Using a Computer Kiosk to Promote Child Safety: Results of a Randomized, Controlled Trial in an Urban Pediatric Emergency Department
OBJECTIVES. The effects of a computer kiosk intervention on parents' child safety seat, smoke alarm, and poison storage knowledge and behaviors were evaluated in a pediatric emergency department serving predominantly low-income, urban families. The effects of parent anxiety and the reason for the child's emergency department visit also were examined.
METHODS. A randomized, controlled trial of a Safety in Seconds program with a 2- to 4-week follow-up interview was conducted with 759 parents of young children (4–66 months of age). The intervention group received a personalized report containing tailored, stage-based safety messages based on the precaution adoption process model. The control group received a report on other child health topics.
RESULTS. The intervention group had significantly higher smoke alarm, poison storage, and total safety knowledge scores. The intervention group was more likely to report correct child safety seat use. Neither parent anxiety nor the reason for the emergency department visit was related to the safety behaviors. Virtually all (93%) intervention parents read at least some of the report; 57% read it all, and 68% discussed it with others. Lower-income intervention parents who read all of the report and discussed it with others were more likely than control parents to practice safe poison storage. Higher-income intervention parents were more likely than control parents to report correct child safety seat use.
CONCLUSIONS. These results bode well for widespread applicability of computer technology to patient education in busy emergency departments and other child health care settings. Reducing financial barriers to certain safety behaviors should continue to be a high priority.
- injury prevention
- child safety seats
- smoke alarms
- poison prevention
- emergency department
- trauma center
- computer kiosk
- precaution adoption process model
- low literacy
Injuries remain the leading cause of death for children in the United States; among children <6 years of age, 3329 died in 2003 and 2.7 million received emergency medical care in 2004.1 In 2000, total childhood injuries cost $50.6 billion.2 Effective safety products exist, such as child safety seats3 and smoke alarms,4 and their use is widely recommended.5 Unfortunately, many families, especially those with low income, do not use safety products and practice safety behaviors consistently and properly.6–9
The emergency department (ED) can provide an appropriate venue for injury prevention. In 2004, there were 12.2 million ED visits by children <5 years of age, 38% of which were neither emergency nor urgent.10 EDs have access to large populations, and some have argued that a visit could provide a “teachable moment.”11,12 This potential remains largely unexplored in early childhood injury prevention, despite the long-standing interest in prevention by the American College of Emergency Physicians.13 One concern is that ED care providers have limited time to devote to counseling. New computer technology has the potential to reduce the time demands on providers, and it has demonstrated promise for other health problems14–16 and for injury prevention in primary care settings.17,18 Only 1 study could be found that evaluated this technology in an ED, and it focused on alcohol misuse among injured adolescents.19 Therefore, investigating the application of computer technology to early childhood injury prevention in the ED setting is both timely and important.
The objective of this study was to evaluate a theory-based, computer-tailored intervention called Safety in Seconds, which was designed to promote parents' car seat, smoke alarm, and poison storage safety knowledge and behaviors. In addition to testing the hypothesis that the intervention would result in higher rates of correct safety knowledge and behaviors, we examined the effects of the intervention on parents whose children were being seen for an injury and parents who were anxious. No previous ED intervention study has compared injured and noninjured patients, leaving it unclear for whom there may be a teachable moment. Because prevention may be more salient for injured children, we hypothesized greater effectiveness for this group. We also hypothesized that anxious parents would benefit less from the intervention because an ED visit can be anxiety provoking20,21 and because we know that anxiety can influence how individuals hear, understand, and react to information.22
A randomized, controlled trial was undertaken in the ED of a level 1 pediatric trauma center. Parents of young children were recruited at the time of the child's ED visit regarding either an injury or a medical complaint. A computer kiosk located in the ED waiting room was used to assign participants to study groups, to collect baseline data, and to generate the parent reports. Participants were assigned randomly to receive 1 of 2 reports, (1) a personalized and stage-tailored safety report (intervention group) or (2) a personalized but otherwise generic report on other child health topics (control group). Follow-up interviews were conducted by telephone 2 to 4 weeks and 4 months after enrollment, by interviewers who were blinded to treatment group. Data available for analysis in this article are from the first outcome assessment at 2 to 4 weeks only. The study was approved by the Johns Hopkins Committee for Human Research.
