This paper explores simple techniques to improve the performance of pattern-based information extraction from biomedical literature. The authors propose several filtering methods that focus on the complexity of patterns and the complexity of the texts they are applied to. These techniques are evaluated using the BioNLP 2009 shared task, which focuses on event extraction in biomedical texts. The results show that these simple filtering techniques can significantly enhance the F-score of the extraction method, with improvements ranging from 6.7% to 100%. The filters also reduce the runtime by decreasing the number of matches that need to be analyzed. The proposed methods are applicable to other pattern-based extraction methods and offer a way to tune the precision-recall trade-off.This paper explores simple techniques to improve the performance of pattern-based information extraction from biomedical literature. The authors propose several filtering methods that focus on the complexity of patterns and the complexity of the texts they are applied to. These techniques are evaluated using the BioNLP 2009 shared task, which focuses on event extraction in biomedical texts. The results show that these simple filtering techniques can significantly enhance the F-score of the extraction method, with improvements ranging from 6.7% to 100%. The filters also reduce the runtime by decreasing the number of matches that need to be analyzed. The proposed methods are applicable to other pattern-based extraction methods and offer a way to tune the precision-recall trade-off.