Teenage Pregnancy and Mental Health: A Systematic Review Notes

Introduction

  • Teenage pregnancy has been a focus of public health programs since the 1970s.
  • Birth rates among adolescents have declined in industrialized countries over the past 20 years but remain high in areas like the UK, New Zealand, and the USA.
  • Teenage pregnancy is linked to marginalization, with teenage mothers more likely to come from deprived socioeconomic backgrounds.
  • Teenage pregnancy can worsen socioeconomic circumstances, leading to low educational attainment, unemployment, and poor social support.
  • Teenage mothers face increased risks of short-term health issues, including pregnancy complications and postpartum depression.
  • Poor mental health beyond the postpartum period is plausible due to poverty, low education, unemployment, poor social support, and chronic stressors associated with young parenting.
  • Mental health problems like depression, anxiety, PTSD, and substance use disorders are significant public health concerns.
  • Children of mothers with mental illness are at higher risk for developmental and behavioral problems, while affected women are at risk for self-harm, relationship issues, and chronic diseases.
  • A systematic evaluation of the literature is needed to understand the association between teenage pregnancy and long-term mental health.
  • Objectives of the systematic review:
    • Determine if women with teenage pregnancies are more likely to experience poor mental health beyond the postpartum period. This involves comparing them to women who delay childbearing.
    • Critically assess the quality of evidence from quantitative and qualitative studies.
    • Identify directions for future research.

Methods

  • The study followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines.
  • Databases searched included MEDLINE, Embase, PsycINFO, CINAHL, Scopus, and Web of Science from inception to June 14, 2017.
  • A librarian was consulted to finalize the search strategy, which was conducted by one author (CX).
  • Bibliographies of original articles and reviews were hand-searched to find articles missed in the database search.
  • Eligibility criteria:
    • Quantitative and qualitative studies examining the association between teenage pregnancy and long-term mental health were included.
    • Teenage mothers were defined as women giving birth to their first child before age 20, with slight variations allowed (e.g., <18 years).
    • Studies focusing on young adults or very young teenagers were excluded.
    • Mental health was measured 1 year or later after childbirth.
    • Outcomes included self-reported or physician-diagnosed depression, anxiety, PTSD, or substance use.
    • Exclusion criteria: studies without a comparison group, studies of children's or fathers' mental health, studies of social or physical health outcomes, conference proceedings, reviews, commentaries, and studies not in English or French.
    • No geographic restrictions were applied.
  • Data collection process:
    • Data were independently extracted by two authors (CX and HKB), and discrepancies were resolved through discussion with a third author (AB).
    • An adapted version of the Cochrane Data Extraction Form was used to collect data on setting, study period, design, data sources, response/follow-up rates, sample size, inclusion/exclusion criteria, definitions, mental health outcomes, confounders, and study findings.
  • Risk of bias assessment:
    • Quantitative studies were evaluated using the Effective Public Health Practice Project critical appraisal tool.
    • This tool rates articles as strong, moderate, or weak based on selection bias, confounding, detection bias, and attrition bias.
    • For confounding, variables associated with teenage pregnancy and maternal mental health were identified: demographics, childhood socioeconomic status (SES), childhood adversity, childhood achievement, and childhood social support.
    • Adult social factors were considered mediators.
    • Qualitative studies were to be evaluated using the Critical Appraisal Skills Programme, but no qualitative articles were included.
    • Articles were appraised by two authors (CX and HKB), with discrepancies resolved by the third author (AB).
  • Synthesis of results:
    • A quantitative synthesis using meta-analysis was planned but not conducted due to methodological diversity.
    • A qualitative synthesis was performed, presenting results along with a narrative description of the studies' risk of bias.