Data Collection and Randomization
To capture the full range of health and injury conditions seen over time and to minimize seasonality concerns, we recruited participants from September 2004 through December 2005. On the basis of an analysis of the previous year's ED visits, we selected the 12 busiest shifts per week for recruitment times. Study recruiters screened triage sheets to identify children who were in the study age range, and they approached potentially eligible parents in the waiting room. The parent of any child whose visit was noted with suspicion of child abuse or neglect or whose child was critically ill or injured was not approached. Eligibility criteria included English-speaking parent or guardian of a child between 4 and 66 months of age being seen for an injury or medical complaint, or an age-appropriate sibling of a child being seen for those reasons; living in Baltimore City; and living with the child “at least some of the time.”
Once the study recruiter obtained written informed consent, participants were escorted to the computer kiosk. When the computer program was activated by the study recruiter, the random number-generation program in FileMaker Pro (FileMaker Inc, Santa Clara, CA) assigned the participant to the intervention or control group, and the appropriate baseline assessment instrument appeared on the screen. Participants in both groups then completed a 12-minute assessment, after which the kiosk printed their reports. Participants were paid $10, to thank them for their time, and they were telephoned 2 to 4 weeks later.
Estimates for sample size calculations were taken from our previous intervention work, showing rates of safe poison storage at 10% and working smoke alarms at 80%.6,23 The desired sample size of 375 to 400 families per group was based on type I error α at .05 and power of 0.80. For 0.10 ≤ p1 ≤ 0.80 (p1 is the proportion of safe practices in the control group), we can detect differences of ≥10% in p2 (proportion of safe practices in the intervention group) with sufficient protection from type I and type II error.24
Intervention Condition (Safety in Seconds)
The Safety in Seconds program drew on theories of information processing and behavior change. First, message tailoring is grounded in the elaboration likelihood model,25 which states that the more personally relevant messages are, the more likely they are to be processed cognitively, remembered, and used. Tailoring is an assessment-based approach in which an individual provides personal data that are used to determine the most-appropriate messages for that individual.26 Computer technology allows tailoring to be used efficiently with large numbers of people, and it is a promising intervention strategy being used to address myriad health behaviors.27 Second, we drew on the precaution adoption process model (PAPM) described by Weinstein and Sandman,28 which describes behavior change as a process that evolves through 6 predictable stages, that is, (1) unaware (does not know about the need for the behavior), (2) unengaged (not thinking about adopting the behavior), (3) undecided (thinking about adopting the behavior), (4) decided not to adopt the behavior, (5) decided to act (planning to adopt the behavior), and (6) acting (adopting the behavior). PAPM has received increasing support for its usefulness in predicting health-related actions, and the Safety in Seconds program seems to be the first study to apply it to parents' child safety behaviors (L.B.M., W.C.S., E.M.M., unpublished data, 2007).
Application of PAPM
On the basis of consultation with Dr Neil Weinstein regarding conceptualizing and measuring child safety seat use, smoke alarm use, and poison storage, we determined that each of these behaviors actually consists of several different behaviors. For example, correct smoke alarm behavior is made up of having ≥1 working smoke alarm, changing the batteries in the smoke alarm at the proper intervals, and, for complete coverage, having a smoke alarm on every level in the home and changing all of the batteries at the correct intervals. We labeled these “behavioral profiles” (Table 1). The number of profiles can differ according to safety topic, because achieving maximal protection is more or less complex depending on the safety behavior. We determined that there were 3 behavioral profiles for safe poison storage (having a locked place, keeping poisons in the locked place, and returning poisons to the locked place after each use) and 4 each for smoke alarms and child safety seats (Table 1). Within each behavioral profile, a person can be in any 1 of the 6 PAPM stages. The assessment used in this study assigned each participant to 1 behavioral profile and 1 PAPM stage for each of the 3 safety topics. The benefit of assessing a person's behavioral profile and PAPM stage is that this allows a more tailored and presumably more effective educational message to be generated.
