Statistical analysis, while at first glance an objective way to extract insights from data, remains at its core a human endeavor. Elements of subjectivity are introduced by the many decisions that... Show moreStatistical analysis, while at first glance an objective way to extract insights from data, remains at its core a human endeavor. Elements of subjectivity are introduced by the many decisions that go into the selection of a statistical method. Such subjectivity may harm the evidentiary value of results from statistical analyses. Standardization of statistical methods decreases the degrees of freedom available to researchers and may thus be seen as a way to increase the objectivity of the analysis. Here, we argue that standardization of methods is not only impossible because statistical methods rely on assumptions that need to be considered on a case-by-case basis but also undesirable because it may block innovation. We propose that the entheseal changes field is better served by standardization of reporting and discuss how reporting guidelines may be developed based on examples from biostatistics. Show less
In this introduction to the special issue, Adaptive Tools for Resilient Bones: Biostatistical Approaches to Past Physical Activity in Osteoarchaeology, we discuss the outcome of the workshop held... Show moreIn this introduction to the special issue, Adaptive Tools for Resilient Bones: Biostatistical Approaches to Past Physical Activity in Osteoarchaeology, we discuss the outcome of the workshop held in Leiden (the Netherlands; November 18–19, 2021). We review statistical approaches to entheseal changes and present a series of new contributions to this field. These research, commentary, and review articles present different statistical approaches to entheseal changes and reflect the current state of research in the field. Show less
Over the last decades, increasingly so in the last years, epidemiological methods have been refined, making it challenging to keep abreast of all methodological developments. The choice of the data... Show moreOver the last decades, increasingly so in the last years, epidemiological methods have been refined, making it challenging to keep abreast of all methodological developments. The choice of the data analytical method directly influences the interpretation and clinical meaning of results of an analysis, yet it is undesirable that technical considerations define the subject of the investigation. Having a deeper understanding of the impact that data analytical decisions can have on the interpretation of numerical results of a study would help to apply analytical tools that are both suitable and appropriate to answer clinical questions. The aim of this thesis was to investigate the impact of choices regarding the design and statistical analysis of a study on the meaning of its numerical results in two sets of case studies in research into causal effects (Part I) and prediction research (Part II). The thesis concludes with a discussion on the role and importance of clinical research questions and estimands. Clearly defining a clinically relevant estimand ensures that data analytical decisions yield meaningful results. Making targeted research questions central to quantitative clinical research can reduce fallacious confidence in (complex) methods and can add to intelligibility of findings. Show less
Large and complex data sets are increasingly available for research in critical care. To analyze these data, researchers use techniques commonly referred to as statistical learning or machine... Show moreLarge and complex data sets are increasingly available for research in critical care. To analyze these data, researchers use techniques commonly referred to as statistical learning or machine learning (ML). The latter is known for large successes in the field of diagnostics, for example, by identification of radiological anomalies. In other research areas, such as clustering and prediction studies, there is more discussion regarding the benefit and efficiency of ML techniques compared with statistical learning. In this viewpoint, we aim to explain commonly used statistical learning and ML techniques and provide guidance for responsible use in the case of clustering and prediction questions in critical care. Clustering studies have been increasingly popular in critical care research, aiming to inform how patients can be characterized, classified, or treated differently. An important challenge for clustering studies is to ensure and assess generalizability. This limits the application of findings in these studies toward individual patients. In the case of predictive questions, there is much discussion as to what algorithm should be used to most accurately predict outcome. Aspects that determine usefulness of ML, compared with statistical techniques, include the volume of the data, the dimensionality of the preferred model, and the extent of missing data. There are areas in which modern ML methods may be preferred. However, efforts should be made to implement statistical frameworks (e.g., for dealing with missing data or measurement error, both omnipresent in clinical data) in ML methods. To conclude, there are important opportunities but also pitfalls to consider when performing clustering or predictive studies with ML techniques. We advocate careful valuation of new data-driven findings. More interaction is needed between the engineer mindset of experts in ML methods, the insight in bias of epidemiologists, and the probabilistic thinking of statisticians to extract as much information and knowledge from data as possible, while avoiding harm. Show less
Timmis, A.; Townsend, N.; Gale, C.P.; Torbica, A.; Lettino, M.; Petersen, S.E.; ... ; Vardas, P. 2020
