Systematic Review of Effective Strategies for Reducing Screen Time Among Young Children
Screen-media use among young children is highly prevalent, disproportionately high among children from lower-income families and racial/ethnic minorities, and may have adverse effects on obesity risk. Few systematic reviews have examined early intervention strategies to limit TV or total screen time; none have examined strategies to discourage parents from putting TVs in their children's bedrooms or remove TVs if they are already there. In order to identify strategies to reduce TV viewing or total screen time among children <12 years of age, we conducted a systematic review of seven electronic databases to June 2011, using the terms “intervention” and “television,” “media,” or “screen time.” Peer-reviewed intervention studies that reported frequencies of TV viewing or screen-media use in children under age 12 were eligible for inclusion. We identified 144 studies; 47 met our inclusion criteria. Twenty-nine achieved significant reductions in TV viewing or screen-media use. Studies utilizing electronic TV monitoring devices, contingent feedback systems, and clinic-based counseling were most effective. While studies have reduced screen-media use in children, there are several research gaps, including a relative paucity of studies targeting young children (n = 13) or minorities (n = 14), limited long-term (>6 month) follow-up data (n = 5), and few (n = 4) targeting removing TVs from children's bedrooms. Attention to these issues may help increase the effectiveness of existing strategies for screen time reduction and extend them to different populations.
The American Academy of Pediatrics has recommended that children over 2 years of age use screen media for no more than 2 h per day (1), yet US children use screen media (e.g., TV, videos, DVDs, video games, and/or computers) for about 7 h per day, on average (2). Most time with media is spent watching TV, with children watching 4½ h per day, and infants watching 1–2 h per day (2,3). There is substantial tracking of TV viewing over childhood, such that heavy viewing in the preschool age group predicts greater TV viewing in later childhood and adolescence (4,5,6). Additionally, over one-third of children under the age of 6 years (7) and more than half of middle school-age children have TVs in their bedrooms (2). Research indicates that having a TV in the room where a child sleeps predicts greater TV use (8,9,10) and that once TVs are placed in a child's bedroom, they are not often removed (11).
Several studies reveal a dose-response relationship between the average hours of TV young children watch and prevalence of obesity (4,12,13,14,15,16). Children who watch commercial TV are exposed to more food and beverage advertisements, which increase obesity risk (17,18). They are also more likely to eat when the TV is on, and sometimes in larger amounts, because satiety cues are often disregarded while viewing (19,20,21).
Few studies have independently examined relationships between obesity and video game use, computer use, or total screen-media use. Some show an association, while others do not (22,23,24,25,26,27,28). This may be because some video games and Internet sites contain food and beverage advertisements, product placements, or “advergames” (in which an advertised product is embedded in a game (29,30)), but many do not. Video game play has also been associated with both increased calorie consumption after gaming (31) and energy expenditure/calorie burning during gaming, especially when children play active games, like those made for Nintendo's Wii gaming system (32,33).
A major public health problem, obesity puts children at risk for heart disease (34), type 2 diabetes (35), asthma (25,36), and depression (34) and is associated with higher morbidity and mortality in adulthood (37,38). Since it is very challenging to reduce obesity once established, early intervention may prove essential for prevention of childhood obesity and its associated complications (39,40). Intervention strategies could include limiting TV or screen time, discouraging parents from putting TVs in their children's bedrooms, or removing the TVs if they are already there.
Many interventions have aimed to prevent or treat obesity by reducing TV viewing, or total screen time, in children (39,41,42). While some have reduced viewing, others have not been successful. To our knowledge, few systematic reviews have solely examined early intervention strategies to limit TV or total screen time, discourage parents from putting TVs in their children's bedrooms, or remove TVs if they are already there. In addition, although prevention efforts may have the largest impact for children <6 years of age, a paucity of studies summarizes what is known about media reduction programs for infants or preschool-aged children. Hence, the purpose of this review was to identify and summarize studies that reduce TV viewing or screen time in children 6–12 years of age, as well as birth to <6 years of age, in order to identify effective strategies for future interventions with these populations and to develop future research priorities based on gaps in the current literature.
