Parsimonious Model Research Articles

Design and validation of the AI literacy questionnaire: The affective, behavioural, cognitive and ethical approach

Artificial intelligence (AI) literacy is at the top of the agenda for education today in developing learners' AI knowledge, skills, attitudes and values in the 21st century. However, there are few validated research instruments for educators to examine how secondary students develop and perceive their learning outcomes. After reviewing the literature on AI literacy questionnaires, we categorized the identified competencies in four dimensions: (1) affective learning (intrinsic motivation and self‐efficacy/confidence), (2) behavioural learning (behavioural commitment and collaboration), (3) cognitive learning (know and understand; apply, evaluate and create) and (4) ethical learning. Then, a 32‐item self‐reported questionnaire on AI literacy (AILQ) was developed and validated to measure students' literacy development in the four dimensions. The design and validation of AILQ were examined through theoretical review, expert judgement, interview, pilot study and first‐ and second‐order confirmatory factor analysis. This article reports the findings of a pilot study using a preliminary version of the AILQ among 363 secondary school students in Hong Kong to analyse the psychometric properties of the instrument. Results indicated a four‐factor structure of the AILQ and revealed good reliability and validity. The AILQ is recommended as a reliable measurement scale for assessing how secondary students foster their AI literacy and inform better instructional design based on the proposed affective, behavioural, cognitive and ethical (ABCE) learning framework. Practitioner notesWhat is already known about this topic AI literacy has drawn increasing attention in recent years and has been identified as an important digital literacy. Schools and universities around the world started to incorporate AI into their curriculum to foster young learners' AI literacy. Some studies have worked to design suitable measurement tools, especially questionnaires, to examine students' learning outcomes in AI learning programmes. What this paper adds Develops an AI literacy questionnaire (AILQ) to evaluate students' literacy development in terms of affective, behavioural, cognitive and ethical (ABCE) dimensions. Proposes a parsimonious model based on the ABCE framework and addresses a skill set of AI literacy. Implications for practice and/or policy Researchers are able to use the AILQ as a guide to measure students' AI literacy. Practitioners are able to use the AILQ to assess students' AI literacy development.

Relations of Lifetime Perceived Stress and Basal Cortisol With Hippocampal Volume Among Healthy Adolescents and Those at Clinical High Risk for Psychosis: A Structural Equation Modeling Approach

BackgroundHippocampal volume (HV) is sensitive to environmental influences. Under normative conditions in humans, HV increases linearly into childhood and asymptotes in early adulthood. Studies of humans and nonhuman animals have provided evidence of inverse relationships between several measures of stress and HV. MethodsUsing structural equation modeling, this study aimed to characterize the relationships of age, basal cortisol, biological sex, and lifetime perceived stress with bilateral HV in a sample of healthy adolescents and adolescents at clinical high risk for psychosis (CHR-P) (N = 571, 43% female; age range = 12–19.9 years). This sample included 469 individuals at CHR-P and 102 healthy comparison participants from the combined baseline cohorts of the second and third NAPLS (North American Prodrome Longitudinal Study). ResultsA structural model that constrained the individual effects of basal cortisol and perceived stress to single path coefficients, and freely estimated the effects of age and biological sex in group models, optimized model fit and parsimony relative to other candidate models. Significant inverse relationships between basal cortisol and bilateral HV were observed in adolescents at CHR-P and healthy comparison participants. Significant sex differences in bilateral HV were also observed, with females demonstrating smaller HV than males in both groups. ConclusionsMultigroup structural equation modeling revealed heterogeneity in the relationships of age and biological sex with basal cortisol, lifetime perceived stress, and bilateral HV in individuals at CHR-P and healthy comparison participants. Moreover, the findings support previous literature indicating that elevated basal cortisol is a nonspecific risk factor for reduced HV.

Tax Avoidance and Firm Risk: New Insights from a Latent Class Mixture Model

ABSTRACT We examine the relations between tax avoidance and two components of firm risk (priced risk and idiosyncratic risk) using latent class mixture models. OLS regressions suggest that tax avoidance is negatively related to both priced risk and idiosyncratic risk. However, results from latent class mixture models reveal that 35.6 (58) percent of the sample exhibits a significant positive association between tax avoidance and priced (idiosyncratic) risk and firm characteristics are strikingly different across the latent classes. We use these differences to develop parsimonious models to predict latent class assignments, which we then use to demonstrate that firms predicted to have a positive relation between tax avoidance and priced risk or idiosyncratic risk have more negative future tax outcomes, including higher relative cash tax rate volatility, larger additions to tax reserves, larger payouts to tax authorities, and more negative tax-related news coverage over the following three years. Data Availability: Data are available from the public sources cited in the text. JEL Classifications: M41; G12; D81.

