TABLE 4—
Estimated Number of Deaths and Serious Injuries Among Cyclists and Pedestrians Prevented by the Zone 30 Policy: Japan, September 2011 to December 2016
| No. Killed or Seriously Injured Cases Prevented, Point Estimatea (95% CI)b | |
| Total | 1704 (1293, 2198) |
| Females | |
| Age, y | |
| 0–14 | 66 (−30, 188) |
| 15–24 | 31 (−38, 113) |
| 25–64 | 215 (44, 406) |
| 65–74 | 152 (−17, 346) |
| ≥ 75 | 263 (125, 411) |
| Males | |
| Age, y | |
| 0–14 | 457 (234, 728) |
| 15–24 | 174 (78, 286) |
| 25–64 | 112 (−50, 290) |
| 65–74 | 100 (19, 190) |
| ≥ 75 | 135 (36, 243) |
Note. CI = confidence interval.
We obtained the point estimates for each subgroup with 
, where m is the number of months after January 2005 (m = 1), i (= 1, 2,…, 10) is the subgroup, b2i (with mean 
and standard error 
) is the estimated monthly change in the rate ratio after September 2011 for subgroup i, and 
is the observed number of killed or seriously injured cases for subgroup i in the mth month. We obtained the total number of killed or seriously injured cases prevented by summing the estimated numbers across subgroup.
We obtained the 95% CI for each subgroup by plugging in the upper and lower limits of b2i to
. We obtained the 95% CI for the total number through a Monte Carlo simulation of 1000 draws from 
for each subgroup, assuming independence of randomness among the subgroups. See the “Estimation of the Impact” subsection in the Methods for details.