Conditional Inference Framework Research Articles

Identifying Predictors of First-Line Subcutaneous TNF-Inhibitor Persistence in Patients with Inflammatory Arthritis: A Decision Tree Analysis by Indication.

Treatment persistence is a proxy for efficacy, safety and patient satisfaction, and a switch in treatment or treatment discontinuation has been associated with increased indirect and direct costs in inflammatory arthritis (IA). Hence, there are both clinical and economic incentives for the identification of factors associated with treatment persistence. Until now, studies have mainly leveraged traditional regression analysis, but it has been suggested that novel approaches, such as statistical learning techniques, may improve our understanding of factors related to treatment persistence. Therefore, we set up a study using nationwide Swedish high-coverage administrative register data with the objective to identify patient groups with distinct persistence of subcutaneous tumor necrosis factor inhibitor (SC-TNFi) treatment in IA, using recursive partitioning, a statistical learning algorithm. IA was defined as a diagnosis of rheumatic arthritis (RA), ankylosing spondylitis/unspecified spondyloarthritis (AS/uSpA) or psoriatic arthritis (PsA). Adult swedish biologic-naïve patients with IA initiating biologic treatment with a SC-TNFi (adalimumab, etanercept, certolizumab or golimumab) between May 6, 2010, and December 31, 2017. Treatment persistence of SC-TNFi was derived based on prescription data and a defined standard daily dose. Patient characteristics, including age, sex, number of health care contacts, comorbidities and treatment, were collected at treatment initiation and 12months before treatment initiation. Based on these characteristics, we used recursive partitioning in a conditional inference framework to identify patient groups with distinct SC-TNFi treatment persistence by IA diagnosis. A total of 13,913 patients were included. Approximately 50% had RA, while 27% and 23% had AS/uSpA and PsA, respectively. The recursive partitioning algorithm identified sex and treatment as factors associated with SC-TNFi treatment persistence in PsA and AS/uSpA. Time on treatment in the groups with the lowest treatment persistence was similar across all three indications (9.5-11.3months), whereas there was more variation in time on treatment across the groups with the highest treatment persistence (18.4-48.9months). Women have low SC-TNFi treatment persistence in PsA and AS/uSpA whereas male sex and golimumab are associated with high treatment persistence in these indications. The factors associated with treatment persistence in RA were less distinct but may comprise disease activity and concurrent conventional systemic disease-modifying anti-rheumatic drug (DMARD) treatment.

A Machine Learning Model for Prediction of Amputation in Diabetics.

Diabetic foot ulcer (DFU) and the resulting lower extremity amputation are associated with a poor survival prognosis. The objective of this study is to generate a model for predicting the probability of major amputation in hospitalized patients with DFU. The National Inpatient Sample (NIS) database from 2008 to 2014 was used to select patients with DFU, who were then further divided by major amputation status. International Classification of Diseases, Ninth Edition, Clinical Modification (ICD-9-CM) and Agency for Healthcare Research and Quality (AHRQ) comorbidity codes were used to compare patient characteristics. For the descriptive statistics, the Student t test, the χ2 test, and the Spearman correlation were utilized. The five most predictive variables were identified. A decision tree model (CTREE) based on conditional inference framework algorithm and a random forest model were used to develop the algorithm. A total of 326 853 inpatients with DFU were identified, and 5.9% underwent major amputation. The top five contributory variables (all with P < .001) were gangrene (odds ratio [OR] = 11.8, 95% confidence interval [CI] = 11.5-12.2), peripheral vascular disease (OR = 2.9, 95% CI = 2.8-3.0), weight loss (OR = 2.6, 95% CI = 2.5-2.8), systemic infection (OR = 2.5, 95% CI = 2.4-2.53), and osteomyelitis (OR = 1.7, 95% CI = 1.6-1.73). The model performance of the training data was 77.7% (76.1% sensitivity and 79.3% specificity) and of the testing data was 77.8% (76.2% sensitivity and 79.4% specificity). The model was further validated with boosting and random forest models which demonstrated similar performance and area under the curve (AUC) (0.84, 95% CI = 0.83-0.85). Utilizing machine learning methods, we have developed a clinical algorithm that predicts the risk of major lower extremity amputation for inpatients with diabetes with 77.8% accuracy.

