Standardizing Patient-reported Outcome in Oncology: Improving Cancer Care Through Patient Experience

Lacking clear PRO objectives results in uncertainty around the optimal approach for measuring PROs in cancer clinical trial adding confusion in their analysis and reporting. This can be observed in the current practice for utilizing patient experience data, typically characterized by inconsistent terminology, inflation of the number of analyses and potentially conflicting results hindering decision-making and cross-validation across studies. Using estimands that reflect relevant and concise PRO objectives can improve design, analysis, interpretation and communication of PRO results. We will highlight and exemplify recommendations in terms of analyses of PROs used in randomized trials based on consensus from various stakeholders. A particular topics is unobserved data which is a persistent problem in PRO analyses due to voluntary patient participation. It can stem from two distinct processes: intercurrent events and missing data. Intercurrent events inherently affect endpoint interpretation and therefore strategies to account for them must be aligned with the intended PRO objective. In particular, how to address terminal events such as death (common in oncology) must be decided a priori in function of the objective. On the other hand, missing data, often a non-random occurrence, requires strategies that consider potential informative relationships and assess overall robustness with respect to the missing data assumptions. of the results to various assumption about the missing data. Reporting the extent of data availability at each assessment timepoint by two standardized different metrics, completion rates and available data rates is recommended. As these rates will differ over time due to patient selection. Attrition and missing data will reduce observed data over time leading to selection bias. This bias impacts common methods where time trends are inferred from repeated cross-sectional analyses presented as a longitudinal series. 

Currently, the terminology around thresholds used to interpret PRO data is heterogeneous and often unclear. This poses challenges when selecting thresholds for a specific purpose and may lead to their inappropriate or misleading application and interpretation. SISAQOL-IMI has, therefore, established a harmonised terminology that differentiates different types of PRO score interpretation thresholds, in particular focusing on thresholds for individual- and group-level data.

The analysis of PRO scores requires various different statistical methods depending on the research question. Some methods compare mean scores between two or more treatment groups for example, or compare how mean scores change from baseline over time within a group. Other methods are more focussed on individual patient changes, such as the proportion of individuals who achieve an improvement (or deterioration) in their PRO score over time or the time to improvement (or deterioration). Commonly, the same interpretation threshold has been applied regardless of whether groups or individual scores are of interest. Our work demonstrates this can lead to invalid interpretation of results and conclusions and the SISAQOL-IMI recommendations ensure adequate consideration of the choice of threshold and improved reporting of how thresholds have been applied. This will allow adequate assessment of the robustness of conclusions from the PRO analyses.

SISAQOL-IMI recommendations provide key criteria for the selection of appropriate thresholds based on the analysis of interest and also for assessment of robustness of how a published threshold has been derived. The recommendations emphasize a preference for thresholds derived from anchor-based methods over distribution-based methods in order to utilize an assessment of meaning to the interpretation over and above the measurement properties of the PRO. The need for patient-centred anchors is highlighted and consideration of the population that thresholds are derived in compared to the study population where interpretation is required is also emphasized.
This session will cover these key recommendations from SISAQOL-IMI on the interpretation of PRO scores. The established key criteria aim to support the application of valid, high-quality thresholds as a cornerstone of PRO data analysis and interpretation.
 

Single arm trials (SAT) play an increasingly important role in cancer research. In certain situations, it serves as alternatives to randomized control trials as well. However, the absence of a randomized control group can limit the interpretation and conclusions, particularly when assessing the effect of a specific treatment on patient reported outcomes (PRO), such as quality of life (QOL) data. Currently, PRO objectives in SATs are often unclear or not mentioned at all. Moreover, different approaches to handle intercurrent events may yield different results and conclusions, even in a descriptive setting Specifically, addressing death should be carefully considered in advance, because patient reported outcomes after death are not defined. To address this, single-arm trials require pre-specified PRO objectives that can be translated into key clinical questions using the the ICH-E9 (R1) estimand framework including pre-specified strategies to handle intercurrent events. The chosen strategy should be defined prior to analysis in line with the pre-defined PRO objective. For example, when describing PROs over time, the while-alive strategy is generally preferred. The population-level summary for this approach includes the PRO score of participants alive and descriptive statistics about death such as the proportion of patients still alive at the time point of assessment.

Making statements on treatments in single arm studies is challenging as changes over time in PROs cannot be solely attributed to the treatment. Various factors such as natural changes over time (e.g., due to disease worsening), response shift and the effects of concomitant therapies and comorbidities may also contribute to observed changes. Use of appropriate external data may address some of the concerns, but it also poses additional challenges including defining relevant estimands, accounting for confounding and different study drop out. Moreover, a key challenge for the QOL data analysis is handling of missing data which potentially introduce bias in the results depending on how cause of missingness related patient's medical condition. We'll discuss some of the recent works done by SISAQOL-IMI project to address these challenges.

The work presented here is part of the European IMI-SISAQOL project. SISAQOL-IMI is an international project, led by the European Organization for Research and Treatment of Cancer and Boehringer Ingelheim. The aim of this four-year project is to establish international standards in the analysis of patient reported outcomes (PRO) and health-related quality of life data in cancer clinical trials.