Video Abstract
Poor mental health is recognized as one of the greatest global burdens of disease. Maternal mental health is crucial for the optimal health of mothers and their children. We examined the effects of an Australian Nurse Home Visiting (NHV) program (right@home), offered to pregnant women experiencing adversity, on maternal mental health and well-being at child age 3 years.
A randomized controlled trial of NHV delivered via universal child and family health services (2013–2016). Pregnant women experiencing adversity (≥2 of 10 risk factors) were recruited from 10 antenatal clinics across 2 states. Intervention comprised 25 home visits until child age 2 years. The outcomes assessed 1-year postintervention completion were maternal self-report of mental health symptoms (Depression Anxiety Stress Scales) and positive aspects of mental health (personal well-being and self-efficacy).
Of the 722 women enrolled in the trial, 255 of 363 (70%) intervention and 240 of 359 (67%) control group women provided data at 3 years. Compared with controls, the intervention group reported better mental health (reverse Depression Anxiety Stress Scales scores): effect sizes of 0.25 (depression; 95% confidence interval [CI]: 0.08 to 0.32), 0.20 (anxiety; 95% CI: 0.05 to 0.30), 0.17 (stress; 95% CI: 0.09 to 0.37), and 0.23 (total score; 95% CI: 0.12 to 0.38); 0.16 (95% CI: 0.04 to 0.29) for personal well-being; and an odds ratio of 1.60 (95% CI: 1.19 to 2.16) for self-efficacy.
An NHV designed to support mothers experiencing adversity can lead to later maternal mental health benefits, even after the program ends.
Maternal mental health is crucial for optimal health of mothers and their children. Nurse home visiting (NHV) is an established model of health care delivery available in multiple countries but has mixed results in relation to maternal mental health.
Benefits of the right@home NHV program were evident for maternal mental health and well-being at 1-year postintervention completion (child age 3 years). An NHV program that is designed for women experiencing adversity can lead to latent mental health benefits.
Poor mental health is recognized as one of the greatest global burdens of disease.1,2 In the United States alone, annual spending on mental illness is estimated at $89 billion, not including the economic cost of lost earnings and productivity.3 In Australia, an estimated 8 million working days are lost annually because of mental illness, and international estimates are similar.4,5 For mothers and children, the estimated annual societal cost in the United States of maternal mental illness from birth to 5 years is $14.2 billion.6 In high income countries like the United States and Australia, poor mental health is 2- to 3-times higher for those in the lowest income quintile compared with those in the highest.4,7 Similarly, mothers experiencing social adversities, such as relationship difficulties, social isolation, unemployment, and low educational attainment, are at higher risk of poorer mental health.3,4,7,8 The economic and psychosocial stressors of the coronavirus disease 2019 pandemic (eg, unemployment, income loss, isolation, and strain on family relationships) are likely to put mothers at heightened risk of poor mental health, with greater impact for those experiencing adversity who are already disproportionately at risk.9 A national survey of Australian households found almost one-half of parents (48%) reported that the pandemic had negatively impacted their mental health, and this was more likely among those who had experienced financial impacts.10
Maternal mental health and well-being are crucial aspects of the optimal health of mothers and their children’s health and well-being.2,11–13 From conception to preschool age, poor maternal mental health and well-being can hinder the provision of optimal care at a time when children are most sensitive to their environments.12 Global prevalence estimates suggest that 9% to 16% of women experience depression and 10% to 15% experience an anxiety disorder in the perinatal period (antenatal to 1-year postpartum),14,15 with limited data on prevalence rates for women with older children.8,16,17
Nurse home visiting (NHV) is an established model of health care delivery with the potential to address inequities in maternal mental health and, subsequently, children’s development. NHV prioritizes women experiencing social adversity and overcomes barriers to health service access through outreach to women’s homes. It shows promise for improving early parent care and home learning environments, with some studies revealing greater benefits of NHV for women with poorer mental health.18,19 However, few NHV studies have reported on these benefits beyond the first year postpartum. For example, only 6 (all from the United States) of the 21 high quality20 (ie, randomized-trial–level evidence) NHV models that followed cohorts of children between the ages of 2 and 4 years examined the impact of NHV on maternal mental health and well-being.21–25 Of these, 3 reported small-to-modest benefits (effect size [ES] 0.10–0.56) for maternal depression symptoms and parenting stress,21–23 and 3 reported no evidence of these benefits.18,24,25
Given the parallel international policy interest in NHV and preventing inequitable rates of poor mental health and the mixed results to date, it is timely to examine the potential benefits of NHV for maternal mental health and well-being. The right@home randomized controlled trial (RCT) is the largest multisite trial of NHV in Australia and has demonstrated benefits to the primary outcomes of parent care, parent responsivity, and the home learning environment when children turned 2.26 The program was designed for women experiencing adversity who may benefit from additional support beginning in pregnancy, rather than as a specific mental health intervention. However, maternal mental health and well-being support (eg, nurse-led discussion of maternal mood, coping, and identification of additional support needs) were offered as components within the program. Fidelity monitoring showed these supports were the most frequently delivered program components, provided in >88% of nurse visits.27,28 As such, in designing the follow-up study, we anticipated that the significant parenting-related benefits (more confident parenting [eg, increased parent warmth, parental involvement, and parenting efficacy]) seen at 2 years may translate to later benefits in maternal mental health and well-being, as an important potential latent effect. We hypothesized that, when compared with women who were offered the universal child and family health (CFH) service (usual care), mothers who received the NHV intervention would report better maternal mental health and well-being outcomes at child age 3 years.
