data: ASP ingestion — strict skin sub filter, full provenance, 86 canonical URLs

Rewrite ingest_asp.py with academic rigor:
- Strict filter to skin_substitute_hcpcs.csv code universe (NO regex)
- 28 non-skin-sub Q4 codes excluded (Q4054 darbepoetin, Q4055 epoetin,
  Q4074-Q4099 ESRD drugs) that were incorrectly included before
- source_url and source_file columns on every output row (100% provenance)
- Documented column mapping rationale across 20 years of CMS format drift
- DuckDB path corrected to data/aco.duckdb

Result: 1,974 observations, 68 quarters (2009-Q1–2025-Q4), 169 codes
Canonical source URLs: data/cms/asp_source_urls.csv (86 CMS.gov URLs)

Refs #233

data/cms/asp_source_urls.csv

@@ -0,0 +1,87 @@
+quarter,url
+2026-Q2,https://www.cms.gov/files/zip/april-2026-medicare-part-b-payment-limit-files-03-24-2026-final-file.zip
+2026-Q1,https://www.cms.gov/files/zip/january-2026-medicare-part-b-payment-limit-files.zip
+2025-Q4,https://www.cms.gov/files/zip/october-2025-asp-pricing-final-file.zip
+2025-Q3,https://www.cms.gov/files/zip/july-2025-asp-pricing-file.zip
+2025-Q2,https://www.cms.gov/files/zip/april-2025-asp-pricing-file.zip
+2025-Q1,https://www.cms.gov/files/zip/january-2025-asp-pricing-file-03/11/25-final-file.zip
+2024-Q4,https://www.cms.gov/files/zip/october-2024-asp-pricing-file.zip
+2024-Q3,https://www.cms.gov/files/zip/july-2024-asp-pricing-file.zip
+2024-Q2,https://www.cms.gov/files/zip/april-2024-asp-pricing-file.zip
+2024-Q1,https://www.cms.gov/files/zip/january-2024-asp-pricing-file.zip
+2023-Q4,https://www.cms.gov/files/zip/october-2023-asp-pricing-file.zip
+2023-Q3,https://www.cms.gov/files/zip/july-2023-asp-pricing-file.zip
+2023-Q2,https://www.cms.gov/files/zip/april-2023-asp-pricing-file.zip
+2023-Q1,https://www.cms.gov/files/zip/january-2023-asp-pricing-file.zip
+2022-Q4,https://www.cms.gov/files/zip/october-2022-asp-pricing-file.zip
+2022-Q3,https://www.cms.gov/files/zip/july-2022-asp-pricing-file.zip
+2022-Q2,https://www.cms.gov/files/zip/april-2022-asp-pricing-file.zip
+2022-Q1,https://www.cms.gov/files/zip/january-2022-asp-pricing-file.zip
+2021-Q4,https://www.cms.gov/files/zip/october-2021-asp-pricing-file.zip
+2021-Q3,https://www.cms.gov/files/zip/july-2021-asp-pricing-file.zip
+2021-Q2,https://www.cms.gov/files/zip/april-2021-asp-pricing-file.zip
+2021-Q1,https://www.cms.gov/files/zip/january-2021-asp-pricing-file.zip
+2020-Q4,https://www.cms.gov/files/zip/october-2020-asp-pricing-file.zip
+2020-Q3,https://www.cms.gov/files/zip/july-2020-asp-pricing-file-updated-06012020.zip
+2020-Q2,https://www.cms.gov/files/zip/april-2020-asp-pricing-file.zip
+2020-Q1,https://www.cms.gov/files/zip/january-2020-asp-pricing-file.zip
+2019-Q4,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/2019-Oct-ASP-Pricing-File.zip
+2019-Q3,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/Downloads/2019-July-ASP-Pricing-File.zip
+2019-Q2,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/April-2019-ASP-Pricing-File.zip
+2019-Q1,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/Downloads/2019-January-ASP-Pricing-File.zip
+2018-Q4,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/Downloads/2018-Oct-ASP-Pricing-File.zip
+2018-Q3,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/Downloads/2018-July-ASP-Pricing-File.zip
+2018-Q2,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/Downloads/2018-April-ASP-Pricing-File.zip
+2018-Q1,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/Downloads/2018-January-ASP-Pricing-File.zip
+2017-Q4,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/Downloads/2017-October-ASP-Pricing-File.zip
+2017-Q3,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/Downloads/2017-July-ASP-Pricing-File.zip
+2017-Q2,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/Downloads/2017-April-ASP-Pricing.zip
+2017-Q1,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/Downloads/2017-January-ASP-Pricing-Files.zip
+2016-Q4,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/Downloads/2016-October-ASP-Pricing-File.zip
+2016-Q3,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/Downloads/2016-July-ASP-Pricing-File.zip
+2016-Q2,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/2016-April-ASP-Pricing-File.zip
+2016-Q1,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/2016-January-ASP-Pricing-File.zip
+2015-Q4,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/2015-October-ASP-Pricing-File.zip
+2015-Q3,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/2015-July-ASP-Pricing-File.zip
+2015-Q2,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/2015-April-ASP-Pricing-File.zip
