Neither seen nor heard: the evidence gap on the effect of covid-19 on mental health in children
BMJ 2024; 387 doi: https://doi.org/10.1136/bmj-2023-078339 (Published 07 October 2024) Cite this as: BMJ 2024;387:e078339Read the collection: Research priorities for future shocks
- Tamsin Ford, professor1,
- Tamsin Newlove-Delgado, associate professor2,
- Ann Kannuralil Sabu, student3,
- Abigail Russell, senior lecturer2
- 1University of Cambridge, Cambridge, UK
- 2University of Exeter, Exeter, UK
- 3public representative, Cambridge Children’s Hospital Youth Forum, Cambridge, UK
- Correspondence to: T Ford tjf52{at}medschl.cam.ac.uk
The impact of health shocks, such as pandemics and human made and natural disasters, on children and young people is too often under-researched, meaning evidence informed responses are inadequate. The original terms of reference of the UK Covid-19 Inquiry omitted children, for example, as did much of the global pandemic response. Yet data indicate that the pandemic had a disproportionate impact on their mental health.1 This was entirely predictable, given that pandemic restrictions resulted in social isolation and disrupted routines and sleep, as well as financial and academic stress, all of which undermine mental health, in addition to fears about loss of educational opportunities and family members becoming ill (box 1).2
Young person’s perspective, by Ann Kannuralil Sabu
I am 17 and in full time education. Over the past few years, I have collaborated with the Cambridge Children’s Hospital (CCH) youth forum. To elicit views about the effects of the pandemic that needed researching, in August 2023 we sent a survey to about 600 people linked to the CCH youth forum, as well as the linked parents and carers forum, and about 20 third sector organisations that work with children and families in the east of England. We also shared it through social media. We received 38 responses. The following are typical of the comments received.
Parent of a pre-school child: “Being able to socialise at a young age is vital to a child’s development, and he missed out on so much. I feel this is the reason he is much less confident than his brother.”
Young person age 14-17: “Not experiencing the typical teenage experiences and feeling [left] behind.”
Young person age 14-17: “Isolation. Being lonely and unable to socialise.”
Parent: “Not being able to mix with others put a huge strain on families and also impacts on behaviour. Behaviour that seems unnatural became natural for them.”
Responses indicated that many young people experienced diminished self-confidence, anxiety, and loneliness. The cancellation of exams caused anxiety and uncertainty, and prevented some young people from pursuing their ideal futures. Those with better-off or more stable family circumstances felt better able to cope with lockdowns. Children in abusive households and with young caregivers found less respite, posing an extra threat to their mental wellbeing.
I was fortunate: lockdown for me was a break—time for reflection and growth. Despite the challenges and loss of two grandparents, my family grew closer, spending quality time together. In contrast, the pandemic took a clear toll on some young people’s mental health. Some are waiting months after referral by a GP, while their problems are worsening. Children need better access to mental health support, and the government must guarantee timely access to care and clear the backlog. We need more research about why some children struggled during the pandemic and others did not, and how to identify and support the young people who most need help.
RETURN TO TEXTCovid has been described as “a systemic shock to the wider determinants of child health, with impacts on family functioning and income, access to healthcare and education.”3 Protecting and promoting health in the early years is crucial in the prevention of adult ill health, and this is particularly true for mental health: three quarters of adults who struggle with poor mental health experienced their initial difficulties in childhood.4 A child who cannot function for several months pays a developmental price in terms of skills and educational development. Similarly, there can be no health without mental health: prevention and containment of communicable diseases must not neglect mental health, nor widen inequalities.
To better protect and promote the mental health of children, responses to future health shocks must be based on sound evidence. However, the evidence base regarding the impact of the pandemic on children’s mental health is not as robust as that for the adult population. There are fewer researchers, who face greater logistical, ethical, and methodological barriers related to ethical approval, recruitment, and data collection, particularly of younger children, which deters involvement and delays outputs. A living systematic review of the mental health impact of covid-19 screened almost 150 000 abstracts, but only 720 were sufficiently robust, of which 35 studies concerned those under the age of 18.5
Given that sound policy results from strong evidence, this evidence gap poses a substantial problem, particularly as a third of the UK population is under 25 years. Poor mental health in childhood and adolescence leads to negative outcomes in health, occupation, social, and economic outcomes, with substantial costs to society as well as those affected and their families.3 The voices of children and their parents and carers must be central to the selection of research priorities.