The sanitary sewer unit hydrograph model: A comprehensive tool for wastewater flow modeling and inflow-infiltration simulations

Sanitary sewer systems are critical urban water infrastructure that protect both human and environmental health. Their design, operation, and monitoring require novel modeling techniques that capture dominant processes while allowing for computationally efficient simulations. Open water flow in sewers and rivers are intrinsically similar processes. With this in mind, we formulated a new parsimonious model inspired by the Width Function Instantaneous Unit Hydrograph (WFIUH) approach, widely used to predict rainfall-runoff relationships in watersheds, to a sanitary sewer system consisting of nearly 10,000 sewer conduits and 120,000 residential and commercial sewage connections in Northern Virginia, U.S.A. Model predictions for the three primary components of sanitary flow, including Base Wastewater Flow (BWF), Groundwater Infiltration (GWI), and Runoff Derived Infiltration and Inflow (RDII), compare favorably with the more computationally demanding industry-standard Storm Water Management Model (SWMM). This novel application of the WFIUH modeling framework should support a number of critical water quality endpoints, including (i) sewer hydrograph separation through the quantification of BWF, GWI, and RDII outflows, (ii) evaluation of the impact of new urban developments on sewage flow dynamics, (iii) monitoring and mitigation of sanitary sewer overflows, and (iv) design and interpretation of wastewater surveillance studies.

Noise in Expectations: Evidence from Analyst Forecasts

Abstract Analyst forecasts outperform econometric forecasts in the short run but underperform in the long run. We decompose these differences in forecasting accuracy into analysts’ information advantage, forecast bias, and forecast noise. We find that noise and bias strongly increase with forecast horizon, while analysts’ information advantage decays rapidly. A noise increase with horizon generates a mechanical reversal in the sign of the error-revision (Coibion-Gorodnichenko) regression coefficient at longer horizons, independently of over-/underreaction. A parsimonious model with bounded rationality and a noisy cognitive default matches the term structures of noise and bias jointly.

A time-course prediction model of global COVID-19 mortality.

The COVID-19 pandemic has caused over 6 million deaths worldwide and is a significant cause of mortality. Mortality dynamics vary significantly by country due to pathogen, host, social and environmental factors, in addition to vaccination and treatments. However, there is limited data on the relative contribution of different explanatory variables, which may explain changes in mortality over time. We, therefore, created a predictive model using orthogonal machine learning techniques to attempt to quantify the contribution of static and dynamic variables over time. A model was created using Partial Least Squares Regression trained on data from 2020 to rank order the significance and effect size of static variables on mortality per country. This model enables the prediction of mortality levels for countries based on demographics alone. Partial Least Squares Regression was then used to quantify how dynamic variables, including weather and non-pharmaceutical interventions, contributed to the overall mortality in 2020. Finally, mortality levels for the first 60 days of 2021 were predicted using rolling-window Elastic Net regression. This model allowed prediction of deaths per day and quantification of the degree of influence of included variables, accounting for timing of occurrence or implementation. We found that the most parsimonious model could be reduced to six variables; three policy-related variables - COVID-19 testing policy, canceled public events policy, workplace closing policy; in addition to three environmental variables - maximum temperature per day, minimum temperature per day, and the dewpoint temperature per day. Country and population-level static and dynamic variables can be used to predict COVID-19 mortality, providing an example of how broad temporal data can inform a preparation and mitigation strategy for both COVID-19 and future pandemics and assist decision-makers by identifying population-level contributors, including interventions, that have the greatest influence in mitigating mortality, and optimizing the health and safety of populations.

Long-term variability of retinal nerve fibre layer thickness measurement in patients with glaucoma of African and European descents

BackgroundTo examine long-term retinal nerve fibre layer thickness (RNFLT) variability and associated clinical factors in African (AD) and European descent (ED) individuals with glaucoma.MethodsThis retrospective cohort study included glaucoma eyes...

Predicting peripartum infection in laboring patients at high risk in Cameroon, Africa

