Pneumonia was defined as ≥1 inpatient or outpatient claims 1) with ICD-10 codes for pneumonia recorded as any diagnosis in inpatient claims or primary ~ 4th secondary diagnosis in outpatient claims, AND 2) with diagnostic test code for chest-X ray or chest- computed tomography (CT), AND 3) with antibiotics prescription during the following period after the index date. To assess the incidence of pneumonia event, the first pneumonia event observed during the follow-up period was considered.

Incidence rate for the first pneumonia event was calculated as below:

Incidence rate per 1000 person-years= Number of patients with event/ sum of event-free period during the follow-up period (1000 person-years)

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

Pneumonia was defined as ≥1 inpatient or outpatient claims 1) with ICD-10 codes for pneumonia recorded as any diagnosis in inpatient claims or primary ~ 4th secondary diagnosis in outpatient claims, AND 2) with diagnostic test code for chest-X ray or chest- computed tomography (CT), AND 3) with antibiotics prescription during the following period after the index date.

This outcome measure reports: Total number of pneumonia events/ sum of follow-up duration (person-years) for all participants

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

Pneumonia was defined as ≥1 inpatient or outpatient claims 1) with ICD-10 codes for pneumonia recorded as any diagnosis in inpatient claims or primary ~ 4th secondary diagnosis in outpatient claims, AND 2) with diagnostic test code for chest-X ray or chest- computed tomography (CT), AND 3) with antibiotics prescription during the following period after the index date. Date of pneumonia event was defined as the starting date of outpatient claim, or as the admission date of inpatient claim with the above 3 conditions to define pneumonia event.

Mean and Standard Deviation of time to first pneumonia event from the index date was estimated using the Kaplan-Meier curve.

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

Pneumonia was defined as ≥1 inpatient or outpatient claims 1) with ICD-10 codes for pneumonia recorded as any diagnosis in inpatient claims or primary ~ 4th secondary diagnosis in outpatient claims, AND 2) with diagnostic test code for chest-X ray or chest- computed tomography (CT), AND 3) with antibiotics prescription during the following period after the index date.

Number of patients with pneumonia event is reported.

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

To assess the incidence of COPD exacerbation event (all types), the first COPD exacerbation event among pre-defined moderate COPD exacerbation and severe COPD exacerbation event observed within the 12-month period after the index date was considered.

Moderate COPD exacerbation was defined as ≥ 1 outpatient claims 1) with ICD-10 codes for COPD recorded as primary ~ 4th secondary diagnosis, AND 2) with systemic steroids and/or antibiotics prescription within 12 months after the index date. Severe COPD exacerbation was defined as ≥ 1 Emergency Room (ER) visit or inpatient claims 1) with ICD-10 codes for COPD or diseases due to COPD worsening as any diagnosis, AND 2) with systemic steroids and/or antibiotics prescription.

Incidence rate of COPD exacerbation events (all types) was calculated as below: Incidence rate per 1000 person-years= Number of patients with event/sum of event-free period during the follow-up period (1000 person-years)

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

To assess the frequency of all types (moderate and severe) COPD exacerbation events, any events observed within the 12-month period after the index date was considered.

Moderate COPD exacerbation was defined as ≥ 1 outpatient claims 1) with ICD-10 codes for COPD recorded as primary ~ 4th secondary diagnosis, AND 2) with systemic steroids and/or antibiotics prescription within 12 months after the index date. Severe COPD exacerbation was defined as ≥ 1 Emergency Room (ER) visit or inpatient claims 1) with ICD-10 codes for COPD or diseases due to COPD worsening as any diagnosis, AND 2) with systemic steroids and/or antibiotics prescription.

Outcome measure reports: Total number of COPD exacerbation events (all types)/sum of follow-up duration (person-years) for all participants

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

Number of patients with Chronic Obstructive Pulmonary Disease (COPD) exacerbation (moderate or severe) events is reported.

