Adolescent Depressive Symptoms and Subsequent Pregnancy, Pregnancy Completion and Pregnancy Termination in Young Adulthood: Findings from the Victorian Adolescent Health Cohort Study

Article Outline

• Abstract
• Introduction
• Materials and Methods
  • Procedure and Sample
  • Measures
  • Potential Confounding Factors
  • Data Analysis
• Results
• Discussion
• Acknowledgments
• References
• Copyright

Abstract 

Study Objective

To examine relationships between depressive symptoms in adolescence (14-18 years of age) and becoming pregnant, completing a pregnancy (live birth) and terminating a pregnancy in young adulthood (21-24 years of age).

Participants and Design

Data from 1000 females were drawn from a larger sample of 1943 young Australians participating in a longitudinal study of adolescent health and development, followed across 8 waves from adolescence (waves 1-6) to young adulthood (waves 7 and 8).

Setting

Victoria, Australia.

Main Outcome Measures

Pregnancy, pregnancy completion and pregnancy termination between 21-24 years of age.

Results

We observed a twofold increase in the odds of becoming pregnant in those reporting persisting patterns of depressive symptoms during adolescence (2+ waves); however, after staged adjustment for adolescent antisocial behaviour, drug use and socioeconomic disadvantage, there was no evidence of association. Of particular note, and consistent with previous research, adolescent antisocial and drug use behavior were strongly associated with becoming pregnant and pregnancy termination in young adulthood.

Conclusions

Adolescent antisocial and drug use behavior, not depressive symptoms, independently predict pregnancy outcomes in young adulthood.

Key Words: Pregnancy, Pregnancy completion, Pregnancy termination, Young Adulthood, Adolescence, Depression, Antisocial Behaviour, and Drug Use, Cohort, Longitudinal

Back to Article Outline

Introduction 

Patterns of sexual and reproductive behavior have changed dramatically. In Australia, from 1971 to 2008, the median age of all women giving birth increased from 25.4 to 30.7-years.¹ Similar trends are apparent in other industrialized countries such as the United Kingdom and Canada.², ³ This generational delay in the age of child bearing means that current concepts of early pregnancy have extended to include pregnancy in young adulthood. Predictors of adolescent pregnancy have been extensively studied; however, it is unclear whether the same factors hold for pregnancy in young adulthood and which additional developmental influences might come into play.

Well-established risks for adolescent pregnancy and parenthood include disrupted family structure and socioeconomic disadvantage,⁴, ⁵, ⁶ conduct problems and antisocial behavior,⁷, ⁸, ⁹, ¹⁰ and alcohol and drug use.⁴, ⁹ The effect of adolescent depression is more controversial. Whereas some studies report associations between adolescent depression, pregnancy, and parenthood,⁹, ¹¹, ¹² others have been unable to demonstrate such links.¹³, ¹⁴, ¹⁵

There have been few studies of predictors of pregnancy in young adulthood. Findings from a New Zealand cohort suggest that many of the same risks for pregnancy and parenthood in adolescence (antisocial behavior, drug use, family climate) also apply to pregnancy and parenthood in young adulthood.¹⁶ To our knowledge, there have been only 2 prospective studies examining the relationship between a history of depression and pregnancy in young adulthood, and neither reported an association.¹⁷, ¹⁸ However, both studies examined pregnancy mainly in adolescents (17-21 years¹⁷; 16-21 years¹⁸), and neither controlled for well-established externalizing risk factors such as antisocial behavior and drug use. Furthermore, neither study examined other outcomes of pregnancy in young adulthood, specifically, termination of pregnancy.

Most studies of associations between depression and termination are either retrospective or cross-sectional in design, limiting causal inference.⁹, ¹⁹, ²⁰ However, a recent prospective study did report an association between adolescent depression and pregnancy termination in adulthood.²¹ The association remained after adjustment for known confounders, including disruptive disorders (attention deficit hyperactive disorder, oppositional defiant disorder, or conduct disorder) and substance use (tobacco, alcohol, or illicit drugs). In addition to clarifying relationships between depression and continuation of a pregnancy to a live birth, there is an equally important need to clarify relationships between depression and the choice to terminate a pregnancy.

