Data Scaling: Implications for Climate Action and Governance in the UK.

Local actors have growing prominence in climate governance but key capacities and powers remain with national policymakers. Coordination between national and local climate action is therefore of increasing importance. Underappreciated in existing academic and policy literature, coordination between actors at different scales can be affected not only by politics and institutional arrangements, but also by methods of data analysis. Exploring two datasets of GHG emissions by local area in England-one of consumption-based emissions and the other of territorial emissions-this paper shows the potential for a data scaling problem known as the modifiable areal unit problem and its possible consequences for the efficacy and equity implications of climate action. While this analysis is conceptual and does not identify specific instances of the modifiable areal unit problem or its consequences, it calls attention to methods of data analysis as possible contributors to climate governance challenges. Among other areas, future analysis is needed to explore how data scaling and other aspects of data processing and analysis may affect our understanding of non-state actors' contribution to climate action.

Understanding individual health-related social needs in the context of area-level social determinants of health: The case for granularity.

Introduction: Screening for health-related social needs (HRSNs) within health systems is a widely accepted recommendation, however challenging to implement. Aggregate area-level metrics of social determinants of health (SDoH) are easily accessible and have been used as proxies in the interim. However, gaps remain in our understanding of the relationships between these measurement methodologies. This study assesses the relationships between three area-level SDoH measures, Area Deprivation Index (ADI), Social Deprivation Index (SDI) and Social Vulnerability Index (SVI), and individual HRSNs among patients within one large urban health system. Methods: Patients screened for HRSNs between 2018 and 2019 (N = 45,312) were included in the analysis. Multivariable logistic regression models assessed the association between area-level SDoH scores and individual HRSNs. Bivariate choropleth maps displayed the intersection of area-level SDoH and individual HRSNs, and the sensitivity, specificity, and positive and negative predictive values of the three area-level metrics were assessed in relation to individual HRSNs. Results: The SDI and SVI were significantly associated with HRSNs in areas with high SDoH scores, with strong specificity and positive predictive values (∼83% and ∼78%) but poor sensitivity and negative predictive values (∼54% and 62%). The strength of these associations and predictive values was poor in areas with low SDoH scores. Conclusions: While limitations exist in utilizing area-level SDoH metrics as proxies for individual social risk, understanding where and how these data can be useful in combination is critical both for meeting the immediate needs of individuals and for strengthening the advocacy platform needed for resource allocation across communities.

The relationship between ambient air pollution and life expectancy - an ecological fallacy revisted.

INTRODUCTION: The results of multiple studies indicate the negative impact of exposure to air pollution on human life expectancy. Epidemiological evidence on this relation is in large proportion provided by ecological studies, what causes interpretation difficulties. Poland is a country characterized by large territorial differences in ambient air pollution and in life expectancy. This promotes analyses of the mentioned relationship based on ecological model. AIM: The aim of the study was to analyse the results of a simple ecological study concerning the relationship between life expectancy and air pollution with the focus on the difficulties in interpretation of the results. MATERIAL AND METHODS: The study used a simple ecological study approach. We used the official data on sex-specific life expectancy for year 2018 and annual average ambient air concentrations of PM2.5 in years 2010-2018 for 10 large metropolitan areas in Poland. The data was used as a marker of long-term air pollution levels in particular areas. Associations between life expectancy and air pollution levels were assessed using Spearman correlation analysis. RESULTS: The analysis concerning exposure to air pollution and life-expectancy in several Polish large agglomerations did not show statistically significant associations. CONCLUSIONS: Our ecological study did not show statistically significant associations between life expectancy and ambient air pollution levels measured by means of PM2.5 concentrations. Ecological nature of the population's exposure marker, without considering many important factors influencing length of life, may explain the negative results of the correlation analysis. This is an example of so called ecological fallacy, affecting the used model of epidemiological study.

ADHD and reification: Four ways a psychiatric construct is portrayed as a disease.

