Abstract
OBJECTIVE
School attendance by children with medical complexity (CMC) may be influenced by parent perceptions of their child’s risk for coronavirus disease 2019 (COVID-19). The authors of this study aimed to quantify in-person school attendance and identify attendance predictors.
METHODS
From June to August 2021, surveys were collected from English- and Spanish-speaking parents of children aged 5 to 17 years with ≥1 complex chronic condition who received care at an academic tertiary children’s hospital in the Midwestern United States and who attended school prepandemic. The outcome, in-person attendance, was defined dichotomously as any in-person attendance versus none. We evaluated parent-perceived school attendance benefits, barriers, motivation, and cues, COVID-19 severity and susceptibility using survey items derived from the health belief model (HBM). Latent HBM constructs were estimated with exploratory factor analysis. Associations between the outcome and the HBM were evaluated with multivariable logistic regression and structural equation models.
RESULTS
Among 1330 families (response rate 45%), 19% of CMC were not attending in-person school. Few demographic and clinical variables predicted school attendance. In adjusted models, family-perceived barriers, motivation, and cues predicted in-person attendance, whereas benefits, susceptibility, and severity did not. The predicted probability (95% confidence interval) of attendance ranged from 80% (70% to 87%) for high perceived barriers to 99% (95% to 99%) for low perceived barriers. Younger age (P <.01) and previous COVID-19 infection (P = .02) also predicted school attendance.
CONCLUSIONS
Overall, 1 in 5 CMC did not attend school at the end of the 2020 to 2021 academic year. Family perceptions of schools’ mitigation policies and encouragement of attendance may be promising avenues to address this disparity.
Children with medical complexity (CMC) experience distinct coronavirus disease 2019 (COVID-19) risks owing to their multisystem chronic conditions, functional limitations, and service needs.1 Early in the pandemic, CMC had nearly 3 times the odds of severe COVID-19 illness and accounted for the majority of COVID-19 inpatient encounters with severe disease.2 These realities caused many families with CMC to question whether and when to have their child attend school in person. Weighing the risks of COVID-19 against the benefits of in-person schooling became a dilemma for families, educators, and health care teams.3–5
Many challenges of in-person schooling during COVID-19 are unique for CMC, particularly mitigation measures. The intensive hands-on, direct care required for CMC throughout their day can contradict physical distancing (e.g., administering medications or enteral tube nutrition, toileting, mobility assistance). The use of barrier masks, although effective, may be contraindicated or impractical and poorly tolerated because of sensory challenges, sialorrhea, or airway/ventilation needs.6,7 Many CMC need daily aerosol-generating procedures, which can increase their risk for COVID-19 transmission (e.g., nebulizer treatments, suctioning, tracheostomy care).8 CMC can also have daily respiratory symptoms that are hard to distinguish from acute COVID-19, which makes it difficult for school personnel to know when to test and prohibit in-person attendance. Many CMC are assisted continuously while attending school, including transportation to, from, and within school, and meals, hygiene, instruction, medical care, and therapy; therefore, the daily number of staff exposures per CMC can be high.9
Despite infection risks, in-person school attendance is an essential component of a healthy life for CMC.10 At school, CMC receive health-promoting services, including physical, occupational, and speech therapy. Virtual learning platforms may not be suitable alternatives to in-person instruction because intellectual and developmental disabilities can impair the ability of CMC to engage with them.9,11–13 When CMC are not in school, parents of CMC, already disproportionately unemployed because of their child’s care needs,14,15 face unrealistic demands to provide educational needs in addition to around-the-clock caregiving duties.
Our objective was to quantify the frequency of in-person school attendance for CMC at the end of the 2020 to 2021 academic year when most schools in Wisconsin had returned to in-person instruction and apply the health belief model (HBM) to identify family perceptions associated with in-person attendance. We hypothesized that parent perceptions of COVID-19 risk to their child (lower susceptibility and severity of COVID-19) would predict more in-person school attendance. Understanding the determinants of decisions to send CMC back to school could help identify parents’ unique COVID-19-related concerns and direct interventions to support safer school attendance for this vulnerable population.
Methods
Study Design, Setting, Participants
This prospective cohort study includes the first wave (June 2, 2021 to August 31, 2021) of a longitudinal study examining a safe return to school for CMC as part of the National Institutes of Health Rapid Acceleration of Diagnostics-Underserved Populations program. Caregivers were asked to complete a survey about their child’s experience in school. We enrolled 1 self-identified primary caregiver (usually a parent) of CMC aged 5 to 17 years who receive care at our academic tertiary children’s hospital in the Midwestern United States. Eligibility criteria included English or Spanish speakers and prepandemic in-person school attendance, which was chosen to focus on CMC who were not homebound. Medical complexity was defined by the presence of at least 1 complex chronic condition (CCC),16 an International Classification of Diseases-based classification system identifying medical conditions that are expected to last at least 12 months and that involve several different organ systems severely enough to require hospitalization in a tertiary care center. When relevant, parents were allowed to respond for multiple children because perspectives may differ for each child. The study was approved by the University of Wisconsin Institutional Review Board.