Methods
Design
This study is an RCT of NHV from pregnancy to child age 2 years, compared with the existing universal CFH service (usual care), conducted as a superiority trial with 2 parallel groups and a primary end point at child age 2 years.26,29 In the current article, we report on mental health and well-being from an extended follow-up of trial outcomes at child age 3 years. The published protocol describes the RCT methods to child age 2 years.29
Participants
Researchers recruited pregnant women attending antenatal clinics of 10 public maternity hospitals across the Australian states of Victoria and Tasmania from April 30, 2013, to August 29, 2014. Eligible women (1) had due dates before October 1, 2014; (2) were <37 weeks’ gestation; (3) had sufficient English to complete face-to-face interviews; (4) lived within travel boundaries specified by participating areas; and (5) had ≥2 of 10 risk factors identified at screening (young pregnancy; not living with another adult; no support in pregnancy; poor health; a long-term illness, health problem, or disability that limits daily activities; currently smokes; stress, anxiety, or difficulty coping; low education; no person in the household currently earning an income; and never having had a job before).29,30 Women were excluded if they (1) were enrolled in an existing Tasmanian NHV program for 15- to 19-year-olds, (2) did not comprehend the recruitment invitation (eg, intellectual disability or insufficient English), (3) had no mechanism for contact, or (4) experienced a critical event (eg, termination of pregnancy, stillbirth, or child death).
Procedures
Eligible women were identified in antenatal clinics and invited into the RCT. Participants provided informed consent for the RCT (initially to 2 years) and completed a home-based baseline interview assessing maternal demographic, economic, and psychosocial factors. The mean gestational age at the baseline assessment was 28.2 weeks. Participants were then randomized to control or intervention arms with a 1:1 allocation according to a computer-generated schedule stratified by site and parity (first-time parent versus those with children) using permuted blocks of sizes 2, 4, or 6. Research managerial staff, participants, and intervention teams were aware of allocation. Researchers who conducted follow-up assessments were blinded to randomization. Participants were reminded not to disclose their randomization allocation before each follow-up, and researchers reported any breach to the research managerial staff; 4 breaches of blinding were reported at the 3-year follow-up. At the 2-year home-based assessment, women were invited to reenroll in the extended follow-up, and informed consent was obtained.
The right@home NHV program was structured around the core Miller Early Childhood Sustained Home-visiting framework and training19,31 and bolstered by 5 evidence-based strategies for content (sleep, safety, nutrition, regulation, and bonding and relationship) and 2 for the delivery process (video feedback and motivational interviewing strategies).27,28 Program implementation was enabled by using program logic that articulated improved long-term (5-year) parent and child outcomes, together with adaptation processes that ensured the program had fidelity to the evidence and worked in the real-world health system.27 The program logic was centered around confident parenting and child developmental outcomes, with anticipated associated maternal mental health benefits. Mothers in the intervention group were offered 25 nurse home visits (mean: 23.2 home visits received), commencing antenatally and delivered mostly by the same nurse trained in the right@home NHV model of care.28,29 Most intervention group women (75.6%) also received ≥1 home visits by a social care practitioner (mean: 1.7 visits),28 who provided brief counseling interventions and/or case management as needed. In contrast, the usual CFH service includes 6 (Tasmania) or 9 (Victoria) free nurse consultations up to 2 years (mean: 7.6 consultations received),26 with some limited program flexibility depending on family need.
Outcomes and Measures
All maternal mental health outcomes were collected via self-report at the 3-year home-based follow-up assessment, in interviews with researchers blinded to randomization allocation. Measures were selected to include mental health symptoms and positive aspects of mental health,32–34 rather than mental ill health only. Maternal mental health symptoms were measured by using the Depression Anxiety Stress Scales (DASS),35 and positive mental health measures of personal wellbeing33 and self-efficacy34 were also assessed (see Table 1 for details). The DASS Total Symptom scores and Depression, Anxiety, and Stress Subscales were each reverse scored to aid interpretation, such that higher scores represent better mental health. Reversed DASS subscale scores were also dichotomized to reflect poorest mental health symptom severity (study defined as the lower 15% of scores according to population reference ranges36 ) versus better mental health (the upper 85% of scores) to estimate the impact of NHV on mental health morbidity. The measures are not diagnostic tools and, therefore, have no agreed minimum clinically important difference.
