Population characteristics and matching
A cohort identification flowchart is presented in Figure 1. Overall, out of 1,018 AATD-associated COPD patients who met inclusion criteria, 953 cases had up to 10 unique matched controls by randomly matching with 7,928 patients non-AATD-associated COPD patients.
Table 1 lists the demographic and clinical characteristics of the analyzed cohorts before matching. Patients with AATD-associated COPD were statistically different in terms of all baseline covariates except for mean age and Charlson Comorbidity Index (Table 1).
Among 953 AATD-associated COPD cases, 114 (12%) patients were diagnosed with AATD before the first COPD diagnosis claim, 37 (4%) patients were diagnosed with AATD and COPD on the same day, and 802 (84%) were diagnosed with AATD after the first COPD diagnosis date. Table 2 presents the temporal relationship between the first AATD and COPD diagnoses dates among AATD-associated COPD patients, stratified by time horizons analyzed in this study.
Healthcare resource utilization: multivariate analyses on matched cohorts
The Poisson model did not offer a good fit for any of the healthcare resource categories; hence, a negative binomial model was used as it offered the best fit across all four models tested.
The adjusted effects of Alpha-1 Antitrypsin Deficiency on health care resource utilization before and after the index date are presented in Table 3.
Comparing the adjusted healthcare resource use across both COPD samples, the AATD-associated COPD cohort utilized more healthcare services across OV and OTH resource categories before the index date, as well as more OV, OP, ER, and OTH healthcare services after the index date (Table 3).
Incremental costs
The GLM model with log link and gamma distribution offered the best model fit for overall healthcare costs and all resource categories costs before and after the index date. A two-stage modeling approach was not used as the number of patients with total pre- and post-index costs equal to zero were much less than 5% among both cases and cohorts. spending stratified by healthcare resource categories with respect to the index date are presented in Table 4. Overall, the AATD-associated COPD cohort incurred 2.036 (95% CI: 1.601 – 2.590) times and 1.976 (95% CI: 1.550 – 2.517) times the total 12-month cost incurred by the non-Alpha-1 COPD cohort before and after the index date, respectively.
Only OTH services used by AATD-associated COPD cohort were more expensive (cost ratio: 6.245; 95% CI: 3.424-11.388) before the index date, resulting in an additional $1,560 (95% CI: $739 – $2,454; data not shown).
After the index date, the cases incurred higher OP costs (cost ratio: 1.490; 95% CI: 1.055 – 2.103), higher OTH costs (cost ratio: 11.834; 95% CI: 7.834 – 17.875) and RX costs (cost ratio: 1.372; 95% CI: 1.087 -1.731; Table 4), resulting in an additional $5,782 (95% CI: $3,982 - $7,582) and $1,974 (95% CI: $141 - $3,447) in OTH and RX expenditures, respectively.
Sensitivity analysis
In the sensitivity analysis, the 12-month total cost estimates were re-calculated for AATD-associated COPD patients with ≥ 2 AATD diagnosis claims to account for the possibility of rule-out diagnoses (testing for AATD).
Of note, the 12-month total cost ratio for AATD-associated COPD patients with ≥ 2 diagnosis claims was even higher and totaled 2.768 (95% CI: 2.087 – 3.672) resulting in an additional $14,753 (95% CI: $8,909 – $20.598) incurred before the index date. After the index date, the cost ratio equaled 2.668 (95% CI: 2.002 – 3.554) resulting in an additional $13,029 (95% CI: $7,232 – $18,825) spent.
When analyzing cases with 10 unique matches only, the cost ratio was very close to the base-case estimate and totaled 2.035 (95% CI: 1.541 – 2.686) resulting in an additional $8,486 (95% CI: 3,690 – 13,282) paid by insurers’ before the index date. After the index date, the cost ratio equaled 1.993 (95% CI: 1.508 – 2.636) resulting in an additional $7,361 (95% CI: $2,592 – $12,130) in costs.
In the multi-way sensitivity analysis, the impact of both ≥ 2 diagnosis claims for AATD and 10 unique matches was investigated. Before the index date, the cost ratio was higher than in the base-case analysis and was equal to 2.873 (95% CI: 2.078 – 3.971), resulting in an additional $17,190 ($8,965 – $25,415) incurred by insurers. After the index date, the cost ratio was equal to 2.796 (95% CI: 2.017 – 3.877), which resulted in an additional $15,278 (95% CI: $7,123 – $23,432).
[Figure 1]
[Table 1, 2, 3, 4]