Objective. To prevent early adolescent health risk behaviors and to maintain or improve safety behaviors, we compared the effects of 2 interventions, delivered through pediatric primary care practices. The interventions, based on an office systems’ approach, sought to prevent early drinking and smoking or to influence bicycle helmet use, gun storage, and seatbelt safety for children who were followed from fifth/sixth grades through eighth/ninth grades.
Design. Settings and Participants. Twelve pediatric practices in New England were paired according to practice size and assigned randomly within pairs to deliver the multicomponent interventions, which built on pediatric primary care clinicians performing as counselors and role models during health supervision visits and other office encounters.
Intervention. One intervention arm focused on alcohol and tobacco use. The other intervention arm focused on gun safety, bicycle helmet, and seatbelt use. Office systems provided infrastructure that supported the clinician’s role. Clinician messages encouraged family communication and rule setting about the issues of the middle school years. The intervention was initiated during a health supervision visit and continued for 36 months. Both child and parent received quarterly newsletters to reinforce the clinician messages.
Outcome Measures. The primary outcomes were ever drinking alcohol, ever smoking, ever using smokeless tobacco, using a bicycle helmet in the previous year, using a seatbelt in the previous 30 days, and guns in the child’s home in locked storage.
Results. The pediatric practices recruited 85% (N = 3525) of the practices’ fifth/sixth grade children and their responding parents. We obtained 36 months’ follow-up data on 2183 child-parent pairs. Chart audit verified that the intervention was implemented. Additional data from interviews and surveys showed that parents, children, and pediatric clinicians found the intervention useful. Despite this, comparisons between the 2 study arms show no significant intervention effects in the prevention of alcohol and tobacco use or gun storage or seatbelt safety. There was a negative effect in the alcohol arm. Only bicycle helmet use showed a positive outcome.
Conclusion. With rigorous evaluation, 2 office interventions failed to produce desired outcomes. Coordinated multiple settings for prevention interventions are probably necessary.
- bicycle helmets
- gun safety
- primary prevention
- health promotion
- behavior change
- primary care
- randomized controlled trial
Prevention is a widely advocated pediatric primary care activity.1 Anticipatory counseling is urged for the important risk factors of adolescence.2 Early adolescence is an appropriate time to address alcohol and tobacco issues and bicycle helmet, seatbelt, and gun safety issues.3,4 The middle school years are a period of developmental transition in our society. Preteens have become more independent and are less closely supervised.5 They have access to alcohol, tobacco, and guns6,7 and may assert independence by not using seatbelts and bicycle helmets.8,a Families not only expect but want their clinicians to give guidance about health behaviors in adolescence.9 These are important issues for pediatricians who care for children in this age group.
An office systems’ approach can enhance implementation of preventive health and screening activities in adult primary care practices.10 The office systems’ approach is based on the premise that the office as a whole delivers preventive health services.11 Actions by physicians are consistently reinforced with the activities of other office staff and written materials over time. An office systems’ approach enhanced pediatric counseling about sun protection.12 The Injury Prevention Program (TIPP) is the only widely known office systems’ approach to prevention in pediatric primary care practices.13 Evaluation of a pediatrician’s single brief message to families at well-child visits, accompanied by written materials showing TIPP, was ineffective in changing gun ownership.14 The results of more intensive primary care-based gun safety interventions have not yet been reported. No evaluation of office-based interventions has previously been conducted for alcohol and tobacco prevention or for bicycle helmet or seatbelt promotion for young adolescents.
Primary care clinicians have been successful in reducing adult alcohol and tobacco use with brief interventions delivered over time in the primary care setting.15,16 Pediatricians, although believing in the importance of preventive counseling, often did not provide even brief prevention counseling in their daily practice.17,18
An intervention designed to promote family communication seemed appropriate based on published prevention studies.19,20 Our pilot data on 50 families in a pediatric primary care practice showed that families talked too late or not at all about drug prevention and safety issues of early adolescence. Neither the parents nor the pediatric primary care clinicians realized that the appropriate time to provide anticipatory guidance was in fifth and sixth grade. Neither realized that experimentation with alcohol and tobacco begins in preadolescence6 and that preadolescents stop wearing bicycle helmets and seatbelts when not monitored.21 Even when families did not own guns, their children had access to guns in friends’ households.22
The Dartmouth Prevention Cohort Study described herein was a large pediatric primary care, office-based, structured prevention intervention. It was designed to test the effectiveness of clinician-delivered advice to promote family communication and to prevent adolescent high-risk behaviors through a randomized, controlled trial. Half the participating practices delivered an alcohol and tobacco use prevention message. The other half delivered an injury prevention message about gun safety, seatbelt use, and bicycle helmet use.
