# Import libraries import pandas as pd import numpy as np # Set file paths input_file_path = '/EXAMPLE_STUDY_DATA/' output_file_path = '/summary_files/' copd = 'CHRONIC OBSTRUCTIVE PULMONARY DISEASE' resp = ['PNEUMONITIS DUE TO FOOD AND VOMIT', 'RESPIRATORY FAILURE, UNSPECIFIED; TYPE UNSPECIFIED', 'CHRONIC RESPIRATORY FAILURE; TYPE II [HYPERCAPNIC]', 'BRONCHOPNEUMONIA, UNSPECIFIED', 'DYSPNOEA', 'PLEURAL EFFUSION IN CONDITIONS CLASSIFIED ELSEWHERE', 'RESPIRATORY FAILURE, UNSPECIFIED; TYPE [HYPERCAPNIC]', 'PLEURAL EFFUSION, NOT ELSEWHERE CLASSIFIED', 'CHRONIC RESPIRATORY FAILURE', 'OTHER BACTERIAL PNEUMONIA', 'ABN MICROBIOLOGICAL FINDINGS IN SPECS FROM RESPIRATORY ORGANS AND THORAX', 'RESPIRATORY FAILURE, UNSPECIFIED', 'PNEUMONIA, UNSPECIFIED', 'LOBAR PNEUMONIA, UNSPECIFIED', 'COUGH', 'PLEURAL PLAQUE WITH PRESENCE OF ASBESTOS', 'PLEURAL PLAQUE WITHOUT ASBESTOS', 'OTHER DISORDERS OF LUNG', 'OTHER SPECIFIED PLEURAL CONDITIONS', 'PULMONARY COLLAPSE', 'ACQUIRED ABSENCE OF LUNG [PART OF]', 'ASPHYXIATION', 'RESPIRATORY FAILURE, UNSPECIFIED; TYPE [HYPOXIC]', 'TRACHEOSTOMY STATUS', 'ACUTE RESPIRATORY FAILURE', 'UNSPECIFIED ACUTE LOWER RESPIRATORY INFECTION', 'OTHER SPECIFIED SYMPTOMS AND SIGNS INVOLVING THE CIRC AND RESP SYSTEMS', 'BACTERIAL PNEUMONIA, UNSPECIFIED', 'PYOTHORAX WITHOUT FISTULA', 'DISEASES OF BRONCHUS, NOT ELSEWHERE CLASSIFIED', 'PNEUMONIA DUE TO HAEMOPHILUS INFLUENZAE', 'ABNORMAL SPUTUM', 'OTHER POSTPROCEDURAL RESPIRATORY DISORDERS', 'OTHER AND UNSPECIFIED ABNORMALITIES OF BREATHING', 'INFLUENZA WITH OTHER RESP MANIFESTATIONS, SEASONAL INFLUENZA VIRUS IDENTIF', 'PERSONAL HISTORY OF DISEASES OF THE RESPIRATORY SYSTEM', 'PNEUMONIA DUE TO STREPTOCOCCUS PNEUMONIAE', 'WHEEZING', 'CHEST PAIN ON BREATHING', 'HAEMOPTYSIS', 'INFLUENZA WITH OTHER MANIFESTATIONS, VIRUS NOT IDENTIFIED', 'OTHER SPECIFIED RESPIRATORY DISORDERS', 'ACUTE UPPER RESPIRATORY INFECTION, UNSPECIFIED', 'T.B. OF LUNG, W/O MENTION OF BACTERIOLOGICAL OR HISTOLOGICAL CONFIRMATION', 'DEPENDENCE ON RESPIRATOR', 'PLEURISY', 'BRONCHITIS, NOT SPECIFIED AS ACUTE OR CHRONIC'] def read_data(file, cols, types): """ Read in data source -------- :param file: string filename :param cols: string list of column names :param types: string list of column types :return: dataframe """ schema = dict(zip(cols, types)) df = pd.read_csv(file, usecols=cols, encoding="cp1252", dtype=schema) return df def update_null_stay(df): """ Calculate the values for any null 'STAY' values using the admission and discharge dates. -------- df : pandas dataframe to be updated """ is_null = df['STAY'].isnull() if sum(is_null) > 0: null_stay = np.where(is_null) for i in null_stay: stay = df.loc[i, 'DISDATE'].item() - df.loc[i, 'ADMDATE'].item() df.loc[i, 'STAY'] = float(stay.days) return df def calculate_total_stay(df): """ Model A: Calculate the cumulative (total) length of stay, given data already grouped by patient ID and sorted by admission date then discharge date. It sums all stays for which the admission date matches the previous discharge date, sets the admission date to the first admission and drops all rows except the final (or only if the patient was not transferred) record for any given stay. Works for any number of transfers. Also adds a 'transfer' column to the existing data (True/False) df : pandas dataframe dataframe to be updated """ df.reset_index(inplace=True, drop=True) rows_to_drop = [] df['transfer'] = df.ADMDATE.eq(df.DISDATE.shift()) for index, row in df.iloc[1:].iterrows(): if row.transfer is True: df.loc[index, 'ADMDATE'] = df.iloc[index - 1].ADMDATE df.loc[index, 'STAY'] = row.STAY + df.iloc[index - 1].STAY rows_to_drop.append(index - 1) df.drop(rows_to_drop, inplace=True) df.drop('transfer', axis=1, inplace=True) return