Description of Maternal Mental Health Outcome Measures
| Item | Description |
|---|---|
| Mental health symptoms | DASS:35 a 21-item measure, rated on a 4-point scale (“not at all” to “most of the time”) assessing the negative emotional states of depression, anxiety, and stress; 3 subscales (7 items each): depression, anxiety, and stress, examined as continuous scores of mental health symptoms. DASS scores were reversed so that higher scores indicate better mental health, ranging from 0 to 21. Reversed DASS subscale scores were also dichotomized to reflect poorest mental health symptom severity (study defined as the lower 15% of scores according to population reference ranges36 ) versus better mental health (upper 85% of scores). DASS subscales are strongly correlated with other self-report mental health measures in Australian postpartum women, such as the Edinburgh Postnatal Depression Scale (Pearson’s correlation [r] = 0.84) and the anxiety and depression subscales of Beck Depression Inventory (r = 0.82 and 0.86, respectively).45 |
| Personal well-being | Personal Wellbeing Index:33 Eight items assessing satisfaction with specific life domains, rated by using a 10-point scale (“no satisfaction at all” to “completely satisfied”). Higher scores indicate better well-being, ranging from 0 to 80. |
| Self-efficacy | Three items assessing mother’s self-efficacy or locus of control, which aimed to capture how the mother felt about her life in general, including the extent to which she felt that she gets what she wants out of life, felt in control, and can run her own life, drawn from the UK Millennium Cohort Study.34 Each item reflected the presence versus absence of self-efficacy and were used to form a single dichotomous item reflecting “any lack of self-efficacy” versus “no lack of self-efficacy.” |
| Item | Description |
|---|---|
| Mental health symptoms | DASS:35 a 21-item measure, rated on a 4-point scale (“not at all” to “most of the time”) assessing the negative emotional states of depression, anxiety, and stress; 3 subscales (7 items each): depression, anxiety, and stress, examined as continuous scores of mental health symptoms. DASS scores were reversed so that higher scores indicate better mental health, ranging from 0 to 21. Reversed DASS subscale scores were also dichotomized to reflect poorest mental health symptom severity (study defined as the lower 15% of scores according to population reference ranges36 ) versus better mental health (upper 85% of scores). DASS subscales are strongly correlated with other self-report mental health measures in Australian postpartum women, such as the Edinburgh Postnatal Depression Scale (Pearson’s correlation [r] = 0.84) and the anxiety and depression subscales of Beck Depression Inventory (r = 0.82 and 0.86, respectively).45 |
| Personal well-being | Personal Wellbeing Index:33 Eight items assessing satisfaction with specific life domains, rated by using a 10-point scale (“no satisfaction at all” to “completely satisfied”). Higher scores indicate better well-being, ranging from 0 to 80. |
| Self-efficacy | Three items assessing mother’s self-efficacy or locus of control, which aimed to capture how the mother felt about her life in general, including the extent to which she felt that she gets what she wants out of life, felt in control, and can run her own life, drawn from the UK Millennium Cohort Study.34 Each item reflected the presence versus absence of self-efficacy and were used to form a single dichotomous item reflecting “any lack of self-efficacy” versus “no lack of self-efficacy.” |
Statistical Analyses
The initial RCT sample size was calculated to detect a minimum ES of 0.3 SD in the primary parent responsivity outcome. A target sample size of 714 participants was estimated to provide 80% power, with a 5% significance level, accounting for clustering by care provider in the intervention group and site of care provision for usual care. This estimate allowed for 40% attrition to 2-year follow-up. A final sample size of 722 participants was achieved. The sample size achieved at the 3-year follow-up was 495 (68.6%), which retained 80% power to detect a minimum ES of 0.3 SD on continuous measures of maternal mental health. For all women retained at 3 years, the baseline characteristics of those in the intervention and usual care groups were compared by using χ2 tests (categorical measures) and t tests (continuous measures) to assess differences arising because of attrition.
In line with the published statistical analysis plan used at the age of 2 years,26 between-group comparisons of mental health outcomes at 3 years were made following intention to treat linear (continuous outcomes) and logistic (binary outcomes) regression models. Initially, these models included families who had participated in the 3-year follow-up and completed the mental health outcome measures (ie, complete case data). Regression models were adjusted for the stratification factors used during randomization (parity and study site) and additional baseline covariates (family’s Socio-Economic Indexes for Areas [SEIFA] score,37 maternal education, maternal age at child’s birth, parity, antenatal risk, maternal self-efficacy, and maternal mental health), as well as the child’s sex and age at the 3-year assessment. All regression analyses accounted for the effects of nurse clustering.29 The adjusted results are presented to ensure treatment effect estimates are corrected for chance imbalances in baseline covariates, appropriate confidence intervals (CIs) are estimated, and statistical power is most efficient.38 The adjusted results are reported as mean differences, standardized ESs, or odds ratios (ORs) with 95% CIs.
Multiple imputation techniques were used to evaluate the sensitivity of the findings on the basis of complete cases to sample attrition. With multiple imputation, we provided estimates of program effects, which included all mothers who were initially randomly assigned. Multiple imputation was conducted by using multivariate normal regression within each of the 2 treatment groups to allow for differing mechanisms by which missing data may have arisen across the groups. Imputation models included all outcomes collected at 3 years, stratification factors, and baseline covariates; 30 data sets were imputed.39 Data were analyzed by using Stata version 15 for Windows (Stata Corp, College Station, TX).
Ethical Approval
right@home was approved by Human Research Ethics Committees (HRECs): Royal Children’s Hospital (HREC 32296), Peninsula Health (HREC/13/PH/14), Ballarat Health Services (HREC/13/BHSSJOG/9), Southern Health (HREC 13084×); and Northern Health (HREC P03/13) in Victoria, and University of Tasmania (HREC H0013113), all Australian. The ethics-approved study protocol included processes for responding to participant or child safety concerns.
