GGGareth James Griffith

Gareth James Griffith

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Gareth James Griffith has a total of 125 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    The effects of early years services (Sure Start) in Northern Ireland on social services interventions

    by Meng Le Zhang, Martin Elliott, Gareth Griffith, Sally Holland

    Sure Start is an area-based intervention which offers support for families with children under four living in some of the most deprived areas in Northern Ireland. This study uses administrative data on almost every child born in Northern Ireland between 2006 and 2013 (n = 108,000). We use a quasi-experiment (Regression Discontinuity design) to measure the impact of Sure Start on interactions with social services in the first five years of life. Our results suggest that Sure Start substantially reduces rates of children who are referred to social services and become a ‘child in need’. We find that these effects are stronger for mothers who were 21 or younger at childbirth and in the first three years of life. However, we find no effects for child protection or a child going into out-of-home care. These findings suggest that area-based early years interventions can be effective at reducing statutory interventions with benefits for families and social services. Further research is recommended to take advantage of the administrative data and design-based evaluation methods to explore the impact of Sure Start in later life.

  • Open Access

    Adjusting for Medication Use in GWAS and Its Impact on Mendelian Randomisation Analyses: An Example of Systolic Blood Pressure in UK Biobank

    by Aaron Jun Yi Yap, Aimee L. Hanson, Gareth J. Griffith, Eleanor Sanderson

    Medication use is common in large‐scale population cohorts, and can modify phenotypic traits of interest. This can potentially bias effect estimates in genome‐wide association studies (GWAS) and impact downstream analyses such as Mendelian randomisation (MR). The best approach to account for medication use in GWAS is unclear. In this study, we compared seven different methods of adjusting for antihypertensive use in a systolic blood pressure (SBP) GWAS of 407,960 White British individuals in the UK Biobank. We found that direct adjustments to measured SBP (adding constants, class‐specific constants, censored normal regression) in general yielded a greater number of genome‐wide significant variant associations and unmasked stronger GWAS effect estimates than unadjusted measures of SBP. Adjustment for class‐specific constants showed the greatest difference relative to unadjusted GWAS. Restriction methods which limit the sample to either untreated individuals or age ranges with low levels of antihypertensive use had less power, due to reduced sample sizes. Effect estimates of treated individuals were deflated relative to untreated individuals, demonstrating the importance of medication adjustment. In MR analyses, we found no substantial differences in inverse‐variance weighted (IVW) estimates when using differing exposure GWAS methods in estimation of the effect of SBP on coronary artery disease. Larger variations in IVW estimates were observed for the causal effect of body mass index on SBP across adjustment approaches. This suggests that bias may arise in MR analyses when the exposure included in the estimation affects the probability of treatment. Finally, we demonstrate that medication adjustment can reveal potentially novel genetic loci, offering additional insight into the biology of a trait.

  • Open Access

    Measuring the impact of area-based early years support in Wales on children’s social care: Preliminary findings

    by Meng Le Zhang, Martin Elliott, Gareth Griffith, Sally Holland

    Flying Start is an area-based intervention which offers support for families with children under four living in disadvantaged areas in Wales. This study uses a quasi-experiment (Regression Discontinuity Design) and administrative data on children born in Wales between 2012 and 2017 (n = 126,705) to measure the impact of Flying Start eligibility on children’s social care at age 4/5. We find that Flying Start reduces rates of children with care and support plans and children looked after. These effects are stronger for male infants and mothers who were 21 or younger at childbirth. However, we find no effects on child protection.

