SEARCH · Search PubMed

Results for “causal inference”

Search indexed PubMed citations on genomics, clinical trials, systematic reviews and public health. Explore titles, authors and supplied subject terms, then open the PubMed record.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 163 records · Page 9Linked to original sources

PubMedDate not supplied

[Popper and the problem of induction in epidemiology].

In this article we are discussing a few of the contributions by the Austro-British philosopher Karl R. Popper, one of our most influential contemporary thinkers, whose epistemological and socio-political theories have also penetrated the sphere of epidemiology. We are focusing mainly on the so-called problem of induction. We sustain, in line with Popper, that the scientific method does not use inductive reasoning, but rather hypothetical-deductive reasoning. Although the movement from the data evaluating a hypothesis to a conclusion on the latter goes from the specific to the general, that is, in an inductive direction, the induction does not exist as a reasoning process or inference. That is, there is no method that enables us to infer or to verify hypotheses or theories (we cannot explore all of the possible situations to see whether the theory stands up), or even to render them very probable. Besides, scientists look for highly informative theories, not highly probable ones. What we actually do is to propose a hypothesis as a tentative solution to a problem, to confront the prediction deduced from the hypothesis with actual experience, and evaluate whether the hypothesis is rejected or not by the facts. As theories cannot be verified, we can only accept them if they withstand an attempt to reject them. Consequently, the test of a theory consists of criticism or a serious attempt at falsification, that is, the elimination of error within a theory, in order to reject it if it is false. The objective is, thus, the search for true theories. For this purpose, the scientific method uses a systematic set of methodological (not logical) rules, that is, decisions. These methodological rules or principles can be summed up in two: [symbol: see text]be inventive and critical!, that is, propose bold hypotheses and subject them to severe tests of experience. Logic plays its role mainly by allowing us to deduce from a hypothesis the predictions to be confronted with the facts or evidence. This is applicable both to statistical inference as well as to causal inference. We argue that the criteria of causality used in epidemiology are none other than rules of the method designed for the same purpose: they are concerned with eliminating or reducing errors (chance, bias...) on testing a causal hypothesis. Consequently, the so-called ausal inference, the step from evidence to causal theory, is not a logical inductive or probabilistic process but rather a decision based on the evaluation of a causal hypothesis thanks to methodological rules such as the criteria of causality. We believe that the interest of the debate between the Popperian and the inductivist epidemiologists is not merely a matter of words, as, if we are aware that we do not operate inductively, that we cannot establish firmly hypotheses, not even affirm them probabilistically, we will presumably adopt a humbler attitude and look more for the errors in our theories than for their facile examples of confirmation.

Causality↗

PubMed2026

Decoding TnsC Filament Assembly in CRISPR-Associated Transposons Using Interpretable Deep Learning and Molecular Simulations.

CRISPR-associated transposons (CASTs) enable programmable DNA integration, yet how the TnsC regulator forms processive filaments on DNA to coordinate RNA-guided transposition in type V-K CAST systems remains unknown. Here, we integrate large-scale molecular simulations, interpretable deep learning using graph attention networks (GATs), and causal inference analyses to define the molecular determinants of TnsC filament nucleation and elongation. We show that TnsC nucleates by inducing localized DNA deformation that propagates along extended filaments, with Granger causality revealing that TnsC motions precede and predict DNA deformation. Interpretable GAT models demonstrate that elongation is determined during early recognition between incoming and DNA-bound subunits, followed by structural reorganization that regenerates the recruitment interface and enables processive assembly. These results elucidate the molecular mechanism of processive TnsC filament assembly and explain why isolated TnsC filaments preferentially elongate in the 5' → 3' direction, while accessory transposition factors can reshape the interaction landscape and alter filament growth polarity. Together, these findings advance our understanding of CAST function and inform the engineering of programmable DNA integration platforms. Beyond CAST systems, this work introduces an interpretable GAT approach as a general and transferable deep learning strategy for uncovering molecular mechanisms in biological systems, while demonstrating the power of causal inference for dissecting directional relationships in molecular dynamics.

Deep Learning↗

PubMedDate not supplied

Threats to internal validity in worksite health promotion program research: common problems and possible solutions.