Results

  • Study selection:
    • The initial search identified 9386 articles; after removing duplicates, 7027 articles remained.
    • 79 articles potentially met inclusion criteria based on title and abstract assessment.
    • Full-text review led to the removal of 70 studies based on exclusion criteria; two additional studies were found through hand-searching bibliographies.
    • Nine quantitative studies comprising analyses of 11 cohorts were included.
  • Study characteristics:
    • Studies were conducted in the USA (n=5), UK (n=2), Australia (n=1), and New Zealand (n=2).
    • Cohorts were born in the 1930s–1960s (n=4) and 1970s–80s (n=5).
    • Study sizes ranged from 98 to 4430 participants.
    • Several studies had samples of largely European descent (n=3); five studies did not describe ethnicity.
    • Most studies excluded individuals with missing data (n=6); three restricted samples to women still caring for their child.
    • Definitions of teenage pregnancy varied, including birth of a first child before 18 (n=2) and 20 years (n=7).
    • Comparison groups included women who delayed childbearing to early (n=5) or later adulthood (n=7).
    • Mental health was measured in women in their 20s (n=4) to 50s (n=2), with most studies measuring outcomes at <40 years (n=6).
    • Measures included self-reported screening tools of general mental health (n=2) and depressive (n=6) or anxiety symptoms (n=1); two studies used diagnostic interviews.
    • Studies varied regarding variables controlled for, including demographics, childhood SES, childhood adversity, childhood achievement, and childhood social support.
    • Several studies also controlled for adult social factors.
  • Risk of bias within studies:
    • Studies were rated as strong (n=4), moderate (n=3), or weak (n=2) quality.
    • Selection bias: Most samples were representative due to sampling strategies, but one study was limited to a low-income sample. Several studies had response rates of <80% or did not report response rates.
    • Confounding: Although all studies controlled for confounders, most controlled for <80% of important categories of confounders. Current social factors controlled for in most studies could be on the causal pathway, potentially attenuating effects.
    • Detection bias: All studies provided validity and reliability data for outcome measures.
    • Attrition bias: Several studies had follow-up rates of <80%, suggesting potential for attrition bias.
  • Synthesis of results:
    • Three studies reported statistically significant associations between teenage pregnancy and poor mental health beyond the postpartum period after adjustment.
      • Aitken et al. showed that teenage mothers born in 1936–1965 had lower mental health scores at >40 years than older mothers, after controlling for demographics and childhood SES.
      • Henretta et al. showed that young mothers in the US (1931–1941) and UK (1946) cohorts had lower levels of depressive and anxiety symptoms than teenage mothers at 51–61 years, after controlling for demographics, childhood SES, teenage and current social support, and current SES.
      • Maughan and Lindelow showed that teenage mothers in their 1958 cohort had higher rates of psychological distress in their 30s than adult mothers after controlling for childhood SES, adversity and achievement, and current social support and SES.
    • Eight studies did not find statistically significant associations after adjustment.
      • Three reported associations that were statistically significant in unadjusted analyses only.
        • Boden et al. found that teenage mothers born in 1977 had a higher mean number of mental disorders in their 20s than young mothers, but this association disappeared after controlling for childhood SES, adversity, and achievement.
        • Falci et al. found that adult mothers born in 1974 had lower depressive symptoms at 29 years than teenage mothers, but this association disappeared after controlling for demographics; childhood SES, adversity and achievements; childhood and current social support; and current SES.
        • Henretta et al. found that older mothers in their US cohort had lower depressive symptoms than teenage mothers, but findings were non-significant after controlling for demographics, childhood SES, teenage and current social support, and current SES.
      • Five studies reported associations that were non-statistically significant in all analyses.
        • Jaffee reported no difference in depressive or anxiety symptoms at 26 years between teenage and adult mothers born in 1972–1973.
        • Maughan and Lindelow showed no difference in affective symptoms between teenage and adult mothers in their 1946 cohort.
        • Taylor reported no difference in depressive symptoms between teenage and older mothers, born in 1939, in their 50s.
        • Vicary and Corneal found no difference in depressive symptoms between teenage and older mothers, born in 1972, in their 20s, but this study was small (n=98).
        • Whitworth found no difference in depressive symptoms between teenage and adult mothers, born in 1976–1980, at 28–32 years.

Discussion

  • Summary of evidence:
    • Of the 11 cohorts in nine studies, three found a statistically significant association between teenage pregnancy and poor mental health beyond the postpartum period after adjustment. Three found a statistically significant association before but not after adjustment, and five found no association.
    • Heterogeneity existed in study eras, definitions, mental health measures, ages, and study quality.
    • Studies reporting statistically significant associations examined earlier cohorts, measured outcomes at older ages, and had stronger designs.
  • Potential explanations for findings:
    • The life course perspective suggests that teenage pregnancy deviates from normative patterns, leading to poorer mental health.
    • Stress process theory posits that teenage pregnancy and its social consequences increase the risk of mental health problems.
    • Birth cohort and current age impact findings.
    • Women born in earlier eras faced greater stigmatization, potentially explaining stronger associations between teenage pregnancy and poor long-term mental health.
    • The sexual revolution of the 1960s increased pressure to participate in the workforce and delay childbearing.
    • Accumulation of chronic strain from low education, underemployment, and poor social support across a lifetime may result in poor mental health.
    • Mediation analyses of adult social factors are needed to test this hypothesis.
    • Long-term impact may be due to 'selection of disadvantaged women into teen childbearing.'
    • Poor mental health in midlife could be a continuation of chronic mental health issues resulting from early adversity.
    • Controlling for social factors may lead to overadjustment.
  • Limitations:
    • Several studies were limited by high risk of selection bias and confounding.
    • Data were collected by self-report, potentially resulting in response bias.
    • Variability in measures could contribute to heterogeneity.
    • All studies were conducted in industrialized nations.
    • Studies did not examine circumstances surrounding the pregnancy.
    • The review only considered studies in English or French and excluded studies focused on very young teenagers or young adults.
    • Heterogeneity prevented meta-analysis.
  • Implications:
    • There is a need to disentangle age, period, and cohort effects.
    • Studies should investigate mediating influences of adult SES and social support.
    • They should also evaluate the role of childhood mental illness.
    • Relationships should be considered in the context of culture and ethnicity.
    • Social factors play an important role in the aetiology of poor mental health beyond the postpartum period.

Conclusion

  • The association between teenage pregnancy and mental health beyond the postpartum period remains unclear.
  • Future studies should employ age–period–cohort frameworks and evaluate modifiable social factors that could explain the association.