Complete Case Analysis Research Articles

Youth Preseason Performance on the Sport Concussion Assessment Tool 5 Across Multiple Sports.

To examine preseason Sport Concussion Assessment Tool 5 (SCAT5) performance of adolescent sport participants by environment (in-person/virtual), sex, age, concussion history, collision/noncollision sport participation, and self-reported medical diagnoses. Cross-sectional. Canadian community and high-school sport settings. Three thousand eight hundred five adolescent (2493 male, 1275 female, and 37 did not disclose; 11- to 19-year-old) sport participants. Sport Concussion Assessment Tool 5 administration method (in-person/virtual), sex (male/female/unreported), age (years), concussion history (0/1/2/3+), collision/noncollision sport participant, and self-reported medical diagnoses [attention deficit disorder or attention-deficit/hyperactivity disorder, headache/migraine, learning disability, and psychiatric disorder (ie, anxiety/depression/other)]. Preseason SCAT5 outcomes including total number of symptoms (TNS; /22), symptom severity score (SSS; /132), Standardized Assessment of Concussion (SAC; /50), and modified Balance Error Scoring System (mBESS; /30). Multiple multilevel linear or Poisson regression complete case analyses adjusting for clustering and robust standard errors, with β-coefficients (95% CI) back-transformed to indicate an increase/decrease in SCAT5 subdomains when relevant for clinical interpretation. Virtual (V) performance was associated with fewer symptoms reported [TNS Difference V-IP = -1.53 (95% CI, -2.22 to -0.85)], lower SSS [-2.49 (95% CI, -4.41 to -0.58)], and fewer mBESS errors (IP) [-0.52 (95% CI, -0.77 to -0.27)] compared with in-person. For every one-year increase in age, more symptoms [TNS = 0.22 (95% CI, 0.01-0.44)], higher SSS [0.52 (95% CI, 0.01-1.06)], higher SAC [0.27 (95% CI, 0.15-0.38), and poorer balance [mBESS = -0.19 (-0.28 to -0.09)] were observed. Differences between males and females were also seen across all SCAT5 outcomes. Individuals reporting any medical diagnosis or 3+ concussion history also reported more symptoms (TNS) and higher SSS than those who did not. Administration environment, sex, age, concussion history, and medical diagnoses were associated with SCAT5 subdomains and are important considerations when interpreting the SCAT5 results.

INVESTIGATING THE RELATIONSHIP BETWEEN HABITUAL ACTIVITY AND COGNITIVE DOMAINS AMONG PEOPLE LIVING WITH DEMENTIA

Abstract Habitual activity preserves cognitive function in people without dementia, but the relationship between habitual activity and cognitive domains among people living with dementia is unclear. The purpose of this study was to explore the association between habitual activity and cognition domains among people living with dementia. Participants living with dementia in residential care facilities (complete case analysis: n=24/42) completed a battery of cognitive tests (global cognition: Montreal Cognitive Assessment; executive function: Trail-Making Test, Digit Span Forward Test; perception and orientation: Benton Judgement of Line Orientation Test; language: Boston Naming Test; learning and memory: Rey Auditory Verbal Learning Test; complex attention: Digit Symbol Substitution Test). Participants wore an actigraphy monitor on their non-dominant wrist over seven days. We conducted a linear regression for total habitual activity (independent variable) with race (white/black), fall risk (Morse Fall Scale), and the number of comorbidities (Functional Comorbidities Index) as covariates, and cognitive tests as variables of interest. Participants were primarily male (75%), white (87.5%), and 50% had unspecified dementia (Alzheimer’s disease: 33%). Greater habitual activity was associated with poorer global cognition, better executive function, and better learning and memory (ps< 0.05). Habitual activity was not related to visuospatial perception, language, or complex attention. Habitual activity may preserve executive function and learning and memory among people living with dementia. Wandering is more common in later stages of dementia, which may explain greater habitual activity observed with lower global cognition. Regularly assessing habitual activity may be useful in screening and monitoring cognitive changes.

Properties of the full random-effect modeling approach with missing covariate data.

