Imputed Datasets Research Articles

The STAR care pathway for patients with chronic pain after total knee replacement: four-year follow-up of a randomised controlled trial

BackgroundThe Support and Treatment After Replacement (STAR) care pathway is a clinically important and cost-effective intervention found to improve pain outcomes over one year for people with chronic pain three months after total knee replacement (TKR). We followed up STAR trial participants to evaluate the longer-term clinical- and cost-effectiveness of this care pathway.MethodsParticipants who remained enrolled on the trial at one year were contacted by post at a median of four years after randomisation and invited to complete a questionnaire comprising the same outcomes collected during the trial. We captured pain (co-primary outcome using the Brief Pain Inventory (BPI) pain severity and interference scales; scored 0–10, best to worst), function, neuropathic characteristics, emotional aspects of pain, health-related quality of life, and satisfaction. Electronic hospital informatics data on hospital resource use for the period of one to four years post-randomisation were collected from participating hospital sites. The economic evaluation took an National Health Service (NHS) secondary care perspective, with a four-year time horizon.ResultsOverall, 226/337 (67%) of participants returned completed follow-up questionnaires, yielding adjusted between-group differences in BPI means of -0.42 (95% confidence interval, CI (-1.07, 0.23); p = 0.20) for pain severity and − 0.64 (95% CI -1.41, 0.12); p = 0.10) for pain interference. Analysis using a multiple imputed data set (n = 337) showed an incremental net monetary benefit in favour of the STAR care pathway of £3,525 (95% CI -£990 to £8,039) at a £20,000/QALY willingness-to-pay threshold, leading to a probability that the intervention was cost-effective of 0.94.ConclusionsThe magnitude of the longer-term benefits of the STAR care pathway are uncertain due to attrition of trial participants; however, there is a suggestion of some degree of sustained clinical benefit at four years. The care pathway remained cost-effective at four years.Trial registrationISRCTN: 92,545,361.

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.

Description of a Clinical Score to Identify PMBL Patients at High Risk of Early-Failure after Rituximab Doxorubicin Back-Bone Chemoimmunotherapy. a FIL Real-World Study

Introduction. Primary Mediastinal B Lymphoma (PMBL) is a rare and aggressive neoplasia that typically occurs in young women and is primarily located in the mediastinum. Although the combination of anthracycline and rituximab immunochemotherapy, and PET-guided consolidation radiotherapy, results in a high cure rate, approximately 10-20% of patients show refractoriness or early failure to front-line treatment with a rapid clinical deterioration and a dismal prognosis. These cases, if promptly identified, could benefit from early treatment shift. Therefore, we addressed the development of a clinical score aimed at estimating the risk of early failure of PMBL patients treated with anthracycline-rituximab-based chemoimmunotherapy in a real-world nationwide setting. Material and Methods This is a retrospective cohort study of an unselected population of 931 adult PMBL patients treated in 37 FIL centers. To qualify for this research, patients had to have a confirmed diagnosis of PMBL, display typical symptoms and have been included in the local databases from 01/01/2007 to 31/12/2019. All patients received treatment with rituximab plus CHOP21 (n=98), CHOP14 (n=181), megaCHOP (n=31), MACOPB (225), VACOPB (n=179), DAEPOCH (n=179). Patients were excluded from this analysis if they received suboptimal treatment (22) or ASCT as consolidation of first-line therapy (12) or were censored within 365 days of treatment initiation (22). We defined early failure (EF) as any relapse or progression within 365 days of treatment initiation. Candidate predictors were clinical factors at the time of diagnosis, including: age, ECOG PS (a 4-level scale), LDH higher than Upper Limit of Normal, Ann Arbor Stage (a 4-level scale), systemic B symptoms, bulky mediastinum > 10 cm, number of Extra-Nodal sites (a 4-level scale: 0, 1, 2, or >=3), and pericardial or pleural effusion. Missing data in the evaluated predictors were multiple-imputed, and combined estimates were obtained from 20 imputed data sets. We used a multivariable logistic model with Backward Selection (level to stay p<0.1). Estimates were adjusted for front-line therapy. Standard errors were adjusted using the clustered sandwich estimator according to the patient's center. The model was internally validated with 1000 bootstrap samples from the original data. The optimism-corrected AUROC was estimated based on the results of the bootstrap process. We constructed a risk score by assigning each factor a weight proportional to the smallest coefficient, rounded to the nearest integer. The discrimination ability of the prognostic score was also evaluated considering the time to progression or relapse (Early Failure Survival) using the Kaplan-Meier method. Results Among the 869 patients included in the present study, 109 (12.5%) were identified as early failures (EF). Four factors were selected after the backward selection process to construct the prognostic score (AUROC 0.744, optimism-corrected 0.710). These factors included Systemic B-symptoms (OR=1.76, p=0.007, weight=1), Bulky mediastinum > 10 cm (OR=2.12, p=0.017, weight=1), AA Stage II (OR=2.20, p=0.018, weight=1), AA Stage III-IV (OR=2.73, p=0.017, weight=2), and Number of Extra-Nodal sites >=2 (OR=2.88, p<0.001, weight=2) (Table 1). Based on the incidence of EF, the score was classified into 3 classes: low (0-1) with 216 patients and an EF incidence of 3.7% (95% CI: 1.6% - 7.2%), intermediate (2-3) with 408 patients and an EF incidence of 11.8% (95% CI: 9.0% - 15.0%), and high (> 4) with 122 patients and an EF incidence of 30.8% (95% CI: 22.3% - 39.3%). (Figure 1A). At 24 months, Early-Failure-Survival (EFS) rates were as follows: High - 63.7% (95% CI: 54.5% - 71.6%), Intermediate - 87.0% (95% CI: 83.6% - 89.7%), and Low - 95.4% (95% CI: 91.5% - 97.5%). Conclusions Based on a large PMBL retrospective case series, we developed a simple and practical prognostic score subdividing patients into three clear-cut groups with different likelihoods of experiencing early failure. We acknowledge that the study has potential biases inherent in retrospective analysis, and further efforts aim to validate these results in external series of patients. If the predictive ability is confirmed, the score could be a valuable tool for treatment planning, interpretation, and comparison of clinical studies.