Survey Instrument
The survey instrument was created from items from the National Institutes of Health Rapid Acceleration of Diagnostics-Underserved Populations Common Data Elements adapted for the pediatric population17 and included novel questions designed by our team to assess constructs adapted from the HBM. The HBM is particularly suited to evaluate parent decisions to send CMC to school. It is theoretically grounded, evidence-based, and models the influence of perceived susceptibility, severity, benefits, barriers, motivations, and external cues on individual actions. Several studies have used HBM to model pandemic behaviors (e.g., behaviors to avoid COVID-19,18 decisions to receive a vaccine,19 contact tracing,20 among others). Before fielding, the novel survey items were refined by methodologist review at the University of Wisconsin Survey Center and pilot tested with 3 parents not enrolled in the study.
Exposures
Exposures were caregiver-rated benefits, barriers, severity, susceptibility, cues, and motivation. The survey questions used Likert scales and were crafted from HBM guidance.21
Primary Outcome
In-person school attendance was the HBM action of interest and was defined dichotomously as parent report of any in-person instruction (fully in-person or hybrid) at the end of the 2020 to 2021 academic year versus no in-person instruction (fully virtual) among those reporting the option to attend school. Those reporting that they did not have the option for in-person school were classified as missing for the primary outcome and were not included in the analyses of this outcome.
Covariates
Covariates assumed to have potential confounding relationships with HBM constructs and the primary outcome were defined a priori. Among respondents, these included age, language spoken at home (English, non-English), income (<$35 000, $35 000–74 999, $75 000–99 999, ≥$100 000), and currently employed (yes, no). Among respondents’ CMC, these included age, grade (preK–5, 6–8, 9–12), sex, race/ethnicity, primary insurance (private, public, both, none), previous COVID-19 testing, and COVID-19 vaccine status. The severity of illness data were abstracted from the electronic health record for CMC and included CCCs and hospitalizations in the year before enrollment. Race/ethnicity data were included because of known racial/ethnic disparities in COVID-19 treatment, morbidity, and mortality.22
Data Collection
The University of Wisconsin Survey Center managed consent, enrollment, and survey data collection. The protocol was designed to maximize response rates, data quality, and representativeness of the pool of participants. We included internet and mail options, dual-language (English and Spanish) options, and pre- and post-incentives ($5 before and $50 after survey completion). Mailed surveys have been shown to increase participation of lower income/lower education households compared with internet-only design.23 Respondents received a mailed request to participate with a paper copy of the survey questionnaire and a printed URL to an internet version of the survey. Two reminder mailings containing paper copies of the survey questionnaire and a printed URL to an internet version of the survey were sent to non-responders at 4-week intervals.
Statistical Analysis
Respondent and CMC characteristics were summarized in terms of frequencies and percentages. To facilitate descriptive interpretation, dichotomous versions of each of the HBM items were created to reflect affirmative versus neutral/negative responses from the Likert scales. χ-squared tests were used to evaluate differences in each participant characteristic and HBM items between participants whose CMC were attending versus not attending in-person school.
Exploratory factor analysis (EFA) was used to reduce HBM items and determine the factor structure of the HBM constructs. Eigenvalues and a scree plot were used to determine the number of factors, with the final structure identifying 3 factors: benefits, barriers, and severity. Factor loadings and the corresponding standard errors were reported. Internal consistency between items within the 3 factors was determined with Cronbach’s α. In the analysis, 3 HBM constructs were outside the 3-factor EFA structure and were therefore represented through the following single items with the largest differences between CMC attending versus not attending in-person school: susceptibility (“How well is your child’s school able to follow the recommendations to keep your child safe while in-person?”), motivation (“Based on the situation right now, how much do you want your child to attend school in-person at least some of the time?”), and cues (“Has a teacher or staff member encouraged your child to attend school in-person?”). In subsequent analyses, HBM constructs were represented categorically in 1 of 3 categories (“low” <25th percentile, “moderate” “25–75th percentile, or “high” >75th percentile) on the basis of the mean of Likert scales of items within the construct.
Univariate and multivariate logistic regression analyses were used to estimate the probability of in-person school attendance. Because 1242 (94%) families were unique, we did not account for the clustering of sibling CMC within families. Multivariate models included the HBM constructs and covariates with statistically significant univariate associations and no collinearity. The final model included the HBM constructs, child grade, and whether the child ever tested positive for COVID-19. Structural equation modeling (SEM) was performed to evaluate the association between in-person school attendance (outcome variable) and observed and latent predictor variables, including modifying factors (latent variables), cues to in-person school attendance (observed variable), child’s grade (observed variable), and the presence/absence of a previous positive COVID-19 test (observed variable). The preliminary structure of the SEM was based on an a priori theoretical construct that included various observed and latent predictor variables. The final structure of the SEM was determined by evaluating the addition or removal of potential predictors based on model fit criteria. Parameter estimation of the SEM was performed by using the maximum likelihood approach. The SEM analysis was conducted by using Mplus software, version 8.4. All other statistical analyses were conducted by using SAS software (SAS Institute, Cary NC), version 9.4. All reported P values are 2-sided, and P <.05 was used to define statistical significance.
Our power estimates were based on a conservative estimate of n = 1000 responses and assumed 90% of enrolled participants would have an option for in-school attendance, making the effective sample size n = 900. This sample was adequate to detect moderate effect sizes (1.5 ≤ odds ratio ≤ 2.0) at the 2-sided 0.05 significance level in the cross-sectional logistic regression models predicting in-school attendance with 80% power under various scenarios and binary predictors (x = 0 − absent/x = 1 − present). Anticipating response rates near 50% from an estimated eligible pool of approximately n = 3000, our study was expected to be adequately powered.
