Context-based interactive health information searching

Introduction. This paper deals with the impact of contextual features, such as sex, age, mother tongue, health status, health literacy, Internet use experience, and frequency of health information seeking on health information searching. Method. An interactive information retrieval approach was used to study users' searching behaviour. An online survey and experiment using simulated situation technique were used as data collection methods. The online survey gathered data about user features, such as sex, age, mother tongue, health status, health literacy, Internet use experience, and health information seeking. An experiment was then carried out using four simulated tasks to collect information about health information searching. Analysis. The multiple linear regression analysis method was used to identify contextual factors affecting query length and number of queries. In addition, binary logistic regression analysis method was used to identify contextual factors affecting result clicking. Results. Frequent health information seeking leads to more queries and long query length, and English as a mother tongue and being healthy contribute to long query length. Queries with spelling errors and those formulated outside task descriptions are found to be ineffective. Conclusion. Contextual features such as frequency of health information seeking, mother tongue, and health status influence query formulation. In addition, spelling errors and source of query affect the effectiveness of queries. The findings could be useful for health information retrieval systems to learn and predict users' information needs to aid effective retrieval.