This paper describes the submissions of the Natural Language Processing (NLP) team from the Australian Research Council Industrial Transformation Training Centre (ITTC) for Cognitive Computing in Medical Technologies to the TREC 2021 Clinical Trials Track. The task focuses on the problem of matching eligible clinical trials to topics constituting a summary of a patient's admission notes. We explore different ways of representing trials and topics using NLP techniques, and then use a common retrieval model to generate the ranked list of relevant trials for each topic. The results from all our submitted runs are well above the median scores for all topics, but there is still plenty of scope for improvement.