Results
We received 1330 responses from families of 2977 CMC (response rate = 45%). Table 1 summarizes respondent and CMC characteristics. In the cohort, 10% of participants spoke a language other than English at home. Respondents’ children were split between elementary (44%), middle (23%), and high (33%) school. Approximately one-quarter (23%) of the CMC had a neuromuscular CCC, 28% had 2 or more CCCs, and 18% had 1 or more hospitalization in the previous year. At the time of the survey (spring/summer 2021), 9% of participants reported that their CMC had ever tested positive for COVID-19 and 29% reported that their child had completed the COVID-19 vaccine course.
TABLE 1.
Respondent and Child Characteristics and Differences by In-Person School Attendance Status
| Attending in-person school*, n (%) | ||||
|---|---|---|---|---|
| Overall, n (%) | Yes | No | P | |
| n = 1049 (81%) | n = 251 (19%) | |||
| Grade | ||||
| PreK–5 | 580 (44) | 476 (45) | 102 (41) | .002 |
| 6–8 | 303 (23) | 255 (24) | 45 (18) | |
| 9–12 | 422 (32) | 318 (30) | 104 (41) | |
| Child age | ||||
| ≤6 y | 142 (11) | 119 (11) | 23 (9) | .16 |
| 7–12 y | 576 (44) | 472 (45) | 102 (41) | |
| 13–17 y | 587 (45) | 458 (44) | 126 (50) | |
| Respondent (parent) age | ||||
| <30 y | 31 (2) | 29 (3) | 2 (1) | .27 |
| 30–40 y | 430 (33) | 351 (33) | 76 (30) | |
| 40–50 y | 634 (49) | 499 (48) | 133 (53) | |
| >50 y | 199 (15) | 162 (15) | 37 (15) | |
| Not reported | 11 (1) | 8 (1) | 3 (1) | |
| Child sex, female | 622 (48) | 499 (48) | 119 (47) | .96 |
| Child race/ethnicity | ||||
| White, non-Hispanic | 1072 (82) | 856 (82) | 214 (85) | .14 |
| Black, non-Hispanic | 51 (4) | 46 (4) | 5 (2) | |
| Hispanic | 93 (7) | 79 (8) | 12 (5) | |
| Multiracial, non-Hispanic | 29 (2) | 24 (2) | 5 (2) | |
| Other, non-Hispanic | 60 (5) | 44 (4) | 15 (6) | |
| Language spoken at home | ||||
| English | 127 (10) | 102 (10) | 23 (9) | .79 |
| Non-English | 1178 (90) | 947 (90) | 228 (91) | |
| Family income | ||||
| <$35 000 | 154 (12) | 127 (12) | 26 (10) | .09 |
| $35 000–74 999 | 270 (21) | 216 (21) | 52 (21) | |
| $75 000–99 999 | 188 (14) | 139 (13) | 48 (19) | |
| >$100 000 | 553 (42) | 458 (44) | 94 (37) | |
| Not reported | 140 (11) | 109 (10) | 31 (12) | |
| Currently employed, yes | 972 (74) | 802 (76) | 166 (66) | .0008 |
| Insurance status | ||||
| Private | 806 (62) | 645 (61) | 158 (63) | .28 |
| Public | 299 (23) | 243 (23) | 54 (22) | |
| Private and public | 178 (14) | 140 (13) | 38 (15) | |
| None | 22 (2) | 21 (2) | 1 (0) | |
| Child ever tested for COVID-19, yes | 936 (72) | 788 (75) | 145 (58) | <.0001 |
| Child ever tested positive for COVID-19, yes | 113 (9) | 104 (10) | 9 (4) | .001 |
| Child completed COVID-19 vaccine course, yes | 374 (29) | 278 (27) | 95 (38) | .0004 |
| Complex chronic conditions | ||||
| 1 | 934 (72) | 771 (74) | 159 (63) | <.0001 |
| 2 | 230 (18) | 188 (18) | 42 (17) | |
| 2+ | 141 (11) | 90 (9) | 50 (20) | |
| Neuromuscular complex chronic condition, yes | 300 (23) | 234 (22) | 64 (26) | .28 |
| Respiratory complex chronic condition, yes | 65 (5) | 42 (4) | 23 (9) | .0008 |
| Technology assistance complex chronic condition, yes | 165 (13) | 109 (10) | 55 (22) | <.0001 |
| Hospitalizations year before enrollment | ||||
| 0 | 1072 (82) | 867 (83) | 200 (80) | .27 |
| ≥1 | 233 (18) | 182 (17) | 51 (20) | |
In-person school “Yes” was defined as full-time or hybrid in-person attendance; “No” was defined as fully virtual school attendance. Those without the option to attend in-person school were defined as “missing.”
Nearly 1 in 5 CMC (19%) with the option to attend in-person school was not attending at the end of the 2020 to 2021 academic year. Few demographic variables had univariate associations with in-person school attendance (Table 1). CMC attending in-person school were more often younger, had been previously tested or had tested positive for COVID-19, were unvaccinated, or only had 1 CCC. Those with respiratory CCCs and technology assistance reported less in-person school.
