BACKGROUND AND OBJECTIVE: To date, most parent-based research has neglected the role of fathers in shaping adolescent sexual behavior and has focused on mothers. The objective of this study was to conduct a structured review to assess the role of paternal influence on adolescent sexual behavior and to assess the methodological quality of the paternal influence literature related to adolescent sexual behavior.
METHODS: We searched electronic databases: PubMed, PsychINFO, Social Services Abstracts, Family Studies Abstracts, Sociological Abstracts, and the Cumulative Index to Nursing and Allied Health Literature. Studies published between 1980 and 2011 that targeted adolescents 11 to 18 years and focused on paternal parenting processes were included. Methodological quality was assessed by using an 11-item scoring system.
RESULTS: Thirteen articles were identified and reviewed. Findings suggest paternal factors are independently associated with adolescent sexual behavior relative to maternal factors. The most commonly studied paternal influence was emotional qualities of the father-adolescent relationship. Paternal communication about sex was most consistently associated with adolescent sexual behavior, whereas paternal attitudes about sex was least associated. Methodological limitations include a tendency to rely on cross-sectional design, nonprobability sampling methods, and focus on sexual debut versus broader sexual behavior.
CONCLUSIONS: Existing research preliminarily suggests fathers influence the sexual behavior of their adolescent children; however, more rigorous research examining diverse facets of paternal influence on adolescent sexual behavior is needed. We provide recommendations for primary care providers and public health practitioners to better incorporate fathers into interventions designed to reduce adolescent sexual risk behavior.
- MQS —
- methodologic quality score
- STI —
- sexually transmitted infection
Adolescence is a developmental stage associated with increased risk-taking behaviors that contribute to negative sexual health outcomes.1,2 The 2011 Youth Risk Behavior Survey documented the prevalence of sexual risk behaviors among US high school students. This survey showed that 47% of students in grades 9 to 12 have ever had sexual intercourse and 40% of currently sexually active high school students did not use a condom at their last sexual intercourse,3 indicating that many adolescents are at risk for pregnancy and sexually transmitted infections (STIs), including HIV.
In response to this area of continued concern regarding adolescent sexual risk behaviors, researchers have investigated the role of families in supporting healthy adolescent development.4 This research has demonstrated a notable association between positive parent-adolescent relationships and reduced adolescent engagement in sexual risk behaviors.5 Specifically, parenting processes such as monitoring and discipline, parent-adolescent communication, and parent-adolescent relationship satisfaction have been associated with adolescents’ involvement in sexual risk behaviors and their subsequent sexual and reproductive health outcomes.6–9
Although the importance of parental effects has been well established, most research has focused on how mothers shape adolescent sexual behavior. Numerous studies report associations between specific attributes of the mother-adolescent relationship, such as maternal closeness and support, and positive adolescent behaviors, such as increased contraceptive use, higher abstinence rates, and decreased sexual risk-taking.10–12 Attributes of mother-adolescent communication, such as greater frequency of communication, openness, and self-disclosure about dating experiences during adolescence, have been associated with more conservative attitudes among adolescents toward sex and delayed sexual debut.12–15
Research has largely concentrated on mothers, in part because mothers have been perceived to be the parent primarily responsible for providing directed education about sexuality.16 When research has explored paternal influences on adolescent behavior, issues of father absence and economic support have dominated the research.17–19 Consequently, research on father involvement has tended to conceptualize fathers with a limited perspective. More recently, however, research on fathers has explored the relationship between paternal parenting practices and child development. Evidence increasingly suggests that mothers and fathers independently shape areas of child development, such as academic success20 and peer relationships.21 Little is known about how paternal parenting processes are associated with adolescent sexual behavior. This lack of research warrants further attention.
This review examines the relationship between father parenting processes and adolescent sexual behaviors that contribute to pregnancy and STIs, including HIV. We also assess the methodological quality of studies included in this review.
A computer-based search of the literature was conducted via 6 databases: PubMed, PsychINFO, Social Services Abstracts, Family Studies Abstracts, Sociological Abstracts, and the Cumulative Index to Nursing and Allied Health Literature. Keywords for adolescent sexual behavior and parenting processes were identified. We used an ancestry approach22 to identify additional studies eligible for inclusion. With this approach, the references of articles identified through the initial computer-based search were examined. For an example of our search strategy, see Fig 1.