During drug development, a key step is the identification of relevant covariates predicting between-subject variations in drug response. The full random effects model (FREM) is one of the full-covariate approaches used to identify relevant covariates in nonlinear mixed effects models. Here we explore the ability of FREM to handle missing (both missing completely at random (MCAR) and missing at random (MAR)) covariate data and compare it to the full fixed-effects model (FFEM) approach, applied either with complete case analysis or mean imputation. A global health dataset (20421 children) was used to develop a FREM describing the changes of height for age Z-score (HAZ) over time. Simulated datasets (n = 1000) were generated with variable rates of missing (MCAR) covariate data (0%-90%) and different proportions of missing (MAR) data condition on either observed covariates or predicted HAZ. The three methods were used to re-estimate model and compared in terms of bias and precision which showed that FREM had only minor increases in bias and minor loss of precision at increasing percentages of missing (MCAR) covariate data and performed similarly in the MAR scenarios. Conversely, the FFEM approaches either collapsed at 70% of missing (MCAR) covariate data (FFEM complete case analysis) or had large bias increases and loss of precision (FFEM with mean imputation). Our results suggest that FREM is an appropriate approach to covariate modeling for datasets with missing (MCAR and MAR) covariate data, such as in global health studies.

Understanding the Implications of a Complete Case Analysis for Regression Models with a Right-Censored Covariate

Despite its drawbacks, the complete case analysis is commonly used in regression models with incomplete covariates. Understanding when the complete case analysis will lead to consistent parameter estimation is vital before use. Our aim here is to demonstrate when a complete case analysis is consistent for randomly right-censored covariates and to discuss the implications of its use even when consistent. Across the censored covariate literature, different assumptions are made to ensure a complete case analysis produces a consistent estimator, which leads to confusion in practice. We make several contributions to dispel this confusion. First, we summarize the language surrounding the assumptions that lead to a consistent complete case estimator. Then, we show a unidirectional hierarchical relationship between these assumptions, which leads us to one sufficient assumption to consider before using a complete case analysis. Lastly, we conduct a simulation study to illustrate the performance of a complete case analysis with a right-censored covariate under different censoring mechanism assumptions, and we demonstrate its use with a Huntington disease data example.

How to enhance the applicability of a risk prediction model for term small-for-gestational-age neonates in clinical settings?

To construct a simple term small-for-gestational-age (SGA) neonate prediction model that is clinically practical. This analysis was based on the Born in Guangzhou Cohort Study (BIGCS). Mothers who had a singleton pregnancy, delivered a term neonate, and had an ultrasonography within 30 + 0 to 32 + 6 weeks of gestation were included. Term SGA was defined with customized population percentiles. Prediction models were constructed with backward selection logistic regression in a four-step approach, where model 1 contained fetal biometrics only, models 2 and 3 included maternal features and a time factor (weeks between ultrasonography and delivery), respectively; and model 4 contained all features mentioned. The prediction performance of individual models was evaluated based on area under the curve (AUC) and a calibration test was performed. The prevalence of SGA in the study population of 21 346 women was 11.5%. With a complete-case analysis approach, data of 19 954 women were used for model construction and validation. The AUC of the four models were 0.781, 0.793, 0.823, and 0.834, respectively, and all were well-calibrated. Model 3 consisted of fetal biometrics and corrected for time to delivery was chosen as the final model to build risk prediction graphs for clinical use. A prediction model derived from fetal biometrics in early third trimester is satisfactory to predict SGA.

Implementation and engagement of the SMART Work & Life sitting reduction intervention: an exploratory analysis on intervention effectiveness

BackgroundTo enhance the impact of interventions, it is important to understand how intervention engagement relates to study outcomes. We report on the level of implementation and engagement with the SMART Work & Life (SWAL) programme (delivered with (SWAL plus desk) and without a height-adjustable desk (SWAL)) and explore the effects of different levels of this on change in daily sitting time in comparison to the control group.MethodsThe extent of intervention delivery by workplace champions and the extent of engagement by champions and participants (staff) with each intervention activity was assessed by training attendance logs, workplace champion withdrawal dates, intervention activities logs and questionnaires. These data were used to assess whether a cluster met defined criteria for low, medium, or high implementation and engagement or none of these. Mixed effects linear regression analyses tested whether change in sitting time varied by: (i) the number of intervention activities implemented and engaged with, and (ii) the percentage of implementation and engagement with all intervention strategies.ResultsWorkplace champions were recruited for all clusters, with 51/52 (98%) attending training. Overall, 12/27 (44.4%) SWAL and 9/25 (36.0%) SWAL plus desk clusters implemented all main intervention strategies. Across remaining clusters, the level of intervention implementation varied. Those in the SWAL (n = 8 (29.6%) clusters, 80 (32.1%) participants) and SWAL plus desk (n = 5 (20.0%) clusters, 41 (17.1%) participants) intervention groups who implemented and engaged with the most intervention strategies and had the highest percentage of cluster implementation and engagement with all intervention strategies sat for 30.9 (95% CI -53.9 to -7.9, p = 0.01) and 75.6 (95% CI -103.6 to -47.7, p < 0.001) fewer minutes/day respectively compared to the control group at 12 month follow up. These differences were larger than the complete case analysis. The differences in sitting time observed for the medium and low levels were similar to the complete case analysis.ConclusionsMost intervention strategies were delivered to some extent across the clusters although there was large variation. Superior effects for sitting reduction were seen for those intervention groups who implemented and engaged with the most intervention components and had the highest level of cluster implementation and engagement.Trial RegistrationISRCTN11618007. Registered on 24 January 2018. https://www.isrctn.com/ISRCTNISRCTN11618007.

