The meaning of assisted feeding for people living with spinal cord injury: a phenomenological study


  • Bente Martinsen,

    1. Bente Martinsen MSN RN Doctoral Student Department of Nursing Science, Institute of Public Health, University of Aarhus, Aarhus, Denmark
    Search for more papers by this author
  • Ingegerd Harder,

    1. Ingegerd Harder PhD RN Associate Professor Department of Nursing Science, Institute of Public Health, University of Aarhus, Aarhus, Denmark
    Search for more papers by this author
  • Fin Biering-Sorensen

    1. Fin Biering-Sorensen MD PhD Head of Department Clinic for Spinal Cord Injuries, Copenhagen University Hospital, Copenhagen, Denmark
    Search for more papers by this author

B. Martinsen: e-mail:


Title. The meaning of assisted feeding for people living with spinal cord injury: a phenomenological study

Aim.  This paper is a report of a study to explore the meaning of assisted feeding through the experiences of people with high cervical spinal cord injury.

Background.  Eating difficulties are known to affect a person’s self-image and transform social lives. Little is known about the experience of assisted feeding as a permanent situation.

Method.  Sixteen people with high cervical spinal cord injury were interviewed twice within a period of 18 months in 2005–2006. The second interview was combined with observation. Transcriptions of interviews and notes from the observations were analysed using the phenomenological guidelines by Dahlberg and colleagues.

Findings.  The essence of the phenomenon assisted feeding was described as a constructed pattern based on coordinated attention between the person with high cervical spinal cord injury and the helper. The constituents of the essence were: paralysis as a condition of life, facing the altered meal, reconciling with diminished bodily anchoring, sensitive cooperation between self and helper, realization of own values around meals, balanced use of meal-related devices and negotiating relationship with helpers.

Conclusion.  Assisted feeding should be adjusted to each individual person. Fixed procedures or routines should be avoided and assistive devices used with care. We recommend that continuity in the cooperation between the parties involved in assisted feeding is given priority, and that personal standard and social norms around meals are acknowledged.

What is already known about this topic

  • • Eating is a complex process requiring sensitive coordination of a number of motor and sensory functions.
  • • Assisted feeding builds on a subtle cooperation between the person receiving help and the helper.
  • • Eating difficulties affect people’s self-esteem and their social lives.

What this paper adds

  • • It is a considerable life change when a person with high spinal cord injury needs assisted feeding permanently.
  • • The construction of a new eating pattern is not merely a re-construction but a careful creation of a new schema that has the former eating pattern and conventions among self-reliant people as a frame of reference.
  • • Assisted feeding leads to serious considerations about appearance and the presence of strangers during a meal may affect the construction of the eating pattern.


Assisted feeding is sometimes required by people in hospitals, nursing homes or their place of residence. People suffering from dementia are often assisted with feeding (Watson 1993), but also those with conditions such as stroke and multiple sclerosis may need assisted feeding (Bäckström et al. 1987). Some will need help for a short period, but others will have to rely permanently on help during meals. Assistance from others during meals may cause low self-confidence and feelings of stigmatization and role-conflict may occur between the helper and the person seeking help (Kayser-Jones & Schell 1997). However, little is known about the experience of needing help during meals on a permanent basis. This article presents the findings of a phenomenological study, exploring assisted feeding as experienced by people living with high cervical spinal cord injury (hcSCI).

The international incidence of SCI lies between 10·4 and 83 per million inhabitants per year (Wyndaele & Wyndaele 2006). Worldwide statistics concerning SCI vary and data are scarce (Wyndaele & Wyndaele 2006). In Denmark, with a population around 5½ million, the incidence of traumatic SCI is 10–15/million inhabitants and the majority are men (4:1) age 15–30 years (Biering-Soerensen 2001). About half will become tetraplegic as a result of cervical SCI, implying paralysis of upper and lower extremities and the trunk, as well as impairment of sensory, respiratory, urinary bladder, bowel and sexual function. Although the life expectancy of people with SCI still remains below that in the general population (Biering-Soerensen 2001), the small group of individuals who are paralyzed in the upper extremities because of hcSCI live with dependency for years. As part of their new life situation, they achieve considerable experience with being dependent on help from others during meals.