Implementation of the Intervention
To assign the participant to a behavioral profile and specific PAPM stage, we developed “staging algorithms” for each safety topic, which were administered as part of the baseline assessment at the computer kiosk. Each safety topic was assessed with 10 to 12 items, depending on the parent's responses. For example, a parent whose responses indicated that she had the wrong type of car seat for her child's age and weight would then be asked if she had heard about the need to have the right type of car safety seat. If she responded “no,” then she would be classified as profile 2, stage 1 (unaware), for child safety seat use. On the basis of responses to the safety topic items and to other questions about the parent and the child, a parent safety report was then printed at the kiosk. This colorful, 4-page report was personalized with the child's name, gender, ethnicity, and age and was tailored to the parent's behavioral profile and PAPM stage. For example, a parent who was in profile 2, stage 1, for child safety seat use, as described above, would receive the following message: “Because Darius weighs 31 pounds and is 4 years old, the best types of car safety seats for him are a forward-facing convertible seat or a high-back booster (with or without a harness). These are the seats that match his size the best. Just like shoes or clothes, different car safety seats are made to fit children of different sizes. Darius will be the safest in a car safety seat that fits him. The good news is that car safety seats are easy to get! You can get the right kind of car safety seat for Darius at the Children's Safety Center or many local stores.”
Messages were designed to move parents to a higher stage within their behavioral profile or to the next behavioral profile if they were already in the highest stage. Parents whose answers put them in both the highest stage and the highest behavioral profile were given a message that congratulated them and reinforced their continued attention to child safety. Because many of the families seen in the ED have limited literacy skills, the report was written by following established guidelines for low-literacy materials and was pilottested extensively before implementation.29,30
At the kiosk, the control group participants completed an assessment of the same length as for the intervention group. Items included sociodemographic characteristics and knowledge and concerns about 4 child health-related topics, namely, development, sleep, neighborhood safety, and dog bites. The control group then received a kiosk-generated report, which used the same 4-page template and followed the same guidelines for low-literacy materials as for the intervention group. The report differed from that for the intervention group in that it was personalized with the child's name only and contained generic information on the 4 child health-related topics.
We did not assess the child safety seat, smoke alarm, and poison storage behavioral profiles and PAPM stages for the control group for 2 reasons. First, the use of a randomized design allowed us to assume equivalence of the 2 groups. Second, the behavioral profile and stage assessments for each safety behavior were an inseparable component of the intervention itself (ie, the assessments determined the message content) and not a baseline measure in the traditional sense of a pretest/posttest design. Therefore, it was neither necessary nor appropriate to assess the behavioral profiles and PAPM stages for the control group.
At enrollment, participants in both groups were asked their child's age and gender and their relationship to the child, ethnicity, education, income, and marital status. Per capita income was calculated as total household income divided by the total number of individuals supported.
To test knowledge of the information provided in the parent safety report, 10 multiple-choice and true/false items were developed and pilottested with families in the same ED (3 related to child safety seats, 3 related to smoke alarms, and 4 related to poison storage). These items were administered to both study groups in follow-up assessments.
Behavioral Profile and PAPM Stage
The follow-up interviews for both study groups assessed behavioral profiles and PAPM stages for each safety topic by using staging algorithms with 10 to 12 items, depending on the parent's responses. The outcome variables for each safety topic were single ordinal variables that incorporated both behavioral profiles and PAPM stages (Table 1). All data were ordered from profile 1, stage 1, through the highest behavioral profile/stage combination, which was profile 4, stage 6, for child safety seat use and smoke alarm use and profile 3, stage 6, for poison storage. It should be noted that stage 4 in the PAPM is “decided not” to adopt the behavior in question; however, because we had few respondents in this category (9 for child safety seats, 1 for smoke alarms, and 6 for poison storage), we retained the data in this stage for the analyses, rather than deleting them or reassigning them to another stage. On the basis of the distributions of these single ordinal variables in the total combined sample for each safety topic, child safety seat use was best categorized into approximate quartiles and poison storage and smoke alarm use were each best categorized into dichotomous variables (highest behavioral profile/stage combination versus all others).
Reason for Child's Visit
The reason for the child's visit was recorded from the triage sheet and was coded as an injury or a medical condition.
The State-Trait Anxiety Inventory described by Spielberger31 was completed by the parent at enrollment, preferably before the child was seen (although, because of patient flow issues, 23% of participants completed it after the child was seen). There was no difference in the mean anxiety scores for those who completed the inventory before versus after the visit (scores: 34.5 vs 35.7; t = −1.27; P = .20).