Development and internal validation of a model predicting patient-reported shoulder function after arthroscopic rotator cuff repair in a Swiss setting

BackgroundPrediction models for outcomes after orthopedic surgery provide patients with evidence-based postoperative outcome expectations. Our objectives were (1) to identify prognostic factors associated with the postoperative shoulder function outcome (the Oxford Shoulder Score (OSS)) and (2) to develop and validate a prediction model for postoperative OSS.MethodsPatients undergoing arthroscopic rotator cuff repair (ARCR) were prospectively documented at a Swiss orthopedic tertiary care center. The first primary ARCR in adult patients with a partial or complete rotator cuff tear were included between October 2013 and June 2021. Thirty-two potential prognostic factors were used for prediction model development. Two sets of factors identified using the knowledge from three experienced surgeons (Set 1) and Bayesian projection predictive variable selection (Set 2) were compared in terms of model performance using R squared and root-mean-squared error (RMSE) across 45 multiple imputed data sets using chained equations and complete case data.ResultsMultiple imputation using data from 1510 patients was performed. Set 2 retained the following factors: American Society of Anesthesiologists (ASA) classification, baseline level of depression and anxiety, baseline OSS, operation duration, tear severity, and biceps status and treatment. Apparent model performance was R-squared = 0.174 and RMSE = 7.514, dropping to R-squared = 0.156, and RMSE = 7.603 after correction for optimism.ConclusionA prediction model for patients undergoing ARCR was developed using solely baseline and operative data in order to provide patients and surgeons with individualized expectations for postoperative shoulder function outcomes. Yet, model performance should be improved before being used in clinical routine.

Missing data imputation, prediction, and feature selection in diagnosis of vaginal prolapse

BackgroundData loss often occurs in the collection of clinical data. Directly discarding the incomplete sample may lead to low accuracy of medical diagnosis. A suitable data imputation method can help researchers make better use of valuable medical data.MethodsIn this paper, five popular imputation methods including mean imputation, expectation-maximization (EM) imputation, K-nearest neighbors (KNN) imputation, denoising autoencoders (DAE) and generative adversarial imputation nets (GAIN) are employed on an incomplete clinical data with 28,274 cases for vaginal prolapse prediction. A comprehensive comparison study for the performance of these methods has been conducted through certain classification criteria. It is shown that the prediction accuracy can be greatly improved by using the imputed data, especially by GAIN. To find out the important risk factors to this disease among a large number of candidate features, three variable selection methods: the least absolute shrinkage and selection operator (LASSO), the smoothly clipped absolute deviation (SCAD) and the broken adaptive ridge (BAR) are implemented in logistic regression for feature selection on the imputed datasets. In pursuit of our primary objective, which is accurate diagnosis, we employed diagnostic accuracy (classification accuracy) as a pivotal metric to assess both imputation and feature selection techniques. This assessment encompassed seven classifiers (logistic regression (LR) classifier, random forest (RF) classifier, support machine classifier (SVC), extreme gradient boosting (XGBoost) , LASSO classifier, SCAD classifier and Elastic Net classifier)enhancing the comprehensiveness of our evaluation.ResultsThe proposed framework imputation-variable selection-prediction is quite suitable to the collected vaginal prolapse datasets. It is observed that the original dataset is well imputed by GAIN first, and then 9 most significant features were selected using BAR from the original 67 features in GAIN imputed dataset, with only negligible loss in model prediction. BAR is superior to the other two variable selection methods in our tests.ConcludesOverall, combining the imputation, classification and variable selection, we achieve good interpretability while maintaining high accuracy in computer-aided medical diagnosis.

The intensity and duration of occupational noise exposure and cardiovascular disease in the United States: a nationally representative study, 2015 to 2020.