ObjectiveTo develop a predictive model for peripartum infection among high risk laboring patients in Cameroon, Africa. Study designWe conducted a secondary analysis of the Cameroon Antibiotic Prophylaxis Trial (NCT 3248297), a multicenter 3-arm double-blind randomized controlled trial of oral azithromycin ± amoxicillin among term pregnancies with prolonged labor or rupture of membranes in Cameroon 1/2018–5/2020. Patients with chorioamnionitis prior to randomization, study drug contraindications, or planned cesarean were excluded. The outcome of interest was a composite of maternal peripartum infection (chorioamnionitis, endometritis, sepsis by World Health Organization criteria, wound infection/abscess) diagnosed up to 6 weeks postpartum. Potential predictors were compared between patients with and without the composite outcome, and evaluated at a 0.05 alpha level. Statistically significant exposures were analyzed using multivariable regression (to generate adjusted odds ratios and 95 % confidence intervals) with backwards selection to generate a parsimonious model. Receiver operating characteristic curves with associated area under the curve assessed the model’s predictive ability. A nomogram based on the final best fit multivariable model was constructed. ResultsOf 756 patients in the parent trial, 652 were analyzed: 45 (7 %) had peripartum infection. Those with infection were more likely to be nulliparous, lower education level, higher gestational age, receive antibiotics per hospital protocols, and undergo cesarean. In our best-fit multivariable model, none/primary education (vs university), cesarean birth, and antibiotic receipt per physician discretion (vs for cesarean prophylaxis) were significantly associated with increased infection risk. This model was moderately predictive (AUC = 0.75, 95 % CI 0.67–0.82). When using this 3 factor model, for a patient with a cesarean birth, receipt of antibiotics per physician discretion, and university education, the probability of peripartum infection was 35 % (95 % CI 0.11–0.73). ConclusionsWhile several variables such as parity are associated with infectious morbidity within 6 weeks among high risk laboring patients in Cameroon, only education level, antibiotic indication, and cesarean birth were independently associated, and a model including these 3 factors was moderately predictive. Validation of our findings in a larger population is warranted.

Modelling individual infancy growth trajectories to predict excessive gain in BMI z-score: a comparison of growth measures in the ABCD and GECKO Drenthe cohorts

BackgroundExcessive weight gain during childhood is a strong predictor for adult overweight, but it remains unknown which growth measures in infancy (0–2 years of age), besides predictors known at birth, are the strongest predictors for excessive weight gain between 2 and 5–7 years of age.MethodsThe Amsterdam Born Children and their Development (ABCD) study formed the derivation cohort, and the Groningen Expert Center for Kids with Obesity (GECKO) Drenthe study formed the validation cohort. Change (Δ) in body mass index (BMI) z-score between 2 and 5–7 years was the outcome of interest. The growth measures considered were weight, weight-for-length (WfL), and body mass index (BMI). Formats considered for each growth measure were values at 1, 6, 12, and 24 months, at the BMI peak, the change between aforementioned ages, and prepeak velocity. 10 model structures combining different variable formats and including predictors at birth were derived for each growth measure, resulting in 30 linear regression models. A Parsimonious Model considering all growth measures and a Birth Model considering none were also derived.ResultsThe derivation cohort consisted of 3139 infants of which 373 (11.9%) had excessive gain in BMI z-score (> 0.67). The validation cohort contained 2201 infants of which 592 (26.9%) had excessive gain. Across the 3 growth measures, 5 model structures which included measures related to the BMI peak and prepeak velocity (derivation cohort area under the curve [AUC] range = 0.765–0.855) achieved more accurate estimates than 3 model structures which included growth measure change over time (0.706–0.795). All model structures which used BMI were superior to those using weight or WfL. The AUC across all models was on average 0.126 lower in the validation cohort. The Parsimonious Model’s AUCs in the derivation and validation cohorts were 0.856 and 0.766, respectively, compared to 0.690 and 0.491, respectively, for the Birth Model. The respective false positive rates were 28.2% and 20.1% for the Parsimonious Model and 70.0% and 74.6% for the Birth Model.ConclusionModels’ performances varied significantly across model structures and growth measures. Developing the optimal model requires extensive testing of the many possibilities.

A predictive model for perinatal hypoxic ischemic encephalopathy using linked maternal and neonatal hospital data

PurposeTo build an evidence-based model to estimate case-specific risk of perinatal hypoxic ischemic encephalopathy. MethodsA retrospective, cross-sectional study of all births in Hawaii, Michigan, and New Jersey between 2010 and 2015, using linked maternal labor/delivery and neonatal birth records. Stepwise logistic regression and competitive Akaike information criterion were used to identify the most parsimonious model. Predictive ability of the model was measured with bootstrapped optimism-adjusted area under the ROC curve. ResultsAmong 836,216 births there were 376 (0.45 per 1000) cases of hypoxic ischemic encephalopathy. The final model included 28 variables, 24 associated with increased risk, and 4 that were protective. The optimism-adjusted area under the ROC curve was 0.84. Estimated risk in the study population ranged from 1 in ∼323,000 to 1 in 2.5. The final model confirmed known risk factors (e.g., sentinel events and shoulder dystocia) and identified novel risk factors, such as maternal race and insurance status. ConclusionOur study shows that risk of perinatal hypoxic ischemic encephalopathy injury can be estimated with high confidence. Our model fills a notable gap in the study of hypoxic ischemic encephalopathy prevention: the estimation of risk, particularly in the United States population which is unique with respect to racial and socioeconomic disparities.

