The COVID-19 pandemic has considerably impacted people across the globe regarding finances, work, education, and health [1–3]. In the general population, the literature suggests increased anxiety, stress, depression, and emotional distress in many countries across the globe [4–8]. Epidemiological and psychological research documents the increased burden on mental health (especially anxiety levels) related to the COVID-19 pandemic and highlights its adverse effects on vulnerable groups, such as those with neurodevelopmental conditions (NDCs) [9–11] and their families [12].
The NDC umbrella includes a wide range of conditions and disabilities often linked to the functioning of the brain and neurological systems, such as attention-deficit / hyperactivity disorder (ADHD), autism, learning disabilities (LD), intellectual disabilities (ID), and developmental language disorder (DLD), as well as genetic and chromosomal conditions including Down syndrome (DS) and Williams syndrome (WS). Due to their cognitive, mental health, and medical profiles, children with NDCs often depend on support and care services and institutions, such as specialized schools, day care services, therapists, clinicians, or medical care, which differ in the amount and type of support provided [13]. The rapid escalation of COVID-19 from an epidemic to a pandemic interrupted numerous mechanisms and systems designed to protect public health and well-being across the globe [14], especially in relation to the cessation of mental health or support services. The effects of the COVID-19 pandemic have been shown to vary depending on country of residence [15], financial standing [16], and family context [17,18], and have significantly impacted the lives of families with a child with an NDC throughout the pandemic, although mainly during the first months [19].
While several studies have documented an increase in emotional and behavioural problems in response to the COVID-19 pandemic in children with NDCs as well as their parents and caregivers [9,20–24], most of these studies relied on small sample sizes, focused on populations within one country such as the UK, Italy, China and / or on one specific diagnostic group, e.g., autism, Down syndrome, or Prader-Willi syndrome [10,21,22,25]. Furthermore, no previous research has examined how factors at the country (e.g., restrictions, health care spending), family (e.g., parental concerns about finances, illness), and child (e.g., children’s concerns, health) level may explain the variability in parental and child anxiety in multiple groups of NDCs. Such an approach is critical to inform research, policy, and the design of support systems for future crises. Therefore, the present study aimed to quantify to what extent families and their children with NDCs were impacted by the COVID-19 pandemic and how their anxiety levels were mitigated or aggravated by country, family, and individual child factors such as their health or other individual differences or concerns.
More specifically, we aimed at examining the anxiety levels and concerns of families and their children with various types of NDCs in the early months of the pandemic and compared their anxiety levels pre-pandemic (retrospective), at the time of the first outbreak (retrospective), and as they completed the survey (from April to August 2020). These anxiety levels and concerns can be seen as indicators of the extent and appropriateness of government responses to the pandemic, an early reaction to how public health systems were coping with the crisis. Since the anxiety levels of the people were likely affected by national demographic characteristics or government policy (country level: e.g., number of deaths, fiscal measures, overall obesity levels, emergency health care investment), family context (family level: e.g., concerns about family safety, family conflict) and child context factors (child level: e.g., concerns about loss of routine, becoming ill themselves), we integrated data concerning how the pandemic evolved, government responses, and structural descriptors of countries from the European Centre for Disease Prevention and Control [26], the University of Oxford [27], and the CIA World Factbook [28] to a multilevel data set. This multilevel data set was used to assess which of these national characteristics, familial contexts, and child context factors predicted anxiety of parents and of their child with an NDC to understand how well the care and support system was perceived to be performing as a buffer to the economic, social, and political disruptions caused by the pandemic.
To summarize, this paper explores for the first time the extent to which the anxiety levels of parents and their children with an NDC can be explained by different levels of variance as we described above (country, family, and child levels). While we expected anxiety levels to increase at the beginning of the pandemic compared to pre-pandemic levels in both, the parents and their child with an NDC, we did not formulate specific hypotheses about which of these different levels (country, family, or child) would explain anxiety best due to the limited research on anxiety in families and their children with NDCs with a cross-country perspective taking into account all of these levels at the same time.
METHODS
Design
A cross-sectional online survey with a non-random sampling method was employed. This allowed us to gather sociodemographic and mental health perspectives of individuals with NDCs as well as of their families, including their typically developing siblings.
Participants
Convenience sampling obtained 8043 participating families from seventy-eight countries (Section A, Table S1 in the Online Supplementary Document). After data cleaning, 6611 families from 70 countries remained, which included families with children with one of the most frequently diagnosed NDCs: autism, attention-deficit / hyperactivity disorder (ADHD), developmental language disorder (DLD), Down syndrome (DS), Williams syndrome (WS), and intellectual disability (ID). About one-third of the families also provided information about a typically developing sibling (TD). We decided to include the TD population in our analytical sample as previous research highlights that the mental health of siblings is often affected, either positively or negatively, by their sibling’s disabilities or neurodevelopmental conditions [29–31]. There were no constraints with respect to the age of the child with an NDC, so respondents could report about adults or children. For full sample descriptive statistics, see Table 1. Due to missingness in data, numbers vary per variable.
