Effects of integrated programs for substance‐involved mothers on infant and child development outcomes: A systematic review and meta‐analysis

Abstract

Maternal substance use is a pressing public health issue that confers risk for maternal health, the parent‐infant relationship, and child development. Integrated interventions that jointly address maternal substance use and child development have shown promise for enhancing child outcomes. No research to date has focused exclusively on the outcomes of young children or examined potential moderators of the effect sizes of integrated programs. This review evaluates the pooled effect of integrated interventions for substance‐involved mothers on the developmental outcomes of their children. A comprehensive search strategy was conducted in seven databases (APA PsycINFO, CINAHL, Cochrane CENTRAL, Embase, MEDLINE, Sociological Abstracts, Web of Science) from January 2011 and May 2023. Studies were included if they reported on an intervention with at least one substance use treatment and one parenting or child treatment service for substance‐involved mothers of children under 6 years of age. A total of 21 studies met inclusion criteria, and 14 nonoverlapping studies reported on effect sizes with a pooled effect size of SMD = .470 (95% CI = .35, .59). There was a trend toward treatment duration being a significant moderator (p = .08). Additional high‐quality studies are needed to demonstrate the long‐term impact of these interventions.

1. INTRODUCTION

Substance use in pregnancy is a pressing public health concern associated with a host of deleterious effects for both mothers and their offspring (Dos Pinto et al., 2010; Santos et al., 2018). Public health surveys in Canada indicate that 3.1% of pregnant women report using cannabis during pregnancy, 2.6% report using cannabis while breastfeeding, and 1.4% report using opioids during pregnancy (Grywacheski et al., 2021). Polysubstance use tends to be common, particularly among individuals reporting higher use (Eiden et al., 2023). Similar prevalence rates have been found around the world, including in the United States (Prince et al., 2024), Europe, Asia, South America, and Oceania (Tavella et al., 2020). Alarmingly, Canada and the United States are currently facing national crises in opioid‐related overdose. Recent reports show sharp increases in substance‐related poisoning hospitalizations and overdose deaths among pregnant people in North America (Government of Canada, 2023; Han et al., 2024). Such findings underscore the urgent need for targeted attention toward programs and interventions that support substance‐involved mothers and pregnant people globally.

Substance use in pregnancy can have direct teratogenic effects on fetal development. Indeed, fetal exposure to drugs and harmful substances can cause important changes in brain structure and function (Ross et al., 2015), as well as affect fetal growth, behavior (e.g., impulsivity and attention), and language and reading skills (Behnke & Smith, 2013). It is important to note that different substances can impact child development in various ways. For instance, alcohol consumption in pregnancy can result in fetal alcohol spectrum disorders (FASD), while opioid use can lead to neonatal abstinence syndrome (NAS). Furthermore, emerging research suggests that certain opioids (e.g., fentanyl) may be linked to specific syndromes (see Wadman et al., 2023).

In addition to these teratogenic effects and the increased risk for pregnancy complications and maternal health difficulties, maternal substance use is associated with pernicious impacts on the parent‐infant relationship and child development. Indeed, substance use can significantly impair a mother's ability to respond contingently to her infant's needs and cues (Cataldo et al., 2019; Suchman et al., 2017), which can negatively impact the formation of a healthy bond. In more severe circumstances, substance use may lead to complete disruption of the attachment relationship if parent‐child separation occurs (Lander et al., 2013). Additionally, children of substance‐involved mothers have been shown to be at increased risk for cognitive (Conradt et al., 2023), behavioral (Guille & Aujla, 2019), emotional, and social difficulties persisting into childhood and adolescence (Irner, 2012). Maternal substance misuse often co‐occurs with other social risk factors that have negative implications for child development (Eiden et al., 2023). For example, a comprehensive review of 30 studies demonstrated that infants exposed to substances are at increased risk of child protective services involvement, exposure to maltreatment, and hospitalization for suspected maltreatment, compared to nonexposed infants (Austin et al., 2022). Taken together, children whose mothers engage in problematic substance use are more likely to be exposed to cumulative risk and consequently more likely to be at risk for poor developmental outcomes (Behnke & Smith, 2013; Ross et al., 2015).

Key Findings

• The pooled effect of integrated interventions on the developmental outcomes of children of substance‐involved mothers (N = 14, SMD = .470, 95% CI = .35, .59) was medium in size.

