Abstract

Abstract

Abstract
The quality of consumer-oriented health information on the Web is usually assessed through the medical certification of websites. These tools are built upon quality indicators but, so far, no standard set of indicators has been defined. The objective of the present study is to explore the popularity of specific document features and their influence on the quality of health web documents, using HON code as ground truth. A set of top-ranked health documents retrieved from a major search engine was characterized in a univariate analysis, and then used in a bivariate analysis to seek features that affect documents' quality. The univariate analysis provides insights into the characteristics of the overall population of the health web documents. The bivariate analysis reveals strong relations between documents' quality and a set of features (namely split content, videos, images, advertisements, English language) that are potential quality indicators. We characterized health web documents and identified specific document features that can be used to assess whether the information in such documents is trustworthy. The main contribution of this work is to provide other features as candidate indicators of quality. Non-health professionals can use these indicators in automatic and manual assessments of health content. © 2018 AISTI.

Abstract
Introduction. The concept and study of relevance has been a central subject in information science. Although research in information retrieval has been focused on topical relevance, other kinds of relevance are also important and justify further study. Motivational relevance is typically inferred by criteria such as user satisfaction and success. Method. Using an existing dataset composed by an annotated set of health Web documents assessed for relevance and comprehension by a group of users, we build a multivariate prediction model for the motivational relevance of search sessions. Analysis. The analysis was based on lasso variable selection, followed by model selection using multiple logistic regression. Results. We have built two regression models; the full model, which considers all variables of the dataset, has a lower estimated prediction error than the reduced model, which contains the statistically-significant variables from the full model. The higher values of evaluation metrics, including accuracy, specificity and sensitivity in the full model support this finding. The full model has an accuracy of 91.94%, and is better at predicting motivational relevance. Conclusions. Our findings suggest features that can be considered by search engines to estimate motivational relevance, to be used in addition to topical relevance. Among these features, a high level of success in Web search and in health information search on social networks and chats are some of the most influencing user features. This shows that users with higher computer literacy might feel more satisfied and successful after completing the search tasks. In terms of task features, the results suggest that users with clearer goals feel more successful. Moreover, results show that users would benefit from the help of the system in clarifying the retrieved documents.

Abstract
Health information is highly sought on the Web by users that naturally have different levels of expertise in the topics they search for. Assisting users with query formulation is important when users are searching for topics about which they have little knowledge or familiarity. To assist users with health query formulation, we developed a query suggestion system that provides alternative queries combining Portuguese and English language with lay and medico-scientific terminology. Here, we analyze how this system affects the precision of search sessions. Results show that a system providing these suggestions tends to perform better than a system without them. On specific groups of users, clicking on suggestions has positive effects on precision while using them as sources of new terms has the opposite effect. This suggests that a personalized suggestion system might have a good impact on precision. © Springer Nature Switzerland AG 2018.

Abstract
Purpose: The quality of consumer-oriented health information on the web has been defined and evaluated in several studies. Usually it is based on evaluation criteria identified by the researchers and, so far, there is no agreed standard for the quality indicators to use. Based on such indicators, tools have been developed to evaluate the quality of web information. The HONcode is one of such tools. The purpose of this paper is to investigate the influence of web document features on their quality, using HONcode as ground truth, with the aim of finding whether it is possible to predict the quality of a document using its characteristics. Design/methodology/approach: The present work uses a set of health documents and analyzes how their characteristics (e.g. web domain, last update, type, mention of places of treatment and prevention strategies) are associated with their quality. Based on these features, statistical models are built which predict whether health-related web documents have certification-level quality. Multivariate analysis is performed, using classification to estimate the probability of a document having quality given its characteristics. This approach tells us which predictors are important. Three types of full and reduced logistic regression models are built and evaluated. The first one includes every feature, without any exclusion, the second one disregards the Utilization Review Accreditation Commission variable, due to it being a quality indicator, and the third one excludes the variables related to the HONcode principles, which might also be indicators of quality. The reduced models were built with the aim to see whether they reach similar results with a smaller number of features. Findings: The prediction models have high accuracy, even without including the characteristics of Health on the Net code principles in the models. The most informative prediction model considers characteristics that can be assessed automatically (e.g. split content, type, process of revision and place of treatment). It has an accuracy of 89 percent. Originality/value: This paper proposes models that automatically predict whether a document has quality or not. Some of the used features (e.g. prevention, prognosis or treatment) have not yet been explicitly considered in this context. The findings of the present study may be used by search engines to promote high-quality documents. This will improve health information retrieval and may contribute to reduce the problems caused by inaccurate information. © 2018, Emerald Publishing Limited.