Investigating the Potential of Extracting Answers to Health-Related Quality of Life Questionnaires From Patient Community Data
Author(s)
David M. Schmidt, MSc1, Brian Po-Han Chen, MSc2, Deborah Kuk, MSc2, Yogesh Vohra, MS, PharmD, PhD2, Valmeek Kudesia, MD2, Philipp Cimiano, PhD1.
1Center for Cognitive Interaction Technology, Bielefeld University, Bielefeld, Germany, 2Inspire, Arlington, VA, USA.
1Center for Cognitive Interaction Technology, Bielefeld University, Bielefeld, Germany, 2Inspire, Arlington, VA, USA.
OBJECTIVES: Health-related quality of life (HRQoL) questionnaires capture the subjective assessment of the impact of disease and treatment on the physical, psychological and social well-being and functioning of a patient. They help identify treatment preferences and quantify the value of a therapy from the patient perspective. As the collection of HRQoL data represents a significant effort, we investigate the extent to which answers to HRQoL questionnaires can be extracted from posts of patients in an online patient community.
METHODS: Breast cancer patients from the Inspire.com community were recruited and administered the EORTC-QLQ-C30 questionnaire with the breast cancer extension (EORTC-QLQ-BR23). Participants provided consent to analyze their online posts in relation to their questionnaire responses. The posts were manually coded in regards to whether the posts contained answers to the questionnaires. The study obtained ethical approval from the University Ethics Committee.
RESULTS: The data from 134 participating patients (756 posts, 19478 comments) was manually coded by up to three independent annotators, yielding a substantial agreement (mean Fleiss’ Kappa of 0.52). Overall, we found answers in the coded data for 50 out of 53 EORTC-QLQ-C30 questions. The information coded in the posts reliably predicted the answers given in the questionnaires (F1 = 0.74), while covering every tenth answer on average. The 5 questions that can be answered best on the basis of the coded posts were: “Did you feel ill or unwell?” (304 of 2683 annotated posts and comments), “Did you worry?” (105 posts), “Have you had pain?” (104 posts), “Did you feel tense?” (85 posts), and “Were you limited in doing either your work or other daily activities?” (77 posts).
CONCLUSIONS: Our feasibility study shows that there is valuable HRQoL-related information in posts of online patient communities. Future research should consider how these insights can be used to complement existing HRQoL instruments.
METHODS: Breast cancer patients from the Inspire.com community were recruited and administered the EORTC-QLQ-C30 questionnaire with the breast cancer extension (EORTC-QLQ-BR23). Participants provided consent to analyze their online posts in relation to their questionnaire responses. The posts were manually coded in regards to whether the posts contained answers to the questionnaires. The study obtained ethical approval from the University Ethics Committee.
RESULTS: The data from 134 participating patients (756 posts, 19478 comments) was manually coded by up to three independent annotators, yielding a substantial agreement (mean Fleiss’ Kappa of 0.52). Overall, we found answers in the coded data for 50 out of 53 EORTC-QLQ-C30 questions. The information coded in the posts reliably predicted the answers given in the questionnaires (F1 = 0.74), while covering every tenth answer on average. The 5 questions that can be answered best on the basis of the coded posts were: “Did you feel ill or unwell?” (304 of 2683 annotated posts and comments), “Did you worry?” (105 posts), “Have you had pain?” (104 posts), “Did you feel tense?” (85 posts), and “Were you limited in doing either your work or other daily activities?” (77 posts).
CONCLUSIONS: Our feasibility study shows that there is valuable HRQoL-related information in posts of online patient communities. Future research should consider how these insights can be used to complement existing HRQoL instruments.
Conference/Value in Health Info
2025-05, ISPOR 2025, Montréal, Quebec, CA
Value in Health, Volume 28, Issue S1
Code
PCR57
Topic
Patient-Centered Research
Topic Subcategory
Patient-reported Outcomes & Quality of Life Outcomes
Disease
SDC: Oncology