validation/parameter_calculation/Time_to_first_event_calculations.py

# Import libraries
import functools as ft
import numpy as np
import pandas as pd

# Set file paths
file_path = '<YOUR_DATA_PATH>/'
input_file_path = file_path + 'data_for_model_e_columns/'


def read_data(file):
    """
    Read in data source
    --------
    :param file: string filename
    :return: dataframe
    """
    df = pd.read_csv(file)
    return df


def format_data(exacerbations_data, admissions_data, onboard, IDs):
    """
    Remove unescessary columns from dataframes,
    merge onboarding, admissions, and exacerbations dataframes,
    convert datetime columns to datetime format,
    filter to include only RECEIVER and scale up 1 IDs,
    and create new column showing date of death for those who died during the study
    --------
    :param exacerbations_data: dataframe containing exacerbations data
    :param admissions_data: dataframe containing admissions data
    :param IDs: dataframe containing RECEIVER and scale up 1 study IDs
    :param onboard: dataframe containing onboarding dates
    :return: formatted dataframe
    """
    admissions_data = admissions_data[['Study_ID', 'admitted_1']]
    exacerbations_data = exacerbations_data[['Study_ID', 'first_exacerbation']]

    dfs = [onboard, exacerbations_data, admissions_data]
    df_combined = ft.reduce(lambda left, right: pd.merge(left, right, on='Study_ID', how="outer"), dfs)
    data = pd.merge(IDs, df_combined, on="Study_ID", how="left")
    
    data['first_exacerbation'] = pd.to_datetime(data['first_exacerbation'])
    data['admitted_1'] = pd.to_datetime(data['admitted_1'])
    data['OB_date'] = pd.to_datetime(data['OB_date'])
    data['censor'] = pd.to_datetime(data['censor'])
    
    conditions_DOD = [data['censor'] != '2021-08-31']
    values_DOD = [data['censor'].dt.date]
    data['DOD'] = np.select(conditions_DOD, values_DOD, default=None)
    data['DOD'] = pd.to_datetime(data['DOD'])
    return data


def time_to_events(data):
    """
    Calculate time to first event (exacerbation, admission, or death) and first admission or death
    for each study ID and save the summary dataframe
    --------
    :param data: dataframe containing admissions data, exacerbations data, and onboarding dates
    :return: dataframe with additional columns showing number of days until first event and number of days
    to first admission/ death
    """
    data['first_event'] = data[["admitted_1", "first_exacerbation", "DOD"]].min(axis=1)
    data['first_event'] = pd.to_datetime(data['first_event'])
    data['first_admission_or_death'] = data[["admitted_1", "DOD"]].min(axis=1)
    data['first_admission_or_death'] = pd.to_datetime(data['first_admission_or_death'])

    data['days_to_first_event'] = (data['first_event'] - data['OB_date']).dt.days
    data['days_to_first_admission_death'] = (data['first_admission_or_death'] - data['OB_date']).dt.days
    
    data.to_csv(file_path + 'Time_to_first_event.csv')


def main():
    # Read data
    PRO_LOGIC_data = input_file_path + "First_exacerbation_data.csv"
    admissions_data_file = input_file_path + "admissions_data_up_to_31082021.csv"
    RC_SU1_IDs_data_file = input_file_path + "RC_SU1_IDs.csv"
    onboard_file = input_file_path + "onboarding_dates.csv"

    PRO_LOGIC_data = read_data(PRO_LOGIC_data)
    admissions_data = read_data(admissions_data_file)
    RC_SU1_IDs = read_data(RC_SU1_IDs_data_file)
    Onboard = read_data(onboard_file)

    # Format data
    RC_combined_data = format_data(PRO_LOGIC_data, admissions_data, Onboard, RC_SU1_IDs)
    
    # Calculate time to first event for each study ID and save the summary dataframe
    time_to_events(RC_combined_data)


main()