Our goal was to prevent or delay the onset of drinking and smoking behaviors and to enhance safety behaviors. The later a child adopts risky behaviors or changes from safe behaviors, the fewer the adverse consequences.4 We hypothesized that the pediatric primary care clinician could act as an effective catalyst for behavior change by encouraging better communication and explicit rule setting within families to prevent or delay children’s alcohol and tobacco use, to maintain or adopt seatbelt and bicycle helmet use, and to keep guns in locked storage.
Study Population and Recruitment
The Dartmouth Prevention Project was conducted in 12 pediatric primary care practices in Massachusetts, New Hampshire, and Vermont. The practices varied from 2 practitioners in a rural small town to 13 practitioners in a city of 90 000. All practices served a broad range of families from the community for primary care. Families were recruited at health supervision visits, often required at this age for camp, sports, or school. At that time, measles vaccine was still given in the sixth grade.
Figure 1 shows the recruitment pattern. Pediatric clinicians in 12 primary care practices attempted to recruit all families with fifth and sixth grade children who visited their practices for well-child care (N = 4096) during a 21-month period. Of the 4096 families approached by their clinicians, 3525 (86%) agreed to participate, 3496 (85%) met the grade eligibility requirements, and 3145 (77%) completed both the children’s and the parents’ baseline questionnaires and were enrolled in the cohort. To be eligible for the study, the child had to be in the fifth or sixth grade and accompanied to the appointment by her/his parent or legal guardian. Only 1 child and parent pair in a family could participate.
All participating children and parents signed informed consents. Fourteen percent of all families with fifth or sixth grade children who presented for a well-child visit refused to participate. Our institutional review board did not allow us to collect any additional information on potentially eligible families who did not sign an informed consent.
The Intervention Model
The intervention focused on the role of the pediatric primary care clinician, who is familiar with, respected, and trusted by both the child and the parent. We changed the focus of the well-child visit from screening and providing facts to engaging the child and parent in a joint discussion and encouraging communication about alcohol and tobacco use, bicycle helmet, and seatbelt use, and safe gun storage. Office staff conveyed similar messages when they interacted with participating families. We supported delivery of this intervention with site visits, telephone calls, newsletters, and informational materials to the clinical sites and with regularly scheduled visual and printed materials mailed to the families.
Specific Intervention Components
Figure 2 shows the sequence in which the intervention was provided to families. The child and parent signed consent at the beginning of a health supervision visit. Depending on their practice’s randomization status, the primary care clinician focused either on alcohol and tobacco use or on safety as key health issues for young adolescents. They discussed risks with the child and parent. The child, parent, and pediatric clinician signed a contract that the family would talk about the risks at home and develop a family policy about alcohol and tobacco or about safety issues (bicycle helmet, seatbelt use, and gun storage). About 10 days later, the family received a letter signed by their clinician reinforcing the agreement and a refrigerator magnet to post the family’s contract.
The child and parent were reminded of the importance of family communication about the specific issues in their intervention arm at subsequent office visits for 36 months. At each visit, the clinician’s role was to provide age-appropriate risk behavior information, encourage family communication, and offer help. We provided the clinicians with brief age- and risk-appropriate messages each month so that they had new information to focus on over time.
The families received a brochure on effective communication immediately after enrolling and annual reinforces of our message for family and child, eg, a card game about communication, magnets, or pens, each of which carried an intervention message.
Continued reinforcement of our messages to families also occurred through newsletters mailed during the 36 months, a total of 12 for children and 12 for parents. Each intervention had its own series of newsletters with parallel discussions of communication issues, but using the intervention-specific risk factors, either safety or alcohol and tobacco. All newsletters were in developmentally appropriate sequence.
Each parent’s newsletter included parenting or communication skills, new risk factor data, suggested family activities, “what if” exercises, and book or Web site reviews.
Each child’s newsletter included communication strategies and intervention-specific risk factor data from journals, books, and the Internet. The risk factor discussed matched the risk factor in the previous parent’s newsletters. Other articles were about communication and age-appropriate decision-making skills, complementing the communication article in the previous parents’ newsletter. Games, puzzles, and quizzes added an interactive component. Practice staff received copies of child and parent newsletters and a series of their own.
