Visual analysis of obesity‐related query terms on HealthLink

A study on health care consumers’ diabetes term usage across identified categories

Purpose

– The purpose of this paper is to investigate health care consumers’ diabetes term usage patterns based on Yahoo!Answers social question and answers (Q&A) forum, identified characteristics and relationships among terms within three pairs of related categories identified from the Q&A log, and revealed users’ diabetes term usage patterns.

Design/methodology/approach

– The Q&A analysis method allowed first-hand investigation of massive data from health care consumers. Visual term clustering analysis across categories was conducted using a multi-dimensional scaling (MDS) visualization method which provides an intuitive and interactive way to explore and discover term association patterns in a visual environment. Closely related categories were identified and corresponding visual term clustering analyses between categories (Sign & Symptom and Organ & Body Part; Diagnosis and Test; and Diagnosis and Medication) as well at the term level were analyzed.

Findings

– The findings show that there are close relationships between terms in two related categories. Related terms were grouped and patterns were revealed. All the stress values of the MDS analyses fall below 0.10 and RSQ for each of the combined categories is over 0.90 which indicate the investigated terms were well clustered in the visual analyses.

Originality/value

– The study provides a unique research methodology for similar consumer health research studies. The results of this study offer insight into consumer health term use behavior, and enrich existing thesauri and subject heading lists, enhance diabetes-related web sites or portals, and improve effectiveness of internal search engines.