Use of Kiosk Reports
In follow-up interviews, 4 questions assessed use of the report by parents in the intervention group, as follows. Do you remember receiving a report from the computer kiosk? Did you read the report? How much of the report did you read (some, most, or all)? Did you discuss the report with family or friends? A summary variable was created to measure what we considered optimal exposure to the intervention; “high exposure” was defined as reading all of the report and discussing it with others, and “low exposure” was characterized as all other responses.
We first compared the intervention and control groups with respect to sociodemographic characteristics, as a check on the randomization and equivalence of the groups at follow-up assessments. Knowledge outcomes were compared between study groups by using t tests for the total mean percentages of correct scores for each safety topic and for the total test. For the behavioral outcome analyses, we used ordinal regression for the quartiles of child safety seat use and logistic regression for the dichotomous poison storage and smoke alarm use variables. For the intent-to-treat analysis, the only variable in the models was study group, and the intervention group consisted of all participants regardless of their reported exposure to the intervention. To address the hypothesis about anxiety, regression models for each of the behavioral outcomes included study group, anxiety, and an interaction term. The hypothesis about the child's reason for the ED visit was tested in the same way. For the intervention exposure analysis, we used ordinal and logistic regression analyses to compare behavioral outcomes across 3 study groups, that is, control group, intervention group with low reported exposure, and intervention group with high reported exposure. To complete the exposure analyses, we examined the data for sociodemographic confounders, including ages of parent and child, maternal education, marital status, ethnicity, relationship to child, child's gender, and per capita income. The latter variable was associated significantly with both high reported exposure and outcomes; therefore, we present final regression analyses for each behavioral outcome with the 3 study groups and per capita income included as the independent variables.
We approached 1412 parents with children who were age-eligible according to the triage sheet; 239 (17%) were ineligible, 201 (14%) refused to participate, and 69 (5%) were missed by the recruiters (Fig 1). The child's age and the reason for the visit were the only data available for parents who refused to participate or were ineligible; they were no different from enrolled subjects with respect to these variables. A total of 901 parents were enrolled (448 in the intervention group and 453 in the control group). Follow-up rates were 86% for the intervention group (n = 384) and 83% for the control group (n = 375). In both study groups, completion rates did not differ according to child's gender, age, and reason for visit and respondent's relationship to child, ethnicity, marital status, employment, income, and education. In both groups, completion rates were lower for older mothers (>30 years), relative to younger mothers (78% vs 85%; P < .05).
The majority of respondents were black (93%), mothers (90%), unmarried (69%), between 20 and 29 years of age (58%), with a high school degree (75%), and with an annual per capita income of less than $5000 (63%). The largest proportion (42%) of children were 1 to 2 years of age, divided equally between boys and girls. Most ED visits were for medical complaints (72%), and the mean anxiety score (score: 34.95) was within the normal range. There were no differences between the intervention and control groups with respect to sociodemographic characteristics, reason for ED visit, and anxiety (Table 2).
The intervention group scored significantly higher on knowledge related to smoke alarms and poison storage and in total (Table 3).
The intervention group was significantly more likely to be in a higher behavioral profile/stage for using child safety seats, relative to the control group (Table 4, model 1). Although not statistically significant, the odds ratios were in the positive direction for smoke alarms and poison storage as well. Neither anxiety nor reason for the visit was related significantly to any of the outcomes, and there were no interactive effects (data not shown).
Virtually all participants in the intervention group remembered receiving the report (98.4%) and reported reading at least some of it (92.9%). More than one half (57.3%) of the intervention group reported reading the entire report, and 68.1% discussed it with family members or friends. In total, 39% reported both reading the entire report and discussing it. This high-exposure group was significantly more likely to be in a higher behavioral profile/stage for all 3 behaviors (Table 4, model 2). Although 23% of the control group reported proper child safety seat use, 40% of the high-exposure group did so. Rates of proper smoke alarm use were higher overall, with 80% of the control group and 89% of the high-exposure group reporting working smoke alarms on all levels. Safe poison storage was reported by 61% of the control group, compared with 76% of the high-exposure intervention group (Fig 2).