Moderate COPD exacerbation was defined as ≥ 1 outpatient claims 1) with ICD-10 codes for COPD recorded as primary ~ 4th secondary diagnosis, AND 2) with systemic steroids and/or antibiotics prescription within 12 months after the index date. Severe COPD exacerbation was defined as ≥ 1 Emergency Room (ER) visit or inpatient claims 1) with ICD-10 codes for COPD or diseases due to COPD worsening as any diagnosis, AND 2) with systemic steroids and/or antibiotics prescription.

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

Mean and Standard Deviation of time to first COPD exacerbation (moderate or severe) event from the index date was estimated using the Kaplan-Meier curve.

Moderate COPD exacerbation was defined as ≥ 1 outpatient claims 1) with ICD-10 codes for COPD recorded as primary ~ 4th secondary diagnosis, AND 2) with systemic steroids and/or antibiotics prescription within 12 months after the index date. Severe COPD exacerbation was defined as ≥ 1 Emergency Room (ER) visit or inpatient claims 1) with ICD-10 codes for COPD or diseases due to COPD worsening as any diagnosis, AND 2) with systemic steroids and/or antibiotics prescription. Date of severe COPD exacerbation event was defined as the starting date of ER claims or admission date of inpatient claims with the above 2 conditions to define moderate COPD exacerbation event.

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

To assess the incidence rate of initiating triple combination therapy, the first triple combination therapy observed during the follow-up period was considered.

Triple combination therapy of inhaled corticosteroids (ICS), long-acting beta agonists (LABA) and long-acting muscarinic antagonists (LAMA) was defined as ≥ 1 outpatient or inpatient claims with a combination of ICS, LABA, and LAMA. Three components of ICS, LABA and LAMA MUST have been prescribed together in a same prescription issued on same date to be considered as triple combination therapy (i.e. ICS/LABA fixed dose combination (FDC) and LAMA, ICS and LABA/LAMA FDC, ICS/LABA FDC and LABA/LAMA FDC, ICS and LABA and LAMA are all considered as triple combination therapy).

Incidence rate of initiating triple combination therapy is reported as: Incidence rate per 1000 person-years= Number of patients with event/ sum of event-free period during the follow-up period (1000 person-years)

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

Number of patients who initiated a triple combination therapy is reported. Triple combination therapy of inhaled corticosteroids (ICS), long-acting beta agonists (LABA) and long-acting muscarinic antagonists (LAMA) was defined as ≥ 1 outpatient or inpatient claims with a combination of ICS, LABA, and LAMA. Three components of ICS, LABA and LAMA MUST have been prescribed together in a same prescription issued on same date to be considered as triple combination therapy (i.e. ICS/LABA fixed dose combination (FDC) and LAMA, ICS and LABA/LAMA FDC, ICS/LABA FDC and LABA/LAMA FDC, ICS and LABA and LAMA are all considered as triple combination therapy).

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

Median and Inter-Quartile Range of time to initiating triple combination therapy from the index date was estimated using the Kaplan-Meier curve.

Triple combination therapy of inhaled corticosteroids (ICS), long-acting beta agonists (LABA) and long-acting muscarinic antagonists (LAMA) was defined as ≥ 1 outpatient or inpatient claims with a combination of ICS, LABA, and LAMA. Three components of ICS, LABA and LAMA MUST have been prescribed together in a same prescription issued on same date to be considered as triple combination therapy (i.e. ICS/LABA fixed dose combination (FDC) and LAMA, ICS and LABA/LAMA FDC, ICS/LABA FDC and LABA/LAMA FDC, ICS and LABA and LAMA are all considered as triple combination therapy).

Date of triple combination therapy was defined as the starting date of outpatient claim or admission date of inpatient claim with a combination of ICS/LABA and LAMA.

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

Mean and Standard Deviation of any medical visit per person per month is reported. Medical visits were calculated as the sum of inpatient and outpatient visits assuming that emergency room (ER) visits, intensive care unit (ICU) visits and visit for pharmacy dispensation would be counted with the inclusion of inpatient and outpatient visit.

