A framework for longitudinal latent factor modelling of treatment response in clinical trials with applications to Psoriatic Arthritis and Rheumatoid Arthritis.

OBJECTIVE: Clinical trials involve the collection of a wealth of data, comprising multiple diverse measurements performed at baseline and follow-up visits over the course of a trial. The most common primary analysis is restricted to a single, potentially composite endpoint at one time point. While such an analytical focus promotes simple and replicable conclusions, it does not necessarily fully capture the multi-faceted effects of a drug in a complex disease setting. Therefore, to complement existing approaches, we set out here to design a longitudinal multivariate analytical framework that accepts as input an entire clinical trial database, comprising all measurements, patients, and time points across multiple trials. METHODS: Our framework composes probabilistic principal component analysis with a longitudinal linear mixed effects model, thereby enabling clinical interpretation of multivariate results, while handling data missing at random, and incorporating covariates and covariance structure in a computationally efficient and principled way. RESULTS: We illustrate our approach by applying it to four phase III clinical trials of secukinumab in Psoriatic Arthritis (PsA) and Rheumatoid Arthritis (RA). We identify three clinically plausible latent factors that collectively explain 74.5% of empirical variation in the longitudinal patient database. We estimate longitudinal trajectories of these factors, thereby enabling joint characterisation of disease progression and drug effect. We perform benchmarking experiments demonstrating our method's competitive performance at estimating average treatment effects compared to existing statistical and machine learning methods, and showing that our modular approach leads to relatively computationally efficient model fitting. CONCLUSION: Our multivariate longitudinal framework has the potential to illuminate the properties of existing composite endpoint methods, and to enable the development of novel clinical endpoints that provide enhanced and complementary perspectives on treatment response.

New strategies for confirmatory testing of secondary hypotheses on combined data from multiple trials.

BACKGROUND: Pivotal evidence of efficacy of a new drug is typically generated by (at least) two clinical trials which independently provide statistically significant and mutually corroborating evidence of efficacy based on a primary endpoint. In this situation, showing drug effects on clinically important secondary objectives can be demanding in terms of sample size requirements. Statistically efficient methods to power for such endpoints while controlling the Type I error are needed. METHODS: We review existing strategies for establishing claims on important but sample size-intense secondary endpoints. We present new strategies based on combined data from two independent, identically designed and concurrent trials, controlling the Type I error at the submission level. We explain the methodology and provide three case studies. RESULTS: Different strategies have been used for establishing secondary claims. One new strategy, involving a protocol planned analysis of combined data across trials, and controlling the Type I error at the submission level, is particularly efficient. It has already been successfully used in support of label claims. Regulatory views on this strategy differ. CONCLUSIONS: Inference on combined data across trials is a useful approach for generating pivotal evidence of efficacy for important but sample size-intense secondary endpoints. It requires careful preparation and regulatory discussion.

Response to anti-IL17 therapy in inflammatory disease is not strongly impacted by genetic background.

Response to the anti-IL17 monoclonal antibody secukinumab is heterogeneous, and not all participants respond to treatment. Understanding whether this heterogeneity is driven by genetic variation is a key aim of pharmacogenetics and could influence precision medicine approaches in inflammatory diseases. Using changes in disease activity scores across 5,218 genotyped individuals from 19 clinical trials across four indications (psoriatic arthritis, psoriasis, ankylosing spondylitis, and rheumatoid arthritis), we tested whether genetics predicted response to secukinumab. We did not find any evidence of association between treatment response and common variants, imputed HLA alleles, polygenic risk scores of disease susceptibility, or cross-disease components of shared genetic risk. This suggests that anti-IL17 therapy is equally effective regardless of an individual's genetic background, a finding that has important implications for future genetic studies of biological therapy response in inflammatory diseases.

Analysis of serum proteomics data identifies a quantitative association between beta-defensin 2 at baseline and clinical response to IL-17 blockade in psoriatic arthritis.