We examine the relationship between adolescent depressive symptoms and pregnancy outcomes in young adulthood using data from a longitudinal study of adolescent health and development (Victorian Adolescent Health Cohort Study, est. 1992). Specifically, we sought to examine associations between adolescent depressive symptoms (14-18 years) and becoming pregnant, pregnancy completion (live birth), and pregnancy termination in young adulthood (21-24 years). Analyses were adjusted for indicators of adolescent antisocial and drug use behavior and socioeconomic status.

Back to Article Outline

Materials and Methods 

Procedure and Sample 

Between August 1992 and March 2003, we conducted an 8-wave cohort study of adolescent and young adult health in the state of Victoria, Australia. Data collection protocols were approved by The Royal Children’s Hospital Ethics in Human Research Committee. Informed consent was obtained from participating schools and parents. Further participant consent was obtained in the young adult waves. The cohort was initially defined with a 2-stage cluster sample in which 2 Year 9 classes were selected at random from each of 44 schools drawn from a stratified frame of all secondary schools (government, Catholic, and independent) in Victoria (total number of Year 9 students 60,905 in 1993). School retention rates to Year 9 in the year of sampling were 98% in Victoria.

One class from each school entered the cohort in the latter part of the ninth school year (wave 1), and the second class entered 6 months later, early in the 10th school year (wave 2). Participants were subsequently reviewed at a further four 6-month intervals during the teens (waves 3 to 6), with 2 follow-up waves in young adulthood (wave 7, 20-21 years; wave 8, 24-25 years). In waves 1 to 6, participants self-administered the questionnaire on laptop computers,²² with telephone follow-up of those absent from school. The seventh and eighth waves of data collection were undertaken via computer-assisted telephone interviews.

From a total sample of 2032 students, 1943 (95.6%) participated at least once during the first 6 (adolescent) waves. In wave 8, 1520 (78% of the 1943 participants) were interviewed between April 2002 and June 2003. Response rates are shown in Figure 1. Reasons for noncompletion at wave 8 were refusal (n = 267), loss of contact (n = 150), and death (n = 6). The mean age at waves 1 and 8 were 14.9 and 24.1 years, respectively. The present study was based on 777 female participants with complete data on depressive symptoms in waves 2-6 (14-18 years) and complete data on pregnancy and pregnancy outcomes in waves 7 and 8 (21-24 years). Missing data were handled using the methods of multiple imputation (described below).

• 
  • View Large Image
  • Download to PowerPoint

• Fig. 1 

  Sampling and ascertainment in the Victorian Adolescent Health Cohort, 1992 to 2003.

Measures 

Young adult pregnancy outcomes were assessed at 21-24 years by asking participants: “Have you ever had any children of your own?” and “Have you ever had a termination or abortion?” As the focus of the analysis was on pregnancy outcomes in young adulthood, pregnancies (live births, terminations, and miscarriage) occurring before age 21 years were excluded from analysis.

Adolescent depressive symptoms were assessed at 14-18 years using a self-administered, computerized form of the revised Clinical Interview Schedule (CIS-R).²³, ²⁴, ²⁵ The CIS-R provides information on frequency, severity, persistence, and intrusiveness of 14 common psychiatric symptoms in nonclinical populations. The internal reliability of the scale across waves 2-6 was good (α = .79). A CIS-R total score was calculated at each wave during adolescence and dichotomized with a threshold of 11/12 corresponding to the point where a general practitioner might be concerned about an individual’s mental health.²⁴, ²⁵, ²⁶ Participants were categorized as having “low depressive symptoms” (never reported depressive symptoms (12+), “moderate depressive symptoms” (depressive symptoms [12+] reported once), and “high depressive symptoms” (depressive symptoms [12+] reported at least twice).

