Patient Pre-Screening

Pre-screening operates on a limited, de-identified dataset—typically age, sex, primary diagnosis code, and a handful of key lab values—rather than the complete medical record. This privacy-by-design approach ensures HIPAA compliance while rapidly filtering out obviously ineligible candidates. The process uses only the minimum necessary information to answer the question: 'Is this patient potentially eligible?'

Unlike full matching algorithms that use semantic similarity, pre-screening relies on strict, binary logic against hard criteria:

• Age range: Is the patient within the protocol-defined window?
• Diagnosis presence: Does the patient have the required ICD-10-CM code?
• Key exclusion triggers: Does the patient have a disqualifying comorbidity? This deterministic approach ensures zero false positives from probabilistic inference at this stage.

Pre-screening engines are designed to operate as synchronous, low-latency services embedded within clinical workflows. When a physician opens a patient's chart, a Trial Pre-Screening API call fires automatically, returning a list of potentially matching trials in under 500 milliseconds. This real-time capability ensures recruitment opportunities are surfaced at the point of care without disrupting the clinical encounter.

Beyond individual patient matching, pre-screening aggregates anonymized results across a site's entire patient population to generate feasibility assessments. By running pre-screening logic against a retrospective cohort, sponsors can estimate the number of potentially eligible subjects at a specific research site before activation. This data-driven approach replaces manual chart review guesswork with quantitative recruitment forecasting.

Before pre-screening can execute, complex protocol criteria must be atomically decomposed into independently evaluable components. A criterion like 'Histologically confirmed non-small cell lung cancer, Stage IIIB-IV, with disease progression after platinum-based chemotherapy' is broken into:

• Diagnosis: NSCLC (SNOMED CT: 254637007)
• Staging: IIIB or IV (AJCC criteria)
• Prior therapy: Platinum-based chemotherapy
• Temporal sequence: Progression after therapy Each atomic unit maps to a discrete, queryable data element.

Pre-screening serves as a cost-containment mechanism by preventing unnecessary full record reviews. By filtering out clearly ineligible patients early, it reduces the burden on clinical research coordinators who would otherwise manually review complete charts. Analysis of screen failure patterns from pre-screening logs also provides sponsors with data to refine overly restrictive protocol criteria before full trial launch.

A machine-processable definition of a clinical condition, expressed as a set of logical expressions and data queries, used to identify patient cohorts from electronic health records. A computable phenotype translates a clinical concept like "Type 2 Diabetes" into executable rules combining ICD-10-CM codes, LOINC lab values, and RxNorm medication orders. These definitions serve as the building blocks for both inclusion criteria and the pre-screening filter logic.

A technique that transforms a patient's clinical profile into a dense numerical vector to enable semantic similarity comparisons with clinical trial requirements. By encoding demographics, diagnoses, and procedures into a shared latent space, pre-screening systems can rapidly compute the cosine similarity between a patient vector and a trial vector. This approach excels at identifying candidates who may not match exact keyword queries but are semantically close to the target profile.

The process of converting parsed, structured eligibility criteria into executable database queries, such as SQL or FHIR API calls, to screen patient repositories. This step bridges the gap between a structured logical rule and the actual data infrastructure. For pre-screening, this translation generates lightweight queries that run against minimal demographic and diagnostic datasets, avoiding the computational cost of full record retrieval until a patient passes the initial filter.

The systematic review of reasons why pre-screened patients failed to meet trial eligibility, used to optimize recruitment strategies and refine protocol inclusion criteria. By aggregating and categorizing failure reasons—such as concomitant medication exclusions or out-of-range lab values—sponsors can identify overly restrictive criteria that may be unnecessarily excluding viable participants and adjust protocols before full enrollment begins.

A programmatic interface that allows external systems, such as EHRs or patient portals, to submit a de-identified patient profile and receive a list of potentially matching clinical trials. These APIs enforce privacy-preserving constraints by accepting only minimal demographic and diagnostic data, returning match probabilities without exposing the underlying trial logic. They serve as the technical integration point between healthcare delivery systems and clinical research recruitment platforms.