Editor,—Routine anti-smoking advice from a physician has been shown to increase the cessation of smoking in adults by about two percentage points compared with controls.1 2However, to date, research using this approach has concentrated on adult smokers. We investigated the likely impact of an equivalent approach with adolescents. As young people are the least likely to use medical services,3 the intervention was in the form of an age-appropriate letter from a general practitioner.

Our intervention was directed at children of 12–14 years, an age when the rate of increase in regular smoking is highest.4Participants were 192 patients (101 boys (53%)) from five metropolitan general practices in Perth, Western Australia. On their birthday, participants were sent a personalised letter that combined a health message and birthday greeting. They were assigned randomly within age strata to receive either a letter on smoking or one on a control topic. The letters on smoking contained information about the dangers of smoking and the benefits of not smoking. The control letters had information on immunisation (12th birthday), minimising exposure to the sun (13th birthday), or dietary information (14th birthday). We conducted follow-up telephone interviews about six weeks after each child’s birthday.

Given the exploratory nature of the project, the assessments were firstly of the recollection of having received a letter, and secondly, of recall of its main topic. Naming the main health topic counted as successful recall—for example: “It was about smoking”; or “About diet and what to eat”—but just recalling the birthday greeting did not.

The table shows the numbers of each type of letter sent out, interviews completed, and the number in each group who correctly identified the main topic. There were no significant differences between the six groups (F (5, 113) = 0.91, p>0.05) or between the experimental and control interventions (F (1, 113) = 2.06, p>0.05) in the intervals between birthday and interview date.

From 192 letters sent, 114 (59%) telephone interviews were completed. There were eight refusals by either the parent or young person, on seven occasions the letter had been intercepted by a parent, and in 43 cases the telephone number (or address) was incorrect or multiple calls were not answered. In 21 instances the letter was not received. There were no significant differences between the groups in the proportion of people for whom an interview was conducted (χ² = 3.00, df = 5, p = 0.70).

Of those interviewed, 98 (86%; 95% confidence interval (CI) = 80% to 92%) remembered having received a letter from their doctor and 75 (66%; 95% CI = 57% to 75%) were able to describe the main health topic. Following a prompt, a further five people remembered receiving it, with three recalling the main topic. Overall 78 (68%; 95% CI = 59% to 76%) were able to describe the main message of the letter.

Across the age groups, 87% (95% CI = 78% to 96%) of those receiving a letter about smoking correctly named the topic whereas only 51% (95% CI = 38% to 64%) correctly identified the other letters, a highly significant difference (χ² =17.48, df = 1, p<0.0001). Eighty-four people said that they read the entire letter (75%; 95% CI = 68% to 83%) and 48% (95% CI = 39% to 57%) still had the letter.

Receiving a personal letter from their family doctor was more memorable than earlier focus groups had envisaged, with over two-thirds of those interviewed able to name the main health topic six weeks later. However, this is likely to be an over-estimate of the magnitude of the effect as those who agreed to be interviewed may have been the most receptive to a letter from their doctor.

Recall of letters about smoking was significantly more likely than recall of the control letters. The reason for this difference is unclear but the former may have been more applicable to adolescents than were the latter. For example, it is unlikely that 12-year-olds were making decisions about immunisation. Yet, this is not the case with diet or sun safety. Outside parental supervision, these behaviours are at least as open to independent actions as smoking, given that they are licit behaviours. All four of the health topics involved in this study had been the subject of local television campaigns.

Having family doctors send letters to their young patients encouraging them not to smoke appears to have considerable potential. Given its simplicity and low cost, this strategy warrants systematic testing in a larger controlled trial with the proportion of regular smokers at an appropriate follow-up point as its principal outcome. Unfortunately, the definitive trial cannot be conducted in Perth as too few general practitioners have computerised record systems that would allow easy identification of all young teenagers who consult the practice.