Prevalence, Response and Management of Self-harm in School Children Under 13 Years of Age: A Qualitative Study

Research suggests that the incidents of self-harm among young people are increasing and age of onset of self-harm is decreasing. There is limited understanding of how widespread the problem of self-harm among younger school students is, and how schools respond to these incidents. This study used an in-depth qualitative approach to understand self-harm in children under 13, and how elementary schools respond including typical actions, support for the child and parental involvement. School psychologists in New South Wales, Australia (n = 17; 78% females), completed in-depth interviews detailing types of self-harm, prevalence and frequency, and how these incidents were managed, including student intervention approaches and participation of parents. Thematic analysis of interviews was conducted. Psychologists estimated the prevalence of self-harm in their schools was 6.5% and was increasing, with an average age of onset of 10.8 years. Self-harm was most often understood as a coping mechanism associated with anxiety, stress and being bullied. Six themes emerged from the interviews. School psychologists reported that self-harm occurs less frequently in primary school children than high school children, but noted these behaviours still require early intervention. Participants felt they were limited in the support they can provide students who self-harm, and wanted more training for all school staff and parents in identifying and responding to student self-harm. School psychologists are important in identification and management of self-harm, and they report they need further support in helping children who are engaging in self-harm behaviours. Upskilling teachers and parents may assist in reducing rates of self-harm among children.

Self-harm is increasing in younger populations and has been identified in children as young as six years old (Australian Institute of Health & Welfare, 2014;de Kloet et al., 2011;Mitchell et al., 2018;Singareddy et al., 2013). An American study found that 7.6% of third graders reported having self-harmed (Barrocas et al., 2012). For those who reported self-harm, 60% reported they had hit themselves, and 13.3% reported cutting or carving their skin. Further, a study in Ireland reported that between 2007 and 2016, the age of onset for self-harm in young people decreased, while rates of self-harm increased (Griffin et al., 2018).
Methods of self-harm in children may typically be less severe than methods of adolescents and adults; however, they can be an indication of stress and poor emotional regulation, which can often develop into more severe mental health disorders (Brunner et al., 2007;Chapman et al., 2006;De Leo & Heller, 2004). In adolescents, self-harm has been linked to underlying mood and emerging personality disorders (Chapman et al., 2006). It is also associated with low self-esteem (De Leo & Heller, 2004), greater impulsivity (Hawton, Bergen, et al., 2012;Hawton, Saunders, et al., 2012), experiences of sexual or physical abuse (Palmer &
A recent Australian study found in a population-based sample that 3% of 11-12 years reported self-harm within the last year. It was found that mental ill-health and difficult peer relationships were commonly associated with self-harm behaviours (Borschmann et al., 2020). However, despite the reported increase in presentations of self-harm for children and young people, the true prevalence of such behaviours remains uncertain-particularly for children under the age of 13. This may be attributed to several factors. Firstly, the stigmatised nature of self-harm may prevent disclosure of these behaviours to trusted adults (Rickwood et al., 2005). Secondly, although there is some emerging literature on self-harm in children, the research in this area is still sparse (Bem et al., 2017;Berger et al., 2013a, b), and current research relies on small sample sizes or self-report measures (Borschmann et al., 2020). Thirdly, if children present to health services such as emergency department for self-harm, it is often not documented as such and therefore rarely reported (Australian Institute of Health & Welfare, 2014). Finally, for studies investigating individuals with known self-harm, there have been differences in methods of data collection (e.g. hospital records, self-reports, parent reports), and this has produced a high amount of variability in estimates (Clements et al., 2016). As such, current estimates may be under-estimating the prevalence of self-harm in children.
The available evidence suggests that self-harm in children is increasing, and as the link between self-harm and adverse outcomes is well established, ensuring that children are appropriately supported for their mental health is an important public health issue to address. Given that children spend a considerable proportion of their time at school, school counsellors, psychologists, and welfare staff are often the first to be involved in cases of children and young people who self-harm (Morgan et al., 2017). One study of parents found that initial confirmation that their child was engaging in self-harm behaviour often came from school staff (Oldershaw et al., 2008). Teachers have found signs of selfharm in student written work and subtle changes in student behaviour, such as a refusal to change clothing in physical education (Dowling & Doyle, 2017).
Previous studies have focused on the perceptions of high school teachers, school psychologists, and other staff members when working with students who self-harm (Berger et al., 2013a, b;Dowling & Doyle, 2017). Berger and colleagues (2013b) administered an online questionnaire to 500 high school staff members, of which 106 were mental health staff, regarding their perceptions and knowledge of self-harm. The study found that the majority of staff members reported a lack of knowledge and confidence in dealing with self-harm in students and feelings of frustration and anger regarding these behaviours. Further, 73.6% of participants identified the need for additional training in the area and only 2.1% of participants reported they provided students with appropriate mental health resources. This study provides some insight into the need for resources and training in self-harm for school staff. It is possible that within this sample different school staff (e.g. teachers and school psychologists) showed different patterns of responding to self-harm; however, this was not discussed in the study.
Understanding primary school-aged children (5-12 years) self-harming behaviours and offering alternative solutions may serve as a preventative measure for further mental health difficulties (Best, 2006;Simm et al., 2008). In order to do this, school staff need to be able to identify and respond appropriately to self-harm behaviours. It is therefore important to understand whether primary school staff-including school psychologists-feel equipped and have the adequate training and resources to respond to self-harm behaviours among children. As most primary schools employ a psychologist or counsellor to respond to issues of student well-being, it is important to understand the specific perceptions of school psychologists, which may differ from other staff and contribute unique psychological understandings to this issue.
The present study aims to understand how school psychologists in primary schools perceive and respond to incidents of self-harm within this setting. A realist epistemological approach (Willig, 2013) was used as the overarching framework to understand school psychologists' experiences.
We aim to answer two research questions-how do school psychologists perceive self-harm in primary school-age students (5-2 years), including prevalence, onset age and rates over time, and how do the schools in which they are located, respond to children who self-harm. A qualitative approach was selected to provide exploratory insights into the nature of self-harm in this age group and to help develop hypotheses for future quantitative research, as such no a priori hypotheses were made. Qualitative research methods can provide a deeper understanding of human experiences and increase the breadth of data received from participants. When investigating school psychologists' perceptions, an interview may allow for a broader exploration of current issues and their possible solutions.