OBJECTIVES: Despite several effective targeted therapies, biomarkers that predict whether a patient with psoriatic arthritis (PsA) will respond to a particular treatment are currently lacking. METHODS: We analysed proteomics data from serum samples of nearly 2000 patients with PsA in placebo-controlled phase-III clinical trials of the interleukin-17 inhibitor secukinumab. To discover predictive biomarkers of clinical response, we used statistical learning with controlled feature selection. The top candidate was validated using an ELISA and was separately assessed in a trial of almost 800 patients with PsA treated with secukinumab or the tumour necrosis factor inhibitor adalimumab. RESULTS: Serum levels of beta-defensin 2 (BD-2) at baseline were found to be robustly associated with subsequent clinical response (eg, American College of Rheumatology definition of 20%, 50% and 70% improvement) to secukinumab, but not to placebo. This finding was validated in two independent clinical studies not used for discovery. Although BD-2 is known to be associated with psoriasis severity, the predictivity of BD-2 was independent of baseline Psoriasis Area and Severity Index. The association between BD-2 and response to secukinumab was observed as early as 4 weeks and maintained up to 52 weeks. BD-2 was also found to predict response to treatment with adalimumab. Unlike in PsA, BD-2 was not predictive of response to secukinumab in rheumatoid arthritis. CONCLUSIONS: In PsA, BD-2 at baseline is quantitatively associated with clinical response to secukinumab. Patients with high levels of BD-2 at baseline reach and sustain higher rates of clinical response after treatment with secukinumab.

A Bayesian approach to clinical trial designs in dermatology with multiple simultaneous treatments per subject and multiple raters.

We consider the statistical analysis of clinical trial designs with multiple simultaneous treatments per subject and multiple raters. The work is motivated by a clinical research project in dermatology where different hair removal techniques were assessed based on a within-subject comparison. We assume that clinical outcomes are assessed by multiple raters as continuous or categorical scores, e.g. based on images, comparing two treatments on the subject-level in a pairwise manner. In this setting, a network of evidence on relative treatment effects is generated, which bears strong similarities to the data underlying a network meta-analysis of clinical trials. We therefore build on established methodology for complex evidence synthesis and propose a Bayesian approach to estimate relative treatment effects and to rank the treatments. The approach is, in principle, applicable to situations with any number of treatment arms and/or raters. As a major advantage, all available data is brought into a network and analyzed in one single model, which ensures consistent results among the treatment comparisons. We obtain operating characteristics via simulation and illustrate the method with a real clinical trial example.

DMC reports in the 21st century: towards better tools for decision-making.

Data Monitoring Committees (DMCs) have the important task to protect the safety of current and future patients during the conduct of a clinical study. Unfortunately, their work is often made difficult by voluminous DMC reports that are poorly structured and difficult to digest. In this article, we suggest improved solutions. Starting from a principled approach and building upon previous proposals, we offer concrete and easily understood displays, including related computer code. While leveraging modern tools, the most important is that these displays support the DMC's workflow in answering the relevant questions of interest. We hope that the adoption of these proposals can ease the task of DMCs, and importantly, lead to better decision-making for the benefit of patients.

Secukinumab in moderate-to-severe hidradenitis suppurativa (SUNSHINE and SUNRISE): week 16 and week 52 results of two identical, multicentre, randomised, placebo-controlled, double-blind phase 3 trials.

