Autistic characteristics and mental health symptoms in autistic youth during the first COVID-19 wave in Canada

Autistic youth are at heightened risk for mental health issues, and pandemic-related stressors may exacerbate this risk. This study (1) described caregiver-reported youth mental health prior to and during the pandemic; and (2) explored individual, caregiver, and environmental factors associated with changes in autistic characteristics, social – emotional symptoms, and overall mental health. 582 caregivers of autistic children (2 – 18 years old) completed an online survey between June and July 2020 in which they provided demographic information, their child ’ s pre-COVID and current mental health, autistic characteristics, and social – emotional symptoms. Caregivers also rated their own perceived stress, and COVID-related household and service disruption. According to caregivers, youth experienced more autistic characteristics and social – emotional concerns during the pandemic. Autistic youth were also reported to experience poorer overall mental health during the pandemic than before the pandemic. Older youth whose caregiver ’ s indicated higher perceived stress and greater

• Less household financial hardship prior to COVID-19, caregiver stress, and service disruption as well as absence of a pre-existing psychiatric condition in youth were associated with better youth pandemic mental health. • Autistic youth who were older and whose caregivers experienced more stress and greater household disruption were reported to experience more autistic traits.

INTRODUCTION
The coronavirus  pandemic has had an unprecedented impact on life around the world. Emergency measures such as stay-at-home directives and closures of schools, clinics, and community programs left many autistic and non-autistic children facing new stressors that threatened their wellbeing (Brooks et al., 2020;Hiraoka & Tomoda, 2020;Spinelli et al., 2020;White et al., 2021). Pandemic-related work interruptions also created financial issues and limited access to essential resources in many households (Kansiime et al., 2021;Martin et al., 2020;Messacar & Morissette, 2020). Autistic youth and their caregivers already experience higher rates of co-occurring mental health conditions and elevated levels of stress prior to the pandemic compared to non-autistic children and caregivers (Hayes & Watson, 2013;Lai et al., 2019). Higher risk for mental health concerns, combined with COVID-19 disruptions, put autistic children at increased risk for negative outcomes (White et al., 2021). Few studies, however, have examined the influence of the pandemic on autistic youth. The pandemic has been associated with increased prevalence of anxiety, depression, and stress-related symptoms in youth without autism (Huang & Zhao, 2020;Liang et al., 2020;Marazziti et al., 2020;Pierce et al., 2020;Racine et al., 2020Racine et al., , 2021. Self-and caregiver-reported prevalence of anxiety and depression in neurotypical youth from various countries during COVID ranges from 19% to 44%, respectively (Liang et al., 2020;Rosen et al., 2020;Xie et al., 2020;Zhou et al., 2020). A meta-analysis including 80,879 neurotypical youth globally showed pooled prevalence estimates of clinically elevated depression and anxiety, 25% and 21%, respectively in youth (Racine et al., 2021). These authors indicated depression and anxiety symptoms during COVID-19 have doubled from pre-pandemic estimates, and rates were higher when collected later in the pandemic, in older adolescents and girls (Racine et al., 2021).
The mental health of autistic youth has also been negatively impacted during the pandemic. In a large Canadian study of youth with and without pre-existing psychiatric and neurodevelopmental conditions, up to 70% of children 6-18 years of age experienced poorer mental health during the pandemic compared to prior to the pandemic (Cost et al., 2021) according to both caregiver and self-report. Autistic youth were identified as the group whose mental health had been negatively impacted the most during the pandemic, with several autistic youth indicating significant deteriorations in depression (55%), irritability (65%), attention (55%), and hyperactivity (55%). Similarly, 50% of autistic youth with pre-existing mental health conditions reported that their anxiety was worse during the pandemic relative to pre-pandemic levels (Cost et al., 2021). At the same time, this study noted that a minority of autistic youth were reported to have no change in depression (30%), irritability (30%), attention (40%), and hyperactivity (40%), and some autistic youth even showed improvements in these mental health domains (5%-15%; Cost et al., 2021). Similarly, White et al. (2021) surveyed 3,502 caregivers of individuals with autism regarding the impact of COVID-related disruptions to therapies on autistic traits and family stress. Almost two-thirds (64%) of caregivers reported disruptions to services had severely impacted their youth and family, with autistic youth showing more prominent autistic traits and their families experiencing moderate distress during the pandemic (White et al., 2021). Huang et al. (2021) also found that 50% of autistic children were reported to have experienced sleep problems, and 36% to have worsened emotional and social wellbeing during the pandemic by their caregivers. Over half (61%) of caregiver respondents reported that their child experienced a decrease in access to therapy services, and consequently, caregivers reported increased pressure to support their autistic child who were experiencing higher levels of distractedness and meltdowns (Huang et al., 2021). Consistent with these findings, other rapid-response COVID-19 studies have highlighted caregiver-report increases autism-related traits (Asbury et al., 2021), and elevated social-emotional distress in children and families with disabilities including autism (Arim et al., 2020;Statistics Canada, 2020a, 2020b. Similar to neurotypical people (Hwang et al., 2020;Loades et al., 2020;Pietrabissa & Simpson, 2020;Rauschenberg et al., 2021), autistic youth and adults, as well as caregivers of autistic children have reported greater social isolation during the pandemic, with dissatisfaction for substitutions, such as online/telephone-based alternatives, for person-to-person connection (Pellicano et al., 2021;White et al., 2021).
