Unordered Categorical Variables Research Articles

Psychosocial Health Among Young Adults With Kidney Failure: A Longitudinal Follow-up of the SPEAK (Surveying Patients Experiencing Young Adult Kidney Failure) Study

Rationale & ObjectiveThere have been no longitudinal studies examining the evolution of psychosocial health of young adults with kidney failure as they age. We aimed to address this in the SPEAK-2 [Surveying Patients Experiencing young Adult Kidney failure-2] study. Study Design5-year follow-up longitudinal survey of the original SPEAK cohort. Setting& Participants: 16- to 30-year-olds in the UK receiving KRT between 2015-2017 who participated in the SPEAK study. ExposuresKidney failure, and KRT modality. OutcomesPsychosocial health and lifestyle behaviours. Analytical ApproachWithin-cohort changes in psychosocial health were analysed using the paired t-test, Wilcoxon signed-rank test and McNemar’s test. We compared responses to the age-matched population, and examined the impact of changes in KRT modality on psychological health, using linear regression for continuous outcome variables and logistic, ordered logistic and multinomial logistic regression for binary, ordered categorical and unordered categorical variables respectively. ResultsWe obtained 158 survey responses; 129 had previously responded to SPEAK. Of these, 90% had a kidney transplant. Compared to the general population, respondents were less likely to be married or have children and were more likely to be living with their parents. They had nearly 15 times greater odds of being unable to work due to health (OR, 14.41; 95% CI, 8.0-26.01; P<0.001). Respondents had poorer quality of life and mental wellbeing, and were more likely to report psychological problems (OR 5.37; 95% CI 3.45-8.35; P<0.001). A negative association between remaining on or moving to dialysis and psychosocial health was observed, although this was attenuated when controlling for the psychosocial state in SPEAK. LimitationsLow response rate resulting in imprecise and potentially biased estimates, and impact of Covid-19 pandemic while survey was active on psychosocial health. ConclusionsYoung adults with kidney failure have persistent poorer psychosocial health compared to their healthy peers as they age. Our findings also suggest a potential causal relationship between KRT modality and psychosocial health.

Prognostic value of preoperative body mass index for diabetic patients with non-metastasis gastric cancer: a single center experience

AimThis study was designed to investigate the prognostic effect of preoperative body mass index (BMI) for Type 2 diabetes mellitus (T2DM) patients with non-metastasis gastric cancer (GC) who underwent D2 gastrectomy.MethodsT2DM patients with pT1–4bN0–3bM0 GC were retrospectively collected in Department of Gastrointestinal Surgical Oncology, Fujian Cancer Hospital & Fujian Medical University Cancer Hospital from January, 2000 to December, 2010. These patients underwent D2 radical resection of the stomach combined with regional lymphadenectomy. Chi-square test was used to analyze unordered categorical variables and ranked data, followed by Kaplan–Meier analysis as well as Cox regression models to detect risk factors for survival outcomes. In addition, the cut-off point was determined by the X-tile program. All analyses were carried out using survival package of R and SPSS Software.ResultsA total of 302 T2DM patients with pT1–4bN0–3bM0 GC were collected and analyzed. The cut-off points of BMI, identified by the X-tile program, was 19 kg/m2. Patients with low BMI (< 19 kg/m2) had a higher percentage of advanced T stage (T4a and T4b), more advanced TNM stage (stage IIIA, IIIB and IIIC), and more elevated level of serum carcinoembryonic antigen (CEA), compared to those with high BMI (> 19 kg/m2) (all P < 0.05). In the low BMI subgroup, the 5-year overall survival rate was 39.02%, which was as high as 58.11% in the high BMI subgroup (P < 0.05). In the multivariate Cox regression model revealed that IIIC stage (OR = 3.101), N3b stage (OR = 3.113) were the most important prognostic indicators, followed by pretreatment BMI (OR = 2.136).ConclusionLow preoperative BMI (< 19 kg/m2) was a poor prognostic marker for T2DM patients with pT1–4bN0–3bM0 GC.

Use of the World Health Organization Checklist—Swedish Health Care Professionals' Experience: A Mixed-Method Study

The purpose was to describe health care professionals' experience of using the World Health Organization (WHO) surgical safety checklist. A descriptive cross-sectional mixed-method study, including health care professionals from two clinics at the same university hospital in the western part of Sweden was conducted. Data were collected from one hundred ninety-six health care professionals using a self-administered questionnaire that contained 12 questions. The Mantel-Haenszel and Pearson χ2 tests were used for ordered and unordered categorical variables. The text analysis was inspired by Malterud. The most statistically significant difference between the clinics related to the responsibility for administering the checklist (P=.0010) and always using the checklist in all emergency situations (P=.045). Among the health care professionals who were educated and trained in using checklists, 63% and 65.5% stated that the checklist was adapted to the department. Sixty four percent stated that the assistant nurses were responsible for implementing the checklist. The health care professionals also mentioned a large number of positive and negative aspects of using the WHO checklist. The health care professionals had difficulties using the WHO checklist in the Swedish health care system. More research is needed to determine why specific items are overlooked and whether these items could form the basis of the further elaboration of a modified checklist. An increased understanding of why the checklist is important, as well as updated knowledge on the content could produce greater compliance and thereby increase patient safety.