• Effect sizes were larger in studies where treatment duration was longer.

• Interventions that address both maternal substance use and child development may help to break cycles of risk for children who have been prenatally exposed to substances.

Statement of Relevance

Substance use during pregnancy and postpartum harms both mothers and children. Integrated interventions addressing maternal substance use, parenting, and child development show preliminary evidence for enhanced child development outcomes. However, previous reviews are now outdated, and there is a need to explore for whom and under what conditions these interventions work best. This review provides an up‐to‐date synthesis, showing that longer interventions may improve outcomes. High‐quality, longitudinal evidence is needed to further determine if these outcomes are maintained over time.

To mitigate poor outcomes associated with maternal substance misuse, there has been an emphasis on interventions to address maternal substance use and provide support to reduce psychosocial risk exposure for both mothers and their infants (Rutman et al., 2021). A national evaluation report conducted in Canada found that intervention programs that use relational, women‐centered, trauma‐informed, and culturally grounded approaches as well as provide women with basic needs support, housing support, substance use services, trauma and violence support, parenting support, and children's health services are most likely to be associated with enhanced outcomes for mothers and their children (Rutman et al., 2021). For example, women who participate in integrated interventions (i.e., interventions which simultaneously address maternal substance use and provide services for child development or the parent‐child relationship) decrease their substance use over time, have healthier births (76% of infants have a healthy gestation) and are less likely to require neonatal intensive care stays.

These findings are aligned with research evidence suggesting that integrated interventions are associated with improved outcomes for both mothers (Milligan et al., 2010) and their offspring (Niccols et al., 2012). Specifically, a review of 13 studies examining the effects of integrated programs for substance‐involved mothers and their children found small to large effect sizes for improvements in child development and large effect sizes for improvements in emotional and behavioral functioning of children prenatally exposed to substances (Niccols et al., 2012). Additionally, when comparing the outcomes of children whose mothers participated in integrated versus nonintegrated programs, those whose mothers were in integrated programs demonstrated greater improvements in emotional and behavioral functioning, although this advantage was small. Therefore, evidence from program evaluations and the scientific literature suggest that programs that address the needs of the mother (e.g., substance use and mental health support) and those of her child (e.g., child development services, parenting supports) have the greatest potential for optimizing the development of children of substance involved mothers.

It is important to note that the substance use and treatment landscape has changed significantly over the past decade. There has been a notable shift from primarily inpatient treatment to community‐based, outpatient programs for opioid use disorders (OUDs). These outpatient programs, such as medication‐assisted treatment and office‐based opioid treatment, enable individuals to receive comprehensive care while minimizing disruptions to their daily routines.

Although a previous review (Niccols et al., 2012) provides initial evidence that integrated interventions are effective in reducing developmental, behavioral, and emotional difficulties among substance‐exposed children (birth to 16 years), there are still gaps in our understanding of the effectiveness of integrated interventions for infants and young children (under age 6 years), specifically. First, previous work has not focused exclusively on the outcomes of infants and young children. Infancy and early childhood are considered sensitive periods where experiences of stress and deprivation may be particularly impactful (Gabard‐Durnam & McLaughlin, 2020). Additionally, pregnancy and the early years are often considered windows of opportunity where mothers are motivated to make changes in their substance use (Flykt et al., 2021). As such, the current review sought to focus exclusively on child outcomes prior to the age of 6 years. Second, previous work has not investigated for whom and under what conditions integrated interventions may be most effective. That is, potential moderating variables have not been examined. This is important as sociodemographic variables (e.g., maternal age, income, single parent status), methodological variables (e.g., study type, study quality), or treatment variables (e.g., treatment duration) may impact the effectiveness of integrated interventions. For example, longer and greater service use have been associated with more positive substance use, parent‐child relationship, and child development outcomes upon service completion (Andrews et al., 2018). Younger maternal age has also been associated with increased likelihood of premature treatment dropout (Scott‐Lennox et al., 2000). As such, the examination of moderator variables will shed light on who may benefit most from integrated interventions and what gaps in research continue to exist.

