Maternal and Child Health Following 2 Home Visiting Interventions vs Control: Five-Year Follow-Up of a Randomized Clinical Trial

Abstract

Importance: Home-based interventions targeting socially disadvantaged families may help to improve maternal and child health. Only a few studies have investigated how different staffing models affect early home visiting program outcomes.

Objective: To assess the effects of 2 staffing models of an early childhood intervention on mother and child outcomes.

Design, setting, and participants: The baseline assessment of this randomized trial was conducted between November 2006 and December 2009 in 15 municipalities in Germany. The follow-up assessment at offspring age 7 years was carried out by interviewers masked to treatment conditions from April 2015 to December 2017. Data analysis was performed from March to August 2023. Pregnant women with no previous live birth, low-income, and at least 1 additional psychosocial risk factor were eligible. A total of 1157 women were referred to the study by gynecologists, psychosocial counseling services, or employment agencies; 755 were randomized to treatment conditions (2 intervention groups and 2 control groups); and 525 completed the follow-up.

Interventions: Based on the Nurse-Family Partnership program, women assigned to the intervention groups received visits by either a midwife (midwife-only model) or by a team consisting of a social worker and a midwife (tandem model) until child age 2 years. Women assigned to control groups had access to the standard health and social services.

Main outcomes and measures: Average treatment effects (ATEs) on the following primary outcomes were assessed using adjusted regression models with inverse probability weighting: developmental disorders, child behavioral problems, adverse, neglectful and abusive parenting, maternal mental health, and life satisfaction.

Results: The mean (SD) age at follow-up was 29.6 (4.36) years for mothers and 7.55 (0.75) years for children; 272 (52.2%) of the children were female. Mothers in the tandem model reported fewer internalizing child behavioral problems compared to their control group (ATE, 2.98; 95% CI, -5.49 to -0.47; absolute reduction, 13.3 percentage points). Beneficial intervention effects were found in the midwife-only group on abusive parenting (ATE, -4.00; 95% CI, -6.82 to -1.18), parenting stress (ATE, -0.13; 95% CI, -0.20 to -0.06), and maternal mental health burden (ATE, -3.63; 95% CI, -6.03 to -1.22; absolute reduction, 6.6 percentage points in depressive symptoms), but not in the tandem group.

Conclusions and relevance: Both staffing models produced positive intervention effects, with more effects seen in the midwife-only model. These insights can guide future early childhood intervention designs and may help improve health care for socially disadvantaged families.

Trial registration: German Clinical Trials Register Identifier: DRKS00007554.