Multiple imputation of unordered categorical missing data: A comparison of the multivariate normal imputation and multiple imputation by chained equations

Missing data are common in survey data sets. Enrolled subjects do not often have data recorded for all variables of interest. The inappropriate handling of them may negatively affect the inferences drawn. Therefore, special attention is needed when analysing incomplete data. The multivariate normal imputation (MVNI) and the multiple imputation by chained equations (MICE) have emerged as the best techniques to deal with missing data. The former assumes a normal distribution of the variables in the imputation model and the latter fills in missing values taking into account the distributional form of the variables to be imputed. This study examines the performance of these methods when data are missing at random on unordered categorical variables treated as predictors in the regression models. First, a survey data set with no missing values is used to generate a data set with missing at random observations on unordered categorical variables. Then, the two methods are separately used to impute the missing values of the generated data set. Their performance is compared in terms of bias and standard errors of the estimates from the regression models that determine the association between the woman’s contraceptive methods use status and her marital status, controlling for the region of origin. The baseline data used is the 2007 Demographic and Health Survey (DHS) data set from the Democratic Republic of Congo. The findings indicate that although the MVNI relies on the statistical parametric theory, it produces more accurate estimates than MICE for nonordered categorical variables.

Multiple Imputation for Continuous and Categorical Data: Comparing Joint Multivariate Normal and Conditional Approaches

We consider the relative performance of two common approaches to multiple imputation (MI): joint multivariate normal (MVN) MI, in which the data are modeled as a sample from a joint MVN distribution; and conditional MI, in which each variable is modeled conditionally on all the others. In order to use the multivariate normal distribution, implementations of joint MVN MI typically assume that categories of discrete variables are probabilistically constructed from continuous values. We use simulations to examine the implications of these assumptions. For each approach, we assess (1) the accuracy of the imputed values; and (2) the accuracy of coefficients and fitted values from a model fit to completed data sets. These simulations consider continuous, binary, ordinal, and unordered-categorical variables. One set of simulations uses multivariate normal data, and one set uses data from the 2008 American National Election Studies. We implement a less restrictive approach than is typical when evaluating methods using simulations in the missing data literature: in each case, missing values are generated by carefully following the conditions necessary for missingness to be “missing at random” (MAR). We find that in these situations conditional MI is more accurate than joint MVN MI whenever the data include categorical variables.

Patterns of Practice in the Prescription of Palliative Radiotherapy for the Treatment of Thoracic Symptoms of Lung Cancer at the Rapid Response Radiotherapy Program between 2006 and 2012

Objective: Radiation therapy can often be used for palliation of thoracic symptoms in patients presenting with locally advanced lung cancer or lung metastases. The aim of this study was to examine whether patterns of practice in prescription of palliative thoracic radiation therapy have changed over time in the Rapid Response Radiotherapy Program. Secondary outcomes were factors that may have influenced the treatment regimen prescribed, including patient, disease, and organisational factors. Methods: This study was a retrospective review of a prospective database of patients with locally advanced lung cancer or lung metastases referred to the Rapid Response Radiotherapy Program for thoracic symptoms between 1 July 2006 and 30 April 2012. Patient demographics, and organisational and disease factors were descriptively analysed. Differences in proportions between unordered categorical variables were examined using chi-squared test. Univariate logistic regression analysis and backward stepwise selection procedure were used to determine the most significant factors in prescription practice. Results: A total of 175 courses of palliative thoracic radiation therapy were prescribed. The median age of the patients was 71 years, and the median Karnofsky Performance Status was 60. The most commonly prescribed treatment regimen was 20 Gy in 5 fractions (20 Gy/5), which made up 64% of all the courses prescribed. There was a significant increase in frequency of the prescription of 20 Gy/5 over time (p = 0.02). The site of radiation (disease factor) and years of certification for independent practice of the treating radiation oncologist (organisational factor) were also significant factors in the prescription of 20 Gy/5 over time (both p = 0.02). Conclusion: A significant increase in the prescription of 20 Gy/5 was observed over time. However, the prescription of a higher dose fractionation schedule for patients with a higher performance status, as seen in other clinical trials and guidelines, was not observed. Future studies should further explore other possible factors such as patient survival, preference, comorbidities, and disease burden that may influence the dose fractionation prescribed.

