Learning Subgroups with Maximum Treatment Effects Without Causal Heuristics

To evaluate the effects of a dexmedetomidine-midazolam-ketamine (DMK) combination administered IM to captive red-footed tortoises (Chelonoidis carbonaria). 12 healthy adult red-footed tortoises. In a prospective experimental study, DMK (0.1, 1.0, and 10 mg/kg, respectively) was administered IM as separate injections into the right antebrachium. Atipamezole (0.5 mg/kg, IM) and flumazenil (0.05 mg/kg, SC) were administered into the left antebrachium 60 minutes later. Times to the first treatment response and maximal treatment effect after DMK administration and time to recovery after reversal agent administration were recorded. Vital signs and reflexes or responses to stimuli were assessed and recorded at predetermined intervals. DMK treatment produced deep sedation or light anesthesia for ≥ 20 minutes in all tortoises. Induction and recovery were rapid, with no complications noted. Median times to first response, maximum effect, and recovery were 4.5, 35, and 14.5 minutes, respectively. Two tortoises required additional reversal agent administration but recovered < 20 minutes after the repeated injections. Mean heart and respiratory rates decreased significantly over time. All animals lost muscle tone in the neck and limbs from 35 to 55 minutes after DMK injection, but other variables including palpebral reflexes, responses to mild noxious stimuli (eg, toe pinching, tail pinching, and saline ([0.9 NaCl] solution injection), and ability to intubate were inconsistent. DMK administration produced deep sedation or light anesthesia with no adverse effects in healthy adult red-footed tortoises. At the doses administered, deep surgical anesthesia was not consistently achieved. Anesthetic depth must be carefully evaluated before performing painful procedures in red-footed tortoises with this DMK protocol.

Introduction: Despite careful design of dose ranging studies, full characterisation of dose-response for clinical endpoints is not always achieved. The absence of a dose-response for reduction in clinically significant exacerbations with mepolizumab in the primary analysis led to a meta-analysis using individualized weight-based measures of dose. Objectives: Evaluate weight-based measures of mepolizumab dose-response for reduction in annualized exacerbation rate. Methods: Inclusion of exacerbation data with dose and bodyweight information from two GSK funded mepolizumab severe asthma pivotal studies (DREAM and MENSA). Intravenous doses were converted to subcutaneous-equivalent dose using absolute bioavailability of 80 and 70% and normalized by individual bodyweight, body mass index and body surface area at baseline. Results: Description of the dose-response showed that although the maximal treatment effect was well-estimated, location of half-maximal efficacious dose (ED 50 ) was poor. Using a Bayesian analysis with informative priors for ED 50 (away from the pharmacological ED 50 (blood eosinophil reduction) [11 mg]), convergence was achieved. Posterior ED 50 distributions showed a shift in prior belief towards the pharmacological ED 50 with narrower credibility interval. The largest effects were noted for bodyweight-adjusted dose-response. Conclusions: Exacerbation data from studies DREAM and MENSA did not show a dose-response by conventional analysis unlike with Bayesian analysis which also showed consistency between the efficacy dose-response and the well-defined dose-response for pharmacology, with an ED 50 of approximately 0.17 mg/kg (12 mg).

Suicide and violence assessment in psychiatry

Background: To compare different surgical correction methods of vesico-ureteric reflux in children using both open surgery and endoluminal (intraluminal) treatment options of this pathology. Patients and methods: 166 patients aged from 4 months to 13 were examined and treated. All children underwent X-ray urological examination through cystography and ultrasound examination of kidneys and urocyst. Cohen's operation, STING procedure involving endoscopic injection of bulking agents was performed. Treatment results were assessed 6 months later through control cystography. Conclusion: Endoscopic correction of vesico-ureteric reflux is an effective and minimally invasive treatment method for this pathology. Maximum treatment effect after biodegradable bulking agents application should be expected in younger age group. Treatment rates of endoluminal correction of vesico-ureteric reflux with the use of sterile viscous gel as a fixing agent are higher than with the use of bovine collagen.