+2015-Q1,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/2015-January-ASP-Pricing-File.zip
+2014-Q4,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/2014-October-ASP-Pricing-File.zip
+2014-Q3,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/Jul-2014-ASP-Pricing-File.zip
+2014-Q2,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/Apr-14-ASP-Pricing-File.zip
+2014-Q1,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/Jan-14-ASP-Pricing-File.zip
+2013-Q4,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/2013-October-ASP-Pricing-File.zip
+2013-Q3,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/2013-July-ASP-Pricing-File.zip
+2013-Q2,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/Apr-13-ASP-Pricing-file.zip
+2013-Q1,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/Jan-2013-ASP-Pricing-File.zip
+2012-Q4,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/October-2012-ASP-Pricing-File.zip
+2012-Q3,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/July-2012-ASP-Pricing-File.zip
+2012-Q2,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/April-2012-ASP-Pricing-file-revised030513.zip
+2012-Q1,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/Jan_2012_ASP_Pricing_File.zip
+2011-Q4,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/Oct_2011_ASP_Pricing_File.zip
+2011-Q3,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/July_2011_ASP_Pricing_File.zip
+2011-Q2,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/April_2011_ASP_Pricing_File.zip
+2011-Q1,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/January2011_ASP_PricingFile.zip
+2010-Q4,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/October_2010_ASP_Pricing_File.zip
+2010-Q3,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/July_2010_ASP_Pricing_File.zip
+2010-Q2,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/April_2010_ASP_Pricing_File.zip
+2010-Q1,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/January_2010_ASP_Pricing_File.zip
+2009-Q4,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/October_2009_ASP_Pricing_File.zip
+2009-Q3,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/July_2009_ASP_Pricing_File.zip
+2009-Q2,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/April_2009_ASP_Pricing_File_by_HCPCS.zip
+2009-Q1,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/JAN_2009_ASP_Pricing_File_by_HCPCS.zip
+2008-Q4,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/October2008ASPPricingFilebyHCPCS.zip
+2008-Q3,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/July2008ASPPricingFilebyHCPCS.zip
+2008-Q2,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/April08ASPbyHCPCS.zip
+2008-Q1,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/Jan08ASPbyHCPCS.zip
+2007-Q4,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/October07ASPbyHCPCS.zip
+2007-Q3,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/July07ASPbyHCPCS_121707.zip
+2007-Q2,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/April07ASPbyHCPCS_121707.zip
+2007-Q1,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/Jan07ASPbyHCPCS_121707.zip
+2006-Q4,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/oct06asp_hcpcs.zip
+2006-Q3,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/july06asp_hcpcs.zip
+2006-Q2,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/apr06pricing.zip
+2006-Q1,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/Jan06pricing.zip
+2005-Q4,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/oct05aspbyhcpcs033106v2.zip
+2005-Q3,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/jul05aspbyhcpcs_032906.zip
+2005-Q2,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/apr05aspbyhcpcsv8_033106.zip
+2005-Q1,https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-B-Drugs/McrPartBDrugAvgSalesPrice/downloads/jan05aspbyhcpcsv6_033106.zip

dev/scripts/ingest_asp.py

@@ -0,0 +1,353 @@
+"""Ingest CMS Average Sales Price (ASP) Drug Pricing Files.
+
+Reads all quarterly ASP ZIP files from data/cms/asp/, normalises column
+names across 20+ years of CMS format drift, filters **strictly** to
+skin substitute HCPCS codes defined in data/cms/skin_substitute_hcpcs.csv,
+and produces:
+
+  * data/cms/asp/skin_subs_asp_quarterly.csv   (flat file with provenance)
+  * skin_subs.asp_quarterly                     (DuckDB table)
+
+Every output row carries ``source_url`` and ``source_file`` columns for
+full provenance — each observation traces back to a specific CMS
+quarterly ZIP and the file within it that was parsed.
+
+Source URLs are read from data/cms/asp_source_urls.csv (canonical URLs
+provided from CMS.gov, one per quarter 2005-Q1 through 2026-Q2).
