pycomorb 0.4.2__tar.gz → 0.5.1__tar.gz

{pycomorb-0.4.2/src/pycomorb.egg-info → pycomorb-0.5.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: pycomorb
-Version: 0.4.2
+Version: 0.5.1
 Summary: Comorbidity and frailty score calculation utilities (Charlson, Elixhauser, Gagne, HFRS) using polars
 Author-email: Finn Fassbender <finn.fassbender@charite.de>
 License-Expression: MIT
@@ -23,6 +23,8 @@ Python package to calculate comorbidity and frailty scores using ICD codes and p
 
 It is inspired by the R package [`comorbidity`](https://github.com/ellessenne/comorbidity) by Alessandro Gasparini and the corresponding Python rewrite [`comorbidipy`](https://github.com/vvcb/comorbidipy) by Vishnu V Chandrabalan.
 
+As a special feature, it also implements the history of the German Modification of the ICD-10 (ICD-10-GM) coding system, allowing users to apply the correct mappings and weights based on the year of diagnosis.
+
 ## Included Indices / Scores
 
 - **Charlson Comorbidity Index** (multiple mappings and weights)
@@ -108,3 +110,4 @@ See the docstrings in each module for details on arguments and supported variant
 
 - Free software: MIT license
 - Documentation: (TODO)
+- _Die Erstellung erfolgt unter Verwendung der maschinenlesbaren Fassung des Bundesinstituts für Arzneimittel und Medizinprodukte (BfArM)._

{pycomorb-0.4.2 → pycomorb-0.5.1}/README.md

@@ -8,6 +8,8 @@ Python package to calculate comorbidity and frailty scores using ICD codes and p
 
 It is inspired by the R package [`comorbidity`](https://github.com/ellessenne/comorbidity) by Alessandro Gasparini and the corresponding Python rewrite [`comorbidipy`](https://github.com/vvcb/comorbidipy) by Vishnu V Chandrabalan.
 
+As a special feature, it also implements the history of the German Modification of the ICD-10 (ICD-10-GM) coding system, allowing users to apply the correct mappings and weights based on the year of diagnosis.
+
 ## Included Indices / Scores
 
 - **Charlson Comorbidity Index** (multiple mappings and weights)
@@ -93,3 +95,4 @@ See the docstrings in each module for details on arguments and supported variant
 
 - Free software: MIT license
 - Documentation: (TODO)
+- _Die Erstellung erfolgt unter Verwendung der maschinenlesbaren Fassung des Bundesinstituts für Arzneimittel und Medizinprodukte (BfArM)._

{pycomorb-0.4.2 → pycomorb-0.5.1}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "pycomorb"
-version = "0.4.2"
+version = "0.5.1"
 authors = [
   { name="Finn Fassbender", email="finn.fassbender@charite.de" },
 ]

{pycomorb-0.4.2 → pycomorb-0.5.1}/src/pycomorb/CharlsonComorbidityIndex.py

@@ -21,6 +21,10 @@
 #    Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data.
 #    Med Care. 2005 Nov;43(11):1130-9.
 #    doi: 10.1097/01.mlr.0000182534.19832.83. PMID: 16224307.
+# 7. Armitage JN, van der Meulen JH; Royal College of Surgeons Co-morbidity Consensus Group.
+#    Identifying co-morbidity in surgical patients using administrative data with the Royal College of Surgeons Charlson Score.
+#    Br J Surg. 2010 May;97(5):772-81.
+#    doi: 10.1002/bjs.6930. PMID: 20306528.
 
 from pathlib import Path
 
@@ -51,7 +55,7 @@ def CharlsonComorbidityIndex(
         age_col (str): Name of the column containing patient ages. Default: "age".
         icd_version (str): ICD version ('icd9', 'icd10', or 'icd9_10'). Default: "icd10".
         icd_version_col (str, optional): Name of the column with ICD version for 'icd9_10'. Default: None.
-        implementation (str): CCI implementation ('quan', 'deyo', 'romano', 'australia', 'sweden', 'uk_shmi'). Default: "quan".
+        implementation (str): CCI implementation ('quan', 'deyo', 'romano', 'australia', 'sweden', 'rcs', 'uk_shmi'). Default: "quan".
         return_categories (bool): If True, also return presence indicators for each CCI category.
 