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

Mean and Standard Deviation of outpatient visits per person per month is reported. Outpatient visit was defined as clinic, hospital, or other medical institution (e.g. public health, etc.) visit as an outpatient.

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

Mean and Standard Deviation of inpatient visits per person per month is reported. Inpatient visit was defined as at least 1 overnight stay in a hospital as an inpatient.

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

Mean and Standard Deviation of ER visits per person per month is reported.

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

Mean and Standard Deviation of ICU visits per person per month is reported.

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

COPD-related HCRU was defined as medical claims with ICD-10 codes for COPD and/or COPD-related diseases. Mean and Standard Deviation of any medical visit (inpatient and outpatient visits which include emergency room (ER) and intensive care unit (ICU) visits) per person per month is reported. For outpatient claims, COPD-related HCRU was defined as claims with ICD-10 codes for COPD as primary ~ 4th secondary diagnosis. For inpatient claims, COPD-related HCRU was defined as claims with ICD-10 codes for COPD and/or for diseases due to COPD worsening as any diagnosis.

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

Mean and Standard Deviation of outpatient visits per person per month is reported. For outpatient claims, COPD-related HCRU was defined as claims with ICD-10 codes for COPD as primary ~ 4th secondary diagnosis.

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

Mean and Standard Deviation of inpatient visits per person per month is reported. For inpatient claims, COPD-related HCRU was defined as claims with ICD-10 codes for COPD and/or for diseases due to COPD worsening as any diagnosis.

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

Mean and Standard Deviation of ER visits per person per month is reported.

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

Mean and Standard Deviation of ICU visits per person per month is reported.

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

Mean and Standard Deviation of total medical costs per person per month is reported. Total medical costs were estimated as the sum of medical costs (inpatient + outpatient) and pharmacy dispensation costs (drug cost + pharmacy service cost). Costs were converted to US dollars using the average 2019 exchange rate: US$ = 1166.51 KRW (South Korean won) and were also adjusted to calendar year 2019 using the Consumer Price Index of South Korea.

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

Mean and Standard Deviation of hospitalization costs per person per month is reported. Hospitalization costs were estimated as the total costs of services from all inpatient claims (e.g. cost for procedure, diagnostic test, medication, patient meal, and bed stay, etc.). Costs were converted to US dollars using the average 2019 exchange rate: US$ = 1166.51 KRW (South Korean won) and were also adjusted to calendar year 2019 using the Consumer Price Index of South Korea.

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

Mean and Standard Deviation of outpatient medical costs per person per month is reported. Outpatient medical costs were estimated as the total costs of services from all outpatient claims (e.g. cost for procedure, diagnostic test, prescription, physician fee, etc.). Costs were converted to US dollars using the average 2019 exchange rate: US$ = 1166.51 KRW (South Korean won) and were also adjusted to calendar year 2019 using the Consumer Price Index of South Korea.

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

Mean and Standard Deviation of outpatient pharmacy dispensation costs per person per month is reported. Outpatient pharmacy dispensation costs were estimated as the total costs of services from all outpatient pharmacy claims (pharmacy dispensing service costs + medication costs). Costs were converted to US dollars using the average 2019 exchange rate: US$ = 1166.51 KRW (South Korean won) and were also adjusted to calendar year 2019 using the Consumer Price Index of South Korea.

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

Mean and Standard Deviation of COPD-related total medical costs per person per month is reported. COPD-related total medical costs were estimated as the sum of medical costs (inpatient + outpatient) of medical claims with ICD-10 codes for COPD and/or COPD-related diseases and pharmacy dispensation costs (drug costs + pharmacy service costs) of pharmacy claims including dispensation of ICS/LABA, LABA, LAMA, LABA/LAMA, short-acting beta-2 agonist (SABAs), short-acting muscarinic antagonist (SAMAs), SABA/SAMA, theophlliyne, or Phosphodiesterase-4 (PDE4) inhibitor. Costs were converted to US dollars using the average 2019 exchange rate: US$ = 1166.51 KRW (South Korean won) and were also adjusted to calendar year 2019 using the Consumer Price Index of South Korea.

