Sources & Citations

The evidence behind every detection engine

StackRx flags signals in your medication stack using primary-source pharmacology. This page lists the FDA, NIH, and peer-reviewed citations every engine is grounded in — so prospects, prescribers, and reviewers can independently verify the science. Not a substitute for medical advice.

Evidence-first posture

What this page is — and what it isn't

Every citation below points to authoritative primary literature maintained by the publishing body — no third-party summaries, no paywalled reprints. StackRx surfaces what those sources say; it does not interpret, prescribe, or replace the judgment of your prescriber. The printable doctor brief inside the app is the practitioner-facing counterpart — it exports your actual stack and the flagging rules it triggered into a single briefing your prescriber can review.

How to read these citations

Each section below corresponds to one of StackRx's three detection engines. The left column describes what the engine looks for and which signals it fires on. The right column lists the primary sources that engine is grounded in.

  • FDA = U.S. Food & Drug Administration label data and prescribing information
  • NIH = U.S. National Institutes of Health open databases (LiverTox, ODS, NCCIH)
  • Peer-reviewed academic publications with stable institutional URLs
  • Society guidelines (e.g. KDIGO, AGS Beers Criteria) where applicable

Reading the citation cards

Every external link resolves to the publishing body's own server (fda.gov, nih.gov, kdigo.org, peer-reviewed journals). Click through to the original document — StackRx is the routing layer between your stack and these sources.
Informational only

Engine output is a flag — never a recommendation. Discuss with your prescriber.

Traceable evidence

Every flag can be traced to a cited primary source.

Engine 1

Overstacking Detection —  services/overstacking-engine.js

Runs four always-on detectors on every stack change. Pairwise drug-drug interactions, therapeutic duplication, lab contraindications, and dosing anomalies. Findings deduplicate into a 0–100 Stack Health Score.

What it detects

  • DDI scan — drug-drug pairs that the FDA labeling or NIH LiverTox flag as clinically significant (e.g. tenofovir × NSAIDs, boosted protease inhibitors × simvastatin, St John's Wort × antiretrovirals)
  • Therapeutic duplication — two or more medications resolving to the same therapeutic class (e.g. two SSRIs, two statins, two ACE inhibitors)
  • Lab contraindications — medication plus lab marker out of safe range (tenofovir with low eGFR ≤60, metformin at reduced eGFR thresholds, statins with elevated ALT, ACE-i/ARB with hyperkalemia, valproate with transaminitis)
  • Dosing anomalies — dose above per-medication standard maximum (caution at >max; critical at ≥1.5×max) using per-med STANDARD_DOSE_MAX tables; injectable and PRN meds are skipped

Every finding carries a discuss with your prescriber disclaimer in its message field and at the API response top level.

Engine 2

Anticholinergic Burden —  services/anticholinergic-engine.js

Scores each medication in your stack on the Anticholinergic Cognitive Burden (ACB) scale and surfaces cumulative burden. Anticholinergic load correlates with cognitive impairment, delirium, and falls in older adults — independent of any single medication's dose.

What it detects

  • Per-med ACB score — 0 (none) through 3 (definite anticholinergic activity) for every medication with a documented anticholinergic effect
  • Cumulative burden — sum of ACB scores across the full stack; a single medication is rarely the problem, but two ACB-3 meds is a known cognitive-risk signal
  • Beers-Criteria overlap — flags individual medications that show up on the 2023 AGS Beers list of medications that should be avoided in older adults in most cases
  • Cross-class combinations — antihistamines plus tricyclic antidepressants plus bladder antispasmodics stack additively even though they're prescribed by different specialists

Burden findings are informational. The scale itself is the gold standard referenced in geriatric guidelines worldwide.

Engine 3

Compound Accumulation —  services/compound-accumulation-engine.js

Tags each medication and supplement by its metabolic pathway, hepatic load, and therapeutic redundancy. Surfaces when supplements are doubling up on the same active ingredient, hitting the same pathway, or competing for hepatic clearance.

What it detects

  • Pathway overlap — multiple medications or supplements hitting the same metabolic pathway (CYP3A4, CYP2D6, etc.) even when prescribed by different clinicians
  • Active-ingredient duplication — two products with the same active molecule (e.g. two magnesium supplements, two fish-oil products with overlapping EPA/DHA)
  • Hepatic load — cumulative load on liver metabolism, with NIH LiverTox-documented hepatotoxic agents flagged at elevated ALT
  • Herb-drug interactions — supplements that interact with prescription medications at the documented mechanism level (St John's Wort, grapefruit, kava, comfrey)

Pathway tagging uses NIH Office of Dietary Supplements ingredient data, with cross-references to LiverTox and NCCIH herb-drug guidance.

See what your own stack has on it

The printable doctor handoff PDF is the in-product counterpart — exports your stack + citations from your own data into a single briefing your prescriber can review.

Start Free — No Payment Required
Free for up to 4 medications. $12/mo for unlimited.