Evidence suggests that teen dating violence (TDV) is associated with long-term adverse outcomes, but these associations have not been systematically assessed.
To conduct a systematic review of the longitudinal associations between TDV and negative outcomes, including mental and physical health, reoccurrence of violence in intimate relationships, and high-risk behaviors (substance use and sexual behaviors).
Peer-reviewed articles published in English were searched in PsycINFO/Eric/PsycArticles, PubMed, and Web of Science database from inception to November 2022.
Prospective studies that assessed TDV during adolescence, had a follow-up of at least ≥1 year, and evaluated the associations of TDV with health, violence in romantic intimate relationships, or risk behaviors were included.
Study characteristics, baseline data, and follow-up outcomes were extracted from included studies.
Thirty-eight studies involving 23 unique samples were analyzed. Findings showed that TDV in adolescence was associated with future teen dating and intimate partner violence in adulthood. Studies also indicated that TDV was longitudinally associated with increasing high-risk behaviors (ie, marijuana and alcohol use) and poor mental health outcomes (particularly for victimization). There was unclear evidence on the longitudinal link between TDV and suicidal attempts. Significant associations between TDV and negative outcomes were reported more frequently among females compared with males.
Length of follow-up varied across studies.
Dating violence in adolescence may represent a risk factor for a wide range of long-term outcomes. Female adolescents reporting TDV may be at higher risk of adverse outcomes compared with males.
Adolescents are at high risk of experiencing some forms of aggression within dating relationships.1,2 Teen dating violence (TDV) has been defined as a variety of harmful partner-directed behaviors (including physical, sexual, psychological, cyber violence) that occur within adolescent romantic relationships.3 Research has shown that TDV is a significant public health problem, with prevalence rates of ∼20% for physical and 9% for sexual TDV.4 Psychological TDV has been estimated to be the most common form of dating violence among adolescents, with a lifetime prevalence ranging from 17% to 88%.5
Research has also shown that TDV is cross-sectionally associated with a wide range of adverse outcomes. A review on the correlates of dating violence perpetration in young individuals (16–24 years old) reported that aggression in dating relationships was positively associated with psychopathology, including negative affect, general distress, impulsivity, hostility, and antisocial personality features.6 Additionally, dating violence perpetration was found to be linked to high-risk behaviors, such as substance use, as well as with dating violence victimization, and poor conflict and anger management skills.6 Research studies have also reported significant associations between TDV, poor physical health,7 and suicidal ideation.8,9
Although informative, evaluation of such correlates does not allow conclusions to be drawn on the direction and temporal relationships between TDV and adverse outcomes. This is particularly important for the longitudinal assessment of violent behaviors (reoccurrence of violence in romantic relationships) and health-related risks, such as suicidal attempts, drug abuse, and a wide range of adverse mental and physical outcomes.10 Additionally, it is unclear whether TDV might have differential impacts depending on sex and the direction of violence (i.e., victimization/perpetration). Prospective studies on adolescents are therefore critical for strengthening the evidence base of the long-term risks of TDV.11 Various mechanisms have been discussed to explain the process through which TDV could be longitudinally associated with negative outcomes. Theoretically, it is possible that TDV could lead in the long term to further violence, drug use, and negative mental and physical health, as well as that these negative factors could increase the risk of TDV, or that bidirectional relationships exist.12 It is also possible that dating violence occurring in adolescents could be associated with negative long-term adverse health outcomes through the increased risk for continuation of violence from youth to adulthood.13,14
In the past decade, numerous longitudinal studies have been conducted to investigate the long-term relationships between TDV and adverse outcomes. However, findings from these studies have not been systematically assessed and the overall long-term risks of TDV remain unclear. There have been several reviews of prospective studies on intimate partner violence (IPV) involving adult populations. Such studies have shown that adult IPV is significantly and longitudinally associated with increased depressive symptoms, postnatal depression, drug use, cardiovascular risks, and suicide attempts.12,15–18 A comprehensive review of findings from prospective studies involving adolescents might provide new and critical insights concerning the long-term risks of TDV.
The aim of this study was to conduct a systematic review of longitudinal studies that evaluated the long-term associations between teen dating violence and adverse outcomes, including negative physical and mental health, reoccurrence of violence within romantic relationships, and high-risk behaviors (substance use and sexual behaviors). The review also addressed questions about the differential impacts of TDV depending on sex and the direction of violence.
Methods
The current study was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-analyses guidelines.19
Data Sources and Search Strategy
We searched PsycINFO/Eric/PsycArticles, PubMed, and Web of Science database from inception to November 2022. Search terms included indexed English terms unique to each database (eg, PubMed MeSH terms used in specific periods, PsycINFO index), as well as selected text words to enhance search effectiveness. We combined search terms pertaining to dating violence with those of adolescents and longitudinal design. Studies were selected independently by 2 of us (A.P. and F.W.). After elimination of the duplicates, titles and relevant abstracts were reviewed. The full texts of the remaining records were retrieved to determine whether they met all inclusion criteria.