In non-autistic children, researchers have frequently used the developmental-ecological framework to better understand what factors contribute to mental health issues (Anderson & Mohr, 2003;Kemp et al., 2016). The developmental-ecological framework proposes that child social-emotional wellbeing and development is a complex system of relationships affected by multiple levels of influence (Bronfenbrenner & Morris, 1998;Garbarino, 2014). From this view, child wellbeing is the result of factors that are proximal to the child (e.g., gender, age, and diagnosis) and broader contextual factors (e.g., characteristics of their caregiver, household, and community) that accumulate to transactionally influence outcomes (Garbarino, 2014). In relation to autism, levels of influence might include the immediate responses of caregivers, school actions and policies, as well as broader cultural, economic, and political practices (Danforth, 2013). Theorists increasingly recognize that consideration of an autistic individual's characteristics and life experiences in isolation is not sufficient for contextualizing the individual in an inherently social world (Danforth, 2013;Loveland, 2001;Predescu et al., 2018). Further, in the context of a global pandemic, external factors existing outside of the child at the family (e.g., fear of illness, job loss, and caregiver stress) and community (e.g., changing public health guidelines, school, and clinic closures) levels surely influence child outcomes. Thus, a holistic ecological approach is beneficial in understanding the mental health in autistic youth during the pandemic as well as to structure an effective support network for autistic children and their families. Consistent with this framework, we studied influences at the child (e.g., autism symptom severity, pre-existing mental health conditions), caregiver (e.g., perceived stress), and environmental (e.g., household and professional service disruption) level. To our knowledge, no study has examined how child, caregiver, and environmental factors may accumulate to understand the mental health of autistic children during the pandemic.
Recently, researchers have started to investigate various factors that put neurotypical youth at risk for experiencing poorer pandemic mental health. These factors have been examined independently of each other and studies have shown mixed results. While older children tend to have higher symptoms of mental health issues (Jiao et al., 2020), no association between child age and mental health have been found . Child gender has also shown mixed results; with some studies suggesting that identifying as female is a risk factor for depression and anxiety in non-autistic children during the pandemic (Oosterhoff et al., 2020;Racine et al., 2021;Zhou et al., 2020), while others have failed to find an association (Xie et al., 2020). Financial strain and higher caregiver distress are associated with higher anxiety and depressive symptoms in non-autistic youth during the pandemic (Rosen et al., 2020); however, the additive impact of these variables with other child-specific variables is unclear. Despite a growing body of literature indicating an association between mental health symptoms in non-autistic and autistic youth and the pandemic, potential risk factors for these outcomes have mainly been studied in isolation without analysis of the additive impact of these factors.
The current study is the first to address this gap in the literature by examining the combination of potential contributing factors to autistic characteristics, socialemotional symptoms, and overall mental health in autistic youth during the pandemic using the developmentalecological framework. We (1) described caregiverreported autistic characteristics, social-emotional symptoms (i.e., anxiety and mood), and the overall mental health during the first wave of the pandemic in a large cohort of caregivers of autistic individuals; and (2) using a developmental-ecological framework, explored factors that might contribute to changes in symptomology and overall mental health. We included autism symptom severity and the presence of a pre-existing co-occurring psychiatric diagnosis at the child level, as these variables are previously associated with autistic traits, socialemotional functioning, and overall mental health in autistic youth (Havdahl & Bishop, 2019;Lai et al., 2019;Matson et al., 2008;Moseley et al., 2011;Simonoff et al., 2008). As caregivers of autistic youth have reported increased stress during the pandemic (Friesen et al., 2021;Lee et al., 2021), caregiver stress was considered as a potential contributing factor to child outcomes. Lastly, prior to the pandemic autistic people received a high number of special education and professional therapy services in schools, clinics, and community programs (Koegel et al., 2012;Volden et al., 2015;Weiss et al., 2018), which were abruptly disrupted due to COVID-19 (Arim et al., 2020;White et al., 2021). As such, service and household disruption were considered as possible environmental factors contributing to youth pandemic outcomes in this study.
Overall, it was expected that according to caregiver report youth would show more autistic characteristics and social-emotional symptoms (i.e., anxiety and mood), and poorer overall mental health during the pandemic relative to pre-pandemic levels. Further, these characteristics and symptoms were anticipated to be associated with caregiver-reported characteristics of the child, the caregiver, and the broader ecological context (e.g., household and service disruption).

