MSMpred: interactive modelling and prediction of individual evolution via multistate models.

BACKGROUND: Modelling the course of a disease regarding severe events and identifying prognostic factors is of great clinical relevance. Multistate models (MSM) can be used to describe diseases or processes that change over time using different states and the transitions between them. Specifically, they are useful to analyse a disease with an increasing degree of severity, that may precede death. The complexity of these models changes depending on the number of states and transitions taken into account. Due to that, a web tool has been developed making easier to work with those models. RESULTS: MSMpred is a web tool created with the shiny R package that has two main features: 1) to allow to fit a MSM from specific data; 2) to predict the clinical evolution for a given subject. To fit the model, the data to be analysed must be upload in a prespecified format. Then, the user has to define the states and transitions as well as the covariates (e.g., age or gender) involved in each transition. From this information, the app returns histograms or barplots, as appropriate, to represent the distributions of the selected covariates and boxplots to show the patient' length of stay (for uncensored data) in each state. To make predictions, the values of selected covariates from a new subject at baseline has to be provided. From these inputs, the app provides some indicators of the subject's evolution such as the probability of 30-day death or the most likely state at a fixed time. Furthermore, visual representations (e.g., the stacked transition probabilities plot) are given to make predictions more understandable. CONCLUSIONS: MSMpred is an intuitive and visual app that eases the work of biostatisticians and facilitates to the medical personnel the interpretation of MSMs.

Characteristics and Outcomes by Ceiling of Care of Subjects Hospitalized with COVID-19 During Four Waves of the Pandemic in a Metropolitan Area: A Multicenter Cohort Study.

INTRODUCTION: The profiles of patients with COVID-19 have been widely studied, but little is known about differences in baseline characteristics and in outcomes between subjects with a ceiling of care assigned at hospital admission and subjects without a ceiling of care. The aim of this study is to compare, by ceiling of care, clinical features and outcomes of hospitalized subjects during four waves of COVID-19 in a metropolitan area in Catalonia. METHODS: Observational study conducted during the first (March-April 2020), second (October-November 2020), third (January-February 2021), and fourth wave (July-August 2021) of COVID-19 in five centers of Catalonia. All subjects were adults (> 18 years old) hospitalized with a proven SARS-CoV-2 infection and with therapeutic ceiling of care assessed by the attending physician at hospital admission. RESULTS: A total of 5813 subjects were analyzed. Subjects with a ceiling of care were mainly older (difference in median age of 20 years), with more comorbidities (Charlson index 3 points higher) and with fewer clinical signs at baseline than patients without a ceiling of care. Some features of their clinical profiles changed among waves. There were differences in treatments received during hospital admission across waves, but not between subjects with and without a ceiling of care. Subjects with a ceiling of care had a death incidence more than four times the death incidence of subjects a without a ceiling of care (risk ratio (RR) ranging from 3.5 in the first wave to almost 6 in the third and fourth). Incidence of severe pneumonia and complications for subjects with a ceiling of care was around 1.5 times the incidence in subjects without a ceiling of care. DISCUSSION: Analysis of hospitalized subjects with SARS-CoV-2 infection should be stratified according to therapeutic ceiling of care to avoid bias and outcome misestimation.

SARS-Cov-2 incubation period according to vaccination status during the fifth COVID-19 wave in a tertiary-care center in Spain: a cohort study.

BACKGROUND: The incubation period of an infectious disease is defined as the elapsed time between the exposure to the pathogen and the onset of symptoms. Although both the mRNA-based and the adenoviral vector-based vaccines have shown to be effective, there have been raising concerns regarding possible decreases in vaccine effectiveness for new variants and variations in the incubation period. METHODS: We conducted a unicentric observational study at the Hospital Universitari de Bellvitge, Barcelona, using a structured telephone survey performed by trained interviewers to estimate the incubation period of the SARS-CoV-2 Delta variant in a cohort of Spanish hospitalized patients. The distribution of the incubation period was estimated using the generalized odds-rate class of regression models. RESULTS: From 406 surveyed patients, 242 provided adequate information to be included in the analysis. The median incubation period was 2.8 days (95%CI: 2.5-3.1) and no differences between vaccinated and unvaccinated patients were found. Sex and age are neither shown not to be significantly related to the COVID-19 incubation time. CONCLUSIONS: Knowing the incubation period is crucial for controlling the spread of an infectious disease: decisions on the duration of the quarantine or on the periods of active monitoring of people who have been at high risk of exposure depend on the length of the incubation period. Furthermore, its probability distribution is a key element for predicting the prevalence and the incidence of the disease.

Circulating carotenoids are associated with favorable lipid and fatty acid profiles in an older population at high cardiovascular risk.