Methods and Procedures
We conducted a systematic review of the published literature on intervention studies to reduce TV viewing in children, from 1947 through June 2011, using the terms “intervention” and “television,” “media,” or “screen time,” to identify relevant studies. Databases searched included MEDLINE/PubMed, Academic Search Premier, RePort, ERIC, NHS EED, the Cochrane Database of Systematic Reviews, and the Cochrane Register of Controlled Trials. We also reviewed reference lists of included papers and other relevant reviews and meta-analyses.
Eligible intervention studies were broadly defined as research studies that aimed to reduce TV viewing, or total screen-media use, in children. Studies that did not directly aim to influence screen-media use but aimed to modify behaviors related to overweight or obesity, physical activity, or nutrition and recorded some measure of frequency of TV viewing/screen-media use, as a primary or secondary outcome were also eligible for inclusion. Eligible studies had to include children under age 12, be peer reviewed, and in English. Case studies and theses were excluded. Studies were not excluded because of their designs, durations, or settings.
One author (M.E.S.) read all the titles and abstracts identified in the initial searches. Since our search criteria were very broad, a large number of studies were excluded because it was obvious from their titles that they did not fit our inclusion criteria. Many studies also repeated across databases. After reviewing all of the study titles, one author (M.E.S.) reviewed all the remaining abstracts, and another author (A.O.) separately reviewed all the abstracts of the papers selected for inclusion by the first author to verify eligibility. Of 144 possible titles, 47 were selected for inclusion. Decisions about excluding studies were based on thorough reading of the selected papers by one author, and, if necessary, discussion among the other authors. Common reasons for exclusion included no intervention component (e.g., correlation studies), results not reported, child TV viewing not a measured outcome, not an original article, case studies, or no participants under age 12.
Tables 1, 2, and 3 show the characteristics of the 47 intervention studies selected for inclusion, categorized by age group and study setting. Almost all studies included a home component and most took place in more than one setting. For this review, we categorized studies into settings based on where the intervention program was primarily delivered. Settings included schools, homes, community settings, and clinics.
Table 1. Summary of school-based studies: design, duration, participants, sample, primary exposures/outcomes, intervention, and results
Table 2. Summary of home- and community-based studies: design, participants, sample, primary exposures/outcomes, intervention, and results
Table 3. Summary of clinic-based studies: design, participants, sample, primary exposures/outcomes, intervention, and results
Among included studies, a wide variety of strategies were used to reduce children's TV viewing. These included educational sessions, physical activity sessions, resources/curricula for teachers, resource kits for parents, parent newsletters, electronic TV time monitors (e.g., TV Allowance, Mindmaster, Miami, FL), incentives, open- or closed-loop contingent systems, activity programs, and community advertising. Most interventions (35 of 47, or 74%) were randomized controlled trials and lasted less than one year (32 of 47, or 68%). Only 11 (23%) evaluated follow-up effects beyond the end of the intervention period. In all, 29 of 47 studies (62%) reported statistically significant reductions in TV viewing or total screen time after intervention.
In the subheadings below, studies are reviewed in greater detail, within setting categories and age groups. Special attention is given to intervention strategies and their outcomes. Given the many differences across cultures in nutrition, physical activity, and educational systems, results are separated into United States and international studies in the tables; however, all programs that share settings and ages are reviewed together.
Early education and school-based studies
Almost all school-based studies involved some kind of in-class instruction about nutrition, physical activity, and/or media use; some programs also included a physical education program, teacher training, and/or school food service modifications. Only four programs were found for children <6 years of age. Most school-based studies had large sample sizes, with participants 8 years of age or older. The majority of programs lasted between 6 months and 1 year; only three lasted 2 years or longer.
Early education and school-based studies, children under 6 years. As reported in Table 1, four school-based studies with children under 6 years of age were included in our review, and only one of these significantly reduced TV viewing. One additional study reduced total screen time. All took place in the United States and were randomized controlled trials. Three lasted for 14 weeks, and one lasted for 7 weeks; three had sample sizes under 500. Three of the interventions were specifically designed to target racial/ethnic minority children (43,44,45).
Only Dennison et al. significantly reduced TV viewing (∼3 h per week), via a 7-session, weekly TV reduction-focused intervention program, which also provided participants with a variety of materials for parents and children to use at home. Video/computer game play, measured separately, was not reduced (46). Three other school-based studies with preschool-aged participants tested the same intervention program (“Hip Hop to Health”) among Black and Latino children (43,44,45). Although the “Hip Hop to Health” program did not reduce TV viewing in any of the studies, total screen time was reduced, by just under 28 min per day, in one study (45).