Training the Pediatric Practices for the Intervention
All pediatricians and nurse practitioners (N = 92) in every practice were trained by the pediatrician coinvestigators and senior project staff during a 3-hour session on site. Many clinicians knew the topic of the alternate intervention, but received no details or training. Dr Olsen trained all alcohol/tobacco sites. Dr Boyle trained all safety sites. Training included an introductory presentation on the project’s intervention components and their rationale and rates of preadolescent risk behaviors in the local geographic area. The clinicians were divided into groups of 2 or 3 and role-played the initial intervention with a child/parent pair of standardized patients. Each clinician received feedback from the pediatrician/trainer, their colleagues, and the pair of standardized patients.23 The training session focused on how clinicians might increase parents’ awareness of the early onset of risky behaviors and gain parent and child commitment to initiate family discussion about these health risks rather than only screening for current early adolescent behaviors or providing educational materials.
All other practice staff were trained by the senior project coordinator who gave a project overview, facts about preadolescent risk behaviors in their communities, and clinical office systems. The intervention was broken down into tasks, and office system tools for accomplishing these tasks were introduced. A flow diagram for implementing the project, tailored to the existing organizational structure and staff roles of each practice, was developed and consensus was gained. Posters and a notebook of project tools and family materials were provided initially and updated throughout the project.
Ongoing Practice Supports During Implementation
The practices received a “message of the month,” feedback from chart audits, calls, and routine visits from research coordinators for problem identification and solving. Staff/clinician newsletters were sent quarterly to all participating practices. The newsletter content varied by study arm and over time. During the recruitment phase, we emphasized engaging families in initial visits and reported recruitment rates by site. Later we focused on reinforcing the intervention message. We reported documentation rates on the project flow sheet for each site and announced the quarterly winner, told success stories, and provided tips based on participating practices’ experiences. We usually provided information from a new article or book on 1 or more of our risk behaviors or on parenting skills and communication.
Outcome Evaluation Measures
Children and parents received self-administered, 21-page surveys to complete independently of each other and return by mail at baseline, 12, 24, and 36 months.
The child’s 107-item survey assessed alcohol and tobacco use, bicycle helmet and seatbelt use, and gun access and use. In addition, the survey included questions about school health programs (eg, DARE), the child’s perception of peer behaviors, and of parental support and control.24
The parent’s 170-item survey assessed demographic items, the reporting parent’s and spouse’s alcohol and tobacco use, parent’s perception of child’s alcohol and tobacco use, bicycle helmet and seatbelt use, gun storage in the home, parenting style,25 and the child’s problem behaviors.26
All survey questions were validated in previously published instruments6,13,24–26,28,30 and were piloted for test-retest reliability in a Northern New England population. Overall, the surveys for children are at reading level 5.9 (Gunning-Fogg index). Questions on outcome measures are at reading level 5.1 (Gunning-Fogg index). All surveys were mailed to homes with prepaid postage return envelopes and when both child and parent surveys arrived at the data coordinating center, the child received a $5.00 cash award. If surveys were not returned within 4 weeks, the child and parent received a reminder postcard. If surveys were not returned within 6 weeks, the family received a reminder telephone call. Ninety percent of eligible families returned the baseline survey, constituting a cohort of 3145 families. Over 36 months, 4% were lost to follow-up and fewer than 1% withdrew. Three thousand three families received the intervention. Seventy-three percent (2183/3003) of parents and children completed all surveys: baseline, 12 months, 24 months, and 36 months.
Practices were randomly assigned within pairs using computer-generated, pseudo-random numbers. Baseline survey results were tabulated by intervention group and unadjusted t tests, and χ2 tests were used to identify potential imbalances. We adjusted for baseline variables showing distinct differences between intervention arms. Data for the primary child outcomes were analyzed using logistic regression analysis to control for baseline factors identified as potentially related to outcomes including child’s age, parental education, family income, gender of child, parent’s marital status, child’s having friends who drink, parental high stress and low self-esteem, and parental drinking problems. To account for the randomization within practice pairs, an indicator term was included for each pair in the logistic regression models. Adjusted odds ratios and P values were generated from the coefficients for the indicators for intervention arm.