Participants
Participant recruitment was undertaken until qualitative data saturation was reached. School psychologists were identified from a pool of 182 public primary schools in New South Wales (NSW). While many school psychologists work across several schools, all have a base school where their school principal had to agree to their involvement in our study and to identify and pass onto the researchers details of their school psychologist as a potential participant. The psychologist then provided written informed consent to be interviewed for this research. The final sample size of the study was 17 psychologists.
Participants had a mean age of 39.39 years (SD = 14.51, range = 24-68), and 14 (77.80%) participants were females. The majority of participants were from the metropolitan areas (n = 11), followed by regional (n = 5) and rural areas (n = 2). On average, there were 745.71 (SD = 238.32) students at each school and participants had an average of 10.36 years (SD = 11.57, range = 1-38) of experience as a school psychologist.

Measures
Demographic information was collected, and participants were asked to provide information on years of experience as a school psychologist, number of children attending the base school they are employed at, and how many children they had been in contact with in the past 12 months who had self-harmed. To provide context about the phenomena under investigation, participants were asked to estimate the percentage of primary school students who self-harm, the typical age of onset for self-harm and whether they perceive self-harm in this age group is increasing. Perception of Self-harm in Children-Participants were also asked to respond to a series of developed statements regarding their perception of self-harm in children, with possible response ranges from 1 = strongly disagree to 5 = strongly agree. Example statements include 'selfharm is a suicide attempt' and 'bullying can lead to selfharm in children'. The purpose of these questions was to understand the possible factors and perceptions related to self-harm in this age group that school psychologists may hold but not discuss in the qualitative interview.
Following this, participants were asked semi-structured interview questions which included providing their own definition of self-harm and what they believe the most common methods of self-harm is for primary school-aged children. They were also asked their perspective on the amount of focus in the primary school environment that is required for students engaging in self-harm behaviours, and how the primary school they are based in responds to student self-harm.

Procedure
Participants provided written informed consent following approval of the local Institutional Review Board and Education Department. Their school principal was also required to consent to their participation. All participants completed the semi-structured interview and survey with one researcher. In-depth interviews were audio recorded and then transcribed verbatim.