Patterns of Practice in the Prescription of Palliative Radiotherapy for the Treatment of Bone Metastases at the Rapid Response Radiotherapy Program between 2005 and 2012

We examined whether patterns of practice in the prescription of palliative radiation therapy for bone metastases had changed over time in the Rapid Response Radiotherapy Program (rrrp). After reviewing data from August 1, 2005, to April 30, 2012, we analyzed patient demographics, diseases, organizational factors, and possible reasons for the prescription of various radiotherapy fractionation schedules. The chi-square test was used to detect differences in proportions between unordered categorical variables. Univariate logistic regression analysis and the simple Fisher exact test were also used to determine the factors most significant to choice of dose-fractionation schedule. During the study period, 2549 courses of radiation therapy were prescribed. In 65% of cases, a single fraction of radiation therapy was prescribed, and in 35% of cases, multiple fractions were prescribed. A single fraction of radiation therapy was more frequently prescribed when patients were older, had a prior history of radiation, or had a prostate primary, and when the radiation oncologist had qualified before 1990. For patients with bone metastasis, a single fraction of radiation therapy was prescribed with significantly greater frequency.

Metamodels for mixed variables based on moving least squares

Surrogate-based optimization has become a major field in engineering design, due to its capacity to handle complex systems involving expensive simulations. However, the majority of general-purpose surrogates (also called metamodels) are restricted to continuous variables, although versatile problems involve additional types of variables (discrete, integer, and even categorical to model technological options). Therefore, the main contribution of this paper consists in the development of metamodels specifically dedicated to handle mixed variables, in particular continuous and unordered categorical variables, and their comparison with state-of-the-art approaches. This task is performed in three steps: (i) considering an appropriate parametrization (integer mapping, regular simplex, dummy, effect codings) for the mixed variable design vector; (ii) defining metrics to compare pairs of design vectors; (iii) carrying out an ordinary or moving least square regression scheme based on the parametrization and metric previously defined. The proposed metamodels have been tested on six analytical benchmark test cases, and applied to the structural finite element analysis model of a rigid frame characterized by continuous and categorical variables. In particular, it is demonstrated that using a standard regular simplex representation for the nominal categorical variables usually outperforms a direct conversion of the nominal parameters to integer values, while offering an efficient and systematic way to encompass all types of variables in a common framework. It is also shown that the choice of a given variable representation has a higher impact on the results than the selected scheme (ordinary or moving least squares), or than the metric used for calculating distances between samples.

Simplex Factor Models for Multivariate Unordered Categorical Data

Gaussian latent factor models are routinely used for modeling of dependence in continuous, binary, and ordered categorical data. For unordered categorical variables, Gaussian latent factor models lead to challenging computation and complex modeling structures. As an alternative, we propose a novel class of simplex factor models. In the single-factor case, the model treats the different categorical outcomes as independent with unknown marginals. The model can characterize flexible dependence structures parsimoniously with few factors, and as factors are added, any multivariate categorical data distribution can be accurately approximated. Using a Bayesian approach for computation and inferences, a Markov chain Monte Carlo (MCMC) algorithm is proposed that scales well with increasing dimension, with the number of factors treated as unknown. We develop an efficient proposal for updating the base probability vector in hierarchical Dirichlet models. Theoretical properties are described, and we evaluate the approach through simulation examples. Applications are described for modeling dependence in nucleotide sequences and prediction from high-dimensional categorical features.

Role of Adjuvant Chemotherapy in Patients With Early Stage Uterine Papillary Serous Cancer

Uterine papillary serous carcinoma (UPSC) is an aggressive subtype of endometrial cancer. We studied survival outcomes in patients with stages I/II UPSC. A retrospective, multi-institutional study of patients with stages I/II UPSC was conducted. Patients underwent surgical staging followed by observation, adjuvant platinum-based chemotherapy (CT), or radiation therapy (RT). Continuous variables were compared via Wilcoxon rank sum test; Fisher exact test was used for the unordered categorical variables. Kaplan-Meier curves were used to estimate survival. Thirty-nine women were diagnosed with stage I (n = 30) or II (n = 9) UPSC, with a median follow-up of 52 months. Of the 26 patients who did not receive adjuvant CT, 9 developed recurrences and 8 died of their disease. Of the 10 patients with no myometrial invasion who did not receive adjuvant CT, 3 developed recurrences and died. Of the 7 patients who underwent RT, 2 developed distant recurrences and died. Of the 13 patients who underwent CT, 1 developed vaginal recurrence. The 5-year overall (OS) and progression-free survival (PFS) rates for the adjuvant CT group were 100% and 92%, respectively, compared with 69% and 65% for those who did not receive CT (P = 0.002 OS, P = 0.002 PFS). The 5-year OS and PFS rates for RT group were both 71%. Patients with stages I/II UPSC are at significant risk for distant recurrence and poor survival. Platinum-based adjuvant CT may decrease recurrence rate and improve survival in women with early and well-staged UPSC.

