No. Half of the INJURED people in USA are stoned or drunk. You would need a control group that isn't injured to support your conclusion. But, presumably, tha

No. Half of the INJURED people in USA are stoned or drunk. You would need a control group that isn't injured to support your conclusion. But, presumably, that population would have a much lower percentage of stoned/drunk people, in line with loads of existing data showing that operating under the influence is a strong cause of crashes & injuries both statistically, and with a well-established causal chain (impaired perception, impaired cognition, slowed reaction times, etc.)