feat: remove cost/费用 statistics from clinical analytics

Per request — drop all monetary statistics (住院费用 is sensitive).
- Backend statistics.py: remove mean_cost KPI + cost_histogram / cost_by_disease
  / cost_vs_los from /inpatient-clinical (models, computation, response)
- Frontend: drop 人均费用 KPI card (now 4 KPIs), 住院费用分布, 各病种平均费用,
  费用×住院天数散点; delete CostByDiseaseChart + CostVsLosScatter components;
  trim statsApi type + e2e fixture + chartColors

Clinical page now: KPI(总人次/中位住院日/治愈好转率/急诊占比) + LOS dist + LOS-by-disease
box + outcome donut + admission-route donut + age-band BMI box.

Gates: backend 106 pytest · tsc 0 · build ok · clinical+user-flows e2e 19/19 ·
live endpoint confirmed cost-free

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

backend/routers/statistics.py

@@ -116,7 +116,6 @@ class KeyValueCount(BaseModel):
 class InpatientKpis(BaseModel):
     total_admissions: int
     median_los_days: float
-    mean_cost: float
     cure_rate: float
     emergency_admit_ratio: float
 
@@ -129,17 +128,6 @@ class LosByDisease(BaseModel):
     n: int
 
 
-class CostByDisease(BaseModel):
-    diagnosis: str
-    mean_cost: float
-    n: int
-
-
-class CostVsLos(BaseModel):
-    los: int
-    cost: float
-
-
 class LabelCount(BaseModel):
     outcome: Optional[str] = None
     route: Optional[str] = None
@@ -168,9 +156,6 @@ class InpatientClinicalResponse(BaseModel):
     kpis: InpatientKpis
     los_histogram: list[KeyValueCount]
     los_by_disease: list[LosByDisease]
-    cost_histogram: list[KeyValueCount]
-    cost_by_disease: list[CostByDisease]
-    cost_vs_los: list[CostVsLos]
     outcome_counts: list[OutcomeCount]
     admission_route_counts: list[RouteCount]
     bmi_by_age_band: list[BmiByAge]
@@ -252,11 +237,10 @@ class TemporalResponse(BaseModel):
 def _empty_inpatient_clinical() -> InpatientClinicalResponse:
     return InpatientClinicalResponse(
         kpis=InpatientKpis(
-            total_admissions=0, median_los_days=0.0, mean_cost=0.0,
+            total_admissions=0, median_los_days=0.0,
             cure_rate=0.0, emergency_admit_ratio=0.0,
         ),
-        los_histogram=[], los_by_disease=[], cost_histogram=[],
-        cost_by_disease=[], cost_vs_los=[], outcome_counts=[],
+        los_histogram=[], los_by_disease=[], outcome_counts=[],
         admission_route_counts=[], bmi_by_age_band=[],
     )
 
@@ -287,8 +271,6 @@ def _compute_inpatient_clinical() -> InpatientClinicalResponse:
 
     total = len(df)
     median_los = float(df_los["los"].median()) if len(df_los) else 0.0
-    cost = pd.to_numeric(df["住院总费用"], errors="coerce")
-    mean_cost = float(cost.mean()) if cost.notna().any() else 0.0
 
     outcome = df["出院情况"].fillna("未知")
     cure_n = int(outcome.isin(["治愈", "好转"]).sum())
@@ -301,7 +283,6 @@ def _compute_inpatient_clinical() -> InpatientClinicalResponse:
     kpis = InpatientKpis(
         total_admissions=total,
         median_los_days=round(median_los, 2),
-        mean_cost=round(mean_cost, 2),
         cure_rate=round(cure_rate, 4),
         emergency_admit_ratio=round(emerg_ratio, 4),
     )
@@ -328,39 +309,6 @@ def _compute_inpatient_clinical() -> InpatientClinicalResponse:
                 n=int(len(grp)),
             ))
 
-    # Cost histogram: 0-2k,2-4k,4-6k,6-8k,8-10k,10k+
-    cost_valid = cost.dropna()
-    cost_bins = [(0, 2000, "0-2k"), (2000, 4000, "2-4k"), (4000, 6000, "4-6k"),
-                 (6000, 8000, "6-8k"), (8000, 10000, "8-10k")]
-    cost_histogram: list[KeyValueCount] = []
-    for lo, hi, label in cost_bins:
-        cost_histogram.append(KeyValueCount(
-            bin_label=label, count=int(((cost_valid >= lo) & (cost_valid < hi)).sum())))
-    cost_histogram.append(KeyValueCount(bin_label="10k+", count=int((cost_valid >= 10000).sum())))
-
-    # Cost by disease (top 8 by n)
-    cost_by_disease: list[CostByDisease] = []
-    df_cost = df[cost.notna()].copy()
-    df_cost["_cost"] = cost[cost.notna()]
-    if len(df_cost):
-        top_cd = df_cost["诊断名称"].value_counts().head(8).index.tolist()
-        for d in top_cd:
-            grp = df_cost[df_cost["诊断名称"] == d]["_cost"]
-            cost_by_disease.append(CostByDisease(
-                diagnosis=str(d),
-                mean_cost=round(float(grp.mean()), 2),
-                n=int(len(grp)),
-            ))
-
-    # cost vs los scatter (up to 500 points)
-    cost_vs_los: list[CostVsLos] = []
-    scatter_df = df_los[cost.reindex(df_los.index).notna()].copy()
-    scatter_df["_cost"] = cost.reindex(scatter_df.index)
-    if len(scatter_df) > 500:
-        scatter_df = scatter_df.sample(n=500, random_state=42)
-    for _, r in scatter_df.iterrows():
-        cost_vs_los.append(CostVsLos(los=int(r["los"]), cost=round(float(r["_cost"]), 2)))
-
     # outcome counts
     outcome_counts = [
         OutcomeCount(outcome=str(k), count=int(v))
@@ -399,9 +347,6 @@ def _compute_inpatient_clinical() -> InpatientClinicalResponse:
         kpis=kpis,
         los_histogram=los_histogram,
         los_by_disease=los_by_disease,
-        cost_histogram=cost_histogram,
-        cost_by_disease=cost_by_disease,
-        cost_vs_los=cost_vs_los,
         outcome_counts=outcome_counts,
         admission_route_counts=admission_route_counts,
         bmi_by_age_band=bmi_by_age_band,