Differences in parent perceptions of school safety for CMC who were and were not attending in-person school are summarized in Table 2. Most respondents (86%) with CMC attending school reported that in-person school is important to their child’s overall health compared with only 47% for those not attending school (P <.001). In unadjusted analyses, every perception was significantly different between groups in expected directions (ie, those with in-person attendance expressed less concern over COVID-19 safety, the school’s ability to provide mitigation measures, or their child’s vulnerability to COVID-19). Items with the greatest difference (all >40%) between those attending versus not attending in-person school were: (1) comfort with the number of people around the child at school, (2) the school’s ability to stop COVID-19 spread, (3) the school’s ability to follow recommendations to keep their child safe, and (4) teachers or staff members encouraging the child’s attendance (all P <.001). Nearly 90% of those attending school reported that school personnel encouraged attendance compared with only 35% of those not attending school.
TABLE 2.
HBM School Attendance Perceptions for Children With Medical Complexity at the End of the 2020 to 2021 Academic Year
| Attending In-Person School,a n (%) | |||||
|---|---|---|---|---|---|
| HBM Construct | Item* | Overall, n (%) | Yes | No | P |
| n = 1049 (81%) | n = 251 (19%) | ||||
| Benefits | Attending school in-person is positive for...your family | 1069 (82) | 922 (88) | 144 (57) | <.001 |
| Attending school in-person is positive for...your child’s classmates | 968 (74) | 821 (78) | 145 (58) | <.001 | |
| Attending school in-person is positive for...the staff and teachers | 1010 (77) | 854 (81) | 153 (61) | <.001 | |
| Attending school in-person is important to your child’s overall health | 1019 (78) | 901 (86) | 117 (47) | <.001 | |
| Compared with fully virtual school, attending any school in-person is better for your child | 1105 (85) | 952 (91) | 149 (59) | <.001 | |
| Your child’s therapy needs are only met by attending school in-person | 449 (34) | 393 (37) | 54 (22) | <.001 | |
| Your family takes more precautions to avoid COVID-19 while your child is attending school in-person | 533 (41) | 365 (35) | 164 (65) | <.001 | |
| Your child attending school in-person help the adults in your family to keep their jobs | 592 (45) | 514 (49) | 76 (30) | <.001 | |
| Barriers | Comfort with... the number of people around your child at school | 638 (49) | 606 (58) | 31 (12) | <.001 |
| Comfort with... how close people have to be to your child at school | 636 (49) | 591 (56) | 44 (18) | <.001 | |
| Comfort with...the amount of personal protective equipment, such as masks and gloves, available at school | 821 (63) | 732 (70) | 88 (35) | <.001 | |
| Comfort with...the amount of COVID-19 testing among school staff and classmates | 624 (48) | 573 (55) | 50 (20) | <.001 | |
| Comfort with...the ability of your child’s school to take all precautions to stop the spread of COVID-19 | 807 (62) | 740 (71) | 64 (26) | <.001 | |
| Comfort with...how closely parents of classmates follow recommendations to keep your child safe | 492 (38) | 456 (43) | 35 (14) | <.001 | |
| Difficult to transport your child to or from school as a result of COVID-19 | 79 (6) | 49 (5) | 30 (12) | <.001 | |
| In your child’s school, they have access to necessary facilities to wash | 1210 (93) | 992 (95) | 213 (85) | <.001 | |
| In school, your child is required to be in close contact (ie, within 6 ft) with others | 571 (44) | 427 (41) | 142 (57) | <.001 | |
| In school, your child is able to wear a mask | 1166 (89) | 960 (92) | 202 (80) | <.001 | |
| In school, your child and their caregivers have access to necessary personal protective equipment | 1145 (88) | 960 (92) | 181 (72) | .0002 | |
| Severity | If your child was sick with COVID-19, their health would be severely affected | 217 (17) | 150 (14) | 66 (26) | <.001 |
| If your child was sick with COVID-19, they would have grave consequences | 156 (12) | 106 (10) | 49 (20) | <.001 | |
| If your child was sick with COVID-19, their health would be permanently reduced | 127 (10) | 83 (8) | 44 (18) | <.001 | |
| Susceptibility | Your child’s school is able to follow the recommendations to keep your child safe while in-person | 843 (65) | 769 (73) | 71 (28) | <.001 |
| Your child is likely to get sick with COVID-19 by attending school in-person | 163 (12) | 83 (8) | 78 (31) | <.001 | |
| Most of the people who interact with your child at school have been fully vaccinated | 415 (32) | 357 (34) | 57 (23) | .0007 | |
| Motivation | You want your child to attend school in person at least some of the time | 586 (45) | 515 (49) | 70 (28) | <.001 |
| Cues | A teacher or staff member encouraged your child to attend school in person | 1024 (78) | 935 (89) | 88 (35) | <.001 |
In-person school “Yes” was defined as full-time or hybrid in-person attendance; “No” was defined as fully virtual school attendance. Those without the option to attend in-person school were defined as “missing.”
Table 3 summarizes the EFA results, identifying 3 latent HBM constructs: perceived attendance benefits and barriers and COVID-19 severity. The benefits construct included 3 survey items (Cronbach α = 0.85) about how positive school attendance is for the child, their classmates, and school personnel. The barriers construct included 5 survey items (Cronbach α = 0.91) about how comfortable the family was with school mitigation measure implementation (e.g., the number and distance of those around the child, the amount of personal protective equipment and testing available at the school, and the ability of the school to take precautions to stop COVID-19 spread). The severity construct included 3 survey items (Cronbach α = 0.94) about the seriousness of COVID-19 infection for their child.