  • Open Access

    Evaluating confounding in rare variant genome wide association studies

    by Aimee L. Hanson, Gareth J. Griffith, Si Fang, Neil M. Davies, George Davey Smith, Daniel J. Lawson, Gibran Hemani

    The theorised risk that confounded rare variant associations will emerge from population based genetic studies has not been investigated empirically. Here, we use 306,991 sequenced exomes from the UK Biobank to demonstrate that recent demography is poorly captured by common and rare variant principal components, and accounting for haplotype sharing does not eliminate false-positive rare variant associations with non-heritable spatially structured traits. Through re-analysis of 155 phenotypes in siblings, we show a trend of higher effect estimates bias for non-uniformly distributed traits, suggesting population stratification is most pervasive in these settings. Despite its spatial structure, bias of rare variant associations with height appeared most strongly influenced by assortative mating. We explore the risk of elevated false discovery rates for recent variants private to extended families sharing polygenic liability to extreme phenotypes, as well as through local linkage with common causal variants. Overall, we consider the complex confounding mechanisms that can impact rare variant studies and demonstrate family-based approaches can enable important sensitivity analyses.


  • Open Access

    Interpreting inequalities and causal considerations – reflections on the application of descriptive intersectional MAIHDA

    by Natalie C Bennett, Andrew Bell, Gareth Griffith, George Leckie, Clare Evans

    Intersectional MAIHDA (multilevel analysis of individual heterogeneity and discriminatory accuracy) is a novel method for analysing intersectional inequalities that is increasingly popular in health-related research. The method is characterised by its unusual application of a standard multilevel modelling framework, which allows users to describe inequalities in an outcome across multiple intersecting characteristics simultaneously. It does this by nesting individuals within the relevant combination of intersecting characteristics, or ‘strata’. However, the aims of such analyses are not always clear, and misunderstandings of the fundamental goals of intersectional MAIHDA are evident. Intersectional MAIHDA and its closely related approaches (including multicategorical MAIHDA and Eco-Intersectional Multilevel models) were initially conceived as descriptive methods, yet some recent applications stray into causal imaginings. Care is needed when designing, conducting and interpreting the results of MAIHDA. There are also several notable risks for analyses situated in the grey area between description and causal inference. In this study, we consider three key issues that warrant attention: (1) how to interpret the coefficients of stratum-defining variables, (2) the potential for inadvertent collider adjustment and generalisability bias, illustrated using simulated data, and (3) the inclusion of ‘control’ variables. We provide recommendations for producing more robust descriptive research using MAIHDA.

  • Open Access

    Adjusting for medication use in GWAS and its impact on Mendelian randomization analyses: an example of systolic blood pressure in UK Biobank

    by Aaron Jun Yi Yap, Aimee L Hanson, Gareth J Griffith, Eleanor Sanderson

    Medication use is common in large-scale population cohorts, and can modify phenotypic traits of interest. This can potentially bias effect estimates in genome-wide association studies (GWAS) and impact downstream analyses such as mendelian randomization (MR). The best approach to account for medication use in GWAS is unclear. In this study, we compared seven different methods of adjusting for antihypertensive use in a systolic blood pressure (SBP) GWAS of 407,960 White British individuals in the UK Biobank. We found that direct adjustments to measured SBP (adding constants, class-specific constants, censored normal regression) in general yielded a greater number of genome-wide significant variant associations and unmasked stronger GWAS effect estimates than unadjusted measures of SBP. Adjustment for class-specific constants showed the greatest difference relative to unadjusted GWAS. Restriction methods which limit the sample to either untreated individuals or age ranges with low levels of antihypertensive use had less power, due to reduced sample sizes. Effect estimates of treated individuals were deflated relative to untreated individuals, demonstrating the importance of medication adjustment. In MR analyses, we found no substantial differences in inverse-variance weighted (IVW) estimates when using differing exposure GWAS methods in estimation of the effect of SBP on coronary artery disease. Larger variations in IVW estimates were observed for the causal effect of body mass index on SBP across adjustment approaches. In general, the effects of medication use do not substantially affect overall findings. However, bias may arise in MR analyses when the exposure included in the estimation affects the probability of treatment.

  • Open Access

    Dissecting the effect of long-term exposure to air pollution on risk of dementia in UK Biobank

    by Ensor Rafael Palacios, Chin Yang Shapland, Levi John Wolf, Liv Tybjærg Nordestgaard, Emma Anderson, Chloe Slaney, Dan Bernie, Dann Mitchell, Patrick Gavin Kehoe, Gareth James Griffith, et al.

