AE attribution matrix — intervention × severity

Build clinical trial SQL on your resume

Clinical trial dataset modeled on real phase 2 and phase 3 protocols. 12 active trials across cardiology, oncology, neurology, and endocrinology. NCT-format trial IDs, real MedDRA Preferred Terms (Nausea, Neutropenia, Pyrexia, etc.), CTCAE v5 severity grades 1–5, and ATC drug classes. Planted patterns include Trial 1’s severe arm imbalance (M6), Trial 3’s ~37% screen-fail rate (M13), three patients enrolled in two trials each (M11), and three enrollments seeded with multiple SAEs (M17). Patient_id space is fully independent of every other path.

Practice the SQL patterns clinical research analysts and trial sponsors actually run: enrollment cohort analysis, arm balance checks, MedDRA-coded adverse event signal detection, visit adherence drift, and Kaplan-Meier-style retention curves. Real-world terminology — NCT-format trial IDs, CTCAE v5 severity grades 1-5, ATC drug classes, and MedDRA Preferred Terms — on a fully isolated 7-table dataset modeled on real phase 2 / phase 3 protocols.

NCT-format trial IDs, MedDRA Preferred Terms for adverse events, CTCAE v5 severity grades, ATC drug classes. The vocabulary every clinical research analyst, biostatistician, and trial sponsor uses daily.

Authentic Trial Terminology

Arm balance checks, screen-fail rates, eligibility strata, and per-arm safety comparisons — the SQL patterns that drive every interim analysis and DSMB report.

Protocol & Cohort Analysis

Severity-by-arm cross-tabs, drug-vs-placebo AE rate deltas, and SAE cohort intersection — the analytics that flag a signal before the next protocol amendment.

Safety Signal Detection

Multi-CTE retention curves, sponsor portfolio scorecards, and Kaplan-Meier-style survival analysis. “Built SQL retention models on phase 3 trial data” lands cleanly with sponsor analytics, CRO, and biotech hiring managers.

Resume-Grade Capstone

The protocol says enroll 200.

Active trials by phase

Enrollments per site

Withdrawal reasons breakdown

Visit completion vs no-show

Adverse events by severity grade

Arm balance — flag the imbalanced trials

Adherence rate per enrollment

Phase-3 trial roster

Average visit drift by site

SAE rate per site

First adverse event per enrollment

Running enrollment by month

Screen-fail rate vs portfolio average

Site rank by enrollment velocity

Visit drift trend per enrollment

High-risk cohort — active enrollments with multiple SAEs or grade-5 AE

Site HHI — concentration of enrollments per therapeutic area

Drug vs placebo AE rate delta — paired per-arm comparison

Adherence quartiles × outcome — NTILE × status pivot

Trial portfolio scorecard — every trial, every metric

Multi-phase safety signal — phase 3 vs phase 2 by therapeutic area

Adverse-event temporal cluster detection

Patient retention curve — Kaplan-Meier-style step function

Sponsor performance benchmark — multi-metric ranks

trial_id

nct_id

trial_name

phase

condition

therapeutic_area

sponsor

status

start_date

end_date

dim_trials

protocol_id

arm_name

intervention_type

atc_class

version

dim_protocols

site_id

site_name

city

state

region

site_type

enrollment_capacity

pi_name

activated_date

dim_sites

patient_id

year_of_birth

race_ethnicity

bmi_baseline

prior_therapy_count

ecog_score

comorbidity_count

consent_date

withdrawal_date

dim_trial_patients

visit_id

enrollment_id

scheduled_date

actual_date

visit_type

completed_flag

days_off_schedule

fact_trial_visits

enrolled_date

withdrawal_reason

fact_enrollments

ae_id

event_date

meddra_term

severity

serious_flag

related_to_intervention

outcome

fact_adverse_events

AE attribution matrix — intervention × severity

The Brief

You'll practice