METHOD Participants
Five hundred and eighty-two Canadian caregivers between 24 and 70 years of age (M = 36.9 years, SD = 6.4; 56.4% women, n = 328) completed an online survey in June-July 2020 regarding their child's coping during the COVID pandemic. Majority of caregivers were married or common law (89.9%, n = 523) and lived in suburban or urban communities (86.6%, n = 505). Only one caregiver per family completed the survey and they had at least one autistic child or youth living in their household. Most respondents (74.2%, n = 432) described themselves as either managing or managing very well financially prior to COVID-19, with the remaining experiencing some to deep financial trouble. Caregivers completed the survey from across 10 Canadian provinces with the majority reporting from Ontario (34.4%, n = 200), British Columbia (26.9%, n = 157), and Alberta (20.4%, n = 119). No significant differences in responses based on region were found, all p's > 0.05.

Procedure
Caregiver participants were recruited through various social media platforms (e.g., Instagram and Facebook) and autism-related agencies across Canada. Inclusion criteria were: (a) able to read in English; and (b) providing care/parenting (i.e., biological, adoptive, or foster parent and/or legal guardian) to at least one person with a autism diagnosis from a health professional (including autism spectrum disorder, autism/ autistic disorder, Asperger's syndrome, pervasive developmental disorder, pervasive developmental disorder not otherwise specified). Data collection began on June 9, 2020 and ended on July 30, 2020. After completing the self-administered survey, caregivers received a gift card of a small amount to reimburse them for their time. This study obtained ethics from the relevant university ethics review board. To align the present study with the interests of the autistic community, caregivers of autistic people were consulted regarding study design (e.g., participant recruitment, data collection, and survey and question development) and interpretation of the study findings.

Demographic information
Caregivers provided basic demographic information about themselves (e.g., own age, gender, and marital status, as well as their autistic youth such as the youth's age, and gender [male, female]). Household financial status was also obtained by asking caregivers whether they experienced financial hardship prior to the pandemic (yes, no). From the Centers for Disease Control and Prevention's National Centre for Health Statistics (NCHS) Survey of Pathways to Diagnosis and Services, caregivers rated the severity of their youth's autism symptoms as mild, moderate, or severe. In previous studies that have used this item (e.g., Zuckerman et al., 2018), caregiverreported autism symptom severity has been found to be associated with increased service use in autistic children. Additionally, whether the youth experienced a preexisting psychiatric condition prior to the pandemic (yes, no), as well as the autism-related and mental health therapeutic services they received prior to the pandemic were provided by caregivers.

Autistic characteristics and social-emotional symptoms
Caregivers responded to Likert-type questions regarding their child's autistic characteristics and social-emotional symptoms, and how these symptoms may be different from prior to the pandemic. Caregivers reported on four items related to autistic characteristics: (1) restricted and repetitive behaviors (my child is experiencing more restricted and repetitive behaviors than usual [e.g., lining up objects, repeating words, phrases, or body movements, insisting on routine, etc.]); (2) social communication (my child is experiencing more challenges with social communication than usual [e.g. avoiding eye contact, avoiding social situations]); (3) sensory issues (my child is experiencing more sensory issues than usual); and (4) interest in specific topics (my child is experiencing greater interest in specific topics/objects than usual [e.g. special interest]). Social-emotional items inquired about: (1) behavior problems (my child is experiencing more behavior problems than usual [e.g., tantrums, aggression, self-harm]; (2) anxiety (my child is experiencing more anxiety than usual); and (3) low mood (my child is experiencing more difficulties with depression or low mood than usual).
Caregivers rated items according to a scale from 0 = strongly disagree to 4 = strongly agree. Likert responses of 3 = agree and 4 = strongly agree were categorized as worsened, since these responses indicate a higher degree of the trait characteristic or symptom than prior to the pandemic. Likert responses of 0 = neither agree nor disagree, 1 = disagree, and 2 = strongly disagree were categorized to not worsened, since these responses do not indicate more of the characteristic or symptom present than prior to the pandemic. Total autistic characteristics and social-emotional symptom scores were created by summing the items (autistic traits total score range = 0-16; social-emotional symptom total score range = 0-12). Higher total scores were indicative of more autistic characteristics present and socialemotional symptoms than prior to the pandemic. The alpha reliability coefficients for these scales were acceptable (Cronbach's alpha = 0.65; 0.67). Results of Spearman correlations indicated significant positive associations between the four autism-related trait items with weak to strong associations (r(582) = 0.24-0.41, all p's = <0.001) as well as the three social-emotional items with moderate to strong associations (r(582) = 0.37-0.42, all p's = <0.001).
Overall youth mental health A global assessment of overall youth mental health was rated by caregivers on a sliding scale where 0 = Very poor, 50 = Fair, 100 = Excellent. Respondents rated their youth's overall mental health on the sliding scale at the time of survey participation (i.e., current pandemic mental health) and retrospectively 2 months prior to the COVID-19 pandemic (i.e., prior overall mental health). Additionally, caregivers completed a similar sliding scale for specific symptoms of behavioral challenges, anxiety, and low mood.

