Post-Modern Epidemiology: When Methods Meet Matter

Sex ratio and age of onset in AQP4 antibody-associated NMOSD: a review and meta-analysis

BACKGROUND

Aquaporin-4 (AQP4) antibody-associated neuromyelitis optica spectrum disorder (NMOSD) is an antibody-mediated inflammatory disease of the central nervous system. We have undertaken a systematic review and meta-analysis to ascertain the sex ratio and mean age of onset for AQP4 antibody associated NMOSD. We have also explored factors that impact on these demographic data.

METHODS

A systematic search of databases was conducted according to the PRISMA guidelines. Articles reporting sex distribution and age of onset for AQP4 antibody-associated NMSOD were reviewed. An initially inclusive approach involving exploration with regression meta-analysis was followed by an analysis of just AQP4 antibody positive cases.

RESULTS

A total of 528 articles were screened to yield 89 articles covering 19,415 individuals from 88 population samples. The female:male sex ratio was significantly influenced by the proportion of AQP4 antibody positive cases in the samples studied (p < 0.001). For AQP4 antibody-positive cases the overall estimate of the sex ratio was 8.89 (95% CI 7.78-10.15). For paediatric populations the estimate was 5.68 (95% CI 4.01-8.03) and for late-onset cases, it was 5.48 (95% CI 4.10-7.33). The mean age of onset was significantly associated with the mean life expectancy of the population sampled (p < 0.001). The mean age of onset for AQP4 antibody-positive cases in long-lived populations was 41.7 years versus 33.3 years in the remainder.

CONCLUSIONS

The female:male sex ratio and the mean age of onset of AQP4 antibody-associated NMOSD are significantly higher than MS. The sex ratio increases with the proportion of cases that are positive for AQP4 antibodies and the mean age of onset increases with population life expectancy.

Chance, ignorance, and the paradoxes of cancer: Richard Peto on developing preventative strategies under uncertainty

During the early 1980s both cancer biology and epidemiological methods were being transformed. In 1984 the leading cancer epidemiologist Richard Peto - who, in 1981, had published the landmark Causes of Cancer with Richard Doll - wrote a short chapter on "The need for ignorance in cancer research", in which the worlds of epidemiology and speculative Darwinian biology met. His reflections on how evolutionary theory related to cancer have become known as "Peto's paradox", whilst his articulation of "black box epidemiology" provided the logic of subsequent practice in the field. We reprint this sparkling and prescient example of biologically-informed epidemiological theorising at its best in this issue of the European Journal of Epidemiology, together with four commentaries that focus on different aspects of its rich content. Here were provide some contextual background to the 1984 chapter, and our own speculations regarding various paradoxes in cancer epidemiology. We suggest that one reason for the relative lack of progress in indentifying novel modifiable causes of cancer over the last 40 years may reflect such exposures being ubiquitous within environments, and discuss the lessons for epidemiology that would follow from this.

Estimating the causal effect of liability to disease on healthcare costs using Mendelian Randomization

Accurate measurement of the effects of disease status on healthcare costs is important in the pragmatic evaluation of interventions but is complicated by endogeneity bias. Mendelian Randomization, the use of random perturbations in germline genetic variation as instrumental variables, can avoid these limitations. We used a novel Mendelian Randomization analysis to model the causal impact on inpatient hospital costs of liability to six prevalent diseases and health conditions: asthma, eczema, migraine, coronary heart disease, Type 2 diabetes, and depression. We identified genetic variants from replicated genome-wide associations studies and estimated their association with inpatient hospital costs on over 300,000 individuals. There was concordance of findings across varieties of sensitivity analyses, including stratification by sex and methods robust to violations of the exclusion restriction. Results overall were imprecise and we could not rule out large effects of liability to disease on healthcare costs. In particular, genetic liability to coronary heart disease had substantial impacts on costs.

