Using genetics to understand the relationship between young people’s health and educational outcomes

Amanda Hughes, Kaitlin H. Wade, Matt Dickson, Frances Rice, Alisha Davies, Neil M. Davies & Laura D. Howe

Follow Amanda, Kaitlin, Matt, Alisha, Neil and Laura on twitter

Young people with health problems tend to do less well in school than other students, but it has never been clear why. One explanation is that health problems directly damage educational outcomes. In that case, policymakers aiming to raise educational standards might want to focus first on health as a means of improving attainment.

But are there other explanations? What if falling behind in school can affect health, for instance causing depression? Also, many health problems are more common among children from less advantaged backgrounds – for example, from families with fewer financial resources, or whose parents are themselves unwell. These children also tend to do less well in school, for reasons that may have nothing to do with their own health. How do we know if their health, or their circumstances, are affecting attainment?

It is also unclear if health matters equally for education at all points in development, or particularly in certain school years. Establishing how much health does impact learning, when, and through which mechanisms, would better equip policymakers to improve educational outcomes.

Photo by Edvin Johansson on Unsplash

Using genetic data helps us understand causality

Genetic data can help us answer these questions. Crucially, experiences like family financial difficulties, which might influence both a young person’s health and their learning, cannot change their genes. So, if young people genetically inclined to have asthma are more absent from school, or do less well in their GCSEs, that would strongly suggest an impact of asthma itself. Similarly, while falling behind in school might well trigger depression, it cannot change a person’s genetic propensity for depression. So, a connection between genetic propensity for depression and worse educational outcomes supports an impact of depression itself. This approach, of harnessing genetic information to better understand causal processes, is known as Mendelian randomization.

To find out more, we investigated links between

  • health conditions in childhood and adolescence
  • school absence in years 10 & 11
  • and GCSE results.

We used data from 6113 children born in the Bristol area in 1991-1992. All were participants of the Avon Longitudinal Study of Parents and Children (ALSPAC), also known as Children of the 90s. We focused on six different aspects of health: asthma, migraines, body mass index (BMI), and symptoms of depression, of attention-deficit hyperactivity disorder (ADHD), and of autism spectrum disorder (ASD). These conditions, though diverse, have two important things in common: they affect substantial numbers of young people, and they are at least in part influenced by genetics.

Alongside questionnaire data and education records, we also analysed genetic information from participants’ blood samples. From this information, we were able to calculate for each young person a summary score of genetic propensity for experiencing migraines, ADHD, depression, ASD, and for having a higher BMI.

We used these scores to predict the health conditions, rather than relying just on reports from questionnaire data. In this way, we avoided bias due to the impact of the young people’s circumstances, or of their education on their health rather than vice versa.

Even a small increase in school absence predicted worse GCSEs.

We found that, for each extra day per year of school missed in year 10 or 11, a child’s total GCSE points from their best 8 subjects was a bit less than half (0.43) of a grade lower. Higher BMI was related to increased school absence & lower GCSE grades.

Using the genetic approach, we found that young people genetically predisposed towards a higher BMI were more often absent from school, and they did less well in their GCSEs. A standard-deviation increase* in BMI corresponded to 9% more school absence, and GCSEs around 1/3 grade lower in every subject. Together, these results indicate that increased school absence may be one mechanism by which being heavier could negatively impact learning. However, in other analyses, we found a substantial part of the BMI-GCSEs link was not explained by school absence. It’s unclear which other mechanisms are at play here, but work by other researchers has suggested that weight-related bullying, and negative effects of being heavier on young people’s self-esteem, could interfere with learning.

*equivalent to the difference between the median (50th percentile) in population and the 84th percentile of the population

Diagram showing the pathways through which higher BMI could lead to lower GCSEs; either through more schools absence aged 14-16, or other processes such as weight-related bullying.
Our results suggest increased school absence may partly explain impact of higher BMI on educational attainment, but that other processes are also involved.

ADHD was related to lower GCSE grades, but not increased school absence.