Potential Confounding Factors 

Adolescent antisocial behavior was assessed at every wave from 14-18 years using “The Self-Report Early Delinquency Instrument,”²⁷ from which 10 items were selected to cover antisocial behaviors that occurred during the 6 months prior to data collection.²⁶ Behaviors related to property damage (vandalism, car damage, graffiti), interpersonal conflict (fighting, carrying weapons, running away from home, expulsion from school), and theft (stealing from parents and others, stealing cars) were assessed. Response categories were “No, never,” “Yes, only once,” and “Yes, more than once.” Antisocial behaviors were stratified into those with antisocial behaviors (1 or more antisocial behavior on more than 1 occasion) and those without antisocial behaviors.

Adolescent substance use was assessed at 14-18 years. Tobacco use was defined as any smoking (yes/no) in the past month. Cannabis use was defined as any use (yes/no) over the past 12 months. Binge drinking was assessed using self-reported alcohol use in the previous 6 months. For participants who reported being at least occasional drinkers, a 7-day retrospective diary of alcohol consumption during the week prior to data collection was completed. The diary used a beverage-specific approach and detailed types of drinks (eg, low-alcohol beer, normal beer, wine, spirits, and mixed drinks), as well as the quantities consumed on each drinking day.²⁸ Binge drinking was defined by the average level of alcohol consumption using a cutoff of 5 standard drinks (10 g of alcohol) per drinking day.

Externalizing behavior. Sample size limitations required thorough assessment of co-linearity between potential confounding factors to ensure well-powered models. Exploratory factor analysis (principal factor analysis with varimax rotation) was used to examine whether indicators of antisocial behavior and drug use were sufficiently co-linear to combine into a single indicator variable of externalizing behavior. Results showed only 1 factor (Eigen value > 1.0; α = .71, see Table 1), supporting the use of a single indicator variable for externalizing behavior (See Table 1).

Table 1. Factor Loadings Based on a Principal Factor Analysis with Varimax Rotation for 4 Items of Adolescent Externalizing Behaviors between 14 and 18 Years

Eigenvalue = 1,468

Response categories for tobacco and cannabis use, binge drinking, and antisocial behavior variables were standardized to ensure equivalence of meaning, whereby each variable was recoded to represent “Never,” “Moderate” (any behavior, 1-3 waves), and “High” (any behavior, 4+ waves). These 4 standardized variables were then combined into a single 3-level variable representing “No externalizing behavior” (no high antisocial and/or drug use at any wave), “Moderate externalizing behavior” (any high antisocial behavior and/or drug use, 1-3 waves), and “High externalizing behavior” (any high antisocial behavior and/or drug use, 4+ waves).

Socioeconomic confounder variables were indicated by metropolitan or regional location of school, parental separation/divorce, and parental education (incomplete secondary school or completed secondary/technical qualification).

Data Analysis 

Multiple imputation (MI) was used to deal with the potential bias arising from missing data. Multiple imputation was conducted using NORM software (Version 2.03, USA, 2000).²⁹ Twenty complete datasets were created by imputation under a multivariable normal model, incorporating all variables of interest measured at all waves of data collection, along with fixed covariates (age, metropolitan or regional school location, and parental education), as well as auxiliary variables (attachment style, impulsivity, neuroticism, parental bonding, parental smoking, language other than English, level of education, and employment status), using adaptive rounding for binary measures.²⁹, ³⁰ Estimates of log odds ratios were obtained within the multiple imputation framework by averaging across the imputed datasets with Wald-type confidence intervals obtained using Rubin’s combination rules.³⁰, ³¹ Final data analysis was conducted using Stata 11, release 1.0 (Stata Corporation LP, College Station, TX, USA, 2007). Associations between adolescent depressive symptoms and young adult pregnancy outcomes are reported as odds ratios (OR) with 95% confidence intervals (CI) and a level of statistical significance of P < .05. All analyses were adjusted for socioeconomic indicators and antisocial and drug use behavior.

Back to Article Outline

Results 

Frequencies of adolescent depressive symptoms (14-18 years) and pregnancy outcomes between 21-24 years of age are displayed in Table 2.