During the recruitment phase of the project, research staff conducted monthly audits of 10% or at least 20 participants’ medical charts (N = 3700). Audits checked for the project participation identification sticker, copies of signed informed consents, the signed contract and the project flow sheet. The number of encounters documented on the flow sheet was recorded. We mailed a monthly report showing each clinician’s recruitment rate during the past month and the average recruitment rate at all participating practices. This comparison generated a friendly competition among practices. When recruitment was completed, staff visited sites quarterly for an audit of a random sample of at least 20 charts (N = 1100). In addition to checking the above items, the number of office visits documented in the progress notes was compared with the number of office visits documented on the research project flow sheet. We consider this ratio a measure of how consistently sites reinforced the original message. Every staff newsletter included a standard report showing this ratio for each practice and announcing the practice with the best ratio, the winner of the quarterly “pizza party.” Clinicians’ self-reported perceived change in counseling behaviors were surveyed at the end of the project.
In the final questionnaire, children and parents were surveyed as to whether they read the newsletters, liked them, and found them useful. Additional comments on all surveys and phone contacts from children and parents regarding the intervention were compiled.
Implementation by Practices
During the 3 years of intervention, practice level compliance with delivering the initial prevention message and subsequent reinforcement was tracked. Over 99% of participants’ charts were labeled with the research project identification sticker and contained the child/family contract signed during the initial counseling at well-child visits and the flow sheet to document subsequent messages. In the 12 practices’ chart audits, 95% of the participating children had been seen for subsequent visits in the office. The proportion of office visits with a documented prevention message by physician or other staff member averaged 47% in safety sites and 51% in alcohol/tobacco sites. This rate may underestimate actual delivery of prevention messages, because only messages documented in the research flow sheet were counted. The audit did not include prevention topics noted as discussed in progress notes or the well-child check lists that were used for early adolescents in 5 of the sites. Children with chronic problems requiring multiple visits also did not have message reinforced at every visit.
Clinician Prevention Counseling Behaviors
We did not collect observation data on clinician counseling during the intervention visits. With our intervention, clinicians’ self-reported perceptions of their prevention counseling changed specifically for the issues targeted in their practice. Fifty-seven percent of clinicians (52/92) in the practices completed the survey of their counseling practices after the intervention. Fifty-eight percent reported being more informed about their intervention specific prevention issues. Forty percent felt they counseled patients about prevention more frequently, and 42% felt they counseled more young adolescent patients. Twenty-three percent of the clinicians made additional written comments indicating they had increased their counseling over the course of the project. Forty-four percent reported being satisfied or very satisfied with their counseling on the final survey versus 34% on the initial survey before the intervention.
Family Response to the Intervention
A random subgroup of 400 parents and 400 children from both intervention arms were interviewed individually after 12, 24, and 36 months to determine changes in topics discussed by families, policy setting, and recall of clinician advice. Their responses at 36 months displayed in Table 1, show that the intervention had continuing fidelity. In addition, families’ evaluation of the newsletter component was determined at 36 months. A survey mailed separately to parents and children was completed by 57% of families. Ninety-three percent of the parents and 71% of the children read at least half the newsletters. Although only 7% of parents said the newsletters covered new information, 89% reported the topics applied to their family situations. Forty-two percent of the children reported that the information was not new to them, but 74% thought the information did apply to them and their lives. Forty-six percent of the children liked getting the newsletters, but 54% had some criticism of the newsletter. The most common response at 38% was the information was not new to them. Sixteen percent answered the newsletter did not apply to their lives. Seventeen percent thought it was too immature, and 19% thought the newsletter was boring. Parents were less critical, with 63% saying there was nothing they disliked about the newsletter.
In the annual survey, children were asked to recall whether they had visited the doctor in the past year and whether the doctor or nurse had discussed any prevention topics. At baseline, children were significantly more likely to recall clinicians discussing the topics in their intervention arm. However, as the children grew older the rates of discussion by the clinician about alcohol and tobacco became similar in both arms of the study. The rates at which children in the safety arm recalled safety discussions significantly higher throughout the project.
Initially, 89% of children in the safety arm recalled clinicians speaking about seatbelt and bike helmet use at their enrollment visit. This declined to about 60% for the previous year in the final survey. Although the number of children who recalled discussions about guns dropped by 50% during the project, 38% of children in the safety arm recalled clinicians discussing gun safety at the end of the project. Rates of discussion about safety issues in all years was significantly higher in the safety arm. In the alcohol and tobacco prevention arm, 51% of children reported clinicians discussing alcohol and 56% discussing tobacco after their enrollment visit. By the third year, rates of alcohol discussion dropped to 39% and 44% for tobacco. After year 1, the rates of discussion about alcohol and tobacco were not significantly different between intervention arms.