Eating is a complex process requiring sensitive coordination of a number of motor and sensory functions. Athlin et al. (1990) found that when a person has to rely on assisted feeding, meals require that patient and caregiver coordinate their behaviour. The caregiver should put the food into the patient’s mouth when it is opened, and the patient should respond when food is approaching. In order to achieve this subtle cooperation, the people involved must be able to perceive and interpret each other’s verbal and non-verbal behaviour. According to Athlin and Norberg (1987), the main responsibility for this cooperation lies with caregivers, whose experiences, educational background and personal beliefs may influence the course of the mealtime.

In a meta-analysis, caregivers were found to have ambiguous views about feeding completely dependent people, as some regard it as an unpleasant but prescribed task, while others see it as a situation where relations are in the foreground (Martinsen et al. 2007). Feeding by hand is time-consuming and labour-intensive work, and some residents in nursing homes have been found to be left to their own devices and without assistance, although their eating difficulties were well-known to staff (Kayser-Jones & Schell 1997). In addition, tube-feeding may be used to reduce nursing workload (Dibartolo 2006).

A study of coping with eating disabilities among stroke survivors showed that continuous confrontation with eating difficulties reminds patients of their incapacity and influences their self-image (Perry & McLaren 2003). Another study of stroke survivors showed that patients feel ashamed of their appearance and try to hide their lack of control when eating (Jacobsson et al. 2000). The inability of patients with head and neck cancer to eat in a normal way has been found to cause feelings of abnormality and not being able to recognize oneself (Larsson et al. 2003). Independence during meals is seen as ideal, and assisted feeding is postponed for as long as possible (Sidenvall & Ek 1993, Sidenvall et al. 1994, Jacobsson et al. 1996). When people who suffer stroke lose their meal-related autonomy it may threaten their hope for the future, whereas hope returns when meals become easier (Jacobsson et al. 2000). Physical problems such as changes of taste and decreased ability to chew and swallow reduce the joy of food, and people find that eating problems may devastate other aspects of life (Jacobsson et al. 1997, Larsson et al. 2003). Eating difficulties also have consequences for people’s social lives. Residents in nursing homes have been found to be reluctant to eat in communal dining rooms if they cannot reach their own standards of eating behaviour at the table (Sidenvall et al. 1994). Because of the changed appearance when eating, people may also isolate themselves at mealtimes (Jacobsson et al. 2000, Larsson et al. 2003).

The study


The aim of the study was to explore the meaning of assisted feeding through the experiences of people who are permanently and completely dependent on others.


Reflective lifeworld research, as developed by Dahlberg et al. (2001), was chosen as an approach capable of dealing with the human experience of a complex phenomenon. It focuses on describing the world exactly as it is experienced. Dahlberg et al. (2001) built on the philosophy of Husserl, Merleau-Ponty and Gadamer. Paramount in this phenomenological approach is an emphasis on a researching openness. The researcher turns to the experienced world with the aim of understanding the phenomenon on its own premises, and not taking its meaning for granted. According to Dahlberg et al., this open attitude is not easily accomplished because of our natural tendency to see events and objects in a specific way. We immediately get the full picture, even if we see only part of the thing or phenomenon. Reflective lifeworld research obliges researchers to hold back their preunderstandings, i.e. their assumptions, theories and other types of knowledge about the phenomenon are ‘bridled’, as Dahlberg et al. (2001) rename Husserl’s notion of ‘bracketing.’ Throughout the research process, we strove to set aside our everyday understanding of the phenomenon and be open to assisted feeding as it emerged in its immediacy. We critically considered how our prior knowledge might influence the research process and the outcome of the study. Making this shift from immediate involvement with the phenomenon into a reflective stance is crucial for the ability to really see the phenomenon and is aimed at achieving scientific objectivity (Dahlberg et al. 2001).


We recruited to the study people who had to rely on help from others for the rest of their lives and were therefore powerful informants of the lived experience of permanent and complete dependency on assisted feeding. They had no cognitive deficits.

To obtain a rich description of the phenomenon under study, participants were selected with variation in gender, age, marital status and length of experience with hcSCI. Criteria for inclusion were capacity to answer questions, reliance on help from others during meals for more than 3 months, and no use of tube-feeding. Two women and 14 men aged 18–65 years and living with hcSCI were included in the study. All participants were living in their own homes with 24-hour assistance from helpers. Seven were permanently phrenic pacer/ventilator dependent, while others had intermittent ventilatory assistance.