     Returns:
@@ -76,6 +80,7 @@ def CharlsonComorbidityIndex(
     elif icd_version == "icd9" and implementation in [
         "australia",
         "sweden",
+        "rcs",
         "uk_shmi",
     ]:
         print(f"Warning: Implementation '{implementation}' only uses ICD-10. Setting ICD version to 'icd10'.") # fmt: skip
@@ -86,12 +91,14 @@ def CharlsonComorbidityIndex(
         "quan",
         "deyo",
         "romano",
+        "dhoore",
         "australia",
         "sweden",
+        "rcs",
         "uk_shmi",
     ], (
         "implementation must be one of: "
-        "'quan', 'deyo', 'romano', 'australia', 'sweden', or 'uk_shmi'."
+        "'quan', 'deyo', 'romano', 'dhoore', 'australia', 'sweden', 'rcs', or 'uk_shmi'."
     )
     assert (
         age_col in df.columns
@@ -140,6 +147,8 @@ def CharlsonComorbidityIndex(
         definition_file = "CHARLSON_AUSTRALIA.csv"
     elif implementation == "sweden":
         definition_file = "CHARLSON_SWEDEN.csv"
+    elif implementation == "rcs":
+        definition_file = "CHARLSON_RCS.csv"
     elif implementation == "uk_shmi":
         definition_file = "CHARLSON_UK_SHMI_v1.55.csv"
     else:

pycomorb-0.5.1/src/pycomorb/ICDModifications.py

@@ -0,0 +1,61 @@
+# Reference for ICD-10-GM modifications:
+# BfArM: https://www.bfarm.de/DE/Kodiersysteme/Services/Downloads/_node.html
+
+import polars as pl
+from pathlib import Path
+
+
+def get_icd10gm(
+    data: pl.DataFrame, code_col="code", year_col="year", target_year=2024
+) -> pl.DataFrame:
+    """
+    Map ICD-10-GM codes to the target year using official transfer files.
+
+    Args:
+        data (pl.DataFrame): Input DataFrame with at least columns [code_col, year_col].
+        code_col (str): Name of the column containing ICD codes. Default: "code".
+        year_col (str): Name of the column containing the year of the ICD code. Default: "year".
+        target_year (int): Target year for ICD-10-GM codes (e.g., 2024). Default: 2024.
+
+    Returns:
+        pl.DataFrame: DataFrame with original columns plus a new column 'icd10gm_{target_year}' with mapped codes.
+    """
+
+    assert (
+        code_col in data.columns
+    ), f"Column '{code_col}' (ICD code) must be present in input DataFrame."
+    assert (
+        data.height
+        == data.filter(pl.col(code_col).str.contains(r"^[A-Z]")).height
+    ), f"All values in column '{code_col}' must start with an uppercase letter (A-Z)."
+    assert (
+        year_col in data.columns
+    ), f"Column '{year_col}' (year) must be present in input DataFrame."
+    assert target_year in range(
+        2004, 2026
+    ), "Target year must be between 2004 and 2025."
+
+    # Load all transfer files into a dictionary of DataFrames
+    transfer_file_path = Path(__file__).parent / "germanmodification/icd10gm.csv" # fmt: skip
+
+    # Unpivot the data to long format
+    transfer_data = (
+        pl.read_csv(transfer_file_path, null_values=["", " "], separator=",")
+        .unpivot(
+            index=str(target_year),
+            variable_name=year_col,
+            value_name=code_col,
+        )
+        .cast({year_col: int})
+        .rename({str(target_year): f"icd10gm_{target_year}"})
+    )
+
+    return (
+        data.join(transfer_data, on=[code_col, year_col], how="left")
+        # Prefer mapped code, but fall back to original if no mapping available
+        .with_columns(
+            pl.coalesce(
+                pl.col(f"icd10gm_{target_year}"), pl.col(code_col)
+            ).alias(f"icd10gm_{target_year}")
+        )
+    )

