Poor mental health affects all domains of life, including work, relationships and physical health. Prevalence of mental illness in the UK is increasing, with 1 in 6 people experiencing mental health problems each week. One way of tackling this mental illness epidemic is through prevention. We can achieve this by maintaining and improving people's general mental health and wellbeing, before they reach a clinical diagnosis of mental illness. In this project we are interested in identifying behaviours in ones life that can be modified to improve mental health and wellbeing.

One possible predictor of mental health is our health behaviours. For example, our diet, how much physical activity we do, how much alcohol we consume and whether or not we smoke. While this may seem intuitive, there are many gaps in the research literature. First, it is commonly thought that these behaviours influence our mental health and wellbeing, but little research has been done to assess whether this is a causal effect rather than just a correlation. In order to identify effective intervention targets, it is critically important to ensure that these are causal. We will use complementary methods, allowing us to approach our research question through a number of different lenses. If we can draw consistent results, then we will be able to obtain the strongest and most informative evidence. Second, there is little research that has looked at the nuances of these health behaviours. For example, does the length or intensity of physical activity matter when considering its effect on mental health? We will use the wealth of secondary data available to us to untangle these specific details. Third, these health behaviours do not occur in isolation, therefore, we will investigate how they work together to affect
mental health. And finally, early intervention (i.e. in childhood) is an important aspect of prevention. We will use novel methods to understand how parental health behaviours influence the child's mental health. This can inform us about whether it is better to intervene on the parents or on the child's health behaviours.

The project will explore how being born as part of a twin birth or being born premature affects school readiness (performance on entry assessments) and academic outcomes during the early primary school years (Key Stage 1).

There is evidence that twins have poorer academic outcomes during early adulthood relative to singletons (Tsou, Tsou, Wu, & Liu, 2008). It is possible that this is actually a consequence of the increased probability of twins being born preterm or small for gestational age.
In addition, this finding is not consistently observed; a large scale Danish study found no significant differenes in adolescents’ academic outcomes between twins and singletons (Christensen, Petersen, Skytthe, & Hersking, 2006). We aim to explore whether being born as a twin affects school readiness and early academic outcomes relative to being an only child or having one older non-twin sibling and if so, whether this effect persists after controlling for factors known to affect academic outcomes (e.g. birth weight, maternal education).

The second research question aims to examine whether the birth order of twins affects academic outcomes. There is some evidence that first born twins weigh more than second born twins (Yokoyama et al., 2016). Whilst first born twins can be delivered in much the same way as a singleton, there are many more opportunities for obstetric complications (and therefore perinatal compromise) for second born twins, particularly if delivered vaginally. Intrapartum monitoring of the second twin can be challenging, there may be a malpresentation of the second twin requiring internal cephalic version and breech delivery; if these manouveres fail an emergency caesarean section may be required (Leung 2002, Leung 2004). The analysis will therefore explore whether first born twins perform better academically during the primary school years, and whether this effect differs depending on the mode of delivery. Also of interest is whether any effects are reduced or eliminated once controlling for factors such as birth weight and gestational age.

A final analysis will be completed to explore whether children born preterm have worse academic performace relative to their term born siblings. Pettinger et al. (2019) found that gestational age was a significant predictor of whether a child attained a good level of development on the early years foundation stage profile (EYFSP) (e.g. Pettinger et al., 2019). The EYFSP is a mandatory assessment carried out in all state-funded schools at the end of reception year, and is seen as a measure of ‘school readiness’.
However, it is possible that this effect may be (in part) explained by the home environment. Although Pettinger et al. (2019) controlled for factors such as maternal education and receipt of means tested benefits, there may be other unknown environmental factors that affect school readiness. Comparing preterm children to term born siblings would allow us to control for these factors. There is published evidence from Scandinavian studies using sibling matching, which shows that extreme prematurity (<27 weeks gestation) is associated with increased mortality, autism and low educational attainment (D’Onofrio BM, 2013; Risnes KR, 2016). Our work could add significantly more than this, since this study only reported on ‘years in education’ and ‘failing’ education.

These analyses have been completed on the Born in Bradford data already (approval code: SP335). However, the number of children in each analysis is small, limiting the conclusions that can be drawn. We are therefore seeking to replicate our findings using the ALSPAC database, as this would allow us to strengthen our conclusions if the outcomes are replicated. If possible, we also plan to run a set of analyses on a combined dataset, which would allow us to increase the statistical power.