Occupational noise exposure may be associated with an increased risk of cardiovascular disease (CVD). Yet the findings are inconclusive. This study aimed to examine the association between self-reported occupational noise exposure and CVD (using a broad composite case definition and by each condition) and identify how these associations vary with the intensity and duration of noise exposure, and combinations thereof. This cross-sectional study included a nationally representative sample (n = 6,266) from the National Health and Nutrition Examination Survey (2015 to 2020), aged 20 and greater, in the United States. Survey-weighted logistic regression models were constructed from multiple imputed datasets. Relative to the unexposed, the adjusted odds ratio (95% confidence interval) of composite CVD was 1.33 (1.05 to 1.67) among the noise-exposed population, and ranged from 1.23 to 1.56 when examining CVD conditions separately. The odds ratios of composite CVD were 1.43 (1.06 to 1.93), 1.43 (1.04 to 1.95), and 1.51 (1.03 to 2.21) among those who had noise exposure with very loud intensity of any duration, with duration ≥10 years at any intensity, and with a combination of very loud noise ≥10 years, respectively, compared to those unexposed. Increased risk of CVD is associated with occupational noise exposure, particularly at higher intensities and longer durations. Policies and interventions for noise mitigation at workplaces are warranted, targeting individuals with chronic exposure to high-level noise.

An Externally Controlled Trial to Indirectly Compare Tazemetostat Plus Lenalidomide and Rituximab Versus Lenalidomide and Rituximab after at Least One Prior Systemic Therapy in Patients with Refractory and Relapsed Follicular Lymphoma

BACKGROUND: Follicular lymphoma (FL) is an indolent, incurable lymphoma with a heterogeneous clinical course characterized by multiple relapses. Recent data indicate a potential synergism between tazemetostat and lenalidomide and an additive effect with rituximab, offering a clinical rationale for this as a combination strategy (Batlevi, Blood 2022). SYMPHONY-1 Phase 1b, Stage 1 (EZH-302; NCT04224493) is an open-label, single-arm, safety run-in trial to evaluate the safety and tolerability of tazemetostat combined with lenalidomide plus rituximab (R2). This study indirectly compares the efficacy results of the SYMPHONY-1 Phase 1b single arm trial with an external control arm (ECA) derived from the Flatiron Health Research Database (FHRD), consisting of real-world FL patients treated with R2 after receiving ≥1 prior systemic therapy. METHODS: This study utilized individual patient data from SYMPHONY-1 based on a data cutoff date of July 10, 2023, and the FHRD, a large database of electronic health records from oncology practices in the United States, including data through December 31, 2022. The treatment arm consisted of patients enrolled in the safety run-in portion of SYMPHONY-1 treated with a combination of tazemetostat plus R2, followed from enrollment to the data cutoff date. The ECA was constructed by applying eligibility criteria adapted from SYMPHONY-1 to real-world patients from the FHRD treated with R2 in second line (2L); patients were followed from the start date of the earliest eligible line of therapy (LOT) to the end of the patient's record. The primary endpoint was objective response rate (ORR), defined as the proportion of patients who achieved a complete or partial response. In SYMPHONY-1, ORR was assessed by an independent review committee according to the 2014 Lugano Classification. In the ECA, real-world ORR (rwORR) was determined by the treating clinician's interpretation of change in disease burden. The analysis set included patients from the treatment arm with ≥1 evaluable response assessment and patients from the ECA with ≥1 radiographic assessment documented in the FHRD. The statistical analysis used propensity score weighting (PSW) to adjust the distribution of seven prespecified baseline patient characteristics of the ECA to align with the treatment arm. These characteristics were age, Eastern Cooperative Oncology Group Performance Status, disease grade at diagnosis, progression of disease within 24 months (POD24) from diagnosis, lactate dehydrogenase (LDH) > upper limit of normal (ULN), size of the largest node, and Follicular Lymphoma International Prognostic Index (FLIPI) risk group. An impute-then-exclude multiple imputation strategy was used to handle missing data for variables required in the eligibility criteria and PSW. The primary endpoint was indirectly compared by estimating the difference between the ORR of the treatment arm and rwORR of the balanced ECA on each imputed dataset, and the associated standard errors were approximated using bootstrapping. Rubin's rules were applied to combine the results into a single estimate with a 95% confidence interval and corresponding two-sided p-value. RESULTS: The primary analysis compared 42 patients in the treatment arm from SYMPHONY-1 and a median of 36 patients (range: 34 to 38) in the ECA among 815 total patients in the FHRD across ten imputed datasets. The unweighted analysis found an ORR of 95.2% in the treatment arm and 75.7% in the ECA, resulting in a difference of 19.6% (95% confidence interval [CI]: 3.6-35.5) (Table 2). After PSW, the distribution of baseline characteristics for the ECA was comparable with the treatment arm (all standardized mean differences < 0.16), and the median effective sample size was reduced to 12.7 (range: 4.9 to 19.1) (Table 1). ORR was 95.2% in the treatment arm and 77.9% in the ECA, resulting in a difference of 17.3% (95% CI: -17.5-52.1; two-sided P = 0.16) (Table 2). CONCLUSIONS: This ECA study found that patients treated with tazemetostat combined with R2 were estimated to have a higher ORR than those treated with R2, though inference is limited due to small sample sizes. These findings help contextualize the SYMPHONY-1 Phase 1b single-arm trial results and combination potential of tazemetostat plus R2. There is a randomized Phase 3 confirmatory trial underway to confirm the clinical benefit of tazemetostat plus R2 (SYMPHONY-1 Stage 2, NCT04224493).

