ctfepi
Sign in

# prologue

Port Meridian, 2031

Port Meridian is a fictional harbor city of 1.4 million people in 2031. You are the newest analyst at its Office of Evidence and Appraisal, the small unit that checks the numbers behind every health decision the city makes. The Office's founder, Professor Theo Marlow, left abruptly two years ago. He left behind a cabinet of annotated case files, and one cryptic note for every problem he ever set.

Small inconsistencies in the numbers of Corvane Therapeutics, maker of the heart failure drug Velquor, surface in book after book. Book VII reveals what Marlow found, and why he left.

The cast

Dr. Imogen Vale

Director of the Office of Evidence and Appraisal, who assigns your cases

Professor Theo Marlow

Founder of the Office, absent, present only in his notes

Nadia Ferreira

Field epidemiologist at the Meridian Institute of Public Health

Dr. Ansel Brook

Cardiologist and trialist at Saltmarsh General Hospital

Odile Kastner

Chair of the Meridian Health Technology Assessment Board

Quill

An anonymous source

Organizations
  • Office of Evidence and Appraisal
  • Meridian Institute of Public Health
  • Saltmarsh General Hospital
  • Meridian Health Technology Assessment Board
  • Corvane Therapeutics
  • Halden Biologics
  • Meridian Medical Review
Districts
  • Old Harbor
  • Saltmarsh
  • Kettle Hill
  • Northgate
  • Lantern Quay
  • the Refinery Flats
Data you will meet
  • Meridian Harbor Cohort: Dockworkers followed since 2009
  • Meridian Health Survey: Biennial household survey
  • Saltmarsh General electronic health records
  • QUAYSIDE trials: The trials of Velquor
  • Birth, cancer, and death registries of Port Meridian