1.1. Current review

The current review aims to provide an update and expansion on the effects of integrated interventions on the developmental, behavioral, and emotional functioning, referred to as developmental outcomes herein, of infants of substance‐involved mothers. A previous review was conducted in 2012 and included 13 studies examining child outcomes up to age 16 years. Thus, the first objective of this current meta‐analysis is to provide an updated pooled effect size as of May 2023 with an emphasis on child outcomes in infancy and early childhood (0 to 6 years). The second objective is to examine potential moderators of the effect of integrated interventions on the outcomes of children of substance‐involved mothers. Understanding the effects of integrated interventions for infants and young children, as well as the circumstances under which they are most successful, can inform the development and delivery of services for substance‐involved mothers and their young children.

2. METHODS

This meta‐analysis is reported following the Preferred Reporting Items for Systematic Reviews and Meta‐Analyses (PRISMA) (Page et al., 2021). Ethics approval was not required for this meta‐analysis as all data were archival and available in peer‐reviewed journals. The protocol for this review was preregistered (PROSPERO #CRD42023444771).

2.1. Definition of predictor and outcome variables

For the purposes of the current meta‐analysis, integrated interventions for maternal substance use were operationalized as interventions where at least one specific substance use treatment (e.g., individual or group therapy; methadone) and at least one parenting or child treatment service (e.g., prenatal care, childcare, parenting classes, home visitation, child development services) were present. Children's developmental outcomes were measured using well‐validated measures and screening tools (e.g., Bayley Scales of Infant Development, Ages and Stages Questionnaire: Social‐Emotional).

2.2. Study identification

A research librarian (P.R.L.), along with the senior author (N.R.), created the preliminary search strategy for APA PsycInfo (Ovid). This strategy was informed, in part, by consulting Niccols and colleagues’ 2012 review and by considering strategies developed for reviews on substance use (Goldberg et al., 2021; McGovern et al., 2021) as well as for reviews focused on children (Hetrick et al., 2021; James et al., 2020). The initial search was peer‐reviewed by a second research librarian using the Peer Review of Electronic Search Strategies (PRESS) guideline for systematic reviews (McGowan et al., 2016). The search was conducted on May 9, 2023, in APA PsycInfo (Ovid), CINAHL (EBSCOhost), Cochrane CENTRAL (Ovid), Embase (Ovid), MEDLINE (Ovid), Sociological Abstracts (ProQuest), and Web of Science (Clarivate; SCIE, SSCI, and ESCI). The search included the following key terms: “early childhood development” OR “infant development” AND “mother child relations” OR “mothers” AND “drug addiction” OR “substance use disorder” AND “addiction treatment” OR “rehabilitation” OR “treatment” OR “intervention”. The search sought to identify studies published between January 2011 and May 2023 to capture any additional contemporary literature published since the Niccols and colleagues (2012) review that examines the impacts of integrated interventions on the development, behavioral functioning, and emotional functioning of infants and young children whose mothers are substance‐involved. No other restrictions were used when searching databases. Results were exported from databases and imported into Covidence, an online tool that facilitates various screening phases for reviews, where duplicates were removed. See Supplementary Material for the complete search strategy used in all databases.

In addition to the studies identified through the search and those included in Niccols and colleagues’ 2012 review, further studies were sought through a hand search with no date limitations to ensure maximum comprehensiveness. The hand search process entailed searching the reference lists of the included studies, as well as conducting a search on Google Scholar using the following key terms: (1) “women” and “substance abuse” and “substance use intervention” and “child development” and (2) “mother” and “substance abuse” and “substance use intervention” and “child development.”

2.3. Study inclusion and exclusion criteria

The first step for establishing which studies met the inclusion criteria involved reviewing titles and abstracts identified from the search strategy. Abstracts were screened independently and in duplicate by two screeners. The following inclusion criteria were used to assess references: (1) integrated intervention for maternal substance use (prenatal and/or postnatal use); (2) a measure of child developmental functioning (e.g., cognitive, behavioral, and emotional functioning); (3) the study reported a statistic that could be transformed into a correlation coefficient examining the association between the integrated intervention and the child's outcomes; (4) English language article; and (5) empirical article (e.g., case reports, review papers, and qualitative studies were excluded). Inclusion criterion 3 is stated specifically for inclusion in the formal analysis to create the pooled effect. The average inter‐rater agreement probability was .93 for the title and abstract review and .87 for the full text review, which are considered strong. Conflicts at both stages of the review were resolved by consensus.