A survival tree for interval-censored failure time data

&lt;abstract&gt;&lt;p&gt;Interval-censored failure time data as a general type of survival data often arises in medicine and other applied fields. Survival tree is a flexible predictive method for survival data because no specific assumptions are required.&lt;/p&gt; &lt;p&gt;Generalized Log-Rank Test have good power with parameters for interval-censored failure time data. We construct a special test statistic of Generalized Log-Rank Tests, and propose a new survival tree with hyper-parameter by combining the test statistic with Conditional Inference Framework for interval-censored failure time data. The effect of tuning hyper-parameter are discussed and hyper-parameter tuning allows the tree method to be more general and flexible. Thus the tree method either improve upon or remain competitive with existing tree method for interval-censored failure time data-ICtree, which is a special case of ours. An extensive simulation is executed to assess the predictive performance of our tree methods. Finally, the tree methods are applied to a tooth emergence data.&lt;/p&gt;&lt;/abstract&gt;

The Impact of the COVID-19 Pandemic on Nursing Care: A Cross-Sectional Survey-Based Study.

The COVID-19 pandemic has had a severe impact on nursing care. This cross-sectional survey-based study compared aspects of nursing care and nurses’ satisfaction with care provided before and during the first wave of the COVID-19 pandemic. A total of 936 registered nurses (RNs) rated the frequency with which they performed fundamental care, nursing techniques, patient education, symptom management, and nurse–patient relationships before and during the pandemic. A recursive partitioning for ordered multivariate response in a conditional inference framework approach was applied. More frequent fundamental cares were associated with their frequency before the pandemic (p < 0.001), caring for COVID-19 patients (p < 0.001), and workplace reassignment (p = 0.004). Caring for COVID-19 patients (p < 0.001), workplace reassignment (p = 0.030), and caring for ≤7.4 COVID-19 patients (p = 0.014) increased nursing techniques. RNs in high-intensity COVID-19 units (p = 0.002) who educated patients before the pandemic, stopped this task. RNs caring for COVID-19 patients reported increased symptom management (p < 0.001), as did RNs caring for more non-COVID-19 patients (p = 0.037). Less frequent nurse–patient relationships before the pandemic and working in high-intensity COVID-19 units decreased nurse–patient relationships (p = 0.002). Despite enormous challenges, nurses continued to provide a high level of care. Ensuring the appropriate deployment and education of nurses is crucial to personalize care and to maintain nurses’ satisfaction with the care provided.

The FICI grading for a dacryocystorhinostomy ostium.

The objective of this study is to propose a simple grading to assess the health of a dacryocystorhinostomy (DCR) ostium. Prospective case series of 237 Ostia evaluated following dacryocystorhinostomy were included in the study. All the ostia were assessed for 10 parameters in detail using the earlier published DCR ostium scoring (DOS scoring). Each of the 10 parameters were scored individually, and final DOS scores were obtained. The anatomical and functional outcomes of each of the surgery were noted. The most significant parameters that influenced the success were determined using the binary recursive partitioning in a conditional inference framework. Data management and statistical tests were performed using the statistical "Software R" and the library "partykit" toolkits. The most significant DCR ostium parameters that influenced the outcomes were anatomical and functional fluorescein dye transit (p < 0.001), dynamicity of internal common opening (ICO) on the blink (p < 0.001), end-on threats to ICO from granuloma, membrane or a synechia, and cicatricial closure of the Ostia. Each of these significant factors (FICI - Fluorescein endoscopy dye test, ICO dynamicity, Cicatricial ostium closure and ICO threats) were assigned simple scoring and the final scores were graded from 0 (poor DCR ostium) to +5 (excellent DCR ostium). FICI grading can guide the physician on the health status of the DCR ostium and the possible need for intervention. FICI DCR ostium grading is a simple, easy to perform, and physician-friendly system. More studies with larger sample size would help in further validating the FICI grading.