{pycomorb-0.4.2 → pycomorb-0.5.1}/src/pycomorb/__init__.py

@@ -2,25 +2,36 @@
 
 __author__ = """Finn Fassbender"""
 __email__ = "finn.fassbender@charite.de"
-__version__ = "0.4.2"
+__version__ = "0.5.1"
 
 from .CharlsonComorbidityIndex import CharlsonComorbidityIndex
 from .ElixhauserComorbidityIndex import ElixhauserComorbidityIndex
 from .GagneComorbidityIndex import GagneComorbidityIndex
+from .CustomComorbidityIndex import CustomComorbidityIndex
 from .HospitalFrailtyRiskScore import HospitalFrailtyRiskScore
+from .ICDModifications import get_icd10gm
 
 
 def comorbidity(
     score: str,
     df,
-    id_col="id",
-    code_col="code",
-    age_col="age",
+    id_col: str = "id",
+    code_col: str = "code",
+    age_col: str = "age",
     icd_version="icd10",
-    icd_version_col=None,
-    implementation=None,
-    weights=None,
-    return_categories=False,
+    year_col: str = "year",
+    icd_version: str = "icd10",
+    icd_version_col: str = None,
+    icd_modification: str = None,
+    icd_modification_target_year: int = 2004,
+    implementation: str = None,
+    weights: str = None,
+    definition_data=None,
+    definition_file_path: str = None,
+    weight_col_name: str = "weights",
+    score_col_name: str = "Custom Comorbidity Score",
+    mutual_exclusion_rules: list[tuple[str, str]] = None,
+    return_categories: bool = False,
 ):
     """
     Unified wrapper to calculate a comorbidity or frailty score.
@@ -31,16 +42,38 @@ def comorbidity(
         id_col (str): Name of the column containing unique identifier. Default: "id".
         code_col (str): Name of the column containing ICD codes. Default: "code".
         age_col (str): Name of the column containing patient ages (Charlson only). Default: "age".
+        year_col (str): Name of the column containing the year of the ICD code (for ICD history modifications). Default: "year".
         icd_version (str): ICD version. Default: "icd10".
         icd_version_col (str, optional): Name of the column with ICD version for 'icd9_10'. Default: None.
+        icd_modification (str, optional): ICD modification to apply ('icd10gm'). Default: None.
+        icd_modification_target_year (int): Target year for ICD modification (if applicable). Default: 2004.
         implementation (str, optional): Implementation variant (see individual index docs).
         weights (str, optional): Weighting scheme (Elixhauser only).
+        definition_data (DataFrame, optional): DataFrame with ICD definitions and weights (Custom only).
+        definition_file_path (str, optional): Path to CSV with ICD definitions and weights (Custom only).
+        weight_col_name (str): Name of the column with weights in definition_data (Custom only).
+        score_col_name (str): Name for the calculated score column (Custom only).
+        mutual_exclusion_rules (list of tuple, optional): List of mutually exclusive category pairs (Custom only).
         return_categories (bool): Whether to return category indicators.
 
