An Intervention to Reduce Television Viewing by Preschool Children

Background  Television viewing has been associated with increased violence in play and higher rates of obesity. Although there are interventions to reduce television viewing by school-aged children, there are none for younger children.

Objective  To develop and evaluate an intervention to reduce television viewing by preschool children.

Design  Randomized controlled trial conducted in 16 preschool and/or day care centers in rural upstate New York.

Patients  Children aged 2.6 through 5.5 years.

Intervention  Children attending intervention centers received a 7-session program designed to reduce television viewing as part of a health promotion curriculum, whereas children attending the control centers received a safety and injury prevention program.

Outcome Measurements  Change in parent-reported child television/video viewing and measured growth variables.

Results  Before the intervention, the intervention and control groups viewed 11.9 and 14.0 h/wk of television/videos, respectively. Afterward, children in the intervention group decreased their television/video viewing 3.1 h/wk, whereas children in the control group increased their viewing by 1.6 h/wk, for an adjusted difference between the groups of −4.7 h/wk (95% confidence interval, −8.4 to −1.0 h/wk; P = .02). The percentage of children watching television/videos more than 2 h/d also decreased significantly from 33% to 18% among the intervention group, compared with an increase of 41% to 47% among the control group, for a difference of −21.5% (95% confidence interval, −42.5% to −0.5%; P = .046). There were no statistically significant differences in children's growth between groups.

Conclusions  This study is the first to show that a preschool-based intervention can lead to reductions in young children's television/video viewing. Further research is needed to determine the long-term effects associated with reductions in young children's television viewing.

TELEVISION VIEWING has increased significantly during the past 20 years, in part because of increased program development, availability, and marketing.¹ Television viewing by children has adverse effects, including attenuated social behavior,^2,3 poor school performance,^4,5 higher rates of violence,^6,7 and increased rates of child obesity.^8-11 Increased viewing hours might lead to increased adiposity because of increased snacking in response to the many food-related commercials¹² and/or to reduced energy expenditure by displacing more physically active behaviors.¹¹

Because of the adverse effects associated with television viewing, several school-based studies have focused on reducing children's viewing. A study that reduced television viewing among sixth through eighth grade students reported a reduction in obesity rates among girls, but not boys,¹³ whereas a reduction in television viewing and meals eaten in front of the television by third and fourth graders was associated with reductions in adiposity measures¹⁴ and relative declines in aggression¹⁵ among both boys and girls. Clinical trials with overweight children aged 8 through 12 years have shown that behavioral interventions targeting sedentary behaviors, specifically television viewing, lead to reductions in adiposity.¹⁶

We know of no interventions focusing on reduction of television viewing by preschool children. However, television viewing hours increase during the preschool years, and habits of relatively long or short viewing begin to develop during these early years.^8,17 Because of concerns about the adverse health effects of television, the Committee on Public Education of the American Academy of Pediatrics has cautioned parents to limit children's exposure to television and other media to a maximum of 2 h/d. For children younger than 2 years, they completely discourage television viewing.¹⁸ As women increasingly work outside the home, the percentage of preschool children being cared for outside the home has also increased. Thus, child care settings offer a unique and emerging opportunity to promote the development of healthy lifestyle behaviors such as limited television viewing.

The purpose of this study was to develop and then implement an intervention in the child care setting to reduce children's television viewing. This report describes the development, implementation, and impact of this intervention on children's television viewing and growth variables.

Preschool and day care centers located within a 45-mile radius of Cooperstown, NY, but excluding Cooperstown, that enrolled children aged 3 through 5 years were invited to participate in the Brocodile the Crocodile health promotion program. Twenty centers were identified, and directors at 19 centers agreed to participate in this study. Before randomization, 1 day care center was excluded to avoid contamination, as several children who attended this center also attended 1 of 3 other preschools participating in the study.

Randomization used the day care or preschool center as the unit of analysis. Before randomization, the centers were stratified by day care (n = 5) or preschool (n = 13) and by mean child age. Randomization was performed in random permutations of the numbers 1 and 2 such that assignment of one center to a group arbitrarily forced its nearest neighbor (based on child age and type of child care center) into the opposite group. After randomization, a preschool center originally randomized to the intervention was dropped because most of the recruited children attended only 3 d/wk, but not on the day of the week when we were able to provide the intervention. Before the beginning of the second school year, another center was dropped for the same reason. Thus, the final sample available for analysis included 8 centers in the intervention group and 8 centers in the control group.

