Objective. To determine if hands-on instruction in child safety seat (CSS) installation decreases the number of errors in installation.
Design. Cross-sectional study.
Setting. Primary care offices, emergency department, CSS checkpoint.
Participants. Parents of children <2 years old receiving medical care or attending a CSS check.
Main Outcome Measure. Errors in CSS use.
Results. Only 6.4% of parents had a correctly installed CSS. Hands-on instruction was associated with fewer errors in seat installation. Increased parent age, completion of college, and having private insurance were also associated with fewer errors in CSS placement. The majority of parents learned to install seats from reading the manual, from friends and relatives, and from figuring it out on their own.
Conclusions. Errors in CSS installation are a significant problem. Hands-on instruction decreases the numbers of errors in CSS installation. However, few parents receive hands-on instruction from experts in CSS installation. Increases in correct CSS use could result from hands-on education by trained professionals.
Motor vehicle crashes are a significant source of morbidity and mortality among children <1 year old, although proper use of child safety seats (CSS) can prevent many injuries. Use of CSS reduces the likelihood of fatal injury followed a crash by 69% in infants, and the number of injuries sustained in children <4 years old by 60%.1–3 A 36% decrease in hospitalizations, 25% decrease in head injuries, and 20% decline in extremity injury resulting from motor vehicle crashes followed the implementation of a child restraint law for children <4 years old in Michigan.4
Despite the clear benefits of CSS use and laws requiring children to be restrained, only 55% to 80% of children <4 years old are restrained in CSS.3,5–8 Infants are more likely to be restrained than older children; studies have shown restraint use ranging from 76% to 96.6% in children <1 year old.1,3,9 The majority of children who are placed in CSS, however, are not properly restrained, with reported misuse rates ranging from 63% to 88.5%.5,7,8 Improper CSS use has been associated with a number of injuries in children, including head injuries, fractures, peripheral neuropathy, liver laceration, and cervical spine injuries.10–13
Given the high rate of CSS misuse and the associated risk of injury, as well as the large variety of CSS and automobiles available for purchase, each requiring slightly different methods of installation, parents should receive adequate training to install CSS properly. The complexity of CSS use makes it difficult for parents, in the absence of expert instruction, to use CSS properly. However, few studies have examined how and where parents learn about CSS use, and whether hands-on training is more effective than nondemonstrated educational methods. Ruffin and Kantor14 determined that prenatal instructors and manufacturers' information were the best sources of general knowledge about CSS use, but ability to install CSS using this knowledge was not evaluated. Other data suggest that most parents learn how to install CSS from reading the manual or from speaking with friends or relatives, but few learn through demonstration.7,8
Whether hands-on instruction in CSS installation leads to fewer errors has not been evaluated, although the American Academy of Pediatrics' (AAP) Committee on Accident and Poison Prevention and others recommend this form of instruction.7,15 This study examined how parents learn to install CSS and whether demonstrated instruction leads to fewer errors in installation. The study also explored specific errors in CSS installation and other factors associated with improper CSS use.
Participants were recruited from 4 separate locations in Chapel Hill and Raleigh, North Carolina, between June 8 and July 30, 1998. These sites included 2 primary care pediatric practices: a combined faculty and resident practice affiliated with the University of North Carolina Hospitals and a private practice not affiliated with the medical center. Participants were also recruited in Raleigh at a CSS check sponsored by SafeKids, a national injury prevention coalition. This CSS check was located at a local battery manufacturing plant, and participants included both community members and employees. The Pediatric Emergency Department at the University of North Carolina served as an additional site for subject recruitment.
Parents who agreed to participate and signed a statement of informed consent received a survey containing questions about how, when, and where they learned how to install their child's safety seat, whether they received hands-on instruction, and whether they felt that their training was adequate. The survey also requested demographic information, including parent age, child age, number of other children, health insurance status, education, and race. Survey questions related to car seats are included in the Appendix.
