Where each tool gets its evidence

Every tool on this site draws on specific, published sources. Citations are shown as short labels (for example, Mills 2026); each label links to the full reference on the References page.

Definitions and terminology. The default framework is the Fifth Universal Definition of MI Mills 2026. Where a guideline or study predates it, the app says so and uses the Fourth Universal Definition Thygesen 2018 vocabulary; the 2025 ACC/AHA guideline Rao 2025 is one example. The background reviews from our group are Maayah 2024 and Gokhan 2024.

Serial high-sensitivity troponin (Bayes interpreter). The sensitivity, specificity, predictive values, and likelihood ratios for pairs of hs-troponin values come from Neumann 2019a, whose primary outcome was myocardial infarction.

Bedside findings (Bayes interpreter). Test characteristics for demographics, comorbidities, symptoms, vitals, exam, and ECG findings come from Fanaroff 2015. Its outcome was acute coronary syndrome.

eGFR. Estimated GFR uses the 2021 CKD-EPI creatinine equation Inker 2021.

ECG criteria checker. ST-elevation thresholds and occlusion-equivalent patterns follow the Fifth Universal Definition (Section 13 and Table 5) Mills 2026, §13. Pattern primary sources are cited on the page: Sgarbossa Sgarbossa 1996, modified Sgarbossa Smith 2012a, de Winter de Winter 2008, and a review of occlusion-MI patterns Ricci 2025.

Troponin kinetics. The schematic curve illustrates Figure 10 of the Fifth Universal Definition Mills 2026, Fig 10. It is not a pharmacokinetic fit of any assay.

Etiologies. The 24 original topics come from Maayah 2024. Mechanism tags and the additional draft topics follow Tables 2 and 3 and Sections 14 and 16 of the Fifth Universal Definition Mills 2026.

Sex-specific 99th percentile cutoffs. The optional assay presets in the UDMI classifier are manufacturer-labeled values, each cited to its FDA decision summary, the IFCC C-CB table Collinson 2019, or the manufacturer, with the date they were last checked.