Assessment of ambient aerosol sources in two important Atlantic Rain Forest hotspots in the surroundings of a megacity

Between 2010 and 2015, an assessment of ambient aerosol sources was carried in two unique fragments of the Atlantic Rain Forest in the surroundings of the Metropolitan Region of Rio de Janeiro (MRRJ). Airborne particulate matter samples were collected at Serra dos Órgãos National Park (43°04′42.1″W and 22°29′16.9″S) and Mário Xavier National Forest (43°42′21.8″W and 22°43′21.7″S). At the former site, PM10 samples were collected, while at the latter TSP samples were collected due to a particular interest on the preservation of an endangered endemic species of tree frog (Physalaemus soaresi). Elemental composition, inorganic and organic water-soluble compounds were analyzed along with local meteorology variables in order to provide the most relevant variables for particulate matter prediction and its potential sources. For TSP, the main predictors were NO3− >Mn >Rad (Global radiation) >Ca2+ >Precipitation >Mg 2+. For PM 10, the main predictors were Gust (Gust wind speed) >NO 3− >Ca2+ >Zn >Cu >Ti. Furthermore, trends in the particulate matter were analyzed considering the prevailing winds and sources were evaluated whether intermittent or continuous, using the conditional bivariate probability function (CBPF). With the use of CBPF, recent developed machine learning algorithms (Conditional inference trees – CIT, and Random Forests using a conditional inference framework), and other standard data analysis techniques tuned for air quality exercises, we provide an example case for planning and evaluation of environmental risk assessment by stakeholders.

An ensemble method for interval-censored time-to-event data.

Interval-censored data analysis is important in biomedical statistics for any type of time-to-event response where the time of response is not known exactly, but rather only known to occur between two assessment times. Many clinical trials and longitudinal studies generate interval-censored data; one common example occurs in medical studies that entail periodic follow-up. In this article, we propose a survival forest method for interval-censored data based on the conditional inference framework. We describe how this framework can be adapted to the situation of interval-censored data. We show that the tuning parameters have a non-negligible effect on the survival forest performance and guidance is provided on how to tune the parameters in a data-dependent way to improve the overall performance of the method. Using Monte Carlo simulations, we find that the proposed survival forest is at least as effective as a survival tree method when the underlying model has a tree structure, performs similarly to an interval-censored Cox proportional hazards model fit when the true relationship is linear, and outperforms the survival tree method and Cox model when the true relationship is nonlinear. We illustrate the application of the method on a tooth emergence data set.

Regression trees for longitudinal data with baseline covariates

ABSTRACTLongitudinal changes in a population of interest are often heterogeneous and influenced by a combination of baseline factors. In such cases, classical linear mixed effects models [Laird NM, Ware JH. Random-effects models for longitudinal data. Biometrics. 1982;38:963–974.] for the mean structure provide poor fit to the data. We propose regression tree methodology for the longitudinal data identifying and characterizing homogeneous subgroups. Currently available regression tree construction methods are either limited to a repeated measures scenario or combine the heterogeneity among subgroups with the random inter-subject variability. We propose a longitudinal classification and regression tree (LongCART) algorithm under conditional inference framework that overcomes these limitations utilizing a two-step approach. The LongCART first selects the partitioning variable via a parameter instability test and then finds the optimal split for the selected partitioning variable. Thus, at each node, the decision of further splitting is type I error controlled, guarding against variable selection bias, over-fitting and spurious splitting. We obtained asymptotic results for the proposed instability test and examined its finite sample behavior through simulation studies. Comparative performance of LongCART algorithm was evaluated empirically via simulation studies. Finally, we applied LongCART to study the longitudinal changes in choline levels among HIV-positive patients.

Childhood maltreatment as risk factor for lifetime depression: The role of different types of experiences and sensitive periods

Childhood maltreatment (CM) is associated with a higher risk for the development of lifetime major depressive disorder (MDD). Initial evidence further suggests a significant role of type and timing of CM on mental health. Thus, this study aimed at investigating which particular subtypes of CM at which age of exposure are the best predictors for the development of lifetime MDD. Three months postpartum, N = 285 women were interviewed with the German interview version of the Maltreatment and Abuse Chronology of Exposure Scale which allows a broad and differentiated assessment of CM and enables the assessment of ages at which CM occured. Lifetime MDD was diagnosed with the Structured Clinical Interview for mental disorders according to the Diagnostic and Statistical Manual of Mental Disorders (5th ed.; DSM-5). To identify the strongest predictor, random forest analyses embedded in a conditional inference framework were applied. Analyses revealed that emotional subtypes of CM by parents, experienced during adolescence, showed the strongest prediction for lifetime MDD. Furthermore, women with a lifetime diagnosis of MDD reported a significantly greater CM severity, more different subtypes of CM, and a longer duration of CM compared to women without lifetime MDD. Given the “neglect” of emotional CM in previous studies, emotional neglect and abuse should be considered more frequently in the prevention and treatment of MDD.