High- and low-exposure intervention groups differed significantly according to income, a variable that was also related to the outcomes. Those with per capita incomes of at least $5000 were more likely to be in the high-exposure group (45%) than in the low-exposure group (35%). Adjustment for income weakened the intervention exposure effects on smoke alarm use, and neither income nor study group remained significant in explaining the smoke alarm outcome. Lower-income families in the high-exposure group were significantly more likely to report safe poison storage (Table 4, model 3). Those in the higher-income groups were 2 to 3 times as likely to have correct child safety seat behavior, relative to the lower-income control group (Table 4, model 3). Among the higher-income families, 53% of the high-exposure intervention group and 46% of the low- exposure intervention group reported proper child safety seat use, compared with 24% of the control group. Among the lower-income families, there was little variation across study groups, with proper child safety seat use being reported by 28% of the high-exposure intervention group, 26% of the low-exposure intervention group, and 25% of the control group.
To our knowledge, this is the first trial of a computer-tailored kiosk intervention to deliver early childhood safety education in an ED. We are encouraged by the findings that this technology is feasible for use in a busy ED that serves predominantly low-income families, many with low literacy. Families were able to complete the assessment in a short time, and virtually the entire intervention group used the kiosk-generated report; more than one half read the entire report, and two thirds discussed it with others.
We are encouraged by the significant improvements in safety knowledge that resulted from the intervention. These results also help us understand parents' existing knowledge and information gaps. For example, knowledge was high (≥85%) in both groups regarding how to transport a child safely, how many smoke alarms are needed and how to know they are working, and causes of poisoning (hair relaxer and adult medications). However, less than one third of parents knew the legal requirements for using child safety seats or the proportion of seats that are estimated to be used incorrectly. Although the intervention group was more knowledgeable than the control group regarding items related to smoke alarms and poison storage, a large proportion of respondents (≥30%) were unaware of the magnitude of house fire-related deaths and thought that keeping poisons on a high shelf was sufficient protection against childhood poisoning.
We have promising findings with respect to short-term safety behavior outcomes. The intervention group was significantly more likely than the control group to report having the correct child safety seat for their child, using it consistently, and having it inspected or installed by a car seat expert. Approximately one half of the intervention group achieved this level of protection, compared with less than one fourth of the control group. Increasing parents' correct use of child safety seats is important, given that high rates of misuse have been observed.7 Parents with per capita incomes of at least $5000 were most likely to report correct child safety seat use, which highlights the need to address financial barriers when serving lower-income families.
Neither smoke alarm use nor safe poison storage was improved significantly in the intent-to-treat analysis, although the odds ratios were positive, favoring the intervention group. In the exposure analysis, families who read all of the parent safety report and discussed it with others did benefit; they were significantly more likely to report safer smoke alarm and poison storage behaviors. However, per capita income was related significantly to intervention exposure, and adjustment for this variable weakened the impact of exposure to the intervention for smoke alarms, although the odds ratios remained positive. It may be that, because large proportions of all study groups reported proper smoke alarm use, cell sizes were too small for this more-detailed examination of exposure and income simultaneously. The high overall rates of reported smoke alarm use are not surprising, because Baltimore City Fire Department has a long-standing program of free smoke alarm distribution for residents. Interestingly, the benefit of high exposure to the intervention for poison storage was positive for both income groups but it was strongest and significant for families with lower incomes, which suggests that our recommendations for safe poison storage were feasible even for families with very limited incomes. It is important to point out that the estimates of intervention effectiveness in the exposure analysis may be conservative because we intentionally used a stringent indicator of exposure. Although there is no standard measure for exposure in studies such as ours, we reasoned that, to be fully engaged in the learning process, parents should report reading all of the report and talking about it with others. Only 39% of the sample reported doing so. It is possible that families who read and discussed only portions of the report (presumably the sections most pertinent to their needs) were also affected positively.
Our hypothesis that anxiety would interfere with learning from the report was not supported. This result could be attributable to the fact that the report could be taken home and read at the participants' leisure, when they were less stressed, or because our sample participants, on average, were not very anxious. The effect of the intervention did not vary according to the child's reason for the visit. Therefore, we did not find evidence of a teachable moment associated with having experienced an injury. Taken together, these results suggest that injury prevention programs in ED settings can be offered to the wide spectrum of families served there.
