# Health Product Evidence Check > Automatically evaluate a health-related product, medicine-like item, supplement, cream/gel, brace, tape, massager, electrotherapy device, wearable, exercise/rehabilitation aid, or wellness product when the user shares a photo, label, link, or asks "işe yarar mı", "almaya değer mi", "10 üzerinden kaç", "zararlı mı", "reçeteli mi", "SGK karşılıyor mu", or "fiyatı iyi mi". Verify exact product/variant, ingredients or technical parameters, regulatory category, claims, finished-product evidence versus ingredient evidence, safety, contraindications, independent testing, advertising/affiliate incentives, current price, prescription and reimbursement status. Use medical evidence, market pricing, and evidence-integrity skills as needed. Do not diagnose, prescribe, or infer an unreadable label. - Skill: `yigityildiz0/health-product-evidence-check` (Agent Skill, multi-file: 4 files) - Install (CLI): `npx skillmds@latest add yigityildiz0/health-product-evidence-check` - Raw SKILL.md: https://api.skillmd.com/api/skills/yigityildiz0/health-product-evidence-check/raw - Safety review: pending - Works with: Claude Code, Claude.ai, OpenAI Codex - Category: Coding & Dev Tools - Author: yigityildiz0 (https://skillmd.com/u/yigityildiz0) - Updated: 2026-09-17 - Page: https://skillmd.com/skills/yigityildiz0/health-product-evidence-check --- # Health Product Evidence Check Answer the practical question: what the exact product is, what credible evidence says it can and cannot do, whether it fits the user's intended use, and whether the price is justified. Read [references/product-evidence-framework.md](references/product-evidence-framework.md). Use `$research-medical-evidence` for clinical effectiveness/safety, `$market-pricing-analysis` for comparable current prices, `$physio-clinical-copilot` for fit within a rehabilitation case, and `$evidence-integrity-guard` for the final claim/citation audit. ## Workflow 1. **Resolve exact identity.** Record brand/manufacturer, product name, model/variant, concentration/strength, form, size/count, lot/expiry when relevant, market/jurisdiction, and intended use. From photos, state what is legible and ask for another image if a decisive field is unreadable. 2. **Classify the product.** Determine whether it is a medicine, medical device, supplement, cosmetic, food, exercise equipment, wellness product, or ambiguous. Verify current regulator/manufacturer records; marketplace category and visual appearance are not enough. 3. **Decompose claims.** Separate objective specifications, manufacturer claims, mechanism claims, clinical outcome claims, safety claims, and user-fit questions. Do not let “clinically tested” stand without an exact study and tested version. 4. **Match evidence to the exact item.** Distinguish: - evidence on the finished product and exact dose/settings; - evidence on one ingredient/component; - evidence on the broader product class; - theory or marketing with no patient-important outcome evidence. 5. **Assess effect and safety.** Match population, indication, comparator, outcomes, duration, dose, adverse effects, contraindications, interactions, device warnings, and meaningful effect size. Distinguish statistical from clinical importance. 6. **Audit incentives.** Identify seller/manufacturer pages, sponsored or affiliate reviews, coupon content, undisclosed commercial ties, and copied claims. Use them for declared specs/claims only; seek independent evidence for effectiveness and quality. 7. **Verify practical status.** For “reçeteli mi?”, reimbursement, recall, registration, allowed claims, warranty, or current price, check the current official Turkish authority/payer/manufacturer record and dated retailers as appropriate. Do not extrapolate from another country or an old rule. 8. **Compare value.** Normalize price per dose/use/unit and compare like-for-like alternatives. Include required consumables, warranty, durability, service, return conditions, and whether a simpler evidence-based option achieves the same goal. 9. **Conclude for the stated use.** State who may benefit, who should avoid or obtain professional review, what outcome is realistic, what is unknown, and the least costly sensible next step. ## Scoring without false precision If the user requests a score, give three labeled rounded scores rather than one decorative number: - **Kanıt gücü /10:** directness, quality, consistency, and clinical relevance of evidence. - **Beklenen fayda /10:** likely magnitude for the stated goal and population, not universal effectiveness. - **Fiyat/değer /10:** current total cost versus credible benefit and alternatives. Then give one overall recommendation—`ALMAYA DEĞER`, `KOŞULLU`, `GEREKSİZ/DAHA İYİ ALTERNATİF VAR`, or `YETERSİZ BİLGİ`—with confidence. A score must not conceal uncertainty or safety limits. ## Hard rules - Never infer product identity, ingredients, concentration, expiry, approval, or authenticity from an unclear image. - Never treat registration or legal sale as proof of effectiveness, or lack of reimbursement as proof of ineffectiveness. - Never transfer evidence from an ingredient to the finished product without dose/formulation/applicability limits. - Never recommend starting, stopping, or changing a medicine. Route medication interactions and individualized use to the prescriber/pharmacist. - For red flags, severe reactions, poisoning, electric/thermal injury, worsening neurological/cardiorespiratory symptoms, or another emergency possibility, stop routine product analysis and give appropriate urgent action. - Do not use marketplace reviews as proof of medical benefit. They may reveal recurring usability or failure patterns only when identity and independence are considered. ## Output Lead with the verdict and exact product identity. Then show evidence for the exact product versus its ingredients/class, realistic benefit, key risks/fit, commercial-incentive limitations, current prescription/reimbursement/price status when requested, alternatives, scores if requested, confidence, and what missing fact could change the verdict.