df def track_copd_resp(df, track_type='both'): """ Search for COPD and/or respiratory admissions -------- df : pandas dataframe dataframe to be updated track_type : str 'copd', 'resp' or 'both' """ diag_columns = ['DIAG1Desc', 'DIAG2Desc', 'DIAG3Desc', 'DIAG4Desc', 'DIAG5Desc', 'DIAG6Desc'] df_diag = df[diag_columns] if track_type in ['copd', 'both']: copd_event = df_diag.apply(lambda x: track_feature(x, copd, True)) copd_event = copd_event.any(axis=1).astype(int) df['copd_event'] = copd_event if track_type in ['resp', 'both']: resp_event = df_diag.apply(lambda x: track_feature(x, resp, False)) resp_event = resp_event.any(axis=1).astype(int) df['resp_event'] = resp_event return df def track_feature(x, desc, single): """ Fill nulls and search to see if x matches a description ------- x : str list feature to track desc : str list string list to compare single : boolean if checking against single description e.g. "COPD" True otherwise False """ x = x.fillna('') if single: result = [desc in s for s in x] else: result = [s in desc for s in x] return result def filter_data(data, date): """ Filter data to only include copd or resp admission events occurring after the index date -------- :param data: dataframe :param date: index date :return: filtered dataframe """ data['ADMDATE'] = pd.to_datetime(data['ADMDATE']) data = data[data['ADMDATE'] >= date] data = data[(data['copd_event'] == 1) | (data['resp_event'] == 1)] return data def calculate_time_to_first_copd_admission(data, date): """ Calculate days to first COPD admission -------- :param data: dataframe :param date: Index date in 'DD-MM-YYYY' format :return: dataframe showing the number of days to the first COPD admission event for each ID since the index date """ copd_data = data[data['copd_event'] == 1] first_copd_admission = copd_data.groupby('SafeHavenID').agg(first_copd_admission=('ADMDATE', np.min)) first_copd_admission['index_date'] = date first_copd_admission['index_date'] = pd.to_datetime(first_copd_admission['index_date']) first_copd_admission['days_to_first_copd_admission'] = (first_copd_admission['first_copd_admission'] - first_copd_admission['index_date']).dt.days return first_copd_admission def calculate_time_to_first_resp_admission(data, date): """ Calculate days to first resp admission -------- :param data: dataframe :param date: Index date in 'DD-MM-YYYY' format :return: dataframe showing the number of days to the first resp admission event for each ID since the index date """ resp_data = data[data['resp_event'] == 1] first_resp_admission = resp_data.groupby('SafeHavenID').agg(first_resp_admission=('ADMDATE', np.min)) first_resp_admission['index_date'] = date first_resp_admission['index_date'] = pd.to_datetime(first_resp_admission['index_date']) first_resp_admission['days_to_first_resp_admission'] = (first_resp_admission['first_resp_admission'] - first_resp_admission['index_date']).dt.days return first_resp_admission def calculate_time_to_first_copd_or_resp_admission(data, date): """ Calculate days to first copd or resp admission -------- :param data: dataframe :param date: Index date in 'DD-MM-YYYY' format :return: dataframe showing the number of days to the first COPD or resp admission event for each ID since the index date """ data['copd_or_resp_event'] = (data['resp_event'] | data['copd_event']) resp_copd_data = data[(data['copd_or_resp_event'] == 1)] first_resp_or_copd_admission = resp_copd_data.groupby('SafeHavenID').agg(first_copd_or_resp_admission=('ADMDATE', np.min)) first_resp_or_copd_admission['index_date'] = date