Parental inflammatory bowel disease and autism in children

Evidence linking parental inflammatory bowel disease (IBD) with autism in children is inconclusive. We conducted four complementary studies to investigate associations between parental IBD and autism in children, and elucidated their underlying etiology. Conducting a nationwide population-based cohort study using Swedish registers, we found evidence of associations between parental diagnoses of IBD and autism in children. Polygenic risk score analyses of the Avon Longitudinal Study of Parents and Children suggested associations between maternal genetic liability to IBD and autistic traits in children. Two-sample Mendelian randomization analyses provided evidence of a potential causal effect of genetic liability to IBD, especially ulcerative colitis, on autism. Linkage disequilibrium score regression did not indicate a genetic correlation between IBD and autism. Triangulating evidence from these four complementary approaches, we found evidence of a potential causal link between parental, particularly maternal, IBD and autism in children. Perinatal immune dysregulation, micronutrient malabsorption and anemia may be implicated.

Interpreting Mendelian-randomization estimates of the effects of categorical exposures such as disease status and educational attainment

BACKGROUND

Mendelian randomization has been previously used to estimate the effects of binary and ordinal categorical exposures-e.g. Type 2 diabetes or educational attainment defined by qualification-on outcomes. Binary and categorical phenotypes can be modelled in terms of liability-an underlying latent continuous variable with liability thresholds separating individuals into categories. Genetic variants influence an individual's categorical exposure via their effects on liability, thus Mendelian-randomization analyses with categorical exposures will capture effects of liability that act independently of exposure category.

METHODS AND RESULTS

We discuss how groups in which the categorical exposure is invariant can be used to detect liability effects acting independently of exposure category. For example, associations between an adult educational-attainment polygenic score (PGS) and body mass index measured before the minimum school leaving age (e.g. age 10 years), cannot indicate the effects of years in full-time education on this outcome. Using UK Biobank data, we show that a higher educational-attainment PGS is strongly associated with lower smoking initiation and higher odds of glasses use at age 15 years. These associations were replicated in sibling models. An orthogonal approach using the raising of the school leaving age (ROSLA) policy change found that individuals who chose to remain in education to age 16 years before the reform likely had higher liability to educational attainment than those who were compelled to remain in education to age 16 years after the reform, and had higher income, lower pack-years of smoking, higher odds of glasses use and lower deprivation in adulthood. These results suggest that liability to educational attainment is associated with health and social outcomes independently of years in full-time education.

CONCLUSIONS

Mendelian-randomization studies with non-continuous exposures should be interpreted in terms of liability, which may affect the outcome via changes in exposure category and/or independently.

Advancing the Social Epidemiology Mission of the American Journal of Epidemiology

Social epidemiology is concerned with how social forces influence population health. Rather than focusing on a single disease (as in cancer or cardiovascular epidemiology) or a single type of exposure (e.g., nutritional epidemiology), social epidemiology encompasses all the social and economic determinants of health, both historical and contemporary. These include features of social and physical environments, the network of relationships in a society, and the institutions, politics, policies, norms and cultures that shape all of these forces. This commentary presents the perspective of several editors at the Journal with expertise in social epidemiology. We articulate our thinking to encourage submissions to the Journal that: 1) expand knowledge of emerging and underresearched social determinants of population health; 2) advance new empirical evidence on the determinants of health inequities and solutions to advance health equity; 3) generate evidence to inform the translation of research on social determinants of health into public health impact; 4) contribute to innovation in methods to improve the rigor and relevance of social epidemiology; and 5) encourage critical self-reflection on the direction, challenges, successes, and failures of the field.

Planning for health equity in the crossfire between science and policy

The persistence of health inequalities means that many public health professionals face an ongoing task to develop and update policies to tackle them. However, although the inequalities might be unchanged, the political priorities in the many policy areas involved are changing and the ambition to reduce the health divide is constantly facing strong forces pushing in the opposite direction. Recent proposals to re-think health inequalities need to be treated with caution because they are disconnected from what is needed for policy-making in this area. From our experience of 35 years in developing strategies to tackle health inequalities, we still see many entry points with space for local and national improvements, but it is crucial to ask the right questions. The aim of this Commentary is to present a new framework of eight questions that might provide a helpful structure for the necessary dialogue between researchers and policy-makers. Even if answers are not yet available for all of them, we believe that discussing them for a specific population in a specific political context will be fruitful to inform policy on the ground.