Statistical Analysis
A deductive analysis of qualitative responses was conducted using NVivo Version 11 (QSR International, 2015). Thematic analysis based on Braun and Clarke's approach (2019) was used to analyse data, allowing interpretation of complex data and accounting for individual experience. Members of the research team immersed themselves in the data by reading and re-reading transcripts. Significant statements within the data were prescribed initial codes and then further grouped into themes. Interpretation continued until themes were refined and organised into a coherent account of the data. Data saturation was reached when 15 interviews were conducted, and a further two interviews were conducted to confirm the redundancy. Data were then searched for disconfirming information to ensure that the themes accurately fit participant responses. Quotes were selected on the basis that they illustrated the particular aspects of the theme. Some subjectivity in coding the data was assumed due to the biases of the researcher. However, reliability was ensured via crosschecks with co-authors.

Results
Participants had worked with an average of 7.47 (SD = 12.09, range = 0-50) children who had self-harmed in the last 12 months. The mean prevalence rate for self-harm in primary school children was estimated as 6.5% (SD = 8.73) by participants; however, there was high variability within the data (range = 0.5-33%). Almost half (44.40%) of participants strongly agreed that self-harming behaviour in primary school children was increasing and estimated the average age of onset of self-harming behaviours as 10.83 years (SD = 1.98).
For the developed questions regarding perception of self-harm among students, participants on average agreed that self-harm is a coping mechanism that can arise from bullying, anxiety and stress, and is a continuum of behaviours from mild to severe. Participants on average disagreed that self-harm is a suicide attempt, is attention-seeking and represents bad behaviour. Participants overall felt neutral about identifying self-harm in children and the relationship between self-harm and other forms of escalated behaviour. Table 1 presents the mean responses to these statements.
Thematic analysis of open-ended questions identified six main themes in terms of school psychologists' perception of and response to self-harm: (1) identifying and managing self-harm in primary school-aged students, (2) self-harm occurs less frequently in younger students but still requires active intervention, (3) school psychologists have limited capacity to support primary school students who self-harm, (4) school staff require further professional development in working with children who self-harm, (5) parents require psychological education for self-harm behaviours in children and (6) children engaging in self-harm face barriers in receiving support outside of the school setting.

Theme 1: Identifying and Managing Self-Harm in Primary School-Aged Children
There were differing views of how self-harm presents in young children. Most participants identified intentional physical harm to the body, including 'pulling out their hair. Or banging their heads.' [Participant 7], 'cutting. But not to the point in like high school where they find razors. It would be things like using scissors… and also not allowing for wounds that they already have to heal.' [Participant 10] and 'head banging. Or hitting themselves…, or scratching themselves, but not necessarily with an implement.' [Participant 14]. Participants also presented more covert behaviours that had a self-harm and destructive intent, such as reckless running across the road, or restricting food intake.
Self-harm behaviours among primary school-aged children were described by psychologists as often having an emotional basis. For example, when discussing a primary school student who banged his head on the table, one participant reported 'well given that he's 11, I'd, unless he's come and requested to speak to me individually, either the principal or myself would ring the parent and just say he sort of expressed feeling emotional…' [Participant 4]. Participants most commonly reported they would attempt to build rapport and engage with students following an episode of self-harm and would confidentially discuss with students to understand the function of the behaviour while ' [coming] up with a safety plan for that person and, um, and alternate means to distract themselves' [Participant 3]. School psychologists also discussed the importance of assessing medical safety with the assistance of other school staff or emergency services, and the need for a calm response from all staff members in order to prevent any escalation of behaviour.