An important yet under-explored question in the teamwork literature concerns how group characteristics affect productivity. Within a given teamwork setting, it is not obvious how group member diversity affects the performance of the individual and the group. The group may gain from knowledge transfer and sharing while it may be crippled by communication and coordination problems that are prevalent in heterogeneous groups. In this study, we combine class performance data from an undergraduate management class with students' personal records to explore diversity and knowledge spillover effects. A major advantage of our dataset is the exogenous assignment of groups, which rules out the troublesome yet common self-selection issue in team literature. Our results indicate that male-dominant groups performed worse both in group work and in individually taken exams than female-dominant and equally-mixed gender groups after controlling for other group characteristics. Individual members from a group with more diversity in age and gender scored higher in exams. However, we did not find any significance of a group's racial composition over group and individual performances. Another novel aspect of this natural experiment is that each group chooses their own group contract form -members of "autonomous" groups receive equal grade for their group work while those in "democratic" groups can adopt differentiated point allocation, thus, providing a proper mechanism to punish free riders. Our estimation results show a significant correlation between the choice of a democratic contract and the group and individual performance. To address the endogeneity problem in groups' contract choices, we use a maximum likelihood treatment effect model and found that the democratic group contract has a positive and significant effect on group performance.

BackgroundCombined chemotherapy is the cornerstone treatment for patients with advanced Hodgkin lymphoma (HL). The objective of our study was to perform a network meta-analysis of the efficacy of different chemotherapy regimens in adults with advanced-stage HL.Materials and methodsWe searched for relevant randomized controlled trials (RCTs) in titles/abstracts in PubMed, Embase, and the Cochrane Library. The search was last updated on April 3, 2018. RCTs that assessed the effectiveness of one of the following treatments were included: doxorubicin, bleomycin, vinblastine, and dacarbazine (ABVD); four cycles of increased dose of bleomycin, etoposide, doxorubicin, cyclophosphamide, vincristine, procarbazine, and prednisone (BEACOPPescalated) followed by two or four cycles of standard dose of BEACOPP (4× BEACOPPescalated + 2 or 4× BEACOPPbaseline); brentuximab vedotin plus doxorubicin, vinblastine, and dacarbazine (A+AVD); doxorubicin, vinblastine, mechlorethamine, vincristine, bleomycin, etoposide, and prednisone combined with radiation therapy (Stanford V); mechlorethamine (cyclophosphamide), vincristine, procarbazine, and prednisone (M[C] OPP); sequential or alternating chemotherapy regimens with ABVD as the footstone (eg, COPP/ABVD or mechlorethamine, vincristine, procarbazine, and prednisone [MOPP]/ABVD); eight cycles of BEACOPPescalated; hybrid MOPP/ABV; and M[C]EC (M[C]OPP with epidoxorubicin, bleomycin, vinblastine [EBV], and lomustine, doxorubicin, and vindesine [CAD]).ResultsOverall, we screened 3,564 citations and deemed 18 reports of 16 trials eligible and included them in our network meta-analysis. A total of 11,928 participants were randomly assigned to one of the 12 combinations of chemotherapy regimens, of which 11,476 participants were analyzed. For the overall survival (OS), no differences were observed within any interventions when the ABVD regimen was used as the reference treatment. Similarly, relative to A+AVD, 8× BEACOPPescalated and 6× BEACOPPescalated also showed no differences (HR =1.07, 95% credible interval (CrI): 0.58–1.95; HR =0.62, 95% CrI: 0.16–1.83; and HR =0.71, 95% CrI: 0.30–1.72, respectively). In terms of complete remission (CR), enough evidence exists to support a maximum clinical treatment effect for 6× BEACOPPescalated (OR =1.88, 95% CrI: 1.20–2.96; and OR =3.43, 95% CrI: 1.87–6.24).ConclusionWhen compared across the 12 combined chemotherapy regimens, six cycles of BEACOPPescalated may be the optimal treatment for patients with advanced-stage HL.