Multiple Imputation of Missing Race/Ethnicity Information in the National Assisted Reproductive Technology Surveillance System.

Background: Missing race/ethnicity data are common in many surveillance systems and registries, which may limit complete and accurate assessments of racial and ethnic disparities. Centers for Disease Control and Prevention's National Assisted Reproductive Technology (ART) Surveillance System (NASS) has a congressional mandate to collect data on all ART cycles performed by fertility clinics in the United States and provides valuable information on ART utilization and treatment outcomes. However, race/ethnicity data are missing for many ART cycles in NASS. Materials and Methods: We multiply imputed missing race/ethnicity data using variables from NASS and additional zip code-level race/ethnicity information in U.S. Census data. To evaluate imputed data quality, we generated training data by imposing missing values on known race/ethnicity under missing at random assumption, imputed, and examined the relationship between race/ethnicity and the rate of stillbirth per pregnancy. Results: The distribution of imputed race/ethnicity was comparable to the reported one with the largest difference of 0.53% for non-Hispanic Asian. Our imputation procedure was well calibrated and correctly identified that 89.91% (standard error = 0.18) of known race/ethnicity values on average in training data. Compared to complete-case analysis, using multiply imputed data reduced bias of parameter estimates (the range of bias for stillbirth per pregnancy across race/ethnicity groups is 0.02%-0.18% for imputed data analysis, versus 0.04%-0.66% for complete-case analysis) and yielded narrower confidence intervals. Conclusions: Our results underscore the importance of collecting complete race/ethnicity information for ART surveillance. However, when the missingness exists, multiply imputed race/ethnicity can improve the accuracy and precision of health outcomes estimated across racial/ethnic groups.

Dealing with missing observations in the outcome and covariates in randomized controlled trials

This article compares different missing data methods in randomized controlled trials, specifically addressing cases involving joint missingness in the outcome and covariates. In the existing literature, it is still unclear how advanced methods like linear mixed model (LMM) and multiple imputation (MI) perform in comparison to simpler methods regarding the estimation of treatment effects and their standard errors. We therefore evaluates the performance of LMM and MI against simple alternatives across a wide range of simulation scenarios for various realistic missingness mechanisms. The results show that no single method universally outperforms the others. However, LMM followed by MI demonstrates superior performance across most missingness scenarios. Interestingly, a simple method that combines complete case analysis for the missing outcome and mean imputation for the missing covariate (CCAME) performs similarly to LMM and MI. All methods are furthermore compared in the context of a randomized controlled trial on chronic obstructive pulmonary disease.

Impact of Missing Data on Correlation Coefficient Values: Deletion and Imputation Methods for Data Preparation

The correlation coefficient is one of the essential statistical techniques used to discover relationships among variables. Various techniques can quantify correlation, such as Pearson's, Spearman's, and Kendall's correlation coefficients, depending on the data type. As with any use of data, missing data will impact the availability of data, reducing it and potentially affecting the results. Furthermore, the removal of missing-value data from the study when using complete case analysis or available case analysis may result in selection biases. In this paper, we investigate the impact of missing data on the correlation coefficient value by calculating the difference between the correlation coefficient of the original complete dataset and that of a dataset with missing data. Two deletion strategies (Listwise and Pairwise) and three imputation strategies (Mean, k-Nearest Neighbors (k-NN), and Expectation-Maximization) were used to prepare the data before calculating the correlation coefficient. Unique correlation coefficient values were created by converting unique values to a one-dimensional array, and RMSE metrics were used to evaluate the experiments. Eight UCI and Kaggle datasets with different sizes and numbers of attributes were used in this study. The experiment results demonstrate that the Pairwise strategy and k-NN give good results on the correlation coefficient, respectively, when the missing rate is moderate or less. Pairwise uses all the available values and discards only the missing values of the related attribute, while k-NN fills the missing values with new values that produce correlation coefficient values close to the actual values.

Cost-effectiveness of Spironolactone for Adult Female Acne (SAFA): economic evaluation alongside a randomised controlled trial

ObjectiveThis study aims to estimate the cost-effectiveness of oral spironolactone plus routine topical treatment compared with routine topical treatment alone for persistent acne in adult women from a British NHS...