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

Mean and Standard Deviation of COPD-related hospitalization costs per person per month is reported. COPD-related hospitalization costs were estimated as the total costs of services (e.g. costs for procedure, diagnostic test, medication, patient meal, and bed stay, etc.) from all inpatient claims (medical claims with ICD-10 codes for COPD and/or for diseases due to COPD worsening as any diagnosis).

Costs were converted to US dollars using the average 2019 exchange rate: US$ = 1166.51 KRW (South Korean won) and were also adjusted to calendar year 2019 using the Consumer Price Index of South Korea.

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

Mean and Standard Deviation of COPD-related outpatient medical costs per person per month is reported. COPD-related outpatient medical costs were estimated as the total costs of services (e.g. costs for procedure, diagnostic test, prescription, physician fee, etc.) from all outpatient claims (medical claims with ICD-10 codes for COPD as primary ~ 4th secondary diagnosis).

Costs were converted to US dollars using the average 2019 exchange rate: US$ = 1166.51 KRW (South Korean won) and were also adjusted to calendar year 2019 using the Consumer Price Index of South Korea.

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

Mean and Standard Deviation of COPD-related outpatient pharmacy dispensation costs per person per month is reported.

COPD related outpatient pharmacy dispensation costs were estimated as the total costs of services (pharmacy dispensing service costs + medication costs) from all outpatient pharmacy claims (pharmacy claims including dispensation of ICS/LABA, LABA, LAMA, LABA/LAMA, short-acting beta-2 agonist (SABAs), short-acting muscarinic antagonist (SAMAs), SABA/SAMA, theophlliyne, or Phosphodiesterase-4 (PDE4) inhibitor).

Costs were converted to US dollars using the average 2019 exchange rate: US$ = 1166.51 KRW (South Korean won) and were also adjusted to calendar year 2019 using the Consumer Price Index of South Korea.

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

Mean and 95% Confidence Interval of adjusted all-cause total medical costs per person per month is reported. All-cause total medical costs were adjusted for baseline characteristics and healthcare resource utilization (HCRU) and healthcare costs occurred during baseline period. Total medical cost were estimated as the sum of medical costs (inpatient + outpatient) and pharmacy dispensation costs (drug costs + pharmacy service costs).

Costs were converted to US dollars using the average 2019 exchange rate: US$ = 1166.51 KRW (South Korean won) and were also adjusted to calendar year 2019 using the Consumer Price Index of South Korea.

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

Mean and 95% Confidence Interval of adjusted all-cause hospitalization costs per person per month.

All-cause hospitalization costs were estimated as the total costs of services from all inpatient claims (e.g. costs for procedure, diagnostic test, medication, patient meal, and bed stay, etc.) and were adjusted for baseline characteristics and healthcare resource utilization (HCRU) and healthcare costs occurred during baseline period.

Costs were converted to US dollars using the average 2019 exchange rate: US$ = 1166.51 KRW (South Korean won) and were also adjusted to calendar year 2019 using the Consumer Price Index of South Korea.

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

Mean and 95% Confidence Interval of adjusted all-cause outpatient medical costs per person per month. All-cause outpatient medical costs were estimated as the total costs of services from all outpatient claims (e.g. costs for procedure, diagnostic test, prescription, physician fee, etc.) and were adjusted for baseline characteristics and healthcare resource utilization (HCRU) and healthcare costs occurred during baseline period.

Costs were converted to US dollars using the average 2019 exchange rate: US$ = 1166.51 KRW (South Korean won) and were also adjusted to calendar year 2019 using the Consumer Price Index of South Korea.

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

Mean and 95% Confidence Interval of adjusted all-cause outpatient pharmacy dispensation costs per person per month is reported. All-cause outpatient pharmacy dispensation cost were estimated as the total costs of services from all outpatient pharmacy claims (pharmacy dispensing service costs + medication costs) and were adjusted for baseline characteristics and healthcare resource utilization (HCRU) and healthcare costs occurred during baseline period.

