The hospital staff) obviously know he's lost weight because they weigh him, but nobody has said anything. (Sid's daughter)

In only four cases could any evidence be found of the provision of information about weight loss and/or dialogue facilitating acceptance of the symptom. It seemed that, in practice, there was a taboo around entering into conversation about weight.

Model of living with weight loss

The themes that emerged from the interview data form the basis of a conceptual model of the experience of living with weight loss and advanced cancer (See Figure 1).

Figure 1.

The experience of living with weight loss and advanced cancer.

The experience of weight varies across time, for example with its stability/instability, yet the six themes cut across all experience. It is the relationship between time along the course of the advanced disease pathway and experience that is important to the likelihood of weight being of concern.

Weight typically declines across the cancer journey. However, whilst weight loss remains invisible and can be interpreted positively as a desirable improvement in health and/or continuity of life before cancer, then it is not used as a measure of illness or advancing disease. But, when weight loss becomes visible, that is the person appears physically wasted, the interpretation of weight changes for the patient and members of their social network. Weight loss comes to be seen as an undesirable physical change, symbolizing progressing disease and approaching death. Each of the themes of the experience of weight can then be sources of concern.

The model was developed from data collected about people believed to be typical of the total population of the case loads of two palliative homecare teams. Ethic minority groups were present in only small numbers across this population. Although the interview study sample included two people from ethnic minorities, the model may not capture the experience of all groups of people with cancer. However, the exploratory nature of the study means that the intention was not to generate generalizable findings, but context specific understanding that may have transferability because many features of the cases are present in other settings (Popay et al. 1998).


Weight loss can be a source of concern for patients when it becomes visible and leads to change in the lived experience of weight. Specialist nurses rarely assessed these concerns. The model of the experience of weight loss developed from the findings could provide the basis of a clinical approach to the identification of concern leading to the support of patients with weight loss related concern.

The importance of studying weight loss in advanced cancer is, in part, because of the large number of people affected (Poole & Froggatt 2002b). Patients who are losing weight have a poorer quality of life, as compared with weight stable patients (Ovesen et al. 1993, O'Gorman et al. 1998, Davidson et al. 2004). Our findings suggest that certain groups of weight losing patients are most likely to report poor quality of life. Those troubled by the symptom had visible weight loss that lead to self-perceived negative consequences. Perhaps this is because patients with greatest weight loss are also those with greatest functional impairment (Sarna et al. 1993). However, our findings show that weight loss not only has physical impact, but psychosocial consequences that may also compromise quality of life. The concerns discovered supports the proposition that weight loss can act as a symbol of advancing disease and can have a negative impact on patients’ sense of control (Poole & Froggatt 2002a,b).

For patients troubled by their weight, weight loss was a metaphor for change in a feeling of control. For healthy people, outward appearance has been suggested to demonstrate worthiness, control and self-discipline (Chamberlain 2004) and changing body states, such as pregnancy to have the potential to self-hate and feeling out of control (Lupton 1996, p142). The people with advanced cancer in this study were concerned about their weight loss because they experienced it as being out of control. However, concerns were more than manifestations of troubling change. It was clear from the data that for some they were acting as triggers to behaviour focused on regaining a sense of control over life. Concerns were perhaps a component of a process of appraisal and adaptation (Lazarus 1991) to a challenged sense of control over life.

To date, interventions to stabilize weight or lead to weight gain have been sought. Such interventions are assumed to be the way to improve quality of life in those living with weight loss (Ovesen et al. 1993, Davidson et al. 2004). Yet, this study has revealed sources of concern that arise in consequence of a psychosocial response to weight loss that may be amenable to change. Opening up the possibility of support for these concerns requires contravening the weight loss taboo.

Weight loss taboo

Overall, the response to visible weight loss of patients, their caregivers and their specialist nurses was one of helplessness. This sense of helplessness led to conversations about weight loss being taboo. There seemed to be no purpose in entering into discussion about a symptom that might provoke feelings of helplessness in others (but also one's self), and for which ‘nothing can be done’.

Nurse specialists experienced helplessness because weight loss undermined them in their role as managers of symptoms. Weight loss, which could only be temporarily arrested or reversed with medication (Tisdale 2002, Strasser 2003), revealed the limitations of medical science in controlling disease that led to an aversion to discussing the symptom. Other studies have similarly found that healthcare professionals are likely to block communication when they feel unable to offer practical help (Booth et al. 1996, Krishnasamy 1999) or when they perceive that they have skill deficits in managing patients’ concerns (Parle et al. 1999).

Patients experienced weight loss to be undermining because it spoke to a world of physical and emotional weakness. These weaknesses were considered to be undesirable as they symbolized a loss of control of the disease and losing the battle to avert death. They compromised the autonomy that patients could exercise through making choices that exerted control over the illness experience. Perhaps healthcare professionals have an awareness that weight loss can be experienced as a loss of control by patients which contributed to their reluctance to initiate discussion of the symptom.

Significance of the weight loss taboo

Observance of the weight loss taboo led to reactive management. As a result the assessment of weight loss and intervention was usually in response to a request from patient or carer. This approach may compromise both the psychosocial well being of patients and the potential for identifying factors which may contribute to weight loss that are amenable to interventions, such as practical difficulties in preparing meals.

Reactive management of weight loss reduced the chance that patients and their carers were forewarned of the potential for weight change, its causation and the availability of practical help for both real and anticipated physical and emotional weakness. Communication and a sense of connectedness are important to psychosocial well being (Lin & Bauer-Wu 2003). The opportunity to talk about experiences has been found to be supportive for those living with the symptoms of both fatigue and breathlessness (Corner et al. 1996, Krishnasamy 1999); this may also be the case for weight loss.

Giving information may be important in aiding emotional adaptation to illness, by presenting possibilities and consequences of actions that could otherwise be overlooked by patients and carers (Lazarus 1991). For example, knowledge of progressive disease being the primary cause of weight loss (Ravasco et al. 2003) may help with biographical appraisal and reconstitution of the rupture between body, self and society that is theorized to occur with chronic illness (Williams 2000). However, tailoring information to respond to individual concerns is likely to be important if adaptation is context specific (O'Driscoll & Cooper 1996, Hopkinson et al. 2005).


This study has important implications for clinical practice in the Western world, where the dominant paradigm in health care is a biomedical model that focuses primarily on physical function. Our findings prompt reflection on patients’ experience of weight loss and raise questions about the potential for alleviating psychosocial suffering by addressing concerns about the symptom. The model of the experience of weight loss provides a framework for assessing these concerns. Such assessment would create the opportunity to offer information proactively. If information can facilitate a sense of control in patients by aiding adaptation to illness, then distress as a consequence of the symptom seems less likely. Proactive management may empower people in an environment where the current practice of reactive management contributes to sustaining a weight loss taboo. In this way, systematic assessment and proactive intervention could both compliment and facilitate the use of pharmacological, nutritional and energy-balance approaches to the management of weight loss in people with advanced cancer. The proposal that weight loss-related concern can be mitigated by intervention now needs to be tested in clinical practice through research that examines patient outcomes when nurses break through the weight loss taboo.


The authors thank Macmillan Cancer Relief for funding this study and all the patients, carers and healthcare professionals who made it possible.

Author contributions

JBH and DNMW contributed to the study conception and design, data collection, data analysis and drafting of the manuscript; DNMW and JLC contributed to critical revisions of the manuscript for important intellectual content; JLC obtained funding and supervised the study.