Data collection

Qualitative interviews were conducted twice by the first author (BM) with all 16 participants within a period of 18 months in 2005–2006. According to Van Manen (1990), lived body, lived space, lived time and lived relations pervade the lifeworld of human beings and may serve as keywords during interviews with a phenomenological approach. In the present study, after an opening question, ‘How is a typical day to you?’, participants were encouraged to elaborate on their experiences of assisted feeding using prompts such as ‘Please tell me more?’ or ‘Can you give another example of this?’

Before the second interview, several listenings and readings of the first interview tapes and transcripts were conducted, with the intention of continuing the conversations with participants. To deepen understanding of the phenomenon and maintain openness (Dahlberg et al. 2001), the second interview was combined with observation of a meal. The idea of observing meals arose during the preliminary analysis of the first interviews and permission was obtained from 14 participants. It was emphasized that the intention was to generate additional questions for the second interview, which was conducted immediately after the observation. The presence of the interviewer might have affected the way in which the meals were conducted. However, acquaintance from the first interviews is believed to have eased the situation. Notes were taken immediately afterwards and the observations provided a background for describing the lived experience of assisted feeding.

Ethical considerations

The local ethics committee confirmed that approval for the study was not required because of the non-biomedical character of the study, according to the law in Denmark. A physician who knew potential participants well confirmed that they were mentally fit to contribute to the study, and then they were contacted by the first author. Written informed consent was obtained and the ethical guidelines for nursing research in the Nordic countries were respected throughout the study (Northern Nurses Federation 2003).

Data analysis

The goal of a phenomenological analysis is to arrive at a structure where essences and their relationships can be described. In the search for essences, some meanings occur in all cases and help to clarify the phenomenon. Verbatim transcriptions of interviews and notes from the observations were analysed according to guidelines given by Dahlberg et al. (2001). The overall principle behind this phenomenological analysis is to understand each part of the text in terms of the whole and to understand the whole in terms of its parts. This constant movement between the whole and the parts is directed towards finding meaning, in this case the meaning of assisted feeding.

The analysis began with repeated and thorough readings of all the text to gain a global sense of each participant’s whole description. When it was possible to articulate an overall theme of a particular text as well as the demographic information contained in it, the character of the reading was changed to focusing on the meaning of the parts, i.e. the text was slowly reread and divided into meaning units. The various meanings were clustered into patterns of understanding that mirrored their interrelationships. Identification of the essence of the phenomenon of assisted feeding was achieved by synthesizing the clustered meaning units so that the phenomenon’s inner structure was illuminated. This search for structure was based on imaginative variations of the phenomenon in order to discover the fundamental meaning without which the phenomenon could not present itself as it did, e.g. parts of the phenomenon were freely changed to see if the phenomenon remained identifiable with the changed part or not (Dahlberg et al. 2001). When the essence was identified, its constituents were explicated and described.


The phenomenon assisted feeding as experienced by people with hcSCI is presented below as a general structure with seven constituents: 1) the paralyzed body as a condition of life, 2) facing the altered meal, 3) reconciling with diminished bodily anchoring, 4) sensitive cooperation between self and helper, 5) realization of own values around meals, 6) balanced use of meal-related devices, 7) negotiating relationship with helpers.

The structure of assisted feeding

The essence of the phenomenon of assisted feeding as experienced by people with hcSCI can be described as a constructed pattern based on coordinated attention between the parties involved. When a person suffers hcSCI and needs feeding for the rest of their life, the meal loses its bodily anchoring. The eating pattern used to be inextricably linked with the body, independent of conscious reflection and not possible to articulate. Now that the person is paralysed following an accident, the eating pattern is carefully thought through and explicated in order to be handled by other people. The person with hcSCI borrows the helper’s body to eat and drink and gradually teaches the other body to follow their eating pattern, which is constructed after the accident. Cooperation has to be based on mutual respect and involve a balance between closeness and distance. The person with hcSCI needs to sense a readiness on the part of the helper to obey even their smallest wishes and the affected person should never feel intimidated while being helped. Both helper and any assistive devices have to fit into the lifeworld of the person with hcSCI in order to make the assisted meal as similar to the original meal as possible. People with hcSCI must be able to recognize themselves in the helper’s actions and therefore the helper needs to consider their previous habits and preferences as well as the present situation. Assisted feeding involves a risk of being helped in a manner that does not correspond with the person’s self-image. Help that is provided in subtle ways allows the person with hcSCI to focus on other things than the actual feeding process. At the same time, however, personal wishes must be communicated to such an extent that the helper considers the affected person’s needs. When people with hcSCI are engaged in creating a new eating pattern, common conventions among people are still important.