first_resp_or_copd_admission['index_date'] = pd.to_datetime(first_resp_or_copd_admission['index_date']) first_resp_or_copd_admission['first_copd_or_resp_admission'] = pd.to_datetime(first_resp_or_copd_admission['first_copd_or_resp_admission']) first_resp_or_copd_admission['days_to_first_copd_or_resp_admission'] = (first_resp_or_copd_admission['first_copd_or_resp_admission'] - first_resp_or_copd_admission['index_date']).dt.days return first_resp_or_copd_admission def calculate_ad_count_1_year(data, year_censor, first_admission_df, adm_col): """ Calculate the number of COPD or respiratory admissions in the year following the index date and join this data to the time to first admissions data for each ID -------- :param data: dataframe containing admissions dates :param year_censor: date 1 year following Index date 'DD-MM-YYYY' format :param first_admission_df: dataframe showing days to first admission :param adm_col: binary column showing if an admission was copd or respiratory related or not :return: dataframe showing the number of days to the first COPD or resp admission event for each ID since the index date """ admission_year = data[data['ADMDATE'] < year_censor] year_admission_count = admission_year.groupby('SafeHavenID').agg(admission_count_year_post_index=(adm_col, 'sum')) all_admissions_data = pd.merge(year_admission_count, first_admission_df, on="SafeHavenID", how="outer") all_admissions_data['admission_count_year_post_index'] = all_admissions_data['admission_count_year_post_index'].fillna(0) return all_admissions_data def main(): adm_file = input_file_path + "SMR01_Cohort3R.csv" adm_cols = ['SafeHavenID', 'ETHGRP', 'ADMDATE', 'DISDATE', 'DIAG1Desc', 'DIAG2Desc', 'DIAG3Desc', 'DIAG4Desc', 'DIAG5Desc', 'DIAG6Desc', 'STAY'] adm_types = ['int', 'object', 'object', 'object', 'str', 'str', 'str', 'str', 'str', 'str', 'int'] adm = read_data(adm_file, adm_cols, adm_types) # Nulls dropped later in process, only drop duplicates adm = adm.drop_duplicates() # Convert date columns to correct type adm['ADMDATE'] = pd.to_datetime(adm['ADMDATE']) adm['DISDATE'] = pd.to_datetime(adm['DISDATE']) # Update any null STAY data using ADM and DIS dates adm = update_null_stay(adm) # Correct stays for patients passed across departments adm = adm.sort_values(['SafeHavenID', 'ADMDATE', 'DISDATE']) adm = adm.groupby('SafeHavenID').apply(calculate_total_stay) adm = adm.reset_index(drop=True) # Prepare text data - strip string columns adm = adm.apply(lambda x: x.str.strip() if x.dtype == 'object' else x) # Track COPD and respiratory events adm = track_copd_resp(adm) # Filter to only include copd or resp admission events occuring after index adm = filter_data(adm, '01-01-2020') # Calculate time to first respiratory and COPD admissions first_copd_admission = calculate_time_to_first_copd_admission(adm, '01-01-2020') first_resp_admission = calculate_time_to_first_resp_admission(adm, '01-01-2020') first_resp_or_copd_admission = calculate_time_to_first_copd_or_resp_admission(adm, '01-01-2020') # Calculate number of respiratory and COPD admissions in the follow up year and join this to the time to admission data first_copd_admission = calculate_ad_count_1_year(adm, '01-01-2021', first_copd_admission, 'copd_event') first_resp_admission = calculate_ad_count_1_year(adm, '01-01-2021', first_resp_admission, 'resp_event') first_resp_or_copd_admission = calculate_ad_count_1_year(adm, '01-01-2021', first_resp_or_copd_admission, 'copd_or_resp_event') # Save data adm.to_pickle(output_file_path + 'all_COPD_and_resp_admissions_from_index_date.pkl') first_copd_admission.to_pickle(output_file_path + 'copd_admissions_cohort_summary.pkl') first_resp_admission.to_pickle(output_file_path + 'resp_admissions_cohort_summary.pkl') first_resp_or_copd_admission.to_pickle(output_file_path + 'copd_or_resp_admissions_cohort_summary.pkl') main()