The Dennison et al. study specifically targeted reducing TV viewing or promoting alternatives to TV viewing, like reading. Further, the Dennison et al. intervention included two “TV turnoff week” components, and parents were provided with materials and incentives to facilitate achieving their child's TV reduction targets. In contrast, the “Hip Hop to Health” interventions primarily targeted diet and physical activity, and devoted only 1 week of each of the 14-week programs to reducing TV viewing.
School-based studies, 6- to 12-year-old children. Fifteen studies implemented in grade schools were selected for inclusion, as detailed in Table 1. Six took place in the United States, and nine took place internationally. Most school-based studies with 6- to 12-year-old children were randomized controlled trials and had over 500 participants. Almost half lasted for 1 year or longer. Three of 15 studies specifically recruited ethnic minorities (47,48,49); five studies specifically recruited from low socioeconomic status (SES) schools (47,50,52,53,54).
As measured at the end of the intervention programs, eight of 15 school-based intervention programs significantly reduced TV and/or screen-media use. Robinson et al. reported the largest reductions (∼1 h per day) in screen-media use (56), followed by Gortmaker et al. (∼40 min per day (51)) and Muller et al. (20 min per day, TV only) (40). Jones et al. (57), Sprujit-Metz et al. (48), and Salmon et al. (54) all reported reducing screen media by less than 20 min per day. Gentile et al. reported a 2-h per week reduction in screen time, but only according to parent, and not child, report (58). One additional study, by Simon et al., reduced the proportion of children spending >3 h per day in sedentary activity, by 7% in girls and 3% in boys (55).
All but one of these programs specifically focused on screen-media reduction or had content on reducing screen media as a major part of classroom lessons. Like Dennison et al., many also utilized household TV reduction campaigns, e.g., Robinson included a 10-day TV turnoff campaign (56) and Gortmaker et al. included a 2-week “power down” campaign (51). The use of electronic TV time monitors was a unique component of the Robinson trial that reported the largest statistically significant reductions in screen media (56).
A variety of intervention programs have been tested in homes, including contingent feedback systems, TV time monitors, and parent education programs. Most home-based studies lasted less than 6 months and had less than 50 participants. Unlike school-based programs, most home-based interventions reviewed here specifically recruited obese or overweight participants (59,60,61,62), and/or participants who watched above average amounts of TV (60,62,63,64). No home-based studies specifically recruited ethnic minorities or low-SES participants. Only two home-based interventions included children under 6.
Home-based studies, children under 6 years. As shown in Table 2, two home-based studies included children under age 6. Only Epstein et al. successfully reduced screen-media use by using TV Allowance devices, monetary incentives, and sticker charts to reward children for reducing their TV time to 50% of baseline. After 6 and 24 months, children in the intervention group recorded 17.5 fewer hours of TV and computer use per week (59). In contrast, the second home-based study for children under 6, by Essery et al., utilized weekly newsletters or a 52-page booklet to improve preschoolers' feeding practices and physical activity (65). Reducing TV viewing was recommended in the newsletter/booklet; however, the program did not specify particular goals for reducing TV viewing, nor was TV reduction its primary focus.
Home-based studies, 6- to 12-year-old children. Three of five home-based studies significantly reduced TV viewing, or screen-media use, in 6- to 12-year-old children (see Table 2). Faith et al. employed a “closed-loop” feedback system to reduce TV viewing in school-age children and recorded the largest reductions (∼20 h per week) among home-based studies. Participants could only watch TV while they pedaled a custom-made stationary bicycle, attached to a home TV, at a prescribed level of intensity (60). Goldfield et al. also used a contingent feedback system, albeit an “open-loop” one, in which intervention participants' TV access was made contingent on their level of physical activity—for each 400 counts of physical activity on a pedometer, they could watch 1 h of TV/VCR/DVDs, as managed by an electronic device. Goldfield reduced TV/VCR/DVD/video game playing time by ∼2 h per day (62). Using electronic TV time monitors, TV time budgets, and TV viewing diaries, Todd also reduced screen-media use, by about 1 h per day (64).