BACKGROUND: Few therapeutic options are available for patients with moderate-to-severe hidradenitis suppurativa. We aimed to assess the efficacy of secukinumab in patients with moderate-to-severe hidradenitis suppurativa in two randomised trials. METHODS: SUNSHINE and SUNRISE were identical, multicentre, randomised, placebo-controlled, double-blind phase 3 trials done in 219 primary sites in 40 countries. Patients aged 18 years old or older with the capacity to provide written informed consent and with moderate-to-severe hidradenitis suppurativa (defined as a total of ≥5 inflammatory lesions affecting ≥2 distinct anatomical areas) for at least 1 year were eligible for inclusion. Included patients also agreed to daily use of topical over-the-counter antiseptics on the areas affected by hidradenitis suppurativa lesions while on study treatment. Patients were excluded if they had 20 or more fistulae at baseline, had ongoing active conditions requiring treatment with prohibited medication (eg, systemic biological immunomodulating treatment, live vaccines, or other investigational treatments), or met other exclusion criteria. In both trials, patients were randomly assigned (1:1:1) by means of interactive response technology to receive subcutaneous secukinumab 300 mg every 2 weeks, subcutaneous secukinumab 300 mg every 4 weeks, or subcutaneous placebo all via a 2 mL prefilled syringe in a double-dummy method as per treatment assignment. The primary endpoint was the proportion of patients with a hidradenitis suppurativa clinical response, defined as a decrease in abscess and inflammatory nodule count by 50% or more with no increase in the number of abscesses or in the number of draining fistulae compared with baseline, at week 16, assessed in the overall population. Hidradenitis suppurativa clinical response was calculated based on the number of abscesses, inflammatory nodules, draining fistulae, total fistulae, and other lesions in the hidradenitis suppurativa affected areas. Safety was assessed by evaluating the presence of adverse events and serious adverse events according to common terminology criteria for adverse events, which were coded using Medical Dictionary for Regulatory Activities terminology. Both the SUNSHINE, NCT03713619, and SUNRISE, NCT03713632, trials are registered with ClinicalTrials.gov. FINDINGS: Between Jan 31, 2019, and June 7, 2021, 676 patients were screened for inclusion in the SUNSHINE trial, of whom 541 (80%; 304 [56%] women and 237 [44%] men; mean age 36·1 years [SD 11·7]) were included in the analysis (181 [33%] in the secukinumab every 2 weeks group, 180 [33%] in the secukinumab every 4 weeks group, and 180 [33%] in the placebo group). Between the same recruitment dates, 687 patients were screened for inclusion in the SUNRISE trial, of whom 543 (79%; 306 [56%] women and 237 [44%] men; mean age 36·3 [11·4] years) were included in the analysis (180 [33%] in the secukinumab every 2 weeks group, 180 [33%] in the secukinumab every 4 weeks group, and 183 [34%] in the placebo group). In the SUNSHINE trial, significantly more patients in the secukinumab every 2 weeks group had a hidradenitis suppurativa clinical response (rounded average number of patients with response in 100 imputations, 81·5 [45%] of 181 patients) compared with the placebo group (60·7 [34%] of 180 patients; odds ratio 1·8 [95% CI 1·1-2·7]; p=0·0070). However, there was no significant difference between the number of patients in the secukinumab every 4 weeks group (75·2 [42%] of 180 patients) and the placebo group (1·5 [1·0-2·3]; p=0·042). Compared with the placebo group (57·1 [31%] of 183 patients), significantly more patients in the secukinumab every 2 weeks group (76·2 [42%] of 180 patients; 1·6 [1·1-2·6]; p=0·015) and the secukinumab every 4 weeks group (83·1 [46%] of 180 patients; 1·9 [1·2-3·0]; p=0·0022) had a hidradenitis suppurativa clinical response in the SUNRISE trial. Patient responses were sustained up to the end of the trials at week 52. The most common adverse event by preferred term up to week 16 was headache in both the SUNSHINE (17 [9%] patients in the secukinumab every 2 weeks group, 20 [11%] in the secukinumab every 4 weeks group, and 14 [8%] in the placebo group) and SUNRISE (21 [12%] patients in the secukinumab every 2 weeks group, 17 [9%] in the secukinumab every 4 weeks group, and 15 [8%] in the placebo group) trials. No study-related deaths were reported up to week 16. The safety profile of secukinumab in both trials was consistent with that previously reported, with no new or unexpected safety findings detected. INTERPRETATION: When given every 2 weeks, secukinumab was clinically effective at rapidly improving signs and symptoms of hidradenitis suppurativa with a favourable safety profile and with sustained response up to 52 weeks of treatment. FUNDING: Novartis Pharma.

Comparing algorithms for characterizing treatment effect heterogeneity in randomized trials.

The identification and estimation of heterogeneous treatment effects in biomedical clinical trials are challenging, because trials are typically planned to assess the treatment effect in the overall trial population. Nevertheless, the identification of how the treatment effect may vary across subgroups is of major importance for drug development. In this work, we review some existing simulation work and perform a simulation study to evaluate recent methods for identifying and estimating the heterogeneous treatments effects using various metrics and scenarios relevant for drug development. Our focus is not only on a comparison of the methods in general, but on how well these methods perform in simulation scenarios that reflect real clinical trials. We provide the R package benchtm that can be used to simulate synthetic biomarker distributions based on real clinical trial data and to create interpretable scenarios to benchmark methods for identification and estimation of treatment effect heterogeneity.

Lower Limbs are the Most Difficult-to-Treat Body Region of Patients with Psoriasis: Pooled Analysis of CLEAR and CLARITY Studies of Secukinumab Versus Ustekinumab by Body Region.