Sequential feature selection and inference using multi-variate random forests.

Random forest (RF) has become a widely popular prediction generating mechanism. Its strength lies in its flexibility, interpretability and ability to handle large number of features, typically larger than the sample size. However, this methodology is of limited use if one wishes to identify statistically significant features. Several ranking schemes are available that provide information on the relative importance of the features, but there is a paucity of general inferential mechanism, particularly in a multi-variate set up. We use the conditional inference tree framework to generate a RF where features are deleted sequentially based on explicit hypothesis testing. The resulting sequential algorithm offers an inferentially justifiable, but model-free, variable selection procedure. Significant features are then used to generate predictive RF. An added advantage of our methodology is that both variable selection and prediction are based on conditional inference framework and hence are coherent. We illustrate the performance of our Sequential Multi-Response Feature Selection approach through simulation studies and finally apply this methodology on Genomics of Drug Sensitivity for Cancer dataset to identify genetic characteristics that significantly impact drug sensitivities. Significant set of predictors obtained from our method are further validated from biological perspective. https://github.com/jomayer/SMuRF. souparno.ghosh@ttu.edu. Supplementary data are available at Bioinformatics online.

Survival trees for interval-censored survival data.

Interval-censored data, in which the event time is only known to lie in some time interval, arise commonly in practice, for example, in a medical study in which patients visit clinics or hospitals at prescheduled times and the events of interest occur between visits. Such data are appropriately analyzed using methods that account for this uncertainty in event time measurement. In this paper, we propose a survival tree method for interval-censored data based on the conditional inference framework. Using Monte Carlo simulations, we find that the tree is effective in uncovering underlying tree structure, performs similarly to an interval-censored Cox proportional hazards model fit when the true relationship is linear, and performs at least as well as (and in the presence of right-censoring outperforms) the Cox model when the true relationship is not linear. Further, the interval-censored tree outperforms survival trees based on imputing the event time as an endpoint or the midpoint of the censoring interval. We illustrate the application of the method on tooth emergence data.

Impact of gene expression profiles on clinical predictors of survival in patients (pts) with BRAFV600-mutated metastatic melanoma (mM).

9556 Background: Treatment of BRAFV600-mutated mm with V and C+V has clinical benefit. Prior analyses established PFS/OS prognostic models in pts with mm based on clinical variables. This exploratory analysis evaluated the impact of immune and cell cycle gene signatures (GS) from baseline mm (Wongchenko et al, Pigment Cell Melanoma Res2015;28:822) on survival outcomes and associated prognostic models. Methods: Data from all eligible pts with GS data in the BRIM-2, -3, -7, and coBRIM studies were pooled for analysis. The independent effect of GS on PFS/OS outcomes was tested by multivariate Cox proportional hazards models. Recursive partitioning (RP) for censored response variables in a conditional inference framework was performed in the pooled dataset to model relationships between prespecified covariates, GS, and PFS/OS. Prognostic subgroups identified by the model for all pooled pts were applied to pooled treatment cohorts (dacarbazine [D], V, and C+V). Results: GS data were available for 608 pts across pooled studies. Immune GS was associated with improved survival vs cell cycle for PFS (HR = 0.75, 95% CI 0.62–0.89, P= 0.0015) and OS (HR = 0.74, 95% CI 0.61–0.90, P= 0.0024) in multivariate models. HR point estimates were consistent across treatment cohorts. PFS RP models identified baseline LDH, tumor size, and GS as prognostic factors, giving 4 groups with distinct outcomes. GS was not identified as a prognostic factor in OS RP models. C+V improved PFS/OS outcomes vs V across prognostic subgroups. Immune GS was significantly (one-sided Cochran-Armitage trend test) more prevalent in the favorable prognostic groups previously defined for PFS and OS for the V ( P&lt; 0.0001 for both) and C+V cohorts ( P= 0.0240 and P= 0.0269, respectively), but not the D cohort. Conclusions: PFS/OS was significantly improved for pts with immune GS. RP modeling confirmed the importance of GS as an independent prognostic factor for PFS but not OS. The preferential association of immune GS with favorable prognostic subgroups in the C+V and V but not D cohorts and known effects of MAPK signaling on immune response merit further exploration in prospective studies.