Results
Of 5586 women screened, 1427 (25.5%) were eligible (Fig 1). Of these, 722 (50.6%) enrolled in the trial. Of 722 enrolled, 558 (77.3%) reenrolled in the extended follow-up, and 495 (68.6%) provided data at 3-year follow-up. Table 2 presents the baseline characteristics for women who participated at 3 years compared with women lost to follow-up. The women participating in the 3 year follow-up, across both the intervention and usual care group, had slightly better mental health at baseline, were more likely to have completed high school, and were less likely to have reported a drug problem or history of family violence. Among those retained, comparison of baseline characteristics between the intervention and usual care group showed minimal differences.
Baseline Characteristics According to Follow-Up Status (ie, Retained or Lost in right@home Study) at Child Age 3 Years
| Baseline Characteristics (Pregnancy) | Intervention (n = 363) | Control (n = 359) | Pa | ||
|---|---|---|---|---|---|
| Retained (n = 255) | Lost (n = 108) | Retained (n = 240) | Lost (n = 119) | ||
| Mother | |||||
| Age, mean (SD), y | 27.6 (5.9) | 27.1 (6.4) | 28.3 (6.4) | 26.9 (6.2) | .22 |
| DASS (reversed: higher scores indicate better mental health) | |||||
| Total score, mean (SD) | 51.2 (9.7) | 49.8 (11.1) | 51.5 (8.8) | 50.8 (10.4) | .69 |
| Depression scale, mean (SD) | 18.0 (3.5) | 17.7 (4.0) | 18.2 (3.1) | 17.9 (3.6) | .42 |
| Anxiety scale, mean (SD) | 17.5 (3.4) | 17.1 (3.5) | 17.7 (3.1) | 17.2 (3.7) | .49 |
| Stress scale, mean (SD) | 15.7 (4.0) | 15.0 (4.8) | 15.6 (3.9) | 15.7 (4.3) | .77 |
| DASS (dichotomized: within 85th percentile score of better mental health according to norms) | |||||
| Depression scale | 82.7 | 81.5 | 86.7 | 79.0 | .23 |
| Anxiety scale | 59.2 | 50.9 | 61.2 | 52.9 | .64 |
| Stress scale | 80.8 | 77.8 | 80.0 | 83.2 | .83 |
| Education status | |||||
| Did not complete high school | 21.3 | 33.7 | 26.5 | 22.3 | .43 |
| Completed high school and/or vocational training | 67.0 | 58.7 | 63.2 | 66.0 | |
| Completed a university degree | 11.7 | 7.6 | 10.3 | 11.7 | |
| Marital status | |||||
| Single and/or not living with partner | 29.0 | 26.9 | 22.9 | 31.1 | .54 |
| Married and/or living with partner | 69.4 | 70.4 | 75.4 | 66.4 | |
| Separated or divorced | 1.6 | 2.8 | 1.7 | 3.5 | |
| Currently unemployed | 62.8 | 73.2 | 62.9 | 74.0 | .97 |
| Family income from benefit or pension | 42.4 | 47.2 | 41.3 | 42.9 | .26 |
| Ever had a drug problem | 12.2 | 18.5 | 13.0 | 24.6 | .78 |
| Experienced domestic violence in past year | 10.7 | 15.9 | 10.6 | 13.5 | .97 |
| Total adversity risk count (from screening), mean (SD) | 3.0 (1.3) | 3.5 (1.4) | 3.2 (1.2) | 3.3 (1.2) | .10 |
| Child | |||||
| First born | 38.8 | 34.3 | 34.6 | 40.3 | .33 |
| Female | 57.3 | 46.9 | 45.8 | 41.9 | .01 |
| Family | |||||
| SEIFA index of social disadvantage quintile | .55 | ||||
| 1 (most disadvantaged) | 44.2 | 45.2 | 39.8 | 40.9 | |
| 2 | 6.8 | 9.6 | 8.7 | 8.7 | |
| 3 | 39.4 | 32.7 | 39.0 | 36.5 | |
| 4 | 6.4 | 11.5 | 10.0 | 7.8 | |
| 5 (least disadvantaged) | 3.2 | 1.0 | 2.6 | 6.1 | |
| Language other than English | 6.8 | 10.3 | 7.6 | 13.0 | .71 |
| Baseline Characteristics (Pregnancy) | Intervention (n = 363) | Control (n = 359) | Pa | ||
|---|---|---|---|---|---|
| Retained (n = 255) | Lost (n = 108) | Retained (n = 240) | Lost (n = 119) | ||
| Mother | |||||
| Age, mean (SD), y | 27.6 (5.9) | 27.1 (6.4) | 28.3 (6.4) | 26.9 (6.2) | .22 |
| DASS (reversed: higher scores indicate better mental health) | |||||
| Total score, mean (SD) | 51.2 (9.7) | 49.8 (11.1) | 51.5 (8.8) | 50.8 (10.4) | .69 |
| Depression scale, mean (SD) | 18.0 (3.5) | 17.7 (4.0) | 18.2 (3.1) | 17.9 (3.6) | .42 |
| Anxiety scale, mean (SD) | 17.5 (3.4) | 17.1 (3.5) | 17.7 (3.1) | 17.2 (3.7) | .49 |
| Stress scale, mean (SD) | 15.7 (4.0) | 15.0 (4.8) | 15.6 (3.9) | 15.7 (4.3) | .77 |
| DASS (dichotomized: within 85th percentile score of better mental health according to norms) | |||||
| Depression scale | 82.7 | 81.5 | 86.7 | 79.0 | .23 |
| Anxiety scale | 59.2 | 50.9 | 61.2 | 52.9 | .64 |
| Stress scale | 80.8 | 77.8 | 80.0 | 83.2 | .83 |
| Education status | |||||
| Did not complete high school | 21.3 | 33.7 | 26.5 | 22.3 | .43 |