The construction of a new eating pattern takes place on two levels. From the moment the accident happens, an ongoing process starts. The person with hcSCI gradually finds out on what principles meal-related help should be based. A long-term construction is created, such that the meal is perceived as acceptable. However, in connection with every meal a temporary construction is also created, closely related to the present situation and context. The long-term as well as the temporary construction are based on the person’s expressed wishes and only becomes familiar to both parties after a long period of cooperation.


Paralysis as a condition of life

The phenomenon assisted feeding has to be understood in the light of major physical reduction of functions. Because of hcSCI, our participants depended on other people’s help in almost all of life’s business. It was nothing special that they also needed help with meals, it was a condition of life. Without help, the person with hcSCI would not be able to live. If the participants had special requests about their help, they had to make them explicit: ‘I’m already in an explanatory situation.

Having a paralyzed body was a full-time job, and living with hcSCI was not something which added to life. Life did not just go on, now with a paralyzed body. Life changed when the body changed. Daily life implied that the person with hcSCI had to try to make clear their needs for help in such a way that the helper would understand and provide the right kind of care, having the person’s overall well-being in mind.

Facing the altered meal

The assisted meal was one of very few situations where the participants cooperated with their helper by being physically active. They could turn and bend their heads, open and close their mouths. The self-reliant body was the frame of reference, which participants used when planning their meals, even after having lived with hcSCI for a number of years, ‘because that’s the way I’ve always done it.

The paralyzed body required that participants engaged in numerous meal-related considerations. Their physical dependency changed their relationship to food and meals for ever. The meal developed from being a pleasure to becoming a technique following specific rules. It might be experienced as an intimate situation where the participant and helper sat closely together and communicated verbally and non-verbally. Assisted feeding required peace and concentration, and participants could not leave the situation before the meal was over: ‘It is a situation you cannot get away from. You are locked in it. You are not free of it until the meal is over. It is very intense in some way.

Reconciling with diminished bodily anchoring

Acquiring hcSCI was a major change in participants’ lives. Immediately and without any warning they had to accept that they needed help with many activities, help they had not received since childhood. After the initial difficult period, they gradually reconciled themselves with the physical dependency and accepted that they needed help. This was an ongoing process.

Just after the accident, participants preferred to get help from close relatives or helpers with whom they already felt comfortable. They preferred to eat in safe places. Being watched when getting help was a vulnerable situation where the person could feel they were regarded as ‘weak.’ Only gradually did they expand their comfort zone when eating. The thought of needing help when eating among unknown people could make them prefer to stay at home away from family reunions etc., especially in the beginning. Even after having lived with hcSCI for several years, some participants did not want to eat in certain places, ‘… that place (burger restaurant) was probably the last place I would suggest. I wouldn’t dream of it.

Sensitive cooperation between self and helper

It was a condition for a new eating pattern that sensitive cooperation between the person with hcSCI and the helper was created. Sensitive cooperation took time to build and was not possible when a helper was still new. It had to be based on the participant’s wishes and needs, and mutual attention was crucial.

At the beginning of the cooperation, the construction of the meal was characterized as a technical procedure. Participants had to explicate all that they wanted in order for the helper to learn the technical part of the construction. Teaching new helpers could be burdensome, and participants could not enjoy the food while constantly having to turn their attention to procedures around the meals: ‘It is pretty damn troublesome and damn annoying sometimes if it takes the helper too long to get the hang of it.

After a varying period of cooperation, assisted feeding could take place as a silent ‘dance’, where communication between the person with hcSCI and the helper became invisible. This was the ultimate goal: ‘I usually throw a glance, nod my head or do something like that instead of having to say each time, “I would like to have another bite”.

Realization of own values around meals

People with hcSCI have their first experiences with assisted feeding in hospital. Routines and norms from the hospital became the basis for the way participants initially constructed their meals, and for some of them this continued for several years after the injury. They strongly wished that their own values around meals could be realized, and they wanted to recognize themselves in the routines and rituals attached to the meal. They resisted being reduced to a ‘handicapped’ person or ‘patient’ who was being fed according to a standardized procedure. Certain foodstuffs were avoided because they were difficult to eat in a non-stigmatizing manner. Assisted feeding should take place as normally as possible and under no circumstance attract attention: ‘I must be able to recognize myself in the actions of the helpers.