Theme 2: Self-Harm Occurs Less Frequently in Younger Students But Still Requires Active Intervention
Approximately half of participants indicated the need for a greater focus on self-harm within primary schools. They referenced the possible long-term effects and escalation of self-harming behaviours in adolescence. For example, one participant stated, 'if self-harm isn't, there's no early intervention for it, it can escalate to the point where it becomes severe or, you know, it might develop into suicidal ideation.' [Participant 10]. These participants broadly discussed self-harm as an expression of emotional distress by a child. One participant remarked 'I think it's an indication of a child's-well it can be an indication of a child's level of distress that they can't otherwise express. And before it becomes a habit, so sort of intervening early in primary school would be really helpful.' [Participant 11]. However, other participants identified that while self-harm behaviours are concerning, they were less frequent in primary schools-'broadly speaking it's not an issue that gets referred to me a lot…I think it's not often seen at school, in primary school, in the same way it might be in high school ' [Participant 7]. Participants also reported that self-harm was more prevalent in older children (aged 14 and over) in the high school environment. A school psychologist who worked across both primary-and School psychologists reported that if they did see a student for support they would focus on rapport building, harm minimisation, implementation of alternative skilful behaviour and safety planning including the family and external services. Several participants noted that 'beyond immediate crisis management' [Participant 8], working with self-harm in primary school-aged students was 'really outside our bounds of, um, expertise and an external counsellor is generally warranted at least to deal with whatever issues are fuelling the self-harm.' [Participant 8]. Given the limited school psychologist availability in primary schools and complexity of presentations, most participants (82%, n = 14) identified the importance of making external referrals to general practitioners or child psychologists, 'I would not probably position myself as her primary, um, psychological support, I would probably want to link her into external services. '

Theme 4: School Staff Require Further Professional Development in Working with Children Who Self-Harm
School psychologists conveyed the importance of other school staff-including teachers, principals and school office staff-to understand and appropriately react to incidences of self-harm in children. A number of participants expressed the need for teachers to approach incidents in a calm and non-judgemental manner, 'because I think the staff's reac-tion…would play a key role in helping her feel contained and supported, rather than an overreaction.' [Participant 11]. Participants identified barriers for primary school staff to respond appropriately to self-harm, including a lack of education and knowledge about distinguishing self-harm from defiant behaviours requiring disciplinary action. Some participants also reported that school staff may ignore or refuse to assist in incidences of self-harm due to the lack of confidence and willingness in responding to self-harm adequately. This was reflected by a participant who stated 'some staff just go, no, it's not my role, I'm not going to have anything to do with it even though they've got a good relationship with the child and can probably do some positive stuff with them.' [Participant 6]. Participants identified that staff need to be provided with psychological education and training to support identifying and responding to self-harm as 'I think staff probably are a little bit more overwhelmed by things like self-harm and they automatically think self-harm is suicide' [Participant 6]. It was suggested training could focus on an 'understanding what constitutes self-harm, how teachers should respond to it.' [Participant 7]. Participants reported that if staff were appropriately trained in self-harm management, they may be able to detect early signs of selfharm and implement coping strategies for students when psychologists are not available at school.

Theme 5: Parents Require Psychological Education for Self-Harm Behaviours in Children
Participants identified that a lack of knowledge about child self-harm behaviours among parents was the biggest barrier to students receiving adequate support. Participants emphasised the need for psychological education for parents and the implementation of information seminars in schools so that students can be supported by health professionals and also in the home environment. One participant suggested 'coaching the parents in how to respond. So, giving the parents psychological education and helping the parents understand that there's a difference between self-harm and suicidal intent, but also normalising to the parents their own level of distress, and really supporting the parents. Supporting the parents to remain calm, but also have a safety plan at home. ' [Participant 11]. 1 3

Theme 6: Children Engaging in Self-Harm Face Barriers in Receiving Support Outside of the School Setting
Following the recognition that many primary school psychologists had limited time to provide comprehensive care and the complexity of the issues that some students present with, many students who engaged in self-harm were also referred to external services. However, a barrier to students receiving appropriate support identified by participants was that external care may not be sought by caregivers; 'there's a fine line between the recommendation and the referral and the actual intervention taking place so the school can't necessarily be the place, you know, they can't take the child to the GP, they can't take the child to CAMHS [Child and Adolescent Mental Health Service], they can't take the child to Headspace [National Youth Mental Health Foundation]' [Participant 6]. Participants also reported being wary of disclosing child self-harm to parents when there are child protection or welfare issues, including 'if we told the parents she's self-harming because of the domestic violence that's happening at home, then she would get into-like they would physically bash her for opening her mouth so that was putting her in more risk.' [Participant 9]. Therefore, participants indicated that some children may not be able to be supported by their parents to access external services they need, leaving school psychologists as their only source of psychological support.