+
+Column mapping rationale
+------------------------
+CMS has changed column names across 20+ years of ASP files:
+
+  2005–2006: "HCPCS Code", "Short Description", "HCPCS Code Dosage",
+             "Estimated ASP", "Payment Limit"
+  2007–2011: Same but sometimes "Dosage" instead of "HCPCS Code Dosage"
+  2012–2019: "HCPCS Code", "Short Description", "Dosage",
+             "Payment Limit"
+  2020–2025: "HCPCS Code", "Short Description", "HCPCS Code Dosage",
+             "Payment Limit"
+  2026+:     "HCPCS Code", "Short Description", "Dosage or Unit",
+             "Payment Limit" (now "Payment Limit" for skin subs is
+             flat $127.14/cm² under reclassification)
+
+All variants are mapped to: hcpcs_code, short_description,
+dosage_or_unit, payment_limit.
+
+The ``asp_per_unit`` column is derived: payment_limit / 1.06, since
+CMS payment limit = ASP + 6% for drugs/biologicals. For 2026+ skin
+subs under reclassification, this derivation no longer applies (flat
+rate), but we retain it for consistency and flag it in notes.
+
+Filtering
+---------
+We filter STRICTLY to codes present in skin_substitute_hcpcs.csv
+(the code universe built from CMS rulemaking in issue #232). This
+is critical because the Q4xxx prefix includes non-skin-sub codes
+(Q4054 darbepoetin alfa, Q4055 epoetin alfa, Q4074-Q4082 various
+ESRD drugs) that must be excluded.
+
+Usage:
+    uv run python dev/scripts/ingest_asp.py
+"""
+
+from __future__ import annotations
+
+import csv
+import io
+import re
+import zipfile
+from pathlib import Path
+
+import duckdb
+import pandas as pd
+
+# ---------------------------------------------------------------------------
+# Paths
+# ---------------------------------------------------------------------------
+ROOT = Path(__file__).resolve().parents[2]
+ASP_DIR = ROOT / "data" / "cms" / "asp"
+HCPCS_REF = ROOT / "data" / "cms" / "skin_substitute_hcpcs.csv"
+URL_REF = ROOT / "data" / "cms" / "asp_source_urls.csv"
+DUCKDB_PATH = ROOT / "data" / "aco.duckdb"
+OUTPUT_CSV = ASP_DIR / "skin_subs_asp_quarterly.csv"
+
+
+def _load_source_urls() -> dict[str, str]:
+    """Load canonical source URLs: {quarter: url}."""
+    urls = {}
+    if URL_REF.exists():
+        with URL_REF.open() as f:
+            for row in csv.DictReader(f):
+                urls[row["quarter"]] = row["url"]
+    return urls
+
+
+# ---------------------------------------------------------------------------
+# Column normalisation
+# ---------------------------------------------------------------------------
+# Mapping from observed CMS column names (lowercased) to our standard names.
+# Each entry documents which CMS file years use that variant.
+COL_MAP = {
+    # HCPCS code — consistent across all years
+    "hcpcs code": "hcpcs_code",
+    "hcpcs_code": "hcpcs_code",
+    # Short description — consistent
+    "short description": "short_description",
+    "short_description": "short_description",
+    # Dosage/unit — varies by year
+    "hcpcs code dosage": "dosage_or_unit",  # 2005-2006, 2020-2025
+    "hcpcs_code_dosage": "dosage_or_unit",
+    "dosage": "dosage_or_unit",  # 2007-2019
+    "dosage or unit": "dosage_or_unit",  # 2026+
+    # Payment limit — consistent (ASP + 6%)
+    "payment limit": "payment_limit",
+    "payment_limit": "payment_limit",
+}
+
+KEEP_COLS = ["hcpcs_code", "short_description", "dosage_or_unit", "payment_limit"]
+
+
+def normalise_columns(df: pd.DataFrame) -> pd.DataFrame:
+    """Lower-case columns, apply COL_MAP, keep only KEEP_COLS."""
+    df.columns = [str(c).strip().lower() for c in df.columns]
+    df = df.rename(columns=COL_MAP)
+    present = [c for c in KEEP_COLS if c in df.columns]
+    return df[present]
+
+
+# ---------------------------------------------------------------------------
+# Readers
+# ---------------------------------------------------------------------------
+
+
+def _find_header_row(lines: list[str]) -> int:
+    """Return the 0-based row index of the column-header row.
+
+    The header row starts with 'HCPCS Code,' (possibly quoted).