Table 2. Frequency of Adolescent Depression and Pregnancy, Pregnancy Completion, and Termination of Pregnancy in 988 Females

Abbreviation: CI, confidence interval

Associations between adolescent depressive symptoms and becoming pregnant in young adulthood are presented in Table 3. In the unadjusted model (model 1), high adolescent depressive symptomatology (2+ waves) was associated with a 2-fold increase in the odds of reporting pregnancy between 21-24 years of age. This association was attenuated after adjustment for externalizing behavior (model 2). Of particular note, both moderate (1-3 waves) and high (4+ waves) persistence of externalizing behavior in adolescence were associated with 2- and 3-fold increases in the odds of reporting early adult pregnancy, respectively (model 2). When adjusting for socioeconomic indicators, these effects remained (model 3). There was also a 1.7-fold increase in the odds of reporting early adult pregnancy for adolescents with separated or divorced parents.

Table 3. Associations between Adolescent Depressive Symptoms and Becoming Pregnant in Young Adulthood

Adolescent predictors (14 to 18 years)	Pregnancy in young adulthood (21 to 24 years)
	Model 1			Model 2			Model 3
	OR	95% CI	P	OR	95% CI	P	OR	95% CI	P
Depressive symptoms
Low	1.0			1.0			1.0
Moderate (1 wave)	1.4	(0.71-2.7)	.334	1.2	(0.62-2.4)	.556	1.3	(0.64-2.5)	.493
High (2-6 waves)	2.0	(1.2-3.3)	.005	1.6	(1.0-2.6)	.064	1.5	(0.92-2.6)	0.100
Antisocial/drug use behavior
Never				1.0			1.0
Moderate (1-3 waves)				2.2	(1.2-4.1)	.012	2.1	(1.1-3.9)	.021
High (3-6 waves)				3.3	(1.8-6.1)	<0.000	2.9	(1.5-5.5)	.001
School location
Regional							1.0
Metro							0.72	(0.45-1.2)	.186
Parent education
No HS							1.0
HS/Tertiary							0.80	(0.61-1.1)	0.110
Parent separation/divorce
No							1.0
Yes							1.7	(1.0-2.7)	.031

Abbreviations: CI, confidence interval; OR, odds ratio

Associations between adolescent depressive symptoms and pregnancy completion and termination are presented in Table 4, Table 5. In the unadjusted models, there was a trend toward an association between adolescent depressive symptoms and both completion or termination of pregnancy (p=0.089 and 0.06, respectively). There was no association between any adolescent predictors and completion of pregnancy in the final models (Table 4: models 2 and 3). However, consistent with findings of becoming pregnant, adolescent antisocial and drug use behavior (4+ waves) was associated with a 4-fold increase in the odds of reporting a termination (Table 5: model 2) and attenuated any trend toward association with depressive symptoms. This effect persisted after adjusting for socioeconomic indicators (Table 5: model 3). Experiencing parental separation/divorce was associated with a 2-fold increase in the odds of reporting a termination.

Table 4. Associations between Adolescent Depressive Symptoms and Pregnancy Completion in Young Adulthood

Adolescent predictors (14 to 18 years)	Completion of pregnancy in young adulthood (21 to 24 years)
	Model 1			Model 2			Model 3
	OR	95% CI	P	OR	95% CI	P	OR	95% CI	P
Depressive symptoms
Low	1.0			1.0			1.0
Moderate (1 wave)	1.0	(0.31-3.5)	.945	0.95	(0.28-3.2)	.934	1.0	(0.29-3.5)	.995
High (2-6 waves)	1.8	(0.92-3.5)	.089	1.5	(0.76-3.0)	.230	1.5	(0.75-3.1)	.247
Antisocial/drug use behavior
Never				1.0			1.0
Moderate (1-3 waves)				2.2	(0.93-5.1)	.071	2.0	(0.86-4.8)	.107
High (3-6 waves)				2.3	(0.93-5.6)	.070	2.0	(0.81-5.0)	.134
School location
Regional							1.0
Metro							0.61	(0.33-1.1)	.108
Parent education
No HS							1.0
HS/Tertiary							0.75	(0.52-1.1)	.124
Parent separation/divorce
No							1.0
Yes							1.6	(0.84-2.9)	.155