We included studies that (1) were peer-reviewed articles published in English between 1980 and 2011, (2) targeted adolescents 11 to 18 years old residing in the United States, (3) examined adolescent sexual behaviors or reproductive outcomes, and (4) focused on paternal parenting processes. We recognize that the adolescent age range of 11 to 18 years is broad; however, most studies of paternal influences recruit fathers whose children span this range and do not stratify results by adolescent age. We included studies that involved male caregivers who were biological, step, or adoptive fathers, whether they resided with their children or not. We also included studies of other male figures, such as uncles or grandfathers, if these men fulfilled a primary male caregiver role. Articles were included only if adolescent sexual behaviors were examined as distinct dependent variables. In addition, we included studies that examined paternal influences on adolescent sexual behaviors and resultant reproductive outcomes, such as unintended pregnancy and STIs, including HIV.23 We defined paternal influence variables as general parenting processes (eg, monitoring and discipline and parent-adolescent communication). Father-related family structure variables (ie, factors such as employment and absence from the home) were not the focus of the review and were excluded. Because of the focus on identifying independent paternal influences, studies that analyzed parent data in aggregate form without distinguishing paternal and maternal influences were excluded. Studies with positive, negative, or no paternal effects were included.
Eligible articles identified in each database were exported to an EndNote library. The following information was abstracted: sample characteristics, paternal variables studied, and adolescent sexual behaviors and reproductive outcomes. Abstracted paternal variables include the following thematic groupings: (1) paternal attitudes about adolescent sexual activity; (2) monitoring and discipline-related behaviors; (3) paternal involvement in their adolescent’s life; (4) emotional qualities of the father-adolescent relationship (eg, warmth, closeness, attachment); and (5) father-adolescent communication about sex. Each study’s analysis of the association between a paternal influence and an adolescent sexual behavior or outcome was abstracted as a separate finding. Therefore, a study could contribute multiple findings. In studies where unadjusted and adjusted analyses were reported, only the controlled finding was abstracted. A structured review protocol24 was followed and we created a database of articles that met the inclusion criteria (Fig 2).
Given heterogeneity in data available for each paternal influence examined, it was not possible to estimate effect sizes of paternal variables on each area of adolescent sexual behavior. The empirical approaches and variables were too varied to conduct a formal meta-analysis. We report data on 6 adolescent sexual behaviors and 1 reproductive outcome: (1) ever had sex, (2) frequency of sex, (3) number of sexual partners, (4) contraceptive use, (5) frequency of condom use, (6) pregnancy, and (7) a composite measure of adolescent sexual behavior. The outcome “composite measure of sexual behavior” was created by the study authors to refer to scale constructs reported in studies that measured multiple adolescent sexual behaviors simultaneously. We synthesize findings by 3 types of study design: cross-sectional, longitudinal observation, and longitudinal intervention, with data from the cross-sectional category providing the weakest evidence and data from longitudinal intervention category providing the best evidence of paternal effects.
Thirteen articles examining paternal process variables on adolescent sexual behavior were identified (Table 1). Six studies were cross-sectional, 6 were longitudinal observational (henceforth “longitudinal”), and 1 was a longitudinal intervention study (henceforth “intervention”). The paternal process most frequently examined in the reviewed literature (7 of the 13 studies) was emotional qualities of the father-adolescent relationship. The least examined paternal process variables were paternal involvement and father-adolescent communication about sex (2 of 13). We provide an overview of each hypothesized paternal variable on each adolescent sexual behavior or outcome in the following sections. Additionally, we compare data for father versus mother parenting practices where such comparisons were possible.