Predictive Factors for Infusion Site Induration After Outpatient Chemotherapy in Japan: A Secondary Analysis

Even in the absence of extravasation, some patients develop induration at the peripheral intravenous catheterization site prior to the next day's treatment. Infusion site induration commonly affects patients who undergo repeated chemotherapy administrations. Vessel health is crucial for the continuation of chemotherapy. However, there is no effective method to prevent induration. Hence, this study aimed to investigate the factors that could cause induration for preventing its occurrence. This study was a secondary analysis of a prospective observational study. All participants were undergoing outpatient chemotherapy. Participant characteristics and related catheterization data were collected on the treatment day as baseline, and induration incidence was recorded on the subsequent treatment day. Receiver operating characteristic (ROC) analysis was performed to determine the sensitivity and specificity of cutoff points of the vein and catheter diameter ratios for distinguishing between developed induration and not developed induration. Additionally, cox regression analysis with multiple imputation was used to investigate the factors that predicted induration. Seventy-one patients participated in the study. The cutoff point of the vein/catheter diameter ratio calculated using ROC analysis was ≥3.7. The ratio of larger-diameter veins to catheter diameter of ≥3.7 times was negatively associated with induration in both complete case analysis (HR: 0.11; p=0.034) and multiple imputation analysis (HR: 0.12; p=.049). Selecting the vein with 3.7 times higher diameter than the catheter diameter for the catheterization site may help prevent induration on the next treatment day.

Review on data imputation methods in machine learning

Data is an important element in the analysis of machine learning. It is usually measured based on observations and is also an indispensable element in training a model. Good preparation of data helps enhance the performance of analysis and is able to deliver reliable final results. However, lots of factors influence the dataset and some lead to the loss of some data. When some portion of the data is missing, it causes biases in the final prediction outcomes. In order to minimize the consequences of missing data, several data imputation methods are established to solve the problem. This paper will first talk about some basic concepts about missing data. In the following sections, the paper will present several popular data imputation methods, including complete case analysis, single imputation, and multiple imputations. Applications of some methods will be presented to see how they can be used in real analysis situations. Finally, the paper will talk about the limits of current data imputation methods.

Bayesian Network structure learning algorithm for highly missing and non imputable data: Application to breast cancer radiotherapy data

It is not uncommon for real-life data produced in healthcare to have a higher proportion of missing data than in other scopes. To take into account these missing data, imputation is not always desired by healthcare experts, and complete case analysis can lead to a significant loss of data. The algorithm proposed here, allows the learning of Bayesian Network graphs when both imputation and complete case analysis are not possible. The learning process is based on a set of local bootstrap learnings performed on complete sub-datasets which are then aggregated and locally optimized. This learning method presents competitive results compared to other structure learning algorithms, whatever the mechanism of missing data.

Developing an AI-predictive model for predicting the likelihood of post-operative infection in surgical patients

Abstract Introduction Antimicrobial resistance was listed among the top ten global health issues by the World Health Organization in 2019.1 By 2050, it is projected to have caused 10 million deaths, surpassing cancer, diabetes, and road traffic accidents.2 The rise in resistant bacterial strains and reduced availability of novel antibiotics makes tackling antimicrobial resistance even more challenging. Preventing infections is the most effective approach, and Artificial Intelligence (AI) has the potential to help predict the likelihood of a patient experiencing a post-operative infection.3 This can help clinicians identify those who might be at higher risk and implement clinical interventions to potentially reduce this risk, during the pre-operative stage. Aim To develop and internally validate an AI-predictive model for predicting the likelihood of post-operative infection in surgical patients. Methods A secondary care dataset from the Pioneer Research Hub was used to develop the predictive model, which included a range of data such as demographic information, vital signs, microbiological culture results, lab investigations, comorbidities, surgical details, and infection diagnoses. Machine learning techniques and clinician input were used to select predictors, both modifiable (e.g. blood glucose) and non-modifiable (e.g. age). Recursive feature elimination with cross-validation was performed, and the dataset was split into training (80%) and internal validation (20%) sets. Missing data was managed through complete case analysis and multiple imputation. Various supervised machine learning classifier algorithms were trained, and an ensemble model was created. Ethical approval was obtained (20/EM/0183, IRAS 280077). Results Data on 2,716 patients who underwent elective abdominal surgery were included in the model training and validation. The infection incidence in this cohort was 38%. Out of 74 initial predictors, 19 were selected. These predictors included demographic data (gender), comorbidities (diabetes, COPD, cardiac conditions, obesity, anaemia), microbiology data (positive cultures within 90 days before surgery, multi-drug resistance), and laboratory investigations within 90 days before surgery: WBCs, lymphocytes, blood glucose, platelets, haematocrit, mean cell haemoglobin, haemoglobin, total bilirubin, aspartate aminotransferase, alanine transaminase, and C-reactive protein. The gradient boosting classifier performed the best among the supervised machine learning algorithms. The ensemble model showed high performance in training (sensitivity: 85.3%, specificity: 74.6%, AUROC: 88.6%) and internal validation (sensitivity: 96.9%, specificity: 74.1%, AUROC: 85.5%). The PROBAST checklist was used to address bias before predictor selection. Discussion/Conclusion The developed AI-predictive model has the potential to predict the likelihood of an elective abdominal surgery patient experiencing a post-operative infection. Identifying patients at high risk of post-surgical infections can guide clinical decision making and help implement interventions during the pre-operative stage, (i.e. as part of nutritional prehabilitation screening), to reduce infection risk. However, it is important to acknowledge the limitations of this model. One limitation is the presence of missing data in the training dataset, which may introduce biases and affect the accuracy of the model. Additionally, potential inaccuracies in data entry could impact the reliability of the model's predictions. Furthermore, it is crucial to consider any historical biases that might have influenced data collection, as these biases could have adverse effects on the developed model.