Abbreviations: CI, confidence interval; OR, odds ratio

Table 5. Associations between Adolescent Depressive Symptoms and Pregnancy Termination in Young Adulthood

Adolescent predictors (14 to 18 years)	Termination of pregnancy in young adulthood (21 to 24 years)
	Model 1			Model 2			Model 3
	OR	95% CI	P	OR	95% CI	P	OR	95% CI	P
Depressive symptoms
Low	1.0			1.0			1.0
Moderate (1 wave)	1.3	(0.59-3.0)	.499	1.1	(0.49-2.6)	.795	1.1	(0.47-2.5)	.848
High (2-6 waves)	1.8	(1.0-3.4)	.060	1.3	(0.69-2.5)	.430	1.2	(0.61-2.3)	.615
Antisocial/drug use behavior
Never				1.0			1.0
Moderate (1-3 wave)				1.7	(0.69-4.4)	.245	1.7	(0.65-4.2)	.285
High (3-6 waves)				4.5	(1.92-11)	.001	4.0	(1.7-9.6)	.002
School location
Regional							1.0
Metro							1.1	(0.55-2.1)	.826
Parent education
No HS							1.0
HS/Tertiary							0.94	(0.66-1.3)	.709
Parent separation/divorce
No							1.0
Yes							1.9	(1.0-3.6)	.038

Abbreviations: CI, confidence interval; OR, odds ratio

Back to Article Outline

Discussion 

This study examined relationships between adolescent depressive symptoms (14-18 years) and pregnancy, pregnancy completion, and pregnancy termination in young adult women (21-24 years). We observed an association between adolescent depressive symptoms and pregnancy in young adulthood after adjusting for socioeconomic background; however, these effects were substantially reduced after adjustment for externalizing behaviors. We further observed that externalizing behavior (modeled as a confounder) was associated with marked increases in risk for becoming pregnant and termination of pregnancy in early adulthood.

The findings we report are based on longitudinal data and replicate previous longitudinal findings that have shown no relationship between adolescent depression and becoming pregnant and completion of pregnancy in young adulthood.¹⁷, ¹⁸ This replication is particularly notable because it occurred under more stringent definitions of adolescent depression (ie, persistent depressive symptoms), using a more precise age classification of young adulthood (21-24 years vs 16/17 to 21 years), and after adjusting for additional and important confounding factors, in particular, externalizing behaviors such as antisocial and drug use behavior. Furthermore, our findings extend previous longitudinal work by showing no relationship between adolescent depression and pregnancy termination.

Our final models consistently showed that externalizing behaviors posed more important risks for pregnancy in young adulthood than any other factor. Results of the principal factor analysis we report suggest that antisocial and drug use behaviors are highly co-morbid and may reflect an underlying latent disposition to impulsive personality style more prone to immediate gratification and risk taking. Such a disposition could be manifested in a range of behaviors; however, sexual risk taking would be particularly relevant to a peak time for sexual maturation and experimentation in adolescence and young adulthood.

Furthermore, our results show that these externalizing behaviors are associated with young adult pregnancy and termination but less convincingly with pregnancy progressed to delivery (or the choice to become a parent). This finding would again be consistent with the idea that antisocial and drug use behavior may reflect an underlying disposition to risk taking. Additionally, although pregnancy might indicate an impulsive act, a subsequent choice to go to term (and adopt the responsibilities of parenthood) might not. Lack of association with pregnancy completion could be interpreted as being consistent with such a risk-taking hypothesis.

Strengths of the current study include its prospective multiwave design with a sample from a representative pool of schools and high participation rates, which enable us to examine temporal relationships between adolescent behaviors and pregnancy outcomes in young adulthood. However, sample sizes were small for some of the pregnancy outcomes, which might explain lack of association between variables in some cases. Multiple imputation of missing data will have reduced but not eliminated this limitation. Furthermore, despite being able to adjust for a broad range of factors, we lacked data on some potential confounders or effect modifiers including more specific clinical features of depression (eg, history of treatment). However, emerging capabilities for data linkage with external health registries (eg, prescription, hospital, and general practice) may allow such factors to be modeled in future research.