Perceived stress scale (PSS-10)
This 10-item scale is a self-report questionnaire used to measure the degree to which caregivers experienced stress (Cohen et al., 1983;Cohen & Williamson, 1988). Caregivers rated how often they have experienced each statement in the past month on a Likert-type scale from 0 = never to 4 = often. Total scores were calculated by summing each item and ranged from 0 to 40, with higher scores indicating higher perceived stress. Total scores between 0 and 13 are considered low perceived stress, 14-26 are considered moderate perceived stress, and scores 27 and above indicate high perceived stress (Cohen & Kessler, 1997). Acceptable psychometric properties for this scale are reported in a systematic review (Lee, 2012): internal consistency (Cronbach's alpha = >0.70); testretest reliability (>0.70); factorial validity; and hypothesis validity (moderately or strongly correlated with emotional variables such as anxiety and depression). The alpha reliability coefficient for this scale in the present study was questionable (Cronbach's alpha = 0.62); however, authors support interpretation of composite reliability values of 0.60-0.70 as acceptable in exploratory research similar to the current study (Nunnally & Bernstein, 1994).

Pandemic-related household and service disruption
Caregivers responded to Likert-type scales indicating the degree to which aspects of their family life had been disrupted due to the pandemic. Seven items inquired about: (1) overall family life disruption (to what extent has your family's life been disrupted by the COVID-19 pandemic); (2) financial hardship (has your family experienced financial hardship); (3) difficulty getting food (has your family had difficulty getting food); (4) shortage of essential goods (has your family experienced a shortage of essential goods [e.g., toilet paper, cleaning supplies]); (5) difficulty accessing childcare (has your family had difficulties finding childcare); (6) caregiver fear for safety (have you feared for your own or your family's health or safety); and (7) child fear for safety (has your child feared for their own or their family's health or safety). Respondents rated items according to a scale from 0 = not at all to 4 = extremely. A total household disruption score ranging from 0 to 28 was created by summing items. A higher total score was indicative of higher household disruption. The alpha reliability coefficient for this scale was acceptable (Cronbach's alpha = 0.73).
Caregivers also indicated how the therapeutic services their child was receiving prior to the pandemic were impacted. They indicated the nature of change to each service they were receiving as follows: 0 = nothing changed, 1 = canceled altogether, 2 = became temporarily unavailable, and 3 = moved to remote service delivery (e.g., online, telehealth). The maximum number of services any one respondent receiving was 10, and therefore, 10 items were summed to create a total service disruption score ranging from 0 to 30. A higher total score represented higher service disruption.

Data analysis
All analyses were carried out using SPSS version 26.0 with a significance level p ≤ 0.01 given the large sample size and number of pathways included in the models. Descriptive statistics (i.e., frequencies, means, and standard deviations) were calculated to summarize demographic information about caregivers and their youth. Total sample mean scores of perceived caregiver stress as well as household and service disruption were summarized to describe developmental-ecological characteristics during the pandemic. A paired samples t-test was used to examine differences in overall caregiver-reported child mental health prior to and during the COVID-19 pandemic.
A series of hierarchical linear regressions were completed for each of the primary outcomes of interest (i.e., autistic characteristics total score; social-emotional symptoms total score; caregiver-reported overall youth mental health). For each regression, the following variables were included in each step: (1) demographics (i.e., child age and gender, household financial status prior to COVID-19); (2) child variables (i.e., pre-existing psychiatric condition, severity of autism symptoms); (3) caregiver perceived stress; and (4) environmental variables (i.e., household and service disruption due to . Variables were entered in order of child, caregiver, and environment according to the developmental-ecological perspective, which posits influences on the youth to move from their own personal characteristics outward to increasingly external influences (Garbarino, 2014). Importantly, all influences are theorized to play a crucial role, with the additive influence of child, caregiver, and broader environmental characteristics central to youth outcomes.

