The existing literature on the prevalence of drug driving, the effects of drugs on driving performance, risk factors and risk perceptions associated with drug driving was reviewed. The 12-month prevalence of drug driving among the general population is approximately 4%. Drugs are detected commonly among those involved in motor vehicle accidents, with studies reporting up to 25% of accident-involved drivers positive for drugs. Cannabis is generally the most common drug detected in accident-involved drivers, followed by benzodiazepines, cocaine, amphetamines and opioids. Polydrug use is common among accident-involved drivers. Studies of impairment indicate an undeniable association between alcohol and driving impairment. There is also evidence that cannabis and benzodiazepines increase accident risk. The most equivocal evidence surrounds opioids and stimulants. It is apparent that drugs in combination with alcohol, and multiple drugs, present an even greater risk. Demographically, young males are over-represented among drug drivers. Although there is an association between alcohol use problems and drink driving, it is unclear whether such an association exists between drug use problems and drug driving. Evidence surrounding psychosocial factors and driving behaviour is also equivocal at this stage. While most drivers perceive drug driving to be dangerous and unacceptable, there is less concern about impaired driving among drug drivers and drink drivers than from those who have not engaged in impaired driving. Risk perceptions differ according to drug type, with certain drugs (e.g. cannabis) seen as producing less impairment than others (e.g. alcohol). It is concluded that drug driving is a significant problem, both in terms of a general public health issue and as a specific concern for drug users.