Treatment Intensity and Outcomes in Elderly Patients with DLBCL Receiving First Line Therapy

Background: One-third of patients (pts) with newly diagnosed DLBCL are 70 and older. Although a standard of care for first-line (1L) DLBCL is anthracycline-based chemoimmunotherapy with R-CHOP, in elderly DLBCL pts treatment approaches vary and outcomes of these are poorly understood. In this study, we aimed to evaluate long-term outcomes among elderly pts with DLBCL across 1L regimens with a focus on those who received reduced dose (RD) R-CHOP, standard dose (SD) R-CHOP, and non-anthracycline alternative regimens (AR). Methods: We evaluated a cohort of pts from the Flatiron Health electronic health record (EHR)-derived, de-identified database with DLBCL diagnosis (dx) after 1/1/2011, age 70 or greater at dx, and with no evidence of prior indolent lymphoma or other malignancy within 2 years of 1L therapy. RD R-CHOP was defined as <80% SD doxorubicin or cyclophosphamide. Time-to-event analysis was performed with the Kaplan-Meier method and differences were evaluated using the log rank test. Real world overall survival (rwOS) was defined as the interval, in months (mos), from DLBCL dx to death or last follow-up. Real world progression free survival (rwPFS) was defined as the interval, in mos, from initiation of 1L therapy to progression, death, or last follow-up. Multiple imputation was performed (m = 50). Multivariate Cox regression analysis (MVA) was conducted within each imputed dataset and pooled. MVA was additionally performed across age strata. Results: There were 434 pts included in the complete cohort. Eighty-six percent (N=372) received 1L therapy and 62 pts (14%) had no therapy documented. Among treated pts, 47% were female, median age was 79 (IQR 75-82), and median CCI score 7 (IQR 6-9). Twenty-six percent and 47% had limited and advanced stage at dx, respectively; stage was unknown in 27%. Cell of origin was germinal center (GCB), non-GCB, and unknown in 34%, 24%, and 43%, respectively. Forty-eight percent of pts (N=212) received R-CHOP as 1L. Best real world response (rwR), defined as either complete response (CR) or partial response (PR) to 1L, was 89.0%, 89.5%, and 81.6% (p = 0.20) in pts who received SD R-CHOP, RD R-CHOP, and AR, respectively. CR rates were 78.0%, 58.1%, and 60.2% (p = 0.008) among these cohorts, respectively. Median rwOS and rwPFS among all treated pts was 39.0 mos (32.4-47.3) and 15.2 mos (12.3-23.4), respectively. Among patients who received SD R-CHOP, RD R-CHOP, and AR, median age (IQR) was 72 (71-77), 81 (79-83), and 80 (77-81), respectively (p<0.001). Median Charlson comorbidity index (CCI) score (IQR) was 6 (5-8), 8 (6-9), and 7 (6-9) across these groups, respectively (p<0.001). SD R-CHOP was associated with better rwPFS and rwOS in the entire treated cohort. Real world OS and rwPFS were 95.6 mos (67.0-NE) and 77.3 mos (40.3-NE) for SD R-CHOP, 37.6 mos (27.9-NE) and 15.8 mos (11.9-35.8) RD R-CHOP, and 25.7 mos (16.0-32.6) and 9.6 mos (8.1-12.0) for AR, respectively. When stratifying by age, pts 70-79 years had PFS of 77.3 mos (41.2-NE), 22.4 mos (9.86-NE), and 11.0 mos (8.6-23.4) with SD R-CHOP, RD R-CHOP, and AR, respectively (p <0.0001), whereas pts >=80 had PFS 6.14 mos (4.21-NE), 14.7 mos (8.87-NE), and 7.3 mos (4.47-12.0) with these regimens (p=0.03). In Cox regression MVA, CCI was associated with rwPFS in pts >=80 (HR 1.18, p=0.004). High-risk IPI score was associated with rwPFS (HR 2.13, p=0.05) and rwOS (HR 2.51, p=0.033), but only in pts age 70-79. When stratifying by age and adjusting for comorbidities, IPI, and MYC alteration, treatment with RD R-CHOP and AR were associated with inferior rwPFS and rwOS compared to SD R-CHOP among pts age 70-79, however, there was no association between treatment intensity and outcome among pts >=80 (Table). Conclusions: This study represents one of the largest real-world studies evaluating contemporary outcomes among elderly pts with DLBCL receiving 1L therapy. In MVA, SD R-CHOP is significantly associated with improved rwPFS and rwOS compared to RD-RCHOP and AR, but only among pts aged 70-79. There was no effect of treatment intensity or regimen on either rwPFS or rwOS among pts aged >=80. Other covariates significantly associated with outcome included CCI among pts >=80 and high-risk IPI in pts 70-79. These results suggest that pts aged 70-79 should receive SD R-CHOP whenever feasible, but that treatment intensity is not the primary determinant of outcomes among pts aged >=80. CCI should be part of routine evaluation in elderly DLBCL pts.