RESULTS
Overall, approximately half of the caregivers in this study reported that their youth experienced more autismrelated characteristics (ranging from 51.7% for social communication challenges to 54.3% in reference to restricted and repetitive behaviors and sensory issues; Table 1). Likewise, 49.7%-50.5% of caregivers reported that their youth experienced more social-emotional symptoms, such as low mood and anxiety, respectively, since the onset of the pandemic (Table 1).
Caregiver-reported youth overall mental health prior to the pandemic (M = 64.58, SD = 18.27) was significantly higher than ratings at the time of survey completion (M = 62.59, SD = 17.94), with worsening overall mental health, though the effect size was small (t = À2.59, p = 0.010, Cohen's d = 0.11). In both cases, mean ratings prior to and during the first wave of the pandemic remained close to fair on the sliding scale.
Regarding disruption to professional therapeutic services, caregivers reported changes to at least one or several of their services. Sixty-two percent (n = 345) of caregivers indicated that their youth's services were canceled altogether, and just under half of caregivers reported that services became temporarily unavailable (48.6%, n = 270), or moved to remote service delivery (e.g., online, telehealth; 47.4%, n = 263).

Factors associated with greater overall autisticrelated characteristics
The final set of independent variables accounted for approximately 23% of the variation in greater autistic characteristics during the pandemic. Models progressively explained a significant amount of variance (i.e., R 2 = 2%-23%, p's = 0.057-0.008; Table 2). Older youth age (β = 0.13, p < 0.001) was significantly associated with more autistic characteristics during the pandemic at the final stage, as was higher caregiver perceived stress (β = 0.10, p = 0.008), and greater household disruption due to COVID-19 (β = 0.41, p < 0.001). Youth gender, experiencing financial difficulty prior to COVID-19, the presence of a youth pre-existing psychiatric condition, autism symptom severity, and greater service T A B L E 1 Frequencies of caregiver-reported autistic characteristics and social-emotional symptoms of the sample.  Note: Worsened represents Likert responses of 3 = agree and 4 = strongly agree that the child is experiencing more of the trait or symptom than prior to the pandemic. Not worsened represents Likert responses of 0 = neither agree nor disagree, 1 = disagree, and 2 = strongly disagree that the child is experiencing more of the trait or symptom than prior to the pandemic.
T A B L E 2 Results of the hierarchical linear regression predicting greater overall autistic characteristics of the sample. disruption due to COVID-19 were not significantly associated with greater autistic characteristics during the pandemic (Table 2).

Factors associated with greater overall socialemotional symptoms
Regarding greater social-emotional symptoms (i.e., behavior difficulties, anxiety, and low mood) during the pandemic, models progressively explained a significant amount of variance (i.e., R 2 = 4%-20%, all p's = <0.001; Table 3) and the final model accounted for almost 20% of the total variation. Older youth age (β = 0.12, p = 0.003) and presence of a youth pre-existing psychiatric condition (β = 0.11, p = 0.003) were significant demographic predictors of greater social-emotional symptoms during the pandemic. Ecological factors including higher caregiver perceived stress (β = 0.25, p = <0.001), greater household disruption (β = 0.21, p = <0.001), and greater service disruption due to COVID-19 (β = 0.10, p = 0.008) were also significant predictors at the final stage of the model. Youth gender, experiencing financial difficulty prior to COVID-19, and autism symptom severity did not emerge as predictors of greater social-emotional symptoms (Table 3).

Factors associated with child overall pandemic mental health
As shown in Table 4, together the final set of independent variables accounted for approximately 16% of the variation in better overall youth mental health (i.e., sliding scale global assessment of overall mental health). Models progressively explained a significant amount of variance (i.e., R 2 = 3%-16%, p's = 0.008-<0.001; Table 4). Experiencing less household financial difficulty prior to COVID-19 was a significant demographic factor related to better overall youth mental health during the pandemic (β = À0.12, p = 0.002). Having no pre-existing youth psychiatric condition (β = À0.19, p < 0.001) was also a significant predictor of better youth pandemic mental health at the final stage of the model. Finally, reduced caregiver stress (β = À0.26, p < 0.001) and less service disruption due to COVID-19 (β = À0.11, p = 0.005) were significant predictors. Youth age and gender, autism symptom severity, and household disruption due to COVID-19 were not predictors of better overall youth mental health during the pandemic (Table 4).