A variety of intervention programs were conducted at the community level, including family workshops, an after-school dance program, and an employee wellness program. Seven interventions were delivered in community settings, and all targeted children over 6 years of age. Five were conducted in the United States, while two occurred in the United Kingdom and in Australia. Most had sample sizes over 100 and lasted less than 1 year. Three studies recruited overweight or obese participants specifically (67,68,69), and three specifically targeted African American or Hispanic children from low-income communities (66,67,68). One additional study, by de Silva-Sanigorski et al., was delivered to all children <5 years in an entire community in Australia (N = 12,000), via preschools, daycares, community health centers, immunization, and civic programs (70).
Three of seven community-based programs significantly reduced TV viewing or screen-media use. Sacher et al. reduced sedentary activity (TV and computer use) by almost 5 h per week, through a family education, physical activity, and provision of a 12-week free swimming pass to families. Participants were all obese at baseline (69). In de Silva-Sanigorski, media use was significantly lower in the intervention as compared to a control community after a community-wide program for children <5 years that trained early childhood workers and provided resources to parents and teachers (70). In Sepulveda et al., 7% more parents who completed an online parent education program offered to employees of a large corporation were more likely, after the program, to report that their children watched <1 h of screen time per day (71).
Clinic- and WIC-based studies
Clinic-based studies relied primarily on parent and child training/counseling (Table 3). Counseling or training was usually offered by health professionals (e.g., doctors, dietitians, Women, Infants, and Children (WIC) staff). Most studies had over 300 participants and lasted less than 1 year. Seven of the 14 clinic-based studies targeted children under 6. Five specifically recruited low-SES, ethnic minority participants (72,73,74,75,76).
Clinic- and WIC-based studies, children under 6 years. Seven of 14 clinic-based studies specifically targeted children <6 years of age. All took place in the United States, and three were part of the Women, Infants, and Children (WIC) program. Half of the studies ranged between 200 and 600 participants; two included over 3,000 participants. Most programs lasted between 6 months and 1 year. Three of seven studies targeted low-SES participants and/or ethnic minorities (72,74,75). Only one targeted overweight or obese participants (77).
All seven clinic-based studies conducted with children <6 years of age reported significant differences in screen-media use or parenting practices around screen-media use. Three were conducted as part of WIC programs. In Johnson et al., Whaley et al., and Davison et al., families received counseling by WIC staff on reducing TV, physical activity, and/or nutrition. In Johnson and Davison, more intervention parents were likely to report that their children watch <2 h of TV per day after the program (6% and 9 % more parents in Johnson and Davison, respectively (72,78)). In Whaley et al., increases in TV viewing at 12 months were ∼6 min less in the intervention group than in the control group (75).
Four studies were based in primary care clinics. In two studies, participants received parenting advice from a program specialist. In Johnston et al., 16% more intervention, compared to control, parents were less likely to allow more than 1 h of TV viewing daily (79). In Mendelsohn et al., media exposure was reduced for children in the intervention group, by about 20 min per day (74). In two other studies, motivational interviewing techniques were used to encourage participants to reduce media use. In Barkin et al., parents received 2–3 min of counseling from a primary care physician, after which media time was reduced by 30 min per day, and parents were more likely to limit media use to <2 h per day (5.7% increase for the intervention group (80)). In Taveras et al., participants received counseling from a nurse practitioner, and there was a decrease in TV or video viewing, of about 22 min per day, in the intervention group (77).
Clinic-based studies, 6- to 12-year-old children. As reported in Table 3, five of seven clinic-based interventions with children over 6 years of age had sample sizes under 50. Four took place in the United States, and three were conducted internationally. Most were of relatively short duration, lasting for 3 months or less. Three programs recruited obese participants (81,82,83), and two targeted low-SES, ethnic minority children (73,76).
Five of seven clinic-based studies reported statistically significant reductions in TV viewing. Two programs by Nemet et al. reduced screen time by about 2 h per day. In both studies, parents and children attended regular meetings with a dietitian, and children participated in regular physical activity sessions. Parents were specifically encouraged to decrease their children's sedentary behavior, including TV viewing (82,83). Deforche et al. also significantly reduced screen time, by approximately 2 h per day, via an inpatient residential obesity treatment program, in which TV viewing was restricted (81). In Stahl et al., pediatric residents received web-based training, about a program to encourage healthy eating, physical activity, and screen-media reduction. Twelve percent more parents in the intervention group than in the control group reported reducing their children's TV viewing after the intervention (84). Perrin increased the percentage of children in the intervention group who used less than 2 h of screen media per day by 22% by briefly training and providing toolkits to pediatric residents (76).