Efficient imputation of missing data using the information of local space defined by the geometric one-class classifier

Datasets gathered from actual systems may include missing data owing to unintentional faults, such as the breakdown of equipment as well as intentional reasons such as sampling inspection. Because missing data can result in incorrect and distorted results when analyzed, they should be addressed before the analysis is performed. Imputation of missing data involves replacing missing entries of data with values calculated from observed features, which is a more reasonable alternative than simple methods, including a complete case analysis. Although various imputation methods exist for missing data, most ignore the local space around it, which may be closely related to missing values. Furthermore, the imputation method, which can partially reflect local relationships, is susceptible to overfitting and has parameter tuning issues owing to the lack of a systematic definition of the local space. Thus, we propose a composite fuzzy hyper-rectangle (H-RTGL) imputation (CFHRI) method with the following characteristics: (i) it defines the local space using an H-RTGL-based one-class classifier to thoroughly describe the data of the target class, and (ii) it imputes the missing entries using a fuzzy model comprising imputation models calculated from H-RTGLs. These features enable CFHRI to formulate the local space adjacent to missing data systematically and alleviate the hazards of overfitting into a certain region of the dataset. We validated our method based on numerical experiments conducted using a dataset gathered from an actual system and comparison of the imputation performance of our method with that of other imputation methods. CFHRI showed statistically significant improvement in 5 datasets among 7 datasets used, and around 10% enhanced in terms of Mean Absolute Error (MAE). Moreover, we could achieve 3–5% of increased classification accuracy of imputed dataset, which indicates CFHRI can be a useful pre-processor of dataset whose purpose is classification.

Multicenter Study of a Risk Prediction Model for Critically Ill Children at High-Risk for Hospital-Acquired Venous Thromboembolism: Findings from the Children's Hospital-Acquired Thrombosis (CHAT) Consortium