In conclusion, this study expands on findings about predictors of pregnancy and pregnancy outcomes in adolescence to young adulthood. Adolescent depressive symptoms did not emerge as a significant risk factor for young adult pregnancy; however, adolescent antisocial and drug use behaviors are associated with significant risk for pregnancy and termination of pregnancy in young adulthood. Selective preventive intervention in young women with a history of antisocial and drug use behaviors would have the potential to improve their sexual and reproductive health outcomes.

Back to Article Outline

Acknowledgments 

This research is based on data from the Victorian Adolescent Health Cohort Study (Australia), a study funded primarily by the Australian National Health and Medical Research Council. The authors thank the participants for their contributions to the study and acknowledge funding from the Australian Health Management for cohort data analysis. W.N. and E.K. are supported by the Norwegian Research Council and the Norwegian ExtraFoundation for Health and Rehabilitation for their doctoral and postdoctoral positions, respectively. G.C.P. is supported by a National Health and Medical Research Council Senior Principal Research Fellowship.

Back to Article Outline

References 

1. Australian Bureau of Statistics: Births. Canberra, Australia: Australian Bureau of Statistics; 2008;
   • View In Article
2. Office for National Statistics: Birth Statistics 2008: Newport. United Kingdom: Crown; 2008;pp 1–79
   • View In Article
3. Statistics Canada: Births 2007: Ottawa. Canada: Ministry of Industry; 2007;pp 1–53
   • View In Article
4. Coley RL, Chase-Lansdale PL. Adolescent pregnancy and parenthood: recent evidence and future directions. Am Psychol. 1998;53:152–166
   • View In Article
   • CrossRef
5. Hillis SD, Anda RF, Dube SR, et al. The association between adverse childhood experiences and adolescent pregnancy, long-term psychosocial consequences, and fetal death. Pediatrics. 2004;113:320–327
   • View In Article
   • CrossRef
6. Imamura M, Tucker J, Hannaford P, et al. Factors associated with teenage pregnancy in the European Union countries: a systematic review. Eur J Public Health. 2007;17:630–636
   • View In Article
   • CrossRef
7. Cavazos-Rehg PA, Spitznagel EL, Krauss MJ, et al. Understanding adolescent parenthood from a multisystemic perspective. J Adolesc Health. 2010;46:525–531
   • View In Article
   • CrossRef
8. Keenan K, Loeber R, Green S. Conduct disorder in girls: a review of the literature. Clin Child Fam Psychol Rev. 1999;2:3–19
   • View In Article
   • MEDLINE
   • CrossRef
9. Kessler RC, Berglund PA, Foster CL, et al. Social consequences of psychiatric disorders. 2. Teenage parenthood. Am J Psychiatry. 1997;154:1405–1411
   • View In Article
10. Miller-Johnson S, Winn DMC, Coie JD, et al. Risk factors for adolescent pregnancy reports among African American males. J Res Adolesc. 2004;14:471–495
    • View In Article
11. DiClemente RJ, Wingood GM, Crosby RA, et al. A prospective study of psychological distress and sexual risk behavior among black adolescent females. Pediatrics. 2001;108:E85
    • View In Article
    • CrossRef
12. Miller-Johnson S, Winn DM, Coie J, et al. Motherhood during the teen years: a developmental perspective on risk factors for childbearing. Dev Psychopathol. 1999;11:85–100
    • View In Article
    • MEDLINE
    • CrossRef
13. Kowaleski-Jones L, Mott FL. Sex, contraception and childbearing among high-risk youth: do different factors influence males and females?. Fam Plann Perspect. 1998;30:163–169
    • View In Article
    • MEDLINE
    • CrossRef