Introduction: The descriptive classification Attention-Deficit/Hyperactivity Disorder (ADHD) is often mistaken for a disease entity that explains the causes of inattentive and hyperactive behaviors, rather than merely describing the existence of such behaviors. The present study examines discourse on ADHD to analyze how authors passively and actively contribute to reification-a fallacy in which a concept is represented as a thing existing on its own. Methods: Critical Discourse Analysis and Qualitative Content Analysis of academic textbooks, scientific articles, websites and videos were used to analyze how ADHD is reified. Results: The analyses reveal four ways in which inattentive and restless behaviors are presented as an entity by means of the ADHD classification: language choice, logical fallacies, genetic reductionism, and textual silence. First, language choice, such as medical jargon and metaphors aid in representing ADHD as a disease entity. Second, several logical fallacies do the same, including the relatively unknown "ecological fallacy" that refers to the erroneous belief that average group findings, such as average brain size of groups of those with an ADHD classification, can be applied on an individual level. Third, genetic reductionism is often achieved by overstating the results of twin studies and being silent about the disappointing molecular genetic research. Such textual silence is the last identified mechanism of reification and includes instances in which societal factors that affect the ADHD construct are often omitted from texts, thereby obscuring the extent to which ADHD is a limited heuristic. Discussion: It is essential that discourse communities do not repeat these four ways of reifying behavior and social relations into an alleged entity with the acronym ADHD. The errors and habits of writing may be epistemologically violent by influencing how laypeople and professionals see children and ultimately how children may come to see themselves in a negative way. Beyond that, if the institutional world shaped to help children is based on misguided assumptions, it may cause them harm and help perpetuate the misguided narrative. To counter the dominant, reifying and medicalizing view, guidelines such as the recently published "Dutch ADHD Psychoeducation Guidelines" might be helpful.

Estudios ecológicos: herramienta clave para la salud pública / Ecológica! studies: a key tool for public health

RESUMEN Los estudios ecológicos (EE) frecuentemente son analizados con actitud escéptica por investigadores, por considerar que las asociaciones obtenidas en estudios de tipo individual presentan mayor nivel de evidencia científica que las referidas en los estudios poblacionales. Además, se argumenta que solo son utilizados para generar o probar hipótesis etiológicas. Sin embargo, los EE constituyen una alternativa útil para el abordaje de determinantes contextuales, ya que las variables agrupadas proporcionan información sobre atributos no captados por las variables individuales. Esta particularidad facilita la incorporación de los EE en estudios multinivel o contextuales de gran utilidad en la salud pública. Sobre estos fundamentos, se sustenta el propósito del presente ensayo en el cual se señalan los desafíos y las oportunidades presentes y futuras de los EE como herramienta clave para el cumplimiento de las funciones de la salud pública.

ABSTRACT Ecological studies (ES) are frequently analyzed with a skeptical attitude by researchers, who consider that the associations obtained in individual-level studies present a higher grade of scientific evidence than those referred to in population studies. Furthermore, it is argued that ES are only used to generate or test etiological hypotheses. However, ES are a useful alternative for addressing contextual determinants since grouped variables provide information on attributes that are not captured by individual variables. This particularity facilitates the incorporation of ES in multilevel or contextual studies of great utility in public health. The purpose of this paper is based on these foundations, in which the present and future challenges and opportunities of ES as a key tool for the fulfillment of public health functions are pointed out.

Comparing the use of aggregate data and various methods of integrating individual patient data to network meta-analysis and its application to first-line ART.