| Completed high school and/or vocational training | 67.0 | 58.7 | 63.2 | 66.0 | |
| Completed a university degree | 11.7 | 7.6 | 10.3 | 11.7 | |
| Marital status | |||||
| Single and/or not living with partner | 29.0 | 26.9 | 22.9 | 31.1 | .54 |
| Married and/or living with partner | 69.4 | 70.4 | 75.4 | 66.4 | |
| Separated or divorced | 1.6 | 2.8 | 1.7 | 3.5 | |
| Currently unemployed | 62.8 | 73.2 | 62.9 | 74.0 | .97 |
| Family income from benefit or pension | 42.4 | 47.2 | 41.3 | 42.9 | .26 |
| Ever had a drug problem | 12.2 | 18.5 | 13.0 | 24.6 | .78 |
| Experienced domestic violence in past year | 10.7 | 15.9 | 10.6 | 13.5 | .97 |
| Total adversity risk count (from screening), mean (SD) | 3.0 (1.3) | 3.5 (1.4) | 3.2 (1.2) | 3.3 (1.2) | .10 |
| Child | |||||
| First born | 38.8 | 34.3 | 34.6 | 40.3 | .33 |
| Female | 57.3 | 46.9 | 45.8 | 41.9 | .01 |
| Family | |||||
| SEIFA index of social disadvantage quintile | .55 | ||||
| 1 (most disadvantaged) | 44.2 | 45.2 | 39.8 | 40.9 | |
| 2 | 6.8 | 9.6 | 8.7 | 8.7 | |
| 3 | 39.4 | 32.7 | 39.0 | 36.5 | |
| 4 | 6.4 | 11.5 | 10.0 | 7.8 | |
| 5 (least disadvantaged) | 3.2 | 1.0 | 2.6 | 6.1 | |
| Language other than English | 6.8 | 10.3 | 7.6 | 13.0 | .71 |
All values are percentages, except where otherwise stated. Range of intervention n = 351–363; control n = 345–359 because of missing data.
P value for χ2 tests (categorical measures) and t tests (continuous measures) comparing those retained in the intervention and usual care groups.
Table 3 presents the descriptive statistics of the maternal mental health outcomes and the adjusted complete case analyses. Compared with usual care, benefits of the intervention were evident for mental health in the DASS total score (ES: 0.25; 95% CI: 0.12 to 0.38) and consistently across the DASS subscales. These benefits translated to higher odds of better mental health (better 85% of symptom scores relative to norms) for depression (OR: 1.68; 95% CI: 1.08 to 2.60), anxiety (OR: 1.38; 95% CI: 0.92 to 2.08), and stress (OR: 2.09; 95% CI: 1.28 to 3.42). Similarly, program benefits were evident in better personal well-being (ES: 0.16; 95% CI: 0.04 to 0.29) and not lacking self-efficacy (OR: 1.60; 95% CI: 1.19 to 2.16).
Adjusted Complete Case Regression Analyses Comparing the 2 Trial Arms on Maternal Mental Health Outcomes at Child Age 3 Years
| Outcome | Intervention | Control (Usual Care) | Comparative Statistica: Intervention Compared With Control | ||||||
|---|---|---|---|---|---|---|---|---|---|
| n | Summaryb | n | Summaryb | Statistica | 95% CI | P | ES | 95% CI | |
| DASS (reversed: higher scores indicate better mental health) | |||||||||
| Total score | 251 | 54.15 (8.38) | 236 | 51.77 (9.45) | 2.23 | 1.08 to 3.39 | .001 | 0.25 | 0.12 to 0.38 |
| Depression scale | 252 | 18.52 (3.22) | 236 | 17.88 (3.45) | 0.67 | 0.26 to 1.09 | .003 | 0.20 | 0.08 to 0.32 |
| Anxiety scale | 252 | 18.98 (2.72) | 236 | 18.36 (3.13) | 0.51 | 0.15 to 0.88 | .008 | 0.17 | 0.05 to 0.30 |
| Stress scale | 253 | 16.53 (3.67) | 236 | 15.54 (4.10) | 0.90 | 0.35 to 1.46 | .002 | 0.23 | 0.09 to 0.37 |
| DASS (dichotomized: within 85th percentile score of better mental health according to norms) | |||||||||
| Depression scalec | 252 | 222 (88.10) | 236 | 195 (82.63) | 1.68 | 1.08 to 2.60 | .020 | — | — |
| Anxiety scalec | 252 | 214 (84.92) | 236 | 184 (77.97) | 1.38 | 0.92 to 2.08 | .12 | — | — |
| Stress scalec | 253 | 229 (90.51) | 236 | 193 (81.78) | 2.09 | 1.28 to 3.42 | .003 | — | — |
| Personal well-being | 247 | 58.77 (12.93) | 228 | 56.07 (12.56) | 2.11 | 0.47 to 3.76 | .014 | 0.16 | 0.04 to 0.29 |
| Self-efficacyc,d | 249 | 192 (77.11) | 233 | 157 (67.38) | 1.60 | 1.19 to 2.16 | .002 | — | — |
| Outcome | Intervention | Control (Usual Care) | Comparative Statistica: Intervention Compared With Control | ||||||
|---|---|---|---|---|---|---|---|---|---|
| n | Summaryb | n | Summaryb | Statistica | 95% CI | P | ES | 95% CI | |
| DASS (reversed: higher scores indicate better mental health) | |||||||||
| Total score | 251 | 54.15 (8.38) | 236 | 51.77 (9.45) | 2.23 | 1.08 to 3.39 | .001 | 0.25 | 0.12 to 0.38 |
| Depression scale | 252 | 18.52 (3.22) | 236 | 17.88 (3.45) | 0.67 | 0.26 to 1.09 | .003 | 0.20 | 0.08 to 0.32 |
| Anxiety scale | 252 | 18.98 (2.72) | 236 | 18.36 (3.13) | 0.51 | 0.15 to 0.88 | .008 | 0.17 | 0.05 to 0.30 |