After completing the survey, parents were accompanied to their cars, and CSS installation was evaluated by 1 of 3 National Highway Traffic Safety Administration (NHTSA)-trained observers. Errors in installation were recorded on 1 of 2 forms that contained similar assessment criteria. These documents were developed by the North Carolina Highway Safety Research Institute and the National SafeKids Campaign. To create a reasonable number of error categories, data from both assessment sheets were collapsed into 12 broad categories of errors. Both documents also requested information regarding manufacturer certification of the safety seat and recall status. Recall status was evaluated using a list published by NHTSA available on the Internet and updated through March 1998.16
All parents of children <2 years old presenting to 1 of the above sites during the study period, when researchers were available, were eligible to participate. An eligible participant was excluded if she left a site before a previous CSS check was completed (ie, no researcher was available to examine the safety seat). Parents of children requiring specialized safety seats and parents who had traveled to the site via public transportation were also excluded. A total of 115 participants were enrolled. Six parents were later excluded for improper completion of the survey, leaving 109 participants in the final data analysis. Approximately 10 parents refused to participate in the project. All but 2 cited lack of time as the reason for refusal. One parent cited lack of a CSS. One mother stated that she knew that her seat was installed incorrectly, but she was not interested in moving her safety seat from the car's front seat to the back seat.
Approval for this project was obtained through the Institutional Review Board of the University of North Carolina School of Medicine and Hospitals.
Degree of CSS misuse was measured by the number of errors in CSS installation. Bivariate relationships were examined between degree of CSS misuse and each independent variable using t tests and 1-way analysis of variance. Pearson correlation coefficients were used to look for collinearity between variables. Bivariate analyses were followed by multivariate analyses using a linear regression model to estimate adjusted mean differences in continuous variables. Independent variables found to be nonsignificant in the overall regression model were excluded. Collinearity was addressed by excluding all but 1 collinear variable from the model and developing separate models to evaluate the relationship between the degree of CSS misuse and any collinear variables.
All data were analyzed using STATA statistical software.17
One hundred nine patients were included in the data analysis. Demographic characteristics of the study population are listed in Table 1. Mean parent age was 30.1 years with a range of 18 to 49 years. Mean child age was 8.7 months with a range of 0 to 24 months. Sixty-five percent of children were <1 year old. Thirty-six percent of children were insured by Medicaid. More than 75% of parents had received education beyond high school. The racial background of participants (white vs non-white) reflected that of the counties from which they were recruited, based on data from July 1997.18 Rates of Medicaid insurance reflected statewide data from 1993.19 The study population was more highly educated than the general population of the counties from which they were recruited.20
All 109 participants had a CSS in the car at the time of the study. However, only 7 of the 109 participants (6.4%) had a correctly installed CSS. Approximately 84% of parents had between 1 and 3 errors in CSS installation. Table 2 indicates specific problems with installation. A total of 219 problems were found, or 2 errors, on average, per child. The average number of mistakes did not differ by age of the child (<12 months vs >12 months). The most common problems in CSS installation for children of all ages were harness straps being too loose, safety belts being too loose, and either failure to use or improper use of locking clips. Only 1 seat was placed in front of a passenger air bag. Ninety-three percent of CSS had been certified by the manufacturer, meaning that a label was present on the CSS, containing the name of the manufacturer, the model number, and the date of production. Ten safety seats had been recalled. Eleven seats had been used previously by another family. No parent reported that her CSS had been involved in a crash.
Table 3 indicates how, when, and where parents learned to install their safety seats. The majority of parents (51.4%) learned how to install safety seats by reading the CSS manual. Nearly 25% of parents learned how to install the seat on their own, and 15.6% were taught by a friend or relative. Less than 5% of parents reported being taught by a health care worker. Most learning took place at home (75.2%). The majority of parents learned about the CSS before taking their newborn home from the hospital. Over 25% of parents reported having had hands-on instruction when learning how to use a CSS, meaning that someone showed the parent how to install the safety seat in the parent's own car. Twelve parents learned from a friend or relative and 3 parents learned from a health care worker. Hands-on training was associated with an increased likelihood of having the CSS installed correctly, with no errors (P = .02). There was no difference between any of the specific types of errors in CSS installation and whether or not a parent had received hands-on instruction.
