Introduction

The need to protect vulnerable people from COVID-19, and not overwhelm the health service, has led to major social change in the UK over a very short period of time. Since 23rd March 2020, UK residents have been required to stay at home and observe social distancing. In cases where they have symptoms of COVID-19, or are particularly vulnerable to its effects, they are required to fully self-isolate. At the same time, UK schools were closed to all but keyworkersā€™ children and vulnerable groups. These vulnerable groups include some children with Special Educational Needs and Disabilities (SENDs), particularly those with an Education and Health Care Plan (EHCP). However, due to the risks involved, many families with a child eligible to attend school have not chosen to take up a place.

There are good reasons to believe that this pandemic, and our response to it, may affect children with SENDs and their families disproportionately, and that this may have negative implications for their mental health. Some mental health problems are known to be associated with SENDs in normal circumstances, such as anxiety in Autism Spectrum Conditions (ASCs: van Steensel and Heeman 2017). It seems likely that the scale and speed of the social change that has occurred since 23rd March could exacerbate existing mental health problems and trigger new ones.

Families who are raising children with SENDs face more stressors, on average, than those with neurotypical children (McConnell and Savage 2015; McStay et al. 2014). Such stressors can influence the quality of family relationships, making patient and empathic parenting challenging at times (Osborne et al. 2008). Teachers also experience stress in working with children with SENDs, leading to a pattern of some burning out and leaving special education (Brunsting et al. 2014), with some research suggesting that perceived support may be a protective factor (Langher et al. 2017).

Staying at home, and in most cases not attending school, creates a uniquely stressful situation for children with SENDs and their families. Carefully developed routines have been disrupted; support networks have disintegrated; and parents have been asked to do a job that trained teachers find challenging, without any training. These changes have happened abruptly and the consequences could be particularly profound in the SEND community. It is therefore important to ask how COVID-19 is affecting the mental health of these families, with a view to gaining insight into how schools and society can support them over the coming months.

Methods

Ethics

Ethical approval for this study was granted by the Department of Education Ethics Committee, University of York (Reference: 20/05).

Participants

Participants were 241 parents or carers of school-aged children with SENDs in the United Kingdom (M ageā€‰=ā€‰9Ā years; range 5ā€“18). Of the children, 71% were boys; 88% were White British (remainder: 6% Mixed, 3% Asian, 2% White non-British, 1% Other); 44% were in mainstream schools and 70% had an EHCP. School places were made available to 73% of the sample but only 8% had taken up their place. The range of SENDs that were represented can be seen in Table 1.

Table 1 Type of special educational needs and disabilities as reported by the parent
Full size table

Of the 241 parents/carers who provided data on their own and their childā€™s mental health, 92% were mothers, 95% were from England (remainder from Scotland and Wales) and 63% had a pre-tax household income of less than Ā£40,000 (approximate UK median income). Participants were recruited between 22nd March and 1st April 2020 during the first fortnight after school closures, via existing research networks, special schools and online platforms.

Measures

As part of a wider project on the impact of COVID-19 on families with children with SEND, participants were asked to answer the following free response question: Please describe in your own words how the coronavirus outbreak is affecting your mental health and your childā€™s mental health.

Coding and Analysis

An exploratory, inductive content analysis was undertaken (Bengtsson 2016). As an initial step all five authors independently coded data from the same 15 participants and discussed this ā€˜open codingā€™ process in detail. On the basis of these discussions, the authors independently coded data from a further 15 participants and, following detailed discussions, agreed codes. This process was used to develop a codebook made up of 48 codes (see Table 2). Some of these codes were specific to parents e.g. Anxiety (parent) and Anxiety (child), recognising that at some points participants were describing their own experiences and at others reporting on their child. We divided the remaining data (from 211 participants) between three researchers who each used the codebook to code one-third of the data. A second rater independently coded 20% of the data, selecting every fifth participant in all three batches of data (nā€‰=ā€‰43). We tested for intercoder reliability using Fleissā€™ kappa and the results of this analysis are shown in Table 2.

Table 2 Intercoder reliability for first cycle coding
Full size table

We found moderate to strong intercoder agreement for most of the 48 codes applied to the data. Fleissā€™ kappa measures the level of agreement over and above chance with 0 suggesting that agreement is no better than chance and 1 indicating perfect agreement. Altman (1999) suggests that Īŗā€‰<ā€‰0.20 is poor, 0.21ā€“0.40 is fair, 0.41ā€“0.60 is moderate, 0.61ā€“0.80 is good and 0.81ā€“1.00 is very good. Table 2 shows that agreement on 41 of the 48 codes was either very good (26) or good (15). Of the remaining seven codes agreement on three was moderate/fair and agreement was poor for four. This was largely explained by these codes being applied in few cases (<ā€‰=ā€‰3), exaggerating the effects of disagreement. The second rater reviewed the full dataset to ensure that the seven codes with agreement that was neither good nor very good were applied consistently and discussed areas of disagreement with the original coders, achieving group consensus through this team discussion process (Harry et al. 2005). Final tweaks were made to the codebook and the agreed codes were clustered into categories. Our analysis was manifest rather than latent, attempting to stay close to the participantsā€™ own words rather than identifying underlying meanings (Braun and Clarke 2013).

Results and Discussion

The frequencies for the agreed codes are shown in Table 3. It is clear that a large proportion of families report that COVID-19 has affected their mental health, often leading to an increase in anxiety and fear. Smaller numbers also reported increases in distress, low mood and stress. The data suggests that more parents than children have experienced increased anxiety (44% vs 25%) and stress (12% vs 5%), although it is important to note that this is parent-reported data, but frequencies for parent-reported fear, distress and low mood were similar for parents and children.