Comparing implications of distinct emotion, network, and dimensional approaches for co-occurring emotions.

Many situations elicit multiple emotions at the same time. Therefore, emotion theories should explain when and how emotions co-occur. We compared four parsimonious, formal theories that could explain emotion co-occurrence inspired by distinct emotion, network, and dimensional approaches to emotions. In three studies (N = 1,038), diverse participants rated the intensity of awe and kama muta (Study 1; US community sample; conducted in 2020), shame and guilt (Study 2; Dutch students; conducted in 2006), or awe and fear (Study 3; GB community sample; conducted in 2022) on multi-componential scales in response to emotion-eliciting stimuli (i.e., vignettes or videos). All three studies indicated that the network theory explains the data best. The theory implies that emotions co-occur because the networks of interacting components representing different emotions partly overlap. The parsimonious model can serve as a starting point for a comprehensive formal theory that is able to integrate seemingly inconsistent findings in research on co-occurring emotions. (PsycInfo Database Record (c) 2023 APA, all rights reserved).

Examining Complex Relationships among Cerebrovascular, β‐Amyloid, Tau and Cognitive Measures in a Cognitively Unimpaired Late‐Middle Aged Sample

AbstractBackgroundRecent research linking vascular risk with promotion of tau pathology in the inferior temporal cortex (ITC) suggests a potential mechanism connecting cerebrovascular risk to AD pathology and cognitive decline among cognitively unimpaired (CU) individuals (Yau et al, 2022).We used MRI White Matter Hyperintensity (WMH), a neuroimaging marker of chronic cerebral small‐vessel disease/cerebrovascular disease, to further elucidate longitudinal relationships between cerebrovascular risk, PET AD pathological burden, and cognition.MethodWe examined baseline CU late middle‐aged adults from the University of Wisconsin’s Alzheimer’s disease research cohorts. β‐Amyloid burden was measured with 11C‐Pittsburgh compound B (“PiB”; DVR derived from a 0‐70 min dynamic PET scan; PET A+ threshold∼1.16 DVR/∼CL 17). Cerebrovascular risk was quantified using WMH volumes (T2‐FLAIR images; SPM Lesion Segmentation Tool). Tau burden was measured with florquinitau PET (MK6240 SUVR from 70‐90 min post‐injection). Tau outcomes included bilateral volume‐weighted averages for ITC, Medial temporal lobe (MTL) and Meta‐temporal (MTC) composites. Biomarker values were z‐transformed relative to CU/PET A‐ values. Random intercept models examined interactive effects of baseline WMH and Aβ levels on longitudinal tau (covarying baseline tau age, sex, APOE ε4 count). We modeled time two ways: years since baseline PiB (TimePiB1; replicating Yau, et al models); and, in parallel models, amyloid duration (years) at tau scans was substituted for years since PiB1 in the A+ subset (TimePiB1/PiB+). We also used LME to examine Aβ, WMH, and time interactions in longitudinal cognition (PACC3 and Digit Symbol Substitution).ResultThe mean(SD) tau baseline age was 67.9(6.8) years (sample details in Table1). The most parsimonious models using TimePiB1 included WMH*Aβ*Time for ITC and MTC and WMH*Aβ for MTL. Patterns showed a synergistic effect of WMH and Aβ on tau accumulation that was attenuated using TimePiB1/PiB+ where the most parsimonious models included Aβ_status*TimePiB1/PiB+ (ITC, MTC) or WMH*Aβ_status*TimePiB1/PiB+ (MTL; Figure1). AICc model fits were substantially better using TimePiB1/PiB+. Both Aβ and WMH contributed to cognitive trajectories (Figure2).ConclusionAlthough cerebrovascular risk’s apparent contribution to tau accumulation varied depending on how time was modeled, results suggest that cerebrovascular risk may complement imaging biomarkers in assessing risk of tau accumulation and cognitive decline in preclinical Alzheimer disease.