+    Must NOT match note lines that incidentally mention 'HCPCS code'.
+    """
+    for i, line in enumerate(lines):
+        stripped = line.strip().lower()
+        if stripped.startswith("hcpcs code,") or stripped.startswith('"hcpcs code",'):
+            return i
+    return -1
+
+
+def read_csv_from_zip(zf: zipfile.ZipFile, name: str) -> pd.DataFrame:
+    """Read a section-508 CSV inside a ZIP, skipping preamble rows."""
+    raw = zf.read(name).decode("latin-1")
+    lines = raw.splitlines()
+    hdr = _find_header_row(lines)
+    if hdr < 0:
+        return pd.DataFrame()
+    buf = "\n".join(lines[hdr:])
+    df = pd.read_csv(io.StringIO(buf), dtype=str, on_bad_lines="skip")
+    return normalise_columns(df)
+
+
+def read_xls_from_zip(zf: zipfile.ZipFile, name: str) -> pd.DataFrame:
+    """Read .xls inside a ZIP, skipping preamble rows."""
+    import xlrd
+
+    data = zf.read(name)
+    wb = xlrd.open_workbook(file_contents=data)
+    sh = wb.sheet_by_index(0)
+    hdr_row = -1
+    for r in range(min(30, sh.nrows)):
+        first_cell = str(sh.cell_value(r, 0)).strip().lower()
+        if first_cell == "hcpcs code":
+            hdr_row = r
+            break
+    if hdr_row < 0:
+        return pd.DataFrame()
+    headers = [str(sh.cell_value(hdr_row, c)).strip() for c in range(sh.ncols)]
+    rows = []
+    for r in range(hdr_row + 1, sh.nrows):
+        row = [str(sh.cell_value(r, c)).strip() for c in range(sh.ncols)]
+        if any(row):
+            rows.append(row)
+    df = pd.DataFrame(rows, columns=headers)
+    return normalise_columns(df)
+
+
+def pick_pricing_file(names: list[str]) -> str | None:
+    """From a list of files in a ZIP, pick the ASP pricing file.
+
+    Prefer CSV (section 508) over XLS. Skip NOC, crosswalk, NDC,
+    and "not payable" files.
+    """
+    skip_patterns = ["noc", "crosswalk", "ndc", "not payable", "not_payable"]
+    candidates: list[str] = []
+    for n in names:
+        low = n.lower()
+        if any(s in low for s in skip_patterns):
+            continue
+        if low.endswith((".csv", ".xls", ".xlsx")):
+            candidates.append(n)
+
+    csvs = [c for c in candidates if c.lower().endswith(".csv")]
+    if csvs:
+        for c in csvs:
+            low = c.lower()
+            if "payment limit" in low or "pricing" in low:
+                return c
+        return csvs[0]
+
+    xlss = [c for c in candidates if c.lower().endswith((".xls", ".xlsx"))]
+    if xlss:
+        for c in xlss:
+            low = c.lower()
+            if "payment limit" in low or "pricing" in low or "asp" in low:
+                return c
+        return xlss[0]
+
+    return None
+
+
+# ---------------------------------------------------------------------------
+# Numeric cleaning
+# ---------------------------------------------------------------------------
+
+
+def clean_numeric(series: pd.Series) -> pd.Series:
+    """Strip $ signs and convert to float, coercing errors to NaN."""