DISCUSSION
This large cross-sectional study offers insight into caregiver-reported autistic characteristics, socialemotional symptoms (i.e., behavior challenges, anxiety, and low mood), and overall mental health of Canadian autistic youth during the first COVID-19 wave (i.e., June/July 2020). Approximately half of caregivers reported their youth experienced increased autistic characteristics and social-emotional concerns at the beginning of the pandemic. Autistic characteristics such as restricted and repetitive behaviors (54%), social communication challenges (52%), sensory issues (54%), and interest in special topics (53%) were reported by caregivers to generally increased for half of autistic youth. Severity of behavioral difficulties (50%), anxiety (51%), and low mood (50%) also increased for half of our study sample. Problematically, this increase in symptomology came at a time when major disruptions to professional support services occurred due to COVID-19 safety measures. Many caregivers in this study reported their therapeutic services were canceled (62%), became temporarily unavailable (49%), or moved to virtual or telehealth service delivery (47%). Disruption to services and therapies at a time when many autistic youth experienced increased symptoms, some of which may be extreme (e.g., aggression, self-harm), is stressful for families. Consequently, a large majority of caregivers reported themselves to be moderately stressed (86%) and approximately half reported their overall family life was disrupted (52%).
Poorer overall youth mental health was rated by caregivers during the pandemic compared to retrospectively prior to the pandemic, though the difference was of a small magnitude. These findings are consistent with other rapid-response pandemic studies showing poorer overall mental health and elevated social-emotional symptoms (e.g., anxiety, depression) in neurotypical youth (Cost et al., 2021;Racine et al., 2020Racine et al., , 2021Rosen et al., 2020;Xie et al., 2020;Zhou et al., 2020) as well as autistic children (Arim et al., 2020;Asbury et al., 2021;Cost et al., 2021;Huang et al., 2021;White et al., 2021). Consistent with other researchers (Cost et al., 2021;White et al., 2021), we found that an abrupt termination of familiar school-based and other therapeutic services was reported by caregivers, along with a high degree of household disruption that may include loss of structure and consistency in daily routines. These disruptions may present unique challenges for autistic youth possibly contributing to poorer pandemic mental health. Our study findings provide further evidence for increased socialemotional concerns and poorer mental health in autistic youth during the pandemic, and extend previous findings by examining the child, caregiver, and environmental factors associated with these changes in wellbeing.
Indeed, we found several developmental-ecological factors significantly associated with caregiver-reported autistic youth outcomes during the pandemic in the present study. Our findings are consistent with others who have demonstrated factors proximal to the child (e.g., age, pre-existing diagnosis), combined with broader contextual factors (e.g., caregiver stress, socioeconomic T A B L E 3 Results of the hierarchical linear regression predicting greater overall social-emotional symptoms of the sample. status) are associated with youth mental health in neurotypical children during non-pandemic times (e.g., Kemp et al., 2016). Regarding autistic characteristics during the pandemic, we found factors proximal to the youth such as older age, combined with broader contextual factors, including higher caregiver stress and greater household disruption due to COVID-19 were related to youth showing more autistic characteristics during the pandemic. Interestingly, other hypothesized factors such as level of autism severity and degree of service disruption were not associated with increased autistic characteristics. While other studies have also shown an increase in caregiverreported autistic traits in autistic youth during the pandemic (e.g., Asbury et al., 2021;White et al., 2021), our study is the first to extend this finding to show that older autistic youth whose caregivers experienced a high degree of stress and more household disruption may be more likely to demonstrate increased restricted and repetitive behaviors, social communication challenges, and sensory issues during the pandemic. Greater caregiver-reported social-emotional symptoms (i.e., behavior challenges, anxiety, and low mood) was significantly associated with older youth age and the presence of a youth pre-existing psychiatric condition in the present study. Further, higher caregiver stress as well as greater household and service disruption were significant ecological factors associated with greater socialemotional symptoms during the pandemic. Again, youth gender and level of autism severity did not play a significant role in this model. Likewise, youth age and gender as well as level of autism symptom severity were not significant predictors of better overall youth mental health during the pandemic in our study, although experiencing less household financial difficulty prior to COVID-19, having no youth pre-existing psychiatric condition, less caregiver stress, and less service disruption were significantly associated with better youth pandemic mental health.