In line with previous research, young people genetically predisposed to ADHD did less well in their GCSEs.  Interestingly, they did not have increased school absence, suggesting that ADHD’s impact on learning works mostly through other pathways. This is consistent with previous research highlighting the importance of other factors on the academic attainment of children with ADHD, including expectations of the school environment, teacher views and attitudes, and bullying by peers.

We found little evidence for an impact of asthma, migraines, depression or ASD on school absence or GCSE results

Our genetic analyses found little support for a negative impact of asthma, migraines, depression or ASD on educational attainment. However, we know relatively little about the genetic influences on depression and ASD, especially compared to the genetics of BMI, which we understand much better. This makes genetic associations with depression or ASD difficult to detect. So, our results should not be taken as conclusive evidence that these conditions do not affect learning.

What does this mean for students and teachers?

Our findings provide evidence of a detrimental impact of high BMI and of ADHD symptoms on GCSE attainment, which for BMI was partially mediated by school absence. When students sent home during the pandemic eventually return to school, the impact on their learning will have been enormous.  And while all students will have been affected, our results highlight that young people who are heavier, who have ADHD, or are experiencing other health problems, will likely need extra support.

Further reading

Hughes, A., Wade, K.H., Dickson, M. et al. Common health conditions in childhood and adolescence, school absence, and educational attainment: Mendelian randomization study. npj Sci. Learn. 6, 1 (2021). https://doi.org/10.1038/s41539-020-00080-6

A version of this blog was posted on the journal’s blog site on 21 Jan 2021.

Contact the researchers

Amanda Hughes, Senior Research Associate in Epidemiology: amanda.hughes@bristol.ac.uk

Conference time at the MRC Integrative Epidemiology Unit!

Dr Jack Bowden, Programme Leader

Follow Jack on Twitter

 

Every two years my department puts on a conference on the topic of Mendelian Randomization (MR), a field that has been pioneered by researchers in Bristol over the last two decades. After months of planning, including finding a venue, inviting speakers from around the world and arranging the scientific programme, it’s a week and a half to go and we’re almost there!

But what is Mendelian Randomization research all about I hear you ask? Are you sure you want to know? Please read on but understand there is no going back…..

Are you sure you want to know about Mendelian Randomisation?

Have you ever had the feeling that something wasn’t quite right, that you are being controlled in some way by a higher force?

Well, it’s true. We are all in The Matrix. Like it or not, each of us has been recruited into an experiment from the moment we were born. Our genes, which are given to us by our parents at the point of conception, influence every aspect of our lives: how much we eat, sleep, drink, weigh, smoke, study, worry and play. The controlling effect is cleverly very small, and scientists only discovered the pattern by taking measurements across large populations, so as individuals we generally don’t notice. But the effect is real, very real!

How can we fight back?

We cannot escape The Matrix, but we can fight back by extracting knowledge from this unfortunate experiment we find ourselves in and using it for society’s advantage. For example, if we know that our genes predict 1-2% of variation in Low-Density Lipoprotein cholesterol (LDL-c – the ‘bad’ cholesterol) in the population, we can see if genes known to predict LDL-c also predict later life health outcomes in a group of individuals such as an increased risk of heart disease. If they do, then it provides strong evidence that reducing LDL-c will reduce heart disease risk, and we can then take steps to act. This is, in essence, the science of Mendelian randomization. See here for a nice animation of the method by our Unit director, George Davey Smith – our Neo if you like.

An example of the mathematical framework that leads to our analysis (honest)

Mendelian randomization is very much a team effort, involving scientists with expertise across many disciplines. My role, as a statistician and data scientist is to provide the mathematical framework to ensure the analysis is performed in a rigorous and reliable manner.

We start by drawing a diagram that makes explicit the assumptions our analysis rests on. The arrows show which factors influence which. In our case we must assume that a set of genes influence LDL-c, and can only influence heart disease risk through LDL-c. We can then translate this diagram into a system of equations that we apply to our data.