Nonparametric Bayes Modeling of Multivariate Categorical Data

Modeling of multivariate unordered categorical (nominal) data is a challenging problem, particularly in high dimensions and cases in which one wishes to avoid strong assumptions about the dependence structure. Commonly used approaches rely on the incorporation of latent Gaussian random variables or parametric latent class models. The goal of this article is to develop a nonparametric Bayes approach, which defines a prior with full support on the space of distributions for multiple unordered categorical variables. This support condition ensures that we are not restricting the dependence structure a priori. We show this can be accomplished through a Dirichlet process mixture of product multinomial distributions, which is also a convenient form for posterior computation. Methods for nonparametric testing of violations of independence are proposed, and the methods are applied to model positional dependence within transcription factor binding motifs.

Bayesian semiparametric analysis of structural equation models with mixed continuous and unordered categorical variables

Recently, structural equation models (SEMs) have been applied for analyzing interrelationships among observed and latent variables in biological and medical research. Latent variables in these models are typically assumed to have a normal distribution. This article considers a Bayesian semparametric SEM with covariates, and mixed continuous and unordered categorical variables, in which the explanatory latent variables in the structural equation are modeled via an appropriate truncated Dirichlet process with a stick-breaking procedure. Results obtained from a simulation study and an analysis of a real medical data set are presented to illustrate the methodology.

Non‐linear structural equation models with correlated continuous and discrete data

Structural equation models (SEMs) have been widely applied to examine interrelationships among latent and observed variables in social and psychological research. Motivated by the fact that correlated discrete variables are frequently encountered in practical applications, a non-linear SEM that accommodates covariates, and mixed continuous, ordered, and unordered categorical variables is proposed. Maximum likelihood methods for estimation and model comparison are discussed. One real-life data set about cardiovascular disease is used to illustrate the methodologies.

Has the Pattern of Practice in the Prescription of Radiotherapy for the Palliation of Thoracic Symptoms Changed Between 1999 and 2006 at the Rapid Response Radiotherapy Program?

Eleven randomized controlled trials (RCT) comparing various radiotherapy (RT) schedules for locally advanced lung cancer published since 1991 found no difference in palliation of intrathoracic symptoms. The most commonly prescribed schedule by Canadian Radiation Oncologists (RO) (20 Gy in five fractions [20 Gy/5]), when first evaluated versus 10 Gy/1 in a 2002 RCT, showed a significant survival benefit. A subsequent RCT assessing 20 Gy/5 found worse survival versus 16 Gy/2. This study examines whether the RT prescription for lung cancer palliation in the Rapid Response Radiotherapy Program (RRRP) has changed over time. Chart review was conducted for patients treated with palliative thoracic RT across three periods (1999-2006). Patient demographics, tumor, treatment, and organizational factors were analyzed descriptively. Chi-square test was used to detect differences in proportions between unordered categorical variables. Continuous variables were tested using analysis of variance. Multivariate logistic regression was used to identify independent predictors of RT schedule prescribed. A total of 117 patients received 121 courses of palliative thoracic RT. The most common dose (20 Gy/5) comprised 65% of courses in 1999, 68% in 2003, and 60% in 2005-2006 (p = 0.76). The next most common dose was 30 Gy/10 (13%). Overall, the median survival was 14.9 months, independent of RT schedule (p = 0.68). Multivariate analysis indicated palliative chemotherapy and certification year of RO were significant predictors of prescription of 20 Gy/5. RT schedule for palliation of intrathoracic symptoms did not mirror the results of sequential, conflicting RCTs, suggesting that factors other than the literature influenced practice patterns in palliative thoracic RT.

Mosaic Displays for Multi-Way Contingency Tables

Abstract Mosaic displays represent the counts in a contingency table by tiles whose size is proportional to the cell count. This graphical display for categorical data generalizes readily to multi-way tables. This article discusses extensions of the mosaic display to highlight patterns of deviations from various models for categorical data. First, we introduce the use of color and shading to represent sign and magnitude of standardized residuals from a specified model. For unordered categorical variables, we show how the perception of patterns of association can be enhanced by reordering the categories. Second, we introduce sequential mosaics of marginal subtables, together with sequential models for these tables. For a class of sequential models of joint independence, the individual mosaics provide a graphic representation of a partition of the overall likelihood ratio G2 for complete independence in the full table into portions attributable to hypotheses about the marginal subtables.