Health care professionals frequently advise patients to improve their health by stopping smoking. Such advice may be brief, or part of more intensive interventions. The aims of this review were to assess the effectiveness of advice from physicians in promoting smoking cessation; to compare minimal interventions by physicians with more intensive interventions; to assess the effectiveness of various aids to advice in promoting smoking cessation and to determine the effect of anti-smoking advice on disease specific and all cause mortality. We searched the Cochrane Tobacco Addiction Group trials register and the Cochrane Controlled Trials Register. Date of the most recent searches: October 1998. Randomised trials of smoking cessation advice from a medical practitioner in which abstinence was assessed at least six months after advice was first provided. We extracted data in duplicate on the setting in which advice was given, type of advice given (minimal or intensive), and whether aids to advice were used, the outcome measures, method of randomisation and completeness of followup. The main outcome measure was abstinence from smoking after at least six months follow-up. We used the most rigorous definition of abstinence in each trial, and biochemically validated rates where available. Subjects lost to follow-up were counted as smokers. Where possible, meta-analysis was performed using a fixed effects model. We identified thirty-one trials, conducted between 1972 and 1997, including over 26,000 smokers. In some trials, subjects were at risk of specified diseases (chest disease, diabetes, ischaemic heart disease), but most were from unselected populations. The most common setting for delivery of advice was primary care. Other settings included hospital wards and outpatient clinics, and industrial clinics. Pooled data from 16 trials of brief advice versus no advice (or usual care) revealed a small but significant increase in the odds of quitting (odds ratio 1.69, 95% confidence interval 1.45 to 1.98). This equates to an absolute difference in the cessation rate of about 2.5%. There was insufficient evidence, from indirect comparisons, to establish a significant difference in the effectiveness of physician advice according to the intensity of the intervention, the amount of follow-up provided, and whether or not various aids were used at the time of the consultation in addition to providing advice. However, direct comparison of intensive versus minimal advice showed a small advantage of intensive advice (odds ratio 1.44, 95% confidence interval 1.23 to 1.68). In one study which determined the effect of smoking advice on mortality at twenty years, there were no statistically significant differences in death rates in the group receiving advice. Simple advice has a small effect on cessation rates. Additional manoeuvres appear to have only a small effect, though more intensive interventions are marginally more effective than minimal interventions.