Journal of Biopharmaceutical Statistics最新文献

Symptoms post-SARS-CoV-2 infection may persist for months and cause significant impairment and impact to quality of life. Acute symptoms of SARS-CoV-2 infection are well studied, yet data on clusters of symptoms over time, or post-acute sequelae of SARS-CoV-2 infection (PASC), are limited. We aim to characterize PASC phenotypes by identifying symptom clusters over a six-month period following infection in individuals vaccinated (boosted and not) and those unvaccinated. Subjects with ≥1 self-reported symptom and positive RT-PCR for SARS-CoV-2 at CVS Health US test sites were recruited between January and April 2022. Patient-reported outcomes symptoms, health-related quality of life (HRQoL), work productivity and activity impairment (WPAI) were captured at 1 month, 3 months, and 6 months post-acute infection. Phenotypes of PASC were determined based on subject matter knowledge and balanced consideration of statistical criteria (lower AIC, lower BIC, and adequate entropy) and interpretability. Generalized estimation equation approach was used to investigate relationship between QoL, WPAI and number of symptoms and identified phenotypes, and relationship between phenotypes and vaccination status as well. LCA identified three phenotypes that are primarily differentiated by number of symptoms. These three phenotypes remained consistent across time periods. Subjects with more symptoms were associated with lower HRQoL, and worse WPAI scores. Vaccinated individuals were more likely to be in the low symptom burden latent classes at all time points compared to unvaccinated individuals.

Objectives: The FDA recommends the use of anchor-based methods and empirical cumulative distribution function (eCDF) curves to establish a meaningful within-patient change (MWPC) for a clinical outcome assessment (COA). In practice, the estimates obtained from model-based methods and eCDF curves may not closely align, although an anchor is used with both. To help interpret their results, we investigated and compared these approaches.

Methods: Both repeated measures model (RMM) and eCDF approaches were used to estimate an MWPC on a target COA. We used both real-life (ClinicalTrials.gov: NCT02697773) and simulated data sets that included 688 patients with up to six visits per patient, target COA (range 0 to 10), and an anchor measure on patient global assessment of osteoarthritis from 1 (very good) to 5 (very poor). Ninety-five percent confidence intervals for the MWPC were calculated by the bootstrap method.

Results: The distribution of the COA score changes affected the degree of concordance between RMM and eCDF estimates. The COA score changes from simulated normally distributed data led to greater concordance between the two approaches than did COA score changes from the actual clinical data. The confidence intervals of MWPC estimate based on eCDF methods were much wider than that by RMM methods, and the point estimate of eCDF methods varied noticeably across visits.

Conclusions: Our data explored the differences of model-based methods over eCDF approaches, finding that the former integrates more information across a diverse range of COA and anchor scores and provides more precise estimates for the MWPC.

Regulatory agencies are advancing the use of systematic approaches to collect patient experience data, including patient-reported outcomes (PROs), in cancer clinical trials to inform regulatory decision-making. Due in part to clinician under-reporting of symptomatic adverse events, there is a growing recognition that evaluation of cancer treatment tolerability should include the patient experience, both in terms of the overall side effect impact and symptomatic adverse events. Methodologies around implementation, analysis, and interpretation of "patient" reported tolerability are under development, and current approaches are largely descriptive. There is robust guidance for use of PROs as efficacy endpoints to compare cancer treatments, but it is unclear to what extent this can be relied-upon to develop tolerability endpoints. An important consideration when developing endpoints to compare tolerability between treatments is the linkage of trial design, objectives, and statistical analysis. Despite interest in and frequent collection of PRO data in oncology trials, heterogeneity in analyses and unclear PRO objectives mean that design, objectives, and analysis may not be aligned, posing substantial challenges for the interpretation of results. The recent ICH E9 (R1) estimand framework represents an opportunity to help address these challenges. Efforts to apply the estimand framework in the context of PROs have primarily focused on efficacy outcomes. In this paper, we discuss considerations for comparing the patient-reported tolerability of different treatments in an oncology trial context.