Machine-learning vs. logistic regression for preoperative prediction of medical morbidity after fast-track hip and knee arthroplasty—a comparative study

BackgroundMachine-learning models may improve prediction of length of stay (LOS) and morbidity after surgery. However, few studies include fast-track programs, and most rely on administrative coding with limited follow-up and information on perioperative care. This study investigates potential benefits of a machine-learning model for prediction of postoperative morbidity in fast-track total hip (THA) and knee arthroplasty (TKA).MethodsCohort study in consecutive unselected primary THA/TKA between 2014–2017 from seven Danish centers with established fast-track protocols. Preoperative comorbidity and prescribed medication were recorded prospectively and information on length of stay and readmissions was obtained through the Danish National Patient Registry and medical records. We used a machine-learning model (Boosted Decision Trees) based on boosted decision trees with 33 preoperative variables for predicting “medical” morbidity leading to LOS > 4 days or 90-days readmissions and compared to a logistical regression model based on the same variables. We also evaluated two parsimonious models, using the ten most important variables in the full machine-learning and logistic regression models. Data collected between 2014–2016 (n:18,013) was used for model training and data from 2017 (n:3913) was used for testing.Model performances were analyzed using precision, area under receiver operating (AUROC) and precision recall curves (AUPRC), as well as the Mathews Correlation Coefficient. Variable importance was analyzed using Shapley Additive Explanations values.ResultsUsing a threshold of 20% “risk-patients” (n:782), precision, AUROC and AUPRC were 13.6%, 76.3% and 15.5% vs. 12.4%, 74.7% and 15.6% for the machine-learning and logistic regression model, respectively. The parsimonious machine-learning model performed better than the full logistic regression model. Of the top ten variables, eight were shared between the machine-learning and logistic regression models, but with a considerable age-related variation in importance of specific types of medication.ConclusionA machine-learning model using preoperative characteristics and prescriptions slightly improved identification of patients in high-risk of “medical” complications after fast-track THA and TKA compared to a logistic regression model. Such algorithms could help find a manageable population of patients who may benefit most from intensified perioperative care.

Prognostic algorithms for post-discharge readmission and mortality among mother-infant dyads: an observational study protocol.

In low-income country settings, the first six weeks after birth remain a critical period of vulnerability for both mother and newborn. Despite recommendations for routine follow-up after delivery and facility discharge, few mothers and newborns receive guideline recommended care during this period. Prediction modelling of post-delivery outcomes has the potential to improve outcomes for both mother and newborn by identifying high-risk dyads, improving risk communication, and informing a patient-centered approach to postnatal care interventions. This study aims to derive post-discharge risk prediction algorithms that identify mother-newborn dyads who are at risk of re-admission or death in the first six weeks after delivery at a health facility. This prospective observational study will enroll 7,000 mother-newborn dyads from two regional referral hospitals in southwestern and eastern Uganda. Women and adolescent girls aged 12 and above delivering singletons and twins at the study hospitals will be eligible to participate. Candidate predictor variables will be collected prospectively by research nurses. Outcomes will be captured six weeks following delivery through a follow-up phone call, or an in-person visit if not reachable by phone. Two separate sets of prediction models will be built, one set of models for newborn outcomes and one set for maternal outcomes. Derivation of models will be based on optimization of the area under the receiver operator curve (AUROC) and specificity using an elastic net regression modelling approach. Internal validation will be conducted using 10-fold cross-validation. Our focus will be on the development of parsimonious models (5-10 predictor variables) with high sensitivity (>80%). AUROC, sensitivity, and specificity will be reported for each model, along with positive and negative predictive values. The current recommendations for routine postnatal care are largely absent of benefit to most mothers and newborns due to poor adherence. Data-driven improvements to postnatal care can facilitate a more patient-centered approach to such care. Increasing digitization of facility care across low-income settings can further facilitate the integration of prediction algorithms as decision support tools for routine care, leading to improved quality and efficiency. Such strategies are urgently required to improve newborn and maternal postnatal outcomes. https://clinicaltrials.gov/, identifier (NCT05730387).

Influence of model complexity, training collinearity, collinearity shift, predictor novelty and their interactions on ecological forecasting

AbstractAimEcological forecasting is critical in understanding of ecological responses to climate change and is increasingly used in climate mitigation plans. The forecasts from correlative models can be challenged by model complexity, training collinearity, collinearity shift and novel conditions of predictors that are common during model extrapolation. The individual effect of these four factors has been investigated, but it is still unclear how these four factors interactively affect forecasting. To fill this gap, we conducted a comprehensive simulation experiment to quantify how the four factors interactively influence model forecasting.LocationSimulated regions.Time PeriodSimulated scenarios.MethodsWe modelled three response variables commonly used in ecological forecasting following normal, Poisson and binomial distributions as a function of three functional relationships that represented model complexity under three levels of training collinearity using generalized linear models. By calculating prediction error under 3,780,000 testing scenarios, we partitioned its variance to model complexity, training collinearity, collinearity shift, predictor novelty and their interactions.ResultsWe found that increased predictor novelty and collinearity shift degraded model performance, leading up to double prediction errors when a predictor's range increased by ~22% or when the correlation r between two predictors changed >~0.8 for the combination of high training collinearity and interaction functional relationship. Predictor novelty reduced the influence of collinearity shift on model performance, suggesting a negative interaction between them. This pattern was more pronounced under high model complexity and high training collinearity.Main ConclusionsThe accuracy of ecological forecasting using correlative models depends on model complexity, training collinearity, collinearity shift, predictor novelty and their interactions. Besides the consideration of parsimonious models and r of 0.7 in model training, our study further recommends a threshold of <22%–50% increased predictor range depending on training collinearity and/or <0.8 correlation change for making reliable forecasting.