A Predictive Score Based on Cytokines, CD4/CD8 Ratio and Age Predicts Hematological Neoplasm-Associated Hemophagocytic Lymphohistiocytosis (H-HLH)

Introduction Early diagnostic of Hematological neoplasm (H)-associated hemophagocytic lymphohistiocytosis (H-HLH) is a major goal in the management of adult (sporadic) HLH (HLHa) to allow appropriate and early intervention and reduce the risk of fatal outcome. At diagnosis, it is often difficult to distinguish H-HLH from other associated diseases autoimmune/inflammatory (MAS/AID), infectious like leishmaniosis or idiopathic forms. Here we have developed a predictive score for H-HLH that is based on clinical parameters and cytokines levels. Methods Serum levels of thirty cytokines were analysed by Elisa or by Luminex Technology, in 112 patients (age≥18 years old and “HScore”≥169) of the French HLH cohort data base (clinicaltrials.gov number NCT02113917). Multiple imputation chain equation was performed to account for the missingness in covariates. All following analyses were carried out based on the imputed datasets by combining estimates from each completed dataset using Rubin's rules. A logistic regression was used to develop and validate the risk score predicting H-HLH. Model calibration and discrimination were assessed using the Hosmer-Lemeshow and receiver operating characteristic (ROC) curves, respectively. A 10-fold cross-validation was carried out to consider overfitting bias. Risk scores were derived using the model parameters and individual covariates values. A threshold on the risk score was determined to maximize the Youden index (sensitivity+specificity-1). H-HLH high-risk and low-risk patients were then delineated according to this score threshold, respectively, defining an high and low risk H-HLH group (score ≥ or < threshold). Results The cohort (112 patients) included 45 H-HLH and 67 AID or idiopathic HLHa. Malignancies were distributed as Hodgkin Lymphoma (HL) (n=3), B cell-Non Hodgkin Lymphoma (NHL) (n=22, (EBV+, n=9)), and T cell-NHL (n=13 (EBV+, n=7)). The median [IQR] age of patients in the cohort was 48 years [36-61] old and 64 (57%) were men. From twenty-one potential predictors (including 9 HLH-related cytokines), eight were associated with H-HLH in univariate logistic regression: age > 48y (p<0.004), hemoglobin > 9g/l (p=0.03), platelets > 82G/l (p=0.03), Il-10 > 42 pg/mL (p=0.0003), TNF-α>43 pg/mL (p=0.0006), IL-18>574 pg/mL (p=0.0002), lymphocytes CD4/CD8 >1 (p=0.03), and five, independent in a multivariate analysis, were included in the risk score : age > 48y (OR [95%CI]: 3.0 [1.1 - 8.3], p=0.03), Il-10 > 42 pg/mL (OR [95%CI]: 4.5 [1.4-14.6], p=0.01), TNF-α>43 pg/mL (OR [95%CI]: 0.2 [0.07-0.57] ,p=0.003), IL-18>574 pg/mL (OR [95%CI]: 3.1[1.0-9.5], p=0.04) and lymphocytes CD4/CD8 >1 (OR [95%CI]: 0.3 [0.1-1.0], p=0.04). The score ranged from -28 and 37 with an attribution of points of 11 for age>48, 15 for Il-10 > 42, -16 for TNF-α > 43, 11 for Il-18 > 574 and -12 for CD4/CD8 > 1. The mean AUC in the cohort was 0.87 (95%CI, 0.79-0.92) and the AUC in the internal validation was 0.82 (95%CI, 0.61-1.00). The goodness-of-fit test showed a p-value of 0.30. The logistic model demonstrated that patients with a score ≥ 6 (sensitivity [95%CI]: 0.80 [0.65-0.90] and specificity [95%CI]: 0.75 [0.63-0.85] at the threshold of 6), the risk of having H-HLH increases by a factor of 12 compared to patients with a score < 6. (OR [95% CI] = 12 [4.7-29.7], p < 0.0001). Conclusion In this cohort, a risk score based on five parameters (age, IL10, TNF-a, IL-18, lymphocytes CD4/CD8) was developed that may help predict a patient's risk to have H-HLH.

Developing a prognostic model for stress reduction in patients with prolonged work-related stress.