Table 1. Sample demographics and descriptive statistics
| Variable | All (8043 / 10 200) | ADHD (417) | AUTISM (2796) | DLD (502) | DS (1330) | ID (1646) | WS (427) | TD (3083) | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Mean (count) | SD (%) | Mean (count) | SD (%) | Mean (count) | SD (%) | Mean (count) | SD (%) | Mean (count) | SD (%) | Mean (count) | SD (%) | Mean (count) | SD (%) | Mean (count) | SD (%) | |
| Age (P), years | 42,93 | 10,03 | 41,35 | 8,82 | 41,81 | 8,8 | 38,78 | 9,25 | 46,13 | 10,34 | 43,01 | 11,21 | 45,89 | 10,9 | 43,16 | 9,72 |
| Age (C), years | 13,18 | 9,42 | 11,45 | 7,46 | 10,96 | 6,61 | 9,91 | 8,46 | 13,56 | 9,91 | 18,03 | 12,46 | 15,67 | 10,95 | 12,9 | 8,57 |
| Anxiety now (P), 1-5 | 2,96 | 1,29 | 2,91 | 1,3 | 3,03 | 1,26 | 2,55 | 1,37 | 2,99 | 1,29 | 2,9 | 1,31 | 3,02 | 1,2 | 2,96 | 1,29 |
| Anxiety now (P), 1-5 | 2,44 | 1,39 | 2,78 | 1,38 | 2,62 | 1,45 | 2 | 1,31 | 2 | 1,26 | 2,23 | 1,45 | 2,76 | 1,23 | 2,51 | 1,33 |
| Sex (P) | ||||||||||||||||
| Male | 2858 | 36% | 115 | 28% | 837 | 30% | 186 | 37% | 456 | 34% | 825 | 50% | 66 | 15% | 1100 | 36% |
| Female | 5146 | 64% | 298 | 71% | 1949 | 70% | 308 | 61% | 869 | 65% | 809 | 49% | 361 | 85% | 1975 | 64% |
| Sex (C) | ||||||||||||||||
| Male | 6122 | 60% | 275 | 66% | 2175 | 78% | 315 | 63% | 731 | 55% | 1492 | 91% | 237 | 56% | 1492 | 48% |
| Female | 3741 | 37% | 131 | 31% | 580 | 21% | 176 | 35% | 581 | 44% | 1393 | 85% | 187 | 44% | 1393 | 45% |
| Relation to child | ||||||||||||||||
| Father | 2516 | 31% | 100 | 24% | 787 | 28% | 162 | 32% | 400 | 30% | 677 | 41% | 63 | 15% | 1004 | 33% |
| Mother | 4641 | 58% | 271 | 65% | 1841 | 66% | 290 | 58% | 768 | 58% | 633 | 38% | 355 | 83% | 1836 | 60% |
| Other | 860 | 11% | 44 | 11% | 157 | 6% | 46 | 9% | 161 | 12% | 328 | 20% | 9 | 2% | 238 | 8% |
| Education (P) | ||||||||||||||||
| School leaving certificate | 2046 | 25% | 115 | 28% | 604 | 22% | 167 | 33% | 316 | 24% | 498 | 30% | 99 | 23% | 771 | 25% |
| University bachelor or equivalent | 2902 | 36% | 151 | 36% | 1041 | 37% | 147 | 29% | 500 | 38% | 639 | 39% | 120 | 28% | 1106 | 36% |
| University master or equivalent | 1392 | 17% | 61 | 15% | 616 | 22% | 36 | 7% | 263 | 20% | 144 | 9% | 100 | 23% | 8592 | 19% |
| Further vocational training | 859 | 11% | 40 | 10% | 306 | 11% | 68 | 14% | 141 | 11% | 131 | 8% | 82 | 19% | 325 | 11% |
| No formal qualification | 277 | 3% | 18 | 4% | 68 | 2% | 47 | 9% | 28 | 2% | 91 | 6% | 4 | 1% | 71 | 2% |
| Other | 529 | 7% | 28 | 7% | 149 | 5% | 33 | 7% | 79 | 6% | 132 | 8% | 19 | 4% | 212 | 7% |
| Work situation (P) | ||||||||||||||||
| Full time paid employment | 3316 | 41% | 180 | 43% | 1184 | 42% | 196 | 39% | 499 | 38% | 726 | 44% | 148 | 35% | 1253 | 41% |
| Part time paid employment | 1232 | 15% | 58 | 14% | 444 | 16% | 45 | 9% | 272 | 20% | 160 | 10% | 114 | 27% | 520 | 17% |
| Prime homemaker | 1725 | 21% | 94 | 23% | 618 | 22% | 169 | 34% | 243 | 18% | 344 | 21% | 70 | 16% | 675 | 22% |
| Retired | 678 | 8% | 24 | 6% | 147 | 5% | 32 | 6% | 165 | 12% | 203 | 12% | 38 | 9% | 242 | 8% |
| Student | 109 | 1% | 8 | 2% | 36 | 1% | 12 | 2% | 15 | 1% | 20 | 1% | 2 | 0% | 38 | 1% |
| Unemployed | 393 | 5% | 22 | 5% | 123 | 4% | 30 | 6% | 45 | 3% | 102 | 6% | 23 | 5% | 125 | 4% |
| Volunteer work | 83 | 1% | 2 | 0% | 33 | 1% | 4 | 1% | 15 | 1% | 13 | 1% | 6 | 1% | 35 | 1% |
| Other | 478 | 6% | 27 | 6% | 200 | 7% | 12 | 2% | 73 | 5% | 67 | 4% | 26 | 6% | 190 | 6% |
| COVID-19 infection (P) | ||||||||||||||||
| No | 6527 | 81% | 319 | 76% | 2244 | 80% | 390 | 78% | 1083 | 81% | 1285 | 78% | 373 | 87% | 2787 | 90% |
| Yes | 640 | 8% | 38 | 9% | 241 | 9% | 26 | 5% | 112 | 8% | 103 | 6% | 30 | 7% | 287 | 9% |
| COVID-19 infection (C) | ||||||||||||||||
| No | 8861 | 87% | 343 | 82% | 2401 | 86% | 402 | 80% | 1149 | 86% | 1356 | 82% | 395 | 93% | 2815 | 91% |
| Yes | 305 | 3% | 15 | 4% | 93 | 3% | 12 | 2% | 49 | 4% | 34 | 2% | 10 | 2% | 92 | 3% |
| Anxiety disorder (P) | ||||||||||||||||
| No | 4771 | 59% | 223 | 53% | 1594 | 57% | 272 | 54% | 845 | 64% | 883 | 54% | 300 | 70% | 2223 | 72% |
| Yes | 1433 | 18% | 82 | 20% | 553 | 20% | 72 | 14% | 180 | 14% | 268 | 16% | 37 | 9% | 671 | 22% |
| Anxiety disorder (C) | ||||||||||||||||
| No | 6112 | 60% | 168 | 40% | 1338 | 48% | 288 | 57% | 887 | 67% | 780 | 47% | 216 | 51% | 2435 | 79% |