BACKGROUND: The impact of psoriasis, response to treatment, and patients' perceptions of treatment satisfaction vary by body area. OBJECTIVES: We aimed to evaluate the level of response in lower limbs versus other body regions in patient with moderate to severe psoriasis treated with secukinumab and ustekinumab. METHODS: Data were pooled from CLEAR and CLARITY trials, which included patients with moderate to severe plaque psoriasis (Psoriasis Area and Severity Index [PASI] score of ≥ 12) aged ≥ 18 years and a diagnosis of ≥ 6 months before randomisation. Patients received either secukinumab 300 mg or ustekinumab 45/90 mg. The PASI 100 responders and mean PASI scores at weeks 4, 12, 16, 28, 40 and 52 in the head and neck, trunk, upper limbs and lower limbs were measured. RESULTS: At baseline, analysis of PASI scores for each body region revealed that the lower limbs were the most severely affected body region in both treatment arms (mean PASI scores: secukinumab 24.0; ustekinumab 24.4). For both drugs, the highest clearance rates at week 52 were achieved in the trunk (secukinumab 85.2% vs ustekinumab 68.7%) and head and neck (80.7% vs 68.9%), followed by the upper limbs (72.6% vs 61.9%) and lower limbs (68.1% vs 57.2%). At week 52, the mean PASI score was higher in the lower limbs in both treatment arms versus other body regions. CONCLUSIONS: Lower limbs were the most severely affected and most difficult-to-treat regions in patients with psoriasis. Consistent with the individual results of both studies, secukinumab demonstrated numerically faster and higher skin clearance than ustekinumab in all body regions. CLINICAL TRIAL REGISTRATION: CLEAR: NCT02074982; CLARITY: NCT02826603.

Innovative Pediatric Development for Secukinumab in Psoriasis: Faster Patient Access, Reduction of Patients on Control.

This paper describes the innovative pediatric development strategy of secukinumab in psoriasis. Initially, this pediatric development had been planned early during the adult development program. This is common for most development programs as it is required by law in major regions worldwide. At that time, only limited experience with secukinumab's mechanism of action was available, especially in the pediatric population for which no data had been collected yet. Pediatric extrapolation frameworks were not in use. As a result, a standard pediatric program had initially been proposed requiring two relatively large placebo-controlled pivotal trials. However, as experience with the use of secukinumab grew, so did experience with-and regulatory openness to-innovative quantitative approaches, such as extrapolation. It became clear that the pediatric development of secukinumab could be substantially accelerated by these novel ideas. Here, we report how these concepts were implemented in the pediatric development program of secukinumab in psoriasis, using a wide range of innovative quantitative techniques. This strategy led to the removal of the placebo arm from one trial, a substantial reduction of the overall sample size in this program, and a significantly faster approval of secukinumab for the pediatric psoriasis population.

Nail Involvement as a Predictor of Differential Treatment Effects of Secukinumab Versus Ustekinumab in Patients with Moderate to Severe Psoriasis.

INTRODUCTION: Patients with plaque psoriasis may experience varying levels of treatment response to different biologics, based on phenotypic characteristics and underlying genetic factors. Nail psoriasis is a common manifestation of psoriasis (approx. 50% of patients) and has been linked to the human leukocyte antigen-C*0602 (HLA-C*0602) allele, which in turn has been associated with differential treatment responses to certain drugs. Here we investigate whether nail involvement in patients with psoriasis can predict differential skin responses to two biologics with different modes of action, namely secukinumab (anti-interleukin-17A) and ustekinumab (anti-interleukin-12/23), to ultimately guide treatment choice. METHODS: Data were pooled from the CLEAR and CLARITY studies and stratified post hoc by nail involvement status at baseline. Psoriasis Area and Severity Index (PASI) 75 and 90 responses over 52 weeks and absolute PASI ≤ 3, ≤ 1, and 0 values at weeks 16 and 52, were assessed. RESULTS: Based on the medical history, 30.4% (269/886) of the patients in the secukinumab arm and 29.7% (265/891) of patients in the ustekinumab arm presented with nail involvement. Nail involvement status had little to no impact on the efficacy of secukinumab, as comparable responses were achieved for patients with and without nail involvement in terms of PASI 75/90, ≤ 3, and 0 responses; slightly lower PASI ≤ 1 reponses were achieved in patients with nail involvement. In the ustekinumab arm, patients with nail involvement achieved lower responses across all endpoints. CONCLUSIONS: These findings indicate that nail involvement can serve as an observable prognostic factor for efficacy in skin psoriasis treatment and guide the choice between secukinumab and ustekinumab. TRIAL REGISTRATION: CLEAR: NCT02074982; CLARITY: NCT02826603.

Instability and results after non-operative treatment of large anterior glenoid rim fractures: is there a correlation between fragment size or displacement and recurrence?

