Annotating Question Types for Consumer Health Questions

This paper presents a question classification scheme and a corresponding annotated corpus of consumer health questions. While most medical question classification methods have targeted medical professionals, the 13 question types we present are targeted toward disease questions posed by consumers. The corpus consists of 1,467 consumer-generated requests for disease information, containing a total of 2,937 questions. The goal of question type classification is to indicate the best strategy for selecting answers from publicly available health information resources, such as medical encyclopedias and medical research websites. The annotated question corpus, along with detailed annotation guidelines, are publicly available.