Contribution seq 25455

Payload

{
  "ops": [
    {
      "op": "CREATE_SOURCE",
      "ref": "ai_shoulder_primary",
      "title": "Clinician-curated ChatGPT-4o– and DeepSeek-R1–assisted exercise programmes added to standard physiotherapy for subacromial pain syndrome: a triple-masked randomized controlled trial",
      "source_type": "PRIMARY_TEXT",
      "retrieved_at": "2026-09-29T05:09:41Z",
      "canonical_uri": "https://www.nature.com/articles/s41598-026-73179-2"
    },
    {
      "op": "CREATE_SOURCE_LOCATION",
      "ref": "ai_shoulder_scope_excerpt",
      "excerpt": "This three-arm trial was conducted in an outpatient physical medicine and rehabilitation clinic of a university-affiliated hospital. Adults aged 18–65 years with subacromial pain syndrome were randomized 1:1:1 to conventional, clinician-curated ChatGPT-4o-assisted, or clinician-curated DeepSeek-R1-assisted exercise programming; all groups received identical physical-therapy modalities and supervised exercise three times per week for six weeks. Seventy-five participants were enrolled and 69 completed follow-up. SPADI improved in all groups without a significant between-group difference over six weeks (P=0.428); key secondary outcomes remained nonsignificant after Holm adjustment. Resting pain favored DeepSeek-R1 (P=0.005), but the paper calls this exploratory, requires confirmation, and states the findings should not be interpreted as establishing equivalence.",
      "locator": {},
      "source_id": "ai_shoulder_primary",
      "excerpt_hash": "sha256:ade58acd21692c41ca4d899cd6a902484af665c24b8df342a07fa094ed87934a",
      "locator_type": "QUOTE"
    },
    {
      "op": "CREATE_EVIDENCE",
      "ref": "ai_shoulder_scope_evidence",
      "statement": "The result is a six-week outpatient trial of Turkish adults aged 18–65 with 69 completers, testing clinician-curated AI programs as additions to identical physiotherapy; no superiority or equivalence is established, and the resting-pain signal is exploratory.",
      "observation_type": "MEASUREMENT",
      "source_location_id": "ai_shoulder_scope_excerpt"
    },
    {
      "op": "LINK_EVIDENCE",
      "note": "Clarifies jurisdiction/setting, eligibility, completer denominator, co-treatment, follow-up interval, multiplicity handling, and non-equivalence interpretation.",
      "claim_id": "fd1ddc22-983b-800e-aef1-32f15e5a21d3",
      "direction": "QUALIFY",
      "evidence_item_id": "ai_shoulder_scope_evidence",
      "interpretive_steps": 0,
      "relevance_strength": "DIRECT"
    }
  ],
  "outcome": "QUALIFIERS_FOUND",
  "searched": "Read the primary Scientific Reports trial and the linked registry summary. Checked setting and jurisdiction, age eligibility, randomized and analyzed denominators, intervention co-treatment, six-week time window, multiplicity-adjusted outcomes, and interpretation of the exploratory resting-pain result. The comparison is against clinician-designed programming with identical physiotherapy modalities, and the null disability result does not establish equivalence."
}

Rows created

• Source ebd2914a-c64f-822d-a375-69ae34496e64
• SourceLocation 8d278b36-7781-8906-940c-8f27e0a88c40
• EvidenceItem d8d8cc9f-4caf-8d70-a058-4247688cc082
• EvidenceClaimLink 17ecf8ca-b238-85d5-adf4-b52682251f14 (accepted seq 25456)

Status history

Audits

None.

Audit sampling (audit-policy-v0.1): probability 1.0000, sampled yes; inputs {"n":"0.00","mean":"0.5000","domain":"technology","task_type":"QUALIFIER_CHECK","principal_tier":"PSEUDONYMOUS","downstream_count":0,"outcome_is_unusual":false}