Table 2 compares the 2 intervention groups for baseline child and family characteristics potentially related to outcomes. The intervention groups were well balanced. The alcohol and tobacco intervention group of children was slightly more female (50% vs 46%; P = .04). Gender was included in all subsequent adjusted analyses.
Table 3 shows the baseline rates of alcohol, tobacco use, and safety behaviors in subjects for each intervention group. The majority of children did not always wear bike helmets and lived in homes with unlocked guns. There were no notable differences across intervention groups.
Table 4 shows the intervention effects based on logistic regression analyses with adjustment for child and family baseline characteristics and clinic pairs at 12, 24, and 36 months. The table provides comparison of cumulative outcomes, ie, any such behavior over the period. Analyses of children who had new onset of the behavior were also completed and did not show any significant intervention effects.
At 24 and 36 months, there was a moderate increase in children’s drinking among all children who received the alcohol intervention: odds ratio (OR): 1.27, 95% confidence interval (CI): 1.30, 1.55; and OR: 1.30, 95% CI: 1.07, 1.57, respectively. At 36 months’ follow-up, there was a significant effect of the safety intervention on bicycle helmet use in the last year. Children receiving the safety intervention were more likely to use bicycle helmets than children in the alcohol and tobacco prevention intervention: OR: 0.76, 95% CI: 0.63, 0.92. There were no significant smoking, smokeless tobacco, seatbelt use, or gun storage outcomes in any follow-up.
This intervention was specifically designed to enhance early adolescent prevention counseling in a pediatric primary care practice setting. Significant risks of early adolescence were addressed in an anticipatory guidance model during health maintenance and illness visits. There was one positive result: adolescents in the safety intervention arm increased their use of bicycle helmets. There was also a negative alcohol intervention result at 24 and 36 months.
The strength of this study is as a prospective, randomized intervention with a large community practice sample of preadolescents and parents. We conducted a thorough process evaluation, which confirmed that the intervention was implemented in the practices and that the families in the cohort did receive the intervention.
Although anticipatory guidance is frequently advocated,1 it is rare for outcomes to be evaluated. We were surprised to see that, despite being popular, the primary care interventions did not produce the hypothesized results at the end of middle school. There are several possible reasons.
Our intervention was based on promoting parent-child communication. If the project’s emphasis on supporting family communication stimulated broad family discussion about issues beyond the focus of their intervention arm, both groups could behave similarly and have equivalent rates of alcohol, tobacco use and safety behaviors.
The most substantial increases in adolescent alcohol and tobacco use occur later in high school. There is the possibility that communication supports changed family communication but has its impact not on the age of first experimentation but on later regular use. Follow-up of this cohort in high school will be necessary.
A third possibility is that enhancing general child and parent communication is essential, but not sufficient, to produce a significant effect. Recent findings suggest that more specific kinds of child and parent communication such as explicit monitoring of the child’s activities and explicit role modeling may be required to achieve positive behavioral outcomes.27–29 These directive approaches need to be evaluated in primary care settings.
The most likely explanation for this study’s findings is found in other recent studies about changing health behavior. Comprehensive multisetting approaches including clinical practices, schools, and community organizations have been effective in adolescent drinking and smoking30 and childhood solar protection.31 and bicycle helmets 32. At the practice level, the widely promoted American Academy of Pediatrics’ TIPP program of injury prevention materials has never been evaluated to show changes in injury prevention behaviors. Pediatric practices striving to provide acute and chronic care as well as health promotion with limited time and resources may not be able to change complex health behaviors in isolation. Pediatric practice preventive efforts need to be coordinated with other community components involving both parents and children.
Pediatricians need to assess carefully the health promotion activities advocated nationally in light of limited time and resources. Our project that had appeal to both families and clinicians did not result in significant improvement in safety and substance use behaviors when rigorously studied. Prevention efforts may be better if pediatricians rather combine their efforts with others in the community to deliver a coherent message to patients and families in multiple community settings, not just the office.
