ZAS pilot insights: the practical challenges of queries and the promise of GenAI

In a recent project at Ziekenhuis aan de Stroom (ZAS), we compared Monsana’s GenAI assistant to a traditional query-based solution, focusing on the complete process of matching patients to study protocols. The results underscore the real-world value of GenAI when applied in a clinically relevant way: drastically reducing time, effort, and missed opportunities.

Many hospitals have access to structured query tools, yet clinical teams often avoid using them. 

Why? Because despite their potential, these tools face practical barriers that limit their effectiveness in everyday clinical practice. These are the key insights we gathered from interviews with several Belgian and Dutch hospitals:

• Not clinician-friendly: Clinicians without a technical background often find these tools hard to use, leading to significant reliance on IT specialists.

• Time-consuming setup: Translating free-text eligibility criteria into structured queries demands extensive collaboration between clinical and IT teams, due to a knowledge gap between clinical complexities and an abstract data models. In our pilot at ZAS, this could take up to 8 hours per trial.

• Limited matching accuracy: Traditional queries only cover structured data like age or diagnosis codes, leaving details and nuances from the patient’s narrative, such as prior therapy lines, or radiology findings, unassessed.

• Inefficient review process: Even after generating patient lists, clinical teams must manually review each candidate in detail, often sifting through multiple reports per patient. At ZAS, this step took up to 10 minutes per patient.

  
  

ZAS Pilot results: A 10x efficiency gain using GenAI vs query-based tools

In our collaboration with ZAS, we directly compared Monsana’s GenAI-based workflow to a commonly used query-based tool. The findings highlight how much more efficient screening with unstructured text can be:

	Monsana	Query-based tool
Time to add trial	0	8 hours
Time to review one patient-trial recommendation	2 min	10 min
Number of patients to screen manually	45	At least 45 patients*
Total manual workload	1u30 min	At least 15u30 min

*The query-based solution assesses criteria using the structured data model, requiring some criteria to be omitted and resulting in more patients needing manual review. While exact numbers were not obtained in this study, it is likely that the query-based tool required at least as many manual verifications than the Monsana solution.

So where does this time saving come from? In the Monsana platform, eligibility criteria are interpreted directly from trial protocols; no coding, query building, or manual translation needed. Protocols can be uploaded in natural language, and the assistant automatically integrates them into the patient screening process, eliminating the setup phase and reducing configuration time from 8 hours to zero.

When reviewing matches, Monsana goes beyond generating a basic list: each patient recommendation includes a clear summary of met and unmet criteria, the reasoning behind each assessment, and direct excerpts from the patient file. This gives clinicians immediate insight without searching multiple documents, cutting review time per patient from 10 minutes to 2.

Additionally, GenAI likely reduces the number of patients needing manual review by assessing more criteria and better excluding non-eligible cases. However, this was not measured in the current study and can’t be claimed with certainty. Nonetheless, the pilot results already demonstrate at least a 10x increase in time efficiency.

Summary: GenAI isn’t just smarter, it’s faster, too

The time comparison at ZAS reflects what many clinical teams experience: while traditional tools may seem promising on paper, their limitations in flexibility, speed, and usability make them less practical at scale for pre-screening.

Monsana changes that. By combining automated protocol interpretation, full-context patient matching, and explainable results, Monsana transforms a cumbersome process into a streamlined one, freeing up valuable time and helping hospitals identify more eligible patients with fewer resources.

Interested in staying updated on our progress? Reach out to the Monsana team or follow us on LinkedIn to learn more.