Mindfulness-based stress reduction (MBSR) is a 9-session group-treatment programme for managing stress. Research suggests variability in the outcomes of MBSR among participants. This prognostic (not causal) study develops a multivariable model that may support clinicians in forecasting expected MBSR outcomes. We used data of 763 patients collected from MBSR programs conducted between October 2015 and March 2022. Candidate prognostic factors at baseline included psychosocial work environment, sociodemographic, and clinical information. Multiple imputation was used to handle missing data (imputations=200). Important prognostic factors were backward selected in ≥5% of the imputed datasets. The final prediction model including the selected prognostic factors was evaluated using linear regression with a four-fold internal cross-validation procedure. Reductions in perceived stress from baseline to end of the MBSR programme were predicted by a lower General Severity Index (β=2.00, p<0.01), higher baseline levels of stress (β=-0.88, p<0.01), and somewhat by having managerial responsibility in the latest job (vs. no; β=-2.53, p=0.07). The remaining prognostic factors were weaker predictors, for example, gender and income. Internal validity of the final model was indicated by consistent results from four randomly folded subsamples. This study developed a prognostic model predicting changes in stress levels in relation to the MBSR programme. A reduction in stress level was particularly predicted by milder psychological symptoms and higher baseline levels of perceived stress. These predictions cannot be taken as evidence of causal associations. Forecasting of the illness course should be cautiously practiced using clinical judgement regarding individual patients.

Pooling methods for likelihood ratio tests in multiply imputed data sets.

Likelihood ratio tests (LRTs) are a popular tool for comparing statistical models. However, missing data are also common in empirical research, and multiple imputation (MI) is often used to deal with them. In multiply imputed data, there are multiple options for conducting LRTs, and new methods are still being proposed. In this article, we compare all available methods in multiple simulations covering applications in linear regression, generalized linear models, and structural equation modeling. In addition, we implemented these methods in an R package, and we illustrate its application in an example analysis concerned with the investigation of measurement invariance. (PsycInfo Database Record (c) 2023 APA, all rights reserved).

A longitudinal transition imputation model for categorical data applied to a large registry dataset.

Imputation of longitudinal categorical covariates with several waves and many predictors is cumbersome in terms of implausible transitions, colinearity, and overfitting. We designed a simulation study with data obtained from a general practitioners' morbidity registry in Belgium for three waves, with smoking as the longitudinal covariate of interest. We set varying proportions of data on smoking to missing completely at random and missing not at random with proportions of missingness equal to 10%, 30%, 50%, and 70%. This study proposed a 3-stage approach that allows flexibility when imputing time-dependent categorical covariates. First, multiple imputation using fully conditional specification or multiple imputation for the predictor variables was deployed using the wide format such that previous and future information of the same patient was utilized. Second, a joint Markov transition model for initial, forward, backward, and intermittent probabilities was developed for each imputed dataset. Finally, this transition model was used for imputation. We compared the performance of this methodology with an analyses of the complete data and with listwise deletion in terms of bias and root mean square error. Next, we applied this methodology in a clinical case for years 2017 to 2021, where we estimated the effect of several covariates on the pneumococcal vaccination. This methodological framework ensures that the plausibility of transitions is preserved, overfitting and colinearity issues are resolved, and confounders can be utilized. Finally, a companion R package was developed to enable the replication and easy application of this methodology.

Clinical outcomes of Staphylococcus capitis isolation from neonates, England, 2015–2021: a retrospective case–control study

ObjectiveStaphylococcus capitis, a coagulase-negative staphylococci (CoNS) species, has been increasingly detected from UK sterile site samples and has caused neonatal unit outbreaks worldwide. We compared survival to discharge and 30-day...

Severity of influenza-associated hospitalisations by influenza virus type and subtype in the USA, 2010–19: a repeated cross-sectional study

Influenza burden varies across seasons, partly due to differences in circulating influenza virus types or subtypes. Using data from the US population-based surveillance system, Influenza Hospitalization Surveillance Network (FluSurv-NET), we aimed to assess the severity of influenza-associated outcomes in individuals hospitalised with laboratory-confirmed influenza virus infections during the 2010-11 to 2018-19 influenza seasons. To evaluate the association between influenza virus type or subtype causing the infection (influenza A H3N2, A H1N1pdm09, and B viruses) and in-hospital severity outcomes (intensive care unit [ICU] admission, use of mechanical ventilation or extracorporeal membrane oxygenation [ECMO], and death), we used FluSurv-NET to capture data for laboratory-confirmed influenza-associated hospitalisations from the 2010-11 to 2018-19 influenza seasons for individuals of all ages living in select counties in 13 US states. All individuals had to have an influenza virus test within 14 days before or during their hospital stay and an admission date between Oct 1 and April 30 of an influenza season. Exclusion criteria were individuals who did not have a complete chart review; cases from sites that contributed data for three or fewer seasons; hospital-onset cases; cases with unidentified influenza type; cases of multiple influenza virus type or subtype co-infection; or individuals younger than 6 months and ineligible for the influenza vaccine. Logistic regression models adjusted for influenza season, influenza vaccination status, age, and FluSurv-NET site compared odds of in-hospital severity by virus type or subtype. When missing, influenza A subtypes were imputed using chained equations of known subtypes by season. Data for 122 941 individuals hospitalised with influenza were captured in FluSurv-NET from the 2010-11 to 2018-19 seasons; after exclusions were applied, 107 941 individuals remained and underwent influenza A virus imputation when missing A subtype (43·4%). After imputation, data for 104 969 remained and were included in the final analytic sample. Averaging across imputed datasets, 57·7% (weighted percentage) had influenza A H3N2, 24·6% had influenza A H1N1pdm09, and 17·7% had influenza B virus infections; 16·7% required ICU admission, 6·5% received mechanical ventilation or ECMO, and 3·0% died (95% CIs had a range of less than 0·1% and are not displayed). Individuals with A H1N1pdm09 had higher odds of in-hospital severe outcomes than those with A H3N2: adjusted odds ratios (ORs) for A H1N1pdm09 versus A H3N2 were 1·42 (95% CI 1·32-1·52) for ICU admission; 1·79 (1·60-2·00) for mechanical ventilation or ECMO use; and 1·25 (1·07-1·46) for death. The adjusted ORs for individuals infected with influenza B versus influenza A H3N2 were 1·06 (95% CI 1·01-1·12) for ICU admission, 1·14 (1·05-1·24) for mechanical ventilation or ECMO use, and 1·18 (1·07-1·31) for death. Despite a higher burden of hospitalisations with influenza A H3N2, we found an increased likelihood of in-hospital severe outcomes in individuals hospitalised with influenza A H1N1pdm09 or influenza B virus. Thus, it is important for individuals to receive an annual influenza vaccine and for health-care providers to provide early antiviral treatment for patients with suspected influenza who are at increased risk of severe outcomes, not only when there is high influenza A H3N2 virus circulation but also when influenza A H1N1pdm09 and influenza B viruses are circulating. The US Centers for Disease Control and Prevention.