Fake news and youngsters’ decision journey: An evaluation of the influence of misinformation on social media

Purpose: Social media has changed the way users interact with each other, and has become an important part of numerous lives. However, there is an increasing flow of implausible content circulating on social media, which points to the need for some categorization and regulation. This study will examine how the proliferation of fake news on social media impacts students and their choice of university. To answer this question, market research was conducted on the precedents that affect the acceptance of fake news among university students when choosing to study for a master's degree that will help them in their professional careers.Design/methodology/approach: The study used a quantitative method. A parsimonious model of causal relationships was proposed based on scales taken from the literature, assessed by a convenience sample of students, and adjusted by structural equation modelling (SEM).Findings: Results show that the parsimonious model explains 35% of fake news acceptance and that media dependency (ISMD) and parasocial interaction (PSI) are the main direct effects, while perceived media richness (PR) has a significant indirect influence on the attitude towards fake news and, consequently, on its acceptance. Furthermore, fake news literacy plays a correct moderating role with the most relevant source of influence, SNS dependency.Research limitations: A convenience sample was used, and a parsimonious model with three antecedent factors and one mediating factor was proposed. Other social factors could have been considered, including multicultural variables.Practical implications: The results point to students' expressed dependence on social networks as the main factor explaining their attitude towards fake news, negatively moderated by students' level of knowledge about the importance of this phenomenon in social networks. Therefore, it is relevant to promote knowledge about this phenomenon among students to reduce its influence on decision-making processes.Originality/value: This paper provides a novel context for the study of the proliferation of fake news on social networks: the process of choosing a university by students addicted to the news circulating on social media.

Time to Treatment for Patients with Suspected Acute Promyelocytic Leukemia (APL)

Introduction: Acute promyelocytic leukemia (APL) is a rare subtype of acute myeloid leukemia characterized by dysplastic promyelocytes which cause aberrations in coagulation, including hemorrhage and disseminated intravascular coagulation (DIC), leading to a rapidly fatal course if treatment is not initiated promptly. Early administration of all-trans retinoic acid (ATRA) in patients suspected of having APL is crucial for reducing morbidity and mortality. However, APL still exhibits a high early death rate in part due to failure to rapidly initiate ATRA. Therefore, we conducted a retrospective study to investigate the time it takes for ATRA to be ordered and administered to patients with suspected APL at our tertiary care University hospital. We also examined demographic and clinical variables to identify any opportunities for quality improvement. Methods: We queried our electronic medical record (EMR) for all adult patients without a prior diagnosis of APL from 2017-2021 who received at least one dose of ATRA at Thomas Jefferson University Hospital. Our analysis included 91 patients. We collected a wide range of demographic and clinical data (Table 1). The time of admission was defined as the time the admission order was placed in the EMR, and the time of administration was defined as the time the nurse marked the medication as administered. A simple general linear model was used to evaluate univariate association between each single covariate and each log transformed outcome. Model selection by Akaike information criterion (AIC) was conducted to identify the best model according to goodness of fit criteria. The smaller the AIC value, the better the model fit. The parsimonious models were obtained using backward elimination of non-significant predictors (p-values > 0.05) from the best models according to goodness of fit criteria (AIC). Results: The mean time for interval A (time from hospital admission to ordering ATRA) was 284 minutes, the mean time for interval B (time from ATRA order being placed to ATRA administration) was 104 minutes, and the mean time for interval C (time from admission to ATRA administration) was 388 minutes. Further analysis of the time intervals is presented in Table 2. The results of the demographic and clinical variables (Table 1) show that the majority of patients were transferred from an outside hospital (89%) and admitted overnight (87%). Consequently, ATRA was most often ordered (76%) and administered (70%) overnight. Only 26 (29%) patients were ultimately confirmed to have APL. During the index admission, patients required a mean of 6 units of packed red blood cells and 9 units of platelets. A quarter of patients were noted to have signs or symptoms of bleeding upon admission. The results of the multiple linear regression analysis are shown in Table 2. Only signs or symptoms of bleeding were associated with a statistically significant decrease in time to ATRA. Patients with signs of symptoms of bleeding experienced, on average, 39% less time from admission to ATRA being ordered (95% CI: 0.396-0.946, p=0.030) and 31% shorter time from admission to ATRA administration compared to those without signs or symptoms of bleeding (95% CI: 0.489-0.967, p=0.034). Conclusions: Our study highlights that, even at a large, resource-rich, academic medical center, the time to ATRA administration in patients with suspected APL is longer than optimal. The presence of bleeding was the only variable that significantly decreased the time to ATRA. Earlier identification of bleeding risk (including evolving DIC or thrombocytopenia), increased recognition of possible acute leukemia, and more rapid hematology/oncology consultation could help to decrease the time to ATRA and to avoid bleeding complications before they start. Additionally, we found that most patients (89%) were transferred from an outside hospital. Although we were not able to quantify the time from outside hospital admission to transfer, we suspect that the time was substantial and that there are even greater opportunities to improve the care of these patients by increasing ATRA availability at community hospitals, enhancing the education of all front-line providers regarding the prompt recognition of coagulopathy in the setting of suspected acute leukemia, and improving transfer center acceptance documentation that emphasizes the concern for APL and appropriate next steps in evaluation and treatment.

