AI SMS pre-screening is the automated process of evaluating an interested patient's eligibility for a clinical trial through a structured text message conversation — before any coordinator time is spent and before any in-person visit is scheduled. The patient receives a series of questions calibrated to the protocol's inclusion and exclusion criteria. An AI model evaluates the responses in real time and returns a routing recommendation: schedule a visit, request clarification, or decline with a note explaining why the patient does not qualify.
How the SMS Conversation Is Structured
The pre-screening SMS sequence is built from the protocol's eligibility criteria, translated into plain-language questions a patient can answer without clinical knowledge. For a type 2 diabetes study, the sequence asks: current diagnosis, how long they have had the condition, current medications, approximate most recent HbA1c if known, and any major recent cardiovascular events. The questions are short, sequential, and designed to identify disqualifying answers early so clearly ineligible patients exit the sequence quickly without being asked unnecessary questions.
The entire SMS conversation typically takes 3 to 5 minutes for the patient to complete and runs through Twilio or a similar SMS API connected to the site's workflow automation. No app download is required. No portal login is required. The patient responds to a text message from a number associated with the research site.
What the AI Evaluates
The AI model — Claude, GPT-4o, or a similar large language model — receives the patient's responses alongside the protocol's eligibility criteria configured as evaluation rules. It identifies whether the patient's answers meet the primary inclusion criteria, flags any answers that contradict exclusion criteria, and produces a structured pre-screening summary with a routing recommendation. Clearly eligible patients are routed to the scheduling queue. Uncertain cases are routed to coordinator review with the specific concern highlighted. Clearly ineligible patients receive a decline message.
What This Changes for Screen Failure Rate
Screen failure rates of 20 to 80 percent at many sites reflect the proportion of patients who reach the full screening visit and fail eligibility there — after consuming a coordinator's time for scheduling, preparation, and the visit itself. AI SMS pre-screening moves the eligibility filter earlier in the pipeline, before any coordinator time is consumed and before the patient makes the trip to the site. The patients who reach the full screening visit have already passed the pre-screening filter, and their screen failure rate is substantially lower.