Discussion
Interviews with school psychologists were undertaken to understand the prevalence of self-harm, onset age and rates over time in primary school-aged children, and associated management challenges. Participants estimated the average age of onset of self-harming behaviours in primary schoolaged student as 10.83 years, and 44% indicated self-harm in primary school-aged children was increasing. Qualitative responses were analysed to explore how school psychologists and schools respond to children who self-harm including typical actions, support for the child and parental involvement. Six major themes emerged from the data in relation to the research questions.
Firstly, there were varying perspectives of how self-harm presents in primary school-aged children across the school psychologists interviewed. While most participants focused on behaviours directly intended to cause harm including scratching and head banging, several participants also discussed the importance of identifying and managing more covert behaviours with a self-harm or destructive intent. This may reflect the differences between identifying and responding to self-harm issues in primary school-aged children compared with older children and adolescents, where it may be more common to conceptualise self-harm as behaviours causing immediate visible physical harm (Simm et al., 2008). This also suggests the need for a comprehensive definition of self-harm suitable for professionals working with younger age groups to ensure they get the support they need.
Secondly, school psychologists identified that self-harm occurs less frequently in primary school children compared with high school children. This may be reflective of the lower rates of self-harm in primary aged children compared to students in high school (Barrocas et al., 2012) or indicative of difficulties with accurate identification of these behaviours (Rickwood et al., 2005). Participants estimated the prevalence of self-harm among primary school students as 6.5%, and half of the sample indicated a need for greater focus on self-harm in primary schools. The results suggest that selfharm in this age group is used as a coping mechanism to address difficulties with bullying, anxiety and stress. These findings are consistent with several of the triggers identified by adolescents who engage in self-harm (Hetrick et al., 2020). Furthermore, child maltreatment in particular is associated with an increased risk of self-harm (Lang & Sharma-Patel, 2011;Serafini et al., 2017). Participants identified that self-harm may be indicative of mental health difficulties and is a predictor of suicidal behaviours, which reflects quantitative research (O'Connor et al., 2012). As such, they reported that identifying self-harm early may prevent progression to mental health disorders during adolescence.
Thirdly, school psychologists identified that they feel limited in the support they can provide to primary school students who self-harm. Their professional capacity to respond to incidents of self-harm was limited when they worked in a part-time capacity at schools, with an average of 745 students at each school for participants in this sample. School psychologists reported their interventions with primary school-aged students who self-harm would include rapport building, safety planning, implementation of alternative behaviours to self-harm, and liaison with other staff, family and external service providers. Participants also reported they were likely to make external referrals for children who self-harm, due to both school psychologist's availability and limited capacity, and the need for specialist referrals for complex and severe presentations. However, research has indicated that there is a lack of access to appropriate mental health services for many children and young people-particularly in low socio-economic status areas (McGorry et al., 2013). Accessing private mental health care is beyond the financial capacity of many families and over 40% of people with mental health conditions in Australia forgo treatment because of the high costs involved (Callander et al., 2017). Also, there is generally poor uptake of these services among children and adolescents when they are available, possibly as families who are under stress find it difficult to engage in and commit to accessing help (Merikangas et al., 2011). In Australia, one study found that only 25% of children and adolescents meeting a clinical threshold for a mental health disorder sought help (Sawyer et al., 2001). However, counselling at school was the second most frequently accessed service for mental health among the sample, after family doctors (Sawyer et al., 2001). This suggests that children and young people may be more inclined to seek help for mental health disorders from school psychologists who they may already have consistent contact with in the familiar environment of school. This finding is important in highlighting the necessary and unique role that school psychologists have for students and supports the need for an increase in psychologists within primary school contexts, to adequately support students.
These findings also highlight the need for further support and training for school psychologists, especially when working with children facing more complex mental health issues and childhood adversities. Assessing and responding to student self-harm must take place in a way that is trauma informed (Schmaal & Bendall, 2018). School psychologists should assess for childhood maltreatment, stressful life events, mental health issues as well as exploring the functions of self-harm for the child and potential suicidal ideation (Townsend, Gray, et al., 2018;Townsend, Hasleton, et al., 2018). There are a range of evidence-based practices available to school psychologists including dialectical behaviour therapy (DBT-A), cognitive-behavioural therapy (CBT) and mentalisation-based therapy (MBT) (Asarnow & Mehlum, 2019;Ougrin et al., 2015).
Fourthly, school psychologists identified the importance of the response by other school staff to incidents of selfharm. Participants reported school staff members may find it difficult to differentiate self-harm from defiant behaviour in children or may feel uncomfortable with self-harm and ignore incidents. This is consistent with literature which found that teachers may have misconceptions about selfharm, and feel overwhelmed, frustrated, angry and anxious when working with students who engage in these behaviours (Berger et al., 2015) and some self-harm can be attributed as bad behaviour by schools staff (Simm et al., 2010). In order to appropriately and compassionately support children who self-harm, further trauma informed psychological education and training regarding identifying and responding to self-harm may be necessary for all staff, so teachers can provide students with adequate support, especially when school psychologists are not available (Simm et al., 2010;Townsend, Gray, et al., 2018;Townsend, Hasleton, et al., 2018). Guidelines have recently been published that provide guidance to schools to support effective responses to students who engage in self-harm (Matthews et al., 2021). Previous research has also suggested that pre-career training for teachers in responding to self-harm would be beneficial (Best, 2006).
Fifthly, there emerged a need to provide parents with psychological education about self-harm in children. Participants reported that parents did not understand the function and nature of self-harming behaviours and were hesitant in seeking out professional help. This reflects previous literature, indicating parents feel a significant amount of distress, hopelessness, confusion and guilt when their children self-harm (Raphael et al., 2006). As such, education targeted towards parents regarding the function of self-harm and how to identify and appropriately respond to self-harm may be beneficial. Lastly, participants also identified barriers to children receiving support for self-harm behaviours and expressed their concern about parents who did not follow up the external referrals suggested for their children, as ultimately it was the parent's decision to access professional support for their child. Previous studies have found parents may delay seeking help for their children as they view self-harm to be a phase or typical adolescent behaviour (Oldershaw et al., 2008). The findings of the present study suggest that parents may have limited understanding about the nature of self-harm and the mental health issues which can underlie these behaviours. School-based programs and initiatives which provide education about child mental health and the functions of self-harm may be beneficial for parents.
The findings of the present study continue to highlight a need for further support for schools, school psychologists, and parents, from earlier in childhood. Training and employing more school psychologists is important, alongside the training of teachers and other school staff to work with children who self-harm. Providing teachers with mental health skills during training years may better equip them for the workplace. Psychologists also identified the need for additional support for parents and caregivers. Schools may consider providing information sessions targeted for parents in order to raise awareness of self-harm in children and how to respond appropriately. Such programs can also help to reduce the stigma around self-harming behaviours and provide parents with available support systems, perhaps reducing the associated negative feelings of distress, anxiety and frustration which parents may experience (Raphael et al., 2006).