| Yes | 1922 | 19% | 117 | 28% | 748 | 27% | 46 | 9% | 126 | 9% | 323 | 20% | 145 | 34% | 417 | 14% |
| Living arrangement now (C) | ||||||||||||||||
| At home with family | 9280 | 91% | 384 | 92% | 2638 | 94% | 466 | 93% | 1271 | 96% | 1461 | 89% | 413 | 97% | 2647 | 86% |
| In a group home | 199 | 2% | 13 | 3% | 61 | 2% | 11 | 2% | 26 | 2% | 73 | 4% | 5 | 1% | 10 | 0% |
| In a supported living setting | 109 | 1% | 4 | 1% | 43 | 2% | 1 | 0% | 15 | 1% | 32 | 2% | 8 | 2% | 6 | 0% |
| On his / her own | 110 | 1% | 0 | 0% | 22 | 1% | 5 | 1% | 0 | 0% | 14 | 1% | 1 | 0% | 68 | 2% |
| With a significant other | 157 | 2% | 4 | 1% | 8 | 0% | 4 | 1% | 2 | 0% | 21 | 1% | 0 | 0% | 118 | 4% |
| With roommate(s) | 66 | 1% | 4 | 1% | 8 | 0% | 1 | 0% | 6 | 0% | 18 | 1% | 0 | 0% | 29 | 1% |
| School now (C) | ||||||||||||||||
| At home with family | 7696 | 75% | 391 | 94% | 1989 | 71% | 397 | 79% | 1063 | 80% | 1317 | 80% | 320 | 75% | 2329 | 76% |
| Day care centre | 190 | 2% | 7 | 2% | 50 | 2% | 1 | 0% | 27 | 2% | 39 | 2% | 9 | 2% | 57 | 2% |
| Pre-school | 183 | 2% | 7 | 2% | 55 | 2% | 16 | 3% | 22 | 2% | 12 | 1% | 13 | 3% | 58 | 2% |
| School (mainstream) | 689 | 7% | 79 | 19% | 259 | 9% | 30 | 6% | 42 | 3% | 35 | 2% | 25 | 6% | 219 | 7% |
| School (special education) | 613 | 6% | 17 | 4% | 215 | 8% | 25 | 5% | 78 | 6% | 74 | 4% | 23 | 5% | 181 | 6% |
| Working in a job in the community | 4 | 0% | 0 | 0% | 0 | 0% | 0 | 0% | 0 | 0% | 2 | 0% | 1 | 0% | 1 | 0% |
| Working in a protected environment | 123 | 1% | 3 | 1% | 25 | 1% | 2 | 0% | 16 | 1% | 27 | 2% | 7 | 2% | 43 | 1% |
| Paid full or part time work | 86 | 1% | 2 | 0% | 23 | 1% | 9 | 2% | 6 | 0% | 19 | 1% | 4 | 1% | 23 | 1% |
| Other | 474 | 5% | 11 | 3% | 147 | 5% | 6 | 1% | 63 | 5% | 85 | 5% | 24 | 6% | 138 | 4% |
| Environment now (C) | ||||||||||||||||
| City / urban area | 5677 | 56% | 240 | 58% | 1582 | 57% | 272 | 54% | 711 | 53% | 1019 | 62% | 138 | 32% | 1715 | 56% |
| Town / suburban area | 2232 | 22% | 88 | 21% | 672 | 24% | 97 | 19% | 293 | 22% | 267 | 16% | 165 | 39% | 650 | 21% |
| Village / rural area | 2059 | 20% | 76 | 18% | 488 | 17% | 114 | 23% | 295 | 22% | 300 | 18% | 121 | 28% | 665 | 22% |
| Additional diagnosis (NDC child) | ||||||||||||||||
| No | 4622 | 57% | 232 | 56% | 1638 | 59% | 362 | 72% | 797 | 60% | 791 | 48% | 302 | 71% | , | , |
| Yes | 3388 | 42% | 179 | 43% | 1147 | 41% | 136 | 27% | 532 | 40% | 846 | 51% | 125 | 29% | , | , |
| Intellectual disabilities (NDC child) | ||||||||||||||||
| No | 2222 | 28% | 214 | 51% | 1147 | 41% | 284 | 57% | 105 | 8% | 86 | 5% | 8 | 2% | , | , |
| Mild | 3983 | 50% | 167 | 40% | 1247 | 45% | 190 | 38% | 891 | 67% | 832 | 51% | 349 | 82% | , | , |
| Severe | 1764 | 22% | 25 | 6% | 378 | 14% | 22 | 4% | 322 | 24% | 711 | 43% | 69 | 16% | , | , |
SD – standard deviation, Perc – percentage, ADHD – attention-deficit / hyperactivity disorder, A – autism, DLD – developmental language disorder, DS – Down syndrome, ID – intellectual disability-not otherwise specified, TD – typically developing sibling, WS – Williams syndrome, NDC – neurodevelopmental conditions
Materials
This study relies on a questionnaire that we developed at the beginning of the COVID-19 pandemic to assess how families and their children with NDCs were affected by the COVID-19 pandemic [32,33]. All materials, including the survey and related supplementary files can be accessed at: https://osf.io/5nkq9/. The first part of the questionnaire requested demographic information from a parent- or caregiver-respondent (for simplicity, we use the term “parent” here), including the primary diagnosis of their child with an NDC, presence or absence of intellectual disability, etc., see Table 2 for a full list of included variables. Anxiety and specific concerns (e.g., about COVID-19, health, loss of institutional support) were assessed on a scale from 1 (not at all) to 5 (extremely) at three time points: retrospectively before the pandemic began, retrospectively at the start of the pandemic, and at the time of survey completion (between April and August 2020). Respondents reported anxiety and concerns for themselves, their child with an NDC, and optionally for their TD child (the complete survey, including further questions not considered here, can be found here [32,33]).
