Methodology
How Alternative Health evaluates, scores, and ranks health products. Every number on this platform is derived from the process described below.
Principles
Data over opinion. Product rankings are computed from lab data by a deterministic formula. No editorial judgment, no paid placement, no subjective weighting. The same inputs always produce the same score.
Show the math. Every product page shows the raw lab data, every analyte, every detection limit. The score breakdown is visible on the rankings page. Nothing is hidden.
Independent verification only. We only assign full scores to products with independent third-party COA (Certificate of Analysis) data from accredited laboratories. Manufacturer self-reported data is disclosed and score-capped.
No COA, no top-tier ranking. Any product without a public, downloadable COA or equivalent product-level lab report is automatically docked 50 points and cannot score above 50. Those products may still appear in evidence-based watchlists, but they are excluded from the COA-verified tier.
Category integrity. Products are ranked within their category only. Still water is compared to still water. Spring water is compared to spring water. We never mix product types in a single ranking.
Data sources
Certificates of Analysis (COAs)
Third-party lab reports from ISO 17025 and NELAP accredited laboratories. Labs include SimpleLab, NSF International, Edge Analytical, Envirolab Services, and Sapienza University of Rome. We extract every analyte, value, unit, method, detection limit, and regulatory maximum from these reports.
PubMed / NIH
Peer-reviewed studies sourced directly from the National Library of Medicine. 12,000+ studies indexed across photobiomodulation, cold exposure, sauna therapy, autophagy/fasting, EMF exposure, and grounding. Every study links to its PubMed DOI.
EPA, EWG, UCMR5
Tap water data from the Environmental Protection Agency violation database, Environmental Working Group contaminant data, and UCMR5 (Unregulated Contaminant Monitoring Rule) PFAS detections covering 32,000+ US zip codes.
COA Purity Score
The COA Purity Score is a 0-100 numeric score reflecting how clean a product is based on its contaminant test results. It is computed from the following formula, applied only to contaminant panels (heavy metals, VOCs, pesticides, radiologicals, disinfection byproducts, PFAS, microplastics, emerging contaminants). Mineral content is excluded because minerals detected is a positive indicator, not a contaminant.
// Formula
clean_pct = (analytes_not_detected / total_analytes_tested) * 100
score = clean_pct
+ 3 if PFAS not detected
+ 2 if microplastics not detected
+ 2 if heavy metals all clear
+ 2 if 50+ contaminant analytes tested
+ 1 if 20-50 contaminant analytes tested
+ source integrity bonus for untreated / unfiltered products
- penalty for missing PFAS / microplastics testing
- 5 per limit exceedance
final = min(score, verification_tier_cap)
Component breakdown
Clean percentage is the core metric: what fraction of tested contaminant analytes were not detected (ND). A product tested for 80 contaminant analytes with 72 ND scores 90% on this component. Testing more analytes with ND results is rewarded, not penalized.
Category bonuses reward products that have been tested for specific high-concern categories. PFAS testing with ND results adds +3 because PFAS is difficult and expensive to test for. Microplastics ND adds +2. Heavy metals all clear adds +2. These bonuses exist because not all products are tested for these categories, and those that are (and pass) demonstrate additional rigor.
Coverage bonus rewards breadth of testing. Products tested for 50+ contaminant analytes receive +2. Products tested for 20-50 receive +1. This prevents a product tested for only 5 analytes (all ND) from scoring higher than one tested for 80 analytes (72 ND).
Missing-test penalties prevent incomplete COAs from outranking comprehensively tested products. A brand that skips PFAS or microplastics testing is penalized and may also be capped below the top tier until those panels are tested.
Source integrity bonus rewards products that are demonstrably untreated, unfiltered, and additive-free. For bottled water, that matters because processing can alter the product even when contaminant levels are low.
Exceedance penalty is -5 per analyte that exceeds its regulatory maximum contaminant level (MCL). This is the strongest negative signal. A single exceedance drops the score significantly.
Verification tiers
The verification tier determines the maximum score a product can achieve. This is a hard constraint enforced at the system level, not a guideline.
COA Verified
max 100Independent third-party COA from an accredited lab. Full panel data extracted and verified. The only tier eligible for a perfect score when all high-priority tests are present and clean.
Partially Verified
max 70Some lab data exists but panel coverage is incomplete. For example, a product with heavy metals and minerals testing but no PFAS, microplastics, or VOC data.
Self-Reported
max 60Data provided by the manufacturer, not independently verified. Disclosed on the product page. Score capped until independent verification is obtained.
Unverified
max 50No lab data available. Product is listed for informational purposes only with a prominent disclosure. Cannot appear in the top tier of any ranking.
Evaluation process
- 1Source COA reports. We obtain lab reports directly from manufacturers, independent labs, or public filings. Each report must come from an accredited laboratory.
- 2Extract and structure data. Every analyte, value, unit, detection limit, method, and regulatory limit is extracted from the PDF report into a typed data structure. This is done programmatically with manual verification.
- 3Classify analytes into panels. Each analyte is assigned to a panel: minerals, physical properties, heavy metals, VOCs, pesticides, radiologicals, disinfection byproducts, PFAS, microplastics, or emerging contaminants.
- 4Compute status per analyte. Each analyte receives a status: ND (not detected), pass (detected but within regulatory limits), or flag (exceeds regulatory limit).
- 5Apply the scoring formula. The COA Purity Score is computed from contaminant panel data using the formula above. The verification tier cap is applied. The score is final.
- 6Publish with full transparency. The product page shows every analyte, every value, every panel, the score breakdown, and a link to the original lab report PDF when available.
What we do not do
- xAccept payment from brands to influence scores or rankings
- xApply subjective editorial judgment to product evaluation
- xScore products without lab data (they are capped and disclosed)
- xMix product types in a single ranking
- xMake health claims or provide medical advice
- xHide the data behind a paywall or subscription
Common questions
Can a brand pay to improve their score?
No. Scores are computed from lab data by a deterministic formula. There is no editorial override, no paid placement, and no way to influence the output. The formula is public and the data is visible on every product page.
When can a product score 100?
A score of 100 is reserved for products that meet the strictest standard: full third-party COA coverage, zero contaminant exceedances, PFAS not detected, microplastics not detected, heavy metals all clear, and strong source integrity (for water: untreated, unfiltered, no additives). In practice this is extremely rare.
What happens when a product has multiple COA reports?
We use the most recent report for scoring. If multiple reports cover different panels (e.g., one for PFAS and one for heavy metals), all panels contribute to the score. Older reports are archived but not used in the current ranking.
How often are scores updated?
Scores are updated whenever new COA data becomes available. We actively seek updated reports from manufacturers and labs. The lastUpdated field on each product page shows when the data was last refreshed.
Why does my favorite brand have a lower score than I expected?
Scores are driven entirely by lab data coverage and results. A brand with excellent water quality but limited testing (e.g., only tested for 15 analytes instead of 80+) will score lower because the coverage bonus is smaller and the clean percentage is computed from fewer data points. More comprehensive testing generally yields higher scores.