Strengthening research into mediating, moderating, and causal pathways to poor mental health is particularly important. This includes ascertaining why some young people do “better than expected” compared with peers with the same mental health condition, to identify targets for prevention and intervention.
What the evidence shows
Polemic headlines about a mental health crisis in children, plus surging demand and long waiting lists for services, pre-date covid-19. Research indicates that children’s mental health was deteriorating in the UK before the pandemic, particularly among teenage girls and young women.6 Furthermore, pre-pandemic studies suggest that more recent generations of children with a mental health condition seemed to experience increased levels of symptoms with greater effect on their ability to function and worse outcomes than their recent predecessors.7
Systematic reviews of longitudinal studies8910 provide robust evidence for increased depression rather than anxiety among young people living through the pandemic and resulting restrictions, although the prevalence of both disorders would be expected to increase in adolescence.4
Some of the current authors conducted a systematic review of 51 studies from 15 countries that assessed children’s mental health using a validated measure before and during the pandemic.8 This review reported a slight deterioration in mental health among children during the pandemic. Only four studies were of high methodological quality, with huge variation in samples, methods, and timing in relation to the pandemic, which may explain the failure to detect any clear pattern of changes. Data on children under 10 years are lacking, and few studies consider neurodevelopmental conditions and conduct and eating disorders.8 This may be related to a decline in academic paediatrics, the more complex ethical and logistical considerations of research with younger children, which requires parental consent, and the need for teacher and parent reported data for under 8 year olds.
A 2023 systematic review found increased symptoms of attention deficit-hyperactivity disorder (ADHD) during the pandemic in people with sub-clinical or clinical levels of difficulties before the pandemic.11 Globally, presentations to clinical services for children with eating disorders have increased.12 Evidence is mixed on suicide rates and suicidal behaviours.1314 Given the time lag in the release of suicide statistics, future investigation of the potential impact of the pandemic is required. The prevalence of self-harm in UK children and adolescents is about 20%, and data beyond the 2023 systematic review is lacking8; this is a priority area for research, but one that often faces challenges in obtaining ethical approval.
Studies consistently show that children facing socioeconomic adversity, and with pre-existing poor mental health or special educational needs, have worse trajectories.15 School closures are particularly difficult for families facing other adversities (box 1), which could be alleviated with group based support.216 Furthermore, one in five children were persistently absent from school (>10% of sessions) in England in the autumn term of 2024, which is approximately double the pre-pandemic levels.32 We urgently need to ensure outcome data collection around innovative practice to evaluate its impact and to develop evidence based interventions to support reintegration into school.
Studies over time and across the age range are essential to determine whether there are developmental periods where exposure to social restrictions was more detrimental.
Increasing prevalence of anxiety, depression, and self-harm in teenage girls and young women were concerns going into the pandemic.617
Children who experienced disrupted transitions, such as starting school or moving from primary or elementary, to secondary or high school, or from school to work or university, may be particularly at risk for mental health conditions; we need to research this. Children who were toddlers, babies, or born during the pandemic had greatly reduced pre-school social exposure. Anecdotal reports suggest increases in delayed language and social skills among children entering school, emphasising the need for follow-up studies.
We should not forget the impact of infection with covid. England’s Children with Long COVID (CLoCk) study identified more symptoms of greater severity among children who had tested positive at all data points compared with those who tested negative.18 In individuals, symptoms declined over time, and no difference in mental health outcomes was found between children who initially tested positive or negative. Vaccination did not seem to prevent development of long covid. Older teenagers, girls, and children with poorer health before covid had worse outcomes. Data linkage to health and education records would help to understand pre-pandemic health and education and in following up these cohorts. The experience of, and treatment of, children with sustained symptoms should be researched.
