eds_scikit.event.diabetes

DEFAULT_DIABETE_FROM_ICD10_CONFIG module-attribute

DEFAULT_DIABETE_FROM_ICD10_CONFIG = dict(
    codes=dict(
        DIABETES_TYPE_I=dict(prefix="E10"),
        DIABETES_TYPE_II=dict(prefix="E11"),
        DIABETES_MALNUTRITION=dict(prefix="E12"),
        DIABETES_IN_PREGNANCY=dict(prefix="O24"),
        OTHER_DIABETES_MELLITUS=dict(prefix=["E13", "E14"]),
        DIABETES_INSIPIDUS=dict(exact=["E232", "N251"]),
    ),
    date_from_visit=True,
    additional_filtering=dict(
        condition_status_source_value={"DP", "DAS"}
    ),
)

Default parameters feeded to conditions_from_icd10()

diabetes_from_icd10

diabetes_from_icd10(
    condition_occurrence: DataFrame,
    visit_occurrence: DataFrame,
    date_min: Optional[datetime] = None,
    date_max: Optional[datetime] = None,
    codes: Dict[
        str, Union[str, List[str]]
    ] = DEFAULT_DIABETE_FROM_ICD10_CONFIG["codes"],
    date_from_visit: bool = DEFAULT_DIABETE_FROM_ICD10_CONFIG[
        "date_from_visit"
    ],
    additional_filtering: Dict[
        str, Any
    ] = DEFAULT_DIABETE_FROM_ICD10_CONFIG[
        "additional_filtering"
    ],
) -> DataFrame

Wrapper around the conditions_from_icd10() function. Check the default configuration to see the used parameters

PARAMETER	DESCRIPTION
condition_occurrence	OMOP-like condition occurrence DataFrame TYPE: DataFrame
visit_occurrence	OMOP-like visit_occurrence DataFrame TYPE: Optional[DataFrame]
date_min	Lower temporal bound TYPE: Optional[datetime] DEFAULT: None
date_max	Upper temporal bound TYPE: Optional[datetime] DEFAULT: None
codes	Dictionary of ICD-10 used for phenotyping TYPE: Optional[Dict[str, Union[str, List[str]]]] DEFAULT: DEFAULT_DIABETE_FROM_ICD10_CONFIG['codes']
date_from_visit	If true, use the visit_[start/end]_datetime for filtering. Else, use condition_start_datetime TYPE: bool, by default True DEFAULT: DEFAULT_DIABETE_FROM_ICD10_CONFIG['date_from_visit']
additional_filtering	A dictionary to perform additional filtering. Each key should be a valid column name from condition_occurrence Each value should be a value / set of values / list of values For each pair (key, value), filtering is done as condition_occurrence[condition_occurrence[k].isin(v)] TYPE: Dict[str, Any] DEFAULT: DEFAULT_DIABETE_FROM_ICD10_CONFIG['additional_filtering']

RETURNS	DESCRIPTION
DataFrame	Event DataFrame in long format (with a concept and a value column). The concept column contains one of the following: DIABETES_TYPE_I DIABETES_TYPE_II DIABETES_MALNUTRITION DIABETES_IN_PREGNANCY OTHER_DIABETES_MELLITUS DIABETES_INSIPIDUS The value column contains the corresponding ICD-10 code that was extracted

Source code in eds_scikit/event/diabetes.py

def diabetes_from_icd10(
    condition_occurrence: DataFrame,
    visit_occurrence: DataFrame,
    date_min: Optional[datetime] = None,
    date_max: Optional[datetime] = None,
    codes: Dict[str, Union[str, List[str]]] = DEFAULT_DIABETE_FROM_ICD10_CONFIG[
        "codes"
    ],
    date_from_visit: bool = DEFAULT_DIABETE_FROM_ICD10_CONFIG["date_from_visit"],
    additional_filtering: Dict[str, Any] = DEFAULT_DIABETE_FROM_ICD10_CONFIG[
        "additional_filtering"
    ],
) -> DataFrame:
    """
    Wrapper around the [conditions_from_icd10()][eds_scikit.event.icd10.conditions_from_icd10] function.
    Check the [default configuration][eds_scikit.event.diabetes.DEFAULT_DIABETE_FROM_ICD10_CONFIG] to see
    the used parameters

    Parameters
    ----------
    condition_occurrence
        OMOP-like condition occurrence DataFrame
    visit_occurrence : Optional[DataFrame]
        OMOP-like visit_occurrence DataFrame
    date_min : Optional[datetime]
        Lower temporal bound
    date_max : Optional[datetime]
        Upper temporal bound
    codes : Optional[Dict[str, Union[str, List[str]]]]
        Dictionary of ICD-10 used for phenotyping
    date_from_visit : bool, by default True
        If true, use the `visit_[start/end]_datetime` for filtering. Else, use `condition_start_datetime`
    additional_filtering : Dict[str, Any]
        A dictionary to perform additional filtering.

        - **Each key** should be a valid column name from `condition_occurrence`
        - **Each value** should be a value / set of values / list of values
        For each pair (key, value), filtering is done as `condition_occurrence[condition_occurrence[k].isin(v)]`

    Returns
    -------
    DataFrame
        Event DataFrame in **long** format (with a `concept` and a `value` column).
        The `concept` column contains one of the following:

        - DIABETES_TYPE_I
        - DIABETES_TYPE_II
        - DIABETES_MALNUTRITION
        - DIABETES_IN_PREGNANCY
        - OTHER_DIABETES_MELLITUS
        - DIABETES_INSIPIDUS
        The `value` column contains the corresponding ICD-10 code that was extracted
    """
    diabetes = conditions_from_icd10(
        condition_occurrence=condition_occurrence,
        visit_occurrence=visit_occurrence,
        date_min=date_min,
        date_max=date_max,
        codes=codes,
        date_from_visit=date_from_visit,
        additional_filtering=additional_filtering,
    )

    diabetes["value"] = diabetes["concept"]
    diabetes["concept"] = "DIABETES_FROM_ICD10"

    return diabetes