In situations where more than one effect size could be extracted from a study, a protocol was followed so that each sample of participants was only represented once in the meta‐analysis. If a study provided more than one timepoint, the timepoint closest to the intervention time was selected (e.g., if the case of multiple follow‐up measurements, the follow‐up with the shortest measurement was selected). In situations where multiple child outcome domains were reported, a pooled effect size was calculated. In a situation where multiple publications reported on the same sample of participants, the publication with the most comprehensive and complete data reporting was selected. One study (Suchman et al., 2011) was not included in the pooled effect due to an overlapping sample size with an earlier study (Suchman et al., 2010). Finally, to ensure comprehensiveness, studies were retained for narrative description when an effect size was not reported or usable, given the small number of studies available. All included studies are listed in Table S1.

2.4. Data extraction

A standardized data extraction protocol was developed for identifying included studies and for extracting relevant data. Three study authors (S.B., L.F., N.R.) extracted study characteristics and effect sizes. Effect sizes refer to the magnitude of association between the integrated intervention and the child's development score. Potential moderators were also extracted and included: study design (cross‐sectional vs. longitudinal), geographical location, time of data collection, sample size, socioeconomic status (low vs. other), racial/ethnic minority status, age of mothers, age of infants/young children, type of child outcome, measurement type, time (in months) between the start of the intervention and developmental outcome measurement, diagnosis of substance use disorder, single parent home, parent mental health status, length of treatment, intensity of treatment, treatment completion, treatment setting (inpatient vs. outpatient) and study quality. Socioeconomic status of the sample was coded as “low” if the majority of the sample (80% or more) was of low socioeconomic status within the country where the study was conducted. Studies were coded as “other” for socioeconomic status if most of the sample (80% or more) was of “high” socioeconomic status or was mixed, whereby the majority of participants had varying levels of socioeconomic status. All data were extracted twice by two independent coders for reliability, and discrepancies were resolved by consensus.

2.5. Study quality

Study quality was assessed by two independent coders for each study based on a 14‐point tool from the National Institutes of Health Quality Assessment Tool for Observational Cohort and Cross‐Sectional Studies (National Heart, Lung & Blood Institute, 2021) (see Table S2). Conflicts were resolved by consensus, and no studies were excluded based on quality scores.

2.6. Data analysis

All extracted data were entered and analyzed using Comprehensive Meta‐Analysis (CMA, Version 4.0) software. Overall effect sizes, publication bias, and moderator analyses were calculated. Random effects modelling was used to calculate the overall pooled effect size. All effect sizes were converted to the standard mean difference (SMD), and 95% lower and upper confidence limits are provided. The SMD is often used in meta‐analysis when included studies use different scales or measurement instruments. The SMD can be interpreted like Cohen's d, whereby a small effect size is .2, a medium effect size is .5, and a large effect size is .8 (Andrade, 2020). We used the Egger test (Egger et al., 1997) and Duval and Tweedie's trim and fill method (Duval & Tweedie, 2000) to assess for publication bias, which refers to the absence of studies in the published literature that demonstrate null findings. If publication bias was indicated, the trim and fill method was used to correct for publication bias. Heterogeneity of effect sizes was examined using the Q statistic. Although the significance of the Q‐statistic can be used to determine whether moderators should be examined, it is heavily influenced by the number of primary studies, and, as such, moderators of interest were examined in the current study regardless of the statistical significance of the Q statistic (Parr et al., 2019). We examined moderators using meta‐regressions to examine whether continuous moderators could explain variability across studies. In line with recommendations by Borenstein et al. (2011), we only examined categorical moderators when there were at least three studies per category. Meta‐regressions were only calculated when there were more than four effect sizes.

Due to the small number of studies reporting effect sizes (n = 14), we considered whether to pool effect sizes across study designs (e.g., integrated intervention vs. control, integrated intervention vs. nonintegrated intervention, and pre‐post design). The Q‐statistic, which identifies the variation between each study's estimated effect and the overall effect size estimate, was not significant (Q = .241, p = .887); therefore, we opted to pool effect sizes across study types. The overall pooled SMD for studies that compared integrated interventions to a control group, studies that compared integrated interventions to a nonintegrated intervention, and studies that used a pre‐post design was .535 (95%CI = .21–.86), .462 (95%CI = .15–.77), .445 (95%CI = .30–.60), respectively. There was no difference among these effective sizes.

Not all studies that met the inclusion criteria provided an effect size that could be extracted for the purpose of the meta‐analysis. As such, for studies where a usable effect size could not be extracted, the findings were summarized narratively.