Reducing TV time is a potential strategy to prevent or treat childhood overweight and obesity. Of the 47 intervention studies we reviewed, 29 (62%) reported statistically significant reductions in children's TV viewing or screen-media use. Of these, 18 measured BMI and 9 reported reductions in BMI. The most effective interventions specifically targeted and set explicit goals for reduced TV viewing or screen-media use, used electronic monitoring devices, contingent feedback systems or clinic-based counseling, had high levels of parental involvement, and/or recruited participants who were already overweight or obese at baseline.
To our knowledge, few published systematic reviews have summarized effective strategies for reducing TV viewing in young children. A recent meta-analysis by Maniccia et al. revealed a small but statistically significant effect of screen-media interventions to reduce children's screen time. Twenty-nine studies were identified as eligible for meta-analysis (85). In a systematic review of interventions to reduce sedentary behavior (defined as recreational screen time) by DeMattia et al., 12 studies were identified to successfully reduce sedentary behavior (86). Similar to the results of our review, DeMattia et al. found that study approaches and settings varied. DeMattia et al. concluded that targeting sedentary behaviors is an effective way to intervene on obesity and overweight outcomes in children and adolescents (86). Our review updates the literature on reducing TV and video viewing since the publication of DeMattia et al. in 2007 and the meta-analysis of Maniccia (which included studies to 2008), and includes older studies not selected by DeMattia et al. or not eligible for inclusion in meta-analysis in Maniccia et al. We report a similar intervention success rate to DeMattia et al., wherein just over 60% of studies successfully reduced TV viewing in children. Our study also extends Maniccia et al.'s and DeMattia et al.'s work by reporting additional strategies to reduce TV viewing and providing added narrative detail about which intervention strategies and settings were most effective.
We identified electronic monitoring systems as one strategy that most effectively reduced TV viewing among children. Intervention programs that used electronic TV monitors reported significant, large decreases in TV viewing, from 1.5 to 3 h per day. However, in two studies included in our review, about half of families offered electronic TV time monitors either did not use them (56) or reported, after using them, that they would not want to use them in the future (63). Thus, while electronic monitors seem to be an effective strategy for TV reduction, further research is needed to understand how to increase their acceptability in households with children. In addition, more research is needed to determine the long-term effectiveness and sustainability of electronic TV time monitors (63).
Another strategy that had considerable effects on the reduction of TV viewing was the use of contingent feedback systems. For example, Faith et al. (60) used a closed-loop feedback system where TV viewing was made contingent on stationery cycling and saw a decrease in TV viewing by 20 h per week, one of the largest reductions reported in this review. Another effective contingent feedback system was an open-loop feedback in which TV viewing was made contingent on physical activity, as recorded by pedometer or accelerometer. Goldfield used this open-loop design and reduced TV by 116 min per day (62).
Counseling by physicians, nurse practitioners, or Women, Infants, and Children (WIC) providers also had measurable effects on TV viewing. Twelve of 14 clinic-based studies in this review reported significant findings. Most of these studies (seven of 14) were with children under 6 years. However, only one clinic-based study with children under 6 measured follow-up outcomes (80), and only one measured BMI (77). Since the primary care setting offers unique access to large numbers of parents of young children, and parents may be especially receptive to messages delivered in this setting, further research should examine the role of the primary care provider in TV reduction counseling, particularly with regard to effects on BMI, or other weight-based outcomes (74,88). Future clinic-based research should also measure long-term outcomes, in order to determine whether early intervention can have beneficial effects on long-term TV viewing trajectories.
In Table 4, we list screen-media reduction strategies, including but not limited to electronic TV time monitors and feedback systems, that reduced TV/screen media viewing by statistically significant amounts, across multiple studies. Among the different settings in our review, we noted that the largest reductions in TV viewing occurred in home- or clinic-based settings. This may be because parents are required to be involved in home- and clinic-based programs; prior research suggests that high levels of parental involvement are very important, if not essential, for intervention success (87,89,90,91). In addition, most of the home- and clinic-based studies in this review specifically targeted overweight or obese children or children who watched large amounts of TV, whereas interventions in other settings were typically delivered to all participants, randomly assigned to intervention groups regardless of weight or TV viewing habits.