Aims The 2019 report from the European Society of Cardiology (ESC) Atlas provides a contemporary analysis of cardiovascular disease (CVD) statistics across 56 member countries, with particular... Show moreAims The 2019 report from the European Society of Cardiology (ESC) Atlas provides a contemporary analysis of cardiovascular disease (CVD) statistics across 56 member countries, with particular emphasis on international inequalities in disease burden and healthcare delivery together with estimates of progress towards meeting 2025 World Health Organization (WHO) non-communicable disease targets.Methods and results In this report, contemporary CVD statistics are presented for member countries of the ESC. The statistics are drawn from the ESC Atlas which is a repository of CVD data from a variety of sources including the WHO, the Institute for Health Metrics and Evaluation, and the World Bank. The Atlas also includes novel ESC sponsored data on human and capital infrastructure and cardiovascular healthcare delivery obtained by annual survey of the national societies of ESC member countries. Across ESC member countries, the prevalence of obesity (body mass index >= 30 kg/m(2)) and diabetes has increased two- to three-fold during the last 30years making the WHO 2025 target to halt rises in these risk factors unlikely to be achieved. More encouraging have been variable declines in hypertension, smoking, and alcohol consumption but on current trends only the reduction in smoking from 28% to 21% during the last 20years appears sufficient for the WHO target to be achieved. The median age-standardized prevalence of major risk factors was higher in middle-income compared with high-income ESC member countries for hypertension {23.8% [interquartile range (IQR) 22.5-23.1%] vs. 15.7% (IQR 14.5-21.1%)}, diabetes [7.7% (IQR 7.1-10.1%) vs. 5.6% (IQR 4.8-7.0%)], and among males smoking [43.8% (IQR 37.4-48.0%) vs. 26.0% (IQR 20.9-31.7%)] although among females smoking was less common in middle-income countries [8.7% (IQR 3.0-10.8) vs. 16.7% (IQR 13.9-19.7%)]. There were associated inequalities in disease burden with disability-adjusted life years per 100 000 people due to CVD over three times as high in middle-income [7160 (IQR 5655-8115)] compared with high-income [2235 (IQR 1896-3602)] countries. Cardiovascular disease mortality was also higher in middle-income countries where it accounted for a greater proportion of potential years of life lost compared with high-income countries in both females (43% vs. 28%) and males (39% vs. 28%). Despite the inequalities in disease burden across ESC member countries, survey data from the National Cardiac Societies of the ESC showed that middle-income member countries remain severely under-resourced compared with high-income countries in terms of cardiological person-power and technological infrastructure. Under-resourcing in middle-income countries is associated with a severe procedural deficit compared with high-income countries in terms of coronary intervention, device implantation and cardiac surgical procedures.Conclusion A seemingly inexorable rise in the prevalence of obesity and diabetes currently provides the greatest challenge to achieving further reductions in CVD burden across ESC member countries. Additional challenges are provided by inequalities in disease burden that now require intensification of policy initiatives in order to reduce population risk and prioritize cardiovascular healthcare delivery, particularly in the middle-income countries of the ESC where need is greatest. Show less
Research on terrorism has long been criticized for its inability to overcome enduring methodological issues. These include an overreliance on secondary sources and the associated literature review... Show moreResearch on terrorism has long been criticized for its inability to overcome enduring methodological issues. These include an overreliance on secondary sources and the associated literature review methodology, a scarcity of statistical analyses, a tendency for authors to work alone rather than collaborate with colleagues, and the large number of one-time contributors to the field. However, the reviews that have brought these issues to light describe the field as it developed until 2007. This article investigates to what extent these issues have endured in the 2007–2016 period by constructing a database on all of the articles published in nine leading journals on terrorism (N = 3442). The results show that the use of primary data has increased considerably and is continuing to do so. Scholars have also begun to adapt a wider variety of data-gathering techniques, greatly diminishing the overreliance on literature reviews that was noted from the 1980s through to the early 2000s. These positive changes should not obscure enduring issues. Despite improvements, most scholars continue to work alone and most authors are one-time contributors. Overall, however, the field of terrorism studies appears to have made considerable steps towards addressing long-standing issues. Show less