Reviews of the research on worksite health promotion programs reveal that most studies are limited in their ability to draw clear inferences about program effects because the studies employed flawed research designs and/or analyses. Conclusions are often drawn about program effectiveness with little consideration given to alternative explanations for the findings. In an effort to promote improved research, this article uses the Cook and Campbell delineation of threats to valid causal inference to illustrate how the threats can operate in worksite health promotion program research as well as how they can be examined and controlled. Researchers, even those attempting to conduct true experiments, must consider the existence of all plausible threats to validity and control or rule them out before valid causal inferences can be drawn. The theoretical and design issues involved in worksite health promotion program research are presented, followed by a discussion of threats to internal validity.

Data Collection↗

PubMed1999

Instrumental variables when evaluating screening trials: estimating the benefit of detecting cancer by screening.

When evaluating the benefit of detecting cancer by screening we try to answer the question, 'what would a screen detected subject's outcome have been if his/her cancer had progressed to clinical detection'. By 'outcome' we mean survival time, cancer size and stage, lead time effects and more. Because only an unethical study can answer it directly, researchers have attempted to answer the question indirectly using data from randomized cancer screening studies (subjects randomized to study (screened) or control (not screened)). Inferences are made by first selecting the cancer cohort (those subjects who are found to have cancer), then comparing subjects having screen detected cancers to subjects having clinically detected cancers. However, there are two difficulties with this approach: (i) because screening (intends to) detect cancers early, at the trial's end the study group contains more cancer cases than the control group and so the cancer cohort has some unidentified control subjects missing (that is, subjects having cancer during the screening period that have not yet been clinically detected); (ii) because screen detected cancers (may) differ from clinically detected cancers, the comparison group should include only a (non-identified) subset of the cancer cohort's control subjects (that is, only those control subjects having cancers that would have been screen detected). Statistical literature acknowledges these difficulties and attempts to solve them separately, but without success; those methods do not yield meaningful causal inferences and admit substantial bias. Recently, Angrist, Imbens and Rubin and Imbens and Rubin provide a framework for instrumental variable methods that we interpret as allowing us to make causal inferences with incompletely identified comparison groups. We apply their framework to evaluating cancer screening trials and find that we may simultaneously accommodate both difficulties while giving a meaningful answer to the question posed above. Using data from a breast cancer screening trial we demonstrate the general method with a variety of outcome measures and extensions.

Adult↗

PubMed1984

Causal thinking in practice: strengths and weaknesses of the clinical vantage point.

The place of causal inference in clinical practice is considered. One function relates to the validation of existing knowledge, as in the evaluation of the medical literature, and is exemplified by the process of editorial selection. A second function relates to the validation of potential knowledge, as in the testing of hypotheses, and is exemplified by the pursuit of a nutritional cause of retarded mental development. Features of clinical practice that contribute strengths or weaknesses to causal inference are then examined. These include the diagnostic process, the effect of cumulative clinical experience, continuity of observation and follow-up, and the focus on disease.

Child↗

PubMed2026

Genetics of sensory nutrition.

Sensory nutrition is an emerging research area that examines how chemosensory perception, particularly taste and smell, shapes dietary behaviours, nutritional status, and disease risk. Variation in how individuals perceive the same foods may help explain differences in diet quality and responsiveness to behavioural dietary interventions, yet chemosensory phenotypes are rarely measured at the population level. Genetic variation contributes to this perceptual diversity and provides a framework for investigating sensory determinants of diet using genomic approaches. This review summarises evidence linking chemosensory genetics to perception and dietary behaviours, and discusses applications for causal inference and for precision and personalised nutrition. Twin studies reveal moderate to high heritability for bitter taste traits, with more modest and phenotype-dependent estimates for sweetness, sourness, saltiness, fat-related traits, and olfactory measures. Genome-wide association studies have identified loci in taste and olfactory receptor genes associated with specific chemosensory traits as well as liking and intake of various foods, although the evidence remains concentrated on bitter taste and populations of European ancestry. These genetic variants have been used in Mendelian randomisation, a genetics-based approach that strengthens causal inference, to test whether sensory traits influence dietary behaviour. For precision nutrition, evidence for taste genotype-stratified interventions remains limited and mixed. Realising the promise of sensory nutrition will require scalable and standardised chemosensory phenotyping, Findable, Accessible, Interoperable, and Reusable (FAIR) data infrastructure, expanded research in diverse populations, and integration with broader biological and sociocultural determinants of dietary intake.

Genetics↗

PubMed1991

Conceptus viability, malformation, and suspect mutagens or teratogens in humans. The Yule-Simpson paradox and implications for inferences of causality in studies of mutagenicity or teratogenicity limited to human livebirths.

