Table 2.
Pairwise tests of parallelism for ischemic heart disease mortality trends by demographic and geographic subgroups.
| Comparison | K max | Numerator df | Denominator df | Permutations | p-value | Result |
|---|---|---|---|---|---|---|
| Sex (1999–2023) | ||||||
| Female vs. male | 4 | 9 | 30 | 4,500 | *** | Rejected |
| Race/ethnicity (1999–2023) | ||||||
| Hispanic vs. NH Black | 4 | 9 | 30 | 4,500 | ** | Rejected |
| Hispanic vs. NH Other | 2 | 5 | 38 | 4,500 | ** | Rejected |
| Hispanic vs. NH White | 4 | 9 | 30 | 4,500 | *** | Rejected |
| NH Black vs. NH Other | 4 | 9 | 30 | 4,500 | ** | Rejected |
| NH Black vs. NH White | 4 | 9 | 30 | 4,500 | * | Rejected |
| NH Other vs. NH White | 4 | 9 | 30 | 4,500 | 0.3964 | Failed to reject |
| Census region (1999–2023) | ||||||
| Midwest vs. Northeast | 4 | 9 | 30 | 4,500 | *** | Rejected |
| Midwest vs. South | 4 | 9 | 30 | 4,500 | *** | Rejected |
| Midwest vs. West | 4 | 9 | 30 | 4,500 | *** | Rejected |
| Northeast vs. South | 4 | 9 | 30 | 4,500 | ** | Rejected |
| Northeast vs. West | 4 | 9 | 30 | 4,500 | * | Rejected |
| South vs. West | 4 | 9 | 30 | 4,500 | ** | Rejected |
| Urbanization (1999–2020) | ||||||
| Metropolitan vs. Non-metropolitan | 3 | 7 | 28 | 4,500 | *** | Rejected |
NH, not Hispanic; Kmax, assumed number of joinpoints for the comparison test; df, degrees of freedom. The maximum number of joinpoints (Kmax) for each pairwise test was determined based on the WBIC-optimal model of each subgroup, leading to varied Kmax across comparisons. This software-driven setting should be considered when interpreting pairwise parallelism outputs due to differing joinpoint specifications between contrasts. Urbanization analysis limited to 1999–2020.
*p < 0.05; **p < 0.01; ***p < 0.001 by test of parallelism (4,500 permutations).