     Returns:
         - DataFrame with [id_col, score].
         - DataFrame with category indicators if return_categories is True, else None.
     """
+    # apply ICD modification if requested
+    if icd_modification is not None:
+        if icd_modification.lower() == "icd10gm":
+            df = get_icd10gm(
+                data=df,
+                code_col=code_col,
+                year_col=year_col,
+                target_year=icd_modification_target_year,
+            )
+            code_col = f"icd10gm_{icd_modification_target_year}"
+        else:
+            raise ValueError(f"Unknown icd_modification: '{icd_modification}'. Currently, only 'icd10gm' is supported.") # fmt: skip
+
+    # calculate scores
     score = score.lower()
     if score in (
         "cci",
@@ -49,9 +82,7 @@ def comorbidity(
         "charlson_comorbidity_index",
     ):
         if age_col not in df.columns:
-            raise ValueError(
-                f"Column '{age_col}' (age) must be present in input DataFrame for Charlson calculation."
-            )
+            raise ValueError(f"Column '{age_col}' (age) must be present in input DataFrame for Charlson calculation.") # fmt: skip
         return CharlsonComorbidityIndex(
             df=df,
             id_col=id_col,
@@ -108,8 +139,25 @@ def comorbidity(
             icd_version=icd_version,
             return_categories=return_categories,
         )
+    elif score in (
+        "custom",
+        "customcomorbidityindex",
+        "custom_comorbidity_index",
+    ):
+        return CustomComorbidityIndex(
+            df=df,
+            id_col=id_col,
+            code_col=code_col,
+            icd_version=icd_version,
+            icd_version_col=icd_version_col,
+            definition_data=definition_data or definition_file_path,
+            weight_col_name=weight_col_name,
+            score_col_name=score_col_name,
+            mutual_exclusion_rules=mutual_exclusion_rules,
+            return_categories=return_categories,
+        )
     else:
         raise ValueError(
             f"Unknown score: '{score}'. Must be one of: "
-            "'charlson', 'elixhauser', 'gagne', 'hfrs'."
+            "'charlson', 'elixhauser', 'gagne', 'hfrs' or 'custom'."
         )

pycomorb-0.5.1/src/pycomorb/common/CHARLSON_RCS.csv

@@ -0,0 +1,16 @@
+index,category,icd9_codes,icd10_codes,weights
+0,Age,,YYYY|YYYY,0
+1,Myocardial infarction,,I21|I22|I23|I252,1
+2,Congestive heart failure,,I11|I13|I255|I42|I43|I50|I517,1
+3,Peripheral vascular disease,,I70|I71|I72|I73|I770|I771|K551|K558|K559|R02|Z958|Z959,1
+4,Cerebrovascular disease,,G45|G46|I60|I61|I62|I63|I65|I66|I67|I68|I69,1
+5,Dementia,,A810|F00|F01|F02|F03|F051|G30|G31,1
+6,Chronic pulmonary disease,,I26|I27|J40|J41|J42|J44|J43|J45|J46|J47|J60|J61|J62|J63|J66|J64|J65|J66|J67|J684|J701|J703,1
+7,Rheumatological disease,,M05|M06|M09|M120|M315|M32|M33|M34|M35|M36,1
+8,Liver disease,,B18|I85|I864|I982|K70|K71|K721|K729|K76|R162|Z944,1
+9,Diabetes,,E10|E11|E12|E13|E14,1
+10,Hemiplegia or paraplegia,,G114|G81|G82|G83,1
+11,Renal disease,,I12|I13|N01|N03|N05|N07|N08|N171|N172|N18|N19|N25|Z49|Z940|Z992,1
+12,Any malignancy,,C0|C1|C20|C21|C22|C23|C24|C25|C26|C30|C31|C32|C33|C34|C37|C38|C39|C40|C41|C43|C45|C46|C47|C48|C49|C50|C51|C52|C53|C54|C55|C56|C57|C58|C60|C61|C62|C63|C64|C65|C66|C67|C68|C69|C70|C71|C72|C73|C74|C75|C76|C80|C81|C82|C83|C84|C85|C88|C90|C91|C92|C93|C94|C95|C96|C97,1
+13,Metastatic solid tumor,,C77|C78|C79,1
+14,AIDS/HIV,,B20|B21|B22|B23|B24,1