If a genetic disorder or a congenital malformation is associated with embryonic or fetal death, then this may distort reported associations of environmental or biological factors with genetic and congenital disorders. We illustrate how such distortion may occur in humans as a manifestation of the Yule-Simpson (or Simpson) statistical paradox. We analyze the reported negative association of maternal smoking and Down syndrome in livebirths and demonstrate that under plausible conditions, valid negative associations both in livebirths and in embryonic or fetal deaths can occur, despite the fact that in all conceptuses there may be no association whatsoever. A similar analogous manifestation of the Yule-Simpson paradox may hold for positive associations. That is, positive relative risks may occur for an alleged mutagen or teratogen in separate studies both of livebirths, and of embryonic and fetal deaths, despite the absence of any effect in all conceptuses. These considerations illustrate the need for data on genetic disorders and malformations in embryonic and fetal deaths, and the need for a synthesis of such data with those from studies in livebirths before valid inferences may be made about mutagenic or teratogenic effects of environmental or biological factors.

Congenital Abnormalities↗

PubMedDate not supplied

The causal role for genital ulcer disease as a risk factor for transmission of human immunodeficiency virus. An application of the Bradford Hill criteria.

BACKGROUND AND OBJECTIVES: Genital ulcer disease (GUD) has been reported to increase the risk for the acquisition of human immunodeficiency virus (HIV). Although many investigators have reported an increased risk for HIV infection in persons with concurrent or previous GUD, not all studies have been designed to determine whether GUD causes an increased risk for HIV infection or acts only as a risk marker for infection. The evidence from the literature is discussed, and the criteria for causal inference proposed by Sir Austin Bradford Hill are applied. GOAL: To evaluate the strength of the association between GUD and infection by HIV. STUDY DESIGN: Case-control, cross-sectional, and cohort studies that examined the association between HIV seroconversion and GUD were chosen from the literature. Twenty-seven epidemiologic studies were selected for analysis, many of which reported separate analyses of the association between HIV infection and herpes simplex virus infection, syphilis, or nonspecified GUD. The studies were analyzed to investigate the magnitude of association between GUD and HIV, and the evidence evaluated using Hill's criteria. RESULTS: Approximately two thirds of the analyses reported a statistically significant association between GUD and HIV infection. Fourteen studies reported 29 separate analyses using a case-control design, 18 of which reported a statistically significant association between GUD (GUD, herpes, and syphilis) and HIV infection, four analyses were of varying significance depending on the analytical technique used, and seven were nonsignificant. Thirteen studies reported 23 separate longitudinal analyses that used a nested case-control or cohort design: 11 reported a significant association, 11 had nonsignificant findings, and results of one study varied. No study reported a statistically significant negative association. When applying the literature to Hill's criteria, all nine criteria for causal inference were met, providing additional evidence that genital ulcers are associated with an increased risk for the development of HIV infection. CONCLUSIONS: The published evidence suggests that GUD increases the risk for HIV acquisition. Few studies, however, have examined carefully the temporal association between preexisting GUD and subsequent HIV acquisition. The analyses that simultaneously controlled for additional risks for HIV infection, such as lifetime sex partners or history of injection drug use, report a generally lower risk for HIV associated with GUD. It is likely that studies that adequately control for risk factors will find a lower risk associated with GUD than was reported in the literature earlier in the HIV epidemic. Future research needs and the problems associated with conducting these types of studies are discussed.

Case-Control Studies↗

PubMed2004

Open-minded scepticism: inferring the causal mechanisms of human ageing from genetic perturbations.

Given the myriad of age-related changes and the many proposed mechanistic theories of ageing, a major problem in gerontology is distinguishing causes from effects. This review aims to identify and evaluate those mechanisms which have gathered experimental support in favour of seeing them as a cause rather than an effect of ageing. Recent results related to energy metabolism and ageing, the free radical and the DNA damage theories of ageing are reviewed and their predictions evaluated through a systems biology rationale. Crucial in this approach are genetic manipulations in animal models that enable researchers to discriminate causes from effects of ageing and focus on the causal structure of human ageing. Based on a system-level interpretation, the GH/IGF-1 axis appears the most likely explanation for caloric restriction and a possible causal mechanism of human ageing. Although much work remains to fully understand the human ageing process, there is little evidence that free radicals are a causal factor in mammalian ageing, though they may be involved in signalling pathways related to ageing. On the other hand, studying how the DNA machinery affects ageing appears a promising avenue for disclosing the human ageing process.