Table 2. Variables in stepwise regression, grouped by stepwise block (0-4), selection (yes or no), order at which the variable entered the model, and interactional (yes or no)
| Variable | Stepwise block | Block N | Parental Anxiety Model selected | Parental Anxiety Model order | Parental Anxiety Model interaction | Child Anxiety Model selected | Child Anxiety Model order | Child Anxiety Model interaction |
|---|---|---|---|---|---|---|---|---|
| Diagnosis (C) | Base | 0 | Yes | 1 | Yes | Yes | 1 | Yes |
| Time | Base | 0 | Yes | 1 | Yes | Yes | 1 | Yes |
| Age (P) | Base | 1 | Yes | 1 | No | Yes | 1 | No |
| Sex (P) | Base | 1 | Yes | 1 | No | Yes | 1 | Yes |
| Age (C) | Base | 1 | Yes | 1 | Yes | Yes | 1 | Yes |
| Sex (C) | Base | 1 | Yes | 1 | No | Yes | 1 | No |
| Education (P) | Base | 1 | Yes | 1 | No | Yes | 1 | Yes |
| Relation to child | Base | 1 | Yes | 1 | Yes | Yes | 1 | No |
| Work situation (P) | Base | 1 | Yes | 1 | No | Yes | 1 | Yes |
| Anxiety disorder (C) | Base | 1 | Yes | 1 | No | Yes | 1 | Yes |
| Anxiety disorder (P) | Base | 1 | Yes | 1 | Yes | Yes | 1 | Yes |
| Days since the national peak | Base | 1 | Yes | 1 | Yes | Yes | 1 | Yes |
| Days since the pandemic | Base | 1 | Yes | 1 | Yes | Yes | 1 | Yes |
| Anxiety (P) | Family | 3 | Yes | 1 | Yes | 5 | Yes | |
| Concerns own possible illness (P)* | Family | 3 | Yes | 8 | No | Yes | 7 | Yes |
| Anxiety (C) | Child | 4 | Yes | 11 | Yes | Yes | 1 | No |
| Public debt | Country | 2 | Yes | 2 | No | No | ||
| Fiscal measures | Country | 2 | Yes | 3 | No | No | ||
| Emergency health care investment | Country | 2 | Yes | 4 | No | No | ||
| Concerns child approach (P)† | Family | 3 | Yes | 9 | No | No | ||
| Concerns own safety (P)‡ | Family | 3 | Yes | 7 | No | No | ||
| Concerns C-19 (P)§ | Family | 3 | Yes | 6 | Yes | No | ||
| Concerns child health (P)‖ | Family | 3 | Yes | 8 | No | No | ||
| Concerns own illness (P)¶ | Family | 3 | Yes | 5 | Yes | No | ||
| Airports | Country | 2 | No | Yes | 2 | No | ||
| Cancelation of public events | Country | 2 | No | Yes | 4 | No | ||
| Obesity | Country | 2 | No | Yes | 3 | No | ||
| Concerns child ability (P)** | Family | 3 | No | Yes | 6 | Yes | ||
| Concerns family conflict (P)†† | Family | 3 | No | Yes | 8 | Yes | ||
| Concerns child motivation (P)‡‡ | Family | 3 | No | Yes | 9 | No | ||
| Concerns COVID-19 (C)§§ | Child | 4 | No | Yes | 11 | Yes | ||
| Concerns family conflict (C)‖‖ | Child | 4 | No | Yes | 14 | No | ||
| Concerns illness (C)¶¶ | Child | 4 | No | Yes | 10 | Yes | ||
| Concerns routine (C)*** | Child | 4 | No | Yes | 13 | Yes | ||
| Concerns safety (C)††† | Child | 4 | No | Yes | 12 | Yes | ||
| Country area | Country | 2 | No | No | ||||
| Budget surplus | Country | 2 | No | No | ||||
| School closing at peak | Country | 2 | No | No | ||||
| School closing max | Country | 2 | No | No | ||||
| Workplace closing at peak | Country | 2 | No | No | ||||
| Work closing max | Country | 2 | No | No | ||||
| Cancelation of public events | Country | 2 | No | No | ||||
| Restrictions on gathering at peak | Country | 2 | No | No | ||||
| Restrictions on gathering max | Country | 2 | No | No | ||||
| Close public transportation at peak | Country | 2 | No | No | ||||
| Closing public transportation max | Country | 2 | No | No | ||||
| Stay at home at peak | Country | 2 | No | No | ||||
| Stay at home max | Country | 2 | No | No | ||||
| Restrictions on international travel at peak | Country | 2 | No | No | ||||
| Restrictions on international travel max | Country | 2 | No | No | ||||
| Cellphone users | Country | 2 | No | No | ||||
| Containment health index at peak | Country | 2 | No | No | ||||
| Containment health index max | Country | 2 | No | No | ||||
| Death rate | Country | 2 | No | No | ||||
| Death rate total | Country | 2 | No | No | ||||
| Deaths per month | Country | 2 | No | No | ||||
| Deaths per month total | Country | 2 | No | No | ||||
| Income support at peak | Country | 2 | No | No | ||||
| Income support max | Country | 2 | No | No | ||||
| Contract relief at peak | Country | 2 | No | No | ||||
| Contract relief max | Country | 2 | No | No | ||||
| Fiscal measures max | Country | 2 | No | No | ||||
| International support max | Country | 2 | No | No | ||||
| Economic support index at peak | Country | 2 | No | No | ||||
| Economic support index max | Country | 2 | No | No | ||||
| Education | Country | 2 | No | No | ||||
| Gross domestic product per capita | Country | 2 | No | No | ||||
| Government response index at peak | Country | 2 | No | No | ||||
| Government response max | Country | 2 | No | No | ||||
| Public information campaigns at peak | Country | 2 | No | No | ||||
| Public information campaigns max | Country | 2 | No | No | ||||
| Testing policy at peak | Country | 2 | No | No | ||||
| Testing policy max | Country | 2 | No | No | ||||
| Contact tracing max | Country | 2 | No | No | ||||
| Emergency health care investment max | Country | 2 | No | No | ||||
| Investment in vaccines max | Country | 2 | No | No | ||||
| Inflation | Country | 2 | No | No | ||||
| Internet users | Country | 2 | No | No | ||||
| Median age | Country | 2 | No | No | ||||
| Net migration | Country | 2 | No | No | ||||
| Population | Country | 2 | No | No | ||||
| Railways | Country | 2 | No | No | ||||
| Roadways | Country | 2 | No | No | ||||
| Stringency index at peak | Country | 2 | No | No | ||||
| Stringency index max | Country | 2 | No | No | ||||
| Country taxes | Country | 2 | No | No | ||||
| Telephone users | Country | 2 | No | No | ||||
| Unemployment | Country | 2 | No | No | ||||
| Youth unemployment | Country | 2 | No | No | ||||
| School closing now | Country | 2 | No | No | ||||
| Workplace closing now | Country | 2 | No | No | ||||
| Cancelation of public events now | Country | 2 | No | No | ||||
| Restrictions on gatherings now | Country | 2 | No | No | ||||
| Close of public transportation now | Country | 2 | No | No | ||||
| Stay at home now | Country | 2 | No | No | ||||
| Restrictions on international movement now | Country | 2 | No | No | ||||
| International travel controls now | Country | 2 | No | No | ||||