Ever Had Sex
Eight of the 13 studies included a measure of adolescent sexual debut. Cross-sectional data were the most common type of data available (Table 2), followed by longitudinal (Table 3) and intervention (Table 4). In general, paternal attitudes toward adolescent sexual behavior were significantly associated with adolescent sexual debut. Only cross-sectional data showed that paternal approval of adolescent sexual activity was associated with earlier sexual debut, whereas paternal disapproval was associated with delayed sexual debut.25,26 Cross-sectional data also supported a curvilinear relationship between both low or high levels of paternal discipline and earlier adolescent sexual debut.27 For emotional qualities of the father-adolescent relationship, findings varied by type of study design. Longitudinal studies found positive emotional qualities of the father-adolescent relationship, such as higher levels of connectedness or closeness, were significantly associated with delayed sexual debut.28,29 Conversely, cross-sectional data demonstrate negative emotional qualities were associated with earlier sexual debut. For example, 1 qualitative study found poor father-adolescent relationship quality was associated with the onset of sexual activity among teenage mothers30; however, prospective data provided less support for this relationship. For example, although paternal closeness31 and connectedness29 were significantly associated with delayed sexual debut at baseline, only paternal closeness was associated with delayed sexual debut for daughters 1 year later.31 Mixed findings were also observed for paternal involvement. Longitudinal data suggest that paternal shared activities were negatively associated with adolescent sexual debut at baseline, with no significant association observed prospectively.31 Higher levels of problem-focused interactions, however, were a form of paternal involvement positively associated with adolescent sexual debut both concurrently and prospectively.31 For the influence of paternal communication, intervention data from 1 study only are available. A father-son HIV prevention intervention (REAL men) showed higher levels of paternal communication about sex were associated with increased rates of adolescent abstinence.32 This association, however, was observed at the 6-month postintervention time point only, limiting the generalizability of the finding.
Taken together, 8 studies documented a significant relationship between the hypothesized paternal processes: attitudes about sex,25,26 monitoring and discipline,27 emotional qualities of the father-adolescent relationship,28–31 involvement,31 and communication on adolescent sexual debut.32 However, the vast majority of these studies relied on cross-sectional design, therefore providing only limited correlational evidence of paternal relevance.
Frequency of Sex
Our review yielded no support for the statistically significant association between paternal variables and adolescent frequency of sex; however, cross-sectional data from 1 study that examined the association between monitoring and discipline (ie, paternal strictness) and adolescent frequency of sex were available for this outcome (Table 2).33 Given the paucity of research, it is premature to make conclusions about the ability of paternal variables to influence frequency of sexual intercourse in youth.
Number of Sexual Partners
Cross-sectional and longitudinal data were available for the adolescent outcome number of lifetime sexual partners (Tables 2 and 3).33,34 Cross-sectional data indicate paternal monitoring and discipline (ie, paternal strictness) was unrelated to adolescent number of sexual partners,33 whereas longitudinal data demonstrate some emotional qualities of the father-adolescent relationship (ie, paternal self-restraint) were associated with the reduced number of adolescent sexual partners 4 years later.34 These studies provide some evidence to support the association between paternal monitoring and discipline and quality of the father-adolescent relationship on adolescent number of sexual partners, but the evidence derives from somewhat weak designs.
Both cross-sectional and intervention data were available for the outcome of adolescent contraceptive use (Tables 2 and 4). We observed mixed findings for the association of paternal variables on adolescent contraceptive use. For example, only some paternal attitudes toward adolescent sexual activity (ie, paternal tolerance of adolescent sexual behavior), were statistically associated with contraceptive use, whereas paternal attitudes of approval of adolescent sexual behavior and approval of teenage social activities without adult supervision were unrelated.25 Intervention data from 1 study showed there was a positive association between paternal communication and adolescent contraceptive use at the 12-month postintervention time point only.32 These results provide weak support for the association between paternal attitudes toward adolescent sexual behavior and paternal communication about sex on adolescent contraceptive use.
Frequency of Condom Use
Our review documented no support for the significant association between paternal influence and adolescent frequency of condom use (Table 2). Specifically, cross-sectional data from 1 study found paternal monitoring and discipline (ie, paternal strictness) was unrelated to frequency of condom use.33
Only cross-sectional data from 1 study were available for measuring the association between paternal variables on the outcome of adolescent pregnancy (Table 2). Specifically, no significant relationship was found between paternal monitoring and discipline (ie, paternal strictness) and if a sample of African American adolescent male teenagers reported having ever caused a pregnancy.33
Composite Measure of Adolescent Sexual Behavior
Both cross-sectional (Table 2) and longitudinal (Table 3) data were available for the composite measure of adolescent sexual behavior outcome, which across studies included measures of frequency of sex, number of sexual partners, and frequency of contraceptive use/nonuse. For the father variable of paternal communication, only cross-sectional data were available.35 The study found greater paternal communication was associated with increased frequency of sexual behavior among older adolescents.35 In contrast, less paternal communication was associated with decreased frequency of sexual behavior among younger adolescents, suggesting differential results depending on adolescent age. Longitudinal studies suggest both positive father variables related to paternal monitoring and discipline practices (ie, monitoring knowledge)36 and negative father variables, such as poor emotional quality of the father-adolescent relationship (ie, weaker father-adolescent mutual attachment)37 were significantly associated with increased adolescent sexual risk behavior. For example, although paternal monitoring knowledge was linked to reduced risk behavior at baseline, this association waned and reversed direction over time, potentially because of fathers becoming more aware of their teenager’s risky behavior as it increases over time.36 In sum, all 3 studies examining a composite measure of adolescent sexual risk behavior documented a significant relationship with the father-specific variables of paternal communication, paternal monitoring and discipline practices, and emotional qualities of the father-adolescent relationship, although 1 study’s use of cross-sectional design and differences in the orientation of the composite scale make it difficult to summarize the relationship between paternal influence and this outcome.