Our study adds the unique element of a computer kiosk to the small body of literature on providing early childhood injury prevention in an ED. Four studies of counseling in an ED showed promise for addressing home safety and bicycle helmet use among children and adolescents being seen for injuries and medical complaints.32–35 Our study and the only one of those studies to focus on the preschool-aged population32 provide evidence that both home safety behaviors and child safety seat use can be affected positively by intervention in an ED setting. Because ED settings often have limited personnel for patient education,36 computer kiosk interventions may have particular utility.
The use of a stage-based behavior change theory was another innovative aspect of this study that was facilitated by using a computer kiosk. The 2-step process of assessing the parents' current behavioral stage and beliefs and then generating a tailored message could be accomplished efficiently and effectively with computer technology. This study is the first, to our knowledge, to apply the PAPM to child safety and to computer-tailored interventions. The benefit of this approach was that we could address efficiently elements of parents' current safety behavior and their thoughts and beliefs about that behavior, which enhanced the persuasiveness of the message. The results presented here are those achieved 2 to 4 weeks after the intervention. Stage-based behavior change models, such as the one used here, describe behavior change as a process that occurs over time. Therefore, our follow-up period might have been too short for some parents to have taken action on the information we provided. Future analyses of 4-month follow-up interviews should shed light on this question and allow us to test elements of the PAPM more thoroughly, because we will be able to examine predictors of change over time.
A strength of the study is its randomized design. We also examined both overall effects of the intervention and effects among those who used the report fully. We sampled during the busiest shifts over a 14-month period, which helped ensure that we obtained a reasonable cross-section of families and typically seen conditions. We were able to accrue and to monitor a large sample of families representing a high-priority audience for injury prevention. The generalizability of our results should be restricted to similar types of EDs serving urban, low-income families.
The 2 study groups spent the same amounts of time interacting with the computer kiosk, which is a strength of the design. However, we compared generic non–safety-related messages and tailored safety messages, which maximizes the likelihood of finding differences between the 2 groups. Although we cannot comment on the relative advantage of tailored versus generic safety materials, the health communication literature is replete with data supporting the advantages of tailoring for messages focused on other health issues.26,27
A limitation of potential concern for this study is the use of self-reported outcome data, although this is the state of the art in similar studies.32–35 Although self-report data have been found to overestimate some safety behaviors,37,38 one study found that self-reports of owning a car seat were reliable.38 Another study in the same community as the current work found no differential overreporting in a comparison of intervention and control groups.37 Therefore, although our actual rates may be inflated, the difference between the 2 study groups may be accurate.
Our results, coupled with the small body of similar research in other pediatric settings, should generate enthusiasm for innovative communication strategies to promote child safety. Our finding of a different impact on child safety seat use according to family income suggests that reducing financial barriers to safety behaviors should be a high priority. The fact that low-income families seemed to benefit from the safe poison storage messages is encouraging and demonstrates that financial barriers may not be as relevant a concern for less-complex or -expensive safety behaviors. The positive results we found with minimal intrusion into patient flow or health care provider time bodes well for widespread applicability of computer technology to patient education in busy EDs.
This research was funded by grant 5RO1 HD042777-03 to Johns Hopkins University from the National Institute of Child Health and Human Development and by a subcontract to the Health Communication Research Laboratory at St Louis University.
We thank the following individuals who contributed to this project: Heather Jacobsen, Matthew Kreuter, and Eric Wheetley at St Louis University; Kari Burgess, Arnell Carr, and Akisha Price at Johns Hopkins University; and Neil Weinstein at Rutgers University. We also thank the ED staff members and patients' parents who participated in this study.
- Accepted March 19, 2007.
- Address correspondence to Andrea Carlson Gielen, ScD, ScM, Center for Injury Research and Policy, Johns Hopkins Bloomberg School of Public Health, 624 N Broadway, Baltimore, MD 21205. E-mail:
Portions of this work were presented at the annual meeting of the American Public Health Association; November 8, 2006; Boston, MA.
Dr McKenzie's current affiliation is Center for Injury Research and Policy, Columbus Children's Research Institute, Columbus Children's Hospital and Ohio State University College of Medicine, Columbus, OH.