Carotenoid intake has been reported to be associated with improved cardiovascular health, but there is little information on actual plasma concentrations of these compounds as biomarkers of cardiometabolic risk. The objective was to investigate the association between circulating plasma carotenoids and different cardiometabolic risk factors and the plasma fatty acid profile. This is a cross-sectional evaluation of baseline data conducted in a subcohort (106 women and 124 men) of an ongoing multi-factorial lifestyle trial for primary cardiovascular prevention. Plasma concentrations of carotenoids were quantified by liquid chromatography coupled to mass spectrometry. The associations between carotenoid concentrations and cardiometabolic risk factors were assessed using regression models adapted for interval-censored variables. Carotenoid concentrations were cross-sectionally inversely associated with serum triglyceride concentrations [-2.79 mg/dl (95% CI: -4.25, -1.34) and -5.15 mg/dl (95% CI: -7.38, -2.93), p-values = 0.0002 and <0.00001 in women and men, respectively], lower levels of plasma saturated fatty acids [-0.09% (95% CI: -0.14, -0.03) and -0.15 % (95% CI: -0.23, -0.08), p-values = 0.001 and 0.0001 in women and men, respectively], and higher levels of plasma polyunsaturated fatty acids [(0.12 % (95% CI: -0.01, 0.25) and 0.39 % (95% CI: 0.19, 0.59), p-values = 0.065 and 0.0001 in women and men, respectively] in the whole population. Plasma carotenoid concentrations were also associated with higher plasma HDL-cholesterol in women [0.47 mg/dl (95% CI: 0.23, 0.72), p-value: 0.0002], and lower fasting plasma glucose in men [-1.35 mg/dl (95% CI: -2.12, -0.59), p-value: 0.001].

A class of two-sample nonparametric statistics for binary and time-to-event outcomes.

We propose a class of two-sample statistics for testing the equality of proportions and the equality of survival functions. We build our proposal on a weighted combination of a score test for the difference in proportions and a weighted Kaplan-Meier statistic-based test for the difference of survival functions. The proposed statistics are fully non-parametric and do not rely on the proportional hazards assumption for the survival outcome. We present the asymptotic distribution of these statistics, propose a variance estimator, and show their asymptotic properties under fixed and local alternatives. We discuss different choices of weights including those that control the relative relevance of each outcome and emphasize the type of difference to be detected in the survival outcome. We evaluate the performance of these statistics with small sample sizes through a simulation study and illustrate their use with a randomized phase III cancer vaccine trial. We have implemented the proposed statistics in the R package SurvBin, available on GitHub (https://github.com/MartaBofillRoig/SurvBin).

Design of phase III trials with long-term survival outcomes based on short-term binary results.

Pathologic complete response (pCR) is a common primary endpoint for a phase II trial or even accelerated approval of neoadjuvant cancer therapy. If granted, a two-arm confirmatory trial is often required to demonstrate the efficacy with a time-to-event outcome such as overall survival. However, the design of a subsequent phase III trial based on prior information on the pCR effect is not straightforward. Aiming at designing such phase III trials with overall survival as primary endpoint using pCR information from previous trials, we consider a mixture model that incorporates both the survival and the binary endpoints. We propose to base the comparison between arms on the difference of the restricted mean survival times, and show how the effect size and sample size for overall survival rely on the probability of the binary response and the survival distribution by response status, both for each treatment arm. Moreover, we provide the sample size calculation under different scenarios and accompany them with the R package survmixer where all the computations have been implemented. We evaluate our proposal with a simulation study, and illustrate its application through a neoadjuvant breast cancer trial.

A nonparametric test for the association between longitudinal covariates and censored survival data.

Many biomedical studies focus on the association between a longitudinal measurement and a time-to-event outcome while quantifying this association by means of a longitudinal-survival joint model. In this article we propose using the $LLR$ test - a longitudinal extension of the log-rank test statistic given by Peto and Peto (1972) - to provide evidence of a plausible association between a time-to-event outcome (right- or interval-censored) and a time-dependent covariate. As joint model methods are complex and hard to interpret, it is wise to conduct a preliminary test such as $LLR$ for checking the association between both processes. The $LLR$ statistic can be expressed in the form of a weighted difference of hazards, yielding a broad class of weighted log-rank test statistics known as $LWLR$, which allow a specific emphasis along the time axis of the effects of the time-dependent covariate on the survival. The asymptotic distribution of $LLR$ is derived by means of a permutation approach under the assumption that the censoring mechanism is independent of the survival time and the longitudinal covariate. A simulation study is conducted to evaluate the performance of the test statistics $LLR$ and $LWLR$, showing that the empirical size is close to the nominal significance level and that the power of the test depends on the association between the covariates and the survival time. A data set together with a toy example are used to illustrate the $LLR$ test. The data set explores the study Epidemiology of Diabetes Interventions and Complications (Sparling and others, 2006) which includes interval-censored data. A software implementation of our method is available on github (https://github.com/RamonOller/LWLRtest).

A new approach for sizing trials with composite binary endpoints using anticipated marginal values and accounting for the correlation between components.