TABLE 3.
In-Person School Attendance Benefits, Barriers, and Severity Constructs Derived Through Exploratory Factor Analysis
| Factor Analysis | |||
|---|---|---|---|
| HBM Construct* | Item | Factor Loading (SE) | Cronbach’s α |
| Benefits | How positive is your child attending school in-person for...your family? | 0.67 (0.02) | 0.85 |
| How positive is your child attending school in-person for...your child’s classmates? | 0.90 (0.01) | ||
| How positive is your child attending school in-person for...the staff and teachers? | 0.87 (0.01) | ||
| Barriers | How comfortable are you with... the number of people around your child at school? | 0.92 (0.01) | 0.91 |
| How comfortable are you with... how close people have to be to your child at school? | 0.93 (0.01) | ||
| How comfortable are you with...the amount personal protective equipment, such as masks and gloves, available at school? | 0.74 (0.01) | ||
| How comfortable are you with...the amount of COVID-19 testing among school staff and classmates? | 0.80 (0.01) | ||
| How comfortable are you with...the ability of your child’s school to take all precautions to stop the spread of COVID-19? | 0.75 (0.01) | ||
| Severity | If your child was sick with COVID-19, how likely would their health be severely affected | 0.87 (0.01) | 0.94 |
| If your child was sick with COVID-19, how likely would they have grave consequences | 0.98 (0.01) | ||
| If your child was sick with COVID-19, how likely would their health be permanently reduced | 0.91 (0.01) | ||
SE, standard errors of factor loading.
HBM constructs outside the 3-factor structure were represented through the following single items: (1) susceptibility, “How well is your child’s school able to follow the recommendations to keep your child safe while in-person?”; (2) motivation, “Based on the situation right now, how much do you want your child to attend school in-person at least some of the time?”; and (3) cues, “Has a teacher or staff member encouraged your child to attend school in-person?”
In adjusted logistic regression models (Table 4), perceived barriers, motivation, and cues showed “dose-dependent” relationships with in-person school attendance (i.e., incrementally more affirmative perceptions were associated with higher predicted probabilities of school attendance). The adjusted predicted probability (95% confidence interval) of attending school was 80% (70% to 87%) for high perceived barriers, 81% (68% to 90%) for low motivation, and 99% (95% to 99%) for low perceived barriers. After adjustment for motivation, we did not observe associations between in-person school attendance and the HBM constructs of benefits, susceptibility, or severity. Younger child age (P <.01) and previous COVID-19 infection (P = .02) were significantly associated with in-person school attendance. These findings were similarly illustrated in the structural equation path analysis (Fig 1).
TABLE 4.
Adjusted Predicted Probability of In-Person School Attendance Among Children with Medical Complexity at the End of the 2020–2021 Academic Year
| Predicted Probability (95% CI) | ||||
|---|---|---|---|---|
| Unadjusted | P | Adjusted* | P | |
| Benefits | ||||
| Higha | 94% (92% to 96%) | <.0001 | 94% (88% to 97%) | .54 |
| Moderateb | 79% (75% to 82%) | <.0001 | 95% (91% to 98%) | .10 |
| Lowc | 67% (61% to 71%) | Ref | 93% (87% to 97%) | Ref |
| Barriers | ||||
| Higha | 54% (49% to 59%) | <.0001 | 80% (70% to 87%) | <.0001 |
| Moderateb | 87% (84% to 90%) | <.0001 | 93% (88% to 96%) | .008 |
| Lowc | 98% (96% to 99%) | Ref | 99% (95% to 99%) | Ref |
| Severity | ||||
| Higha | 69% (62% to 75%) | <.0001 | 94% (87% to 97%) | .37 |
| Moderateb | 79% (76% to 82%) | <.0001 | 93% (88% to 97%) | .18 |
| Lowc | 91% (87% to 93%) | Ref | 96% (91% to 98%) | Ref |
| Susceptibility | ||||
| Higha | 45% (38% to 53%) | <.0001 | 95% (88% to 98%) | .34 |
| Moderateb | 84% (81% to 86%) | <.0001 | 95% (90% to 97%) | .97 |
| Lowc | 94% (90% to 96%) | Ref | 92% (84% to 96%) | Ref |
| Motivation | ||||
| Higha | 95% (93% to 96%) | <.0001 | 98% (96% to 99%) | <.0001 |
| Moderateb | 79% (73% to 84%) | <.0001 | 96% (92% to 98%) | <.0001 |
| Lowc | 41% (36% to 47%) | Ref | 81% (68% to 90%) | Ref |
| Cues | ||||
| Yes | 88% (85% to 90%) | <.0001 | 95% (92% to 98%) | .006 |
| No | 75% (71% to 78%) | Ref | 91% (86% to 96%) | Ref |
| Grade | ||||
| PreK–5 | 82% (78% to 85%) | .0134 | 95% (90% to 97%) | .005 |
| 6–8 | 85% (81% to 89%) | .0018 | 97% (93% to 98%) | .0005 |
| 9–12 | 75% (71% to 79%) | Ref | 90% (82% to 95%) | Ref |
| Child ever tested positive for COVID-19 | ||||
| Yes | 92% (85% to 96%) | .0109 | 97% (91% to 99%) | .02 |
| No | 83% (81% to 86%) | Ref | 91% (85% to 95%) | Ref |
CI, confidence interval.