The authors have indicated they have no financial relationships relevant to this article to disclose.
- ↵Centers for Disease Control and Prevention. WISQARS injury mortality report. Available at: www.cdc.gov/ncipc/wisqars/default.htm. Accessed July 22, 2006
- ↵Finkelstein EA, Corso PS, Miller TR. The Incidence and Economic Burden of Injuries in the United States. New York, NY: Oxford University Press; 2006:161
- ↵National Highway Traffic Safety Administration. Motor vehicle occupant protection facts. Available at: www.nhtsa.dot.gov/people/injury/airbags/OccupantProtectionFacts/index.htm. Accessed June 4, 2007
- ↵American Academy of Pediatrics. The Injury Prevention Program (TIPP). Chicago, IL: American Academy of Pediatrics; 1994
- ↵National Highway Traffic Safety Administration. Misuse of Child Restraint. Washington, DC: National Highway Traffic Safety Administration; 2005. Publication DOT HS 809 851
- McDonald E, Girasek D, Gielen AC. Home injuries. In: Liller K, ed. Injury Prevention for Children and Adolescents: Research, Practice, and Advocacy. Washington, DC: American Public Health Association; 2006:123– 162
- ↵Runyan CW, Casteel C. Unintentional home injury death. In: Runyan CW, Casteel C, eds. The State of Home Safety in America: Facts About Unintentional Injuries in the Home. 2nd ed. Washington, DC: Home Safety Council; 2004:21– 24
- ↵McCaig LF, Nawar EW. National Hospital Ambulatory Medical Care Survey: 2004 emergency department summary. Adv Data.2006;(372) :1– 29
- Revere D, Dunbar PJ. Review of computer-generated outpatient health behavior interventions: clinical encounters “in absentia. ” J Am Med Inform Assoc.2001;8 :62– 79
- ↵Parkinson GW, Gordon KE, Camfield CS, Fitzpatrick EA. Anxiety in parents of young febrile children in a pediatric emergency department: why is it elevated? Clin Pediatr (Phila).1999;38 :219– 226
- ↵Perloff RM. The Dynamics of Persuasion. Hillsdale, NJ: Lawrence Erlbaum Associates; 1993
- ↵Fleiss JL. Statistical Methods for Rates and Proportions. 2nd ed. New York, NY: Wiley-Interscience; 1981
- ↵Petty RT, Cacioppo JT. Attitudes and Persuasion: Classic and Contemporary Approaches. Boulder, CO: Westview Press; 1996
- ↵Krueter MW, Jacobsen HA, McDonald EM, Gielen AC. Developing computerized tailored health messages. In: Bensley RJ, Brookins-Fisher J, eds. Community Health Education Methods: A Practical Guide. 2nd ed. Sudbury, MA: Jones and Bartlett Publishers; 2003:255– 289
- ↵Doak C, Doak L, Root J. Teaching Patients With Low Literacy Skills. 2nd ed. Philadelphia, PA: Lippincott Williams & Wilkins; 1996
- ↵Spielberger CD. The State-Trait Anxiety Inventory. Palo Alto, CA: Consulting Psychologists Press; 1970
- ↵Posner JC, Hawkins LA, Garcia-Espana F, Durbin DR. A randomized, clinical trial of a home safety intervention based in an emergency department setting. Pediatrics.2004;113 :1603– 1608
- Johnston BD, Rivara FP, Droesch RM, Dunn C, Copass MK. Behavior change counseling in the emergency department to reduce injury risk: a randomized, controlled trial. Pediatrics.2002;110 :267– 274
- ↵Claudius IA, Nager AL. The utility of safety counseling in a pediatric emergency department. Pediatrics.2005;115(4) . Available at: www.pediatrics.org/cgi/content/full/115/4/e423
- ↵Chen L, Gielen AC, McDonald EM. Validity of self-reported home safety practices. Inj Prev.2003;9 :73– 75
- ↵Robertson AS, Rivara FP, Ebel BE, Lymp JF, Christakis DA. Validation of parent self-reported home safety practices. Inj Prev.2005;11 :209– 212
- Copyright © 2007 by the American Academy of Pediatrics