Table 3 Frequencies for final agreed codes
Full size table

Our final 35 codes were organised into six categories that represent the impact of COVID-19 on mental health as reported by these SEND families: Worry; Loss; Mood, Emotions & Behaviour; Knowing what is going on; Overwhelmed and Minimal or Positive Impact. Table 4 shows that codes were organised into categoriesā€”groups of codes that clustered together. For example, the category of worry is made up of three sub-categories (worry for self, worry for others and general worry) which, in turn are made of clusters of codes e.g. worry for self represents four codes, namely, parent and child anxiety for self and parent and child fear for self. Clustering codes into categories is the process via which the researchers identified patterns in the data (Bengtsson 2016).

Table 4 Categories, sub-categories, codes and examples
Full size table

Worry was sub-divided into three sub-categories: worry for self, worry for others and general worry. Although some of the worry described is likely to be shared by parents and children in general, the majority was specific to families with a child with SENDs. The examples given in Table 4 illustrate the point that participants often described extreme anxiety reactions that are not likely to be commonplace in the general population, and that are characterised by known features of SENDs e.g. restricted food preferences. Also, parental worry for others, particularly concern for their childā€™s future, focused on their children falling even further behind in school because they did not know how to meet their needs, and worries about who would look after them if they (the parent) died as a result of COVID-19. The level of worry many SEND families report appears to be substantial and serious, and supporting these families in ways that will help to alleviate or reduce their anxiety should be a priority for education, health and social care professionals.

Loss was also described by many participants as a result of COVID-19, and these losses were organised into four sub-categories: loss of routine, loss of support network and structures, loss of specialist input and, for a minority, financial loss. Some of these losses are likely to be widespread in the community, such as loss of access to support networks, changes in routine and financial losses. However, the data suggests that, in some cases, the effects of these widespread losses are amplified in families with a child with SENDs because the challenge of meeting the childā€™s needs is simply greater. Also, for some children with SENDs it is not possible to explain why these losses have occurred, creating further difficulties. We have several examples in the data of lone parents who are isolated with a child who displays very challenging behaviour without access to any of the support and respite that usually helps them to fulfil their parental role effectively. These lone parent SEND families appear particularly vulnerable in the current situation. Furthermore, there is a suggestion that some parents feel that during the first fortnight of school closures their children were insufficiently supported. In some cases this was because no tailored support was offered (e.g. children in mainstream classrooms being provided with the same input as the rest of the class). Parents also mentioned a pressing need for some children to see familiar faces, such as those of their teacher or teaching assistant, a need that while shared by some neurotypical children is likely to be amplified for children with SENDs such as ASCs.

Moods, emotions and behaviour, including low mood, acting out and behaviour change, were mentioned by many participants. While low mood and distress are likely to be widespread the data suggests they may be experienced more severely within the SEND community. The types of challenging behaviour described in Table 4 are clearly difficult for both parents and children to cope with, involving police, violence to self and others and a level of destructiveness that is likely to be uncommon among children of equivalent ages without SENDs. The distress this causes suggests that support is needed before crisis point is reached in some families.

Knowing what is going on was a key element of some participantsā€™ responses. Parents described situations in which a childā€™s low level of understanding led to distress because they could not understand why everything had changed. In the cases of minimally verbal children their disorientation was sometimes expressed in challenging behaviour. The data suggests that better understanding was associated with better outcomes. One way to address this issue may be to prioritise children known to have limited understanding and to create resources that may help to enhance their understanding, such as social stories, while working to ensure that those aspects of life that can stay the same (or similar) do stay the same (or similar) e.g. providing remote face to face support. Child awareness of the situation was described as having both positive and negative effects in that while some children seem oblivious to COVID-19 and have not minded the changes, others are frustrated for the reasons described above.

Overwhelmed was the penultimate category used to describe participantsā€™ responses. A substantial minority of parents described themselves as being stressed by the new demands that were placed on them, including meeting all of their childā€™s special or additional needs without support, a break or respite, often alongside working and meeting the needs of others in the family. Given that this data was collected within two weeks of school closures we might expect this situation to worsen over time although it also remains possible that, as families become more acclimatised to the situation, they may become less overwhelmed or there may be no discernible change.

Minimal or positive impact was described by a substantial minority of families for whom the impact of COVID-19 was not perceived as harmful. It was notable that positive emotions were often expressed in families where the child has a hard time at school and feels safest at home. For these children, self-isolation and social distancing may lead to a period of calm respite, creating a more relaxed environment for them and their families.

Conclusion

There is evidence in the literature that children with SENDs and their families are likely to be at greater risk of experiencing poor mental health, and to be under substantially greater pressure, than less vulnerable families during COVID-19. The data presented here provides some support for this likelihood, although not definitive support as the study did not include a comparison group of neurotypical children and their families. It is therefore important to assess the specific needs of these families and take concrete steps to meet them. These families were also asked what support they would like during COVID-19 and their responses are reported elsewhere (Toseeb et al. 2020). Frequently mentioned suggestions included specialist professional advice for parents focused on how to meet their childā€™s educational and mental health needs; setting appropriate tasks and resources for home learning; and providing opportunities to see familiar faces, albeit remotely. The data presented here, from the same group of families, indicates that support along these lines would be beneficial to many families with a child who has SENDs. Also, it is important to note that some families will not need or want additional support but that identifying those who are struggling, and providing tailored support, should represent a priority for education, health and social care during the coming weeks.