Trends in comorbid physical and mental health conditions in children from 1999 to 2017 in England

Participants

We used data from the British Child and Adolescent Mental Health Surveys and the Mental Health of Children and Young People in England Survey. These are cross-sectional, population-based surveys of CYP conducted in 1999, 2004, and 2017. The 1999 (age 5–15 years; Great Britain) and 2004 (age 5–16 years; Great Britain) samples were drawn from the Child Benefit Register. The 2017 sample (age 2–19 years; England) was drawn from the NHS Patient Register, as Child Benefit was no longer universal. A detailed description of all three surveys is reported elsewhere [8]. The 1999 and 2004 data are publicly available via the UK Data Service (references 4227 and 5269, respectively), and the 2017 data were accessed via NHS Digital’s Data Access Request Service (DARS, reference DARS-NIC-424336-T7K7T-v0.6). The original surveys were approved by Research Ethics Committees [9, 10], while the University of Cambridge Ethics Committee does not require applications for secondary data analysis. Informed consent was obtained from legal guardians (for children < 11 years old) and from adolescent participants.

In all three surveys, data were collected from parents (94% from the biological mothers), children (aged ≥ 11 years), and teachers (the family nominated the teacher who knew the child best). Parents and CYP were interviewed face-to-face by trained lay interviewers using computer-assisted interviews, with self-reported questionnaires completed for sensitive topics. Teachers responded to a mailed questionnaire. To increase the comparability between the three surveys, we included children from England only (excluded n = 2514) aged 5–15 years old (excluded n = 3502). Participants with incomplete data on study measures were also excluded (n = 234).

Measures

Mental health conditions were measured using the validated Development and Well-being Assessment [11], a standardized diagnostic tool that combines highly structured and semi-structured questions about psychiatric disorders, based on the International Classification of Diseases-10 (ICD-10) criteria [12]. The DAWBA incorporated information from parents (via interview), CYP (via interview), and teachers (via questionnaire). Computer-generated predictions of disorders were produced and reviewed by trained clinicians (including co-authors TF and TND) who assigned diagnoses based on ICD-10 criteria [12]. The DAWBA covered the following: (emotional disorders [separation anxiety disorder, specific phobia, social phobia, panic disorder, agoraphobia, post-traumatic stress disorder, obsessive compulsive disorder, generalized anxiety disorder, other anxiety disorder, depressive episode, other depressive episode], hyperactivity disorders [hyperkinetic disorder, other hyperactivity disorder], conduct disorders [oppositional defiant disorder, conduct disorder confined to family, unsocialized conduct disorder, socialized conduct disorder, other conduct disorder], less common psychiatric disorders [autistic spectrum disorder, tic disorder, eating disorder, selective mutism, psychosis]).

pLTCs were assessed using parent report on whether the child has any of the following conditions at the time of assessment (yes/no): asthma, eczema, hayfever, epilepsy, cerebral palsy, muscle disease, coordination problems, heart problems, food allergy, kidney/urinary tract problems, a condition present since birth (e.g., club foot or cleft palate), deformities, spina bifida, cystic fibrosis, blood disorders, missing limb(s), diabetes, cancer, vision problems, hearing problems. This selection was based on the consensus definition of chronic health conditions in childhood [13] and discussion with our Study Steering Committee.

Statistical analysis

We estimated the population prevalence of CYP with comorbid pLTCs and mental health conditions in 1999, 2004, and 2017. We also estimated the population prevalence of children without identified pLTCs or mental health conditions (healthy group), with pLTCs but no mental health conditions, and with mental health conditions but no pLTCs at the three time points. Prevalence analyses used survey weighting (provided by the three national surveys) to account for selection and non-response bias [8]. Not all teachers responded to the mailed questionnaire [14]; For the 2017 survey, weighting does not fully adjust for the teacher report sub-sample. Nevertheless, applying the teacher adjustment factor, provided for the group of children with mental health conditions [14], had minimal impact relative to our estimated 95% Confidence Interval (CI). Given that the teacher adjustment factor is provided for children with mental health conditions and it is not broken down for our sub-groups (children with pLTCs only, children with mental health conditions only, children with comorbid pLTCs and mental health conditions, healthy children), by applying this factor would not give correct estimates.

In the sub-group of children with comorbid pLTCs and mental health conditions, we undertook linear regression analysis to test for a linear trend in survey weighted prevalence across the three surveys (the trend analysis was weighted by the standard errors for each survey to account for uncertainty in the estimates). In sensitivity analyses, we examined whether there were significant changes in the proportion of children with comorbidities by including the whole Great Britain sample (the 1999 and 2004 surveys included children from Scotland and Wales, while the 2017 survey only sampled from England). Analyses were carried out in STATA 17.0 and R programming language.

留言 (0)

沒有登入
gif