Introduction: Critically ill children are at high-risk of developing a hospital-acquired venous thromboembolism (HA-VTE) as well as being at an increased risk for bleeding. To better understand which subset of children may benefit from thromboprophylaxis measures, (without contraindications for potential associated hemorrhagic complications), the Children's Hospital Acquired Thrombosis (CHAT) Consortium previously derived a HA-VTE risk assessment model (RAM) for critically ill children from 8 centers. This RAM included five variables: (1) immobility defined using a Braden score ≤2, (2) length of hospitalization prior to PICU admission, (3) central venous catheterization, (4) presence of congenital heart disease (CHD) and (5) history of autoimmune/inflammatory disorder or infection. For this study we aimed to externally evaluate our previous HA-VTE RAM in critically ill children in an independent cohort via the CHAT Consortium. Methods: Critically ill children aged 0-21 years admitted or transferred into a pediatric intensive care unit (PICU) at 32 U.S. centers were randomly selected to be enrolled in the study. Participants with a HA-VTE at admission or cardiac surgery within 2 weeks of PICU admission/transfer were excluded. Participants were prospectively followed via chart review for the development of a HA-VTE throughout their hospitalization and up to 30 days post discharge. Twenty-one variables that had been evaluated in univariate analyses from the previously-reported ICU-RAM derivation study were again analyzed in the present prospective cohort via univariate logistic regression. Variables with P-value of &amp;lt;0.1 were included in a multivariable logistic regression model, with P-values of &amp;lt;0.05 considered significant. Missing values were not imputed and complete case analysis was utilized. Results: From January 2020 to July 2022, 4,504 participants were enrolled and 93 developed a HA-VTE (2.1%). Complete data were available for 65 HA-VTE cases and 3056 controls without HA-VTE. Median age for HA-VTE participants was 5.2 years (interquartile range [IQR] 1.0-15.2) compared to 6.6 years (IQR 1.7-13.8) for those without HA-VTE; 55% and 53% of each group, respectively, were males. Among seven variables that met criteria for evaluation in the updated multivariable model, central venous catheter (odds ratio [OR] 4.2, 95% confidence interval [CI] 2.4-7.4), cancer (OR 3.1, 95% CI 1.4-6.7), immobility (OR 2.0, 95% CI 1.2-3.4, prior hospitalization (OR 2.0, 95% CI 1.1-3.5) and mechanical ventilation (OR 1.9, 95% CI 1.1-3.4) were each independently associated with HA-VTE risk (Table 1). The validation RAM had an area under the receiver operating characteristic curve of 0.81 (95% CI, 0.75-0.86) (Figure 1), compared to 0.78 (95% CI=0.73-0.84) for the previous model. Conclusions: The present work represents the largest external validation study of a HA-VTE RAM to date in critically ill children. We identified five independent risk factors for HA-VTE for children admitted to a PICU. Two of these factors were identified in the derivation study RAM, while three (cancer, prior hospitalization, and mechanical ventilation) were newly identified via this validation study.

Outcomes in Minorities with PTCL By Clinical Trial Enrollment

Background Peripheral T-cell Lymphoma (PTCL) affects a diverse cohort of patients for which survival disparities are evident in minorities. Clinical trial enrollment does not match the demographic diversity represented in PTCL and minorities are historically underrepresented. In this single-center retrospective analysis, we aim to answer if clinical trial enrollment can close the gap in survival outcomes in minorities compared to non-minorities in the modern era. Methods After IRB approval, 225 patients with PTCL were identified between 1991-2022, of which 199 had sufficient data for review. Medical records were reviewed for baseline characteristics, treatment parameters, and demographic data. Income quintile was calculated for the cohort using zip code level census data. The social deprivation index ( SDI) was calculated as a composite measure of demographic characteristics in the 2015-2019 American Community Survey. Overall survival ( OS) was calculated from diagnosis to time of death or last follow-up. Kaplan-Meier method was used to estimate survival probability. Survival difference was tested by log-rank and Cox regression complete case analysis. Due to the small sample size, Black and Hispanic patients were combined ( minority) and White with Asian/unknown/other ( non-minority) to delineate differences between groups. The OS was compared across four groups: ( 1) minority with first-line clinical trial, ( 2) minority without first-line clinical trial, ( 3) non-minority with first-line clinical trial, and ( 4) and non-minority without first-line clinical trial. Results From the initial cohort of 199 patients, 118 (59%) had relapsed/refractory (R/R) disease. The median age of diagnosis was 58 years. A total of 48.2% of patients were White (n=96), 22.1% Black (n=44), 18.5% Hispanic (n=37), and 11.1% Asian/other/unknown (n=22). The major PTCL subtypes included AITL 24% (n=48), ALCL 19% (n=38), PTCL_NOS 17% (n=33), and ATLL 15% (n=29). Importantly, 55% (n=24) of the black patient population had ATLL. Baseline characteristics are shown in Table 1. In total, 27% (n=53) were exposed to a novel agent in the first-line, and 64% (n=76) in second-line. Clinical trial enrollment accounted for 21% (n=42) first-line, and 34% (n=40) second-line. Black and Hispanic patients were more likely to be in the lower income quintile, 30% and 38%, respectively. The SDI score (0-100; higher score indicates more deprivation) was higher for Black (88), and Hispanic (95) patients compared to White (42) and Asian/other/unknown (63). However, SDI score was not associated with survival in cox models. In the first-line setting, there was a survival difference for both minorities and non-minorities enrolled in a clinical trial compared to those who were not ( Figure 1, p=0.0045), but there was no difference for novel agents (p=0.18).Given the aggressive nature of ATLL and differences in baseline covariates, a cox model was used to compare the four groups and OS was analyzed controlling for age, sex, insurance, stage, ECOG, and ATLL. Minorities performed similarly to non-minorities when participating in a clinical trial ( HR: 1.19 95% CI: [0.32, 4.4]), while minorities and non-minorities not enrolled on a clinical trial performed worse overall ( HR: 1.97, 95% CI: [0.77, 5.0], and HR: 1.49, 95% CI: [0.64, 3.4], respectively). In the second-line setting, both minorities and non-minorities appeared to experience OS benefit favoring clinical trials compared to those without (p=0.026). Conclusion Black and Hispanic patients have inferior outcomes compared to Whites across most subtypes of PTCL in data published by Adams et al. over a decade ago. Our data represents a trend toward similar outcomes in minoritized compared to non-minoritized patients with clinical trial enrollment in the first-line and second-line settings. Clinical trial enrollment may close the gap in survival outcomes; however interpretation is limited by small sample size. Multi-center efforts are needed to better elucidate these findings. It is imperative to place emphasis on improved clinical trial enrollment in diverse patient populations. *This work has been funded by ASH RTAF