BACKGROUND: The 2018 World Health Organization HIV guidelines were based on the results of a network meta-analysis (NMA) of published trials. This study employed individual patient-level data (IPD) and aggregate data (AgD) and meta-regression methods to assess the evidence supporting the WHO recommendations and whether they needed any refinements. METHODS: Access to IPD from three trials was granted through ClinicalStudyDataRequest.com (CSDR). Seven modelling approaches were applied and compared: 1) Unadjusted AgD network meta-analysis (NMA) - the original analysis; 2) AgD-NMA with meta-regression; 3) Two-stage IPD-AgD NMA; 4) Unadjusted one-stage IPD-AgD NMA; 5) One-stage IPD-AgD NMA with meta-regression (one-stage approach); 6) Two-stage IPD-AgD NMA with empirical-priors (empirical-priors approach); 7) Hierarchical meta-regression IPD-AgD NMA (HMR approach). The first two were the models used previously. Models were compared with respect to effect estimates, changes in the effect estimates, coefficient estimates, DIC and model fit, rankings and between-study heterogeneity. RESULTS: IPD were available for 2160 patients, representing 6.5% of the evidence base and 3 of 24 edges. The aspect of the model affected by the choice of modeling appeared to differ across outcomes. HMR consistently generated larger intervals, often with credible intervals (CrI) containing the null value. Discontinuations due to adverse events and viral suppression at 96 weeks were the only two outcomes for which the unadjusted AgD NMA would not be selected. For the first, the selected model shifted the principal comparison of interest from an odds ratio of 0.28 (95% CrI: 10.17, 0.44) to 0.37 (95% CrI: 0.23, 0.58). Throughout all outcomes, the regression estimates differed substantially between AgD and IPD methods, with the latter being more often larger in magnitude and statistically significant. CONCLUSIONS: Overall, the use of IPD often impacted the coefficient estimates, but not sufficiently as to necessitate altering the final recommendations of the 2018 WHO Guidelines. Future work should examine the features of a network where adjustments will have an impact, such as how much IPD is required in a given size of network.

Be careful with ecological associations.

Ecological studies are observational studies commonly used in public health research. The main characteristic of this study design is that the statistical analysis is based on pooled (i.e., aggregated) rather than on individual data. Thus, patient-level information such as age, gender, income and disease condition are not considered as individual characteristics but as mean values or frequencies, calculated at country or community level. Ecological studies can be used to compare the aggregated prevalence and incidence data of a given condition across different geographical areas, to assess time-related trends of the frequency of a pre-defined disease/condition, to identify factors explaining changes in health indicators over time in specific populations, to discriminate genetic from environmental causes of geographical variation in disease, or to investigate the relationship between a population-level exposure and a specific disease or condition. The major pitfall in ecological studies is the ecological fallacy, a bias which occurs when conclusions about individuals are erroneously deduced from results about the group to which those individuals belong. In this paper, by using a series of examples, we provide a general explanation of the ecological studies and provide some useful elements to recognize or suspect ecological fallacy in this type of studies.

Many continuous variables should be analyzed using the relative scale: a case study of ß2-agonists for preventing exercise-induced bronchoconstriction.

BACKGROUND: The relative scale adjusts for baseline variability and therefore may lead to findings that can be generalized more widely. It is routinely used for the analysis of binary outcomes but only rarely for continuous outcomes. Our objective was to compare relative vs absolute scale pooled outcomes using data from a recently published Cochrane systematic review that reported only absolute effects of inhaled ß2-agonists on exercise-induced decline in forced-expiratory volumes in 1 s (FEV1). METHODS: From the Cochrane review, we selected placebo-controlled cross-over studies that reported individual participant data (IPD). Reversal in FEV1 decline after exercise was modeled as a mean uniform percentage point (pp) change (absolute effect) or average percent change (relative effect) using either intercept-only or slope-only, respectively, linear mixed-effect models. We also calculated the pooled relative effect estimates using standard random-effects, inverse-variance-weighting meta-analysis using study-level mean effects. RESULTS: Fourteen studies with 187 participants were identified for the IPD analysis. On the absolute scale, ß2-agonists decreased the exercise-induced FEV1 decline by 28 pp., and on the relative scale, they decreased the FEV1 decline by 90%. The fit of the statistical model was significantly better with the relative 90% estimate compared with the absolute 28 pp. estimate. Furthermore, the median residuals (5.8 vs. 10.8 pp) were substantially smaller in the relative effect model than in the absolute effect model. Using standard study-level meta-analysis of the same 14 studies, ß2-agonists reduced exercise-induced FEV1 decline on the relative scale by a similar amount: 83% or 90%, depending on the method of calculating the relative effect. CONCLUSIONS: Compared with the absolute scale, the relative scale captures more effectively the variation in the effects of ß2-agonists on exercise-induced FEV1-declines. The absolute scale has been used in the analysis of FEV1 changes and may have led to sub-optimal statistical analysis in some cases. The choice between the absolute and relative scale should be determined based on biological reasoning and empirical testing to identify the scale that leads to lower heterogeneity.