Table 4. TV reduction strategies that should be considered for future interventions
Table 5 lists priorities for future research. In general, we recommend future studies include greater recruitment of racial/ethnic minority children and children under 6. Although studies suggest that electronic TV monitors and feedback systems are the most effective TV reduction strategies, little is known about their long-term feasibility and effectiveness. Furthermore, to our knowledge, only Epstein et al. (59) has used electronic TV monitors with children under 6, and no published findings report on the use of open or closed-loop feedback systems in preschool-aged children. Finally, while electronic TV monitors can be effective, they may be prohibitively expensive, at $60 to $90 each on average (92), for use in large-scale public health intervention programs.
Table 5. Research priorities and recommendations for intervention planning based on gaps in the current literature
Studies have shown that time-use diaries or electronic monitoring systems provide the most accurate data on TV viewing (93), yet most studies in this review used global, retrospective estimates of TV use, e.g., “How many hours, on average, do you watch TV per week?” Future research should employ TV viewing measurement techniques with greater validity, such as electronic monitors, time-use diaries, or momentary sampling (94). An additional limitation to current studies is that only one measured TV content or intervened on specific TV content (74). Since one of the hypothesized mechanisms for associations between TV watching and obesity is increased food intake, either through food and beverage advertising or increased eating during viewing, future research should investigate whether interventions that target specific TV or media content (e.g., food advertising) are effective at reducing TV and/or BMI.
We also were unable to identify any studies that aimed to prevent or treat obesity by specifically reducing computer use or video game use. These kinds of media use have been included in aggregate measures of screen time in intervention studies, but they have not been independently tested as intervention targets. Future research should also explore the potential for interventions targeting specific media, including new media platforms that deliver TV content and advertising (e.g., Internet, mobile “smart” phones). These have not been well studied.
Numerous studies have indicated that children with TV in their bedrooms watch more TV than children without TV in their bedrooms (9) and are at an increased risk for obesity (9,95,96), sleep problems (97,98,99), and other health risk behaviors, e.g., smoking (100) and alcohol use (101). Although two-thirds (68 %) of 8- to 18-year-old children have TV in their bedrooms (2), only 10 studies in this review measured whether children had TV sets in their bedrooms (46,52,56,59,66,68,72,73,77,102) and recommending removal of bedroom TV sets was listed as a component of only four intervention programs (76,77,84,102). We recommend future interventions specifically address the removal of TV from children's bedrooms, both as a behavioral outcome in itself, and in order to reduce overall TV and video viewing.
The vast majority of studies in this review were randomized controlled trials. Due to the wide variety of methods, outcomes, and measures reviewed here, a meta-analysis was not possible. Our conclusions are based on qualitative analysis of broad patterns in the body of published literature and are not definite. In particular, the diverse measurement techniques across studies in this review made it very difficult to compare findings. Although research suggests that different kinds of sedentary media behaviors contribute to obesity differently, if at all (25), half of the studies in this review aggregated screen-media use as a single outcome. When intervention programs report aggregate screen-media reductions, it is unclear which specific media are reduced. When BMI is an outcome of interest, aggregate screen-media measures do not distinguish which specific media may be implicated in any BMI changes. Media use or TV viewing also was variously measured by child report, parent report, or both, and sometimes these measures did not agree (59). For all of these reasons, we were able to draw limited conclusions about the comparative effectiveness of different interventions to reduce TV viewing or screen-media use.
Forty-seven studies have been conducted to reduce TV viewing among children <12 years of age, in various settings, using different strategies, and with variable success. Our findings suggest that electronic TV monitors, contingent feedback systems, and clinic-based counseling were three of the most effective strategies for TV reduction. More work is needed to understand the potential for interventions in children <6 years of age and in low-income and ethnic/minority participants. Future research should also further explore the potential for primary care counseling to reduce children's long-term media use. Finally, no published research to date has evaluated whether removing TV sets from children's bedrooms could measurably impact TV viewing. This should be a priority for future research.
This work was supported by the National Center for Chronic Disease Prevention and Health Promotion (Prevention Research Centers Grants, 1U48DP00194). The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention.
The authors declared no conflict of interest.