Limitations and Future Directions
There were a number of limitations in this study. The use of a semi-structured interview may be subject to recall bias, in which there is a systematic difference between participants' responses and their actual experiences (Sedgwick, 2012). Participant responses may have also been impacted by a social desirability bias, in which participants alter their answers so that they are socially more acceptable or 1 3 favourable (Fisher, 1993). There was a wide range in definitions of self-harm-most participants indicated the physical nature of self-harm, but some discussed covert behaviours such as restricting eating. While this allowed a broad understanding of self-harm, the differing definitions may have impacted participant responses to other questions in the semi-structured interview. Estimated prevalence rates in this study may be consistent with participants' work context, but further multi-informant studies are needed to determine prevalence of self-harm in primary school-aged children across a larger sample. Research assessing the rate of selfharm among primary school students would be beneficial along with longitudinal studies examining child characteristics and outcomes over time, in particular the relationship of self-harm and childhood maltreatment. Future studies may also benefit from including both school psychologists and other school staff members within the sample to explore types of support available in different environments such as the classroom, and to assess what support and training would most benefit school staff members. Future research with children and parents of children is also required to understand their perspectives and needs, as well as the reasons that children engage in these behaviours and the extent they differ from adolescents.

Conclusions
Self-harm is an important target for psychological intervention in children attending school under the age of 13. Findings indicate that self-harm occurs less frequently in primary school-aged children compared with high school-aged children; however, there remains a need for an increased focus on self-harm in primary schools. School psychologists are limited in the support they can provide due to the complexity of these difficulties. Many school psychologists identified the need for multidisciplinary care and often suggested to parents further referral options for support in the community. Further training providing education, understanding and practical strategies is needed for both school staff and parents to more effectively respond to children who selfharm. This study highlights the need to continue to provide psychological education and training in responsiveness to self-harm in order to best support children in need.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.