It was difficult to identify or even imagine what was valuable to them in connection with the meal and participants had to ‘feel their way.’ After a while, they decided whether a particular behaviour from the helper fitted into the construction of their personal eating pattern.

Balanced use of meal-related devices

Assistive devices, e.g. straws, bibs, special cups, played a significant role when constructing a new eating pattern and made it easier to give and receive food and drink. Likewise, devices minimized the risk of placing the participants in situations where they felt humiliated because their helpers spilled food or drink on their clothes or smeared it on their face. Using assistive devices could also negatively influence participants’ self-respect and the respect they received from other people. Meals eaten at home were constructed in another way than meals eaten outside the home: ‘For practical reasons I use a straw for coffee when I’m home. I don’t use a straw when I’m out. Then I drink from a cup.

Participants’ self-respect depended on whether a given device was used or not. Things which were associated with people or situations not consistent with their self-respect were not used. Using an electric wheelchair in public was unusual in itself, and they did not wish to draw attention to their complete dependency on help from others. They felt exposed to other people’s gaze and in danger of becoming an object of contempt. Therefore, the construction of the meal was adjusted to deviate as little as possible from the way self-reliant people eat. On the other hand, the helper should not spill food on the clothes, which might emphasize the disability in a degrading way. Regard for what was practically possible was constantly weighed up against the risk of feeling ashamed: ‘It is not humiliating for me to receive help with my meals. But it is humiliating to look like a pig.

Negotiating relationship with helpers

Formally, an employer–employee relationship existed between the person with hcSCI and the helper. In reality, the relationship was described as a conditional cooperation where it was up to the person to define its character. The helper was described as an extension of the body: ‘the helper is my arms and legs.’ The context was decisive for the variety of ways in which cooperation was achieved. In some contexts, the helper’s position was comparable to the position of a family member, whereas in others the relationship with the same helper might change and become an employer–employee relationship. There was movement between closeness and distance, orchestrated by the participants. A simultaneous meal might be experienced as a way of maintaining and strengthening the relationship with the helper: ‘When we are sitting together as a family and eat, I think it’s nice when the helper is here.’ However, there were meals where the helper’s only task was to assist the person with hcSCI, creating a distance that stressed an employer-employee relationship: ‘They have to consider it a job when they walk through my door, and they bring their own food.

When participants had dinner with friends or relatives, they had to consider whether they preferred to have assistance from the helper or from a relative. The presence of the helper at the meal might threaten the confidentiality which previously existed between the person with hcSCI and family and friends: ‘So far, it is a great strain not to be able to sit and talk with your friends as you used to do.


Study limitations

Assisted feeding was not studied as part of an institutionalized context where meals might be controlled by certain structures, and where nurses play a role in the initial phases of this profound life change. The participants all lived at home and were assisted by helpers who were mostly untrained. The significance of assisted feeding in institutions, the role of nurses, and helpers’ training in general need further investigation. The study was conducted within a Western European country and the meaning of assisted feeding may be different in other cultural settings.

Discussion of findings

Assisted feeding is a complex phenomenon, and for our participants the construction of a new eating pattern was closely connected with their ability to recognize themselves in every single feature of this creative process. Habits and standards did not change fundamentally after the injury, and the acts of the helper, assistive devices and the community around meals must be recognizable to the person with hcSCI. This is in accordance with the findings of Perry and McLaren (2003), who showed that individuals with eating difficulties may have trouble recognizing themselves. The importance of recognition can be further understood in the light of Merleau-Ponty’s notion of the acquisition of habit as a rearrangement and renewal of the corporeal schema. According to Merleau-Ponty (2004/1945), my body appears to me as an attitude directed towards a certain existing or possible task. He describes two distinct layers in the body, the habit-body and the body of this moment. When a person becomes unable to use part of their body, the habit-body still exists as a background for the body of this moment. The manipulative movements that the person could execute before the illness occurred are still present in the habit-body, but the body of this moment is not able to perform them (Merleau-Ponty 2004/1945). For people with hcSCI, their former eating pattern exists as a frame of reference for the construction of a new pattern inherent in the body, although it is difficult to verbalize all aspects. It was necessary for them to try out a specific procedure in order to feel whether it was recognizable by the body and useful for the acquisition of a new habit. For nurses and other caregivers, it is important to note that people who need assisted feeding use the self-reliant meal as their standard of reference. Assisted feeding does not set aside meal-related conventions. On the contrary, it may make people who receive this kind of care particularly aware of prevailing norms.