Variable
Taken together, we found that across the final models higher caregiver stress was consistently associated with all outcome variables in the present study (i.e., autisticrelated traits, social-emotional symptoms, overall youth pandemic mental health). This finding suggests that higher caregiver stress may be especially related to youth autistic characteristics and social-emotional symptoms during the pandemic as well as overall youth pandemic mental health. This finding highlights the need to consider caregiver stress and wellbeing when supporting autistic youth during the pandemic, and following the pandemic.
Youth gender and level of autism symptom severity were also non-significant across all outcome variables of interest in the present study. Youth gender has shown mixed results in previous studies with some researchers suggesting being female may contribute to greater socialemotional concerns and poorer mental health during the pandemic in non-autistic youth (Oosterhoff et al., 2020;Racine et al., 2021;Zhou et al., 2020), while others (e.g., Xie et al., 2020), like us, have failed to find an association between youth gender and caregiver-reported pandemic mental health. Caregiver-reported autism symptom severity was also surprisingly non-significant across all final models in the present study. Consistent with previous research, we anticipated that more severe autistic symptoms prior to the pandemic to be related to greater autistic characteristics during the pandemic as well as increased social-emotional symptoms and poorer overall pandemic mental health. Importantly, our study sample included youth with primarily mild or moderate autism symptoms. As such, further investigation of autism symptom severity as a potentially influential youth factor with more severe presentations included in study samples is needed.
The presence of a youth pre-existing psychiatric condition according to caregivers as well as household and service disruption due to COVID-19 showed variable results across the final models in the present study. For instance, autistic youth having a pre-existing psychiatric condition was a significant predictor of increased social-emotional symptoms and poorer overall pandemic mental health, although it was not significantly associated with more autistic-related symptoms during the pandemic. Likewise, greater household disruption due to COVID-19 was a significant ecological factor for greater autistic-related traits and increased socialemotional symptoms during the pandemic; however, household disruption was not associated with overall youth pandemic mental health. Similarly, while service disruption did not significantly impact autistic traits during the pandemic, it was associated with increased social-emotional symptoms and poorer youth pandemic mental health. We suggest that the varied associations between child, caregiver, and environmental factors across outcome variables of interest in the present study represent the highly complex and multi-facet nature of wellbeing in autistic individuals. While we accounted for a notable degree of total variance in autistic characteristics, social-emotional symptoms, and overall youth mental health during the pandemic (i.e., 16%-23%), other important developmental-ecological factors clearly exist. Exploring additional factors not included in the present study and further clarifying mixed associations is crucial for understanding symptomology and mental health specific to autistic youth during the pandemic, and our findings underscore the importance of considering the additive influence of the child, caregiver, and environment in future investigations. The developmental-ecological factors identified in this study offer an important starting place to help identify those autistic youth who are most vulnerable to poorer outcomes, and therefore, require timely and targeted supports.

Implications
Our findings highlight the importance of clinicians and other service providers monitoring autistic youth for worsening autistic traits and social-emotional symptoms that may interfere with their quality of life during a widespread crisis such as the COVID-19 pandemic. During times of widespread crisis, resources targeting socialemotional symptoms and mental health are especially important for already vulnerable autistic youth and their families. For example, mitigation strategies that ensure continuous access to familiar therapeutic services, caregiver focused stress reduction tools, and autism-informed mental health care (e.g., trained mental health therapists, sensory-reduced urgent mental health clinics, tailored telephone/virtual hotlines) may be beneficial during subsequent pandemic waves and following the pandemic. Additional research should seek patient-informed feedback from youth and families about the supports needed to ensure resources adequately address their unique needs and lived experience. Research on the effectiveness of supports designed to mitigate symptom escalation in autistic youth during major stressful events is also needed.
Clinicians may wish to consider the demographic and developmental-ecological factors revealed in this study to be associated with increased autistic traits and socialemotional symptoms (i.e., behavior problems, anxiety, and low mood), and poorer overall pandemic mental health. Preliminary factors demonstrated in our study could help identify autistic youth who are especially vulnerable to reduced wellbeing in the face of major stressors in order to facilitate timely and targeted supports. For instance, supports appear especially important for older autistic youth with pre-existing psychiatric conditions, household financial difficulty prior to the major stressor (i.e., , caregivers with high perceived stress, and those who experience greater disruption to their household and professional services. Autistic youth with these features may be at greater risk for poorer outcomes than other autistic and non-autistic youth. As such, they may benefit from proactive screening and targeted support services.