Inclusion and Exclusion Criteria
We included published research articles meeting the following criteria:
the study was prospective;
had multiple time points (follow-up at least ≥1 year);
baseline dating violence was assessed during adolescence (mean age of participants 10–18);
the longitudinal association of dating violence as a risk factor for an outcome was reported; and
the longitudinal outcome included either a measure of mental health, physical health, high-risk behaviors (substance use and sexual behaviors), or reoccurrence of violence in intimate relationships.
Retrospective studies, literature reviews, dissertations, conference abstracts, questionnaire validation studies, and papers published in languages other than English were excluded. Studies or analysis evaluating the trajectories of TDV behaviors were excluded. Research articles reporting solely moderator and mediator analyses without describing the direct associations of TDV and the longitudinal outcome were also excluded.
Data Extraction and Quality Assessment
Two reviewers extracted data independently using a standardized study form, which included:
study information;
characteristics of participants;
baseline outcomes, including types of dating violence and control variables;
outcomes at follow-ups, including measures of mental health, physical health, high-risk behaviors (i.e., substance use, sexual behaviors), or future episodes of violence in intimate relationships; and
results (adjusted analyses were preferred over unadjusted).
The risk of bias (RoB) of each included study was assessed with a modified version of the Newcastle-Ottawa Quality Assessment Scale (NOS).20 The NOS includes a “star system” in which a study is judged on 3 broad perspectives: the selection of the study groups, the comparability of the groups, and the ascertainment of the outcome of interest.20 We excluded from our evaluation the questions that were concerned with the method of assessment of the predictor and outcomes (“ascertainment of exposure” and “assessment of outcome”), given that almost all studies evaluated TDV and dependent variables by self-report scales. We determined a score of 7 on the NOS to indicate “low RoB,” between 5 and 6 indicating “medium RoB,” and <4 “high RoB.”
Nonindependence of Study Samples
Nonindependence of studies was determined by name of the study, author group similarity, and by report of common sample or methodology. If nonindependent studies reported on the same outcome, only the article with the longest follow-up was selected. If nonindependent studies reported on the same outcome using the same follow-up, we selected the article that specifically focused on such outcome.
Results
Selection and Characteristics of Included Studies
We screened 2045 titles and abstracts, removed 1461 (833 duplicates, 628 not relevant), and subsequently retrieved 584 full-text articles (Supplemental Fig 2). Of these, 38 articles from 23 unique studies met the inclusion criteria (Table 1). Articles were published from 2004 to 2022, with 86.8% (n = 33) published since 2010. Most studies, 89.5% (n = 34), were conducted in the United States.
Selected Characteristics of the Included Studies
| Name of Unique Study | Source | Sizea and Characteristics of Sample | Mean Ageb | TDV at Baseline | Follow-Upe | Outcomesf |
|---|---|---|---|---|---|---|
| 1. Project EAT | Ackard et al, 200740 | 1516 adolescents (54% females), predominantly white, recruited from schools | 15.4 y | Composite TDVc | 5 y | High-risk behaviors, mental health |
| 2. National Longitudinal Study of Adolescent Health | Teitelman et al, 200857 | 2629 female adolescents, predominantly white, recruited from schools | 16 yb | Physical violence, psychological violencec | 5.5 y | High-risk behaviors |
| Clark et al, 201449 | 4638 female adolescents, predominantly white, recruited from schools | 16.0 y | Composite TDVc | 5.5 y and 13.5 y | Physical health | |
| Exner-Cortens et al, 201723 | 2161 adolescents (59.1% females), predominantly white, recruited from school | 16 y | Composite TDVc | 5.5 y and 13.5 y | Reoccurrence of violence in intimate relationships | |
| Van Dulmen et al, 201246 | 4675 adolescents (58,9% females), predominantly white, recruited from schools | 16.4 y | Composite TDVc | 5.5 y and 13.5 y | Mental health | |
| Exner-Cortens et al, 201342 | 5681 adolescents (52.3% females), predominantly white, recruited from schools | 16.0 y | Composite TDV, psychological violencec | 5.5 y | High-risk behaviors, mental health | |
| Madkour et al, 201414 | 558 female adolescents, predominantly white, recruited from schools | 16 y | Physical violence, psychological violencec | 13.5 y | Physical health | |
| Deutsch, 201953 | 2097 female adolescents, predominantly white, recruited from schools | 16 yb | Physical violencec | 5.5 y | High-risk behaviors | |
| 3. Toledo Adolescent Relationships Study | Copp et al, 201650 | 1321 adolescents (50.6% females), mixed ethnicity, recruited from the community | 15 y | Physical violencec,d | 1 y, 3 y, 5 y, and 11 y | Physical health |
