python-icd10 1.0.2

Downloads ICD10 codes from the CDC and makes them available in a searchable database

Python ICD10

Python-icd10 can be used to download icd10 diagnoses codes from the US CDC website. The data is reformatted and validated using Pydantic, to ensure all notes and additional information is captured.

Once downloaded, the data can be accessed to locate codes via their descriptions, via their partial code, or you can pull specific records using the exact icd10 code.

Installing Python-icd10

To install Python-icd10 from PyPI, run

pip install python_icd10

Basic Usage

Python-icd10 comes in two main classes. One to download and build the database, and one used to read/extract data from it.

Creating/Updating the ICD10 database

import pathlib
from python_icd10.icd10_load import ICD10Load

ICD10Load(pathlib.Path("/path/tp/save/db/")).load_from_cdc()

The above code will download the ICD10 codes from the CDC and create the database. The path does not need to exist before running this command. The code will create directories etc for you, as long as it has access to do so. This class will examine the CDC website and download the most recent file, based on the year. It will not take any notice of minor update files though.

Using the database

import pathlib
from python_icd10.icd10_warehouse import ICD10Warehouse

icd10 = ICD10Warehouse(pathlib.Path("/path/to/the/database/"))

This will create an object with access to the database.

A single record can be fetched by find a record by its direct ICD10 code.

record = icd10.get_by_code("M30")

It should be noted that the returned value is a TinyDB Document object. This is a child of a Dictionary with some additions such as obj.doc_id

Finding records via the diagnosis description can be done like this:

records = icd10.search_by_descriptipn("arthritis")

The search is case insensitive and will return a list of Document objects

You can also search by partial icd10 code

records = icd10.search_by_code("M30")

All ICD10 codes can be returned in a list of strings using:

codes = icd10.get_all_codes()

Directly accessing the database

The TinyDB database is accessable via this object, so you can run custom queries directlt against the database

import pathlib
from tinydb import Query
from python_icd10.icd10_warehouse import ICD10Warehouse

icd10 = ICD10Warehouse(pathlib.Path("/path/to/db/"))
q = Query()

recs = icd10.db.search(q.section_name.matches("*hand*"))

See TinyDB for details on creating your own queries.

Record format

Data from the CDC comes as a multi level XML file with lots of embedded and additional data. Python-icd10 flattens this and makes it more accessible. Each record is shaped as follows

• name: ICD10 code as a string
• description: Description attached to the code
• inclusion_term: Notes on inclusion terms for the code. List of Strings
• excludes1: Exclusion details(1). List of Strings
• excludes2: Exclusion details(2). List of Strings
• includes: Inclusion notes. List of Strings
• code_first: List of Strings
• additional_codes: List of Strings
• parent: Parent code. String or None
• section_name: string
• section_description: String or None
• chapter: String

A lot of this data, I am unsure of its use, but it's included in the CDC data, so I have included it here