Participatory Design of a Clinical Trial Eligibility Criteria Simplification Method

Abstract

Clinical trial eligibility criteria are important for selecting the right participants for clinical trials. However, they are often complex and not computable. This paper presents the participatory design of a human-computer collaboration method for criteria simplification that includes natural language processing followed by user-centered eligibility criteria simplification. A case study on the ARCADIA trial shows how criteria were simplified for structured database querying by clinical researchers and identifies rules for criteria simplification and concept normalization.

Combining human and machine intelligence for clinical trial eligibility querying

“…Human intelligence is needed to tease out and simplify the complex text, as we previously reported. 24 Another reason for manual deletions is that evaluators wanted to correct the span of the entity to which the concept is mapped. For example, for the phrase “mood stabilizer” in trial NCT04249869, only “mood” was automatically recognized as an entity and mapped to a wrong concept, and one of the evaluators deleted the machine output and added the concept “Mood Stabilizer” to the query.…”

“…The model developed by Tseo et al 14 incorrectly mapped “left main stem stenosis” to the overly general concept “stenosis.” Furthermore, our prior study observed the real-world practice for simplifying eligibility criteria text when translating criteria into cohort queries among clinical researchers. 24 For instance, explanatory text or measurement concepts without a value threshold cannot be queried and hence are excluded. 24 Such simplification cannot be automated and still requires human intervention given the state of NLP.…”

“… 24 For instance, explanatory text or measurement concepts without a value threshold cannot be queried and hence are excluded. 24 Such simplification cannot be automated and still requires human intervention given the state of NLP.…”