| Johnson et al, 201444 | 1273 adolescents (52.0% females), mixed ethnicity, recruited from the community | 15 y | Composite TDVc,d | 1 y, 3 y, and 5 y | Mental health | |
| 4. Dating Violence Among Latino Adolescents Study | Cuevas et al, 202035 | 1525 adolescents (53.0% females), predominantly Latino, recruited from the community | 14.7 y | Composite TDVc | 1 y | Reoccurrence of violence in intimate relationships |
| 5. [No name] | Herrenkohl and Jung, 201626 | 457 adolescents (45.7% females), predominantly white, recruited from child welfare and the community | 18.2 y | Composite TDVc,d | 4.5 y, 14.5 y, and 35 y | Reoccurrence of violence in intimate relationships |
| 6. Dating it save | Temple et al, 2016a33 | 780 adolescents (57% females), mixed ethnicity, recruited from schools | 18.1 y | Cyber dating abuse, physical violence, psychological violencec,d | 1 y | Reoccurrence of violence in intimate relationships |
| Temple et al, 2016b41 | 1042 adolescents (57% females), mixed ethnicity, recruited from schools | 15.1 y | Psychological violenced | 3 y | Mental health | |
| Jouriles et al, 201727 | 843 adolescents (57% females), mixed ethnicity, recruited from schools | 16.1 y | Physical violencec | 5 y | Reoccurrence of violence in intimate relationships | |
| Shorey et al, 201951 | 1042 adolescents (57% females), mixed ethnicity, recruited from schools | 15.1 y | Physical violence, psychological violence, sexual violencec | 5 y | High-risk behaviors | |
| Vanwoerden et al, 201948 | 818 adolescents (57% females), mixed ethnicity, recruited from schools | 16.1 y | Physical violence, psychological violence, sexual violencec | 5 y | Mental health | |
| Shorey et al, 202138 | 1042 adolescents (57% females), mixed ethnicity, recruited from schools | 15.1 y | Physical violence, psychological violenced | 6 y | Reoccurrence of violence in intimate relationships, mental health | |
| 7. National Survey of Adolescents-Replication | Walsh et al, 201755 | 3614 adolescents (48.8% females), predominantly white, recruited from the community | 14.6 y | Composite TDVc | 1 y and 2 y | High-risk behaviors |
| 8. Context Study | Foshee et al, 201343 | 3328 adolescents (51% females), mixed ethnicity, recruited from schools | 15 yb | Physical violence, psychological violencec | 2.5 y | High-risk behaviors, mental health |
| Foshee et al, 201645 | 3979 adolescents (51% females), mixed ethnicity, recruited from schools | 15 yb | Physical violenced | 2.5 y | High-risk behaviors, mental health | |
| 9. [No name] | Goodcase et al, 202025 | 285 pregnant or parenting adolescents (69.1% females), predominantly Latino, recruited from schools | 16.9 y | Physical violence, psychological violenced | 1.5 y | Reoccurrence of violence in intimate relationships |
| 10. Second Step: Student Success Through Prevention | Holt et al, 201856 | 220 adolescents (45% females), mixed ethnicity, recruited from schools | 15 y | Psychological violencec | 1 y | High-risk behaviors |
| 11. Healthy Teens Longitudinal Study | Nahapetyan et al, 201447 | 745 adolescents (49.8% females), mixed ethnicity, recruited from schools | 15 y | Physical violencec,d | 6 y | Mental health |
| Orpinas et al, 201752 | 745 adolescents (48% females), mixed ethnicity, recruited from schools | 15 y | Physical violenced | 6 y | High-risk behaviors | |
| 12. [No name] | Reyes et al, 201254 | 2636 adolescents (53% females), mixed ethnicity, recruited from schools | 13 y | Physical violenced | 2 y | High-Risk behaviors |
| 13. [No name] | Sabol et al, 202028 | 1240 adolescents (51.9% females), mixed ethnicity, recruited from schools | 15 yb | Physical violence, psychological violence, sexual violenced | 4 y | Reoccurrence of violence in intimate relationships |
| 14. [No name] | Yu et al, 201832 | 238 adolescents (58% females), recruited from schools in Canada | 14 y | Composite TDVd | 2 y | Reoccurrence of violence in intimate relationships, mental Health |
| 15. [No name] | Fernández-González et al, 202024 | 991 adolescents (52.4% females), predominantly Spanish, recruited from schools | 14.8 y | Composite TDVc,d | 4 y | Reoccurrence of violence in intimate relationships |
| 16. [No name] | Timmons Fritz and Smith Slep, 200930 | 664 adolescents (50.3% females), mixed ethnicity, recruited from schools | 16.3 y | Physical violence, psychological violencec,d | 1 y | Reoccurrence of violence in intimate relationships |
| 17. [No name] | Williams et al, 200831 | 621 adolescents (59% females), recruited from schools in Canada | 15.2 y | Physical violencec,d | 1 y | Reoccurrence of violence in intimate relationships |
| 18. [No name] | Wolfe et al, 200422 | 1317 adolescents (55% females), predominantly white, recruited from schools | 16.1 y | Composite TDVd | 1 y | Reoccurrence of violence in intimate relationships |
| 19. Safe Dates | Reyes et al, 201521 | 1965 adolescents (50% females), predominantly white, recruited from schools | 15 yb | Physical violenced | 1y | Reoccurrence of violence in intimate relationships |