| Stress scale | 253 | 16.53 (3.67) | 236 | 15.54 (4.10) | 0.90 | 0.35 to 1.46 | .002 | 0.23 | 0.09 to 0.37 |
| DASS (dichotomized: within 85th percentile score of better mental health according to norms) | |||||||||
| Depression scalec | 252 | 222 (88.10) | 236 | 195 (82.63) | 1.68 | 1.08 to 2.60 | .020 | — | — |
| Anxiety scalec | 252 | 214 (84.92) | 236 | 184 (77.97) | 1.38 | 0.92 to 2.08 | .12 | — | — |
| Stress scalec | 253 | 229 (90.51) | 236 | 193 (81.78) | 2.09 | 1.28 to 3.42 | .003 | — | — |
| Personal well-being | 247 | 58.77 (12.93) | 228 | 56.07 (12.56) | 2.11 | 0.47 to 3.76 | .014 | 0.16 | 0.04 to 0.29 |
| Self-efficacyc,d | 249 | 192 (77.11) | 233 | 157 (67.38) | 1.60 | 1.19 to 2.16 | .002 | — | — |
Adjusted for baseline characteristics of child sex, family’s SEIFA score, maternal education, maternal age at child’s birth, parity, antenatal risk, maternal self-efficacy, and maternal mental health, plus child age at the 3-year assessment. —, not applicable.
The comparative statistic is the mean difference (intervention minus control) and ES for continuous outcomes and OR (relative odds for intervention compared with receiving usual care) for dichotomous outcomes.
Summary statistics are mean (SD) except where specified as dichotomous, in which they are n and %.
Outcome is dichotomous, n (%).
“No lack of self-efficacy” versus “any lack of self-efficacy.”
Results of the multiple imputation analyses (Table 4) are similar to the complete case analyses in estimated effects, confirming that the use of complete case analyses had not led to bias because of attrition between the treatment groups. Estimates had slightly wider CIs because the imputation of a large proportion of the study outcomes increased the uncertainty of estimation; thus, these analyses using multiple imputation are considered the more conservative.39
Adjusted Multiple Imputed Regression Analyses Comparing the 2 Trial Arms on Maternal Mental Health Outcomes at Child Age 3 Years
| Outcome | Intervention | Control (Usual Care) | Comparative Statistica: Intervention Compared With Control | ||||||
|---|---|---|---|---|---|---|---|---|---|
| n | Summaryb | n | Summaryb | Statistica | 95% CI | P | ES | 95% CI | |
| DASS (reversed: higher scores indicate better mental health) | |||||||||
| Total Score | 363 | 53.79 | 359 | 51.74 | 1.85 | 0.05 to 3.65 | .045 | 0.18 | 0.00 to 0.36 |
| Depression Scale | 363 | 18.51 | 359 | 17.89 | 0.59 | −0.05 to 1.24 | .069 | 0.16 | −0.01 to 0.34 |
| Anxiety scale | 363 | 18.82 | 359 | 18.31 | 0.40 | −0.28 to 1.07 | .23 | 0.12 | −0.08 to 0.32 |
| Stress Scale | 363 | 16.46 | 359 | 15.54 | 0.86 | 0.07 to 1.65 | .035 | 0.20 | 0.02 to 0.38 |
| DASS (dichotomized: within 85th percentile score of better mental health according to norms) | |||||||||
| Depression scalec | 363 | 91.74 | 359 | 88.58 | 1.53 | 1.05 to 2.24 | .028 | — | — |
| Anxiety scalec | 363 | 89.53 | 359 | 85.52 | 1.40 | 1.00 to 1.94 | .047 | — | — |
| Stress scalec | 363 | 93.39 | 359 | 88.02 | 1.95 | 1.20 to 3.16 | .007 | — | — |
| Personal well-beinga | 363 | 58.95 | 359 | 56.23 | 2.37 | −0.59 to 5.34 | .10 | 0.17 | −0.04 to 0.37 |
| Self-efficacyc,d | 363 | 75.68 | 359 | 65.44 | 1.74 | 1.26 to 2.40 | .001 | — | — |
| Outcome | Intervention | Control (Usual Care) | Comparative Statistica: Intervention Compared With Control | ||||||
|---|---|---|---|---|---|---|---|---|---|
| n | Summaryb | n | Summaryb | Statistica | 95% CI | P | ES | 95% CI | |
| DASS (reversed: higher scores indicate better mental health) | |||||||||
| Total Score | 363 | 53.79 | 359 | 51.74 | 1.85 | 0.05 to 3.65 | .045 | 0.18 | 0.00 to 0.36 |
| Depression Scale | 363 | 18.51 | 359 | 17.89 | 0.59 | −0.05 to 1.24 | .069 | 0.16 | −0.01 to 0.34 |
| Anxiety scale | 363 | 18.82 | 359 | 18.31 | 0.40 | −0.28 to 1.07 | .23 | 0.12 | −0.08 to 0.32 |
| Stress Scale | 363 | 16.46 | 359 | 15.54 | 0.86 | 0.07 to 1.65 | .035 | 0.20 | 0.02 to 0.38 |
| DASS (dichotomized: within 85th percentile score of better mental health according to norms) | |||||||||
| Depression scalec | 363 | 91.74 | 359 | 88.58 | 1.53 | 1.05 to 2.24 | .028 | — | — |
| Anxiety scalec | 363 | 89.53 | 359 | 85.52 | 1.40 | 1.00 to 1.94 | .047 | — | — |
