Table 3 Adolescent, youth and adult health outcome studies—quality assessment
StudyQuality grades assigned by reviewers (NS/TJ)Confounders/variables analysedLimitations/strengthsFindings
Adolescent and youth health outcomes
Crowder and Teachman, 200428Sound/soundSE, age, race, single parenthood, year of observationL: Attrition, described as low, but no details given. S: Longitudinal design, analysis of key confoundersNo of moves and neighbourhood measures of disadvantage associated with teenage pregnancy
DeWit, 199829Adequate/soundSE, sex, location, no of siblings, parental employment education, alcohol abuse, depression, and childhood abuse; parental marital conflict and conduct disorderL: long period of respondent recall in cross-sectional design. Recency and reasons for moves not considered. S: attrition analysed. Comprehensive analysis of confounders⩾4 Moves associated with early initiation of drug abuse and drug-related problems (for marijuana), but not alcohol abuse: accelerated time to first use marijuana by 11.8%, and time to marijuana-related problems by 49%
Adam and Chase-Lansdale, 200230Adequate/adequateSE, age and marital status of caregiver, adolescent perceptions of current relationships and neighbourhoodL: Attrition to study, highly selective: limited generalisability. S: Considered moderating factors at family and neighbourhood levelsRM associated with adjustment problems, adolescent girls. Moderated by maternal relationship quality, availability of adult kin support and perceived neighbourhood quality
Gilman et al, 200331Adequate/adequateSE, family history of mental illness, maternal age, child’s age, ethnicityL: Lack of clarity on attrition from cohort; not representative of population (acknowledged by authors). S: Longitudinally collected data⩾3 Moves before 7 years of age: depression: OR 1.36 (1.05 to 1.78); onset depression under 14 years. OR 2.62 (1.46 to 3.91); remission: OR 1.65 (1.10 to 2.47)
Stack, 199432Adequate/suboptimalSE, age, region, father’s education, measures of social controlL: Self-report bias; attrition to study not reported. No information on recency of last move; data from 1976. S: Social control factors includedRM increased risk of premarital sexual behaviour in adolescent girls. Each move associated with partial odds ratio increase of 1.04. A weak effect overall
Mundy et al, 198933Suboptimal/suboptimalNoneL: Huge potential for bias in this highly selective sample. Problems inherent in archival chart review31% Adolescent inpatient psychiatric patients sample had moved ⩾5 times
Adult health outcomes
Bures, 200334Suboptimal/suboptimalSE, sex, family structureL: Highly selective population limiting generalisability; confounders not fully accounted for. S: Longitudinal dataMortality rate over 26 years of follow-up: childhood residential moves not independently associated with increased mortality
Juon et al, 200335Adequate/suboptimalSE, family stability, ethnicity, marital status in adulthood, income, education, community supports, no of childrenL: Respondent recall, often over long periods of time. Self-assessment of most key variables. “Poverty rate” 3.3%; limited generalisability“Residential stability” associated with self-assessed global health in adulthood: residentially stable vs residentially unstable: OR 1.42 (1.18 to 1.72). Mental health—not statistically significant
  • L:, limitations; OR, odds ratio followed by 95% confidence interval in brackets; RM, residential mobility; S:, strengths; SE, socioeconomic factors or measures.