Malocclusion and treatment need in United States youths 12 to 17 years of age

BackgroundCausal inference has seen an increasing popularity in medical research. Estimation of causal effects from observational data allows to draw conclusions from data when randomized controlled trials cannot be conducted. Although the identification of structural causal models (SCM) and the calculation of structural coefficients has received much attention, a key requirement for valid causal inference is that conclusions are drawn based on the true data-generating model.MethodsIt remains widely unknown how large the probability is to reject the true structural causal model when observational data from it is sampled. The latter probability – the causal false-positive risk – is crucial, as rejection of the true causal model can induce bias in the estimation of causal effects. In this paper, the widely used causal models of confounders and colliders are studied regarding their causal false-positive risk in linear Markovian models. A simulation study is carried out which investigates the causal false-positive risk in Gaussian linear Markovian models. Therefore, the testable implications of the DAG corresponding to confounders and colliders are analyzed from a Bayesian perspective. Furthermore, the induced bias in estimating the structural coefficients and causal effects is studied.ResultsResults show that the false-positive risk of rejecting a true SCM of even simple building blocks like confounders and colliders is substantial. Importantly, estimation of average, direct and indirect causal effects can become strongly biased if a true model is rejected. The causal false-positive risk may thus serve as an indicator or proxy for the induced bias.ConclusionWhile the identification of structural coefficients and testable implications of causal models have been studied rigorously in the literature, this paper shows that causal inference also must develop new concepts for controlling the causal false-positive risk. Although a high risk cannot be equated with a substantial bias, it is indicative of the induced bias. The latter fact calls for the development of more advanced risk measures for committing a causal type I error in causal inference.

Causal intervention is an essential tool in causal inference. It is axiomatized under the rules of do-calculus in the case of structure causal models. We provide simple axiomatizations for families of probability distributions to be different types of interventional distributions. Our axiomatizations neatly lead to a simple and clear theory of causality that has several advantages: it does not need to make use of any modelling assumptions such as those imposed by structural causal models; it relies only on interventions on single variables; it includes most cases with latent variables and causal cycles; and, more importantly, it does not assume the existence of an underlying true causal graph as we do not take it as the primitive object; moreover, a causal graph is derived as a by-product of our theory. We show that, under our axiomatizations, the intervened distributions are Markovian to the defined intervened causal graphs, and an observed joint probability distribution is Markovian to the obtained causal graph; these results are consistent with the case of structural causal models, and as a result, the existing theory of causal inference applies. We also show that a large class of natural structural causal models satisfy the theory presented here. The aim of this paper is axiomatization of interventional families, which is subtly different from causal modelling.

Data-driven root cause analysis via causal discovery using time-to-event data

Causal assertions stem from an asymmetric relation between some variable's causes and effects, i.e., they imply the existence of a function decomposition of a model where the effects are a function of the causes without implying that the causes are functions of the effects. In structural causal models, information is encoded in the compositions of functions that define variables because that information is used to constraint how an intervention that changes the definition of a variable influences the rest of the variables. Current probabilistic models with tractable marginalization also imply a function decomposition but with the purpose of allowing easy marginalization of variables. In this article, structural causal models are extended so that the information implicitly stored in their structure is made explicit in an input–output mapping in higher dimensional representation where we get to define the cause–effect relationships as constraints over a function space. Using the cause–effect relationships as constraints over a space of functions, the existing methodologies for handling causality with tractable probabilistic models are unified under a single framework and generalized.