| Stress scalec | 363 | 93.39 | 359 | 88.02 | 1.95 | 1.20 to 3.16 | .007 | — | — |
| Personal well-beinga | 363 | 58.95 | 359 | 56.23 | 2.37 | −0.59 to 5.34 | .10 | 0.17 | −0.04 to 0.37 |
| Self-efficacyc,d | 363 | 75.68 | 359 | 65.44 | 1.74 | 1.26 to 2.40 | .001 | — | — |
Adjusted for baseline characteristics of child sex, family’s SEIFA score, maternal education, maternal age at child’s birth, parity, antenatal risk, maternal self-efficacy, and maternal mental health, plus child age at the 3-year assessment. —, not applicable.
The comparative statistic is the mean difference (intervention minus control) and ES for continuous outcomes and OR (relative odds for intervention compared with receiving usual care) for dichotomous outcomes.
Summary statistics are mean (SD) except where specified as dichotomous, in which they are n and %.
Outcome is dichotomous, n (%).
“No lack of self-efficacy” versus “any lack of self-efficacy.”
Discussion
Benefits of the right@home NHV program were evident for maternal mental health and well-being outcomes at 1-year post intervention completion (at child age of 3 years). Our findings are aligned with results from 3 previous US NHV studies revealing small-to-modest ESs for maternal mental health from child age 2 to 4 years (ES: 0.1–0.6 for maternal depression symptoms and parenting stress).21–23 We found similar-sized benefits for maternal depression, anxiety, and stress symptoms (ES: 0.17–0.25) but note the additional positive benefits on personal well-being (ES: 0.16) and self-efficacy (OR: 1.60).
These findings also align favorably with programs that directly target maternal mental illness. Psychosocial and psychological interventions delivered by nurses, physicians, psychologists, researchers, or lay people have shown similarly modest positive benefits in the antenatal and first year postpartum periods (ES: 0.06–0.16).40,41 However, these benefits have either not been observed or assessed beyond the child age of 1 year nor were they delivered specifically to women experiencing adversity.
Despite substantial research and policy attention paid to maternal mental health and its known importance for children’s development, there are comparatively few published data on maternal mental health beyond the first year postpartum.8,17 An Australian population cohort of first-time mothers reported depression increasing with child age, with a peak at 4 years.17 Within our study, mothers receiving usual care also reported worse mental health symptoms from 2 to 3 years.26 In contrast, our findings for intervention mothers suggest that the right@home NHV program may have prevented or postponed this decline in mental health. These findings deliver on the initial aspiration of the program as a salutogenic prevention intervention embedded in health care for women who may benefit from additional support, rather than an intervention responding to crisis or illness.28
The strengths of our findings lie in the trial’s rigorous design and research collaboration. Implementation of the NHV program and the research evaluation were led by different institutions. Outcome assessments were completed by researchers who were blinded to intervention status. The trial is also strengthened by the high retention of study participants in both groups (69% over a 4-year study duration), despite the substantial adversity experienced by participants. For context, by 2-year follow-up, the UK Building Blocks study retained 71% of their cohort for self-reported outcomes,42 Pro Kind retained <50%,43 and the French CAPEDP retained only 31% of their original cohort.44 Although those lost to follow-up reported slightly more adversities at baseline, participant characteristics at 3 years remained balanced between the trial arms. In addition, multiple imputation methods were implemented to confirm estimates made by using complete data. Compared with the results of the complete cases analyses, multiple imputation methods provided a more conservative estimation of program effects in which benefits for maternal mental health were still evident. Given the large, multisite design of the trial, high participant retention, and confirmation using multiple imputation, we believe our findings should generalize to pregnant women experiencing adversity, in similar health care systems.