Costs were converted to US dollars using the average 2019 exchange rate: US$ = 1166.51 KRW (South Korean won) and were also adjusted to calendar year 2019 using the Consumer Price Index of South Korea.

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

Mean and 95% Confidence Interval of adjusted COPD-related total medical costs per person per month is reported. COPD-related total medical costs were estimated as the sum of medical costs (inpatient + outpatient) of medical claims with ICD-10 codes for COPD and/or COPD-related diseases and pharmacy dispensation costs (drug costs + pharmacy service costs) of pharmacy claims including dispensation of ICS/LABA, LABA, LAMA, LABA/LAMA, short-acting beta-2 agonist (SABAs), short-acting muscarinic antagonist (SAMAs), SABA/SAMA, theophlliyne, or Phosphodiesterase-4 (PDE4) inhibitor and were adjusted for baseline characteristics and healthcare resource utilization (HCRU) and healthcare costs occurred during baseline period.

Costs were converted to US dollars using the average 2019 exchange rate: US$ = 1166.51 KRW (South Korean won) and were also adjusted to calendar year 2019 using the Consumer Price Index of South Korea.

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

Mean and 95% Confidence Interval of COPD-related hospitalization costs per person per month is reported.

COPD-related hospitalization costs were estimated as the total costs of services (e.g. costs for procedure, diagnostic test, medication, patient meal, and bed stay, etc.) from all inpatient claims (medical claims with ICD-10 codes for COPD and/or for diseases due to COPD worsening as any diagnosis and were adjusted for baseline characteristics and healthcare resource utilization (HCRU) and healthcare costs occurred during baseline period.

Costs were converted to US dollars using the average 2019 exchange rate: US$ = 1166.51 KRW (South Korean won) and were also adjusted to calendar year 2019 using the Consumer Price Index of South Korea.

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

Mean and 95% Confidence Interval of COPD-related outpatient medical costs (per person per month) is reported. COPD-related outpatient medical costs were estimated as the total costs of services (e.g. costs for procedure, diagnostic test, prescription, physician fee, etc.) from all outpatient claims (medical claims with ICD-10 codes for COPD as primary ~ 4th secondary diagnosis) and were adjusted for baseline characteristics and healthcare resource utilization (HCRU) and healthcare costs occurred during baseline period.

Costs were converted to US dollars using the average 2019 exchange rate: US$ = 1166.51 KRW (South Korean won) and were also adjusted to calendar year 2019 using the Consumer Price Index of South Korea.

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

Mean and 95% Confidence Interval of COPD-related outpatient pharmacy dispensation costs (per person per month) is reported. COPD related outpatient pharmacy dispensation costs were estimated as the total costs of services (pharmacy dispensing service costs + medication costs) from all outpatient pharmacy claims (pharmacy claims including dispensation of ICS/LABA, LABA, LAMA, LABA/LAMA, short-acting beta-2 agonist (SABAs), short-acting muscarinic antagonist (SAMAs), SABA/SAMA, theophlliyne, or Phosphodiesterase-4 (PDE4) inhibitor) and were adjusted for baseline characteristics and healthcare resource utilization (HCRU) and healthcare costs occurred during baseline period Costs were converted to US dollars using the average 2019 exchange rate: US$ = 1166.51 KRW (South Korean won) and were also adjusted to calendar year 2019 using the Consumer Price Index of South Korea.

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

Date of death event was defined as the date of death recorded in the National Health Information Database (NHID).

Incidence rate of all-cause death was calculated as below: Incidence rate per 1000 person-years= Number of patients with event/ sum of event-free period during the follow-up period (1000 person-years)

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

Number of patients with the event all-cause death is reported.

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

Median and Inter-Quartile Range of time to death from the index date is reported. Date of death event was defined as the date of death recorded in the National Health Information Database (NHID). Time to death was estimated from the Kaplan-Meier curve.