| Containment health index now | Country | 2 | No | No | ||||
| Deaths now | Country | 2 | No | No | ||||
| Deaths per month now | Country | 2 | No | No | ||||
| Income support now | Country | 2 | No | No | ||||
| Contract relief now | Country | 2 | No | No | ||||
| International support now | Country | 2 | No | No | ||||
| Economic support index now | Country | 2 | No | No | ||||
| Government response now | Country | 2 | No | No | ||||
| Public information campaigns now | Country | 2 | No | No | ||||
| Testing policy now | Country | 2 | No | No | ||||
| Contact tracing now | Country | 2 | No | No | ||||
| Investment in vaccines now | Country | 2 | No | No | ||||
| Stringency index now | Country | 2 | No | No | ||||
| Current government guidance | Family | 3 | No | No | ||||
| Family infected with COVID-19 | Family | 3 | No | No | ||||
| First hear about COVID0-19 | Family | 3 | No | No | ||||
| Following gov. advice | Family | 3 | No | No | ||||
| Concerns family’s safety (P)‡‡‡ | Family | 3 | No | No | ||||
| Self-isolation | Family | 3 | No | No | ||||
| Speaking to expert desire | Family | 3 | No | No | ||||
| Schools closed | Family | 3 | No | No | ||||
| Concerns self-isolation (P) | Family | 3 | No | No | ||||
| Self-isolation pre-gov. guidance | Family | 3 | No | No | ||||
| Self-isolation (P)§§§ | Family | 3 | No | No | ||||
| Expert discussion about child (P) | Family | 3 | No | No | ||||
| Concerns family safety post-social distancing announcement (P) | Family | 3 | No | No | ||||
| Concerns child boredom (P)‖‖‖ | Family | 3 | No | No | ||||
| Concerns child illness (P)¶¶¶ | Family | 3 | No | No | ||||
| Concerns finance (P)**** | Family | 3 | No | No | ||||
| Concerns child institution (P)†††† | Family | 3 | No | No | ||||
| Concerns child social (P)‡‡‡‡ | Family | 3 | No | No | ||||
| Concerns work balance (P)§§§§ | Family | 3 | No | No | ||||
| Occupation before (C) | Child | 4 | No | No | ||||
| Urbanity before (C) | Child | 4 | No | No | ||||
| Urbanity now (C) | Child | 4 | No | No | ||||
| Occupation now (C) | Child | 4 | No | No | ||||
| Fear communication (C) | Child | 4 | No | No | ||||
| Health index (C) | Child | 4 | No | No | ||||
| ID (C) | Child | 4 | No | No | ||||
| Residence before (C) | Child | 4 | No | No | ||||
| Residence now (C) | Child | 4 | No | No | ||||
| Medical issues (C) | Child | 4 | No | No | ||||
| COVID-19 infection (C) | Child | 4 | No | No | ||||
| COVID-19 awareness (C) | Child | 4 | No | No | ||||
| Concerns approach (C)‖‖‖‖ | Child | 4 | No | No | ||||
| Concerns boredom (C)¶¶¶¶ | Child | 4 | No | No | ||||
| Concerns finances (C)***** | Child | 4 | No | No | ||||
| Concerns friends (C)††††† | Child | 4 | No | No | ||||
| Concerns own illness (C)‡‡‡‡‡ | Child | 4 | No | No | ||||
| Concerns health (C)§§§§§ | Child | 4 | No | No | ||||
| Concerns institution (C)‖‖‖‖‖ | Child | 4 | No | No | ||||
| Concerns others illness (C)¶¶¶¶¶ | Child | 4 | No | No | ||||
| Typically developing sibling health problems | Child | 4 | No | No |
*How concerned were / are you about the possibility that you will get ill?
†How concerned were / are you that your child is not able to approach others?
‡How concerned were / are you about your child’s safety with respect to COVID-19?
§How concerned were / are you about COVID-19?
‖How concerned were / are you about your child’s health?
¶How concerned were / are you about illness in general?
**How concerned were / are you about your child’s ability to cope with changes in his / her routine?
††How concerned were / are you about family conflict (fights, aggression)?
‡‡How concerned were / are you about your ability to keep your child with NDCs entertained and motivated?
§§How concerned was / is your child about COVID-19?
‖‖How concerned was / is your child about family conflict?
¶¶How concerned was / is your child about illness in general?
***How concerned was / is your child about his or her loss of routine?
†††How concerned was / is your child about the family’s safety with respect to COVID-19?
‡‡‡How concerned did you feel at the tie for your family’s safety?
§§§Did your relevant government (city, state or country) suggest that you and your family should self-isolate? Self-isolation means stay at home (Yes / No answers)
‖‖‖How concerned were / are you about your child becoming bored?
¶¶¶How concerned were / are you about the possibility that your child will get ill?
****How concerned were / are you about your financial / economic situation?
††††How concerned were / are you about the loss of institutional (eg, school, workplace) support for your child including interventions (language therapists, psychologists etc.)?
‡‡‡‡How concerned were / are you about the fact that your child has fewer occasions for social contact and interaction?
§§§§How concerned were / are you about work / childcare balance related to looking after a child with NDCs?
‖‖‖‖How concerned was / is your child about not being able to approach others?
¶¶¶¶How concerned was / is your child about boredom?
*****How concerned was / is your child about his financial / economic situation?
†††††How concerned was / is your child about not being able to meet peers and friends?
‡‡‡‡‡How concerned was / is your child about the possibility that he / she will get ill?
§§§§§How concerned was / is your child about his / her own health?
‖‖‖‖‖How concerned was / is your child about the loss of institutional (eg, school, workplace) support for your child including interventions (language therapist, psychologist etc.)?
¶¶¶¶¶How concerned was / is your child about the possibility that others will get ill?
Procedure and ethics
The link to the online questionnaire was distributed via the networks of participating researchers worldwide in more than 24 countries and international family associations. The questionnaire was available in 16 languages (see limitations) [32,33]. Recruitment flyers were sent to associations and charities to invite parents and caregivers to report about their child with an NDC. Ethical approval for this anonymous questionnaire was obtained by the institutional review board of UniDistance Suisse.
