MedStatica is a biostatistics consult built for clinicians and trainees who were never taught the stats — the test to run, the numbers with confidence intervals, a methods paragraph to paste, and a read on whether the IRB needs to see it.
| Study design | Retrospective cohort |
| Recommended analysis | Chi-square test (2×2, Yates), with logistic regression if you adjust for age or comorbidity |
| Rationale | Your outcome and exposure are both yes/no, so a 2×2 comparison is the natural primary test. If the groups differ at baseline, a logistic model adjusts without changing the question. |
Not a calculator you have to already understand — a consult that meets you where a busy clinician actually is.
Type the population, what you're comparing, and the outcome you care about. No test names required.
The recommended analysis with a plain-English rationale, the assumptions to check, and a citation trail back to the seminal papers.
A read on likely IRB review — exempt, expedited, full board, or possible QI — with concrete next steps before you touch data.
Export a one-page memo — your question, the plan, the oversight notes — for a statistician or research consult.
Most stats tools stop at the p-value. MedStatica also tells a trainee what a seasoned mentor would: whether this needs the IRB, and what to do first. It flags the likely pathway from your design — retrospective vs. prospective, identifiable data, intervention, vulnerable populations — and lists the next steps.
Orientation only. Exempt, expedited, or full-board status is a determination the IRB makes — never the investigator, and never the app.
Four tiers mirroring a fellows' biostatistics course. Fill the highlighted cells; the margins, intervals, and plain-language reading compute live.
| Tier 1 · Describe | Mean (SD), median [IQR], range — with the symmetry check that tells you which to report. |
| Tier 2 · Compare two groups | Chi-square (Yates), Welch and paired t, Pearson r, treatment 2×2 with RR, OR, and NNT. |
| Tier 3 · Adjust | Logistic, linear, Cox, and Poisson — routed to a guided plan, because the hard part is specification, not arithmetic. |
| Tier 4 · Evaluate a test | Diagnostic 2×2 — sensitivity, specificity, PPV/NPV, Youden's J, and likelihood ratios, all with Wilson intervals. |
| Free | Pro · $6.99/mo | Your own key | |
|---|---|---|---|
| Consults / month | 10 credits | 300 credits | unlimited |
| Engines | Swift | all three | all three |
| Every calculator | ✓ | ✓ | ✓ |
| Oversight read & memo export | ✓ | ✓ | ✓ |
Credits per consult scale with model depth — Swift 1 · Balanced 3 · Deep 15 — mirroring relative cost.
One note when it hits the App Store. No list-sharing, no filler.