Reprogramming of lipid metabolism in hepatocellular carcinoma resulting in downregulation of phosphatidylcholines used as potential markers for diagnosis and prediction

ABSTRACT Background Aberrant methylation and metabolic perturbations may deepen our understanding of hepatocarcinogenesis and help identify novel biomarkers for diagnosing hepatocellular carcinoma (HCC). We aimed to develop an HCC model based on a multi-omics. Research design and methods Four hundred patient samples (200 with HCC and 200 with hepatitis B virus-related liver disease (HBVLD)) were subjected to liquid chromatography-mass spectrometry and multiplex bisulfite sequencing. Integrative analysis of clinical data, CpG data, and metabolome for the 20 complete imputation datasets within a for-loopwas used to identify biomarker. Results Totally, 1,140 metabolites were annotated, of which 125 were differentially expressed. Lipid metabolism reprogramming in HCC, resulting in phosphatidylcholines (PC) significantly downregulated, partly due to the altered mitochondrial beta-oxidation of fatty acids with diverse chain lengths. Age, sex, serum-fetoprotein levels, cg05166871,cg14171514, cg18772205, PC (O-16:0/20:3(8Z, 11Z, 14Z)), and PC (16:1(9Z)/P-18:0) were used to develop the HCC model. The model presented a good diagnostic and an acceptable predictive performance. The cumulative incidence of HCC in low- and high-risk groups of HBVLD patients were 1.19% and 21.40%, respectively (p = 0.0039). Conclusions PCs serve as potential plasma biomarkers and help identify patients with HBVLD at risk of HCC who should be screened for early diagnosis and intervention.

Hypnotherapy for agoraphobia-Feasibility and efficacy investigated in a pilot study.

https://classic.clinicaltrials.gov/ct2/show/NCT03684577, identifier: NCT03684577.

Combining meta-analysis with multiple imputation for one-step, privacy-protecting estimation of causal treatment effects in multi-site studies.

Missing data complicates statistical analyses in multi-site studies, especially when it is not feasible to centrally pool individual-level data across sites. We combined meta-analysis with within-site multiple imputation for one-step estimation of the average causal effect (ACE) of a target population comprised of all individuals from all data-contributing sites within a multi-site distributed data network, without the need for sharing individual-level data to handle missing data. We considered two orders of combination and three choices of weights for meta-analysis, resulting in six approaches. The first three approaches, denoted as RR + metaF, RR + metaR and RR + std, first combined results from imputed data sets within each site using Rubin's rules and then meta-analyzed the combined results across sites using fixed-effect, random-effects and sample-standardization weights, respectively. The last three approaches, denoted as metaF + RR, metaR + RR and std + RR, first meta-analyzed results across sites separately for each imputation and then combined the meta-analysis results using Rubin's rules. Simulation results confirmed very good performance of RR + std and std + RR under various missing completely at random and missing at random settings. A direct application of the inverse-variance weighted meta-analysis based on site-specific ACEs can lead to biased results for the targeted network-wide ACE in the presence of treatment effect heterogeneity by site, demonstrating the need to clearly specify the target population and estimand and properly account for potential site heterogeneity in meta-analyses seeking to draw causal interpretations. An illustration using a large administrative claims database is presented.

Development and Internal Validation of a Prediction Model for Falls Using Electronic Health Records in a Hospital Setting.