Causal Inference with Genetic Data: Past, Present, and Future

The set of methods discussed in this collection has emerged from the convergence of two scientific fields-genetics and causal inference. In this introduction, we discuss relevant aspects of each field and show how their convergence arises from the natural experiments that genetics offer. We present introductory concepts useful to readers unfamiliar with genetically informed methods for causal inference. We conclude that existing applications and foreseeable developments should ensure that we rapidly reap the rewards of this relatively new field, not only in terms of our understanding of human disease and development, but also in terms of tangible translational applications.

Is genetic liability to ADHD and ASD causally linked to educational attainment?

BACKGROUND

The association patterns of attention deficit hyperactivity disorder (ADHD) and autism spectrum disorder (ASD) with educational attainment (EA) are complex; children with ADHD and ASD are at risk of poor academic outcomes, and parental EA has been associated with risk of ADHD/ASD in the offspring. Little is known on the causal links between ADHD, ASD, EA and the potential contribution of cognitive ability.

METHODS

Using the latest genome-wide association studies (GWAS) summary data on ADHD, ASD and EA, we applied two-sample Mendelian randomization (MR) to assess the effects of genetic liability to ADHD and ASD on EA. Reverse direction analyses were additionally performed. Multivariable MR was performed to estimate any effects independent of cognitive ability.

RESULTS

Genetic liability to ADHD had a negative effect on EA, independently of cognitive ability (MVMRIVW: -1.7 months of education per doubling of genetic liability to ADHD; 95% CI: -2.8 to -0.7), whereas genetic liability to ASD a positive effect (MVMRIVW: 30 days per doubling of the genetic liability to ASD; 95% CI: 2 to 53). Reverse direction analyses suggested that genetic liability to higher EA had an effect on lower risk of ADHD, independently of cognitive ability (MVMRIVWOR: 0.33 per SD increase; 95% CI: 0.26 to 0.43) and increased risk of ASD (MRIVWOR: 1.51 per SD increase; 95% CI: 1.29 to 1.77), which was partly explained by cognitive ability (MVMRIVWOR per SD increase: 1.24; 95%CI: 0.96 to 1.60).

CONCLUSIONS

Genetic liability to ADHD and ASD is likely to affect educational attainment, independently of underlying cognitive ability.

Synthesizing Core Outcome Sets for outcomes research in cohort studies: a systematic review

BACKGROUND

Life course studies are designed to "collect once, use multiple times" for observational and, increasingly, interventional research. Core Outcome Sets (COS) are minimum sets developed for clinical trials by multi-stakeholder consensus methodologies. We aimed to synthesize published COS that might guide outcomes selection for early life cohorts with an interventional focus.

METHODS

We searched PubMed, Medline, COMET, and CROWN for COS published before January 2021 relevant to four life stages (pregnancy, newborns, children <8 years, and parents (adults aged 18-50 years)). We synthesized core outcomes into overarching constructs.

RESULTS

From 46 COS we synthesized 414 core outcomes into 118 constructs. "Quality of life", "adverse events", "medication use", "hospitalization", and "mortality" were consistent across all stages. For pregnancy, common constructs included "preterm birth", "delivery mode", "pre-eclampsia", "gestational weight gain", "gestational diabetes", and "hemorrhage"; for newborns, "birthweight", "small for gestational age", "neurological damage", and "morbidity" and "infection/sepsis"; for pediatrics, "pain", "gastrointestinal morbidity", "growth/weight", "breastfeeding", "feeding problems", "hearing", "neurodevelopmental morbidity", and "social development"; and for adults, "disease burden", "mental health", "neurological function/stroke", and "cardiovascular health/morbidity".

CONCLUSION

This COS synthesis generated outcome constructs that are of high value to stakeholders (participants, health providers, services), relevant to life course research, and could position cohorts for trial capabilities.

IMPACT

We synthesized existing Core Outcome Sets as a transparent methodology that could prioritize outcomes for lifecourse cohorts with an interventional focus. "Quality of life", "adverse events", "medication use", "hospitalization", and "mortality" are important outcomes across pregnancy, newborns, childhood, and early-to-mid-adulthood (the age range relevant to parents). Other common outcomes (such as "birthweight", "cognitive function/ability", "psychological health") are also highly relevant to lifecourse research. This synthesis could assist new early life cohorts to pre-select outcomes that are of high value to stakeholders (participants, health providers, services), are relevant to lifecourse research, and could position them for future trials and interventional capability.

