HOW THE REPOSITORY WORKS

Evidence standards & coverage

The ambition is a trusted, comprehensive map of natural and supportive approaches. The present release is a selective, source-linked starting collection, not a complete systematic review.

One record, one clinical question

The basic unit is an approach used for a defined goal in a defined setting. Ginger for acute chemotherapy nausea is a different question from ginger as an anticancer therapy. An ingredient should not inherit every result from every formulation or disease.

What a useful record contains

Name and aliases; intervention family; symptom or cancer-control goal.
Cancer type, treatment, stage and participant characteristics relevant to applicability.
Preparation, route, dose and duration when reliably extracted; study regimens are not personal prescriptions.
Study design, comparator, sample size, primary outcome and follow-up.
Absolute and relative results where available, uncertainty and clinically meaningful thresholds.
Negative findings, missing data, safety concerns, interactions and limits to generalization.
Original source links, record identifier, dates and provenance.

Some fields are not yet fully extracted across this first release, including funding, conflicts, all confidence intervals and registry status. An absent field means not recorded—not none.

Read the evidence labels precisely

Guideline-supported: a cited guideline supports the named use in a defined setting; strength and conditions still matter.
Positive randomized signal: a controlled study found a relevant benefit, with its limitations still visible.
Mixed / uncertain: findings, formulations or methods differ; a universal conclusion is premature.
No demonstrated benefit: the linked studies did not establish the claimed effect. Distinguish imprecision from strong evidence of little benefit.
Investigational: early clinical or mechanistic evidence requires further testing; it is not an established cancer treatment.
Harm / interaction concern: a safety finding can matter even when benefit remains uncertain.
Practical guidance: a clinical or nutrition resource, not a trial establishing efficacy.

These are descriptive editorial labels, not a validated score or formal GRADE assessment. Benefit, confidence and risk are separate judgments.

Sources and provenance

The initial collection combines 33 existing Missing Script reviews with 36 dedicated evidence briefs. Existing summaries link back to their fuller review; new briefs cite the source checked for this release. NCI and professional guidelines support orientation, while original studies support study-specific results. Onco.cc inspired the connected structure; its database has not been imported.

What is outside this first release?

An exhaustive search of every published or unpublished study.
An automated medicine–supplement interaction checker or personalized treatment recommendation.
Continuously verified recruitment status for every trial.
Formal independent clinician sign-off or systematic risk-of-bias assessment of every record.

Release notes

October 11, 2026: Evidence Explorer published; 45 records organized; 12 briefs added; symptom, cancer, source, safety and cancer-control routes introduced.

October 12, 2026: A daily evidence-search and repository-maintenance workflow was scheduled to begin October 13. It will prioritize new human studies, guidance changes, negative findings, corrections and safety signals, while gradually filling gaps in older evidence. Verified entries and revisions will be published; unresolved findings will remain drafts. Each search and substantive change will be logged. A search date is separate from an individual record’s appraisal date and does not mean every record has been rechecked. Scheduled searches are selective, not exhaustive live surveillance or independent clinical sign-off.

Methods to consult

← Return to Evidence Explorer · How to judge a study

Maintenance history

October 13, 2026: First maintenance pass completed. Four records were added, one was strengthened, 17 distinct source URLs were added, and the collection reached 49 records. A recent real-world null result for topical diclofenac was published alongside the earlier positive randomized trial. Read the dated Evidence Updates log.

Search dates remain separate from per-record appraisal dates. Selective maintenance is not exhaustive live surveillance, formal independent sign-off or a complete interaction checker.

Maintenance history

October 14, 2026: Ten new clinical-question records were published after primary-source checking, bringing the collection to 59 records and 226 deduplicated source links. Nutrition and dietary-pattern backfill was paired with a seven-day-overlap surveillance search. A new phase III metastatic colorectal cancer result changed the balance of evidence for high-dose vitamin D, so the earlier phase II signal remains visible beside the null confirmatory result.

A ketogenic-diet topic was held as a private draft because cancer types, diet formulations and endpoints still require separation. Search dates remain separate from per-record appraisal dates; this was selective maintenance, not exhaustive live surveillance or independent clinical sign-off.

Maintenance history

October 15, 2026: Ten botanical and nutraceutical records were published after primary-source and safety checking, bringing the collection to 69 records and 263 deduplicated source links. The backfill deliberately included negative and harmful results, including black cohosh, vitamin E, chamomile, aloe, milk thistle and the stopped Goshajinkigan phase III trial.

The overlap surveillance window was October 7–15, 2026. No material new regulator warning, correction or retraction was identified. Literature-search dates remain separate from per-record appraisal dates; this selective maintenance is not exhaustive live surveillance or independent clinical sign-off.

Maintenance history

October 16, 2026: Ten exercise, rehabilitation and supportive-physical-care records were published after primary-source checking, bringing the collection to 89 records and 328 deduplicated source links. The overlap surveillance window was October 8–16, 2026; FIT4TREAT was the main newly indexed randomized signal. Null primary outcomes were retained for neoadjuvant breast-tumor response, objective cognition, insomnia sleep duration, swallowing safety and manual-lymph-drainage prevention. No material new regulator warning, correction or retraction was identified. Search dates remain separate from per-record appraisal dates; this was selective maintenance, not exhaustive live surveillance or independent clinical sign-off.

Search and appraisal note — October 17, 2026

The overlap surveillance window covered October 9–17, 2026. PubMed/PMC, ClinicalTrials.gov, original reports, professional guidance and regulator/retraction searches were checked. A rotating historical backfill focused on mind–body, sleep and psychosocial supportive care. Ten nonduplicate clinical-question records passed source checks; no material new correction, retraction or regulator safety notice affecting these records was identified in the overlap window.

Reviews were used to assess consistency, not to count their component trials again. Registration was used to verify design and status, not as evidence of benefit. Search coverage was selective rather than exhaustive live surveillance, and the October 17 literature-search date remains distinct from each record’s appraisal date.

Coverage note · October 18, 2026

This update added ten discrete supportive-care questions and 39 previously unindexed source URLs. One report per trial was treated as a trial; full text, registry entries and DOI links are source documents, not extra trials.

The search window overlapped October 10–18, 2026; historical backfill covered acupuncture, acupressure, manual therapy and physical supportive care.
Primary outcomes remained primary: the 2026 acupuncture neuropathy trial and CryoPac limb-cooling trial were both recorded as negative on their prespecified primary endpoints.
Abstract-only limitations and unavailable funding/conflict fields are named in the individual records rather than silently filled in.

Read the dated update log →

Coverage note — October 19, 2026

The interaction / unsupported-cure-claim rotation added ten clinical-question records. Pharmacokinetic endpoints were not translated into recurrence or survival claims; observational associations were not treated as causal; multiple papers from one randomized antioxidant trial were kept as one trial. The overlapping surveillance window was October 11–19, 2026. No material new efficacy trial, correction or retraction changed an existing record; a September 17 FDA adulteration alert was added to Safety.