Aged↗

PubMed1999

Perceiving individuals and groups: expectancies, dispositional inferences, and causal attributions.

Two experiments investigated differences in forming impressions of individual and group targets. Experiment 1 showed that when forming an impression of an individual, perceivers made more extreme trait judgments, made those judgments more quickly and with greater confidence, and recalled more information than when the impression target was a group. Experiment 2 showed that when participants were forming an impression of an individual, expectancy-inconsistent behaviors spontaneously triggered causal attributions to resolve the inconsistency; this was not the case when the impression target was a group. Results are interpreted as reflecting perceivers' a priori assumptions of unity and coherence in individual versus group targets.

Analysis of Variance↗

PubMed2003

Normative and descriptive accounts of the influence of power and contingency on causal judgement.

The power PC theory (Cheng, 1997) is a normative account of causal inference, which predicts that causal judgements are based on the power p of a potential cause, where p is the cause-effect contingency normalized by the base rate of the effect. In three experiments we demonstrate that both cause-effect contingency and effect base-rate independently affect estimates in causal learning tasks. In Experiment 1, causal strength judgements were directly related to power p in a task in which the effect base-rate was manipulated across two positive and two negative contingency conditions. In Experiments 2 and 3 contingency manipulations affected causal estimates in several situations in which power p was held constant, contrary to the power PC theory's predictions. This latter effect cannot be explained by participants' conflation of reliability and causal strength, as Experiment 3 demonstrated independence of causal judgements and confidence. From a descriptive point of view, the data are compatible with Pearce's (1987) model, as well as with several other judgement rules, but not with the Rescorla-Wagner (Rescorla & Wagner, 1972) or power PC models.

Adolescent↗

PubMed2001

Causal judgments about relations between multilevel variables.

Three experiments presented stimulus information about cause and effect variables taking 3 quantitative values. Judgments tended to vary in accordance with considerations of conditions affecting the validity of causal inference from correlational data: whether causal candidates were presented simultaneously or in a temporal order such that one could affect the other and whether candidates were confounded with each other. The results supported a general hypothesis that causal judgments are moderated in accordance with acquired methodological intuitions. The fourth experiment showed that tendencies in correlation judgment were different from those in causal judgment, further supporting the hypothesis that causal judgment from multilevel variable information is, to some extent, determined by processes or conceptual frameworks specific to the domain of causal cognition.

Adult↗

PubMed2003

Rethinking temporal contiguity and the judgement of causality: effects of prior knowledge, experience, and reinforcement procedure.

Time plays a pivotal role in causal inference. Nonetheless most contemporary theories of causal induction do not address the implications of temporal contiguity and delay, with the exception of associative learning theory. Shanks, Pearson, and Dickinson (1989) and several replications (Reed, 1992, 1999) have demonstrated that people fail to identify causal relations if cause and effect are separated by more than two seconds. In line with an associationist perspective, these findings have been interpreted to indicate that temporal lags universally impair causal induction. This interpretation clashes with the richness of everyday causal cognition where people apparently can reason about causal relations involving considerable delays. We look at the implications of cause-effect delays from a computational perspective and predict that delays should generally hinder reasoning performance, but that this hindrance should be alleviated if reasoners have knowledge of the delay. Two experiments demonstrated that (1) the impact of delay on causal judgement depends on participants' expectations about the timeframe of the causal relation, and (2) the free-operant procedures used in previous studies are ill-suited to study the direct influences of delay on causal induction, because they confound delay with weaker evidence for the relation in question. Implications for contemporary causal learning theories are discussed.

Adult↗

PubMed2002

Mechanisms of predictive and diagnostic causal induction.

In predictive causal inference, people reason from causes to effects, whereas in diagnostic inference, they reason from effects to causes. Independently of the causal structure of the events, the temporal structure of the information provided to a reasoner may vary (e.g., multiple events followed by a single event vs. a single event followed by multiple events). The authors report 5 experiments in which causal structure and temporal information were varied independently. Inferences were influenced by temporal structure but not by causal structure. The results are relevant to the evaluation of 2 current accounts of causal induction, the Rescorla-Wagner (R. A. Rescorla & A. R. Wagner, 1972) and causal model theories (M. R. Waldmann & K. J. Holyoak, 1992).