Program staff (an early childhood teacher and a music teacher) visited each day care or preschool center in the intervention group once a week to provide a 1-hour session for a total of 39 weeks. Half of each session was spent in musical activities; 10 minutes, eating a snack; and 20 minutes participating in an interactive educational session. The Brocodile the Crocodile health promotion curriculum consisted of 32 sessions (10 in the spring of the first school year and 22 during the second school year) devoted to healthy eating and 7 sessions designed to reduce children's television viewing (provided during the second school year). The development and impact of these 7 sessions are the subject of this report.

Children aged 2.5 through 5.5 years who attended a participating preschool or day care center were eligible to participate in this study. The study was explained to parents, who gave written, informed consent for their child to participate. Parents were free to withdraw their children from the study at any time. The study was approved by the Institutional Review Board of the Mary Imogene Bassett Hospital,Cooperstown, NY; the boards and directors of the participating preschool and day care centers; and the institutional review boards at the Oneonta and Cobleskill branches of the State University of New York, each of which had a participating preschool or day care center.

Because the intervention to reduce television viewing was implemented in the second year of the study, and because there was considerable student turnover between year 1 and year 2, these analyses relate to changes during this period (fall 2000 through spring 2001). Before the television reduction intervention, the parents of all children involved in the Brocodile the Crocodile program were mailed questionnaires (September 2000), with a second mailing a month later to nonrespondents. The parents or guardians reported their relation to the child, the child's date of birth, sex, race or ethnicity, and maternal and paternal educational attainment. The child's mother and father were asked to report the average amount of time (hours per week) that each spent during the past week watching television or videos, playing video or computer games, or surfing the Internet. They were also asked to estimate for Saturday, Sunday, and an average weekday the number of hours the child spent during the past week viewing television/videos and the number of hours the child had spent playing computer/video games. The child's weekly television/video viewing was compared by multiplying weekday viewing hours by 5 and adding viewing hours for Saturday and Sunday. Weekly computer/video game playing was calculated in the same manner. The Committee on Public Education of the American Academy of Pediatrics recommends that children 2 years and older watch no more than 2 h/d of television, videos, or other media.¹⁸ Thus, children were dichotomized into the following 2 groups on the basis of how many hours of television/videos they watched: up to and including 2 h/d or more than 2 h/d.

Parents reported whether their child had a television set in his or her bedroom, the number of days that the child ate dinner with the television turned on, and the number of days that the family ate dinner together (0-7 days). They also reported, using a 5-point Likert scale (always, usually, sometimes, rarely, or never), the frequency that they and the child ate or snacked while watching television.

At the end of the intervention (spring 2001), the parents completed a shorter questionnaire, in writing, or, for those who failed to respond, by telephone. They used the same format to report the child's television viewing behaviors and the children's television/video viewing and computer/video game playing.

The children's growth was measured during fall 2000, winter 2001, and spring 2001. Child height, in stocking feet, was measured using a portable stadiometer. Child weight, lightly clothed, was measured using a digital scale (model 781; Seca Uniscale, Hamburg, Germany). All measurements were made twice, and the means of the 2 measurements were used in all subsequent analyses. A measurer and an assistant were used to make all measurements according to the guidelines from the World Health Organization.¹⁹ The triceps skin-fold thickness of the left arm was measured in triplicate, according to the protocols of Gibson.²⁰ The mean of the 3 measurements was used in all analyses. Test-retest reliabilities were high for all 3 measurements (ρ>0.99 for height, ρ>0.99 for weight, and ρ = 0.96 for triceps skin-fold thickness). Body mass index (BMI) was computed as weight in kilograms divided by the square of height in meters. Triceps skin-fold thickness and BMI were used as measurements of adiposity, which were significantly correlated with each other (ρ = 0.58). For each child, the age- and sex-standardized BMI z score was abstracted from the Centers for Disease Control and Prevention growth charts.²¹

All data were analyzed using Statistical Analysis Software (Version 8.0; SAS Institute, Cary, NC). All statistical tests were 2-sided; P values >.05 were considered nonsignificant. For all mean differences between the intervention and control groups, 95% confidence intervals (CIs) were calculated.

