SimQ: Real-Time Retrieval of Similar Consumer Health Questions

Abstract

BackgroundThere has been a significant increase in the popularity of Web-based question-and-answer (Q&A) services that provide health care information for consumers. Large amounts of Q&As have been archived in these online communities, which form a valuable knowledge base for consumers who seek answers to their health care concerns. However, due to consumers’ possible lack of professional knowledge, it is still very challenging for them to find Q&As that are closely relevant to their own health problems. Consumers often repeatedly ask similar questions that have already been answered previously by other users.ObjectiveIn this study, we aim to develop efficient informatics methods that can retrieve similar Web-based consumer health questions using syntactic and semantic analysis.MethodsWe propose the “SimQ” to achieve this objective. SimQ is an informatics framework that compares the similarity of archived health questions and retrieves answers to satisfy consumers’ information needs. Statistical syntactic parsing was used to analyze each question’s syntactic structure. Standardized Unified Medical Language System (UMLS) was employed to annotate semantic types and extract medical concepts. Finally, the similarity between sentences was calculated using both semantic and syntactic features.ResultsWe used 2000 randomly selected consumer questions to evaluate the system’s performance. The results show that SimQ reached the highest precision of 72.2%, recall of 78.0%, and F-score of 75.0% when using compositional feature representations.ConclusionsWe demonstrated that SimQ complements the existing Q&A services of Netwellness, a not-for-profit community-based consumer health information service that consists of nearly 70,000 Q&As and serves over 3 million users each year. SimQ not only reduces response delay by instantly providing closely related questions and answers, but also helps consumers to improve the understanding of their health concerns.

A Paraphrase Generation System for EHR Question Answering

“…The proposed corpus consists of questions which can be answered using EHR data. Such a corpus would have utility beyond QA systems as well, like in question similarity (Luo et al, 2015;Nakov et al, 2017), and in particular could serve as a standard paraphrase corpus for the medical domain.…”

Aspiring to Unintended Consequences of Natural Language Processing: A Review of Recent Developments in Clinical and Consumer-Generated Text Processing

“…Cohen and colleagues have shown that interactive question answering sites could efficiently address consumer health question answering through either short answers by a small number of dedicated physicians, enabling high throughput, or physician experts operating as moderators in patient forums [102]. Luo and colleagues used syntactic and semantic analysis to align a new question with the questions previously submitted to NetWellness, a website through which highly qualified volunteers provide answers to consumers' health questions [103].…”

A question-entailment approach to question answering

“…RQE is particularly relevant due to the increasing numbers of similar questions posted online [32] and its ability to solve differently the challenging issues of question understanding and answer extraction. In addition to being used to find relevant answers, these resources can also be used in training models able to recognize inference relations and similarity between questions.…”