Composite binary endpoints are increasingly used as primary endpoints in clinical trials. When designing a trial, it is crucial to determine the appropriate sample size for testing the statistical differences between treatment groups for the primary endpoint. As shown in this work, when using a composite binary endpoint to size a trial, one needs to specify the event rates and the effect sizes of the composite components as well as the correlation between them. In practice, the marginal parameters of the components can be obtained from previous studies or pilot trials; however, the correlation is often not previously reported and thus usually unknown. We first show that the sample size for composite binary endpoints is strongly dependent on the correlation and, second, that slight deviations in the prior information on the marginal parameters may result in underpowered trials for achieving the study objectives at a pre-specified significance level. We propose a general strategy for calculating the required sample size when the correlation is not specified and accounting for uncertainty in the marginal parameter values. We present the web platform CompARE to characterize composite endpoints and to calculate the sample size just as we propose in this paper. We evaluate the performance of the proposal with a simulation study and illustrate it by means of a real case study using CompARE.

Selection of composite binary endpoints in clinical trials.

The choice of a primary endpoint is an important issue when designing a clinical trial. It is common to use composite endpoints as a primary endpoint because it increases the number of observed events, captures more information and is expected to increase the power. However, combining events that have no similar clinical importance and have different treatment effects makes the interpretation of the results cumbersome and might reduce the power of the corresponding tests. Gómez and Lagakos proposed the ARE (asymptotic relative efficiency) method to choose between a composite or one of its components as primary endpoint comparing the efficacy of a treatment based on the times to each of these endpoints. The aim of this paper is to expand the ARE method to binary endpoints. We show that the ARE method depends on six parameters including the degree of association between components, event proportion, and effect of therapy given by the corresponding odds ratio of the single endpoints. A case study is presented to illustrate the methodology. We conclude with efficient guidelines for discerning which could be the best suited primary endpoint given anticipated parameters.

Dos métodos para analizar la evolución de la incidencia de consumo de heroína y cocaína en Barcelona / Two methods to analyze trends in the incidence of heroin and cocaine use in Barcelona [Spain]

Objetivos: Mostrar 2 métodos estadísticos para estimar la evolución de la incidencia de consumo de heroína y cocaína en Barcelona. Métodos: Los inicios de tratamiento por consumo de heroína y cocaína del Sistema de Información de Drogas de Barcelona entre 1991 y 2003. Se seleccionaron 4.367 sujetos con un primer tratamiento por heroína y 2.147 por cocaína. Se ha empleado el método Reporting Delay Adjustment (RDA) y modelos logarítmicos lineales (MLL), considerando para el RDA los inicios de tratamiento cuyo inicio de consumo era entre 1991 y 2003, y para el MLL entre 1967 y 2003 para la heroína y entre 1971 y 2003 para la cocaína. Además, para cada droga y método se estudió la distribución del período de latencia (PL) (años entre el primer consumo y el primer tratamiento). Resultados: Según la distribución del PL, los consumidores de heroína tardan menos tiempo en iniciar su primer tratamiento que los de cocaína, indistintamente del método empleado. En general, la incidencia estimada de consumo de heroína en Barcelona disminuyó progresivamente desde 1982. En cambio, para la cocaína aumentó rápidamente hasta 1998, y fue irregular posteriormente. A principios de los noventa la incidencia de consumo de cocaína empezó a ser importante, años antes de manifestarse como problemática. Conclusión: El RDA subestima la incidencia respecto al MLL, pero la incidencia de heroína con MLL puede estar sesgada antes de 1991 por cambios en la oferta de tratamiento. Aunque la incidencia estimada se refiere a los sujetos que hacen tratamiento alguna vez en la vida, su evolución ayuda a prever acciones futuras

Objectives: To describe 2 statistical methods for estimating trends in the incidence of heroin and cocaine use in Barcelona. Methods: Admissions for treatment of heroin and cocaine consumption recorded by the Barcelona Drug Information System between 1991 and 2003 were used. We selected 4,367 subjects initiating treatment for the first time for heroin use, and 2,147 for cocaine use. Two statistical techniques were employed: Reporting Delay Adjustment (RDA) and the Log-linear Model (LLM). RDA was used in subjects who initiated drug consumption between 1991 and 2003, and LLM for those who began heroin use between 1967 and 2003 and cocaine use between 1971 and 2003. In addition, for each drug and method the latency period (LP) was determined (years between first consumption and first treatment). Results: Comparison of the distributions of the LP for each drug revealed that heroin users initiated treatment for the first time sooner than cocaine users, regardless of the method employed. In general, the estimated incidence of heroin use in Barcelona fell progressively after 1982. In contrast, the incidence of cocaine use rose rapidly until 1998, and has been irregular since. The incidence of cocaine use began to be substantial in the early 1990s, but took several years to manifest itself as problematic. Conclusion: The estimated incidence was underestimated by RDA compared with LLM, but the incidence of heroin use could be biased before 1991 due to changes in treatment provisions. Although the estimated incidence is relative to individuals who are admitted for treatment at some time in their life, trends in incidence can be used to plan future actions