Olayinka I. Arimoro, MSc
Department of Community Health Sciences, Cumming School of Medicine, University of Calgary
Tolulope T. Sajobi, PhD
Departments of Community Health Sciences and Clinical Neurosciences, Cumming School of Medicine, University of Calgary
The journey behind the development of a web application was driven by a desire to bridge the gap between the development of advanced statistical methodologies and their uptake by analysts, researchers and practitioners in health sciences research.
Patient-reported outcomes measures (PROMs) are instruments to elicit patients’ assessment of their perceived symptoms, health status, and health related quality of life. PROMs play a pivotal role in research and routine clinical care by providing valuable insights into the experiences and perceptions of patients about their symptoms, the impact of disease, and the efficacy of treatment interventions. For example, PROMs have been used to examine how people are affected by various health conditions and have allowed comparison across health conditions to assist in evaluating and implementing new methods of providing care and equity of service delivery. However, their validity can be threatened by differential item functioning (DIF), a phenomenon where individuals with similar underlying health statuses interpret and respond differently to PROM items due to differences in demographics, culture, or other characteristics.
Traditional methods for DIF detection rely on group-based approaches and often require pre-specifying covariates, which might not capture or overlook individual differences. Tree-based item-response theory (IRT) models address these limitations by leveraging a data-driven approach to identify DIF and uncover subgroup-specific patterns, without requiring predefined covariates associated with DIF. Previous studies have applied a tree-based IRT model for DIF analysis, including the study used as an example to demonstrate the usage of the web application. Also, previous simulation studies have examined the Type I error and statistical power rate of tree-based IRT models and shown that they yield similar or even better statistical power than the conventional multiple IRT models for detecting DIF in multi-item instruments.
After completing extensive simulation studies and analyses on tree-based IRT using recursive partitioning models for DIF assessment, we recognized the need for a tool that would make these methods accessible to researchers and clinicians lacking advanced programming expertise. This realization inspired the creation of a user-friendly R Shiny web application designed to facilitate the implementation of tree-based IRT models for DIF detection in PROMs.
The web application (https://ucalgary-pcma-lab.shinyapps.io/tree_based_dif_analysis/) we developed provides an interactive platform for applying tree-based IRT models. Users can upload data in common formats (.CSV and .XLSX), customize model parameters, and visualize results. The app supports both dichotomous and polytomous items and includes a walkthrough video to guide users (https://youtu.be/vJ0urasiqk0?si=Pk4ZaNLuP97r16NU). The application leverages tree-based IRT methods to generate decision trees, which provide transparent and highly interpretable results for identifying subgroups exhibiting differential response patterns. By integrating advanced statistical methods into a user-friendly tool, this application allows researchers to conduct DIF analysis and ensures the validity of PROMs across diverse populations, ultimately enhancing the equity and precision of healthcare delivery.
This work aligns closely with ISOQOL’s commitment to promoting the use of scientifically rigorous yet practical tools in improving health related quality of life research. In addition, it represents a step forward in making advanced statistical methods accessible and impactful. By translating complex methodologies into practical tools, we can foster a more equitable approach to healthcare research and practice.
Figure 1: Tree-based IRT Shiny web application overview page
This newsletter editorial represents the views of the author and does not necessarily reflect the views of ISOQOL.
How to Submit a Newsletter Editorial
Do you have something to share about health related quality of life and patient-centered outcomes? We want to hear from you!
Learn More
The International Society for Quality of Life Research (ISOQOL) is a global community of researchers, clinicians, health care professionals, industry professionals, consultants, and patient research partners advancing health related quality of life research (HRQL).
Together, we are creating a future in which patient perspective is integral to health research, care and policy.