INTRODUCTION: There is little data available on non-operative treatment of anterior glenoid rim fractures (GRF). Nothing is known about fracture size and displacement in comparison to clinical outcomes and instability in a mainly middle-aged patient population. The aim of this study was to demonstrate the results of non-operative treatment in anterior glenoid rim fractures with the special focus on potential instability/recurrence. METHODS: The inclusion criteria were non-operatively treated anterior GRF of at least ≥ 5 mm width using the age- and gender-matched Constant/Murley score (a.-/g.-CMS) and the Western Ontario Instability Index (WOSI). Radiographic parameters (fracture morphology, displacement, major tuberosity fractures and Hill-Sachs lesion using initial CT and radiographs) and the proportion of the fractured glenoid were detected (2D-CT-circle-method) and osteoarthritis (A.P. and axial radiographs) was classified according to Samilson/Prieto. Proportion of fractured glenoid and medial displacement were correlated with the recurrence rate and the clinical scores. RESULTS: N = 36 patients could be followed-up after a mean of 4.4 years [12-140 month, average age: 58 (± 13, 33-86) years]. The a.-/g.-CMS was 93 (± 11, 61-100) points, and the WOSI was 81% (± 22%, 35-100%) on average. The mean intraarticular displacement was 4 mm (± 3 mm; 0-14 mm). The 2D-circle-method showed a mean glenoid fracture involvement of 21% (± 11, 10-52%). Two cases of frozen shoulders and one case with biceps pathology were associated with the trauma. Within the followed-up patient group re-instability has occurred in n = 2 patients (6%) within the first two weeks after trauma. Osteoarthritis was found in n = 11 cases. There was no correlation between the scores and the fracture size/displacement [(a.-/g.-CMS vs. displacement: r = - 0.08; p = 0.6; vs. size: r = - 0.29; p = 0.2); (WOSI vs. displacement: r = - 0.14; p = 0.4; vs. size: r = - 0.37; p = 0.06)], but very large (≥ 21%) fractures with displacement ≥ 4 mm showed slightly worse results without significant difference (a.-/g.-CMS p = 0.2; WOSI p = 0.2). The apprehension test was negative in all patients at final follow-up. CONCLUSION: Non-operative treatment of anterior GRF was associated with overall good results within a mainly middle-aged larger patient group. Re-instability is rare and is not associated with fragment size but can occur in the first weeks after trauma. Size and dislocation of the fracture is not a criterion for the prognosis of potential instability. LEVEL OF EVIDENCE: Level IV, retrospective case series.

Win odds: An adaptation of the win ratio to include ties.

The win ratio, a recently proposed measure for comparing the benefit of two treatment groups, allows ties in the data but ignores ties in the inference. In this article, we highlight some difficulties that this can lead to, and we propose to focus on the win odds instead, a modification of the win ratio which takes ties into account. We construct hypothesis tests and confidence intervals for the win odds, and we investigate their properties through simulations and in a case study. We conclude that the win odds should be preferred over the win ratio.

Biomarker-Based Classification and Localization of Renal Lesions Using Learned Representations of Histology-A Machine Learning Approach to Histopathology.

Several deep learning approaches have been proposed to address the challenges in computational pathology by learning structural details in an unbiased way. Transfer learning allows starting from a learned representation of a pretrained model to be directly used or fine-tuned for a new domain. However, in histopathology, the problem domain is tissue-specific and putting together a labelled data set is challenging. On the other hand, whole slide-level annotations, such as biomarker levels, are much easier to obtain. We compare two pretrained models, one histology-specific and one from ImageNet on various computational pathology tasks. We show that a domain-specific model (HistoNet) contains richer information for biomarker classification, localization of biomarker-relevant morphology within a slide, and the prediction of expert-graded features. We use a weakly supervised approach to discriminate slides based on biomarker level and simultaneously predict which regions contribute to that prediction. We employ multitask learning to show that learned representations correlate with morphological features graded by expert pathologists. All of these results are demonstrated in the context of renal toxicity in a mechanistic study of compound toxicity in rat models. Our results emphasize the importance of histology-specific models and their knowledge representations for solving a wide range of computational pathology tasks.

Effective Visual Communication for the Quantitative Scientist.

Effective visual communication is a core competency for pharmacometricians, statisticians, and, more generally, any quantitative scientist. It is essential in every step of a quantitative workflow, from scoping to execution and communicating results and conclusions. With this competency, we can better understand data and influence decisions toward appropriate actions. Without it, we can fool ourselves and others and pave the way to wrong conclusions and actions. The goal of this tutorial is to convey this competency. We posit three laws of effective visual communication for the quantitative scientist: have a clear purpose, show the data clearly, and make the message obvious. A concise "Cheat Sheet," available on https://graphicsprinciples.github.io, distills more granular recommendations for everyday practical use. Finally, these laws and recommendations are illustrated in four case studies.