The great thing about Mendelian randomization is that, even when many other factors jointly influence LDL-c and heart disease risk, the Mendelian randomization approach should still work.

Recently, the validity of the Mendelian randomization approach has been called into question due to the problem of pleiotropy. In our example this would be when a gene affects heart disease through a separate unmodelled pathway.

 

An illustration of pleitropy

This can lead to bias in the analysis and therefore misleading results. My research is focused on novel methods that try to overcome the issue of pleiotropy, by detecting and adjusting for its presence in the analysis. For further details please see this video.

The MR Data challenge

At this year’s conference we are organising an MR Data Challenge, to engage conference participants in exploring and developing innovative approaches to Mendelian randomization using a publicly available data set. At a glance, the data comprises information on 150 genes and their association with

  • 118 lipid measurements (LDL cholesterol)
  • 7 health outcomes (including type II diabetes)

Eight research teams have submitted an entry to the competition, to describe how they would analyse the data and the conclusions they would draw. The great thing about these data is that the information on all 118 lipid traits simultaneously assessed to improve the robustness of the Mendelian randomization analysis.

Genetic data can help us understand how to resolve population health issues. Image credit: www.genome.gov

A key aim of the session is to bring together data scientists with experts from the medical world to comment on and debate the results. We will publish all of the computer code online so that anyone can re-run the analyses. In the future, we hope to add further data to this resource and for many new teams to join the party with their own analysis attempt.

Please come and join us at the MR conference in Bristol, 17-19 July, it promises to be epic!

Why are people who stay in school longer less likely to get heart disease?

Alice Carter, PhD researcher at the IEU, outlines the key findings from a paper published in BMJ today.

Follow Alice on twitter

 

Heart disease remains the leading cause of death globally, causing over 17.5 million deaths annually. Whilst death rates from heart disease are decreasing in high income countries, the most socioeconomically deprived individuals remain at the greatest risk of developing heart disease. Socioeconomic causes and the wider determinants of health (including living and working conditions, health care services, housing and a number of other wider factors) and are suggested to be the most important driver of health. Behavioural and lifestyle factors, such as smoking, alcohol consumption, diet and exercise, are the second most important contributor to health and disease.

Why does education matter?

Staying in school for longer has been shown to lead to better lifelong health, including reducing the risk of heart disease (cardiovascular disease) and dementia. We also know that those who stay in school are more likely to adopt healthy behaviours. For example, they are less likely to smoke, but more likely to eat a healthy diet and take part in physical activity. These factors, can in turn, reduce the risk of heart disease, such as by reducing body mass index (BMI) or blood pressure. We wanted to understand if these risk factors (BMI, systolic blood pressure and lifetime smoking behaviour) could explain why those who stay in school for longer are less likely to get heart disease, and how much of this effect they explained.

What did we find?

We found that individually, BMI, systolic blood pressure and smoking behaviour explained up to 18%, 27% and 34% of the effect of education on heart disease respectively. When we looked at all three risk factors together, they explain around 40% of the effect. This means that up to 40% of the effect of staying in school reducing the risk of heart disease can be explained by the fact that those who stay in school tend to lead healthier lives. In this work we looked at four outcomes, coronary heart disease (gradual build-up of fatty deposits in arteries), stroke, myocardial infarction (heart attack) and all subtypes of heart disease combined. For all the outcomes we looked at, we found similar results. Notably, the 40% combined effect is smaller than the amount estimated simply from summing the individual effects together. This suggests there is overlap between the three risk factors in how they cause heart disease.

How did we do this?

In our work, we used a few different methods and data sources to answer our questions.

  • We started by looking at observational data (that is the data self-reported by the participants of the study) in UK Biobank – a large population cohort study of around 500 000 individuals. Of these, almost 220 000 people were eligible to be in our analysis.
  • We looked at how their education affected their risk of four types of heart disease. We then looked at how the intermediate factors, BMI, blood pressure and smoking, could help explain these results.
  • Secondly, we replicated these analyses using two types of  Mendelian randomisation analyses (a form of genetic instrumental variable analysis, see below), firstly in the UK Biobank group and secondly by using summary data from other studies in the area.

