Our analytic sample consisted of 765 parents of immunosuppressed children who participated in 12 to 78 of our weekly surveys, providing in total of 38,449 observations with a median of 52 observations per parent. Household characteristics for these parents are provided in the ‘Full sample’ column of Table 1a and characteristics for their immunosuppressed child are provided in ‘Full sample column’ of Table 1b.

Our first step in the trajectory analysis was to determine the number of different trajectories that best represent parental concern over the first 18 months of the COVID-19 pandemic. Table 2 shows that the BIC and AIC values improve with increasing number of trajectories. A Bayes factor comparing the BIC of the simpler model with the more complex model (2(ΔBIC)) was calculated to determine the best fitting model. A 10-fold difference in Bayes factor was considered strong evidence that the more complex model best represented the data19. Convergence of the 5-trajectory model failed as the PROC TRAJ procedure found it difficult to estimate a 5-trajectory mathematical model among the mixture of parental concern trajectories provided using the likelihood estimation methodology employed by the procedure. It was concluded that the 4-trajectory model best fitted the data. The average posterior probability for the 4 different concern trajectories was well above the recommended minimum average posterior probability of 0.7 (ranged between 0.9969 and 0.9891), indicating that each pattern includes individuals with quite similar patterns of change in parental concern. In addition, the four distinct trajectories represented change in concern over time for parental group whose size was well over the 5% sample size threshold. The four trajectories represent (1) parents (20.1%) with medium levels of concern at the start of the COVID-19 pandemic but full adaptation over the next 18 months (2) parents (34.2%) with medium levels of concern at the start of the COVID-19 pandemic and some adaptation over the next 18 months (3) parents (34%) with high levels of concern at the start of the COVID-19 pandemic and some adaptation over the next 18 months (4) parents (11.7%) who experienced high levels of concern throughout the 18 months of the COVID-19 pandemic. Household and child characteristics as well as univariate comparisons between these four trajectory groups are presented in Table 1a, b.

The second step in the trajectory analysis was to determine the polynomial growth terms that best represented the true shape of the trajectories. For each of the different trajectories we compared linear, quadratic, and cubic shapes. Based on the significance level of the estimated trajectory parameters, the medium parental concern groups were best represented (Fig. 1) by a cubic trajectory shape represented by decreasing parental concern after the initial concern response at the beginning of the pandemic, followed by increased concern during the second lockdown and decrease of concern during the roadmap out of lockdown. The high parental concern groups were best represented by a quadratic shape of slightly increasing parental concern after initial response at the beginning of the pandemic followed by a decrease of concern during the roadmap out of lockdown.

Parental concern during the first 18 months of the COVID-19 pandemic.

In the third step of the trajectory analysis, we examined the association of time-stable covariates and parental concern trajectory membership by expanding the PROC TRAJ syntax with the ‘RISK’ statement. Significant associations are presented in Table 3. In these analyses the ‘medium concern—full adaptation’ group was chosen as the reference group. No differences were found between the four parental concern trajectory groups with regard to their immunosuppressed child becoming infected with SARS-CoV-2, the number of self-isolations, or missing out on school, sport/leisure activities.

Compared to the ‘medium concern—full adaptation’ group, parents in the ‘18 months high concern’ group were significantly less likely to be employed (p < 0.001) and more likely to live in the North of England (p < 0.01), more specifically their child was more likely to receive medical care in North-East England (p < 0.001), and significantly less likely to receive medical care in South-East England (p < 0.05). In addition, these parents were significantly more likely to care for children with nephrotic disease (p < 0.05), organ transplants (p < 0.01), respiratory diseases (p < 0.05), or children with diagnoses such as Trisomy-21, severe eczema, cardiac disorder with abnormal immune function, neurological disorder, haematological disorder (p < 0.001), and significantly less likely to care for children with Juvenile Idiopathic Arthritis (p < 0.001). These condition associations were mirrored by associations with the different drugs used in each condition (p < 0.05 to P < 0.001).

Compared to the ‘medium concern—full adaptation’ group, parents in the ‘high concern—some adaptation’ group were significantly less like to be single parents (p < 0.001), and more likely to care for children with other primary immunodeficiencies (p < 0.05), respiratory diseases (p < 0.05), children with diagnoses including Trisomy-21, eczema, cardiac disorder, neurological disorders, haematology disorders (p < 0.05) or children who are prescribed drugs like inhalers, eye drops, NSAIDs, folic acid, hypertonic saline, Omeprazole, insuline, hydrochloroquine, colchicine, antihypertensives, IViG, leflunomide, sulfasalazine, mercaptopurine (p < 0.05).

Compared to the ‘medium concern—full adaptation’ group, parents in the ‘medium concern—some adaptation’ group were significantly less likely to be single parents (p < 0.05) and significantly more likely to live in rural or semi-rural areas of the UK (p < 0.05).

Each week the parents were asked to describe their current concerns. The parental concerns for the first 18 months of the pandemic could be grouped into four overarching themes: impact on everyday life child; impact on household; SARS-CoV-2 risk; and long-term health condition. Under these overarching themes we identified several sub-themes (Table 4).

During the study, parents expressed both general and COVID-19-specific concerns about their child’s health. They worried about the impact of (non-SARS-CoV-2) infections and childhood illnesses on the health status of their immunosuppressed child. Specific to the COVID-19 pandemic, parents of immunosuppressed children were worried about the impact of delay in diagnostic procedures, referral, and health care provision on their child’s health status, experiencing first-hand the effects of an overstretched NHS. Children needed to isolate for upcoming medical procedures and parents reported concerns that children were suffering psychologically from parental separation during these hospitalisations as UK hospital rules in general allowed only one parent visitor. Some children experienced flare-ups/exacerbation of the symptoms of their long-term condition, which parents sometimes contributed to COVID-related anxiety and stress or lack of exercise.

Throughout the pandemic, parents’ perceived risk of SARS-CoV-2 infection was in line with their perceptions of various SARS-CoV-2 strains, availability and perceived effectiveness of COVID-19 vaccines, prior SARS-CoV-2 infection, perceived effectiveness of preventative COVID-19 guidelines, and public adherence to COVID-19 guidelines. Parents were worried about the impact of a SARS-CoV-2 infection on the health status of their immunosuppressed child, specifically would COVID-19 cause serious illness or even death? As the pandemic progressed, the fear of death subsided, but a fear of long-term COVID-19 implications emerged.

Some parents mentioned exercising continual vigilance to minimise potential SARS-CoV-2 exposure both for the immunosuppressed child and any other vulnerable individuals living in the household. In general, parents believed that the life of their immunosuppressed child had been put on hold during the pandemic. Clinically vulnerable children and their families have experienced shielding and/or repeated periods of isolation. Parental concerns suggested these events impacted social development, education, and mental health. Families had to persist through uncertain times expressing concerns about the lack of information and inconsistent advice or guidelines. Some parents reported that they were more protective of their immunosuppressed child and have been weighing up the risk of SARS-CoV-2 exposure versus the benefits of social interaction, forsaking the social life of various household members if deemed necessary. As the COVID-19 pandemic persisted, some parents wondered what life would be like for a vulnerable child in the age of the COVID-19 pandemic? Families also reported having missed out on activities, holidays and social events, although this may be true of all children not just those immunosuppressed.