    Mounting evidence links air pollution to dementia, the most prevalent cause of cognitive impairment in older people. Here we investigated individual and compound effects of particulate matters (PM 10 , PM 2.5 , PM coarse , PM abs ) and nitric oxides (NO 2 , NO) on risk of all-cause dementia, and its most common subtypes, Alzheimer’s disease (AD) and vascular dementia (VAD), using data from UK Biobank. We addressed factors that hinder causal interpretation of associations previously shown in the literature and their translation into clear public health policies. Specifically: 1) spatial confounding by area-level covariates, 2) collinearity among and identification of the most relevant air pollutants, and 3) the time window for pollution exposure. Furthermore, we used chronic obstructive pulmonary disease (COPD) and frequency of oily fish intake in positive and negative control analyses. We found NO 2 to be the strongest risk factor for dementia, especially when considering participants with longer permanence at residential address as proxy for longer periods ( $$\ge 5$$ years) of exposure (all-cause dementia hazard ratio HR=1.06, 1.02-1.11 per 9.86 $$\mu g/m^3$$ interquartile range). There was stronger evidence of an effect on risk for AD than VAD. Positive control analysis did not provide any evidence against causality, although the analyses of spatial confounding and negative control analyses revealed the presence of some residual bias, thus warranting care in the interpretation of the results. Together, our results highlight that targeting air pollution, in particular NO 2 levels, could inform preventive public health policies for dementia.


  • Open Access

    Measuring the impact of Sure Start expansion in Northern Ireland on health and social care: Early Results

    by Meng Le Zhang, Gareth Griffith, Sally Holland, Martin Elliott

    Early years interventions have the potential to prevent later life inequalities. Sure Start in Northern Ireland is a place-based early years health and education programme that offers services to some of the most disadvantaged areas in the country. Families with children under four residing in these areas are eligible to access Sure Start services. This study uses a naturally occurring experiment with a Regression Discontinuity (RD) Design and administrative data to measure the impact of Sure Start on health and social services as well as child and maternal health at birth. We find no evidence of an impact on health outcomes at birth. However, we do find that Sure Start reduces the rate of child social services referrals and assessments for a child in need as well as the number of children who become a ‘child in need’. These reductions are most substantial in the first five years of life with no evidence of a rebound after school entry. We also find some evidence these effects are higher for mothers aged 21 and younger. Due to study limitations, the magnitude of these effects are imprecisely measured and we suggest further avenues for future research.

  • Open Access

    Accounting for place: confounding via geography obscures polygenic evidence on mental health and environmental exposures in the UK Biobank

    by Zoe E. Reed, Tim T. Morris, Oliver S.P. Davis, George Davey Smith, Marcus R. Munafò, Gareth J. Griffith

    Clustering of common causes by geography (confounding via geography) can induce bias if ignored. Analyses using polygenic indicators (PGI) for mental health traits are traditionally adjusted for genetic principal components (PCs), but not local area. In this study, we investigate whether accounting for geography more explicitly helps ameliorate such confounding. Using UK Biobank data (N=209,391 to 293,851), we construct PGIs across a range of p-value thresholds and contrast single-level regression with multilevel Mundlak models for the relationship between genetically predicted mental health and greenspace, whilst adjusting for 25 genetic PCs. Our formulation decomposes single level estimates into within-area, between-area and contextual estimates of mental health PGIs on two greenspace measures. Single-level models for the least filtered PGIs find genetically predicted depression and schizophrenia associate with lower greenspace, while genetically predicted wellbeing associates with higher greenspace. However stricter p-value selection on PGI exposures results in estimates largely attenuating and, for schizophrenia, changing sign entirely. Within-area estimates of less-filtered PGIs align more closely with strictly filtered PGIs. Differences appear driven by between-area effects, with single-level models underperforming particularly in the most urbanised areas, and National Parks. We advocate for investigating the inclusion of local context as a routine sensitivity analysis for genetic epidemiologists working with spatially structured populations.