Modelling clinicians' decisions to adjust loop diuretic dose based on patient-reported symptoms to inform decision-aids in daily care of chronic heart failure

Abstract Background To relieve congestion and prevent decompensation in chronic heart failure patients with reduced ejection fraction, the guidelines by the European Society of Cardiology recommend the use of loop diuretics.[1] The patient is encouraged to take an active role by tracking their symptoms and adjusting their dose accordingly.[1] Purpose To develop an exploratory prediction model of clinicians’ decisions to titrate loop diuretics based on patient-reported symptoms to create decision-aids. Methods We conducted a post hoc analysis of TIME-CHF.[2] Of the randomised participants at baseline, we selected those who did not withdraw consent during follow-up, had a furosemide-equivalent diuretic dose ≤600 mg at baseline and the first follow-up moment approximately 1 month after baseline, and had a recommendation registered on diuretic titration at the first follow-up. We developed logistic regression models for two outcomes at the first follow-up: up-titration versus no change, and down-titration versus no change. Predictors considered were: age, gender, furosemide-equivalent loop diuretic dose, NYHA class, angina, orthopnoea, paroxysmal nocturnal dyspnoea, nocturia, dry cough, dizziness, orthostasis, syncope, oedema, body mass index, systolic, diastolic, and systolic minus diastolic blood pressure, heart rate, jugular veins assessment, rales, hepatomegaly, third heart sound, serum-creatinine. These predictors separately and the intervention allocation (NT-proBNP versus symptom-guided therapy) were used to predict the outcomes. We explored symptoms occurring simultaneously with the outcome and change in symptoms between baseline and first follow-up. Ordinal predictors were modelled at the nominal and scale measurement level. We then selected a group of predictors deemed suitable in a daily care context per outcome. Analyses were performed based on complete cases and after multiple imputation. Performance metrics were R²Nagelkerke and C-statistics (apparent values and 95% confidence intervals via bootstrap). Results Of the 622 patients randomised at baseline, 500 fulfilled the selection criteria. There were 101 recommendations to up-titrate, 338 to keep the dose unchanged, and 61 to down-titrate. The complete case analysis did not differ substantively from analysis after multiple imputation due to the low number of missing values, therefore only the complete case analysis is reported here. Table 1 presents the results of the patient-reported symptoms modelled separately, collected simultaneously with the outcomes. For ordinal predictors, the scale measurement level is reported. Table 2 presents the results of the analysis using the groups of predictors. Conclusions This analysis may shape decision-aids by suggesting symptom importance in clinicians’ decisions. Oedema, NYHA class, and orthopnoea are tentatively the most important patient-reported symptoms in recommending up-titration, and orthostasis in recommending down-titration.

Data Missingness Reporting and Use of Methods to Address It in Critical Care Cohort Studies.

Failure to recognize and address data missingness in cohort studies may lead to biased results. Although Strengthening the Reporting of Observational Studies in Epidemiology reporting guidelines advocate data missingness reporting, the degree to which missingness is reported and addressed in the critical care literature remains unclear. To review published ICU cohort studies to characterize data missingness reporting and the use of methods to address it. We searched the 2022 table of contents of 29 critical care/critical care subspecialty journals having a 2021 impact factor greater than or equal to 3 to identify published prospective clinical or retrospective database cohort studies enrolling greater than or equal to 100 patients. In duplicate, two trained researchers conducted a manuscript/supplemental material PDF word search for "missing*" and extracted study type, patient age, ICU type, sample size, missingness reporting, and the use of methods to address it. A total of 656 studies were reviewed. Of the 334 of 656 (50.9%) studies mentioning missingness, missingness was reported for greater than or equal to 1 variable in 234 (70.1%) and it exceeded 5% for at least one variable in 160 (47.9%). Among the 334 studies mentioning missingness, 88 (26.3%) used exclusion criteria, 36 (10.8%) used complete-case analysis, and 164 (49.1%) used a formal method to avoid missingness. In these 164 studies, imputation only was used in 100 (61.0%), an analytic strategy only in 24 (14.6%), and both in 40 (24.4%). Only missingness greater than 5% (in ≥ 1 variable) was independently associated with greater use of a missingness method (adjusted odds ratio 2.91; 95% CI, 1.85-4.60). Among 140 studies using imputation, multiple imputation was used in 87 studies (62.1%) and simple imputation in 49 studies (35.0%). For the 64 studies using an analytic method, 12 studies (18.8%) assigned missingness as an unknown category, whereas sensitivity analysis was used in 47 studies (73.4%). Among published critical care cohort studies, only half mentioned result missingness, one-third reported actual missingness and only one-quarter used a method to manage missingness. Educational strategies to promote missingness reporting and resolution methods are required.

