Table 2. Association of Patient Characteristics with LVAD Usea.
| Variables | Marginal effect size (95% CI) | ||||||
|---|---|---|---|---|---|---|---|
| Model 1 | Model 2 | Model 3 | Model 4 | Model 5 | Model 6 | Model 7 | |
| Black race (vs White) | –0.017 (–0.039 to 0.005) | –0.022 (–0.041 to –0.003) | –0.036 (–0.057 to –0.015) | –0.043 (–0.065 to –0.022) | –0.045 (–0.072 to –0.018) | –0.030 (–0.058 to –0.002) | –0.030 (–0.058 to –0.002) |
| Female (vs male) | –0.125 (–0.146 to –0.104) | –0.089 (–0.108 to –0.070) | –0.090 (–0.109 to –0.071) | –0.090 (–0.109 to –0.071) | –0.079 (–0.102 to –0.056) | –0.079 (–0.102 to –0.056) | –0.079 (–0.102 to –0.056) |
| LVAD propensity | NA | 0.965 (0.941 to 0.989) | 0.962 (0.938 to 0.986) | 0.960 (0.936 to 0.984) | 0.956 (0.926 to 0.986) | 0.953 (0.923 to 0.983) | 0.954 (0.924 to 0.984) |
| Age, per year | NA | NA | –0.001 (–0.002 to 0.0003) | –0.001 (–0.0013 to 0.0001) | –0.002 (–0.003 to –0.001) | –0.002 (–0.003 to –0.001) | –0.002 (–0.003 to –0.001) |
| Distance to LVAD Hospital, per 10 miles | NA | NA | NA | –0.002 (–0.003 to –0.001) | –0.003 (–0.004 to –0.002) | –0.003 (–0.004 to –0.002) | –0.003 (–0.004 to –0.002) |
| Low-income subsidy | NA | NA | NA | NA | –0.036 (–0.060 to –0.012) | –0.029 (–0.053 to –0.005) | –0.030 (–0.054 to –0.005) |
| Social deprivation index | NA | NA | NA | NA | NA | –0.067 (–0.105 to –0.029) | –0.068 (–0.106 to –0.030) |
| Observations | 12 310 | 12 310 | 12 310 | 12 310 | 7711 | 7710 | 7710 |
| Year indicators | Yes | Yes | Yes | Yes | Yes | Yes | Yes |
| Neighborhood RE | No | No | No | No | No | No | Yes |
| Part D enrollees only | No | No | No | No | Yes | Yes | Yes |
Abbreviations: LVAD, left ventricular assist device; NA, not applicable; RE, random effect.
a
Model 1 was conditional on age, sex, interactions of age and sex, and year indicators. After adjusting for expected LVAD use based on patient characteristics (Model 2), incorporating interactions with patient age (Model 3), distance to LVAD centers (Model 4), individual poverty (Model 5), and neighborhood-level social deprivation (Model 6), Black race and female sex were associated with a –3.0 (95% CI, –5.8 to –0.2) and –7.9 (95% CI, –10.2 to –5.6) percentage points decrease use of LVAD. These associations were consistent when neighborhood included as a random effect (Model 7).