| 20. Quebec Youth Romantic Relationships Survey | Smith et al, 202229 | 4923 adolescents (59.6% females), predominantly white, recruited from schools in Canada | 15.5 y | Composite TDVc | 3 yb | Reoccurrence of violence in intimate relationships |
| 21. Nationally representative Survey on Teen Relationships and Intimate Violence | Mumford et al, 201936 | 261 adolescents (47.1% females), mixed ethnicity, conducted with a random sample of households | 15.3 y | Physical violence, psychological violence, sexual violencec | 1 y and 2 y | Reoccurrence of violence in intimate relationships, mental health |
| Taylor et al, 202134 | 346 adolescents (49.1% females), mixed ethnicity, conducted with a random sample of households | 14 y | Physical violence, psychological violence, sexual violencec,d | 1 y | Reoccurrence of violence in intimate relationships | |
| 22. [No name] | Hautala and Sittner, 202139 | 521 adolescents (51% females), indigenous youth, recruited from the community | 17.5 yb | Physical violencec | 1 y | High-risk behaviors, mental health |
| 23. Québec Longitudinal Study of Child Development | Psychogiou et al, 202237 | 974 adolescents (53.2% females), predominantly white, recruited from the community | 15 y | Composite TDVc,d | 2 y | Reoccurrence of violence in intimate relationships, mental health |
| Name of Unique Study | Source | Sizea and Characteristics of Sample | Mean Ageb | TDV at Baseline | Follow-Upe | Outcomesf |
|---|---|---|---|---|---|---|
| 1. Project EAT | Ackard et al, 200740 | 1516 adolescents (54% females), predominantly white, recruited from schools | 15.4 y | Composite TDVc | 5 y | High-risk behaviors, mental health |
| 2. National Longitudinal Study of Adolescent Health | Teitelman et al, 200857 | 2629 female adolescents, predominantly white, recruited from schools | 16 yb | Physical violence, psychological violencec | 5.5 y | High-risk behaviors |
| Clark et al, 201449 | 4638 female adolescents, predominantly white, recruited from schools | 16.0 y | Composite TDVc | 5.5 y and 13.5 y | Physical health | |
| Exner-Cortens et al, 201723 | 2161 adolescents (59.1% females), predominantly white, recruited from school | 16 y | Composite TDVc | 5.5 y and 13.5 y | Reoccurrence of violence in intimate relationships | |
| Van Dulmen et al, 201246 | 4675 adolescents (58,9% females), predominantly white, recruited from schools | 16.4 y | Composite TDVc | 5.5 y and 13.5 y | Mental health | |
| Exner-Cortens et al, 201342 | 5681 adolescents (52.3% females), predominantly white, recruited from schools | 16.0 y | Composite TDV, psychological violencec | 5.5 y | High-risk behaviors, mental health | |
| Madkour et al, 201414 | 558 female adolescents, predominantly white, recruited from schools | 16 y | Physical violence, psychological violencec | 13.5 y | Physical health | |
| Deutsch, 201953 | 2097 female adolescents, predominantly white, recruited from schools | 16 yb | Physical violencec | 5.5 y | High-risk behaviors | |
| 3. Toledo Adolescent Relationships Study | Copp et al, 201650 | 1321 adolescents (50.6% females), mixed ethnicity, recruited from the community | 15 y | Physical violencec,d | 1 y, 3 y, 5 y, and 11 y | Physical health |
| Johnson et al, 201444 | 1273 adolescents (52.0% females), mixed ethnicity, recruited from the community | 15 y | Composite TDVc,d | 1 y, 3 y, and 5 y | Mental health | |
| 4. Dating Violence Among Latino Adolescents Study | Cuevas et al, 202035 | 1525 adolescents (53.0% females), predominantly Latino, recruited from the community | 14.7 y | Composite TDVc | 1 y | Reoccurrence of violence in intimate relationships |
| 5. [No name] | Herrenkohl and Jung, 201626 | 457 adolescents (45.7% females), predominantly white, recruited from child welfare and the community | 18.2 y | Composite TDVc,d | 4.5 y, 14.5 y, and 35 y | Reoccurrence of violence in intimate relationships |
| 6. Dating it save | Temple et al, 2016a33 | 780 adolescents (57% females), mixed ethnicity, recruited from schools | 18.1 y | Cyber dating abuse, physical violence, psychological violencec,d | 1 y | Reoccurrence of violence in intimate relationships |
| Temple et al, 2016b41 | 1042 adolescents (57% females), mixed ethnicity, recruited from schools | 15.1 y | Psychological violenced | 3 y | Mental health | |
| Jouriles et al, 201727 | 843 adolescents (57% females), mixed ethnicity, recruited from schools | 16.1 y | Physical violencec | 5 y | Reoccurrence of violence in intimate relationships | |
| Shorey et al, 201951 | 1042 adolescents (57% females), mixed ethnicity, recruited from schools | 15.1 y | Physical violence, psychological violence, sexual violencec | 5 y | High-risk behaviors | |
| Vanwoerden et al, 201948 | 818 adolescents (57% females), mixed ethnicity, recruited from schools | 16.1 y | Physical violence, psychological violence, sexual violencec | 5 y | Mental health | |
| Shorey et al, 202138 | 1042 adolescents (57% females), mixed ethnicity, recruited from schools | 15.1 y | Physical violence, psychological violenced | 6 y | Reoccurrence of violence in intimate relationships, mental health | |
