Screen Failure Analysis: Using Enrollment Data to Stop Repeated Mistakes

Every screen failure is a $180–450 event: coordinator time for the pre-screen call, PI time for the visit review, lab processing costs, and transportation reimbursement. For a site running 20 screening visits per month at a 45% failure rate, that is 9 failed screens and $1,620–4,050 in monthly waste on patients who were never going to enroll. More importantly, 60–70% of screen failures are predictable — they fail on criteria that appear in your data repeatedly. This analysis converts that pattern into prevention.

The Screen Failure Categorization Framework

Categorize every screen failure into one of five buckets. Track counts for each bucket monthly:

• Bucket 1 — Demographic ineligibility: Age out of range, gender (if protocol-specific), geographic restriction. These should be caught by advertising targeting and AI pre-screening — if they reach a screening visit, your pre-screening has a gap.
• Bucket 2 — Diagnosis disqualification: Patient does not have the required diagnosis, or has it in a form excluded by protocol (e.g., Type 1 vs. Type 2, mild vs. moderate-to-severe). These indicate a messaging problem — your advertising is reaching the right condition but wrong specification.
• Bucket 3 — Medication conflict: Patient is on a disqualifying medication. Often catchable with a pre-screening question. High counts in this bucket mean your medication exclusion question is absent or ineffective in pre-screening.
• Bucket 4 — Lab value out of range: Patient’s lab values (A1c, eGFR, BMI, etc.) are outside protocol range. Partially catchable with patient-reported recent lab values in pre-screening. High counts indicate your advertising is reaching the right diagnosis but wrong severity range.
• Bucket 5 — Other protocol criteria: Prior study participation, comorbidities, logistical issues. These are harder to pre-screen but can often be addressed with better coordinator phone screening questions.

Running the Monthly Analysis

On the first of each month: pull all screen failures from the previous month. Categorize each into one of the five buckets. Calculate the count and percentage per bucket. Compare to the prior month. Track the trend: is any bucket growing? Growth in Bucket 2 or 3 indicates your advertising targeting or pre-screening has degraded — an actionable signal.

Converting Each Bucket Into a Specific Fix

• Bucket 1 spike: Tighten age demographic layer in Google Ads and Meta. Add age question to AI pre-screening flow.
• Bucket 2 spike: Review ad copy and landing page — are you clearly communicating the specific diagnosis type required? Add a diagnosis specification question to pre-screening.
• Bucket 3 spike: Add the disqualifying medication as an explicit question in AI pre-screening and coordinator phone script.
• Bucket 4 spike: Add a “Do you know your most recent [lab value]?” question in pre-screening with acceptable range guidance. This does not verify the value, but eliminates patients clearly outside range who know their numbers.
• Bucket 5 spike: Review your coordinator phone screening script. Add targeted questions for the specific criteria causing failures.

48-Hour Action List

1. Hour 1: Pull every screen failure from the past 60 days. Create a spreadsheet with columns: patient ID, failure date, failure reason (free text). Categorize each into the five buckets above. Count each bucket.
2. Hour 2: Identify your highest-count bucket. For that bucket, identify the specific fix from the list above. Implement the fix today — add the pre-screening question, tighten the demographic layer, or update the coordinator script.
3. Hour 3: Build a monthly screen failure tracking tab in your recruitment dashboard. Starting this month, record bucket counts on the first of each month. You need three months of data before trends become meaningful, so start now.
4. Day 2: Share the screen failure analysis with your coordinator and PI. Present it as “here is where we are losing patients we could have caught earlier, and here is the specific fix.” The conversation usually surfaces additional failure reasons your team has observed anecdotally but not tracked systematically.