The seven books

Book I: Orientationtrack A, Foundations
  • Case A1, The Annual Report (Epidemiologic Measures). Vale hands you the draft Port Meridian Health Report 2031 and asks you to recompute every rate before it reaches the mayor.
  • Case A2, Marlow's Cabinet (Biostatistics Foundations). The first drawer holds the exercises Marlow set every new analyst; each hides a trap he once found in a real report.
  • Case A3, The Model Shop (Regression Modeling). The Institute has reused the same regression templates for a decade; you check what each one actually estimates.
  • Case A4, The Meridian Health Survey (Survey Sampling and Complex Surveys). The biennial household survey drives the city budget, and nobody has checked its weights since 2025.
  • Case A5, Two Nurses, One Chart (Measurement and Agreement). Saltmarsh General is replacing its blood pressure devices and triage scale; you judge whether the new measurements agree with the old.
  • Case A6, The Cipher Drawer (Linear Algebra and Numerical Methods). Marlow encrypted his private notes with matrices. Decoding them is how you earn his trust.
Book II: The Harbor Cohorttrack B, Study Designs
  • Case B1, The Dockworkers (Cohort Studies). The Meridian Harbor Cohort has followed 9,000 dockworkers since 2009. Cardiac deaths are rising, and you rebuild the person time from scratch.
  • Case B2, The Night Shift (Case-Control Studies). Ferreira suspects night shift work. A case-control study inside the cohort is faster than waiting for more deaths.
  • Case B3, The Map in the Paper (Cross-Sectional and Ecological Studies). A newspaper map links district salt sales to stroke; you test what district data can and cannot say.
  • Case B4, After the Jab (Self-Controlled and Case-Only Designs). Seizures cluster after the city's new influenza vaccine, and each patient becomes their own control.
  • Case B5, QUAYSIDE-2 (Clinical Trials). Brook runs the Saltmarsh trial of Velquor for heart failure, and you are the unblinded statistician on its data monitoring committee.
  • Case B6, The Marker (Diagnostic Accuracy and Screening). Corvane sells a cardiac blood test alongside Velquor, and the city wants to screen every dockworker over 50.
Book III: Confoundedtrack C, Causal Inference and Bias
  • Case C1, Outside the Trial (Causal Inference). The Saltmarsh records hold 40,000 heart failure patients. You design the target trial before touching the data.
  • Case C2, Like for Like (Propensity Scores and Weighting). Velquor users are younger and richer than nonusers; you build groups that can be compared.
  • Case C3, The Salt Tax (Quasi-Experimental Methods). The city taxed salty foods in 2028, and the mayor wants to know whether stroke admissions fell because of it.
  • Case C4, Through Which Door? (Interaction and Mediation). Velquor lowers blood pressure, and Brook claims that is all it does. You split the effect.
  • Case C5, What If We Are Wrong? (Bias Analysis). Corvane's critics say the records are confounded by frailty; you quantify how wrong the estimate could be.
  • Case C6, The Empty Columns (Missing Data). A third of the ejection fraction values are missing from the records, and not at random.
Book IV: Forecaststrack D, Statistical Modeling
  • Case D1, How Long? (Survival Analysis). The Board needs ten year survival for Velquor patients from trials that lasted three.
  • Case D2, Wards and Visits (Longitudinal and Multilevel Models). Repeated measurements across 30 clinics; patients within clinics are not independent, and neither are visits.
  • Case D3, Prior Engagements (Bayesian Analysis). Vale wants the Board to borrow strength from earlier trials, and Kastner wants to know what the priors are doing.
  • Case D4, The Triage Bot (Machine Learning). Saltmarsh's emergency department bought a triage algorithm; you find out what it really learned.
  • Case D5, The Risk Score (Prediction Models). The city's cardiac risk score was built somewhere else; you validate it in Meridian.
  • Case D6, Settle It by Simulation (Simulation Studies). Brook and Kastner disagree about which estimator to trust, and you run the simulation that decides it.
Book V: The Boardtrack E, Evidence Synthesis and HTA
  • Case E1, The Search (Evidence Synthesis Methods). Corvane's submission to the Board claims a complete systematic review; you check what it left out.
  • Case E2, Pooling QUAYSIDE (Pairwise Meta-Analysis). Eight trials of Velquor, one forest plot, and a Board meeting on Friday.
  • Case E3, The League Table (Network Meta-Analysis). Velquor, Trevamab, and four older drugs have never all met in one trial.
  • Case E4, Unequal Trials (Population-Adjusted Methods (MAIC, STC, ML-NMR)). Corvane has patient level data; Halden only published summaries. The Board needs a fair comparison anyway.
  • Case E5, Worth It? (Health Economic Evaluation). Velquor costs eleven times more than the drug it replaces, and the Board asks what a QALY costs.
Book VI: The Citytrack F, Applied Epidemiology
  • Case F1, Saltmarsh Fever (Outbreak Investigation and Infectious Disease). Forty guests fall ill after a wedding at Lantern Quay, then their households follow.
  • Case F2, The Curve (Epidemic Modeling). The fever spreads beyond the wedding, and the mayor asks when it will peak.
  • Case F3, Early Warning (Surveillance). The Institute's alarms fired too late last winter; you tune them before the next one.
  • Case F4, The Prescription Trail (Pharmacoepidemiology). Dispensing records show who took Velquor, when, and for how long, and spontaneous reports show a strange signal.
  • Case F5, The Refinery Flats (Environmental and Occupational Epidemiology). Residents near the refinery report more asthma; air monitors and worker records hold the answer.
  • Case F6, What Meridian Eats (Nutritional Epidemiology). The diet survey behind the salt tax used food frequency questionnaires, and everyone underreports.
  • Case F7, The Harbor Genome (Genetic Epidemiology and Mendelian Randomization). A biobank of dockworkers lets you ask whether blood pressure itself causes the cardiac deaths.
  • Case F8, First Breaths (Perinatal and Reproductive Epidemiology). The birth registry shows a paradox among low birth weight babies whose mothers smoke.
  • Case F9, Lantern Quay (Cancer Epidemiology). Residents have drawn a circle on a map around the refinery and call it a cancer cluster.
  • Case F10, The Map Room (Spatial Epidemiology). Every crisis in Book VI ends up on a district map; you decide which patterns are real.
  • Case F11, Two Cities (Social Epidemiology and Health Equity). Northgate lives nine years longer than the Refinery Flats; you measure the gap and its causes.
  • Case F12, Counting Lives (Demography and Burden of Disease). The city's life tables and burden estimates decide where next year's budget goes.
Book VII: The Ledgertrack G, Research Integrity
  • Case G1, Quill (Forensics). An anonymous source sends you Corvane's manuscripts with a two word note: count again. Marlow's cabinet has a file on the same papers.
  • Case G2, Chain of Custody (Data Cleaning and Reproducibility). To prove what happened you rebuild the raw data trail, line by line, from the files Quill sends. The final scene of this case closes the story.