| 7. National Survey of Adolescents-Replication | Walsh et al, 201755 | 3614 adolescents (48.8% females), predominantly white, recruited from the community | 14.6 y | Composite TDVc | 1 y and 2 y | High-risk behaviors |
| 8. Context Study | Foshee et al, 201343 | 3328 adolescents (51% females), mixed ethnicity, recruited from schools | 15 yb | Physical violence, psychological violencec | 2.5 y | High-risk behaviors, mental health |
| Foshee et al, 201645 | 3979 adolescents (51% females), mixed ethnicity, recruited from schools | 15 yb | Physical violenced | 2.5 y | High-risk behaviors, mental health | |
| 9. [No name] | Goodcase et al, 202025 | 285 pregnant or parenting adolescents (69.1% females), predominantly Latino, recruited from schools | 16.9 y | Physical violence, psychological violenced | 1.5 y | Reoccurrence of violence in intimate relationships |
| 10. Second Step: Student Success Through Prevention | Holt et al, 201856 | 220 adolescents (45% females), mixed ethnicity, recruited from schools | 15 y | Psychological violencec | 1 y | High-risk behaviors |
| 11. Healthy Teens Longitudinal Study | Nahapetyan et al, 201447 | 745 adolescents (49.8% females), mixed ethnicity, recruited from schools | 15 y | Physical violencec,d | 6 y | Mental health |
| Orpinas et al, 201752 | 745 adolescents (48% females), mixed ethnicity, recruited from schools | 15 y | Physical violenced | 6 y | High-risk behaviors | |
| 12. [No name] | Reyes et al, 201254 | 2636 adolescents (53% females), mixed ethnicity, recruited from schools | 13 y | Physical violenced | 2 y | High-Risk behaviors |
| 13. [No name] | Sabol et al, 202028 | 1240 adolescents (51.9% females), mixed ethnicity, recruited from schools | 15 yb | Physical violence, psychological violence, sexual violenced | 4 y | Reoccurrence of violence in intimate relationships |
| 14. [No name] | Yu et al, 201832 | 238 adolescents (58% females), recruited from schools in Canada | 14 y | Composite TDVd | 2 y | Reoccurrence of violence in intimate relationships, mental Health |
| 15. [No name] | Fernández-González et al, 202024 | 991 adolescents (52.4% females), predominantly Spanish, recruited from schools | 14.8 y | Composite TDVc,d | 4 y | Reoccurrence of violence in intimate relationships |
| 16. [No name] | Timmons Fritz and Smith Slep, 200930 | 664 adolescents (50.3% females), mixed ethnicity, recruited from schools | 16.3 y | Physical violence, psychological violencec,d | 1 y | Reoccurrence of violence in intimate relationships |
| 17. [No name] | Williams et al, 200831 | 621 adolescents (59% females), recruited from schools in Canada | 15.2 y | Physical violencec,d | 1 y | Reoccurrence of violence in intimate relationships |
| 18. [No name] | Wolfe et al, 200422 | 1317 adolescents (55% females), predominantly white, recruited from schools | 16.1 y | Composite TDVd | 1 y | Reoccurrence of violence in intimate relationships |
| 19. Safe Dates | Reyes et al, 201521 | 1965 adolescents (50% females), predominantly white, recruited from schools | 15 yb | Physical violenced | 1y | Reoccurrence of violence in intimate relationships |
| 20. Quebec Youth Romantic Relationships Survey | Smith et al, 202229 | 4923 adolescents (59.6% females), predominantly white, recruited from schools in Canada | 15.5 y | Composite TDVc | 3 yb | Reoccurrence of violence in intimate relationships |
| 21. Nationally representative Survey on Teen Relationships and Intimate Violence | Mumford et al, 201936 | 261 adolescents (47.1% females), mixed ethnicity, conducted with a random sample of households | 15.3 y | Physical violence, psychological violence, sexual violencec | 1 y and 2 y | Reoccurrence of violence in intimate relationships, mental health |
| Taylor et al, 202134 | 346 adolescents (49.1% females), mixed ethnicity, conducted with a random sample of households | 14 y | Physical violence, psychological violence, sexual violencec,d | 1 y | Reoccurrence of violence in intimate relationships | |
| 22. [No name] | Hautala and Sittner, 202139 | 521 adolescents (51% females), indigenous youth, recruited from the community | 17.5 yb | Physical violencec | 1 y | High-risk behaviors, mental health |
| 23. Québec Longitudinal Study of Child Development | Psychogiou et al, 202237 | 974 adolescents (53.2% females), predominantly white, recruited from the community | 15 y | Composite TDVc,d | 2 y | Reoccurrence of violence in intimate relationships, mental health |
Some studies did not have a unique name; those have been marked as “[No name].” EAT, Eating and Activity Over Time.
At baseline (relative to the outcome analyzed).
The measure was approximated from available data.
TDV victimization.
TDV perpetration.
Included in analysis.
Selected outcome.
The mean age of participants at baseline ranged from 13 to 18. In 8 unique studies (14 articles), the ethnicity of the sample was primarily white, in 10 mixed (18 articles), predominantly Latino in 2 studies (2 articles), and the sample of 1 remaining study was Indigenous youth in Canada. Most studies recruited participants through schools (73.7%, n = 28).