3. RESULTS

3.1. Studies selected

The PRISMA flow diagram presents how studies were selected and included in the meta‐analysis (see Figure 1). The initial electronic search from seven databases yielded 14,297 articles after duplicates were removed. Following the review of titles and abstracts, 97 studies were identified as potentially meeting inclusion criteria, and the full‐text articles were reviewed. After full article review, n = 8 studies published after the Niccols et al. (2012) review met full inclusion criteria. A total of 10 studies were included from the Niccols et al. (2012) systematic review examining the effect of integrated interventions for child development outcomes. A total of three additional studies meeting the inclusion criteria were identified and retrieved through a hand search. In total, 21 studies met the inclusion criteria. Of these, six studies did not report effect sizes, and one had an overlapping sample size. As such, 14 effect sizes were included in the pooled effect size. The characteristics of all 21 included studies are presented in Table 1. References for included studies can be found in Table S1.

FIGURE 1.

Preferred reporting items for systematic reviews and meta‐analyses (PRISMA) flow diagram.

TABLE 1.

Included study characteristics.

First author name, year	N a	Outcome measure	Time point	Study design	Income	% Minority	Maternal mean age a	Child age range	Country	Quality score
Andrews, 2018	160	Counsellors’ ratings	At service ending	Single measurement	Mixed	–	29.8	0–6 years	Canada	11
Belcher, 2005	57	BSID & BSID II	–	Single measurement	Low	69.6%	28.1	–	US	6
Camp, 1995	26	BSID	6 and 12 months of age	–	–	–	–	–	US	–
Cheng, 2022	123	Clinical assessment or diagnosis	Every 2–6 months until 5 years of age	Integrated versus nonintegrated	–	–	25.5	–	China	8
Conners, 2001	72	Refusal Skills scale; SIB‐R; DDST II	Intake, every 3 months after intake, and at 3‐, 6‐, and 12‐months post discharge	Pre‐post	Low	–	–	0–14.9 years	US	13
Field, 1998	126	Littman and Parmelee Obstetric and Postnatal Complications Scales; The Pediatric Complications Scale; BNBAS; The Early Social Communication Scales; BSID; INFANIB	3, 6, and 12 months of age	Integrated versus no treatment	Low	90%	18	3–12 months	–	7
Jansson, 1996	100	BSID; DDST‐II; BNBAS;	NBAS: birth and 2 weeks of age; DDST‐II: 3, 6, 9, 12, 15, 18, and 24 months of age; BSID: 6, 12, and 24 months of age	Pre‐post	Low	85%	27.7	0–24 months	US	4
Kerwin, 2007	11	CBCL	3 months after intake and at discharge	Pre‐post	–	–	–	–	U.S	–
Killeen, 2000	86	Infant and Child Development Inventory; Brigance Screen; CBCL	Intake and 6 months post‐discharge	Pre‐post	Low	71.4%	31.5	–	US	4
Nair, 2003	70	BSID; REEL	6, 12, and 18 months of age	Integrated versus no treatment	–	–	27.5	0–24 months	US	10
Niccols, 2005	13	CDI	Intake, 3 and 6 months in the program	Pre‐post	–	–	–	0–6 years	Canada	8
O'Malley, 2021	220	5‐point Likert scale to assess child health	Intake, 3–7, 9–13, 18–22 months of age, and at discharge	Pre‐post	Low	43.7%	–	0–24 months	US	4
Pajulo, 2011	34	BSID II	4 months of age	Single measurement	–	–	25.1	0–2 years	Finland	9
Pajulo, 2012	34	BSID II	4 months of age	Single measurement	–	–	25.1	0–2 years	Finland	6
Pepler, 2002	171	BDI	Intake, 6, 12, 18 and 24 months after intake	Pre‐post	Low	–	–	0–6 years	Canada	8
Schuler, 2003	54	BSID	Baseline, 6, 12, and 18 months of age	Integrated versus no treatment	Low	91%	28.2	2 weeks‐18 months	US	10
Suchman, 2010	23	NCAST	Baseline (0 months), post‐treatment (3 months), and 6‐week follow up (4.5 months)	Integrated versus nonintegrated	Low	22%	31.4	0–36 months	U.S.	11
Suchman, 2011	23	NCAST	Baseline (0 months), post‐treatment (3 months), and 6‐week follow up (4.5 months)	Integrated versus nonintegrated	Low	Intervention: 22% Control: 38%	31.4	0–36 months	US	11
Suchman, 2017	40	SSP	Baseline (6 weeks), 3 and 12 months	Integrated versus nonintegrated	Low	22.9%	29.68	11 months – 5 years	US	8
Weir, 2021	97	Y‐OQ	Prior to treatment, 1‐month intervals during treatment and at end	Pre‐post	Low	–	28.3	2 days–11 years	US	8
Whiteside‐Mansell, 1999	72	BSID	6, 12, and 18 months of age	Integrated versus no treatment	Low	Intervention: 75% Control: 70%	28.8	0–18 months	US	7