The characteristics of children and parents in the intervention group were compared with those in the control group. To account for the randomization at the level of the preschool or day care center, the mean values of continuous variables computed at each center and weighted by the sample size were compared between the intervention and control groups by means of 2-group analysis of variance. We used χ² tests to compare child sex and race distributions and Mantel-Haenszel χ² tests to compare distributions of ordered variables, such as maternal educational attainment, between intervention and control groups. The characteristics of parents and children for whom follow-up data were not available were compared with those with follow-up data as above.

For each child, the differences before and after the intervention in the number of hours per week of television/videos viewed and the number of hours per week of computer/video games played were computed (ie, the difference between spring 2001 and fall 2000). The mean values, weighted by sample size at each center, were compared between the intervention and control groups by means of 2-group analysis of covariance, controlling for child age, sex, and preintervention value. In a similar manner, differences in the percentage of children watching 2 or more h/d of television/videos, the number of days per week the child ate dinner with the television turned on, and the number of days per week the family ate dinner together were compared.

The changes in children's growth (weight, height, and triceps skin-fold thickness) were computed between fall 2000 and spring 2001 as the change per exact time interval (eg, killigrams per year) for each child by means of regression analysis, weighted by the number of measurements, and the mean slope for each center was computed. The mean values, weighted by sample size at each center, were compared between intervention and control groups using 2-group analysis of covariance, with preintervention values added as covariates. Changes in growth during the entire study were computed between spring 2000 and spring 2001 in a similar manner.

To increase awareness of their children's television and video viewing, parents were asked to keep a diary, where they recorded the name and amount of time of each television program and videotape that their children watched for 1 week. The diary and instructions were mailed in early December 2000, with return by prepaid mail.

Each of the 7 sessions included components for the child, the day care or preschool provider, and the parents as described in Table 1. The weekly 20-minute interactive, educational sessions were led by program staff, but day care and preschool staff were actively encouraged to participate. To reinforce the program goals, additional materials and suggested classroom activities were provided to day care and preschool staff. After each class, materials and activities were sent home with each child to foster discussion between parents and children. Children were also encouraged to ask their parents for help in completing these activities at home.

Materials for day care and preschool staff and parents presented positive aspects of alternative activities to television viewing, such as improved literacy skills from reading books to young children²² and enhanced social skills from eating meals together as a family. The children were active participants, generating a list of activities that they enjoyed other than watching television. The children's book The Berenstain Bears and Too Much TV²³ was used as an impetus for discussions at the child care centers and at home. Day care and preschool staff and parents were provided with the list of alternative activities generated by the children and the brochure Television and the Family by the American Academy of Pediatrics.²⁴ The children also made "no television" signs, which were sent home with a weekly calendar and "no television" stickers with instructions for parents to monitor and reward children by placing a sticker on the calendar every day the child refrained from watching television or video shows. Materials from the National TV-Turnoff Week Internet site were also used.²⁵

At the centers randomized to the control group, curriculum, materials, and ideas for activities about health and safety were provided to day care or preschool staff and information and materials for at-home activities were mailed to parents. Eight monthly sessions, each with a different health or safety topic appropriate for the season, were provided during the second school year. Some sessions were adopted from the Healthy Start Health Education Program.²⁶ Where possible, additional community resources were used to enhance what was provided to children at the day care or preschool centers. Examples include working with local fire departments to arrange for a fireman to visit each child care center or to have the children visit the fire house during Fire Prevention Week and arranging for a dental hygienist to visit each center as part of the dental session.

Prior to the television reduction intervention (fall 2000), complete child anthropometric and parental questionnaires were available for 102 children. Complete follow-up data were available for 77 of these children (spring 2001). As shown in Figure 1, the turnover of children attending these preschool and day care centers was considerable, especially between school years. Most children who did not continue in the study left the preschool or day care center to enter prekindergarten or kindergarten, or the family moved out of the area.