Religiosity and Mental Wellbeing Among Members of Majority and Minority Religions: Findings From Understanding Society: the UK Household Longitudinal Study

It is unclear whether links between religiosity and mental health are found in contexts outside the United States or are causal. We examined differences in mental wellbeing and associations between mental wellbeing and religiosity among the religiously unaffiliated, White and non-White Christians, Muslims of Pakistani, Bangladeshi, and other ethnicities, and other minority ethnoreligious groups. We used 4 waves of Understanding Society: the UK Household Longitudinal Study (2009-2013; n = 50,922). We adjusted for potential confounders (including socioeconomic factors and personality) and for household fixed effects to account for household-level unobserved confounding factors. Compared with those with no religious affiliation, Pakistani and Bangladeshi Muslims and members of other minority religions had worse wellbeing (as measured using the Shortened Warwick-Edinburgh Mental Wellbeing Scale and General Health Questionnaire). Higher subjective importance of religion was associated with lower wellbeing according to the General Health Questionnaire; associations were not found with the Shortened Warwick-Edinburgh Mental Wellbeing Scale. More frequent religious service attendance was associated with higher wellbeing; effect sizes were larger for those with religious affiliations. These associations were only partially attenuated by adjustment for potential confounding factors, including household fixed effects. Religious service attendance and/or its secular alternatives may have a role in improving population-wide mental wellbeing.

What Regional Living Conditions Affect Individual Smoking of Adults in Russia

Objectives: Our study evaluated the impact of a wide range of characteristics of large administrative regions on the individual level of cigarette smoking in the Russian adult population. Methods: The pool of participants included 20,303 individuals aged 25-64 years. We applied 64 characteristics of the 12 Russian regions under study for 2010-2014. Using principal component analysis, we deduced five evidence-based composite indices of the regions. We applied the generalized estimating equation to determine associations between the regional indices and the individual level of smoking. Results: The increased Industrial index in the region is associated with the probability of smoking (odds ratio = 1.15; 95% confidence interval = 1.06-1.24). The other indices show associations with smoking only in separate gender and educational groups. Surprisingly, it was found that the Economic index has no associations with the probability of smoking. Conclusion: We evaluated the key associations of the territorial indices with the individual probability of smoking, as well as the mutual influence between the territorial indices and individual factors.

The use of negative control outcomes in Mendelian randomization to detect potential population stratification

A key assumption of Mendelian randomization (MR) analysis is that there is no association between the genetic variants used as instruments and the outcome other than through the exposure of interest. One way in which this assumption can be violated is through population stratification, which can introduce confounding of the relationship between the genetic variants and the outcome and so induce an association between them. Negative control outcomes are increasingly used to detect unobserved confounding in observational epidemiological studies. Here we consider the use of negative control outcomes in MR studies to detect confounding of the genetic variants and the exposure or outcome. As a negative control outcome in an MR study, we propose the use of phenotypes which are determined before the exposure and outcome but which are likely to be subject to the same confounding as the exposure or outcome of interest. We illustrate our method with a two-sample MR analysis of a preselected set of exposures on self-reported tanning ability and hair colour. Our results show that, of the 33 exposures considered, genome-wide association studies (GWAS) of adiposity and education-related traits are likely to be subject to population stratification that is not controlled for through adjustment, and so any MR study including these traits may be subject to bias that cannot be identified through standard pleiotropy robust methods. Negative control outcomes should therefore be used regularly in MR studies to detect potential population stratification in the data used.