The use of assistive devices was central to participants’ sense of appearance and they clearly wanted the meals to look as normal as possible. Using a variety of devices facilitates assisted feeding, but the fact that devices are used may affect the way people perceive themselves and the way they are perceived by others. The participants in our study, carefully assessed the connotations of assistive devices and decided whether they were consistent with their values. Over time, it may happen that a specific object becomes indispensable. As shown in previous studies (Jacobsson et al. 2000, Larsson et al. 2003), people with other eating difficulties are also concerned about their appearance when eating. According to Merleau-Ponty (2004/1945), an assistive device can be incorporated into the corporeal schema as part of a new habit, and by repeated use the device ceases to be an object and becomes an extension of the body. In our study, some assistive devices, such as straws, particularly fitted into the lifeworlds of the participants and gradually developed into extensions of their body. When they were incorporated into their corporeal schema, participants stopped questioning the object and used it as part of their body, without considering if it affected their appearance or not. The findings show that assisted feeding must be tailor-made for every individual. Although particular devices may seem to ease the process of assisted feeding from a nursing perspective, their value can only be assessed by the person who receives the assistance. Assistive devices may potentially be regarded as insulting to people who need help.

Mutual attention between the person with hcSCI and the helper was crucial in the construction of the new eating pattern. Meals that allowed them to melt into each other were experienced as the most desirable. Words were not needed, techniques not in focus, and the construction seemed to unify them in a common lifeworld. This unification implied a certain level of acquaintance between the person with hcSCI and the helper, and a readiness on the part of the helper to make a shift in consciousness. This is in accordance with a phenomenological analysis of attention described by Schuback (2006), who argues that attention presupposes that the caregiver is able to leave former conventions and focus on other ways of understanding. Techniques and habits of attention are not useful, but rather function as an alertness which enables the caregiver to experience themselves as ‘being-with’ and stay open to the events of the lifeworld (Schuback 2006). This suggests that the significance of continuity in the cooperation between the parties involved in assisted feeding should not be underestimated. Assisted feeding is an important area of care where nurses and other professionals have the opportunity to become engaged and to combine knowledge with authentic human involvement.

The meal is one of all the taken-for-granted activities of everyday life that may be altered when a person is injured. Over time, our participants became reconciled with the diminished bodily anchoring of their eating pattern, and the overall feeling of abnormality decreased. However, getting help in public places could evoke this feeling, although the eating pattern was gradually constructed in a way that would reduce the possibility of attracting attention. Toombs (1993) states that illness and disability imply an altered state of existence, a distinct change of being-in-the-world, and that ill or disabled people often feel as if they are seen by others as an object. The loss of upright posture also causes others to assign certain features to people in wheelchairs (Toombs 1993). In his analysis of shame, Sartre (1943) argues that shame is not a feeling I elicit on my own – it presupposes the presence of another person, not merely because the other person is the one before whom I feel ashamed, but also because the other human being constitutes that of which I am ashamed. In our study, participants feared being mistaken for ill or weak people when their meals were exposed to the gaze of others. This latent feeling of abnormality must be taken into consideration when people with eating problems are gathered in institutional dining rooms (Pierson 1999). Eating with others should occur voluntarily and not just for the convenience of staff.


Assisted feeding should be adjusted to each individual person. Fixed procedures or routines should be avoided and assistive devices used with care. We recommend that continuity in the cooperation between the parties involved in assisted feeding is given priority and that personal standard and social norms around meals are acknowledged. Future researchers should focus on the experiences of larger groups of people living with complete dependency and explore nurses’ direct involvement with and professional responsibility towards people in need of assisted feeding.

Author contributions

BM and IH were responsible for the study conception and design. BM performed the data collection. BM and IH performed the data analysis. BM was responsible for the drafting of the manuscript. BM, IH and FBS made critical revisions to the paper for important intellectual content. BM obtained funding. IH and FBS provided administrative, technical or material support. IH and FBS supervised the study.