Evaluation of Use & Lose Laws

Page 5

III. RESULTS (continued)

Proportional Hazard Analyses: Rate of Incidence of 1^st Subsequent Violation or 1^st Subsequent Crash.

On average, the subjects were followed for 3.4 years (25th percentile=2.7 years, 95th percentile=4.8 years). Overall, the cumulative rate of a subsequent 1st violation per month after the input arrest was higher than the one for a subsequent 1st crash. Figure 1 shows that the slope for the rate of a subsequent 1st violation is greater than the slope for the subsequent 1st crash.

Figure 1. Cumulative Hazard Rate for Subsequent 1^st Violation vs. Subsequent 1^st Crash, Pennsylvania (Total=5,690)
Chart of cumulative hazard for each month

Chart of cumulative hazard for each month

D

Proportional Hazard Analyses: Associations Between Age or Gender or Driver's License Action Group with the Rate of Incidence of 1^st Subsequent Violation

Using Cox proportional hazards regression model, age, gender, and license action group were found to be significant predictors of the rate of incidence of a subsequent 1st traffic violation. The unadjusted hazard ratio (HR) of 0.84 for age showed that, for every increase in age increment, older offenders were committing 1st subsequent traffic violations at 0.84 times the rate of the younger offenders (95% CI, 0.81-0.87). For example, subjects 18.8 years old (25th percentile) had 1.30 times the rate of 20.3 year-olds (75th percentile) offenders (95% CI, 1.23- 1.37).

Further, compared to females, males had a higher incidence of having a 1st subsequent traffic violation over time (HR=1.60, 95% CI, 1.41-1.81). Offenders who received a DWI- related driver's license action committed 1st subsequent traffic violations at 0.65 (95% CI, 0.60- 0.70), those who received a Use/Lose-related driver's license suspension at 0.75 (95% CI, 0.67- 0.83), and those with other license action at 0.82 (95% CI, 0.69-0.97) times the rate of those who had no license actions.

Adjusted associations between age, gender, driver's license action group, and the rate of incidence of subsequent 1st traffic violations confirmed the results of the unadjusted associations discussed above. Older offenders had a lower rate of incidence of a subsequent 1st traffic violation than the younger offenders: HR=1.25 (95% CI, 1.19-1.32) for 18.8 year-olds (25th percentile) vs. 20.3 year-olds (75th percentile). Males committed subsequent 1st traffic violations at 1.52 times the rate of females (95% CI, 1.35-1.73). Finally, compared to the group with no license actions after the input arrest, those with a DWI-related license action had the smallest rate of incidence of a subsequent 1st traffic violation (HR=0.68, 95% CI, 0.63-0.74), those with a Use/Lose-related license action had the second smallest rate (HR=0.75, 95% CI, 0.68-0.84), and those with other license actions had the third smallest rate (HR=0.81, 95% CI, 0.68-0.96). Table 10 presents the results of crude and adjusted associations between the independent and dependent variables.

Table 10. Proportional Hazards Model: Age or Gender or Driver's License Action Group as Predictors of Rate of Incidence of 1st Subsequent Violation, Pennsylvania (Total=5,690)

Age:

Gender:

License Action Group:

µ-2Log likelihood=53540.75, p=0.000 (Model includes age, gender, and license action group)

Figure 2 presents the plots of rates of incidence of having a subsequent 1st traffic violation after the input arrest for males vs. females. Differences in the cumulative rates of having a subsequent 1st traffic violation between the four license action groups are depicted in Figure 3.

Figure 2. Cumulative Hazard Rate for Subsequent 1^st Chart of cumulative hazard for each month

D

Proportional Hazards Analyses: Associations between Age or Gender or Driver's License Action Group with the Rate of Incidence of A Subsequent 1st Crash.

Figure 3. Cumulative Hazard Rate for Subsequent 1^st Violation, By License Action Groups, Pennsylvania (Total=5,690)
Chart of cumulative hazard for each month

D

The results of the bivariate proportional hazards regression models showed that only age and license action group were statistically significant predictors for the rate of incidence of a subsequent 1st motor vehicle crash. The hazard ratio for subsequent 1st crashes over time, after the input arrest date, among 18.8 year-olds (25th percentile) vs. 20.3 year-olds (75th percentile) was 1.13 (95% CI, 1.01-1.25). Subjects who received a Use/Lose-related driver's license suspension were involved in a subsequent 1st crash at 0.66 (95% CI, 0.53-0.81), and those who received a DWI-related license suspension at 0.80 (95% CI, 0.70-0.92) times the rate of those who had no license actions. The rate of incidence of a subsequent 1st crash among those with other license actions did not vary significantly from those with no license actions (HR=0.74, 95% CI, 0.53-1.04). The difference based on gender was not statistically significant (HR=1.15, 0.93- 1.42).

After controlling for all three independent variables, only age and driver's license action group were significantly associated with the rate of incidence of a subsequent 1st crash. Older offenders and those who had some driver's license actions applied to them had a lower rate than those who were younger and did not have any license action. For example, those who committed an input offense at 18.8 year-olds (25th percentile) had 1.12 times the rate of 20.3 year-olds (75th percentile) (HR=1.12, 95% CI, 1.00-1.25). Compared to the group with no license actions applied, the group with Use/Lose-related license actions had the lowest rate (HR=0.66, 95% CI, 0.53-0.82), and the group with DWI-related license actions had the second lowest rate (HR=0.83, 0.72-0.95). The group that received other license actions had a rate that was not significantly different from the rate of those with no license actions applied (HR=0.74, 0.52-1.04). Table 11 summarizes these results.

Table 11. Proportional Hazards Model: Age or Gender or Driver's License Action Group as Predictors of Rate of Incidence of 1st Subsequent Crash, Pennsylvania (Total=5,690)