Data analysis
The primary outcomes were self-reported (parent) anxiety and parent-reported anxiety of the child with an NDC (continuous variable) and of the typically developing (TD) child, which we analysed using stepwise multilevel linear regression. This model allowed us to treat individual families and countries as nested sources of random / sampling variance while handling predictors of fixed variance in the usual fashion (changing them in every block). Four blocks of fixed effects were added progressively to the model, consisting of (i) design or confirmatory variables (e.g., NDC group) in the base model, (ii) country variables (e.g., government response to COVID-19 [25]), (iii) family variables (e.g., family situation, parent concerns about COVID-19; concerns about their health and concerns about their child with an NDC), and (iiii) child variables (e.g., the child with NDC’s reported concerns about COVID-19; and concerns about motivation to name a few). As such, blocks (ii)-(iiii) reflected different levels of sources of anxiety from macro (country) to micro-level (child factors). By adding and eliminating variables within each block, we obtained a parsimonious model containing only the most important predictors for parental anxiety and anxiety of the child with an NDC. In selecting important effects, we focused on how time and the different levels of factors (family, country, and child) modified group dynamics in anxiety and whether they were mitigating or aggravating influences. Model output was quantified as marginal- and conditional-R2 for goodness-of-fit, F- and t tests for the inferential significance of predictors, and standardized regression differences as measures of effect. A full description of the data analysis procedure, including the list of variables in each regression block as well as how the missing data were handled, can be found in the Online Supplementary Document. The stepwise model sequence and fit statistics are also presented in section A, Table S2 in the Online Supplementary Document. In addition, Type II ANOVA breakdown for the final multilevel regression models are presented in section A, Table S3 in the Online Supplementary Document.
RESULTS
Our data show that the mean age for the participating parents was 42.93 years and 13.18 years for their child. The mean level of anxiety for the participating parent was 2.96 and 2.44 for their child (on a scale of 1 to 5). 17% of the parents had a university degree, and among the responding parents, the majority (58%) were mothers of children with an NDC.
Parent anxiety
The final model explained 60% of the marginal variation (R2marg = 0.601) in parent anxiety. The confirmatory predictors explained 21.9%, with country-, family-, and child-related variables incrementally adding 0.5%, 36.8% and 0.9%, respectively. This suggested that parent anxiety was best explained by family-related factors, as shown in Figure 1 for a breakdown of model effects and Table S3A in the Online Supplementary Document for the detailed Type II ANOVA of the final multilevel regression for parent anxiety.
Figure 1. Breakdown of model effects and interactions for Parent Anxiety Model. Black lines represent significant Time interactions.
The final model did not reveal a main effect of diagnosis group, F(6,11368.0) = 1.15, P = 0.3327, βz = 0.060, suggesting that parents experienced on average the same anxiety regardless of the NDC of their child (Figure 2).
Figure 2. Standardised Mean Parent Anxiety per NDCs. The lower the mean the greater the anxiety. Stdz – standardized, NDCs – neurodevelopmental conditions, ADHD – attention-deficit / hyperactivity disorder, A – autism, DLD – developmental language disorder, DS – Down syndrome, ID – intellectual disability-not otherwise specified, TD – typically developing sibling, WS – Williams syndrome
The final model revealed a main effect of time; F(2,17.654) = 160.33, P < 0.0001, βz = 0.286, such that average parental anxiety followed an inverted U-shape profile across the three time points. This profile consisted of a “shock”, a “peak”, and a “recovery”, with the shock represented by the large increase in average anxiety from before to the start of the COVID-19 pandemic; t(22.45) = -11.24, P = 0.000, βz = -.38), and the recovery represented by a decrease in average anxiety from the start of the pandemic to the now moment (ie, the moment the participants responded to the questionnaire), t(17.63) = 2.15, P = 0.045, βz = 0.057). This main effect was modified by nine interaction effects, some of which changed shock, recovery or both, as shown in line for Time in Figure 3. A higher-than-average concern about COVID-19 magnified the shock effect and speeded recovery, vs the average time trend, as shown in line concerns C-19 (P) in Figure 3. By contrast, having lower-than-average concerns about COVID-19 flattened the U-shaped time trend compared to the average, as shown in dashed concerns C-19 (P) in Figure 3.
Figure 3. Parent Anxiety Time Main Effect & Interactions with the Shock vs Recovery Effect. Peak is anchored to 0 to show the relative changes with respect to the U-Shape Turn. (C) refers to variables concerning the child, (P) refers to variables concerning the parent or caregiver. +1SD refers to a positive standard deviation away from the mean, -1SD refers to a negative standard deviation away from the mean. All variables are explained in Table 2.
Although some country factors such as country public debt, government fiscal measures in response to COVID-19 now, and emergency investment health care spending at the time of the participant completed the survey were predictive of parent anxiety and selected during stepwise model building, these factors were all non-significant in the final model (section A, Table S3 in the Online Supplementary Document), suggesting that these differences were better explained by family- and child-related factors. In general, the country’s structural characteristics did not affect parental anxiety, nor did the national government’s response to the COVID-19 pandemic. Parental anxiety was strongly predicted by specific health-related concerns, including the parent’s possibility of becoming ill, as presented with green line in Figure 3, or COVID-19 specifically, as presented with orange line in Figure 3. Parental anxiety was also strongly and solely predicted by the level of anxiety of their child with an NDC, in that parents who were anxious themselves tended to report higher anxiety in their children. Although the final model showed that the factors discussed above impacted parental anxiety, it is important to note that degree to which these four factors influenced anxiety differed between countries as reflected in the final model’s random effect structure. Figure 4 shows which countries were outlying for the differing effects. For example, parent anxiety levels in the Netherlands were best explained by parental concerns about COVID-19 and least explained by their concerns about Possible Illness compared to Luxembourg where concerns about Possible Illness had a stronger impact. It is important to note that these countries were not outliers specifically due to their sample size.
Figure 4. Differing effects of parent anxiety per outlying country. (P) refers to variables concerning the parent or caregiver.
Anxiety reported for children with NDCs and their TD siblings
The final model explained 57.6% of R2marginal variation in levels of anxiety for children with NDCs and their TD siblings. The confirmatory predictors explained 23%, with country-, family-, and child-related variables incrementally adding 2.8%, 9.7% and 21.2%, respectively. This suggested that levels of anxiety for children were best explained by the child factors in addition to the base factors. For a breakdown of model effects see Figure 5 and section A, Table S3B in the Online Supplementary Document for the detailed Type II ANOVA of the final multilevel regression for child anxiety.
Figure 5. Breakdown of model effects and interactions for Child Anxiety Model. Black lines represent significant Time interactions, and red lines represent significant Diagnosis interactions.