Abbreviations: BDI, Battelle Developmental Inventory; BNBAS, The Brazelton Neonatal Behavioural Assessment Scale; BSID, The Bayley Scales of Infant Development; CBCL, Child Behavior Checklist; CDI, Child Development Inventory; DDST II, Denver Developmental Screening Test II; INFANIB, Infant Neurological Infant Battery; NCAST. Nursing Child Assessment Satellite Training; REEL, Receptive‐Expressive Emergent Language Scale; SIB‐R, Scales of Independent Behavior (Revised); SSP, Strange Situation Procedure; Y‐OQ, Youth Outcome Questionnaire 2.01.

^a Value for the intervention group is reported.

3.2. Effects of integrated interventions on child developmental outcomes

3.2.1. Study characteristics and quality

The sample size of the intervention groups ranged from 11 to 220 mother‐child dyads. Mean maternal age was 27.75 years (SD = 3.32). On average, 59.26% of mothers were racially or ethnically minoritized, 59.81% were single parents, and 55.4% were experiencing mental health difficulties. The average age at the time of intervention was 17.25 months (SD = 8.14). Studies were conducted in the United States (n = 14, 71.4%), Canada (n = 3, 14.3%), Finland (n = 2, 9.5%), and China (n = 1, 4.7%). The average study quality score was 8.05 (range: 4–13; see Table S3).

3.2.2. Pooled effect size estimates

A total of 14 nonoverlapping studies were used to estimate the pooled effect size for the effect of integrated interventions (i.e., interventions where there is at least one specific substance use treatment and at least one parenting or child treatment service). Substance use treatments included individual counseling, relapse prevention groups, and detoxification, whereas parenting treatments included new mothers' support groups, parent education, and employment assistance. Child‐specific treatment services included medical follow‐ups, child care, and play therapy (see Table 2 for a detailed overview). A random‐effects meta‐analysis produced a medium effect size of SMD = .470 (95% CI = .35, .59; see Figure 2). The Egger test result for publication bias was significant (p = .048), and the forest plot was largely symmetrical (see Figure S1). Using Duval and Tweedie's trim and fill technique, four studies were identified to be added to adjust for publication bias. The adjusted pooled estimate with the addition of four studies was .38 (95% CI = .25, .52).

TABLE 2.

Overview of integrated interventions.

