Perceived Accessibility as a Predictor of Youth Smoking

We thank Dr. Blum and Ms Johnson for their thoughtful comments about our article. They raised several important issues. However, it would have been interesting to learn about the work being done at their center, particularly how this present research might be used in their work and the role that family physicians may play in reducing youth smoking. We hope that Drs. Alan Blum and John Pierce will provide additional insights for the broader readership. We share with Dr. Blum and his colleagues the passion for curtailing tobacco use in adolescents. The cost of tobacco-related diseases including coronary artery disease, pulmonary disease and cancer is enormous. Youth are very vulnerable to the addictive effects of smoking. It is well known that people who begin smoking as adolescents are more likely to be become addicted and have prolonged exposure to the harmful effects of exposure to the tobacco smoke. Family physicians play an important role in the care of children and in reducing harm from smoking. The pathways to teen smoking and its prevention are far from being fully understood. Our paper published in this journal successfully quantified the relationship between perceived accessibility to cigarettes and initiation of smoking among adolescents, which adds new knowledge to the body of literature on tobacco control. Although the construct of perceived accessibility is well known, to our knowledge, this is the first study to carefully quantify the relationship of increased perceived accessibility with higher levels of tobacco use in the context of a well-designed longitudinal study. Our study shows that increased perceived accessibility increases the risk of future smoking for kids who have never before smoked. We also found that perceived accessibility increases the risk of escalation after kids puff on a cigarette for the first time. This makes perceived accessibility a useful measure to explore for use in clinical settings and public health programs. The sources of cigarettes for adolescents are diverse, temporally dynamic, and vary across communities depending on local enforcement measures and community norms. It has therefore not been feasible to accurately measure and track access to tobacco products in clinical practice. However, perceived accessibility is relatively easy to measure and can be routinely assessed during well-child or even sick visits. Kids might be more willing to tell their physician about whether they think they would be able to get cigarettes than to tell about their own tobacco use or intentions. Therefore, the practical importance of perceived accessibility should not be ignored as Dr. Blum and colleagues have suggested. It is supported by evidence and provides a clinically practical way to assess perceived accessibility by family physicians in primary care settings. As practicing family physicians (CAD, JRD), we agree that the challenges of medical practice are overwhelming and there is precious little time available to attend to the multiplicity of preventative care mandates that currently exist. That said, tobacco use is one of the most important public health challenges of our times. Family physicians are uniquely positioned to identify at-risk kids because of the role we play in providing care to generations within the same family. Therefore, we can use our knowledge of the families we see to help reduce the deleterious effects of smoking. Accessibility to cigarettes and the influences of family and friends are important factors that we should ask and address. We recommend that clinicians modify the way smoking risk is assessed by incorporating our findings to the assessment of kids as part of The Early Periodic Screening, Diagnosis, and Treatment (EPSDT) Program. As said previously, it is feasible and of practical importance for clinicians to routinely evaluate perceived accessibility of their adolescent patients. We agree that the relationship of smoking with perceived accessibility and peer smoking is complex. We did not intend to imply that the relationship is unidirectional. Kids who have friends that smoke are more likely to perceive easy access and easy access promotes tobacco use. In addition, kids who smoke or are interested in smoking may be more likely to associate with others that smoke. Performing studies on youth is complex and our methods have limitations. The DANDY study has contributed to our understanding of youth smoking due to the strengths of the design. It is possible that kids may be reluctant to volunteer information about their smoking habits. That scenario is more likely to result in finding that there is no association between perceived accessibility and smoking. So the fact we found an association strengthens our conclusions. Moreover, the kids in the study underwent 3 rounds of interviews each year for nearly 4 years with a low attrition rate. Therefore, the interviewers were not "strangers" to the participants. This longitudinal design is very strong indeed and the repeated measures may also help reduce reporting bias. Moreover, we do not know of any research that shows that face-to-face interviewing with children in a longitudinal study is inappropriate. The quality control measures instituted for the DANDY study also strengthens the study and the findings. Unfortunately, the word limits for manuscripts limited our ability to describe all the details of the study. Additional details about the study have been published and are included in the reference list of the article. In the course of writing this response, we found several videos of the Joe Camel commercials on YouTube. Although the ads were pulled a long time ago, some children may still remember them and others may continue to be exposed to them– isn’t that what the tobacco industry wanted? This is important if the impact of bans on tobacco advertising is to be studied. It may also take many years for the impact of tobacco advertising to be erased from the minds of children who were exposed, and the nature of today’s new media may make the effects of cigarette advertising hard to undo. However, we considered but did not include the Joe Camel variable in the analyses that produced the results. We agree with the concern about the race/ethnicity categories that we used in the analyses. We were constrained by the demographics of our study population. The kids interviewed were mostly from towns with very few minority groups. This limitation of the study does not support analysis by racial subgroups. We are sorry that the Blum and Johnson were "flummoxed" by the term “person-time.” Like man-hours, this term is used as a measure of the number of subjects and the length of time they were followed and is routinely used in the context of survival analyses. The unit of time chosen is necessarily arbitrary. This could have been explained more clearly to readers who are less statistically oriented. We are open to clarify issues raised by our readers if warranted. We agree that dichotomizing the ordinal variable leads to some loss of information. However, in Table 2 we showed the results of the analyses using the original categories of perceived accessibility. That showed a strong dose-response effect - the higher the perceived access the higher the risk of smoking. In order to simplify the presentation in the paper and make the study results more accessibility to readers with less statistical training, we presented the results based on dichotomizing the variables. Since common logistic regression models were applied, the interpretation of results is straightforward and easier to understand. Our analyses show that neither our findings nor the interpretation are comprised by dichotomization of perceived accessibity. . We share Dr. Blum’s sense of outrage about the display of tobacco products. We are interested in knowing if he has suggestions on how this apparent paradox or hypocrisy might be addressed perhaps through policy changes.

Competing interests:   None declared