Included articles examined 4 different types of TDV: physical violence (n = 22), psychological/emotional violence (n = 15), sexual violence (n = 5), and cyber abuse (n = 1). Furthermore, 14 articles used composite scores of TDV (combining different forms of TDV). Twenty-eight articles assessed TDV victimization and 20 TDV perpetration. The length of follow-up ranged from 1 year to 35 years. Specifically, follow-ups were up to 1 year in 8 articles, 2 to 5 years in 17 articles, and longer than 5 years in 13 articles. Thirty articles used 2 time points, and 8 articles used 3 to 5 time points. Four health outcomes were assessed longitudinally in relation to TDV: reoccurrence of violence in intimate relationships (n = 18), mental health (n = 14), physical health (n = 3), and high-risk behaviors (n = 12).
Evidence Synthesis
Reoccurrence of Violence in Intimate Relationships
Eighteen articles from 15 unique studies21–38 investigated the longitudinal associations between TDV and reoccurrence of violence in adolescent or adult intimate relationships.
Twelve studies found that TDV victimization was significantly associated with future violence in adolescent intimate relationships.21,24,26,27,29–31,33–37 Specifically, 4 studies24,29,35,36 reported significant stability of TDV victimization measured through a composite score, over follow-up periods of 1 to 4 years. Studies also found stability for psychological and physical dating victimization over 130,31 and 234 years. Furthermore, Taylor et al34 showed that sexual or physical dating victimization at baseline predicted victimization 1 year later. Similar results were reported by Temple et al33 relative to a 1-year study on the stability of cyber dating abuse.
Consistent with the findings on TDV victimization, 11 studies found significant stability of TDV perpetration over time.21,22,24,25,28,30–32,34,37,38 Of these, 5 studies22,24,28,32,37 showed that a composite measure of TDV perpetration was stable over follow-ups ranging from 1 year to 4 years. Four studies21,30,31,34 reported stability of physical dating perpetration over periods of 1 to 2 years, and 225,30 articles reported stability for psychological perpetration with a 1-year follow-up period. Shorey et al38 found stability for both physical and psychological violence over a period of 6 years. Furthermore, psychological perpetration was found to predict physical perpetration across all waves of the study, whereas physical perpetration in the first year predicted psychological perpetration only in the second year. Similarly, Taylor et al34 showed that psychological TDV perpetration predicted psychological, physical, and sexual TDV perpetration 1 year later.
Fewer studies assessed the associations of TDV with future IPV in adult relationships.23,26,27 Exner-Cortens et al23 found that adolescents who experienced general TDV were at significantly higher risk of physical IPV victimization 5 years later, compared with adolescents who did not experience TDV at baseline. In the same study, TDV predicted physical IPV victimization also at a 12-year follow-up, but only through the experience of early adulthood IPV victimization. Consistent findings were described by Herrenkohl and Jung.26 In their study, general TDV victimization significantly predicted adult overall IPV victimization and perpetration ∼20 years later. Lastly, another study27 found that physical TDV victimization was associated with recurrent victimization in early adulthood.
Mental Health
Internalizing Symptoms
Five studies found that both TDV victimization and perpetration were significantly and longitudinally associated with internalizing symptoms (eg, depressive and anxiety symptoms) across a period of 2.5 to 5.5 years.32,40–43, The same significant association was also found in a study by Johnson et al44 over a period of 15 years using a composite measure.
Research that examined only victimization found significant links between TDV and internalizing symptoms especially among girls. In particular, Foshee et al43 reported significant associations between psychological TDV victimization and increased internalizing symptoms at a 2.5-year follow-up only among the female sample. Similarly, 2 studies found composite TDV victimization to be associated, over 5 to 6 years, with depressive symptoms only among girls.40,42 On the contrary, a study by Mumford et al36 found TDV victimization to predict a worse mental health status, particularly anxiety and depression, 2 years later only among males. Furthermore, Hautala and Sittner39 found that physical TDV victimization predicted higher depressive symptoms 1 year later.
When perpetration was examined alone, Temple et al41 showed that baseline physical TDV significantly correlated with depressive and anxiety symptoms over a period of 3 years among both males and females. On the contrary, 2 studies32,45 did not find that TDV perpetration was significantly associated with internalizing symptoms over a 2-year period. Lastly, a study by Psychogiou et al37 showed that neither TDV victimization nor perpetration was related to internalizing symptoms over a follow-up of 2 years.
Suicide
Four studies investigated suicidal ideation40,42,46,47 and 2 suicidal attempts.40,46 For suicidal ideation, significant results were reported in 2 studies,42,47 whereas results for suicidal attempts were inconclusive. Specifically, a longitudinal research by Nahapetyan et al47 found that physical TDV perpetration and victimization significantly predicted suicidal ideation over a period of 6 years, and the association was stronger among females compared with males. Exner-Cortens et al42 examined results for males and females separately. They found across a 5.5-year period that psychological TDV victimization was significantly associated with suicidal ideation among males and that a combined score of TDV victimization was significantly linked to suicidal ideation among females. Ackard et al40 showed that both female and male adolescents experiencing TDV victimization at baseline reported higher odds of suicidal ideation and suicidal attempts over a period of 5 years. However, this association only approached significance, because the study may have been underpowered to detect significant effects.40 In another study,46 bivariate associations between TDV victimization and suicide ideation were small but statistically significant. However, TDV victimization did not significantly predict self-reported suicidal ideation or attempts over a period of 13.5 years, when controlling for past history of victimization in a cross-lagged model. Of note, none of the studies on suicidality included health records, but rather relied on self-reports from adolescents.