Ecological Research for Studies of Violence: A Methodological Guide.

Ecological research is important to the study of violence in communities. The phrases "ecological research" and "ecologic study" describe those research studies that use grouped or geographic units of analysis, such as zip codes, cities, or states. This type of research allows for the investigation of group-level effects and can be inexpensive and relatively quick to conduct if the researcher uses existing data. And, importantly, ecological studies are an efficient means for hypothesis generation prior to, and can be used to justify, costlier individual-level studies. Ecological research designs may be employed to study violence outcomes when the research question is at the population level, either for theoretical reasons, or when an exposure or intervention is at the population level, or when individual-level studies are not feasible; however, ecological research results must not be used to make individual-level inferences. This article will discuss reasons to conduct ecological-level research, guidelines for choosing the ecological unit of analysis, frequently used research designs, common limitations of ecological research, including the ecological fallacy, and issues to consider when using existing data.

Tropical Tree Branch-Leaf Nutrient Scaling Relationships Vary With Sampling Location.

Bivariate relationships between plant tissue nutrient concentration have largely been studied across broad environmental scales regardless of their covariation with soil and climate. Comparing leaf and branch wood concentrations of C, Ca, K, Mg, N, Na, and P for trees growing in tropical forests in Amazonia and Australia we found that the concentrations of most elements varied with sampling location, but with foliar and branch woody tissues varying from site to site in different ways. Using a Mixed Effect Model (MEM) approach it was further found that relationships between branch and leaf concentrations within individual plots differed in terms of both slope and/or significance to the ordinary least squares (OLS) estimates for most elements. Specifically, using MEM we found that within plots only K and Mg were correlated across organs, but with the K cross-organ intercept estimates varying significantly between sites. MEM analyses further showed that within-plot wood density variations were also negatively related to wood K and Na, suggesting a potentially important role for these cations in water transport and/or storage in woody tissues. The OLS method could not detect significant correlations in any of the above cases. By contrast, although Ca, N, and P leaf and wood tissue concentrations showed similar patterns when individual elements were compared across sites, MEM analyses suggested no consistent association within sites. Thus, for all these three elements, strong within-tree scaling relationships were inferred when data were analyzed across sites using OLS, even though there was no relationship within individual sites. Thus (as for Ca, N, and P) not only can a pooling of data across sites result in trait (co)variations attributable to the environment potentially being incorrectly attributed solely to the species and/or individual (the so-called "ecological fallacy"), but in some cases (as was found here for K and Na) the opposite can also sometimes occur with significant within-site covariations being obscured by large site-site variations. We refer to the latter phenomenon as "environmental obfuscation."

Commentary on "Predictors of Seclusion or Restraint Use Within Residential Treatment Centers for Children and Adolescents" by Green-Hennessy and Hennessy.

