Differences in MVCs stratified by number of vehicles included in crash during post-stay-at-home order, comparisons between 2020 and average of 2019, 2018 and 2017
| Outcome models | Post-stay-at-home order-2020 compared with average of 2019, 2018 and 2017 | ||
| Incidence rate ratio (IRR) | 95% CI | P value | |
| All type crashes | 0.80 | (0.73 to 0.88) | <0.00 |
| Single | 1.47 | (1.20 to 1.82) | 0.00 |
| Multiple | 0.66 | (0.62 to 0.71) | 0.00 |
| Any injury crash | 0.83 | (0.74 to 0.91) | 0.00 |
| Single | 1.34 | (1.13 to 1.59) | 0.00 |
| Multiple | 0.68 | (0.61 to 0.75) | 0.00 |
| Non-injury crash | 0.79 | (0.72 to 0.87) | 0.00 |
| Single | 1.52 | (1.20 to 1.92) | 0.00 |
| Multiple | 0.66 | (0.61 to 0.70) | 0.00 |
| Fatal crash | 1.66 | (0.97 to 2.84) | 0.06 |
| Single | 2.35 | (1.19 to 4.64) | 0.01 |
| Multiple | 1.08 | (0.68 to 2.42) | 0.84 |
Model specified a Poisson distribution of the outcome with a log transformed VMT as population offset. Robust SEs were specified for a more conservative estimate of coefficients. Model includes linear time, average daily precipitation and average daily highest recorded temperature. Comparison during post-stay-at-home order (order effective in Connecticut on 23 March 2020). VMT, vehicle miles travelled.
Bold Values note statistical significance <=0.05.