Population Phenotype =
  Disease Phenotype ∩
  Eligibility Criteria ∩
  Exclusion Criteria

disease_phenotype:
  name: "{INDICATION}"
  definition:
    primary:
      - type: diagnosis_code
        codes:
          - ICD10: ["C90.0", "C90.00", "C90.01", "C90.02"]
        timeframe: "any time before index"
        occurrence: ">= 1"
    supporting:
      - type: medication
        drug_class: "myeloma_agents"
        timeframe: "any time"
      - type: lab_value
        test: "M_protein"
        value: "> 0"
  validation:
    ppv_target: ">= 0.80"
    sensitivity_target: ">= 0.70"

exposure_phenotype:
  name: "{DRUG}_new_user"
  definition:
    exposure:
      - type: prescription
        drug: "{DRUG}"
        first_occurrence: true
    washout_period:
      duration: "365 days"
      no_prior: "{DRUG}"
    index_date: "first_prescription_date"
  regimen_identification:
    method: "concurrent_meds_window"
    window: "7 days"
    components:
      DRd: [daratumumab, lenalidomide, dexamethasone]
      DVd: [daratumumab, bortezomib, dexamethasone]
      VRd: [bortezomib, lenalidomide, dexamethasone]

outcome_phenotype:
  name: "progression_free_survival"
  definition:
    event:
      type: "composite"
      components:
        - name: "disease_progression"
          criteria:
            - lab_based:
                test: "M_protein"
                change: ">= 25% increase from nadir"
            - clinical:
                source: "progress_notes"
                keywords: ["PD", "progression", "増悪", "進行"]
        - name: "death"
          criteria:
            - type: "death_record"
    censoring:
      - "end_of_follow_up"
      - "treatment_switch"
      - "loss_to_follow_up"

covariate_phenotypes:
  demographics:
    age:
      source: "patient_demographics"
      calc: "index_date - birth_date"
    sex:
      source: "patient_demographics"

  disease_characteristics:
    iss_stage:
      source: "lab_values + clinical_notes"
      definition:
        stage_I: "β2MG < 3.5 AND Alb >= 3.5"
        stage_II: "NOT stage_I AND NOT stage_III"
        stage_III: "β2MG >= 5.5"
      lab_window: "-30 to 0 days from index"

    cytogenetic_risk:
      source: "clinical_summary"
      extraction: "NLP"
      high_risk_markers: ["del(17p)", "t(4;14)", "t(14;16)", "gain1q"]

  comorbidities:
    renal_impairment:
      source: "lab_values"
      definition:
        severe: "eGFR < 30"
        moderate: "eGFR >= 30 AND eGFR < 60"
        mild: "eGFR >= 60 AND eGFR < 90"
        normal: "eGFR >= 90"
      lab_window: "-30 to 0 days from index"

  frailty:
    ecog_ps:
      source: "progress_notes"
      extraction: "NLP"
      keywords: ["PS", "ECOG", "パフォーマンスステータス"]
      values: [0, 1, 2, 3, 4]

lab_specifications:
  - standard_name: "hemoglobin"
    local_codes: ["HB", "Hb", "ヘモグロビン"]
    unit: "g/dL"
    normal_range: [12.0, 18.0]
    clinical_thresholds:
      severe_anemia: "< 8.0"
      moderate_anemia: "8.0 - 10.0"
      mild_anemia: "10.0 - 12.0"

  - standard_name: "eGFR"
    local_codes: ["eGFR", "推算GFR"]
    unit: "mL/min/1.73m2"
    calculation: "CKD-EPI or MDRD"
    clinical_thresholds:
      G5: "< 15"
      G4: "15 - 29"
      G3b: "30 - 44"
      G3a: "45 - 59"
      G2: "60 - 89"
      G1: ">= 90"

  - standard_name: "beta2_microglobulin"
    local_codes: ["β2MG", "β2-MG", "B2M"]
    unit: "mg/L"
    clinical_thresholds:
      elevated: "> 3.5"
      high: "> 5.5"

lab_time_series:
  baseline_value:
    window: "-30 to 0 days from index"
    selection: "closest to index"
    missing_handling: "extend window to -60 days"

  longitudinal_values:
    frequency: "all available"
    time_points: "relative to index date"
    aggregation:
      - method: "last value carried forward"
      - method: "time-weighted average"

  derived_measures:
    nadir:
      definition: "minimum value during treatment"
    change_from_baseline:
      definition: "(current - baseline) / baseline * 100"
    time_to_event:
      definition: "first time threshold crossed"

# Phenotype Specification: {PHENOTYPE_NAME}

## 1. メタデータ

| 項目 | 値 |
|------|-----|
| フェノタイプ名 | {PHENOTYPE_NAME} |
| 種類 | {PHENOTYPE_TYPE} |
| バージョン | v1.0 |
| 作成日 | YYYY-MM-DD |
| 参照定義 | {REFERENCE_DEFINITION} |

## 2. コンピューター実行可能定義

### 2.1 対象集団定義
```yaml
[YAML形式のフェノタイプ定義]

[YAML形式の曝露定義]

[YAML形式のアウトカム定義]

[YAML形式の共変量定義]

[YAML形式の時系列処理仕様]

## クレームDB差別化ポイント

### 本仕様でのみ可能な定義
| フェノタイプ | クレームDB | EHR-RWD | 差別化 |
|-------------|-----------|---------|--------|
| ISS病期 | 不可（検査値なし） | 検査値から算出可 | ○ |
| 細胞遺伝学リスク | 不可 | サマリーから抽出可 | ○ |
| ECOG PS | 不可 | 経過記録から抽出可 | ○ |
| 血清学的奏効 | 不可 | M蛋白/FLC推移から判定可 | ○ |
| AE Grade | 不可 | 経過記録から抽出可 | ○ |

## 注意事項

1. **定義の明確性**: 曖昧さのない実行可能な定義を作成
2. **バリデーションの必須化**: 全フェノタイプにバリデーション計画を含める
3. **欠測への対応**: 欠測率と対応策を事前に計画
4. **バージョン管理**: 定義変更時はバージョン更新と変更履歴記録
5. **臨床的妥当性**: 定義が臨床的に意味のある集団を捉えているか確認