Study	Substance use intervention	Parenting or infant intervention	Duration (average, in months)	Treatment Setting	Intervention outcomes
Andrews, 2018	Individual counseling, relapse prevention group, basic life skills groups, emotion awareness group, connection group and recovery group.	Pediatric medical appointments, home visiting program, new moms support group, learning through play group, and making the connection group.	17.4	Outpatient	‐ Child development and well‐being
Belcher, 2005	Intensive center‐ and home‐based social services, addiction counseling (group and individual), and community‐based services.	Parent education, home visits, and infant/child developmental intervention.	4.4	Outpatient	‐ Child development
Camp, 1995	Group and individual addiction treatment.	Maternal mental health services, group and individual parenting education and counseling, life skills training, prenatal education, medical and nutrition services, education and employment assistance, obstetrical and pediatric care, child care, children's services, and aftercare.	–	Inpatient/Residential	‐ Child development
Cheng, 2022	Parent drug detoxification, parent support network (peer support classes) for pregnant women with illicit drug use; social workers motivate mothers for drug detoxification and stop maternal high‐risk behaviors like smoking and drinking.	Parent education groups, parenting coaching; social workers provide regular home visitations for emotional and parental support to the mother.	60.0	Outpatient	‐ Child development ‐ Learning disorders ‐ Psychological and behavioral outcomes (e.g., ADHD, ASD, and non‐specific behavioral disorders, emotional disorders)
Conners, 2001	Alcohol and other drugs (AOD) education groups, transportation to offsite Cocaine or Narcotics Anonymous groups, weekly relapse prevention groups, daily group counseling on alcohol and drug related issues, and weekly individual counseling sessions.	Group and individual therapy designed to address family and relationship issues, grief and anger, self‐esteem building, and other topics. Educational services including life skills training (including budgeting, household management, etc.), health education, and parenting education. Children also received a variety of educational and mental health services. An onsite licensed child care center was available for full‐day care of young children, and afterschool and full‐day summer care for school‐age children. Early intervention services were provided as needed to young children, and tutoring was provided to school‐age children.	9.9	Inpatient/Residential	‐ Child development ‐ Adaptive behavior
Field, 1998	Drug rehabilitation, group therapy, psychoeducational sessions, self‐help group sessions, and individual and drug counselling.	Parenting classes.	6.0	Outpatient	‐ Child development ‐ Child behavior ‐ Social communication ‐ Neurological development and functioning
Jansson, 1996	Individual and group counseling (including several individualized treatments such as methadone maintenance, psychiatric consultation and therapy for concomitant psychiatric disorders, and specialized primary care and social services for HIV+ women).	Both individual and group parenting classes (e.g., CPR, lactation group), The Developmental Play Program (childcare), home visits, prenatal care, and child development services.	1.78	Hybrid	‐ Child development ‐ Child behavior
Kerwin, 2007	Group and individual addiction treatment.	Maternal mental health services, group and individual parenting education and counseling, life skills training, prenatal education, medical and nutrition services, education and employment assistance, obstetrical and pediatric care, child care, children's services, and aftercare.	–	Outpatient	‐ Child behavior
Killeen, 2000	Relapse prevention, 12‐step participation, substance abuse education, and involvement in a number of community programs.	Parenting classes, counseling for psychological problems, therapeutic child care, structured mother–child interaction groups, medical services, and appropriate referrals to other services in the community.	4.5	Inpatient/Residential	‐ Child development ‐ Child behavior
Nair, 2003	Provided information about drug use and drug treatment.	Home visits, which include developmentally oriented child and parent components and are based on the Infant Health and Development Program (IHDP).	24.0	Outpatient	‐ Child development, including language development
Niccols, 2005	Addiction groups and individual addictions counselling.	Nutrition counselling and skill development, parenting education and counselling, peer support, enriched children's programming and instrumental assistance (e.g., bus tickets, grocery vouchers).	4.0	Outpatient	‐ Child development
O'Malley, 2021	Counseling, crisis intervention, transportation, and support for alcohol and other drug treatment.	Home visitations conducted by a parent resource specialist and a family support specialist, access to a women's support group, and assistance in coordinating services with other community agencies (e.g., drug treatment, child welfare, health care, criminal justice).	24.0	Outpatient	‐ Child health
Pajulo, 2011	Buprenorphine substitution medication (20% during pregnancy; 15% during postnatal phase).	Group meetings aimed at supporting daily situations with child.	4.0	Inpatient/Residential	‐ Child development
Pajulo, 2012	Group meetings on relapse issues and to support abstinence from substances.	Individual counseling, group meetings on parenting themes, support for changing social networks, and encouragement of participation in external services (mental health, detoxification, etc.).	9.0	Inpatient/Residential	‐ Child development
Pepler, 2002	Relapse Prevention Group, Recovery Group, Life Skills Group, Women's Issues Group, and Individual Counseling.	New Moms' Support Group, Nobody's Perfect Parenting Program, Cooking Healthy Together, Parent‐Child Mother Goose Program, Parent‐Child Counseling, Hanen “You Make the Difference” Group, and Child Welfare Access Visits. Children also received Developmental Screening and Assessment, Learning Through Play Group, Home Visits, Pediatric Clinic, and Individual Pre‐Postnatal Counseling.	–	Outpatient	‐ Child development
Schuler, 2003	Provided information about drug use and drug treatment.	Home visits, which include developmentally oriented child and parent components and are based on the Infant Health and Development Program (IHDP).	18.0	Outpatient	‐ Child development
Suchman, 2010	Cognitive‐behavioural therapy, psychiatric services, substance replacement therapy (e.g., methadone, naltrexone), and medical care, among other services to assist them in life.	Child care and child guidance services (e.g., provided with pamphlets on a parenting topic of their choice).	3.0	Outpatient	‐ Parent‐child interactions
Suchman, 2011	Cognitive‐behavioural therapy, psychiatric services, substance replacement therapy (e.g., methadone, naltrexone), and medical care, among other services to assist them in life.	Mothers and Toddlers Program (MTP), an individual psychotherapy intervention.	3.0	Outpatient	‐ Parent‐child interactions
Suchman, 2017	72.4% were enrolled in methadone maintenance and 12.6% were enrolled in suboxone treatment.	Mothering From the Inside Out (MIO) Intervention, a manualized individual therapy to enhance reflective functioning in the parenting role, as well as developmentally informed child‐care.	3.0	Outpatient	‐ Parent‐child attachment
Weir, 2021	Standard individual and group therapy sessions regarding substance abuse.	Theraplay, a counseling approach using play therapy (attachment‐based play).	5.0	Inpatient/Residential	‐ Emotional functioning ‐ Behavioral functioning
Whiteside‐Mansell, 1999	Group and individual counseling covering the disease process of addiction; physiology and pharmacology of tobacco, alcohol, and other drugs as well as effect of tobacco, alcohol, and other drugs on the mother and fetus.	Parenting education and support, health services (prenatal care, delivery services, postpartum and interconceptional care, and child health services), health education (prenatal infant health and development, nutrition, child safety, labor and delivery, and family planning), as well as group and individual counseling.	3.75	Outpatient	‐ Child development