At the first assessment (data now shown) and the second assessment (Table 2) the intervention and control groups were comparable. The children were primarily white. They were similar in age, attended day care or preschool a comparable number of hours per week, and had similar anthropometric measurements. Like other US children, their mean adiposity exceeded the mean of the reference population (ie, BMI z score, >0). The percentages of children with a BMI in at least the 95th and at least the 85th age- and sex-specific percentiles in this study did not differ significantly from that reported for non-Hispanic white children aged 2 through 5 years in the 1999-2000 National Health and Nutrition Examination Survey (4.1% vs 10.1% and 17.6% vs 20.5%, respectively).²⁷ The parents were well educated, with about half having a college degree or higher.

Before the intervention, children in the intervention and control groups spent a comparable number of hours viewing television/videos and playing video/computer games. Boys and girls viewed a similar number of hours of television/videos (12.5 and 13.7 h/wk, respectively) and played video/computer games for a comparable number of hours (2.1 and 2.2 h/wk, respectively). Only 1 (2.4%) of 41 children and 3 (8.8%) of 34 children in the intervention and control groups, respectively, had a television set in their bedroom. The number of hours per week that television/videos were viewed was not significantly different for children with a television set in their bedrooms, compared with those without (14.5 vs 13.1 h/wk; P = .74). However, children with a television set in their bedrooms played 3 times as many hours per day of video/computer games on Sundays than those without (1.0 vs 0.3 h/d; P = .03). The frequency with which children snacked while watching television was significantly associated with the frequency that their parents snacked (Mantel-Haensel χ² = 11.2; P<.001). The demographic, anthropometric, and television/video viewing behaviors of the children did not differ between those for whom we had follow-up data and those for whom we did not (data not shown).

Changes in the children's television/video viewing are shown for those with follow-up data (n = 77) in Table 3. Before the television reduction intervention, children in the 2 groups viewed a similar number of hours per day of television/video on weekdays, Saturday, and Sunday, but after the intervention, they diverged. Significant differences were seen in the changes between the intervention and control groups for television/video viewing on weekdays and on Sunday (Table 3). Changes in Saturday viewing were less, but in the same direction. The mean weekly television/video viewing decreased 3.1 h/wk among children in the intervention group compared with an increase of 1.6 h/wk among those in the control group (adjusted difference between means, −4.7 h/wk; 95% CI, −8.4 to −1.0 h/wk; P = .02) (Figure 2).

Before the intervention, comparable percentages of children in the intervention and control groups watched, on average, more than 2 h/d of television/videos (35% [15/43] and 41% [14/34], respectively; P = .57, χ²). After the intervention, the percentage of children watching more than 2 h/d television/videos was significantly lower among the intervention group compared with the control group (19% [8/43] vs 41% [16/34], respectively; P = .007), with an adjusted difference in the percentages between groups of −21.5% (95% CI, −42.5% to −0.5%; P = .046). The change in the children's television/video viewing did not differ by child sex, paternal or maternal educational attainment, or the number of hours per week the child attended day care or preschool.

The mean changes in the amount of time the children played video/computer games did not differ between the intervention and control groups for weekdays, Saturday, or Sunday.

Table 3 shows the mean changes in the children's anthropometric measures between the intervention and the control groups. As expected, children in the intervention and control groups grew taller (6.5 and 7.3 cm/y, respectively) and gained weight (1.9 and 2.6 kg/y, respectively), while their triceps skin-fold thickness decreased (2.2 and 1.8 mm/y, respectively). Although the BMI decreased among the intervention group and increased among the control group, none of the changes in growth measurements between the 2 groups were significantly different.

Changes in the children's growth measures during the entire study period revealed no significant differences between the 2 groups. There were also no significant changes or differences between the intervention and control groups in the frequency that children snacked while watching television or the number of days the family ate dinner together or watched television during dinner (data not shown).