Paternal Versus Maternal Influence
Several studies included both mothers and fathers and simultaneously examined independent associations of each parenting factor on adolescent sexual behavior (see note in Tables 2 and 3). Specifically, 3 cross-sectional and 3 longitudinal studies compared maternal and paternal parenting effects. Included studies used rigorous analysis techniques, such as canonical correlation,35 multiple regression,26,28,29,33 and multilevel growth analysis,36 to evaluate the effect of parental gender as a predictor of adolescent sexual behavior. Cross-sectional studies found paternal communication was associated with changes in adolescent sexual behavior while controlling for maternal communication,35 paternal attachment was associated with decreased older adolescent sexual behavior, whereas maternal attachment was unrelated,35 and paternal disapproval of adolescent behavior delayed adolescent sexual debut slightly beyond the effect of maternal disapproval. Specifically, adolescents with increasing paternal or maternal disapproval, independently, were less likely to ever have sex.26 Paternal monitoring and discipline (ie, paternal strictness), however, was unrelated to adolescent sexual behavior while controlling for maternal strictness.33 The longitudinal studies found father-daughter relationship quality, connectedness, and father monitoring knowledge influenced adolescent sexual behavior while respective maternal correlates had no effect. Adolescents with higher levels of paternal relationship quality28 or paternal connectedness29 were less likely to initiate sex 1 year later, whereas maternal relationship attributes were unrelated to sexual debut. Additionally, adolescents in families with higher levels of paternal monitoring knowledge were more likely to engage in sexual activity over time, whereas maternal knowledge was unrelated.36
We characterized and scored each of the studies by using the methodologic quality scoring (MQS) system, which has been validated in previous studies.38,39 The MQS is composed of 11 variables and has a possible range of 0 to 20. The frequency distribution for each methodologic criterion is presented in Table 5. By using the guidelines of a previous review,40 scores were grouped to denote low (0–6 score), medium (7–14 score), and high (≥15 score) quality studies. Two primary coders independently scored the studies, with an interrater reliability of 0.84. After initial study coding, the first author met with coders to resolve issues of disagreement. Total MQS scores ranged from 7 to 19. By using the MQS classification, most studies were classified as medium quality,25–27,30,32–35 5 studies were high quality,28,29,31,36,37 and no studies were low quality (see Table 6).
Reports of reliability and validity measures varied widely among studies. Although many studies included validity (4 studies) or reliability data (8 studies) for the paternal process variable examined, or another variable in the study (6 studies), only 1 study reported validity or reliability data for both the paternal process variable and at least 1 main outcome variable.36
Nine studies used a theoretical framework to guide study design and analysis. The most commonly cited theory was social control theory,28,31–33 followed by family systems29 or interaction theory.31,36,37 Other theoretical frameworks used were “the Big 5” domains of adult personality traits34 and ecological systems theory.29 No study used a theoretical framework that was father-specific.