>75th percentile.
25th–75th percentile.
<25th percentile.
Models were adjusted for child grade in school and the presence/absence of previous COVID-19 infection for benefit, barriers, severity, susceptibility, motivation, and cues. For grade, the model was adjusted for the presence/absence of previous COVID-19 infection. For a child who ever tested positive for COVID-19, the model was adjusted for grade.
FIGURE 1.
Structural equation model path analysis depicting relationships among health belief model variables and in-person school attendance.
Shown is a path analysis from structural equation modeling. Numbers correspond to slope estimates of predictors, and statistically significant relationships are illustrated with red text. Squares represent observed variables, and ovals represent latent constructs. Cues, motivation, and perceived barriers were significantly associated with in-person school attendance in the final model. Susceptibility was reflected by the question “How well is your child’s school able to follow the recommendations to keep your child safe while in-person?”. Cues was reflected by the question “Has a teacher or staff member encouraged your child to attend school in-person?”. Motivation was reflected by the question “Based on the situation right now, how much do you want your child to attend school in-person at least some of the time?”.
Discussion
Overall, 1 in 5 CMC with the option to attend in-person school who attended school prepandemic was not attending school at the end of the 2020 to 2021 academic year. The HBM proved to be an informative tool to understand this outcome. For example, although demographic and clinical variables had few associations with school attendance, decisions for CMC to attend school were strongly associated with potentially modifiable concepts, including perceptions about school mitigation measures, encouragement from school personnel, and family motivation. These findings highlight the value of studying the perceptions of families of CMC and practices around school attendance and suggest potential intervention targets.
At the point in the pandemic when this study was conducted, most schools in Wisconsin were open and the Centers for Disease Control and Prevention24 and the American Academy of Pediatrics25,26 recommended that all children attend school, including guidance for in-person school attendance for children with special health care needs. Yet, the observation that nearly 20% of families of CMC opted out of in-person attendance complemented media reports5 and suggested that COVID-19 recommendations (or their implementation) may inadvertently overlook the distinct needs of CMC. This idea is underscored by another study in which personally and professionally diverse stakeholders identified distinct priorities for safer school attendance for CMC during the pandemic.4
Several other studies examining parent attitudes and practices around school attendance during the pandemic27–29 reveal that this may be an underrecognized COVID-19 morbidity for vulnerable populations. For example, it was striking that, even for those whose children were attending school in our study, only one-half responded affirmatively that they actually wanted their child to attend in-person. This begs the question of whether given an option for alternatives, an even greater percentage of CMC would not be attending school. In fact, our sample’s broad definition of medical complexity may even underestimate in-person attendance rates for CMC. In another study we conducted with a narrower population of CMC within a complex care program, 40% of that cohort was not attending in-person school during the same period.30
These data also provide useful contextual insights about COVID-19 mitigation activities for CMC. For example, nearly all families, regardless of school attendance, reported their CMC can mask, which is reassuring because it is one of the most effective mitigation strategies.6,31 Although nearly all families felt that schools had adequate facilities to wash, multiple opportunities remain. Even among the CMC attending school, interactions with unvaccinated individuals were common, and the number and proximity of others around the child at school were uncomfortable for families. Regardless of the impact on school attendance, school districts should attempt to address these concerns through advocacy and communication around vaccine requirements for students and for staff/volunteers at state and district levels. Health care providers can coach families to express concerns to district administrators and state officials.
We were surprised by the lack of associations between the perceived benefits of in-person school and school attendance. One potential explanation is that, for families of CMC, the perceived risk of attendance may simply outweigh the benefits. However, this is inconsistent with the simultaneously observed lack of associations between perceived susceptibility or severity of COVID-19 infection with school attendance. During the analyses, we observed that each of these concepts (benefits, susceptibility, severity) was no longer significant after including motivation in the models. We suspect that the motivation item reflected a simple and holistic expression of families’ synthesis of these risk/benefit calculations.
Motivation, barriers, and cues could serve as future intervention targets because each was associated with in-person attendance. Comfort with mitigation measures (barriers) and encouragement from school personnel (cues) may be particularly amenable to school and public health actions. School districts should not underestimate the magnitude that explicitly welcoming CMC to attend school has on families. It is also important for school districts to recognize that the safety of CMC often depends on practices that promote the safety of all students and staff. As schools navigate policy decisions, they should consider the influence that broad policy decisions (eg, mask use) have on individuals with high-risk medical conditions that increase their vulnerability to COVID-19. Tremendous variability exists in the mitigation measures implemented across school districts throughout the pandemic.32 Despite this variability, all CMC are entitled to reasonable accommodations under Title II of the Americans with Disabilities Act and Section 504 of the Rehabilitation Act. Policies facilitating the full participation of CMC in educational programming that would otherwise be inaccessible are paramount.33,34 Although parent-reported comfort with mitigation measures at school does not necessarily equate to what school mitigation policies are or should be, it is an important signal to which schools should respond. In some cases, simple messaging of what is happening could dispel misconceptions if they exist. In other cases, individualized education plans that support adaptations to promote safety and inclusivity are critical.
