Microtask crowdsourcing for disease mention annotation in PubMed abstracts

TL;DR

This study demonstrates that microtask crowdsourcing via Amazon Mechanical Turk can efficiently produce high-quality disease mention annotations in biomedical literature, matching expert standards and enabling scalable corpus creation.

Contribution

The paper introduces a refined crowdsourcing protocol that achieves high annotation accuracy for disease mentions in PubMed abstracts, validated against a gold standard corpus.

Findings

Achieved an F measure of 0.872 against the gold standard.

Annotations can be tuned for higher precision or recall.

Cost-effective annotation at $0.06 per abstract per worker.

Abstract

Identifying concepts and relationships in biomedical text enables knowledge to be applied in computational analyses. Many biological natural language process (BioNLP) projects attempt to address this challenge, but the state of the art in BioNLP still leaves much room for improvement. Progress in BioNLP research depends on large, annotated corpora for evaluating information extraction systems and training machine learning models. Traditionally, such corpora are created by small numbers of expert annotators often working over extended periods of time. Recent studies have shown that workers on microtask crowdsourcing platforms such as Amazon's Mechanical Turk (AMT) can, in aggregate, generate high-quality annotations of biomedical text. Here, we investigated the use of the AMT in capturing disease mentions in PubMed abstracts. We used the NCBI Disease corpus as a gold standard for refining and benchmarking our crowdsourcing protocol. After several iterations, we arrived at a protocol that reproduced the annotations of the 593 documents in the training set of this gold standard with an overall F measure of 0.872 (precision 0.862, recall 0.883). The output can also be tuned to optimize for precision (max = 0.984 when recall = 0.269) or recall (max = 0.980 when precision = 0.436). Each document was examined by 15 workers, and their annotations were merged based on a simple voting method. In total 145 workers combined to complete all 593 documents in the span of 1 week at a cost of $.06 per abstract per worker. The quality of the annotations, as judged with the F measure, increases with the number of workers assigned to each task such that the system can be tuned to balance cost against quality. These results demonstrate that microtask crowdsourcing can be a valuable tool for generating well-annotated corpora in BioNLP.