PERSONA AND THEORETICAL DECISIONS
“Responses of pregnant women to questions of medical ethical decision making were quite different from those of postpartum women. They consistently and significantly requested less aggressive medical care for critically ill or malformed infants; were more concerned with the infants’ physical pain and with prognosis for the future; and were less concerned with preserving life at all cost. We speculate that the differences observed reflect the intense personalization with which pregnant women related to the cases. Once postpartum, ie, comfortable in the knowledge that their baby was born healthy, the women dealt with the questions in a more theoretical framework.”
Kornbluth LE. Does pregnancy bias medical ethical decision making [abstract]? Pediatr Res. 1996;39:41
Submitted by Student
This study was supported, in part, by the National Institute of Alcohol and Alcohol Abuse grant AA08946. We also received support from the Biostatistical Shared Service at the Norris Cotton Cancer Center.
We thank William Boyle, MD, and our colleagues in the following practices of the Dartmouth Pediatric Research Networks: Dartmouth-Hitchcock Clinic, Concord; Dartmouth-Hitchcock Clinic, Keene; Dartmouth-Hitchcock Clinic, Manchester; Dartmouth-Hitchcock Pediatrics, Nashua; Exeter Pediatrics Associates; Pediatric Associates, Portsmouth; Lahey Clinic Medical Center, Burlington; Plymouth Pediatrics and Adolescent Medicine; and Ottauquechee Health Center for implementing the interventions in their primary care practices. We also thank the children and parents who agreed to participate, the Dartmouth Prevention Project field and data staff, and Anne Olcott.
- Received February 21, 2001.
- Accepted November 10, 2001.
- Reprint requests to (M.M.S.) Norris Cotton Cancer Center, Dartmouth-Hitchcock Medical Center, 1 Medical Center Dr, Lebanon, NH 03756. E-mail:
↵a When we began our intervention, children in New Hampshire, where many members of our cohort reside, could legally cease seatbelt use at age 12.
- ↵Green M, ed. Bright Futures. Guidelines for the Health Supervision of Infants, Children, and Adolescents. Arlington, VA: National Center for Education in Maternal and Child Health;1994
- ↵Guidelines for Adolescent Preventive Services. Chicago IL: American Medical Association;1992
- ↵American Academy of Pediatrics, Committee on Injury and Poison Prevention. Office-based counseling for injury prevention. Pediatrics.1994;94 :566– 567
- ↵Crockett LJ, Petersen AC. Adolescent development. In: Millstein SG, Petersen AC, Nightingale TM, eds. Health Risks and Opportunities for Health Promotion in Promoting the Health of Adolescents: New Directions for the Twenty First Century. New York, NY: Oxford University Press;1993
- ↵American Academy of Pediatrics, Committee on Injury and Poison Prevention. Injury Prevention and Control for Children and Youth. Elk Grove Village, IL: American Academy of Pediatrics;1994
- ↵Fisher M. Parents’ views of adolescent health issues. Pediatrics.1992;90 :335– 341
- ↵American Academy of Pediatrics, TIPP Revision Subcommittee. TIPP. A Guide to Implementing Safety Counseling in Office Practice. Elk Grove Village, IL: American Academy of Pediatrics;1994
- ↵Grossman DC, Cummings P, Koepsell TD, et al. Firearm safety counseling in primary care pediatrics: a randomized controlled trial. Pediatrics.2000;106 :22– 26
- ↵Peterson P, Hawkins J, Abbott R, Catalano R. Disentangling the effects of parental drinking, family management and parental alcohol norms on current drinking by black and white adolescents. J Res Adolesc.1994;42 :203– 227
- ↵Dowswell T, Towner EM, Simpson G, Jarvis SN. Preventing childhood unintentional injuries—what works? A literature review. Inj Prev.1996;2 :140– 149
- ↵Wiley CC, Casey R. Family experiences, attitudes and household safety practices regarding firearms. Clin Pediatr.1993;32 :71– 76
- ↵Jackson C, Henriksen L. Do as I say: parent smoking, antismoking, socialization and smoking onset among children. Addic Behav.1997;22 :107– 114
- ↵Dietrich AJ, Olson AL, Sox CH, Stevens M. A community-based randomized trial encouraging sun protection for children. Pediatrics.1998;102 :1468– 1476
- ↵Rivara FP, Thompson DC, Thompson RS, et al. The Seattle children’s bicycle helmet campaign: changes in helmet use and head injury admissions. Pediatrics.1994;93 :567– 569
- Copyright © 2002 by the American Academy of Pediatrics