Whether family perceptions change if school policies change is an important future research question. Many of the challenges introduced by COVID-19 are relevant to other communicable diseases, including influenza, respiratory syncytial virus, norovirus, and others. For example, even before the pandemic, families of CMC may have been uncomfortable with their child’s risk for transmissible illness during traditional school attendance, and pandemic responses may not affect these long-standing concerns.
The study should be interpreted in light of several limitations. These data were collected at a discrete point in the pandemic shortly after schools had restarted in-person instruction. Perceptions have likely changed over time. Future longitudinal analyses using data we continue to collect will complement this study to investigate changes with time. The study’s findings may have limited generalizability due to unmeasured selection bias and because the cohort was relatively well-educated and less racially/ethnically diverse than other United States geographic regions. Given the relative absence of relationships observed between demographics and in-person school attendance, the impact of nonresponse on the outcome may be modest. During the pandemic, the Wisconsin Department of Public Instruction and Department of Health Services provided guidance; however, decisions on specific COVID-19 mitigation policies were left to individual school districts. Therefore, families expressed perceptions relative to their own districts, and we could not confirm specific school policies in the study. Our study was well-powered to detect most differences and the risk of type II error should be small.
Despite these limitations, our study illustrates that a sizable proportion of families decided not to send their CMC back to in-person school in spring 2021. Parent desire for in-person school attendance was influenced by multiple factors, including a combination of perceived risks and benefits. Family comfort with mitigation measures at school and encouragement from school personnel were independently associated with in-person school attendance. Policies that promote family comfort and encourage attendance are promising strategies to maximize school attendance for CMC.
Acknowledgments
Brooke Walker, MS, Duke Clinical Research Institute, provided editorial review and submission. Ms Walker did not receive compensation for her contributions, apart from her employment at the institution in which this study was conducted.
Glossary
- CCC
complex chronic condition
- CMC
children with medical complexity
- COVID-19
coronavirus disease 2019
- EFA
exploratory factor analysis
- HBM
health belief model
- SEM
structural equation model
Footnotes
Dr Coller conceptualized the study, interpreted the data, drafted the initial manuscript, and critically reviewed and revised the manuscript; Dr Eickhoff and Ms Zhao assisted with study design, analyzed the data, and critically reviewed and revised the manuscript; Drs Kelly, Johnson, Butteris, Ehlenbach, and DeMuri, Ms Katz, and Mr Koval contributed to study conceptualization and data analysis interpretation and critically reviewed and revised the manuscript; Ms Warner and Ms Howell assisted with data collection, interpreted the data, and critically reviewed and revised the manuscript; and all authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.
FUNDING: Funded by the National Institutes of Health (NIH). Research reported in this publication was supported by the Office of the Director of the NIH under award number U24MD016258; NIH Agreement Nos OT2HD107558 and OT2HD108110; the National Center for Advancing Translational Sciences of the NIH under award number U24TR001608; and the National Institute of Child Health and Human Development of the NIH under contract HHSN275201000003I. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
CONFLICT OF INTEREST DISCLOSURES: The authors have indicated they have no potential conflicts of interest relevant to this article to disclose.
References
- 1. Cohen E, Kuo DZ, Agrawal R, et al. Children with medical complexity: an emerging population for clinical and research initiatives. Pediatrics. 2011;127(3):529–538 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Kompaniyets L, Agathis NT, Nelson JM, et al. Underlying medical conditions associated with severe COVID-19 illness among children. JAMA Netw Open. 2021;4(6):e2111182. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Girard S; The Capital Times . Updated: survey shows more than 7,000 MMSD elementary students will return for in-person instruction. Available at: https://captimes.com/news/local/education/local_schools/updated-survey-shows-more-than-7-000-mmsd-elementary-students-will-return-for-in-person/article_f896d273-7d8d-5def-a4be-80f8721ecf55.html. Accessed March 3, 2021
- 4. Kelly MM, DeMuri GP, Barton HJ, et al. Priorities for safer in-person school for children with medical complexity during COVID-19. Pediatrics. 2022;149(3):e2021054434. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. National Public Radio . Parents of children with disabilities join the legal battle over masks in schools. Available at: https://www.npr.org/sections/back-to-school-live-updates/2021/09/07/1034918212/parents-of-children-with-disabilities-join-the-legal-battle-over-masks-in-school. Accessed September 9, 2021
- 6. Zimmerman KO, Akinboyo IC, Brookhart MA, et al. ; ABC Science Collaborative . Incidence and secondary transmission of SARS-CoV-2 infections in schools. Pediatrics. 2021;147(4):e2020048090. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Dorfman D, Raz M. Mask exemptions during the COVID-19 pandemic-a new frontier for clinicians. JAMA Health Forum. 2020;1(7):e200810. [DOI] [PubMed] [Google Scholar]