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

Tuberculosis (TB) was defined as ≥1 inpatient claims or outpatient claims 1) with ICD-10 codes for pulmonary TB recorded as any diagnosis, AND 2) with special benefit costs for TB patients during the following period after the index date. To assess the incidence of TB event, the first TB event observed during the follow-up period was considered. Incidence rate of TB was calculated as below:

Incidence rate per 1000 person-years= Number of patients with event/ sum of event-free period during the follow-up period (1000 person-years)

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

Tuberculosis (TB) was defined as ≥1 inpatient claims or outpatient claims 1) with ICD-10 codes for pulmonary TB recorded as any diagnosis, AND 2) with special benefit costs for TB patients during the following period after the index date. Number of patients with tuberculosis (TB) is reported.

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

Median and Inter-Quartile Range of time to first tuberculosis (TB) event is reported. Tuberculosis (TB) was defined as ≥1 inpatient claims or outpatient claims 1) with ICD-10 codes for pulmonary TB recorded as any diagnosis, AND 2) with special benefit costs for TB patients during the following period after the index date. Date of TB was defined as the admission date of inpatient claim or the starting date of outpatient claim with the above 2 conditions to define TB event. Median and Inter-Quartile Range of time to first tuberculosis (TB) event was estimated using the Kaplan-Meier curve.

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

Lung cancer was defined as ≥1 inpatient claims or outpatient claims 1) with ICD-10 codes for lung cancer recorded as any diagnosis, AND 2) with costs for patients with cancers during the following period after the index date. To assess the incidence of lung cancer event, the first lung cancer event observed during the follow-up period was considered.

Incidence rate of lung cancer was calculated as below:

Incidence rate per 1000 person-years= Number of patients with event/ sum of event-free period during the follow-up period (1000 person-years)

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

Number of patients with lung cancer is reported. Lung cancer was defined as ≥1 inpatient claims or outpatient claims 1) with ICD-10 codes for lung cancer recorded as any diagnosis, AND 2) with costs for patients with cancers during the following period after the index date.

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

Median and Inter-Quartile Range of time to occurrence of lung cancer from the index date.

Lung cancer was defined as ≥1 inpatient claims or outpatient claims 1) with ICD-10 codes for lung cancer recorded as any diagnosis, AND 2) with costs for patients with cancers during the following period after the index date.

Date of lung cancer diagnosis was defined as the admission date of inpatient claim or the starting date of outpatient claim with the above 2 conditions to define lung cancer event.

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

Non-tuberculosis mycobacteria (NTM) lung disease was defined as ≥1 inpatient claims or outpatient claims 1) with ICD-10 codes for NTM recorded as any diagnosis during the following period after the index date. To assess the incidence of NTM event, the first NTM event observed during the follow-up period was considered.

Incidence rate of NTM lung disease was calculated as below:

Incidence rate per 1000 person-years= Number of patients with event/ sum of event-free period during the follow-up period (1000 person-years)

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

Number of patients with non-tuberculosis mycobacteria (NTM) lung disease is reported. NTM lung disease was defined as ≥1 inpatient claims or outpatient claims 1) with ICD-10 codes for NTM recorded as any diagnosis during the following period after the index date.

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.

Median and Inter-Quartile Range of Time to non-tuberculosis mycobacteria (NTM) lung disease from the index date were estimated using the Kaplan-Meier curve.

NTM lung disease was defined as ≥1 inpatient claims or outpatient claims 1) with ICD-10 codes for NTM recorded as any diagnosis during the following period after the index date.

Date of NTM was defined as the admission date of inpatient claim or the starting date of outpatient claim with the above 1 condition to define NTM event.

Propensity score matching population: To address the imbalance of potential confounders between LAMA and ICS/LABA groups, treatment groups were matched using propensity scores estimated as by multiple logistic regression analysis based on age, sex, socioeconomics status, Charlson Comorbidity Index, asthma and history of chronic obstructive pulmonary disease (COPD) exacerbation.