Other Mental Health Outcomes
A study by Shorey et al38 found that psychological perpetration at year 1 and 4 predicted the severity of PTSD symptoms in the following year. Another publication48 from the same unique study showed that psychological and sexual victimization, but not physical TDV, predicted greater borderline personality traits among females across a period of 5 years. These associations were not replicated in the male sample. Ackard et al40 examined the link between a composite measure of TDV victimization and binge eating across a period of 5 years. Both males and females showed higher odds of reporting binge eating over time, but these associations were not statistically significant.
Physical Health
BMI
A study by Clark et al49 showed that composite TDV victimization significantly predicted increased BMI over a period of 13.5 years. This association was only significant among females.
Birth Weight
Madkour et al14 investigated the longitudinal link between physical TDV victimization during prepregnancy and weight of infants at birth, over a period of 13.5 years. Findings revealed that, among the subsample of Black mothers, TDV was significantly associated with lower weight of their infants at birth. No significant associations were found among other ethnic groups of mothers.
Self-Rated General Health
One study by Copp et al50 evaluated TDV and self-rated general health over a period of 11 years. The results showed that physical TDV (perpetration and victimization) was significantly linked to a deterioration in self-reported health status.
High-Risk Behaviors
Cigarette Smoking
Four different studies40,42,43,45 found that TDV victimization predicted future tobacco smoking. Specifically, 2 studies40,43 reported a significant association between composite TDV victimization and cigarette smoking regardless of sex, over a period of 2.5 and 5 years. Exner et al42 found such association to be significant only among females, using a 5.5-year follow-up. One study45 reported on perpetration and found no significant links between TDV and smoking during a 2.5-year longitudinal research when using generalized linear mixed models and controlling for other covariates.
Marijuana Use and Substance Use
Six articles40,42,43,45,51,52 reported significant longitudinal associations between TDV and marijuana use. In 2 studies,40,43 TDV victimization predicted marijuana use only among females, over a period of 5 years. Conversely, Exner-Cortens et al42 found a significant link only among males. Shorey et al51 found significant correlations between psychological, physical, and sexual victimization and the use of marijuana over the course of 5 years regardless of sex. Concerning perpetration, 2 studies45,52 found that physical TDV predicted cannabis use across a period of 2.5 and 6 years, respectively.
Lastly, Hautala and Sittner39 found that physical TDV victimization predicted the likelihood of substance use disorder over a period of 1 year.
Alcohol Use
Nine studies40,42,43,45,51–55 found future consumption of alcohol to be significantly associated with TDV at baseline. Specifically, 1 study51 reported significant correlations between different types of TDV victimization and alcohol use over a period of 5 years. Two articles42,43 showed that psychological TDV victimization significantly predicted increased alcohol use or heavy episodic drinking during a period of 2.5 and 5.5 years. Similarly, Walsh et al55 found that a composite measure of TDV victimization significantly predicted binge drinking during the course of 2 years. Another study,53 revealed that physical TDV victimization in adolescence was significantly associated with higher odds of alcohol-exposed pregnancy at a 5-year follow-up. Only 1 study40 reported no significant associations between baseline composite TDV victimization and alcohol, among both males and females, using a 5-year follow-up. When perpetration was examined, 2 studies52,54 showed that physical TDV was longitudinally associated with alcohol use at a 2- and 6-year follow-up. In contrast, Foshee et al45 did not find physical TDV perpetration to be significantly and prospectively associated with alcohol use at a 2.5-year follow-up.
Sexual Risk Behaviors
Three studies51,56,57 found TDV to be linked with sexual risk behaviors. More specifically, Shorey et al51 showed that psychological and physical TDV victimization, but not sexual TDV, predicted sexual risk behaviors over a time range of 6 years. Psychological victimization was also found to be associated with sex under the influence of alcohol after 1 year among adolescents aged 16 years.56 Teitelman et al57 found no form of TDV victimization to be significantly associated with unprotected sex after 1 year. However, at a 5.5-year follow-up, there was a significant correlation between physical or psychological violence and unprotected sex.
Quality Appraisal of Included Studies
The RoB for each individual study is presented in Supplemental Table 2. To facilitate interpretation of the results, Fig 1 shows significant and nonsignificant results reported by the studies for TDV victimization and perpetration in relation to the RoB. Specifically, 11 studies out of 12 evaluating reoccurrence of TDV victimization found significant results, and they were judged as presenting a low–medium RoB. Eight studies found TDV victimization to be significantly associated with negative mental health outcomes (in all analyses or among subgroups of sample or type of violence, such as psychological, physical, and sexual TDV), whereas 1 study did not find significant results, and they were judged to present a low– medium RoB in the majority of the cases. Furthermore, significant associations were found between TDV victimization and physical health (3 studies) and high-risk behaviors (9 studies), with an overall low– medium RoB.