The Molecular Prognostic Scoring System for Normal Karyotype Myelodysplastic Syndromes

Purpose: Myelodysplastic syndromes (MDS), is characterized by clinical heterogeneity and varying prognoses. Reliable prognostic models are essential to stratify MDS patients and optimize clinical management. Although incorporating molecular information into MDS prognostic models improves their predictive ability, limitations of next-generation sequencing panels and potential heterogeneity among MDS patients with different karyotypes hinder their clinical translation. Methods: In total, 237 normal karyotype (NK) MDS patients from the First Affiliated Hospital of Zhejiang University School of Medicine were used as the derivation set to investigate the molecular frequency and prognostic spectrum. A parsimonious molecular-clinical model was developed with overall survival (OS) as the endpoint and compared with other models using the concordance index as the primary evaluation index. An external independent validation set from the International Working Group for Prognosis in MDS cohort(691 NK MDS patients) was used to evaluate the model generalization ability. Results: In NK MDS, the gene mutation frequency spectrum is similar to the overall MDS population, the differences mainly lie in the prognostic spectrum. Genes such as TP53, previously associated with MDS prognoses have lost their prognostic value. Therefore, a seven-parameter molecular-clinical prognostic model called the Molecular Prognostic Scoring System for NK MDS (NK-PSS-M) was developed based on a specific molecular prognostic spectrum of NK MDS. The model includes age, hemoglobin, platelet count, blast percentage, CEBPA, RUNX1, and U2AF1. The NK-PSS-M model allows for personalized risk assessment based on patient characteristics, with higher scores indicating a greater risk and poorer prognosis. In addition, we developed a five-class classification system based on these scores, which differentiates patients across risk categories and predicts outcomes, such as OS and LFS(Figure 1 A and B). The performance of NK-PSS-M was comparable to that of the Molecular International Prognostic Scoring System(IPSS-M), and it significantly outperformed IPSS-R and other molecular prediction models based on it. The above results were confirmed in the external independent validation cohort(Figure 1C). Our transcriptomics analysis revealed that the low-risk group defined by NK-PSS-M demonstrated significant activation of the inflammatory pathway, and immune activation, which is consistent with the current understanding of the immunological landscape of MDS. Conclusion: The NK-PSS-M model improved the risk stratification of non-molecular models and provided a reliable alternative to the IPSS-M. In addition, we demonstrated the molecular heterogeneity of MDS across different cytogenetics and emphasized the importance of developing cytogenetics-specific prognostic models.

The Post-CAR Prognostic Index (PC-PI): A New Prognostic Score to Predict Overall Survival in Large B-Cell Lymphoma Patients Progressing after Chimeric Antigen Receptor T-Cell Therapy