Background: Daily diaries are an important modality for patient-reported outcome assessment. They typically comprise multiple questions, so understanding their underlying structure is key to appropriate analysis and interpretation. Structural evaluation of such measures poses challenges due to the high volume of repeated measurements. Potential strategies include selecting a single day, averaging item-level observations over time, or using all data while accounting for its multilevel structure.

Method: The above strategies were evaluated in a simulated dataset via exploratory and confirmatory factor modelling by comparing their impact on various estimates (i.e., inter-item correlations, factor loadings, model fit). Each strategy was additionally explored using real-world data from an observational study (the Asthma Nighttime Symptoms Diary).

Results: Both single day and item average strategies resulted in biased factor loadings. The former displayed lower overall bias (single day: 0.064; item average: 0.121) and mean square error (single day: 0.007; item average: 0.016) but greater frequency of incorrect factor number identification compared with the latter (single day: 46.4%; item average: 0%). Increased estimated inter-item correlations were apparent in the item-average method. Non-trivial between- and within-person variance highlighted the utility of a multilevel approach. However, convergence issues and Heywood cases were more common under the multilevel approach (90.2% and 100.0%, respectively).

Conclusions: Our findings suggest that a multilevel approach can enhance our insight when evaluating the structural properties of daily diary data; however, implementation challenges still remain. Our work offers guidance on the impact of data handling decisions in diary assessment.

The impact of chronic diseases on health-related quality of life (HRQOL) in adolescents and young adults (AYAs) is understudied. Latent profile analysis (LPA) can identify profiles of AYAs based on their HRQOL scores reflecting physical, mental, and social well-being. This paper will (1) demonstrate how to use LPA to identify profiles of AYAs based on their scores on multiple HRQOL indicators; (2) explore associations of demographic and clinical factors with LPA-identified HRQOL profiles of AYAs; and (3) provide guidance on the selection of adult or pediatric versions of Patient-Reported Outcomes Measurement Information System® (PROMIS®) in AYAs. A total of 872 AYAs with chronic conditions completed the adult and pediatric versions of PROMIS measures of anger, anxiety, depression, fatigue, pain interference, social health, and physical function. The optimal number of LPA profiles was determined by model fit statistics and clinical interpretability. Multinomial regression models examined clinical and demographic factors associated with profile membership. As a result of the LPA, AYAs were categorized into 3 profiles: Minimal, Moderate, and Severe HRQOL Impact profiles. Comparing LPA results using either the pediatric or adult PROMIS T-scores found approximately 71% of patients were placed in the same HRQOL profiles. AYAs who were female, had hypertension, mental health conditions, chronic pain, and those on medication were more likely to be placed in the Severe HRQOL Impact Profile. Our findings may facilitate clinicians to screen AYAs who may have low HRQOL due to diseases or treatments with the identified risk factors without implementing the HRQOL assessment.

Patient Reported Outcomes (PROs) are widely used in quality of life (QOL) studies, health outcomes research, and clinical trials. The importance of PRO has been advocated by health authorities. We propose this R shiny web application, PROpwr, that estimates power for two-arm clinical trials with PRO measures as endpoints using Item Response Theory (GRM: Graded Response Model) and simulations. PROpwr also supports the analysis of PRO data for convenience of estimating the effect size. There are seven function tabs in PROpwr: Frequentist Analysis, Bayesian Analysis, GRM power, T-test Power Given Sample Size, T-test Sample Size Given Power, Download, and References. PROpwr is user-friendly with point-and-click functions. PROpwr can assist researchers to analyze and calculate power and sample size for PRO endpoints in clinical trials without prior programming knowledge.