Why use genetic data?

Typically, epidemiologists collect data by asking people to report their behaviours, lifestyle characteristics and any diagnoses from a doctor. Alternatively, people in a study may have been to a clinic where their BMI or blood pressure is measured. However, this type of data can lead to inaccuracies in analyses.  This could be because:

  • measures are not reported (or measured) accurately. For example, it can be difficult to get an accurate measure of blood pressure, where it changes throughout the day, and even just going to a clinic can result in higher blood pressure.
  • there may be other variables associated with both the exposure and outcome (confounding). One example of this is suggesting that grey hair causes cancer. Really, age is responsible for i) leading to grey hair and ii) leading to cancer. Without accounting for age, we might suggest a false association exists (see figure 1). In our study using education, this could be ethnicity for example, where it influences both staying in school and risk of heart disease.
  • or an individual with ill health may have been advised to change their lifestyle (reverse causality). For example, an individual with a high BMI may have had a heart attack and have been advised by their doctor to lose weight to avoid having a second heart attack.
Diagram showing a picture of grey hair with an arrow linking to cancer, and a third variable - age - above, which explains both.
Figure 1: Does grey hair really cause cancer?

 

One way to overcome these limitations is to use Mendelian randomisation. This method uses the genetic variation in an individual’s DNA to help understand causal relationships. Every individual has their own unique genetic make-up, which is determined, and fixed, at the point of conception.

We are interested in single changes to the DNA sequence, called single nucleotide polymorphisms (or SNPs). For all of our risk factors of interest (education, BMI, blood pressure and smoking) there are a number of SNPs that contribute towards the observed measures, that are not influenced by factors later in life. This means, Mendelian randomisation estimates are unlikely to be affected by bias such as confounding, reverse causality or measurement error, as we might expect when we rely on observational data. By using these genetic variants, we can improve our understanding of if, or how, a risk factor truly causes an outcome, or whether it might be spurious.

What else might be important?

Although we find BMI, blood pressure and smoking behaviour explain a very large amount of the effect, over 50% of the effect of education on heart disease is still unknown. In some small sensitivity analyses we looked at the role of diet and exercise as intermediate risk factors; however, these risk factors did not contribute anything beyond the three main risk factors we looked at. Other social factors may be involved. For example, education is linked to higher income and lower levels of workplace stress, but these factors may also be related to those we’ve looked at in this analysis.

One further suggestion for what may be responsible is medication prescribing and subsequent adherence (or compliance). For example, individuals with higher education may be more likely to be prescribed statins (cholesterol lowering drugs) compared to someone who left school earlier, but with the same requirement for medication. Subsequently, of those who are prescribed statins for example, perhaps those with higher education are more likely to take them every day, or as prescribed. We have work ongoing to see whether these factors play a role.

What does this mean for policy?

Past policies that increase the duration of compulsory education have improved health and such endeavours must continue. However, intervening directly in education is difficult to achieve without social and political reforms.

Although we did not directly look at the impact of interventions in this area, our work suggests that by intervening on these three risk factors, we could reduce the number of cases of heart disease attributable to lower levels of education. Public health policy typically aims to improve health by preventing disease across the population. However, perhaps a targeted approach is required to reduce the greatest burden of disease.

In order to achieve maximum reductions in heart disease we now need to i) identify what other intermediate factors may be involved and ii) work to understand how effective interventions could be designed to reduce levels of BMI, blood pressure and smoking in those who leave school earlier. Additionally, our work looked at predominantly European populations, therefore replicating analyses on diverse populations will be important to fully understand the population impact.

We hope this work provides a starting point for considering how we could reduce the burden of heart disease in those most at risk, and work to reduce health inequalities.

Read the full paper in the BMJ