There are several limitations. The multiple self-reported mental health measures are likely to represent overlapping constructs. However, this was to encompass both the positive aspects of mental health and symptoms of mental ill health (ie, both positively and negatively framed items).32–34 Although mental health was examined by using both continuous and dichotomized symptom scores across multiple domains of the DASS, this allowed us to estimate the impact of NHV on a scale of mental health symptoms as well as rates of dichotomized mental health morbidity. The DASS is not a diagnostic tool; however, it is one of the only broad-spectrum, self-report mental health measures and is frequently used in research with clinical and population-level cohorts.36 We reported ESs to enable comparability with previous NHV programs and psychological interventions that have reported mental health outcomes using different measures. Although participation in the NHV program in general may have influenced how women reported their mental health (aside from the true benefit), we expect any potential response bias would be minimal, given the measures were collected 1-year after the intervention ended. A further limitation of the current study is that the exclusion criteria mean findings may not generalize to non-English speaking women or women with severe intellectual disability.
Given the crucial role that maternal mental health plays in optimal health of mothers and their children,8,11–13 addressing inequities in maternal mental health can generate substantial societal and mental health benefits.3,5 To realize this goal, the most efficient and equitable approach is to integrate mental health care into existing health services within a prevention and early intervention paradigm.1 Although economic data are not presented in this article, we note that self-reported use of health services after the intervention, from child age 2 years, was similar between the intervention and usual care groups. In future research, researchers will examine the cost consequences of the right@home NHV program outcomes, including maternal mental health benefits, identified at 3 years. With our findings, we provide evidence to support NHV as a potential platform to achieve substantial benefits through maternal mental health.6 Interventions for the prevention of postnatal depression have been estimated to save $23.3 million over 5 years in Australian mental health expenditure.5 Although these estimates are focused predominantly on intervention effects in the antenatal and first year postpartum, the similar benefits identified at 3 years postpartum suggest an additional advantage in focusing policy on maternal mental health beyond the first year.
Conclusions
The right@home NHV program led to emerging benefits for maternal mental health and well-being at child age 3 years, a year after the intervention ended. These findings reveal that an NHV program that is designed for women experiencing adversity can lead to later benefits to mental health, even when implemented within existing CFH services. Although vital for addressing the established mental health burden, the benefits of NHV delivered through the existing health care infrastructure may be most critical as the economic and psychosocial stressors of the coronavirus disease 2019 pandemic emerge for families with young children. At scale, there may be real potential to reduce inequities in maternal mental health.
Acknowledgments
The right@home sustained NHV trial is a research collaboration between the Australian Research Alliance for Children and Youth; the Translational Research and Social Innovation Group at Western Sydney University; and the Centre for Community Child Health, which is a department of the Royal Children’s Hospital and a research group of Murdoch Children’s Research Institute. We thank all families; the researchers, nurses, and social care practitioners working on the right@home trial; the antenatal clinic staff at participating hospitals who helped facilitate the research; and the Expert Reference Group for their guidance in designing the trial.
On request to the corresponding author (sharon.goldfeld@rch.org.au), deidentified participant data, study protocols, and statistical analysis plans can be made available after publication to researchers who provide a methodologically sound proposal for use of the data.
Drs Goldfeld, Price, Hiscock, and Kemp conceptualized and implemented the study design and contributed to the first draft and subsequent revisions of the manuscript; Dr Bryson implemented the study design, coordinated and supervised data collection, conducted the data cleaning and statistical analysis, and contributed to the first draft and subsequent revisions of the manuscript; Dr Mensah conceptualized and implemented the study design, provided statistical expertise in the trial design and conduct of the statistical analysis, and contributed to the first draft and subsequent revisions of the manuscript; Dr Gold conceptualized and implemented the study design, provided health economics expertise in the trial design, and contributed to the first draft and subsequent revisions of the manuscript; Ms Orsini provided statistical expertise in the trial design, conducted the data cleaning and statistical analysis, and contributed to the first draft and subsequent revisions of the manuscript; Dr Perlen implemented the study design, coordinated and supervised data collection, and contributed to the first draft and subsequent revisions of the manuscript; Dr Grobler provided statistical expertise and conducted the statistical analysis and contributed to the first draft of the manuscript; Ms Dakin and Ms Harris implemented the study design and on behalf of the sponsor (Australian Research Alliance for Children and Youth) as the funding holder and contributed to the first draft and subsequent revisions of the manuscript; Ms Harris implemented the study design and contributed to the first draft and subsequent revisions of the manuscript; and all authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.
This trial has been registered with the ISRCTN Register (http://isrctn.org) (identifier ISRCTN89962120).
COMPANION PAPER: A companion to this article can be found online at www.pediatrics.org/cgi/doi/10.1542/2020.032565.
FUNDING: Supported by the state governments of Victoria and Tasmania, the Ian Potter Foundation, Sabemo Trust, Sidney Myer fund, the Vincent Fairfax Family Foundation, and the National Health and Medical Research Council (NHMRC) (1079418). The Murdoch Children's Research Institute administered the research grant for the study and provided infrastructural support to its staff but played no role in the conduct or analysis of the trial. Research at the Murdoch Children's Research Institute is supported by the Victorian Government’s Operational Infrastructure Support Program. Prof Goldfeld was supported by NHMRC Practitioner Fellowship (1155290). Dr Mensah was supported by NHMRC Career Development Fellowship (1111160). Prof Hiscock was supported by NHMRC Practitioner Fellowship (1136222).
Comments