Background: Central nervous system–depressant (CNS-Ds) drugs can impair cognitive functions and driving. They are also the most common drugs taken in overdose in hospital-treated episodes of self-poisoning. In Australia most of these patients are discharged within 48 h, while they still have possible subclinical drug effects. We aimed to determine whether patients treated for self-poisoning with CNS-Ds are impaired in the Trail-Making Test (TMT, parts A and B), a neuropsychological test that is known to correlate with driving performance. Methods: This study was a conducted from November 2008 to April 2011 in a referral center for poisonings in New South Wales, Australia. One hundred seven patients discharged from the clinical toxicology unit following treatment for self-poisoning of CNS-Ds (benzodiazepines, atypical antipsychotics, or opioids) and a control group of 68 discharged following self-poisoning of non-CNS-depressant drugs (acetaminophen or nonsedating antidepressants) were tested with the TMT (parts A and B). Due to the known association of impaired TMT with driving impairment and increased risk of traffic accidents, performance less than the 10th percentile for age was defined as significant impairment in each part of the TMT. The odds ratio (OR) for impairment in each part was calculated in multivariate logistic regression (MLR) models adjusted for gender, education, IQ, and the presence of a major psychiatric illness. A secondary MLR analysis was conducted only for those patients (78 CNS-D and 54 control group participants) who were directly discharged home, after excluding those who were transferred for further psychiatric care. Results: The odds of impairment in the CNS-D group was 2.8 times that of the control group on the TMT-A (38 [35.5%] vs. 11 [16.2%]: adjusted OR = 2.76, 95% confidence interval [CI]: 1.28–5.97), and 4.6 times on the TMT-B (67 [62.6%] vs. 22 [32.4%]: adjusted OR = 4.63, 95% CI: 2.06–10.42). The results were similar in the subgroup of patients discharged home, and the odds of impairment in the CNS-D group was 3.3 times that of the control group on the TMT-A (25 [32.1%] vs. 7 [13.0%]: adjusted OR = 3.30, 95% CI: 1.28–8.52), and 3.6 times on the TMT-B (46 [59.0%] vs. 17 [31.5%]: adjusted OR = 3.64, 95% CI: 1.44–9.20). TMT-B impairment in the CNS-D group remained significant even after adjusting for TMT-A performance. Conclusions: Patients with CNS-D overdose may have significant impairment in cognitive skills underlying driving at the time of discharge from hospitals. Clinicians should warn these patients that their driving skills might still be impaired, even if they are considered clinically recovered and advise them not to drive during the first 1 to 2 days following discharge.