The final model revealed a main effect of the diagnosis group, F(6,13510.4) = 9.90, P < 0.0001, βz = -.154, suggesting that parents reported levels of child anxiety depended on the NDC of their child (Figure 6). This figure shows also that children with WS had the highest anxiety levels. This main effect was modified by 11 interaction effects, which reflected that some variables were more important in predicting anxiety for specific NDC groups with a notable interaction on pre-existing anxiety disorder interacting with the diagnosis; F(6,13533.2) = 4.61, P < 0.0001, βz = 0.066.
Figure 6. Standardised mean child anxiety per NDCs. The lower the mean the greater the anxiety. Stdz – standardized, NDCs – neurodevelopmental conditions, ADHD – attention-deficit / hyperactivity disorder, A – autism, DLD – developmental language disorder, DS – Down syndrome, ID – intellectual disability-not otherwise specified, TD – typically developing sibling, WS – Williams syndrome
Temporal proximity to the start of the pandemic and days since national peak predicted most strongly child anxiety for individuals with WS, whereas this had no impact on the DS and DLD groups. Likewise, whereas concerns about COVID-19 strongly predicted child anxiety in all groups, only for WS, TD, and DLD did child concerns about illness in general play an additional role. For all groups, child concerns about the loss of routine were a strong predictor, but it was the most relevant for individuals with WS, autism, and ADHD (Figure 7).
Figure 7. Diagnosis interactions with continuous predictors. WS – Williams syndrome, TD – typically developing sibling, ID – intellectual disability-not otherwise specified, DS – Down syndrome, A – autism, ADHD – attention-deficit / hyperactivity disorder, DLD – developmental language disorder.
The final model revealed a main effect of time F(2,17.654) = 160.33, P < 0.0001, βz = 0.286, such that average child anxiety followed an inverted U-shape profile across the three time points. Similarly to parental anxiety, this profile consisted of a “shock”, a “peak”, and a “recovery”, with the shock the large increase in average anxiety from before to the start of COVID-19 t(82.5) = -3.73, P = 0.0000, βz = -.069, and the recovery a decrease in average anxiety from the start to the now moment t(37.8) = -.82, P = 0.41, βz = -.016. This main effect was modified by eight interaction effects, some of which changed shock, recovery, or both, as shown in Figure 8.
Figure 8. Child Anxiety Time Main Effect & Interactions with the Shock vs. Recovery Effect. Stdz – standardised, (P) refers to variables concerning the parent or caregiver. +1SD refers to a positive standard deviation away from the mean, -1SD refers to a negative standard deviation away from the mean. All variables are explained in Table 2.
A higher-than-average concern about COVID-19 magnified the shock effect and sped up recovery vs the average time trend. By contrast, having lower-than-average concerns about COVID-19 flattened the U-shaped time trend compared to the average. As for the country effects, country public debt, government fiscal measures in response to COVID-19 up until the now time point, and emergency investment in health care spending now were selected during stepwise model building as predictive of child anxiety, although all were non-significant in the final modelling, suggesting these differences were better explained by family- and child-related factor differences. Finally, several effects from the four levels of influence also differed between countries, as reflected in the final model’s random effect structure. Figure 9 shows which countries were outlying with respect to the differing effects. For example, children’s anxiety was best explained by children’s anxiety disorder in Mexico compared to Greece where children’s concerns about Illness had a bigger effect. Similarly, the cause of these countries being outliers was not specifically due to sample size.
Figure 9. Differing effects of child anxiety per outlying country. (C) refers to variables concerning the child, (P) refers to variables concerning the parent or caregiver.
DISCUSSION
In this international study, we examined the anxiety and concerns of children with NDCs and their families. We assessed which factors relating to the country and government policy, family context, or child contexts were more likely to be related to increased anxiety using a multilevel approach. By adopting a cross-country perspective, this study adds to the mounting evidence of widespread increased stress and mental health concerns, specifically anxiety, in children with NDCs and their caregivers [24]. Parental anxiety was best explained first by family- and child-related predictors and then by country-related predictors. Similarly, children’s anxiety was explained best by child- and family-related predictors followed by country-related predictors. Anxiety levels increased strongly at the beginning of the pandemic in both parents and children which coincided with the first peak of the pandemic in the respective countries (with the peak of the pandemic being defined as the maximum number of deaths between April to August 2020) which aligns with previous research [2,5,9,10,20,21,34,35]. Over time, parental anxiety decreased slightly as the time distance from the first peak increased but it did not return to the perceived pre-pandemic level, indicating chronic increased stress in these families due to the pandemic and pandemic-related issues. However, this was not the case for child anxiety as a decrease in average anxiety from the start of the pandemic to the now moment was revealed in the final model.
Anxiety in both parents and children was linked to pre-existing anxiety conditions, so a pre-existing anxiety condition affected parental anxiety which also impacted the anxiety of their children and vice versa. In addition, we observed that various concerns affected anxiety strongly in both parents and their children with NDCs. Parents were mostly concerned about their children with NDCs not being able to approach others, COVID-19, and possibly getting ill. These concerns likely reflected parents’ fear that their children were at risk of contracting a serious form of COVID-19 and the impact that it would have on their child’s long-term health (e.g., children with WS have higher chances of experiencing severe medical issues [36,37]). In relation to child factors, parental anxiety was best explained through concerns about their children’s motivation, illness, and safety. Nevertheless, it is crucial to analyse country, family, and child characteristics, in addition to concerns that impacted the anxiety of parents and children the most. To date, no research has investigated the effects of COVID-19 using a global sample and employing cross-country models, hence it is difficult to comment on such findings.
Surprisingly, none of the government-level policy measures were significant factors predicting parental anxiety in the final model, which indicates that country-related contexts such as the public health system of a country did not have a direct impact on parental anxiety. However, when we added additional predictors to the analysis, variables other than government-related factors became significant. In the final multilevel analysis model, parental concerns related to their health, their children’s health, and their children’s lack of opportunity to get physically close to others due to anti-transmission measures (e.g., two-meter distance, social distancing) were better predictors of parental anxiety. In sum, health-related worries and concerns about limited opportunities for social interactions due to anti-transmission measures increased parental anxiety. The lack of social contact seems to be one of the main worries for families with a child with an NDC, likely related to concerns about their social development [38]. Interestingly, parental anxiety was not impacted by their child’s condition (group of NDC), showing that parents from different groups had similar anxiety levels. However, there were some differences between countries in how much family-related factors and concerns contributed to the parents’ anxiety. In sum, parental anxiety is mostly mediated by personal concerns and family factors within the context of a country, but not by country level factors per se.