Analysis of Variance↗

PubMed1997

Statistical significance--a misconstrued notion in medical research.

The P-value is the significance probability of obtaining a value of the test statistic that is as extreme, in relation to the null hypothesis, as that observed. Medical researchers may, in some situations, disagree on its appropriate use or on its interpretation as a summary measure of consistency with the null hypothesis in a particular data set. More informative statistical measures such as the likelihood ratio and the Bayesian posterior probability have been suggested for drawing inferences from clinical trials and epidemiologic studies. Causal inference is not statistical in nature; rather it strives to provide scientific explanations or criticisms of proposed explanations that would describe the observed data pattern. In this context, it is important to remember that a finding may not be medically important, or a causal hypothesis may even not be true even if a study shows a significant P-value.

Bayes Theorem↗

PubMed2025

Circulating inflammatory proteins as causal drivers and therapeutic targets in asthma: insights from genetic and pathway-based analyses.

OBJECTIVE: To identify circulating inflammatory proteins with potential causal roles in asthma development through integrated genetic and pathway-based analyses, and to evaluate their potential as therapeutic targets. METHODS: We used genetically anchored instrumental variables from 180 protein quantitative trait loci (pQTLs) to assess the causal effects of 91 circulating inflammatory proteins on asthma risk, using large-scale GWAS datasets. Analytical robustness was evaluated through pleiotropy and heterogeneity testing. Functional enrichment and literature-based pathway analyses were performed to support biological plausibility and validate findings. RESULTS: Four proteins showed significant causal effects on asthma: CCL19 and LIFR were protective (OR = 0.89 and 0.91, p ≤ 6.8E-03), while ARTN and IL6 were associated with increased risk (OR = 1.15 and 1.18, p ≤ 1.1E-04). We also identified reverse causal effects of asthma on 11 cytokines, including MMP10, TGFB1, IL33, and IL18R1. Most of these proteins were enriched in pathways related to cytokine signaling and immune response (p < 0.001). All identified proteins had prior literature support linking them to asthma or airway inflammation. CONCLUSIONS: Our findings highlight a subset of circulating inflammatory proteins that are likely causal in asthma pathogenesis and may serve as promising targets for therapeutic intervention. These results offer novel insights into the immunological mechanisms underlying asthma and support the utility of genetic causal inference in target prioritization.

Asthma↗

PubMed1992

Causality in medicine: towards a theory and terminology.

One of the cornerstones of modern medicine is the search for what causes diseases to develop. A conception of multifactorial disease causes has emerged over the years. Theories of disease causation, however, have not quite been developed in accordance with this view. It is the purpose of this paper to provide a fundamental explication of aspects of causation relevant for discussing causes of disease. The first part of the analysis will discuss discrimination between singular and general causality. Singular causality, as in the specific patient, is a relation between a concrete sequence of causally linked events. General causation, e.g. as in disease etiology, means various categories of causal relations between event types. The paper introduces the concept of a reference case serving as a source for causal inference, reaching beyond the concept of general causality. The second part of the analysis provides exemplification of a theory of causation suitable for discussing singular causation. The chain of events that induce a disease state can be identified as effective causal complexes, each complex composed of non-redundant components, which separately contribute to the effect of the complex, without the individual component being necessary or sufficient in itself to produce the effect. In the third part of the analysis the theory is elaborated further. Causes, defined as non-redundant components, can furthermore be differentiated according to their avoidability, according to theories about human error or by the potential of eradication. Multifactorial models of disease creates a need for systematic approaches to causal factors. The paper proposes a taxonomical terminology that serves this purpose.

Adult↗

PubMed2004

Assessing interactive causal influence.

The discovery of conjunctive causes--factors that act in concert to produce or prevent an effect--has been explained by purely covariational theories. Such theories assume that concomitant variations in observable events directly license causal inferences, without postulating the existence of unobservable causal relations. This article discusses problems with these theories, proposes a causal-power theory that overcomes the problems, and reports empirical evidence favoring the new theory. Unlike earlier models, the new theory derives (a). the conditions under which covariation implies conjunctive causation and (b). functions relating observable events to unobservable conjunctive causal strength. This psychological theory, which concerns simple cases involving 2 binary candidate causes and a binary effect, raises questions about normative statistics for testing causal hypotheses regarding categorical data resulting from discrete variables.

Humans↗