This thesis aims to evaluate the environmental sustainability of European imports of farmed aquatic food products from Asia, using life cycle assessment (LCA). Farming of Asian tiger prawn,... Show moreThis thesis aims to evaluate the environmental sustainability of European imports of farmed aquatic food products from Asia, using life cycle assessment (LCA). Farming of Asian tiger prawn, whiteleg shrimp, freshwater prawn, tilapia and pangasius catfish in Bangladesh, China, Thailand and Vietnam were chosen as representatives of the Asian aquaculture industry. Initial research revealed large discrepancies among LCA results driven by methodological choices and data sourcing. A protocol for quantifying dispersions around unit process data was therefore developed, characterising inherent uncertainty, spread (variability) and unrepresentativeness as the three major sources driving overall discrepancies. Results, propagated using Monte Carlo simulations, highlighted that the uncertainty related to LCA results could range with over an order of magnitude. For comparative purposes, however, only relative uncertainties are of relevance. Defining a hypothesis and using dependent sampling therefore allowed for several significant conclusions to be identified. Among these were significantly lower environmental impacts of Asian tiger shrimp farming in western Bangladesh, tilapia in Guangdong and pangasius in large-scale farms. Common environmental hot-spots included aqua-feeds, eutrophying effluents from farms, the use of benzalkonium chloride and other chlorine releasing compounds as disinfectants, and extensive use of paddle-wheels on shrimp farms. The research identified discrepancies Show less
According to the minimum description length (MDL) principle, data compression should be taken as the main goal of statistical inference. This stands in sharp contrast to making assumptions about an... Show moreAccording to the minimum description length (MDL) principle, data compression should be taken as the main goal of statistical inference. This stands in sharp contrast to making assumptions about an underlying ``true'' distribution generating the data, as is standard in the traditional frequentist approach to statistics. If the MDL premise of making data compression a fundamental notion can hold its ground, it promises a robust kind of statistics, which does not break down when standard, but hard to verify, assumptions are not completely satisfied. This makes it worthwhile to put data compression to the test, and see whether it really makes sense as a foundation for statistics. A natural starting point are cases where standard MDL methods show suboptimal performance in a traditional frequentist analysis. This thesis analyses two such cases. In the first case it is found that although the standard MDL method fails, data compression still makes sense and actually leads to the solution of the problem. In the second case we discuss a modification of the standard MDL estimator that has been proposed in the literature, which goes against its data compression principles. We also review the basic properties of R_nyi's dissimilarity measure for probability distributions. Show less
Type 2 Diabetes (T2D) is a chronic disease, characterized by hyperglycaemia, caused by decreased insulin secretion by beta-cells and insulin resistance of target tissues of insulin. Several risk... Show moreType 2 Diabetes (T2D) is a chronic disease, characterized by hyperglycaemia, caused by decreased insulin secretion by beta-cells and insulin resistance of target tissues of insulin. Several risk factors are known, like decreased exercise, ageing and western diet. Also genetic variance can alter susceptibility to develop T2D. Until recently only four T2D susceptibility genes were identified (PPARG, KCNJ11, CAPN10 and TCF7L2). However, recent Genome Wide Association Studies (GWAS) have increased this number to at least 20. My thesis describes the search for additional T2D susceptibility genes. For this we used a classical candidate gene approach and we case/control studies from the Netherlands, Scandinavia and the UK. We selected 14 nuclear encoded candidate genes, all regarded essential for mitochondrial protein synthesis and biogenesis. Since mitochondrial function was shown to be associated with T2D, we hypothesized that genes in these pathways are good candidates. Tagging SNPs were selected, which should cover all known common variation (minor allele frequency (MAF) > 0.05) in the candidate genes. These tagging SNPs were measured in our first stage, comprising of the Dutch Hoorn study and significant associations were than taking forward for replication in our replication cohorts in the second stage of this study. However, after second stage genotyping non of the signals remained significant, indicating that the selected candidate genes do not play a major role in T2D susceptibility. Furthermore, nuclear encoded mitochondrial genes were not among the top hits of GWAS, which were made available online after completion of our study. Therefore, we conclude that nuclear encoded mitochondrial genes do not have a major contribution to the development of T2D. Next, we analyzed the association of mitochondrial DNA (mtDNA) content with T2D. First we estimated the heritability in Dutch twins and found a heritability of 35% (19%-48%). Next we analyzed the association with prevalent and incident T2D in a Dutch case control study and two prospective studies from the Netherlands and Finland. However, no associations were observed. Therefore, we conclude that mtDNA content does not play a major role in the development of T2D. Finally, we analyzed four known fasting plasma glucose (FPG) influencing genes (GCK, GCKR, G6PC2 and MTNR1B). In a Dutch population based sample we could replicate the association of these genes with FPG levels (except for GCKR). Furthermore, the combined risk alleles (ranging from 0 to 8 risk alleles) were strongly associated with FPG and HbA1C levels. This risk allele score was also associated with T2D susceptibility and age of diagnosis at T2D. We therefore conclude that the FPG influencing genes have a combined effect on FPG and are associated with T2D susceptibility and age at diagnosis of T2D. In conclusion, we could not find evidence that nuclear encoded mitochondrial proteins and mtDNA content are associated with T2D susceptibility. The four known FPG genes do not only influence FPG levels, but also have a combined effect on T2D susceptibility and age at diagnosis of T2D. Show less