In their 2015 study, Sharon Green-Hennessy and Kevin D. Hennessy addressed an important gap in the literature on seclusion and restraint use in child and adolescent residential treatment centers (RTCs). Their analysis revealed that several facility-level characteristics-such as facility size and ownership-predicted the use of seclusion/restraint in child/adolescent RTCs. The authors also examined patient demographic variables that were significant predictors of seclusion/restraint in prior research on individual patients within facilities. However, Green-Hennessy and Hennessy did not find any relationship between these demographic variables and seclusion/restraint. In this commentary I argue that the null relationship between patient demographics and seclusion/restraint was a result of an ecological fallacy. Rather than attempting to use aggregate patient data to infer individual-level processes, this patient data should be used to study aggregate effects. I demonstrate that by re-conceptualizing these patient demographics as indicative of facility characteristics, rather than patient characteristics, new information can be gleaned about the types of facilities that use seclusion/restraint. The arguments presented here have broader implications for future research in this field that relies on aggregate patient data.

Learning to drive: A reconceptualization.

Drivers' population-level crash rates incrementally decrease following licensure, which has led to the implicit assumption that an individual driver's crash risk also decreases incrementally after licensure as they accrue experience. However, in the aggregate data an incremental decrease in crash rate can reflect both incremental reductions in crash risk within individuals and an incremental increase in the proportion of drivers who have experienced an abrupt decrease in crash risk. Therefore, while it is true to say that the population of drivers' crash risk reduces in the months following licensure, it is not necessarily true to say that a driver's crash risk reduces in the months following licensure; that is, it cannot be assumed that individual-level changes in crash risk mirror the population-level changes in crash rates. In statistics, this is known as an ecological fallacy and in formal logic it is known as the fallacy of division, a type of category error. Using computational cognitive modeling methods we demonstrate that aggregating individual-level abrupt decreases in crash risk (i.e., non-incremental change trajectories) accurately fits population-level crash rate data from over 1 million adolescents and uniquely accounts for effects of two interventions found to reduce police-reported MVCs. Thus, we demonstrate that (1) a power-law artifact is readily observable in newly licensed drivers' aggregate crash data, which is not necessarily indicative of individual-level change processes, (2) interventions can alter crash risk trajectories by inducing immediate phase changes in crash risk into a lower risk stratum, or increasing the probability of such a change, (3) a phase transition model provides a stronger and more parsimonious account of the existing data than an incremental-accrual model.

Estimation of the Ecological Fallacy in the Geographical Analysis of the Association of Socio-Economic Deprivation and Cancer Incidence.

Ecological deprivation indices at the level of spatial units are often used to measure and monitor inequalities in health despite the possibility of ecological fallacy. For the purpose of this study, the European Deprivation Index (EDI) was used, which is based on Townsend theorization of relative deprivation. The Slovenian version of EDI (SI-EDI) at the aggregated level (SI-EDI-A) was calculated to the level of the national assembly polling stations. The SI-EDI was also calculated at the individual level (SI-EDI-I) by the method that represents a methodological innovation. The degree of ecological fallacy was estimated with the Receiver Operating Characteristics (ROC) curves. By calculating the area under the ROC curve, the ecological fallacy was evaluated numerically. Agreement between measuring deprivation with SI-EDI-A and SI-EDI-I was analysed by graphical methods and formal testing. The association of the socio-economic status and the cancer risk was analysed in all first cancer cases diagnosed in Slovenia at age 16 and older in the period 2011⁻2013. Analysis was done for each level separately, for SI-EDI-I and for SI-EDI-A. The Poisson regression model was implemented in both settings but adapted specifically for aggregated and individual data. The study clearly shows that ecological fallacy is unavoidable. However, although the association of cancer incidence and socio-economic deprivation at individual and aggregated levels was not the same for all cancer sites, the results were very similar for the majority of investigated cancer sites and especially for cancers associated with unhealthy lifestyles. The results confirm the assumptions from authors' previous research that using the level of the national assembly polling stations would be the acceptable way to aggregate data when explaining inequalities in health in Slovenia in ecological studies.

Ecological fallacy as a novel risk factor for poor translation in neuroscience research: A systematic review and simulation study.