Dose-finding clinical trials in oncology estimate the maximum tolerated dose (MTD), based on toxicity obtained from the clinician's perspective. While the collection of patient-reported outcomes (PROs) has been advocated to better inform treatment tolerability, there is a lack of guidance and methods on how to use PROs for dose assignments and recommendations. The PRO continual reassessment method (PRO-CRM) has been proposed to formally incorporate PROs into dose-finding trials. In this paper, we propose two extensions of the PRO-CRM, which allow continuous enrollment of patients and longer toxicity observation windows to capture late-onset or cumulative toxicities by using a weighted likelihood to include the partial toxicity follow-up information. The TITE-PRO-CRM uses both the PRO and the clinician's information during the trial for dose assignment decisions and at the end of the trial to estimate the MTD. The TITE-CRM + PRO uses clinician's information solely to inform dose assignments during the trial and incorporates PRO at the end of the trial for the estimation of the MTD. Simulation studies show that the TITE-PRO-CRM performs similarly to the PRO-CRM in terms of dose recommendation and assignments during the trial while almost halving trial duration in case of an accrual of two patients per observation window. The TITE-CRM + PRO slightly underperforms compared to the TITE-PRO-CRM, but similar performance can be attained by requiring larger sample sizes. We also show that the performance of the proposed methods is robust to higher accrual rates, different toxicity hazards, and correlated time-to-clinician toxicity and time-to-patient toxicity data.

Performance outcome (PerfO) measures are based on tasks performed by patients in a controlled environment, making their meaningful interpretation challenging to establish. Co-calibrating PerfO and patient-reported outcome (PRO) measures of the same target concept allow for interpretation of the PerfO with the item content of the PRO. The Rasch model applied to the discretized PerfO measure together with the PRO items allows expressing parameters related to the PerfO measure in the PRO metric for it to be linked to the PRO responses. We applied this approach to two PerfO measures used in multiple sclerosis (MS) for walking and manual ability: the Timed 25-Foot Walk (T25FW) and the 9-Hole Peg Test (9HPT). To determine meaningful interpretation of these two PerfO measures, they were co-calibrated with two PRO measures of closely related concepts, the MS walking scale - 12 items (MSWS-12) and the ABILHAND, using the data of 2,043 subjects from five global clinical trials in MS. The probabilistic relationships between the PerfO measures and the PRO metrics were used to express the response pattern to the PRO items as a function of the unit of the PerfOs. This example illustrates the promises of the co-calibration approach for the interpretation of PerfO measures but also highlights the challenges associated with it, mostly related to the quality of the PRO metric in terms of coverage of the targeted concept. Co-calibration with PRO measures could also be an adequate solution for interpretation of digital sensor measures whose meaningfulness is also often questioned.

A statistical methodology named "capture recapture", a Kaplan-Meier Summary Statistic, and an urn model framework are presented to describe the elicitation, then estimate both the number of interviews and the total number of items ("codes") that will be elicited during patient interviews, and present a summary graphical statistic that "saturation" has occurred. This methodology is developed to address a gap in the FDA 2009 PRO and 2012 PFDD guidance for determining the number of interviews (sample size). This estimate of the number of interviews (sample size) uses a two-step procedure. The estimate of the total number of items is then used to estimate the number of interviews to elicit all items. A framework called an urn model is a framework for describing the elicitation and demonstrate the algorithm for declaring saturation "first interview with zero new codes". A caveat emptor is that due to independence assumptions, the urn model is not used as a method for estimating probabilities. The URN model provides a framework to demonstrate that an algorithm such as "first interview with zero new codes" may establish that all codes have been elicited. The limitations of the Urn model, capture recapture, and Kaplan-Meier are summarized. The statistical methods and the estimates supplement but do not replace expert judgement and declaration of "saturation." A graphical summary statistic is presented to summarize "saturation," after expert declaration for two algorithms. An example of a capture-recapture estimate, using simulated data is provided. The example suggests that the estimate of total number of codes may be accurate when prepared as early as the second interview. A second simulation is presented with an URN model, under a strong assumption of independence that an algorithm such as 'first interview with zero new codes" may fail to identify all codes. Potential errors in declaration of saturation are presented. Recommendations are presented for additional research and the use of the algorithm "first interview with zero new codes."