Research Paradigm and Study Design
Most studies used a quantitative research paradigm and nearly half of studies relied on a cross-sectional study design and self-reported survey data. Seven studies collected adolescent reports of perceived paternal influence26,28,29,31,33,35,36 and 2 studies collected father self-reports of paternal influence.25,34 Three studies collected both paternal and adolescent reports of paternal variables.27,32,37 One study used a qualitative research paradigm30 and no study used a mixed-methods approach. More than half of the studies used an observational longitudinal design with at least 2 data-collection time points.28,29,31,32,34,36,37 Only 1 study used a longitudinal randomized control trial design.32
Sample Size and Sampling Methods
Seven studies had large sample sizes (n >300), 4 of which used a randomly selected, nationally representative sample: for example, data collected from a national study such as the National Longitudinal Survey of Adolescent Health.28,29,31,36 Most studies with small or medium sample sizes (n <300) used convenience or nonprobability sampling.27,30,33–35,37 Authors of 1 study using convenience sampling reported a 50% or less response rate.35 No study reported formal power analysis; however, 1 study was qualitative, and statistical methods such as power analysis do not apply.30
Data Analytic Methods
More than one-third of studies used linear or logistic regression analysis,25,26,28,29,33 whereas 3 studies reported using bivariate analysis, such as correlations or t tests.27,32,34 Four studies used higher-order statistical methods, such as structural equation modeling37 or path analysis.31 One study used qualitative analysis and presented no quantitative data.30
Inferences of Causality
Ten studies made appropriate conclusions about their study’s findings given its design limitations, whereas 3 did not.25,27,34 Studies lacking appropriate inferences of causality did not state main internal threats to study validity (such as cross-sectional design), and/or instead discussed limitations related to the generalizability of the study.
Our primary objective was to assess the specific impact of paternal parenting processes on adolescent sexual behavior. Overall, we found a lack of research on the influence of fathers despite the well-documented importance of familial variables in shaping adolescent behavior.4,6 Critical paternal process variables apparent in the literature include paternal attitudes toward adolescent sexual activity, monitoring and discipline, paternal involvement in the life of their adolescent child, emotional qualities of the father-adolescent relationship, and father-adolescent communication about sex.
Our review suggests several plausible paternal mechanisms of import to adolescent sexual behavior and outcomes. The most consistent finding across reviewed studies, including those using both cross-sectional and intervention design, was the significant association between father-adolescent communication and adolescent sexual behaviors such as increased condom use and abstinence from sex.32,35 In general, studies suggest that more positive relationship qualities, including higher emotional quality of the father-adolescent relationship and greater paternal involvement, were associated with decreased adolescent sexual risk behavior; however, findings were mixed and likely varied owing to heterogeneity in defined paternal variables. Studies linking paternal attitudes and adolescent behavior were the most inconclusive and relied on cross-sectional designs only. For example, although paternal disapproval of sexual activity was associated with delayed sexual debut, neither paternal tolerance of adolescent sexual activities nor approval of social activities without supervision were significantly related.25,26 Paternal monitoring and discipline and adolescent outcomes varied by level of parental regulation and study design. Cross-sectional data indicate paternal strictness was associated with delayed adolescent sexual debut when levels were moderate27; however, lenient and overly restrictive paternal strictness was associated with earlier sexual debut, suggesting a curvilinear relationship with varying levels of monitoring and discipline.27,36 In contrast, longitudinal data demonstrate greater paternal monitoring knowledge was also associated with reduced adolescent sexual risk behavior at baseline; however, this association waned and reversed direction over time.36
Several of the reviewed studies compared the distinct role of mother and father parenting practices on adolescent sexual behavior. Although many studies relied on cross-sectional designs, greater emotional quality of the father-adolescent relationship, paternal communication about sex, and paternal disapproval of adolescent sexual behavior were associated with reduced or delayed adolescent sexual behavior independently of the equivalent maternal variable. Although study design limitations should be considered, these results provide some preliminary evidence to suggest fathers independently shape the sexual behavior of their adolescent children relative to mothers.
The overall methodological quality of the reviewed studies is a major limitation in our ability to make causal inferences about how adolescent exposure to paternal process variables contributes to change in sexual behavior. For example, nearly half of the studies collected data at 1 time point, which reduces the ability to rule out reverse causation. Additionally, no study reported formal power analyses, so it was not possible to assess if sample size was sufficient to fully detect changes in adolescent outcomes. There was also inconsistency in the reporting of psychometric properties of measured variables. Validity data were particularly underreported.