Strengths and limitations
This study has several strengths including responses from a large national cohort of caregivers of autistic youth during the first COVID-19 wave in Canada. This study is the first, to our knowledge, to examine the additive influence of several child, caregiver, and environmental characteristics on symptomology and overall mental health in autistic youth during the pandemic. Further, our inclusion of multiple domains of autistic traits, socialemotional symptoms, and overall mental health is especially informative.
This study also has limitations. First, similar to other autism and COVID-19 studies (e.g., White et al., 2021), measures related to autistic characteristics, socialemotional symptoms, and household and professional service disruption were author created. Although these scales were modeled after established-validated survey items and the use of these measures was necessary to rapidly respond to the pandemic as a new area of inquiry, results of these measures should be interpreted cautiously. Second, an established measure used in this study (i.e., PSS-10) was not necessarily designed for caregivers of autistic individuals; however, has been used with this population in previous research examining caregiver mental health during the pandemic (e.g., Friesen et al., 2021). Similarly, while the item used to determine severity of autism symptoms has been used in other national studies (e.g., Zuckerman et al., 2018), caregivers were given no guidance about what was considered mild, moderate, and severe. Thus, caregiver responses may not be an accurate representation of their youth's strengths and abilities. These noted limitations related to the measures used in the current study may have also contributed to the degree of variance explained by the regression models. Third, increases in the severity of autism symptoms during the pandemic reported by caregivers may not necessarily be due to changes in behavior or characteristics but rather caregivers having more opportunity to observe their child due to physical restrictions during the pandemic, and thus should be interpreted with caution. Fourth, there is potential for bias in study participation due to our non-random public survey sampling method. For example, caregivers with available resources to complete a survey during a time of widespread crisis may possess different characteristics (e.g., more time, added caregiving support, less stress) that facilitated their survey completion, whereas highly overwhelmed caregivers may have limited ability to fill out a survey. Alternately, most highly stressed caregivers may seek to share their experience, while those experiencing less difficulties may be less likely to respond. Caregivers stress was consistently associated with all outcome variables; however, it is possible that those caregivers who are highly stressed may have interpreted their youth's behavior and wellbeing as worse due to additional stressors, such as educational or therapeutic services being interrupted or canceled. We also had an under-representation of low-income families, rural families, and non-married/non-common law caregivers who may be more at-risk groups for poorer mental health during major stressors such as the pandemic (Whitehead et al., 2021). Most autistic youth in this study had mild or moderate severity of autistic traits, with fewer severe autistic traits, and it is unknown the proportion of our sample that had a co-occurring intellectual disability. COVID-19 safety measures also varied across provincial Canadian jurisdictions during the pandemic, which may impact individuals from various geographic areas differently (e.g., varying durations of school closures). Future studies should recruit more diverse caregiver and autistic samples to detect possible demographic differences and enhance generalizability of study findings. Most importantly, all survey responses were caregiverreported and not reported directly by autistic individuals themselves. Future studies would largely benefit from incorporating multi-modal evaluation including caregiver-and self-report measures. Finally, our retrospective cross-sectional study design is based on caregiver recollection of symptoms and wellbeing prior to the pandemic compared to presently. This measurement approach is not as accurate at estimating symptom change overtime as longitudinal data with multiinformant standardized measurement. Future longitudinal research is required to understand directionality of the associations presented.

CONCLUSION
Autistic youth were at heightened risk for experiencing mental health concerns prior to the pandemic, and new pandemic-related stressors pose an added threat to their wellbeing. More than half of autistic youth in this study experienced increased autistic characteristics, behavior problems, anxiety, and low mood as well as poorer overall mental health. We found several developmentalecological factors associated with increased symptomology and poorer pandemic mental health such as older youth age, pre-existing mental health conditions, high caregiver perceived stress, and greater household and service disruption. These developmental-ecological factors should be considered in support strategies to improve the wellbeing of autistic people during major stressful events in the future. Support strategies such as maintaining or adapting, rather than canceling professional services, offering caregiver stress reduction supports, and ensuring households are adequately able to meet their essential needs during major stressful events such as the pandemic are needed. Policy/decision-makers, neurodevelopmental and mental health clinicians, and researchers should consider proactive screening and preventive support for autistic youth and their families, especially those with the developmental-ecological risk factors found in this study. Proactive screening and preventive support may protect against worsening symptoms and poorer mental health outcomes in this already vulnerable group of people.

ACKNOWLEDGEMENT
We would like to thank all the caregivers of autsitic people who participated in the Families Facing COVID study.

CONFLICT OF INTEREST STATEMENT
We have no conflicts of interest to disclose.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.