Loss to follow up did not bias associations between early life factors and adult depression

TY - JOUR

T1 - Loss to follow up did not bias associations between early life factors and adult depression

AU - Osler, Merete

AU - Kriegbaum, Margit

AU - Christensen, Ulla

AU - Lund, Rikke

AU - Andersen, Anne-Marie Nybo

PY - 2008

Y1 - 2008

N2 - OBJECTIVE: This study examines the consequences of nonresponse in a follow-up survey for the associations of early life factors with adult depression. STUDY DESIGN AND SETTING: A cohort of 11,532 Danish men born in 1953 had nearly complete follow up for outcomes retrieved from the Danish Psychiatric Register and the National Prescription Register, but only 66% of 9,507 eligible cohort members participated in a follow-up survey in 2004. We examined whether characteristics measured at birth and at ages 12 and 18 years, were associated with survey response. Associations between early life characteristics and four measures of depression were described by odd ratios (OR), estimated by logistic regression. For the register-based measures the effect of nonresponse was described by a relative OR(OR(responders)/OR(entire cohort)=ROR). RESULTS: Nonresponse at 50 years of age was related to having a single mother at birth, low educational attainment at age 18, and low cognitive function at ages 12 and 18. Hospitalizations for depression and having claimed a prescription for an antidepressive drug were also most frequent among men who did not respond in the follow up. However, the effect of this nonresponse on the estimated ORs was small, and all ROR were nonsignificant. CONCLUSION: Although early life characteristics were related to response in a follow-up survey, the ORs for the exposure-risk associations were not biased by nonresponse.

AB - OBJECTIVE: This study examines the consequences of nonresponse in a follow-up survey for the associations of early life factors with adult depression. STUDY DESIGN AND SETTING: A cohort of 11,532 Danish men born in 1953 had nearly complete follow up for outcomes retrieved from the Danish Psychiatric Register and the National Prescription Register, but only 66% of 9,507 eligible cohort members participated in a follow-up survey in 2004. We examined whether characteristics measured at birth and at ages 12 and 18 years, were associated with survey response. Associations between early life characteristics and four measures of depression were described by odd ratios (OR), estimated by logistic regression. For the register-based measures the effect of nonresponse was described by a relative OR(OR(responders)/OR(entire cohort)=ROR). RESULTS: Nonresponse at 50 years of age was related to having a single mother at birth, low educational attainment at age 18, and low cognitive function at ages 12 and 18. Hospitalizations for depression and having claimed a prescription for an antidepressive drug were also most frequent among men who did not respond in the follow up. However, the effect of this nonresponse on the estimated ORs was small, and all ROR were nonsignificant. CONCLUSION: Although early life characteristics were related to response in a follow-up survey, the ORs for the exposure-risk associations were not biased by nonresponse.

U2 - 10.1016/j.jclinepi.2007.11.005

DO - 10.1016/j.jclinepi.2007.11.005

M3 - Journal article

C2 - 18495426

SN - 0895-4356

VL - 61

SP - 958

EP - 963

JO - Journal of Clinical Epidemiology

JF - Journal of Clinical Epidemiology

IS - 9

ER -