‘Dual harm’ is a term that has been used to describe the co-occurrence of self-harm and violence to others. Young people engaging in dual harm are likely to have specific clinical and support needs. However, very little is known about their characteristics and how individuals who have dual harmed differ from those who self-harm and those who are violent. Understanding the characteristics of people with dual harm experiences will help design services to meet their needs.
We will use data from the ALSPAC to look at differences between young people reporting self-harm, violence towards others, dual harm and neither behaviours at age 16. We are interested in differences in depression, attitudes towards violence, being exposed to self-harm and violence in the home or among peers, impulsivity, self-esteem, callous-unemotional traits, behavioural and emotional dysregulation and body image. We will then look at young people who report either self-harm or violence at age 16 and go on to report dual harm in young adulthood. We expect that their characteristics will differ from those who do not transition to dual harm.

The benefits of physical activity (PA) for mental health are well established. However, little longitudinal evidence is available in adolescents regarding the association between PA and anxiety and depressive symptoms. Especially problematic is our lack of knowledge about the intensity of PA that is needed to prevent or treat clinically relevant symptoms in this critical period of life. This lack of knowledge can be strongly attributed to the way we currently analyze the intensity information from accelerometry. Data are ‘simplified’ and collapsed into only a few intensities (e.g. light, moderate-to-vigorous), which are then included in statistical models. This causes substantial loss of information and challenges the detection of the relative importance of specific intensities. However, using a larger number of intensities is not possible since traditional models (i.e. multiple linear regression) cannot handle their closed structure and multicollinearity. A novel approach is multivariate pattern analysis (MPA) (specifically: partial least squares regression [PLSR]) which was introduced to PA research by Aadland et al. in 2018. MPA overcomes these shortcomings and can be used to describe the PA spectrum with many intensities (e.g. > 20 variables) while determining most important ones. This study uses MPA to investigate the association between the entire PA intensity spectrum at age 14 years and future anxiety and depressive symptoms throughout adolescence and young adulthood.

Ref: Aadland E, Kvalheim OM, Anderssen SA, Resaland GK, Andersen LB. The multivariate physical activity signature associated with metabolic health in children. Int J Behav Nutr Phys Act. 2018;15(1):77.

Placental size directly correlates with its ability to provide developing feotues with the nutrients needed for growth while in the womb, differing placenta size has been linked with health issues both during pregnancy and for both the child and the mother later in life. In this project we propose to study the effect of placenta size on health outcomes of the child and mother both post pregnancy and throughout their lifespans.

In high-income countries, underweight is largely ignored in health research, despite a clear link with worse health and higher mortality risk. But in the UK, where changes to welfare policy have been accompanied by an explosion in foodbank use, social differences in underweight must be reassessed. This project will be the first comprehensive investigation of present-day inequalities in low body weight in Britain. Considering adults, children and adolescents, we will describe the extent of inequalities, and identify the most vulnerable groups. We will shed light on causes, by comparing inequalities in present-day UK with inequalities in other policy contexts: the UK in the 1990s and 2000s, and Norway from the 1980s to the present day.
Starkly raised risk of underweight was recently found among British adult jobseekers, and higher risk of thinness among younger disadvantaged children. Reported after substantial changes to welfare policy, it is not clear if such inequalities existed before those changes, if similar patterns are seen in other countries, or which other disadvantaged groups are affected. We will find out when these patterns first emerged in the UK, by comparing data from the 1990s, 2000s and 2010s. To see if they only occur in certain policy contexts – strongly suggesting they are avoidable - we will also use data from Norway, which in the same period had different welfare policies and a lower poverty rate. To identify other vulnerable groups, we will look at underweight among adults in low-income employment, and children and adolescents in households affected by unemployment and low-income employment.
The project therefore fills three urgent needs: to describe the extent of inequalities in low body weight, to identify groups of vulnerable adults and children, and to understand the causes of these inequalities. Policymakers currently do not have the knowledge required to consider inequalities across the full body weight range, and this project will fill that gap.

Alcohol use during adolescence is a major public health concern, in particular because the brain is still developing and undergoing considerable structural and functional changes. Identifying risk factors that influence adolescent alcohol use is important to appropriately target prevention programs. One area of research that has received considerable attention is the role cognitive functioning (e.g., working memory and inhibition) play as risk for involvement in adolescent alcohol use. There is also a large literature showing a relationship between substance use and self-harm. It is possible that alcohol may increase the risk of self-harm by lowering inhibitions and impairing working memory, which is central to decision making. There is also evidence to suggest that there may be a bi-directional relationship, as several longitudinal studies have reported an association between adolescent self-harm and alcohol problems in adulthood. Although previous prospective studies have examined this association, they are often limited by the use of small sample size, different alcohol use phenotype (i.e. bingeing vs frequency), different follow-up periods, or lack of control for relevant confounders.

Few studies have examined a large number of placenta from general population samples. There are therefore many unanswered questions about the links between placental abnormalities and maternal and offspring health in pregnancy and beyond. ALSPAC is unique, with some 10000 stored placentas from the original ALSPAC pregnancies in the early 90s. We will generate pathology data on this collection and study placental findings in relations to short term health in mums and their offspring.