Challenges for research
The pandemic stimulated increased quantity but not quality of mental health research.519 For example, systematic scientometric research in 2022 found 3692 articles on covid and children’s mental health, but only 23% reported new data.20 Studies excluded from the living systematic review already mentioned were mostly cross sectional, lacked pre-pandemic comparison data, and used convenient samples or bespoke measures of mental health, resulting in risks of bias or inaccuracy.5
Consistent use of metrics is crucial. Even small changes in the wording of questionnaires can substantially affect responses, making them invalid.21 Guidance on consistent measures is important to increase both the rigour and the comparability of research. However, the current guidance was developed with insufficient consultation with key stakeholders and focuses only on anxiety, depression, and impairment, ignoring common and important conditions such as eating, neurodevelopmental, and psychotic disorders.22 Such rigid insistence on particular measures risks stasis in research, and regular review of recommendations by research funders, researchers, and policy makers is essential to avoid stifling psychometric innovations, such as computer adaptive testing.
Covid-19 affected researchers’ personal and professional lives, especially female academics with children, potentially limiting their work.23 Pressure for rapid response meant research was not always led by specialists in young people’s mental health or epidemiology.
Furthermore, in the UK, the process to access administrative data and the NHS national surveys for analysis has become more complex and time consuming, hindering secondary analyses.24 Despite pressure to fill the evidence gap, robust methodology is essential to minimise misleading findings.
Nearly all research has reported mean changes across the entire studied population or sub-groups, even if stratified by different levels of risk, which might mask highly variable trajectories within individuals. An increased focus on changes within individuals over time, and their predictors, would reveal more about factors that determine resilience or poor outcomes.
What does the impact on services show?
Even a slight deterioration in mental health, or change in help seeking on a population level, can result in a marked increase in demand on services.12 To respond, policy makers and service commissioners and providers need to understand the drivers and vulnerable groups that lack access.25
For example, fewer children were referred to specialist services or admitted to mental health units during the first lockdown, when there were also reduced mental health presentations and prescribing for young people in primary care.26 This was rapidly followed by a sharp rise in referrals and crisis presentations, with increasing pressure on all levels of the NHS.1 Research that improves our understanding of the influences on help seeking and access are important to facilitate targeting of limited resources to those most in need.
Recommendations for research
Optimise study design–National and international funders should increase the number of developmental studies that follow children over time. Only longitudinal data can show whether a health shock is causing harm compared with other exposures. Cross sectional surveys, which ask those experiencing a health shock to recall previous health states, are biased, inaccurate, and wasteful to study. Cohort studies are expensive but are an important investment that repays in the breadth and depth of policy relevant data that can be obtained.
Improve data linkage–Longer follow-up and more detailed pre-pandemic knowledge could be achieved by the mandatory linkage of epidemiological data to routinely collected health, social care, and education data, which could drive service improvement across the UK at relatively low cost. It would also provide information about health and education function before the initial study, supporting identification of vulnerable groups. However, we lack funding streams to support this work, and the bureaucracy in linking and accessing the data deters researchers and funders. Initiatives such as University College London’s ECHILD and the UK Longitudinal Linkage Collaboration are welcome.2728 Data linkage should be a condition of funding; researchers need financial support to revisit studies after participants have passed key developmental milestones, such as leaving school or transitioning to adult health services. Researchers also need training and supervision to ensure that these sophisticated techniques are more widely deployed.
Build capacity—Paediatric research generally needs greater collaboration, engagement, and capacity building.29 Global epidemiological capacity building is needed in child mental health, including training and peer mentoring. Work by some of the current authors shows greatly reduced activity in the UK.24 Funders should provide dedicated training opportunities that focus on acquiring epidemiological expertise.
Collaborate internationally— Cross border collaborations can be helpful in studying the impact of different responses to health shocks on different service systems and to study risk and protective factors; they also increase sample size where context, response, and settings are similar. We hope the UK’s re-entry to Horizon Europe, the European Union’s research funding programme, will facilitate and fund international studies. International bodies such as WHO and Unicef should encourage and fund collaborative research.