BACKGROUND: Translational neuroscience is largely concerned with establishing causal links between biological processes and functional outcomes. Exciting new methods have emerged and top-tier biomedical journals are placing increasingly high demand for experiments that link outcomes. One pitfall to making these connections is the "ecological fallacy"-establishing a relationship between outcomes based on aggregate (averaged) results (a distinct issue from correlation vs causation). METHODS: To showcase the ecological fallacy, we first used simulated data to define and demonstrate the problem. Next, we performed a systematic review to determine the prevalence of the fallacy in top-tier biomedical journals (Science, Nature Medicine, Neuron, Nature, Nature Neuroscience, Cell). Based on our own research interests and specializations, we specifically focused on recent publications in the area of spinal cord injury and regenerative medicine. RESULTS: Of the articles reviewed which examined a relationship between central nervous system regeneration and a behavioural outcome, 100% (21/21) were subject to possible ecological fallacy. CONCLUSIONS: Ecological fallacy is highly prevalent in neuroscience research and could partially account for translation failures in this field. Reporting guidelines for in vivo experiments should include subject-level correlation analyses for the primary outcomes.

Lack of group-to-individual generalizability is a threat to human subjects research.

Only for ergodic processes will inferences based on group-level data generalize to individual experience or behavior. Because human social and psychological processes typically have an individually variable and time-varying nature, they are unlikely to be ergodic. In this paper, six studies with a repeated-measure design were used for symmetric comparisons of interindividual and intraindividual variation. Our results delineate the potential scope and impact of nonergodic data in human subjects research. Analyses across six samples (with 87-94 participants and an equal number of assessments per participant) showed some degree of agreement in central tendency estimates (mean) between groups and individuals across constructs and data collection paradigms. However, the variance around the expected value was two to four times larger within individuals than within groups. This suggests that literatures in social and medical sciences may overestimate the accuracy of aggregated statistical estimates. This observation could have serious consequences for how we understand the consistency between group and individual correlations, and the generalizability of conclusions between domains. Researchers should explicitly test for equivalence of processes at the individual and group level across the social and medical sciences.

Resolving the vulnerability paradox in the cross-national prevalence of posttraumatic stress disorder.

Socioeconomically disadvantaged individuals are at heightened risk for developing posttraumatic stress disorder (PTSD) following exposure to trauma. Yet a study of cross-national lifetime prevalence rates of PTSD revealed that countries scoring high on an index reflecting cultural and socioeconomic disadvantage exhibited lower rates of PTSD in response to trauma, evincing what the authors called "a vulnerability paradox in the cross-national prevalence of post-traumatic stress disorder" Dückers, Alisic, & Brewin (2016a, p. 300). Drawing on classic studies in sociology and political science concerning the ecological fallacy, the author suggests ways to resolve the striking paradox discovered by Dückers et al.

Determinants of health care costs in the senior elderly: age, comorbidity, impairment, or proximity to death?

Ageing is assumed to be accompanied by greater health care expenditures but the association is also viewed as a 'red herring'. This study aimed to evaluate whether age is associated with health care costs in the senior elderly, using electronic health records for 98,220 participants aged 80 years and over registered with the UK Clinical Practice Research Datalink and linked Hospital Episode Statistics (2010-2014). Annual costs of health care utilization were estimated from a two-part model; multiple fractional polynomial models were employed to evaluate the non-linear association of age with predicted health care costs while also controlling for comorbidities, impairments, and death proximity. Annual health care costs increased from 80 years (£2972 in men, £2603 in women) to 97 (men; £4721) or 98 years (women; £3963), before declining. Costs were significantly elevated in the last year of life but this effect declined with age, from £10,027 in younger octogenarians to £7021 in centenarians. This decline was steeper in participants with comorbidities or impairments; £14,500 for 80-84-year-olds and £6752 for centenarians with 7+ impairments. At other times, comorbidity and impairments, not age, were main drivers of costs. We conclude that comorbidities, impairments, and proximity to death are key mediators of age-related increases in health care costs. While the costs of comorbidity among survivors are not generally associated with age, additional costs in the last year of life decline with age.