Another notable limitation to our review includes the dearth of father-specific research addressing varied aspects of adolescent sexual behavior. For example, most articles (8 of 13 studies) assessed paternal influence on only 1 area of adolescent sexual behavior: sexual debut. Less emphasis was given to other domains of adolescent sexual behavior, such as frequency of sex (1), number of sexual partners (2), contraceptive use (2), frequency of condom use (1), pregnancy (1), and composite measures of adolescent sexual behavior, including sexual risk-taking (3). For example, a small number of studies report that paternal variables are associated with changes in adolescent contraceptive use and number of sexual partners. The dearth of research in this area, however, limits our understanding of which paternal influences are most influential in modifying these behaviors.
Also important to mention are limitations of our structured review. For example, it is possible our literature search did not yield all relevant studies and our analysis included only peer-reviewed articles, which may introduce publication bias.41 Additionally, we lacked an adequate number of studies to focus on a specific adolescent developmental period or impose a father residential status requirement. These factors should be considered in future studies. Findings related to which paternal influence was most or least influential to adolescent sexual behavior should be interpreted with care, as the magnitude of each paternal influence was not taken into consideration.
Despite these limitations, our review provides some evidence to support the role of father-specific variables as an important correlate of adolescent sexual behavior. Important to note is although extant research on mothers has developed greatly over the years, the state of mother-specific research on adolescent sexual behavior was once similar to that of fathers. For example, as research on families and mothers has evolved, more complex conceptual models that delineate the specific mediating mechanisms by which parental process variables shape adolescent behavior have been developed.42–44 In contrast, few prevention studies focus on fathers and no reviewed study examined a father-specific theory of influence. A well-defined theoretical framework can inform intervention development by identifying what parenting practices are most amenable to change and likely to yield the biggest impact on adolescent sexual behavior. Although a growing evidence base supports interventions that target parenting practices as a mechanism to decrease sexual risk-taking among adolescents,45,46 these programs are likely to focus on mothers and not fathers.47 This represents a critical missed opportunity to improve the sexual and reproductive health of adolescents, as fathers are key figures in their children’s lives and may parent in ways that differ from mothers. Future intervention development should take advantage of unique paternal influences, as fathers represent an additional opportunity to influence teenage sexual behavior. Greater research adhering to a higher methodological standard should isolate paternal-specific influences by simultaneously evaluating the maternal and paternal process variables examined in this review, as well as other potential mechanisms of father influence. Future research is also needed to identify strategies to best target, recruit, and retain fathers in intervention programs. This research can better inform pediatricians and public health practitioners on how to best involve fathers in interventions to reduce adolescent sexual risk behaviors that lead to unintended pregnancy, STIs, and HIV infection.
Our review makes clear that fathers have the potential to uniquely influence adolescent sexual behavior, yet have been overlooked in family-based intervention development. We argue for greater research to identify potential mechanisms of father-specific influence that will support successful development of father-based interventions and expand the opportunity to support adolescent health and well-being.
- Accepted July 30, 2012.
- Address correspondence to Vincent Guilamo-Ramos, PhD, MPH, LCSW, Silver School of Social Work, New York University, 1 Washington Square North, Rm 320, New York, NY 10003-6654. E-mail:
The views and conclusions in this paper are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention.
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: Supported by the Centers for Disease Control and Prevention Parenting Synthesis Project.
- Guilamo-Ramos V,
- Bouris A,
- Jaccard J,
- Gonzalez B,
- McCoy W,
- Aranda D
- Guilamo-Ramos V,
- Jaccard J,
- Dittus PJ
- Aronowitz T,
- Morrison-Beedy D
- Guilamo-Ramos V
- ↵Garfinkle I, McLanahan S, Meadows S, Mincy R. Unmarried Fathers’ Earning Trajectories: Does Marriage Matter? Princeton, NJ: Princeton University, Woodrow Wilson School of International and Public Affairs, CRCW; 2009. Working Papers 1133
- White HD
- Jemmott LS,
- Jemmott JB
- Akers AY,
- Holland CL,
- Bost J
- ↵Littell JH, Corcoran J, Pillai VK. Systematic reviews and meta-analysis. New York, NY: Oxford University Press; 2008
- Guilamo-Ramos V,
- Jaccard J,
- Dittus PJ,
- Gonzalez B,
- Bouris A
- DiClemente RJ,
- Crittenden CP,
- Rose E,
- et al
- Downing J,
- Jones L,
- Bates G,
- Sumnall H,
- Bellis MA
- Copyright © 2012 by the American Academy of Pediatrics