This report describes the first intervention of which we are aware that targets reductions in television viewing by preschool children. Before the intervention, the children viewed a similar number of hours as those reported in a national study,²⁸ but fewer hours than those reported in a study of children with less educated parents.⁸ This is consistent with previous findings that children with more compared with less educated parents^8,17,29 and children who attend day care or preschool compared with nonattendees watch fewer hours of television.¹⁷

This intervention to reduce television viewing was well accepted by children, parents, and staff, and all 7 sessions were successfully implemented at the child care centers. Our finding that children in the intervention group, compared with those in the control group, had a relative mean reduction by parental report of 4.7 h/wk in their television/video viewing is clinically and statistically significant, representing a reduction of about one third in weekly viewing hours. Strict comparisons are difficult because we did not ask our questions in exactly the same manner as Robinson et al¹⁵ or Gortmaker et al.¹³ Nonetheless, it appears that our results are similar to those reported by Robinson et al,¹⁵ in which the intervention group reduced their television/video viewing by 4.5 h/wk according to parental report and 7.1 h/wk according to child report, compared with the control group. The reduction also appears comparable to that reported by girls (4.1 h/wk) in a study of sixth through eighth grade students.¹⁶ The validity and the sensitivity of the different questionnaires, however, to detect change in television viewing habits is not known, and may vary by the age of the child, and whether the parent or the child does the reporting.

We sought to reach and influence the parents through the children and via take-home educational materials and parent-child activities. Although this method has been shown to be successful with elementary school–aged children,²⁹ additional strategies may be needed to reach the parents of younger children.

Limitations to this study include reliance on parental report of their children's television/video viewing and other television-related behaviors. We found that parental report of their children's usual television/video viewing on Saturday, Sunday, and weekdays and the viewing hours recorded on a 1-week television/video diary were significantly correlated and yielded comparable viewing estimates (B.A.D. and P.A.D., unpublished data, June 2003). It is possible, however, that after the intervention, parents in the intervention group might have felt more pressure to report socially desired behaviors, leading them to underestimate their children's television/video viewing, which would have increased the apparent difference between the 2 groups. We have no evidence of this, and the fact that so many other behaviors did not change over time or differ between the intervention and control groups leads us to believe that this was not the case.

The relatively high rate of incomplete follow-up and turnover between the 2 school years reflects the transient nature of the preschool population. Children transferred in and out of preschool and day care centers on the basis of parental work schedules and convenience. They also left the preschool or day care centers when parents lost a job, a sibling was born, and/or parents separated or divorced. Between the 2 school years, a significant proportion also left to go to prekindergarten or kindergarten. Because the groups were similar before the intervention and unavailability for follow-up was similar between groups, there is no evidence that subject dropout may have biased the comparison groups.

Additional limitations of this study include the relatively small sample size. This study was conducted in relatively small day care and preschool centers located in rural communities. Larger centers with more children might be available in urban or suburban settings. The relative homogeneity of the study population could also be viewed as a strength (ensuring comparable groups) or a weakness (limiting generalizability.)

Larger, more intensive interventions that extend over longer time and interventions that more directly involve parents and other caretakers might yield greater and longer-lasting reductions in children's television/video viewing. Moreover, studies with more racially, ethnically, and socioeconomically diverse populations are needed to ascertain the generalizability of this intervention. Despite these limitations, these findings suggest that interventions in the preschool or day care setting to reduce television viewing are practical and feasible. Furthermore, because some parents are receptive to making changes to reduce their children's television/video viewing, further research is warranted to understand how to best reach parents.

Unlike studies with older children,^13,15 however, we did not find a significant difference in children's growth or adiposity between the intervention and control groups. Additional research with larger study samples is needed to assess the contribution that reduced television viewing might play in preventing the development of child overweight and obesity.

Article

Television viewing is associated with adverse effects, including higher obesity rates and increased violence, both of which have been shown to decrease after reductions in viewing hours. Television viewing habits begin to develop and increase during the preschool years.

To our knowledge, this is the first study to describe an intervention to reduce television viewing by preschool-aged children. The program was instituted in preschool and day care settings, a place where young children are spending increasing amounts of time.

Corresponding author and reprints: Barbara A. Dennison, MD, Research Institute, Bassett Healthcare, 1 Atwell Rd, Cooperstown, NY 13326 (e-mail: barbara.dennison@bassett.org).

Accepted for publication July 31, 2003.

This study was supported in part by grant 1-R01-HL65144 from the National Institutes of Health, National Heart, Lung, and Blood Institute, Bethesda, Md (Dr Dennison).

This study was previously presented at the Pediatric Academy Societies' Meeting; May 3, 2003; Seattle, Wash.

We thank Elizabeth Jennings and Christina Arnold Meenan for their assistance in developing this intervention, and Kristina Laskovski and Cosby Gibson for their assistance in preparing this manuscript.