Pretreatment 18F-FDG PET Textural Features in Locally Advanced Non-Small Cell Lung Cancer: Secondary Analysis of ACRIN 6668/RTOG 0235.

In a secondary analysis of American College of Radiology Imaging Network (ACRIN) 6668/RTOG 0235, high pretreatment metabolic tumor volume (MTV) on (18)F-FDG PET was found to be a poor prognostic factor for patients treated with chemoradiotherapy for locally advanced non-small cell lung cancer (NSCLC). Here we utilize the same dataset to explore whether heterogeneity metrics based on PET textural features can provide additional prognostic information. Patients with locally advanced NSCLC underwent (18)F-FDG PET prior to treatment. A gradient-based segmentation tool was used to contour each patient's primary tumor. MTV, maximum SUV, and 43 textural features were extracted for each tumor. To address overfitting and high collinearity among PET features, the least absolute shrinkage and selection operator (LASSO) method was applied to identify features that were independent predictors of overall survival (OS) after adjusting for MTV. Recursive binary partitioning in a conditional inference framework was utilized to identify optimal thresholds. Kaplan-Meier curves and log-rank testing were used to compare outcomes among patient groups. Two hundred one patients met inclusion criteria. The LASSO procedure identified 1 textural feature (SumMean) as an independent predictor of OS. The optimal cutpoint for MTV was 93.3 cm(3), and the optimal SumMean cutpoint for tumors above 93.3 cm(3) was 0.018. This grouped patients into three categories: low tumor MTV (n = 155; median OS, 22.6 mo), high tumor MTV and high SumMean (n = 23; median OS, 20.0 mo), and high tumor MTV and low SumMean (n = 23; median OS, 6.2 mo; log-rank P < 0.001). We have described an appropriate methodology to evaluate the prognostic value of textural PET features in the context of established prognostic factors. We have also identified a promising feature that may have prognostic value in locally advanced NSCLC patients with large tumors who are treated with chemoradiotherapy. Validation studies are warranted.

Combat high or traumatic stress: violent offending is associated with appetitive aggression but not with symptoms of traumatic stress.

Former members of armed groups in eastern DR Congo had typically witnessed, experienced, and perpetrated extreme forms of violence. Enhanced trauma-related symptoms had been shown in prior research. But also lashing out in self-defense is a familiar response to threat defined as reactive aggression. Another potential response is appetitive aggression, in which the perpetration of excessive violence is perceived as pleasurable (combat high). What roles do these forms of aggressive behavior play in modern warfare and how are they related to posttraumatic stress symptoms? To answer the question, we sought to determine predictors for appetitive aggressive and trauma-related mental illness, and investigated the frequency of psychopathological symptoms for high- and low-intensity conflict demobilization settings. To this end, we interviewed 213 former members of (para)military groups in the eastern Democratic Republic of Congo in regard to their combat exposure, posttraumatic stress, appetitive aggression, depression, suicidality, and drug dependence. Random forest regression embedded in a conditional inference framework revealed that perpetrated violent acts are not necessarily stressful. In fact, the experience of violent acts that typically implicated salient cues of hunting (e.g., blood, suffering of the victim, etc.) had the strongest association with an appetite for aggression. Furthermore, the number of lifetime perpetrated violent acts was the most important predictor of appetitive aggression. However, the number of perpetrated violent acts did not significantly affect the posttraumatic stress. Greater intensity of conflict was associated with more severe posttraumatic stress symptoms and depression. Psychotherapeutic interventions that address appetitive aggression in addition to trauma-related mental illness, including drug dependence, therefore seem indispensible for a successful reintegration of those who fought in the current civil wars.

Conditional Maximum Likelihood Estimation Following a Group Sequential Test

AbstractWe consider estimation after a group sequential test. An estimator that is unbiased or has small bias may have substantial conditional bias (Troendle and Yu, 1999, Coburger and Wassmer, 2001). In this paper we derive the conditional maximum likelihood estimators of both the primary parameter and a secondary parameter, and investigate their properties within a conditional inference framework. The method applies to both the usual and adaptive group sequential test designs. (© 2004 WILEY‐VCH Verlag GmbH &amp; Co. KGaA, Weinheim)