Twelve out of 14 studies found significant stability of TDV perpetration; 2 of these studies were judged as high RoB, and the remaining as low–medium RoB. Five studies found significant associations between TDV perpetration and long-term negative mental health, whereas 4 did not find significant results. These 4 studies were judged as presenting a medium–high RoB. Three studies found significant associations between TDV perpetration and physical health (1 study) and high-risk behaviors (2 studies), and they were judged as having a medium RoB.
Discussion
This study systematically reviewed the associations between dating violence occurring during adolescence and specific long-term adverse outcomes. Understanding longitudinal associations is important to shed light on the long-term risks of teen dating violence and, particularly, clarify the role of victimization and perpetration. Overall, evidence shows that TDV victimization and perpetration in adolescence were longitudinally associated with a series of adverse outcomes, and that female adolescents reporting dating violence may be at higher risk for experiencing such outcomes compared with males.
The largest and more consistent amount of evidence was found for reoccurrence of violence in intimate relationships, which supported the long-term association of TDV victimization and perpetration with future partner-related violence, both in adolescence and adulthood. These findings suggest that violence in intimate relationships may form part of a continuum whereby individuals experience and/or perpetrate such violence from adolescence through adulthood.58
Evidence on TDV and poor mental health, such as depressive/anxiety symptoms and suicidal ideation, provided support on their long-term association, particularly for dating victimization. Findings on TDV perpetration were more contrasting, and the majority of studies were evaluated as presenting medium–high RoB. These results might also suggest that victimization and perpetration could be associated with different forms of mental health outcomes. Evidence on the links between TDV victimization and suicidal attempts was also inconclusive, because of the limited number of available studies.
The associations of TDV with physical health have been investigated by few studies, indicating an important gap in the literature. Nonetheless, the evidence suggests that dating violence experienced during adolescence might be associated with a decline in self-rated physical health, and with worse birth outcomes and increased BMI among diverse groups. These longitudinal associations might be explained by the continuation of relationship violence from adolescence to adulthood. Among adult women, the experience of long and continuous IPV is associated with increased levels of depression, PTSD, and somatization,13 which could influence weight, birth outcomes, and overall health via maladaptive behaviors and stress pathways.14
Several long-term associations between TDV and high-risk behaviors emerged on the basis of our systematic synthesis of studies. Specifically, the majority of the studies reported that TDV experienced during adolescence significantly predicted future cigarette smoking, and alcohol and marijuana use. TDV was also significantly associated with increased sexual risk behaviors, such as unprotected sex or sexual intercourses under the influence of alcohol. Such findings were consistent across both TDV victimization and perpetration. Substance use can be viewed as both an outcome and a coping mechanism of dealing with an abusive relationship.42 Additionally, a recent study has highlighted a genetic predisposition that, interacting with the experience of dating violence, increases the odds of substance use.59 Concerning sexual behaviors, the link between TDV and unprotected sex might be partially explained by the lower self-efficacy of dating violence/IPV in terms of condom use negotiation.60 Adolescents experiencing dating violence may further reduce self-confidence in addressing future abusive situations during sexual intercourse.61
Lastly, evidence from this review indicates that females reporting TDV might be at higher risk of longitudinal adverse outcomes, compared with males. Several studies found significant associations between TDV and poorer mental health only among females, which might be explained by the greater severity of victimization experienced by females, as well as greater importance attributed by females to having positive intimate relationships.43
The results of the present review should be interpreted in the context of some limitations, including the potential for publication bias and the nonindependence of some study samples. Studies also varied in the time frame and the type of model and covariates that were included, which may explain some of the discrepancies across studies. Moreover, it is possible that our search strategy did not identify all studies on the long-term associations of TDV and adverse outcomes. In this review, we aimed to provide a global assessment of the long-term associations of TDV with health-related outcomes, and we did not specifically analyze the outcome differences associated with the various forms of dating violence.
Future research should further clarify whether poor long-term health-related outcomes are mediated by the persistence and reoccurrence of partner-related violence. Furthermore, there is also need for research to assess the protective factors against long-term negative outcomes among adolescents exposed to dating violence. This review further highlights the importance of prevention programs for dating violence among adolescents. Prevention programs have demonstrated efficacy in reducing teen dating violence,62,63 thereby showing that TDV can be effectively targeted.
Acknowledgments
We thank the research assistant, Anca Ratiu, BSc (University of Klagenfurt), for conducting the second screening and data extraction.
Dr Piolanti conceptualized and designed the study; Ms Waller screened the articles and extracted the data; Mr Schmid extracted the data; Dr Foran conceptualized the study and critically reviewed the manuscript for intellectual content; and all authors drafted the manuscript and approved the final version as submitted and are accountable for all aspects of the work.
FUNDING: No external funding.
CONFLICT OF INTEREST DISCLOSURES: The authors have no conflicts of interest to disclose.
Comments