Introduction: Chimeric antigen receptor (CAR) T-cell therapy fails to achieve durable responses in 60% of relapsed/refractory (R/R) large B-cell lymphoma (LBCL) patients in the third or later line setting. There is no standard treatment after CAR T-cell therapy progression and a wide range of outcomes are observed in this patient population. Besides the interval between CAR T-cell infusion and progressive disease (PD), data regarding prognostic factors at time of progression are scarce. Our aim was to develop a new prognostic tool to predict overall survival (OS) after CAR T-cell therapy progression with easily-available markers from routine clinical practice. Methods: First, we performed a retrospective data collection at 12 Spanish centers of R/R LBCL patients who progressed after CAR T-cell therapy in the third or later line setting from September 2018 until June 2022 (training cohort, TC). We analyzed a total of 15 variables, including pre-CAR T-cell therapy characteristics (gender, histology, primary refractory disease, previous hematopoietic transplant, number of previous lines) and values collected at time of progression to CAR T-cells (age, stage, extranodal sites, ECOG, hemoglobin, neutrophils, platelets, LDH, best response to CAR T-cells, time from CAR T-cell infusion to PD). The primary endpoint was OS from date of progression to CAR T-cell therapy. A stratified Cox model was used to estimate hazard ratios (HRs) using post-progression treatment as a stratification factor. We used LASSO regression with minimum lambda to identify which variables had the highest prognostic impact on OS. Additionally, the factors with a lower contribution were eliminated to create a parsimonious model. The C-statistic was used to evaluate its discrimination. We examined the performance of the International Prognostic Index (IPI) score and Revised IPI (R-IPI) in this setting. Finally, we tested the score in an external validation cohort (VC) which included a comparable patient population from 3 European countries. Results: Among the 216 LBCL patients included in the TC, most were male (66%), had an ECOG of 1 (48%) and stage IV disease (71%) at time of CAR T-cell therapy progression. Median time from CAR T-cell infusion to PD was 2.5 months (95CI% 1.9-2.9) and median follow-up from progression was 15 months. Salvage treatment was classified into 3 subgroups, including immunotherapy or targeted agents (43%), chemotherapy or radiotherapy (20%) and palliative care (38%). To build the prognostic score, a total of 5 variables were selected. Each marker received 1 point (given the similar HR [1.48-1.77]), if they met defined criteria: ECOG (>0), hemoglobin (<10 g/dL), LDH (>2 x upper normal limit), number of extranodal sites (>1) and time from CAR T-cell infusion to PD (<4 months). Patients with 0-1 points were classified as low risk, 2 points as intermediate-low risk, 3 points as intermediate-high risk and 4 or 5 points as high risk. In the TC, the 4 risk groups showed statistically significant differences in OS (Figure 1). In the low-risk group (n=39 [18%]), the median OS [mOS] was not reached; in the intermediate-low risk (n=56 [26%]) mOS was 7.3 months (HR=2.89, p=0.002); in the intermediate-high risk (n=57 [26%]) mOS was 4.9 months (HR=4.81, p<0.001) and in the high risk (n=64 [30%]) mOS was 1.8 months (HR=6.69, p<0.001). In terms of post-relapse therapies, both the chemo/radiotherapy and the immunotherapy groups showed a balanced patient distribution, from low to high risk (32%, 35%, 14% and 19% vs 22%, 32%, 30% and 16%, respectively). The VC included 204 patients with a similar patient distribution in the 4 prognostic risk groups (35%, 25%, 12%, 27%). The mOS for each of these groups was 15.2, 5.3, 2.9 and 0.9 months, respectively. Each group had distinct OS outcomes when compared with all the other risk groups (p<0.05 for each comparison) (Figure 2). Finally, our model presented a C-index of 0.712 for the TC and 0.811 for the VC, outperforming both the IPI (0.647 [TC] and 0.691 [VC]) and R-IPI (0.632 [TC] and 0.683 [VC]). Conclusions: The Post-CAR Prognostic Index (PC-PI) is a clinically useful tool for OS prediction and risk-adapted treatment planning in LBCL patients progressing after CAR T-cell therapy. In addition, our results will help stratification in clinical trials which include patients with prior CAR T-cell therapy.

Strong phylogenetic signal and models of trait evolution evidence phylogenetic niche conservatism for seagrasses

Abstract Phylogenetic signal (PS) is the propensity of closely related species to resemble each other. PS has been tested across clades of terrestrial plants; however, insight for seagrasses is still lacking. Signatures of PS and models of niche (trait) evolution can help to detect phylogenetic niche conservatism (PNC), that is, close relatives live in comparable niches. The initial goal of this study was to assess the pattern of PS for the world's seagrasses, by testing the non‐independence of phylogenetic relatedness and seagrass species traits. A phylogeny of 49 seagrasses was constructed, together with a matrix of nine traits covering morphological, life‐history and reproductive attributes. PS of traits was tested through complementary indices (Pagel's λ, Blomberg's K, Moran's I and Abouheif's Cmean). Three models of niche evolution (Brownian Motion, BM; Ornstein–Uhlenbeck, OU; Early Burst, EB) were then fitted to each trait and the multivariate trait matrix. Results supported the existence of strong PS for seagrasses, with a particularly large effect size for seagrass reproductive traits, which followed an EB evolution model. Local Indicators of Phylogenetic Association (local autocorrelation metrics that can help to identify areas of large autocorrelation) supported the presence of PS across seagrass lineages/clades, supported by the dominance of OU as the most parsimonious trait evolution model. The pattern of strong PS seems to be a consequence of long‐term PNC of seagrass traits after initial radiation. Synthesis: Our study highlights the relevance of evolution from common ancestors and shared history underpinning large seagrass phylogenetic structuring.