ABSTRACTIntroductionCardiac arrhythmias remain a leading cause of cardiovascular morbidity and mortality worldwide, and conventional diagnostic tools such as electrocardiography and Holter monitoring may fail to detect transient or asymptomatic events. Recent advances in artificial intelligence (AI) and machine learning have enhanced arrhythmia detection, risk stratification, and treatment planning; however, most existing models rely primarily on statistical associations rather than underlying physiological mechanisms.MethodsThis narrative review was conducted through a structured but non‐systematic literature search of PubMed, Scopus, and Google Scholar, covering studies published between 2000 and 2025. Search terms included combinations of “cardiac arrhythmia,” “atrial fibrillation,” “causal inference,” “structural causal models,” “digital twins,” “mechanistic modeling,” “cardiac electrophysiology modeling,” and “artificial intelligence.” Peer‐reviewed articles were included if they demonstrated methodological depth and addressed causal inference or mechanistic modeling approaches in cardiovascular research, particularly in arrhythmia detection, risk prediction, treatment optimization, or clinical validation frameworks. Studies were prioritized based on methodological rigor, translational relevance, and recency. Editorials lacking methodological detail, non‐English publications, and studies relying solely on predictive models without incorporating causal or mechanistic components were excluded.ResultsCausal artificial intelligence (Causal AI), offers a more mechanistically grounded framework for understanding arrhythmogenesis and therapeutic outcomes. Emerging evidence suggests that integrating clinical data with structural causal models, mechanistic modeling, and patient‐specific digital twins can bridge the gap between predictive performance and physiological interpretability. These approaches show promise in predicting ablation success, guiding therapy, and improving individualized care.ConclusionDespite this potential, clinical implementation remains limited due to data heterogeneity, validation challenges, and regulatory constraints.

Large language models (LLMs) have shown great potential in decision-making due to the vast amount of knowledge stored within the models.However, these pre-trained models are prone to lack reasoning abilities and are difficult to adapt to new environments, further hindering their application to complex real-world tasks. To address these challenges, inspired by the human cognitive process, we propose Causal-Aware LLMs, which integrate the structural causal model (SCM) into the decision-making process to model, update, and utilize structured knowledge of the environment in a "learning-adapting-acting" paradigm.Specifically, in the learning stage, we first utilize an LLM to extract the environment-specific causal entities and their causal relations to initialize a structured causal model of the environment. Subsequently, in the adapting stage, we update the structured causal model through external feedback about the environment, via an idea of causal intervention. Finally, in the acting stage, Causal-Aware LLMs exploit structured causal knowledge for more efficient policy-making through the reinforcement learning agent. The above processes are performed iteratively to learn causal knowledge, ultimately enabling the causal-aware LLM to achieve a more accurate understanding of the environment and make more efficient decisions. Experimental results across 22 diverse tasks within the open-world game "Crafter" validate the effectiveness of our proposed method.

Large language models (LLMs) have shown great potential in decision-making due to the vast amount of knowledge stored within the models.However, these pre-trained models are prone to lack reasoning abilities and are difficult to adapt to new environments, further hindering their application to complex real-world tasks. To address these challenges, inspired by the human cognitive process, we propose Causal-Aware LLMs, which integrate the structural causal model (SCM) into the decision-making process to model, update, and utilize structured knowledge of the environment in a "learning-adapting-acting" paradigm.Specifically, in the learning stage, we first utilize an LLM to extract the environment-specific causal entities and their causal relations to initialize a structured causal model of the environment. Subsequently, in the adapting stage, we update the structured causal model through external feedback about the environment, via an idea of causal intervention. Finally, in the acting stage, Causal-Aware LLMs exploit structured causal knowledge for more efficient policy-making through the reinforcement learning agent. The above processes are performed iteratively to learn causal knowledge, ultimately enabling the causal-aware LLM to achieve a more accurate understanding of the environment and make more efficient decisions. Experimental results across 22 diverse tasks within the open-world game "Crafter" validate the effectiveness of our proposed method.

In the big data era, data are usually described by multiple information sources in many practical application fields, and it is infeasible way to mine interesting knowledge from a centralized huge source directly after aggregating all information sources, due to the problems of massive data, privacy concerns, and potential transmission cost. To mine decision rules from multiple information sources, therefore, we can mine local decision rules at different local information sources, and then put these rules to form a set of global decision rules. In this paper, we first present a formal representation of decision rule, which depends on the neighborhood granulation of each sample from the viewpoint of granular computing. Then, we obtain the weight of each local information source based on the consensus measure principle between local information sources. Finally, a weighting model for mining global decision rules via synthesizing all local decision rules is proposed. Extensive experimental results demonstrate that the proposed decision rules synthesization model is effective and scalable.