Multistate modeling and simulation of patient trajectories after allogeneic hematopoietic stem cell transplantation to inform drug development.

We present a case study for developing clinical trial scenarios in a complex progressive disease with multiple events of interest. The idea is to first capture the course of the disease in a multistate Markov model, and then to simulate clinical trials from this model, including a variety of hypothesized drug effects. This case study focuses on the prevention of graft-versus-host disease (GvHD) after allogeneic hematopoietic stem cell transplantation (HSCT). The patient trajectory after HSCT is characterized by a complex interplay of various events of interest, and there is no established best method of measuring and/or analyzing treatment benefits. We characterized patient trajectories by means of multistate models that we fitted to a subset of the Center for International Blood and Marrow Transplant Research (CIBMTR) database. Events of interest included acute GvHD of grade III or IV, severe chronic GvHD, relapse of the underlying disease, and death. The transition probability matrix was estimated using the Aalen-Johansen estimator, and patient characteristics were identified that were associated with different transition rates. In a second step, clinical trial scenarios were simulated from the model assuming various drug effects on the background transition rates, and the operating characteristics of different endpoints and analysis strategies were compared in these scenarios. This helped devise a drug development strategy in GvHD prevention after allogeneic HSCT. More generally, multistate models provide a rich framework for exploring complex progressive diseases, and the availability of a corresponding simulation machinery provides great flexibility for clinical trial planning.

How can we make better graphs? An initiative to increase the graphical expertise and productivity of quantitative scientists.

Graphics are at the core of exploring and understanding data, communicating results and conclusions, and supporting decision-making. Increasing our graphical expertise can significantly strengthen our impact as professional statisticians and quantitative scientists. In this article, we present a concerted effort to improve the way we create graphics at Novartis. We provide our vision and guiding principles, before describing seven work packages in more detail. The actions, principles, and experiences laid out in this paper are applicable generally, also beyond drug development, which is our field of work. The purpose of this article is to share our experiences and help foster the use of good graphs in pharmaceutical statistics and beyond. A Graphics Principles "Cheat Sheet" is available online at https://graphicsprinciples.github.io/.

The stratified win ratio.

The win ratio was first proposed in 2012 by Pocock and his colleagues to analyze a composite endpoint while considering the clinical importance order and the relative timing of its components. It has attracted considerable attention since then, in applications as well as methodology. It is not uncommon that some clinical trials require a stratified analysis. In this article, we propose a stratified win ratio statistic in a similar way as the Mantel-Haenszel stratified odds ratio, derive a general form of its variance estimator with a plug-in of existing or potentially new variance/covariance estimators of the number of wins for the two treatment groups, and assess its statistical performance using simulation studies. Our simulations show that our proposed Mantel-Haenszel-type stratified win ratio performs similarly to the Mantel-Haenszel stratified odds ratio for the simplified situation when the win ratio reduces to the odds ratio, and our proposed stratified win ratio is preferred compared to the inverse-variance weighted win ratio and unweighted win ratio particularly when the data are sparse. We also formulate a homogeneity test following Cochran's approach that assesses whether the stratum-specific win ratios are homogeneous across strata, as this method is used frequently in meta-analyses and a better test for the win ratio homogeneity is not available yet.

A Longitudinal Item Response Theory Model to Characterize Cognition Over Time in Elderly Subjects.

For drug development in neurodegenerative diseases such as Alzheimer's disease, it is important to understand which cognitive domains carry the most information on the earliest signs of cognitive decline, and which subject characteristics are associated with a faster decline. A longitudinal Item Response Theory (IRT) model was developed for the Basel Study on the Elderly, in which the Consortium to Establish a Registry for Alzheimer's Disease - Neuropsychological Assessment Battery (with additions) and the California Verbal Learning Test were measured on 1,750 elderly subjects for up to 13.9 years. The model jointly captured the multifaceted nature of cognition and its longitudinal trajectory. The word list learning and delayed recall tasks carried the most information. Greater age at baseline, fewer years of education, and positive APOEɛ4 carrier status were associated with a faster cognitive decline. Longitudinal IRT modeling is a powerful approach for progressive diseases with multifaceted endpoints.