Use validated measures—Research should use the many validated measures of child and adolescent mental health. If a bespoke measure is used, validity and reliability should be established before any publication using the measure as an outcome. This requires cooperation and good methods of training at every level, from researchers to funders and journal editors. Routine use of validated measures in clinical practice with outcome measurement would facilitate clinical research, which the UK government’s Mental Health Missionhopes to achieve.30
Increase research funding–With 20% of children persistently absent from school in England, and a 24% increase in the number of emerging adults unable to work owing to poor mental health,33 governmentcannot afford not to increase the funding allocated to research the mental health of children. Although a third of the UK population is under 25 years, the health research allocation is only 5% of the budget.3 Furthermore, mental health research, of which only a quarter is focused on children, is seriously underfunded compared with other conditions with a comparable burden of disease: annual spending is £9 (€11; $12) per affected person compared with £24 for cardiovascular disease, £127 for dementia, and £228 for cancer.31 Research into children receives a tiny proportion of this. All funders, including industry, which has tended to exclude under 18s, have a duty to contribute. Targeted funding could encourage inclusion of younger children in studies and ensure a focus across all common mental health conditions, while mandated involvement of researchers with appropriate expertise would boost research capacity and activity.
Proportionate data governance—Research has been hobbled by over-restrictive data governance.24 Strong governance around access to epidemiological and administrative data is essential, but application and data transfer processes need to be streamlined and adequately staffed. Current applications often take months or even years, deterring researchers and funders and depriving policy makers and service providers of essential evidence. Recent examples, like ECHILD, which allows researchers to apply to study linked health and education data, and transfer of the Mental Health of Children and Young People Surveys data from NHS England to the UK Data Archive, with a streamlined process for access, show what discussion and collaboration across stakeholders can achieve.
Key messages
Children and young people’s interests and voices must be represented and respected in prioritisation of research questions to tackle the wide evidence gap
Funding calls and streams should encourage longitudinal studies across the range of mental health conditions affecting children and young people, and also ensure that funded research spans life stages from pre-school to emerging adulthood
Linkage of study data to administrative education and health and social care to extend follow-up should be mandated as part of funding agreements, with streamlined processes to expedite both linkage and access for analysis
Funders, supervisors, and journal editors should reject low quality studies, particularly those with bespoke, unvalidated measures, and cross sectional surveys of convenience samples. The considerable resources saved could support more rigorous and reliable research for policy and planning
Acknowledgments
We thank all the young people, parents, and carers who responded to our survey, and Anna Todd, from Cambridge and Peterborough NHS Foundation Trust, who introduced us to Ann and supported her involvement. All research at the department of psychiatry in the University of Cambridge is supported by the NIHR Cambridge Biomedical Research Centre (BRC-1215-20014) and NIHR Applied Research Centre. TND and AER were supported by the National Institute for Health Research (NIHR) advanced fellowships during the preparation of this paper (NIHR300056 and NIHR300591). The views expressed are those of the authors and not necessarily those of the NIHR or the Department of Health and Social Care.
Footnotes
Contributors and sources Our invitation to submit this analysis arose from our systematic review of this topic (TND, AER, TF) and our involvement in the English Mental Health of Children and Young People Survey series (TF, TND). AKS volunteers with the Cambridge Children’s Hospital Youth Forum and led the informal survey of young people and parents/ carers perspectives distributed via Cambridge Children’s Network, Cambridge Children’s Hospital social media platforms, and the third sector. All authors contributed to conceptualisation and drafting of the manuscript. AER updated the literature review, and TND led the sections describing service impact and public health recommendations. TF coordinated the authors and led the drafting of the rest of the manuscript.
Patient and public involvement: One of the authors, AKS, is a young person in full time education. She was involved in structuring our article, reviewing the responses from young people and parents/carers to our survey questions about their experience of the pandemic and comparing it with her own. She reviewed and refined drafts of this manuscript in collaboration with the other authors.
Competing interests: We have read and understood BMJ policy on declaration of interests and have the following interests to declare: TF’s research groups receives funding for her research advice to Place2Be, a third sector organisation that provides mental health training and support in schools across the UK.
Provenance and peer review: Commissioned; externally peer reviewed.
This article is part of a collection proposed by the Health Foundation. The Health Foundation provided funding for the collection, including open access fees. The BMJ commissioned, peer reviewed, edited, and made the decision to publish this article. Richard Hurley was the lead editor for The BMJ.
This is an Open Access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.