Bivariate analyses indicated a number of factors associated with degree of CSS misuse. Hands-on training in safety seat installation was associated with fewer errors in use (hands-on instruction, 1.46 errors; no hands-on instruction, 2.15 errors; P = .005). Children insured privately had fewer mistakes in installation than children insured by Medicaid (private insurance, 1.68 errors; Medicaid, 2.54 errors; P = .001). Increased parental education was also associated with fewer problems in CSS placement (college degree, 1.75 errors; no college degree, 2.38 errors; P= .01). Degree of CSS misuse was not associated with the age of the child, the frequency of removal of the CSS from the car, or parental assessment of the adequacy of training to install the safety seat. Neither how (read manual, taught by friend/relative, taught by health care worker, never taught), where, or when parents learned to install CSS was associated with degree of CSS misuse. Ethnic/racial background was also not associated with degree of misuse.
As noted above, separate models were developed to address collinearity among several variables. Race was retained as a covariate in all models because in the multivariate analysis Asians were found to have significantly fewer errors in safety seat installation as compared with whites. Hands-on training was associated with fewer errors in CSS installation in 2 separate models. The first model used parent age, insurance status, and race as confounders; the second used parent age, education, and race as confounders. Results of selected models are presented in Table 4. Parents receiving hands-on instruction in CSS installation had, on average, 1.34 mistakes in installation while parents without hands-on instruction had, on average, 2.19 errors, after adjusting for parent age, insurance status, and race. Insurance status was associated with fewer problems with safety seat placement, after adjustment for parent age, hands-on instruction, and race. Parents of children with private insurance had, on average, 1.67 errors in CSS installation, while parents of children receiving Medicaid had, on average, 2.55 mistakes in safety seat placement. Increased education was also associated with fewer problems with CSS installation after adjustment for hands-on instruction and race. Parents who had finished college had, on average, 1.71 mistakes in CSS placement, while parents who had not graduated from college had, on average, 2.39 errors in CSS installation. All of the above comparisons were statistically significant (P < .05). Finally, increased parent age was associated with fewer problems with CSS installation after adjustment for hands-on instruction and race.
This study demonstrates that hands-on instruction is associated with a decrease in the number of errors in CSS installation by approximately 1 error, on average, as compared with teaching without hands-on instruction. These results suggest that organizations such as the AAP should continue to encourage hands-on education to install CSS.
Unfortunately, although parents who received hands-on instruction had fewer errors in CSS installation, the fact that only 6.4% of CSS in this study were installed correctly is alarming. This small number of correctly installed seats may be associated with the methods by which most participants learned to install their CSS. Most participants learned about CSS by reading the safety seat manual, suggesting that reading the manual alone is insufficient to ensure proper seat installation. Nearly 25% of parents never received any instruction. Over 15% of parents were taught by friends or relatives, persons who are unlikely to be experts on CSS. Few parents received instruction from anyone who might be considered an expert in CSS installation. While 5 parents did receive instruction from health care workers, only 2 received hands-on instruction, and it is not clear whether these health care providers were knowledgeable about CSS installation. Because of the complexity of safety seat installation, instruction by experts may be particularly important. Had parents received hands-on instruction from persons trained in CSS installation, the mean number of errors in CSS placement may have dropped sharply, with a larger proportion of safety seats installed correctly.
The fact that all parents in the study did have CSS is encouraging. However, as participation was voluntary, parents without CSS may have chosen not to participate in the project or did not frequent the study sites. Parents may also be more likely to use a CSS when visiting the doctor or attending a voluntary CSS check than on other trips. Though usage rates as high as 96.6% in infants have been reported, rates for toddlers have tended to be much lower.9
Several other problems may have affected the validity of the data. First, some parents may have had difficulty remembering how they were taught to install their safety seats. This problem may have been more significant for parents of older children, who may not have received any education about CSS after their children were no longer infants. We attempted to reduce difficulties with recall by limiting the sample to children <2 years old. Results of the study may also have been influenced by how we chose to ask about CSS training. Parents were asked to recall their most recent experience with CSS education because it was assumed that current CSS installation techniques would be influenced primarily by the most recent CSS training. However, earlier educational interventions may also have affected installation techniques. We also did not ask about use of the car owner's manual in learning how to use the safety seat, but parents could have listed this method under “other.” A small number of parents may have misinterpreted the meaning of hands-on education, as 7 parents reporting hands-on teaching stated that they learned how to install their CSS either on their own or by reading the manual. Also, the fact that only 7 parents had correctly installed CSS made it impossible to determine the factors associated with proper safety seat use (ie, no errors). Finally, parents attending safety seat checks are a self-selected population and may be different from other populations in ways we did not or could not measure. The numbers at each site were too small to perform separate analyses at each study site.