Fall prevention is important in many hospitals. Current fall-risk-screening tools have limited predictive accuracy specifically for older inpatients. Their administration can be time-consuming. A reliable and easy-to-administer tool is desirable to identify older inpatients at higher fall risk. We aimed to develop and internally validate a prognostic prediction model for inpatient falls for older patients. Retrospective analysis of a large cohort drawn from hospital electronic health record data. Older patients (≥70years) admitted to a university medical center (2016 until 2021). The outcome was an inpatient fall (≥24hours of admission). Two prediction models were developed using regularized logistic regression in 5 imputed data sets: one model without predictors indicating missing values (Model-without) and one model with these additional predictors indicating missing values (Model-with). We internally validated our whole model development strategy using 10-fold stratified cross-validation. The models were evaluated using discrimination (area under the receiver operating characteristic curve) and calibration (plot assessment). We determined whether the areas under the receiver operating characteristic curves (AUCs) of the models were significantly different using DeLong test. Our data set included 21,286 admissions. In total, 470 (2.2%) had a fall after 24hours of admission. The Model-without had 12 predictors and Model-with 13, of which 4 were indicators of missing values. The AUCs of the Model-without and Model-with were 0.676 (95% CI 0.646-0.707) and 0.695 (95% CI 0.667-0.724). The AUCs between both models were significantly different (P= .013). Calibration was good for both models. Both the Model-with and Model-without indicators of missing values showed good calibration and fair discrimination, where the Model-with performed better. Our models showed competitive performance to well-established fall-risk-screening tools, and they have the advantage of being based on routinely collected data. This may substantially reduce the burden on nurses, compared with nonautomatic fall-risk-screening tools.

Using Multiple Imputation to Account for the Uncertainty Due to Missing Data in the Context of Factor Retention.

Although parallel analysis has been found to be an accurate method for determining the number of factors in many conditions with complete data, its application under missing data is limited. The existing literature recommends that, after using an appropriate multiple imputation method, researchers either apply parallel analysis to every imputed data set and use the number of factors suggested by most of the data copies or average the correlation matrices across all data copies, followed by applying the parallel analysis to the average correlation matrix. Both approaches for pooling the results provide a single suggested number without reflecting the uncertainty introduced by missing values. The present study proposes the use of an alternative approach, which calculates the proportion of imputed data sets that result in k (k = 1, 2, 3 . . .) factors. This approach will inform applied researchers of the degree of uncertainty due to the missingness. Results from a simulation experiment show that the proposed method can more likely suggest the correct number of factors when missingness contributes to a large amount of uncertainty.

Predictors of antimicrobial use in intensive care unit patients

Background: Identification of predictors of antibiotic use can inform targeted antimicrobial stewardship initiatives and can account for sources of bias in before-and-after interventional stewardship studies. To date, no study has identified clinical predictors of antimicrobial use within intensive care units (ICUs), where antimicrobial resistance is most prevalent and problematic. Methods: As part of an ongoing prospective, single-arm, pilot feasibility trial of an ICU diagnostic stewardship intervention, we performed a nested retrospective cohort study to explore associations between patient clinical variables and ICUs antimicrobial use and resistance rates (AURs). We included all patients hospitalized in 3 ICUs (surgical, medical, and cardiac) from 2017 to 2021 at Michigan Medicine, a large, tertiary-care, academic medical center. Data were extracted from the electronic medical record using a structured query. Admission-level data were captured, including patient demographics, medical comorbidities, International Classification of Disease, Tenth Revision (ICD-10) admission diagnoses, as well as calendar day-level data including vital signs, clinical and microbiologic laboratory data, measures of acute severity of illness, ventilator–supplemental oxygen metrics, and procedural interventions using current procedural terminology (CPT) codes. ICU AURs were defined as total antibiotic days of therapy per patient per 100 ICU days. Associations between clinical variables and ICU AURs were calculated as rate ratios (RRs). Multiple imputation using fully conditional specification was performed to create 25 imputation data sets. Negative binomial regression models were constructed for each data set using backward selection. Variables retained in &gt;50% of models were included in a final multivariate model. Results: In total, 15,177 ICU patient admissions were captured. Age, sex assigned at birth, and race did not independently associate with ICU AURs. Comorbidities, medical interventions, admission diagnoses, and laboratory data that independently associated with ICU-AURs are shown in Table 1. The clinical variables most strongly associated with increased ICU-AURs were pneumonia (RR, 1.55; 95% CI, 1.451.64), bacteremia (RR, 1.35; 95% CI, 1.25– 1.46), intraabdominal infection (RR, 1.35; 95% CI, 1.18–1.55), SOFA score (RR, 1.27; 95% CI, 1.14–1.42), abnormal WBC (RR, 1.26; 95% CI, 1.20–1.32), and immunocompromised status (RR, 1.20; 95% CI, 1.10–1.31). Clinical variables most strongly associated with decreased ICU-AURs were cardiac ICU (RR, 0.56; 95% CI, 0.52–0.60), COVID-19 (RR, 0.62; 95% CI, 0.56–0.70), and receipt of an invasive nonsurgical procedure (RR, 0.90; 95% CI, 0.82–0.98). Conclusions: In this single-center retrospective cohort study, several clinical variables were independently associated with ICU-AURs. These results may be used to identify patient subgroups for potentially high-yield ICU-based stewardship interventions and to account for sources of bias in before-and-after studies of ICU-based stewardship interventions.Disclosures: None