pycomorb-0.5.1/src/pycomorb/germanmodification/germanmodification.py

@@ -0,0 +1,87 @@
+# 1. download all the relevant ICD-10-GM versions from https://www.bfarm.de/DE/Kodiersysteme/Services/Downloads/_node.html
+# 2. unzip them into a folder, e.g. data/icd10gm
+# 3. take only the transfer files (e.g. icd10gm2013syst_umsteiger_2012_2013.txt for 2012->2013) and put them into a separate folder, e.g. data/icd10gm/transfers
+# 4. run this script in that folder
+
+import os
+import polars as pl
+
+files = [
+    "umsteiger20042005.txt",  # 2004->2005, renamed from 'umsteiger.txt' for clarity
+    "umsteiger20052006.txt",  # 2005->2006, renamed from 'umsteiger.txt' for clarity
+    "umsteiger20062007.txt",  # 2006->2007, renamed from 'Umsteiger.txt' for clarity
+    "umsteiger20072008.txt",  # 2007->2008
+    "umsteiger_icd10gmsyst2008_icd10gmsyst2009.txt",  # 2008->2009
+    "umsteiger_icd10gmsyst2009_icd10gmsyst2010.txt",  # 2009->2010
+    "umsteiger_icd10gmsyst2010_icd10gmsyst2011.txt",  # 2010->2011
+    "umsteiger_icd10gmsyst2011_icd10gmsyst2012.txt",  # 2011->2012
+    "icd10gm2013syst_umsteiger_2012_2013.txt",  # 2012->2013
+    "icd10gm2014syst_umsteiger_2013_2014.txt",  # 2013->2014
+    "icd10gm2015syst_umsteiger_2014_2015.txt",  # 2014->2015
+    "icd10gm2016syst_umsteiger_2015_2016.txt",  # 2015->2016
+    "icd10gm2017syst_umsteiger_2016_2017.txt",  # 2016->2017
+    "icd10gm2018syst_umsteiger_2017_2018.txt",  # 2017->2018
+    "icd10gm2019syst_umsteiger_2018_2019.txt",  # 2018->2019
+    "icd10gm2020syst_umsteiger_2019_2020.txt",  # 2019->2020
+    "icd10gm2021syst_umsteiger_2020_2021.txt",  # 2020->2021
+    "icd10gm2022syst_umsteiger_2021_2022.txt",  # 2021->2022
+    "icd10gm2023syst_umsteiger_2022_2023_20221206.txt",  # 2022->2023
+    "icd10gm2024syst_umsteiger_2023_20221206_2024.txt",  # 2023->2024
+    "icd10gm2025syst_umsteiger_2024_2025.txt",  # 2024->2025
+]
+years = range(2004, 2025)
+
+df = None
+for year, file in zip(years, files):
+    print(f"Processing {year}->{year+1}: {file}")
+    if df is None:
+        df = pl.DataFrame(schema={str(year): pl.Utf8})
+
+    year_df = (
+        pl.read_csv(
+            file,
+            separator=";",
+            has_header=False,
+            null_values=["", " "],  # , "UNDEF"],
+            schema={
+                str(year): pl.Utf8,  # Alte_Kodiernummer
+                str(year + 1): pl.Utf8,  # Neue_Kodiernummer
+                "A1": pl.Utf8,
+                "A2": pl.Utf8,
+            },
+        )
+        .drop("A1", "A2")
+        # Clean up codes: remove dots, uppercase
+        .with_columns(
+            pl.col(str(year))
+            .str.replace_all(r"\.", "")
+            .replace("UNDEF", f"UNDEF{year}")
+            .str.to_uppercase(),
+            pl.col(str(year + 1))
+            .str.replace_all(r"\.", "")
+            .replace("UNDEF", f"UNDEF{year}")
+            .str.to_uppercase(),
+        )
+        .filter(pl.col(str(year)).ne_missing(pl.col(str(year + 1))))
+        .unique()
+    )
+    df = df.join(year_df, on=str(year), how="full", coalesce=True)
+    
+# Propagate codes forward to fill in unmapped codes
+for year in years:
+    df = df.with_columns(
+        pl.when(pl.col(str(year + 1)).is_null())
+        .then(pl.col(str(year)))
+        .otherwise(pl.col(str(year + 1)))
+        .alias(str(year + 1))
+    )
+# Propagate codes backward to fill in unmapped codes
+for year in reversed(years):
+    df = df.with_columns(
+        pl.when(pl.col(str(year)).is_null())
+        .then(pl.col(str(year + 1)))
+        .otherwise(pl.col(str(year)))
+        .alias(str(year))
+    )
+
+df.write_csv("icd10gm.csv")