Despite these limitations, this study reinforces the fact that misuse of CSS remains a significant problem. Parents continue to have difficulties both in placing their children in CSS (eg, not fully tightening harness straps) and placing the seat in the car (eg, not fully tightening seat belts and not using locking clips). NHTSA has begun to address the issue of correct placement of the seat in the car by moving to require universal child restraint anchorage systems.21 Although this approach may help, it will not address problems associated with placement of the child in the safety seat, and it may not guarantee proper usage if parents are left to their own devices in learning to install CSS.
Earlier research has shown that demonstration of correct CSS use is associated with increased percentages of parents using CSS and may increase correct use.22,23 This paper indicates that hands-on instruction may reduce the number of errors in CSS usage. Hands-on education by trained professionals, such as newborn nursery nurses or other health care or highway safety workers could be effective in ensuring that infants and toddlers are properly and safely restrained in automobiles. Therefore, we reinforce the AAP recommendation that all parents receive hands-on education by NHTSA-trained professionals to properly install CSS. Additionally, harder-to-reach populations, such as Medicaid recipients, should be targeted for enhanced intervention, as these persons may be at higher risk for incorrectly installed CSS.
We would like to thank Jonathan Kotch, MD, MPH, for his assistance with project development, data analysis, and manuscript critique, and Lewis Margolis, MD, MPH, for his review and critique of the manuscript.
- Received December 29, 1999.
- Accepted May 24, 2000.
- Address correspondence to Wendy G. Lane, MD, MPH, Children's Hospital of Philadelphia, Division of General Pediatrics, 34th St and Civic Center Blvd, Philadelphia, PA 19104.
Presented at the Ambulatory Pediatric Association meeting, May 3, 1999; San Francisco, CA.
- CSS =
- child safety seats •
- AAP =
- American Academy of Pediatrics •
- NHTSA =
- National Highway Traffic Safety Administration
- National Highway Traffic Safety Administration. Traffic Safety Facts 1996: Children. Washington, DC: US Department of Transportation; 1997. Available on-line at: http://www.nhtsa.dot.gov Accessed May 19, 1998
- Johnston C,
- Rivara FP,
- Soderberg R
- Eby DW,
- Kostyniuk LP,
- Christoff C
- Bull MJ,
- Stroup KB,
- Gerhart S
- Fuchs S,
- Barthel MJ,
- Flannery AM,
- Christoffel KK
- American Academy of Pediatrics, Committee on Injury and Poison Prevention
- The National Highway Traffic Safety Administration. The National Highway Traffic Safety Administration's Child Seat Safety Recall Campaign Listing January 1988 Through Mid-March 1998. Washington, DC: US Department of Transportation. No date available. Available on-line at: http://www.nhtsa.dot.gov/people/injury/childps/recall/cssinter.html Accessed June 7, 1998
- StataCorp. Stata Statistical Software: Release 5.0. College Station, TX: Stata Corporation; 1997
- Office of State Planning. State Demographics. Raleigh, NC: Office of State Planning. No date available. Available on-line at: http://www.ospl.state.nc.us/demog/ Accessed November 29, 1998
- Office of State Planning. Current Population Survey, NC Health Insurance Coverage, 1993. April, 1996. Raleigh, NC: Office of State Planning; Available on-line at: http://www.ospl.state.nc.us/sdn/hlthin93.html Accessed November 29, 1998
- US Census Bureau. 1990 US Census Data. Washington, DC: US Census Bureau; No date available. Available on-line at: http://www.census.gov/ Accessed November 23, 1998
- Mathews AW, White JB. Car makers face new rules to make kid seats safer. Wall Street Journal. November 6, 1998:Section B:1(Column 1)
- Allen DB,
- Bergman AB
- Berger LR,
- Saunders S,
- Armitage K,
- Schauer L
- Copyright © 2000 American Academy of Pediatrics