Development and External Validation of a Lymphoma-Specific Venous Thromboembolism Risk Assessment Model

Introduction: Venous thromboembolism (VTE) significantly affects cancer patients undergoing systemic therapy. Existing risk models developed in the pan-cancer population, such as the Khorana score, offer limited discrimination in cancer-specific setting. We aimed to develop and validate a risk assessment model (RAM) tailored for lymphoma patients, accounting for the clinical and sociodemographic heterogeneity across three distinct healthcare systems in the United States. Methods: Electronic health records (EHR) linked to cancer registry 2006-2021 from the Veterans Affairs national healthcare system (VA) with a total of 10,313 lymphoma patients were randomly divided into an 80% derivation cohort and a 20% internal validation cohort. Further external validation was carried out using data from two other healthcare systems - Harris Health System (HHS, N= 854, 2011-2020) and MD Anderson Cancer Center (MDACC, N=1,858, 2017-2020). Patients were included if they had newly diagnosed lymphoma requiring first-line systemic therapy within 1 year of diagnosis. Patients were excluded if they had recent diagnosis of acute VTE within the last 6 months or were prescribed anticoagulant within 1 month before index date. The index date was the time of systemic therapy, and all covariates were extracted on or before the index date. Incident VTE was defined using our published computable phenotype algorithm (PMID 36626707; 37067102). For RAM derivation (Figure 1), 32 candidate predictors were initially chosen based on clinical relevance and EHR data availability. Missing values were imputed via chain random forest. Parsimonious variable categories were selected by a combination of Least Absolute Shrinkage and Selection Operator (LASSO), Recursive Feature Elimination (RFE), and random forest to generate the final logistic regression model. For external validation, complete case analysis was performed using the derived beta coefficients without model refitting. Bootstrapped time-dependent c statistics and calibration curves for 6-month VTE were used to assess discrimination and fit. A pre-determined threshold of 8% VTE at 6-month from the predicted probability was used to stratify high vs. low-risk groups. Competing risk models were used to assess cumulative incidence. Results: At 6 months, the VTE incidence were 5.75% (n=469) in VA derivation and 6.12% (n=124), 8.24% (n=69), and 6.55% (n=112) in the VA, HHS, and MDACC validation cohorts, respectively. The three healthcare systems had different baseline characteristics in demographics, lymphoma types, systemic therapies, and comorbidities. Patients from VA were mostly older males with more comorbidities and remote history of VTE, patients from HHS were mostly uninsured minorities who were younger but with more aggressive histology, while those from MDACC were patients treated at a comprehensive cancer center. The variables in the final RAM included lymphoma type (4 groups ranging from very-low to very-high risk), body mass index, indwelling catheter, cytotoxic vs. targeted therapy, history of VTE, recent hospitalization, and paralysis/immobilization. Table 1 shows the variable prevalence in each cohort. As shown in Figure 1, the new RAM showed promising discrimination for VTE with a C-statistic of 0.69 in the derivation cohort and 0.67, 0.70, 0.72 in the internal and two external validation cohorts. By splitting patients using an 8% predicted risk threshold, the high- vs. low-risk group achieved a clear differentiation in risk stratification despite significant differences in baseline characteristics. This represented a noticeable improvement for lymphoma patients over the Khorana Score, which yielded C-statistics of 0.53-0.56. The overall calibration of the RAM appeared adequate. Conclusions: Using multiple cancer-EHR cohorts with heterogeneous patients and a wide range of lymphoma subtypes, we successfully derived and externally validated an improved RAM for VTE specifically in lymphoma patients requiring early initiation of systemic therapy. Along several well-known patient-specific risk factors, lymphoma type was an important predictor of VTE risk. The new RAM could improve VTE risk stratification and prevention efforts in lymphoma patients.