The 2010 i2b2/VA Workshop on Natural Language Processing Challenges for Clinical Records featured three tasks: concept extraction, assertion classification, and relation classification. Annotated reference standards were provided for these tasks, and 22 systems were developed for concept extraction, 21 for assertion classification, and 16 for relation classification. The challenge aimed to support the development of natural language processing (NLP) systems in the clinical domain, which is often expensive and challenging due to patient privacy concerns. The data included discharge summaries and progress reports from three institutions, totaling 394 training reports, 477 test reports, and 877 unannotated reports. Systems used various methods, including conditional random fields (CRFs) and support vector machines (SVMs), with some leveraging rule-based systems and external knowledge sources. The most effective system for concept extraction achieved an exact F-measure of 0.852, while the best performance in relation extraction was 0.757. The challenge demonstrated the potential for combining complementary approaches to improve performance and highlighted the need for extensive feature engineering. The results showed that assertion classification was the easiest task, concept extraction was relatively complex, and relation classification was the most challenging. The challenge also facilitated collaboration between clinical and open-domain NLP communities and contributed to the advancement of NLP in healthcare.The 2010 i2b2/VA Workshop on Natural Language Processing Challenges for Clinical Records featured three tasks: concept extraction, assertion classification, and relation classification. Annotated reference standards were provided for these tasks, and 22 systems were developed for concept extraction, 21 for assertion classification, and 16 for relation classification. The challenge aimed to support the development of natural language processing (NLP) systems in the clinical domain, which is often expensive and challenging due to patient privacy concerns. The data included discharge summaries and progress reports from three institutions, totaling 394 training reports, 477 test reports, and 877 unannotated reports. Systems used various methods, including conditional random fields (CRFs) and support vector machines (SVMs), with some leveraging rule-based systems and external knowledge sources. The most effective system for concept extraction achieved an exact F-measure of 0.852, while the best performance in relation extraction was 0.757. The challenge demonstrated the potential for combining complementary approaches to improve performance and highlighted the need for extensive feature engineering. The results showed that assertion classification was the easiest task, concept extraction was relatively complex, and relation classification was the most challenging. The challenge also facilitated collaboration between clinical and open-domain NLP communities and contributed to the advancement of NLP in healthcare.