Mapping drug epidemiology futures

Epidemiology is a core discipline generating evidence to inform and drive drug policy. In this essay, we speculate on what the future of drug epidemiology might become. We highlight for attention two areas shaping the future of drug epidemiology: nesting epidemiology within a 'syndemic' and 'relational' approach; and innovating in relation to causal inference in the face of complexity. We argue that shifts towards a more relational approach emphasise contingency, including in relation to how drugs might constitute benefit or harm. This leads us to speculate on a 'positive epidemiology'; one that is configured not merely in relation to harm but also in relation to the potential benefits of drugs in relation to well-being. In responding to the complex challenges of delineating contingent causalities, we emphasise the potential of carefully conducted observational study designs that go beyond statistical associations to test causal inference. We acknowledge that each of these developments we describe - a shift towards more relational approaches which emphasise contingent causation, and methodological innovations in relation to establishing causal inference - can be at odds with the other in how they imagine drug epidemiology futures.

Point: Reconciling Epidemiology's Aspirations and Capabilities

Interpreting the results of epidemiologic studies calls for objectivity and rigorous scrutiny, acknowledging the limitations that temper the applicability of the findings to public health action. Current trends have posed new challenges to balancing goals of scientific objectivity and validity with public health applications. The ongoing tension between epidemiology's aspirations and capability has several sources: the need to overpromise in research proposals, compromising methodological rigor because of public health importance, defending findings in the face of hostile critics, and appealing to core constituencies who have specific expectations from the research.

Correlation without a cause: an epidemiological odyssey

In the 1980s debate intensified over whether there was a protective effect of high-density lipoprotein cholesterol (HDL-C) or an adverse effect of triglycerides on coronary heart disease (CHD) risk. In a 1991 paper reprinted in the IJE we suggested that the high degree of correlation between the two, together with plausible levels of measurement error, made it unlikely that conventional epidemiological approaches could contribute to causal understanding. The consensus that HDL-C was protective, popularly reified in the notion of ‘good cholesterol’, strengthened over subsequent years. Reviewing the biostatistical and epidemiological literature from before and after 1991 we suggest that within the observational epidemiology pantheon only Mendelian randomization studies—that began to appear at the same time as the initial negative randomized controlled trials—made a meaningful contribution. It is sobering to realize that many issues that appear suitable targets for epidemiological investigation are simply refractory to conventional approaches. The discipline should surely revisit this and other high-profile cases of consequential epidemiological failure—such as that with respect to vitamin E supplementation and CHD risk—rather than pass them over in silence.

Mendel's laws, Mendelian randomization and causal inference in observational data: substantive and nomenclatural issues

We respond to criticisms of Mendelian randomization (MR) by Mukamal, Stampfer and Rimm (MSR). MSR consider that MR is receiving too much attention and should be renamed. We explain how MR links to Mendel's laws, the origin of the name and our lack of concern regarding nomenclature. We address MSR's substantive points regarding MR of alcohol and cardiovascular disease, an issue on which they dispute the MR findings. We demonstrate that their strictures with respect to population stratification, confounding, weak instrument bias, pleiotropy and confounding have been addressed, and summarise how the field has advanced in relation to the issues they raise. We agree with MSR that "the hard problem of conducting high-quality, reproducible epidemiology" should be addressed by epidemiologists. However we see more evidence of confrontation of this issue within MR, as opposed to conventional observational epidemiology, within which the same methods that have demonstrably failed in the past are simply rolled out into new areas, leaving their previous failures unexamined.

Fifteen years of epidemiology in BMC Medicine

BMC Medicine was launched in November 2003 as an open access, open peer-reviewed general medical journal that has a broad remit to publish "outstanding and influential research in all areas of clinical practice, translational medicine, medical and health advances, public health, global health, policy, and general topics of interest to the biomedical and sociomedical professional communities". Here, I discuss the last 15 years of epidemiological research published by BMC Medicine, with a specific focus on how this reflects changes occurring in the field of epidemiology over this period; the impact of 'Big Data'; the reinvigoration of debates about causality; and, as we increasingly work across and with many diverse disciplines, the use of the name 'population health science'. Reviewing all publications from the first volume to the end of 2018, I show that most BMC Medicine papers are epidemiological in nature, and the majority of them are applied epidemiology, with few methodological papers. Good research must address important translational questions that should not be driven by the increasing availability of data, but should take appropriate advantage of it. Over the next 15 years it would be good to see more publications that integrate results from several different methods, each with different sources of bias, in a triangulation framework.