When examining children’s anxiety levels, we see a similar picture. The government policies and family practices, which can be seen as vital components of the global public health system in protecting vulnerable children, did not impact immediately on children’s anxiety once family and child factors were taken into account. Instead, children’s anxiety was specifically related to base factors such as their sex and age to name a few. In addition, child factors played a key role in explaining anxiety in children with NDCs such as concerns about COVID-19. A detailed overview is presented in Figure 5. In addition, the analysis revealed how a child’s diagnostic group made a difference in how they coped with the onset of the pandemic, indicating that syndrome-specific vulnerability or proneness to anxiety impacted individual and group anxiety levels. In line with previous research [39,40], we identified the highest anxiety levels in children with WS. As revealed in several interactions within the NDC group, certain variables, including anxiety disorder and the child’s and parent’s specific concerns (e.g., about COVID-19, illness in general), seemed to affect the child’s anxiety in several diagnostic groups (section A, Table S3 in the Online Supplementary Document). Moreover, interactions with the diagnostic group were revealed (e.g., concerns about changes in their routine) which is in line with the generally elevated need of children with NDCs for consistency [41,42]. This finding indicates that the sudden changes to the daily structures of children with NDCs during the COVID-19 pandemic, due to the closure of schools and institutions, the discontinuation of mental health care services, as well as the necessary familial reorganization, greatly impacted these children and their families, independently of the country they lived in.
While this study’s strength is undoubtedly the large sample size and the number of countries included, several limitations need to be mentioned: The study (a) used self-reported NDC diagnoses; (b) relied on parent-reported anxiety for children; (c) did not use standardized tools to assess anxiety and specific concerns due to methodological constraints (e.g., the total length of the questionnaire, the availability of these tools in all languages, the lack of pilot data and validation of the tools); (d) the translation of the survey in other languages was not conducted by professionals but by the collaborators, (e) the study involved retrospective reporting about the course of the anxiety rather than contemporaneous rating. However, new research shows that humans can operationalize a scale for their emotions, hence the measure of anxiety through a single scale should still give us insights into their anxiety and stresses [43,44]. Nevertheless, the way anxiety has been assessed does not allow us to draw conclusions concerning the extent to which the anxiety reached clinical relevance. Despite these limitations, we believe that the current study can contribute to understanding the factors that influenced families with a child with an NDC during the first months of the COVID-19 pandemic, as previous research for single countries showed [9,21,45,46], and how country-specific, family, and child factors contributed to parent and child anxiety.
Implications
Whereas family factors such as concerns about COVID-19, the child’s safety, and the child’s health, to name a few examples, explained parental anxiety best, governmental measures and policies (Table 2) did not directly impact parental anxiety once some specific fixed-factor predictors were added in the final model. Similarly, for children with NDCs, country-level factors did not explain variance in anxiety once family and child-level factors were considered. Child factors such as concerns about COVID-19, family conflict, and loss of routine explained children with NDCs’ anxiety and showed that despite there being group-specific concerns and anxiety levels, the lack of routine was a significant contributor to all children’s anxiety. Taking these findings together, this study provides insight for future policy recommendations and is also informative for developing interventions and toolkits to help parents (who were more affected by anxiety), and children cope with future crises. For instance, toolkits that the public health system could provide to support families with children with NDCs through events such as a pandemic should emphasize the importance of re-establishing a new family routine, or to help children as well as their parents to regulate their anxiety, as other researchers have argued in the literature [10,47]. In addition, the findings show that parents and children had very specific concerns that are context specific and should be addressed in future crises. However, future studies should examine this context-specific nature and whether the fact that we did not identify an influence of country-level factors (e.g., public debt; fiscal measures and emergency health care investment) was because parents did not feel supported or because there was another common factor across different countries (e.g., lack of sustainable support through the pandemic).
Additional material
Online Supplementary Document
Acknowledgements
The authors would like to thank all the families who participated in the survey and parent associations and institutions willing to distribute our recruitment flyer. We would like to thank the chEERS Lab (Principal Investigator: Andrea C Samson) for coordinating the study and to various people involved in the translations of the survey (Arabic: Zuber Mohammed, Faisal Al Nemary; Chinese (Simplified): Xueyun Su & Hungtzu Claire Tai; Czech: Lenka Hrncirova; Dutch: Paulien Eijkeler, Jo van Herwegen; French: Anouk Papon, Noemie Treichel; German: Lina Stallmann, Ingolf Prosetzky; Greek: Angeliki Tampaki, Athina Papageorgiou, Panagiota Kovani, Evangelia Sarvani, Danai Orlandatou, Pinelopi Xoundri, Sofia Arvaniti; Italian: Giona Di Poi; Persian: Saied Sadeghi; Portuguese (Portugal): Guida Veiga; Portuguese (Brazil): Roberto Tadeu Iaochite; Romanian: Adela Chirita-Emandi, Puiu Maria, Dorica Dan, Alexandra Zaharia; Russian: Liudmila Gamaiunova; Spanish: Ruth Campos, Pastora Martínez-Castilla; Swedish: Sven Bölte, Eric Zander. We acknowledge Prof Dr Puiu Maria from the Regional Center of Medical Genetics Timis in the Clinical Emergency Hospital for Children “Louis Turcanu” Timisoara, Romania – a member of ERN ITHACA and Dorica Dan from NoRO Center Zalau (a member of ERN ITHACA) for their contribution to promote the questionnaire to the Romanian families. Many thanks go also to: Natalie Doppelt, Mary T Hanley, Anne Hugon, Robin Libove, Diego Lopergolo, Maria Cristina Nedelcu, and Margarida Tome. Finally, this work has been supported by the European Reference Network on Rare Congenital Malformations and Rare Intellectual Disability (ERN-ITHACA).
Ethics statement: Ethical approval for this anonymous questionnaire was obtained by the institutional review board of UniDistance Suisse and informed consent was obtained from all participants involved in the study.
Data availability: Data collected for this anonymous online study will be made available on open science framework (https://osf.io/uykdc/), including data pooled from the European Centre for Disease Prevention and Control, the University of Oxford, the CIA World Factbook, and the R scripts used for this study.