{pycomorb-0.4.2 → pycomorb-0.5.1/src/pycomorb.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: pycomorb
-Version: 0.4.2
+Version: 0.5.1
 Summary: Comorbidity and frailty score calculation utilities (Charlson, Elixhauser, Gagne, HFRS) using polars
 Author-email: Finn Fassbender <finn.fassbender@charite.de>
 License-Expression: MIT
@@ -23,6 +23,8 @@ Python package to calculate comorbidity and frailty scores using ICD codes and p
 
 It is inspired by the R package [`comorbidity`](https://github.com/ellessenne/comorbidity) by Alessandro Gasparini and the corresponding Python rewrite [`comorbidipy`](https://github.com/vvcb/comorbidipy) by Vishnu V Chandrabalan.
 
+As a special feature, it also implements the history of the German Modification of the ICD-10 (ICD-10-GM) coding system, allowing users to apply the correct mappings and weights based on the year of diagnosis.
+
 ## Included Indices / Scores
 
 - **Charlson Comorbidity Index** (multiple mappings and weights)
@@ -108,3 +110,4 @@ See the docstrings in each module for details on arguments and supported variant
 
 - Free software: MIT license
 - Documentation: (TODO)
+- _Die Erstellung erfolgt unter Verwendung der maschinenlesbaren Fassung des Bundesinstituts für Arzneimittel und Medizinprodukte (BfArM)._

{pycomorb-0.4.2 → pycomorb-0.5.1}/src/pycomorb.egg-info/SOURCES.txt

@@ -6,6 +6,7 @@ src/pycomorb/CustomComorbidityIndex.py
 src/pycomorb/ElixhauserComorbidityIndex.py
 src/pycomorb/GagneComorbidityIndex.py
 src/pycomorb/HospitalFrailtyRiskScore.py
+src/pycomorb/ICDModifications.py
 src/pycomorb/__init__.py
 src/pycomorb.egg-info/PKG-INFO
 src/pycomorb.egg-info/SOURCES.txt
@@ -16,6 +17,7 @@ src/pycomorb/common/CHARLSON_AUSTRALIA.csv
 src/pycomorb/common/CHARLSON_DEYO.csv
 src/pycomorb/common/CHARLSON_DHOORE.csv
 src/pycomorb/common/CHARLSON_QUAN.csv
+src/pycomorb/common/CHARLSON_RCS.csv
 src/pycomorb/common/CHARLSON_ROMANO.csv
 src/pycomorb/common/CHARLSON_SWEDEN.csv
 src/pycomorb/common/CHARLSON_UK_SHMI_v1.55.csv
@@ -25,4 +27,5 @@ src/pycomorb/common/ELIXHAUSER_AHRQ_ICD9.csv
 src/pycomorb/common/ELIXHAUSER_QUAN.csv
 src/pycomorb/common/ELIXHAUSER_WEIGHTS.csv
 src/pycomorb/common/GAGNE.csv
-src/pycomorb/common/HFRS.csv
+src/pycomorb/common/HFRS.csv
+src/pycomorb/germanmodification/germanmodification.py