- 8. Tran K, Cimon K, Severn M, et al. Aerosol generating procedures and risk of transmission of acute respiratory infections to healthcare workers: a systematic review. PLoS One. 2012;7(4):e35797. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Constantino JN, Sahin M, Piven J, et al. The impact of COVID-19 on individuals with intellectual and developmental disabilities: clinical and scientific priorities. Am J Psychiatry. 2020;177(11):1091–1093 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Barnert ES, Coller RJ, Nelson BB, et al. A healthy life for a child with medical complexity: 10 domains for conceptualizing health. Pediatrics. 2018;142(3):e20180779. [DOI] [PubMed] [Google Scholar]
- 11. Asbury K, Fox L, Deniz E, et al. How is COVID-19 affecting the mental health of children with special educational needs and disabilities and their families? J Autism Dev Disord. 2021;51(5):1772–1780 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Eshraghi AA, Li C, Alessandri M, et al. COVID-19: overcoming the challenges faced by individuals with autism and their families. Lancet Psychiatry. 2020;7(6):481–483 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Rose J, Willner P, Cooper V, et al. The effect on and experience of families with a member who has intellectual and developmental disabilities of the COVID-19 pandemic in the UK: developing an investigation. Int J Dev Disabil. 2020;68(2):234–236 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Bitsko RH, Visser SN, Schieve LA, et al. Unmet health care needs among CSHCN with neurologic conditions. Pediatrics. 2009;124(Suppl 4):S343–S351 [DOI] [PubMed] [Google Scholar]
- 15. Okumura MJ, Van Cleave J, Gnanasekaran S, Houtrow A. Understanding factors associated with work loss for families caring for CSHCN. Pediatrics. 2009;124(Suppl 4):S392–S398 [DOI] [PubMed] [Google Scholar]
- 16. Feudtner C, Feinstein JA, Zhong W, et al. Pediatric complex chronic conditions classification system version 2: updated for ICD-10 and complex medical technology dependence and transplantation. BMC Pediatr. 2014;14:199. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. NIH RADx-UP . Common data elements. Available at: https://radx-up.org/research/cdes/. Accessed July 2, 2022
- 18. Fathian-Dastgerdi Z, Khoshgoftar M, Tavakoli B, Jaleh M. Factors associated with preventive behaviors of COVID-19 among adolescents: applying the health belief model. Res Social Adm Pharm. 2021;17(10):1786–1790 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Zampetakis LA, Melas C. The health belief model predicts vaccination intentions against COVID-19: a survey experiment approach. Appl Psychol Health Well-Being. 2021;13(2):469–484 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Walrave M, Waeterloos C, Ponnet K. Adoption of a contact tracing app for containing COVID-19: a health belief model approach. JMIR Public Health Surveill. 2020;6(3):e20572. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Champion VL. Instrument development for health belief model constructs. ANS Adv Nurs Sci. 1984;6(3):73–85 [DOI] [PubMed] [Google Scholar]
- 22. Tai DBG, Shah A, Doubeni CA, et al. The disproportionate impact of COVID-19 on racial and ethnic minorities in the United States. Clin Infect Dis. 2021;72(4):703–706 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Dillman DA, Smyth JD, Christian LM. Internet, Phone, Mail, and Mixed-Mode Surveys: The Tailored Design Method. Hoboken, NJ: John Wiley & Sons; 2014 [Google Scholar]
- 24. Centers for Disease Control and Prevention . Guidance for COVID-19 prevention in K-12 schools. Available at: https://www.cdc.gov/coronavirus/2019-ncov/community/schools-childcare/k-12-guidance.html. Accessed July 3, 2022
- 25. American Academy of Pediatrics . Caring for children and youth with special health care needs during the COVID-19 pandemic. Available at: https://www.aap.org/en/pages/2019-novel-coronavirus-covid-19-infections/clinical-guidance/caring-for-children-and-youth-with-special-health-care-needs-during-the-covid-19-pandemic/. Accessed July 3, 2022
- 26. American Academy of Pediatrics . COVID-19 guidance for safe schools. Available at: https://www.aap.org/en/pages/2019-novel-coronavirus-covid-19-infections/clinical-guidance/covid-19-planning-considerations-return-to-in-person-education-in-schools/. Accessed July 3, 2022
- 27. Sherby MR, Kalb LG, Coller RJ, et al. Supporting COVID-19 school safety for children with disabilities and medical complexity. Pediatrics. 2022;149(12 Suppl 2):e2021054268H. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Driansky A, Pilapil M, Bianco K, et al. Caring for children with medical complexity in the early COVID-19 pandemic in an ambulatory primary care setting. Front Pediatr. 2022;10:813393. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Castro-Kemp S, Mahmud A. School closures and returning to school: views of parents of children with disabilities in England during the Covid-19 pandemic. Frontiers Ed. 2021;6:666574 [Google Scholar]
- 30. Coller RJ, Kelly MM, Howell KD, et al. In-home COVID-19 testing for children with medical complexity: feasibility and association with school attendance and safety perceptions. Am J Public Health. 2022;112(S9):S878–S882 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Boutzoukas AE, Zimmerman KO, Inkelas M, et al. School masking policies and secondary SARS-CoV-2 transmission. Pediatrics. 2022;149(6):e2022056687. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. MCH Strategic Data . COVID-19 IMPACT: school district operational status: updates for fall 2021. Available at: https://www.mchdata.com/covid19/schoolclosings. Accessed September 9, 2021
- 33. U.S. Department of Education . Department of Education’s Office for Civil Rights opens investigations in five states regarding prohibitions of universal indoor masking. Available at: https://www.ed.gov/news/press-releases/department-educations-office-civil-rights-opens-investigations-five-states-regarding-prohibitions-universal-indoor-masking. Accessed September 9, 2021
- 34. Houtrow A, Harris D, Molinero A, et al. Children with disabilities in the United States and the COVID-19 pandemic. J Pediatr Rehabil Med. 2020;13(3):415–424 [DOI] [PubMed] [Google Scholar]