14. Lee SH, O’Riordan MA, Lazebnik R. Relationships among depressive symptoms, sexually transmitted infections, and pregnancy in African-American adolescent girls. J Pediatr Adolesc Gynecol. 2009;22:19–23
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (117 KB)
    • CrossRef
15. Rubin AG, Gold MA, Primack BA. Associations between depressive symptoms and sexual risk behavior in a diverse sample of female adolescents. J Pediatr Adolesc Gynecol. 2009;22:306–312
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (132 KB)
    • CrossRef
16. Woodward LJ, Fergusson DM, Horwood LJ. Gender differences in the transition to early parenthood. Dev Psychopathol. 2006;18:275–294
    • View In Article
    • MEDLINE
17. Bardone AM, Moffitt TE, Caspi A, et al. Adult physical health outcomes of adolescent girls with conduct disorder, depression, and anxiety. J Am Acad Child Adolesc Psychiatry. 1998;37:594–601
    • View In Article
    • Abstract
    • Full-Text PDF (6504 KB)
    • CrossRef
18. Fergusson DM, Woodward LJ. Mental health, educational, and social role outcomes of adolescents with depression. Arch Gen Psychiatry. 2002;59:225–231
    • View In Article
    • CrossRef
19. Larsson C, Sydsjo G, Josefsson A. Health, sociodemographic data, and pregnancy outcome in women with antepartum depressive symptoms. Obstet Gynecol. 2004;104:459–466
    • View In Article
    • MEDLINE
    • CrossRef
20. Nelson DB, McMahon K, Joffe M, et al. The effect of depressive symptoms and optimism on the risk of spontaneous abortion among innercity women. J Womens Health Gend Based Med. 2003;12:569–576
    • View In Article
21. Jonsson U, Bohman H, Hjern A, et al. Intimate relationships and childbearing after adolescent depression: a population-based 15 year follow-up study. Soc Psychiatry Psychiatr Epidemiol. 2011;46:711–721
    • View In Article
    • CrossRef
22. Paperny DM, Aono JY, Lehman RM, et al. Computer-assisted detection and intervention in adolescent high-risk health behaviors. J Pediatr. 1990;116:456–462
    • View In Article
    • Abstract
    • Full-Text PDF (588 KB)
    • CrossRef
23. Goldberg DP, Cooper B, Eastwood MR, et al. A standardized psychiatric interview for use in community surveys. Br J Prev Soc Med. 1970;24:18–23
    • View In Article
    • MEDLINE
24. Lewis G, Pelosi AJ, Glover E, et al. The development of a computerized assessment for minor psychiatric disorder. Psychol Med. 1988;18:737–745
    • View In Article
    • MEDLINE
    • CrossRef
25. Lewis G, Pelosi AJ, Araya R, et al. Measuring psychiatric disorder in the community: a standardized assessment for use by lay interviewers. Psychol Med. 1992;22:465–486
    • View In Article
    • MEDLINE
    • CrossRef
26. Patton GC, Harris R, Carlin JB, et al. Adolescent suicidal behaviours: a population-based study of risk. Psychol Med. 1997;27:715–724
    • View In Article
    • MEDLINE
    • CrossRef
27. Moffitt TE, Silva PA. Self-Reported Delinquency: Results From an Instrument for New Zealand. Aust NZ J Crim. 1988;21:227–240
    • View In Article
28. Hilton ME. A comparison of a prospective diary and 2 summary recall techniques for recording alcohol consumption. Br J Addict. 1989;84:1085–1092
    • View In Article
    • MEDLINE
29. Schafer JL. Analysis of incomplete multivariate data. London, England: CRC Press; 1997;
    • View In Article
30. Bernaards CA, Belin TR, Schafer JL. Robustness of a multivariate normal approximation for imputation of incomplete binary data. Stat Med. 2007;26:1368–1382
    • View In Article
    • MEDLINE
    • CrossRef
31. Carlin JB, Galati JC, Royston P. A new framework for managing and analyzing multiply imputed data in Stata. Stata J. 2008;8:49–67
    • View In Article