+    return (
+        series.astype(str)
+        .str.replace("$", "", regex=False)
+        .str.replace(",", "", regex=False)
+        .str.strip()
+        .pipe(pd.to_numeric, errors="coerce")
+    )
+
+
+# ---------------------------------------------------------------------------
+# Main
+# ---------------------------------------------------------------------------
+
+
+def main() -> None:
+    # Load skin-substitute HCPCS codes (strict filter — no regex fallback)
+    hcpcs_ref = pd.read_csv(HCPCS_REF, dtype=str)
+    skin_codes = set(hcpcs_ref["hcpcs_code"].str.strip().str.upper())
+    print(f"Loaded {len(skin_codes)} skin-substitute HCPCS codes from reference")
+
+    # Load canonical source URLs
+    source_urls = _load_source_urls()
+    print(f"Loaded {len(source_urls)} source URLs")
+
+    # Ingest all quarters
+    zips = sorted(ASP_DIR.glob("asp_*.zip"))
+    print(f"Found {len(zips)} quarterly ZIP files\n")
+
+    frames: list[pd.DataFrame] = []
+    for zpath in zips:
+        m = re.match(r"asp_(\d{4}-Q\d)", zpath.name)
+        if not m:
+            print(f"  SKIP {zpath.name} (cannot parse quarter)")
+            continue
+        quarter = m.group(1)
+
+        try:
+            with zipfile.ZipFile(zpath) as zf:
+                target = pick_pricing_file(zf.namelist())
+                if target is None:
+                    print(f"  SKIP {quarter}: no pricing file found in ZIP")
+                    continue
+
+                if target.lower().endswith(".csv"):
+                    df = read_csv_from_zip(zf, target)
+                else:
+                    df = read_xls_from_zip(zf, target)
+
+                if df.empty:
+                    print(f"  SKIP {quarter}: empty after parsing {target}")
+                    continue
+
+                df["quarter"] = quarter
+                df["source_file"] = target
+                df["source_url"] = source_urls.get(quarter, "")
+                frames.append(df)
+                print(f"  OK   {quarter}: {len(df):>5} rows from {target}")
+        except Exception as e:
+            print(f"  ERR  {quarter}: {e}")
+
+    if not frames:
+        raise RuntimeError("No data parsed from any file")
+
+    all_df = pd.concat(frames, ignore_index=True)
+    print(f"\nTotal rows across all quarters: {len(all_df)}")
+
+    # Normalise HCPCS codes
+    all_df["hcpcs_code"] = all_df["hcpcs_code"].astype(str).str.strip().str.upper()
+
+    # Filter STRICTLY to skin substitute code universe
+    # NO regex fallback — only codes in skin_substitute_hcpcs.csv
+    skin_df = all_df[all_df["hcpcs_code"].isin(skin_codes)].copy()
+    print(f"Skin substitute rows (strict filter): {len(skin_df)}")
+
+    # Show what we excluded
+    q4_but_not_skin = all_df[
+        all_df["hcpcs_code"].str.match(r"^Q4\d{2,3}$")
+        & ~all_df["hcpcs_code"].isin(skin_codes)
+    ]["hcpcs_code"].unique()
+    if len(q4_but_not_skin) > 0:
+        print(f"  Excluded Q4 codes NOT in skin sub universe: {sorted(q4_but_not_skin)}")
+
+    if not skin_df.empty:
+        skin_df["payment_limit"] = clean_numeric(skin_df["payment_limit"])
+        # ASP per unit = payment_limit / 1.06
+        # (CMS payment limit = ASP + 6% for drugs/biologicals)
+        skin_df["asp_per_unit"] = (skin_df["payment_limit"] / 1.06).round(3)
+
+    # Reorder columns
+    out_cols = [
+        "quarter",
+        "hcpcs_code",
+        "short_description",
+        "asp_per_unit",
+        "payment_limit",
+        "dosage_or_unit",
+        "source_url",
+        "source_file",
+    ]
+    for c in out_cols:
+        if c not in skin_df.columns:
+            skin_df[c] = None
+    skin_df = skin_df[out_cols].sort_values(["quarter", "hcpcs_code"])
+
+    # Write CSV
+    skin_df.to_csv(OUTPUT_CSV, index=False)
+    print(f"\nWrote {len(skin_df)} rows to {OUTPUT_CSV}")
+
+    if len(skin_df) > 0:
+        quarters = sorted(skin_df["quarter"].unique())
+        codes = sorted(skin_df["hcpcs_code"].unique())
+        print(f"  Quarters: {len(quarters)} ({quarters[0]} to {quarters[-1]})")
+        print(f"  Unique HCPCS codes: {len(codes)}")
+
+    # Load into DuckDB
+    print(f"\nLoading into DuckDB at {DUCKDB_PATH} ...")
+    con = duckdb.connect(str(DUCKDB_PATH))
+    con.execute("CREATE SCHEMA IF NOT EXISTS skin_subs")
+    con.execute("DROP TABLE IF EXISTS skin_subs.asp_quarterly")
+    if not skin_df.empty:
+        con.execute(
+            """
+            CREATE TABLE skin_subs.asp_quarterly AS
+            SELECT * FROM read_csv_auto(?, header=true)
+            """,
+            [str(OUTPUT_CSV)],
+        )
+        count = con.execute(
+            "SELECT count(*) FROM skin_subs.asp_quarterly"
+        ).fetchone()[0]
+        print(f"  Loaded {count} rows into skin_subs.asp_quarterly")
+    else:
+        print("  WARNING: No skin substitute data found in ASP files")
+        print("  This may be expected — skin subs may not appear in main ASP pricing files")
+        print("  They may be in separate NOC or tissue coding files")
+    con.close()
+
+
+if __name__ == "__main__":
+    main()