Table 3.
Full Log Binomial Regression Results of Association Between Transportation Mode and Home Dialysis Transition (1,148,616 30-day Observation Periods, 109,272 Persons)
| Variable | aIRR | 95% Confidence Interval | P Value |
|---|---|---|---|
| Age, y | 0.97 | 0.96-0.97 | < 0.001 |
| Female sex | 1.09 | 0.99-1.19 | 0.07 |
| Race (reference: White) | |||
| American Indian/Alaska Native | 1.23 | 0.82-1.83 | 0.32 |
| Asian | 0.67 | 0.50-0.89 | 0.006 |
| Black | 0.86 | 0.77-0.95 | 0.003 |
| Hispanic | 0.59 | 0.51-0.68 | < 0.001 |
| Native Hawaiian/other Pacific Islander | 0.74 | 0.47-1.16 | 0.19 |
| Missing/unknown | 0.40 | 0.27-0.61 | < 0.001 |
| RUCA (reference: urban) | |||
| Large rural | 1.15 | 1.00-1.33 | 0.049 |
| Small town/rural | 1.41 | 1.20-1.65 | < 0.001 |
| Suburban | 1.44 | 1.24-1.66 | < 0.001 |
| Binary logarithm of missed dialysis treatment ratea | 1.03 | 1.01-1.05 | 0.015 |
| Social deprivation index (in 20-point units)b | 0.41 | 0.20-0.84 | 0.014 |
| Comorbidities | |||
| Heart failure | 1.01 | 0.90-1.14 | 0.81 |
| Dementia | 0.56 | 0.28-1.13 | 0.10 |
| Diabetes | 0.94 | 0.85-1.03 | 0.19 |
| Hypertension | 1.11 | 0.99-1.25 | 0.07 |
| Dialysis duration (reference: 2-7 y) | |||
| 3-<9 mo | 7.66 | 6.55-8.96 | < 0.001 |
| 9-<14 mo | 3.39 | 2.93-3.93 | < 0.001 |
| 14 mo to <2 y | 1.89 | 1.67-2.13 | < 0.001 |
| 7-<12 y | 0.72 | 0.61-0.85 | < 0.001 |
| 12-<17 y | 0.66 | 0.51-0.87 | 0.003 |
| ≥17 y | 0.83 | 0.62-1.11 | 0.21 |
| Transportation mode (reference: private transportation) | |||
| Medicaid | 0.42 | 0.35-0.50 | < 0.001 |
| Paratransit | 0.53 | 0.41-0.67 | < 0.001 |
| Private pay NEMT | 0.47 | 0.27-0.81 | 0.006 |
| Public transit | 0.47 | 0.35-0.63 | < 0.001 |
Abbreviations: IRR, incident rate ratio; NEMT, non-emergency medical transportation; RUCA, rural urban commuting area; SDI, social deprivation index.
The missed dialysis treatment rate is primarily based on the number of missed dialysis treatments divided by the number of expected dialysis treatments during the study period. In the regression model, the binary logarithm of the missed dialysis treatment rate is entered. As a result, the IRR can be interpreted as describing the adjusted relative rate ratio when comparing two patients where the first patient has twice the missed treatments as the second. Said differently, for example, a home dialysis transition IRR of 1.2 means that each doubling in the missed dialysis percentage corresponds to an approximately 20 percent relative increase in home dialysis transition.
The coefficient describes the IRR associated when holding all things constant when switching between the neighborhoods containing one facility to another, where the difference between one facility and another is 20-point difference.