Note: The intervention outcomes column includes only child‐related outcomes measured, as this was the focus of the current study.

FIGURE 2.

Forest plot for study effect sizes.

An additional meta‐analysis of the four studies (Field et al., 1998; Nair et al., 2003; Schuler et al., 2003; Whiteside‐Mansell et al., 1999) comparing integrated interventions to a control group that received no treatment yielded an effect size of SMD = .528 (95% CI = .20, .85), indicating a moderate positive effect.

3.2.3. Moderator analyses

Examining continuous moderators, maternal age, racial/ethnic minority status, single parent status, publication year, and study quality did not explain between‐study heterogeneity (see Table 3). There was a trend toward treatment duration explaining heterogeneity (p = .08), whereby studies with greater treatment duration had higher effect sizes (see Figure S2). There was no difference in treatment effects for inpatient versus outpatient settings (p = .52). There was an insufficient number of studies reporting on maternal mental health status to examine moderation. Additionally, most studies (n = 14) included women who were low‐income, and as such, there was not enough variability in this variable to examine it as a moderator.

TABLE 3.

Results of moderator analyses for the effect of integrated interventions on child development outcomes.

Child developmental outcomes
Continuous moderators	k	b	95% CI	Z	p
Maternal age	10	.01	−.04, .06	.36	.72
Treatment duration	12	.007	−.001, .02	1.76	.08
Racial/ethnic minority status	7	.004	−.004, 0.012	1.05	.30
Single parent	6	−.003	−.01, .005	−.75	.46
Publication year	14	−.01	−.02, .001	−1.29	.20
Study quality	13	.004	−.05, .06	.15	.88

Note: There were insufficient studies to test whether maternal mental health status or income were significant moderators. **p < .05.

Abbreviations: b, estimate; CI, confidence interval.

3.2.4. Narrative findings

Six studies (Andrews et al., 2018; Belcher et al., 2005; Camp & Finkelstein, 1995; Jansson et al., 1996; Pajulo et al., 2011, 2012) did not report effect sizes that could be included in the current review, and one study (Suchman et al., 2011) reported on an overlapping sample. Andrews and colleagues (2018) found no significant relationship between the duration of service use and counsellors’ ratings of improvement in child development (r = .23, p > .05). However, they did report significant associations between improvements in child development and the number of woman‐focused services (r = .34, p = .04), the number of parent/child focused services (r = .36, p = .03), the number of group services (r = .47, p = .003), the proportion of group services (r = .36, p = .02) and the range of services (r = .35, p = .03).

Several studies reported that children of substance‐involved mothers had average developmental scores when mothers were engaged in an integrated intervention. For example, Belcher and colleagues (2005) found that children participating in the STRIVE program had an average range of scores on the Bayley Scales of Infant Development (BSID). Specifically, the mean scores on the Mental Developmental Index (MDI) were 100.9 (SD = 21.5) for infants below 6 months of age, 103.2 (SD = 14.6) for children 6 to 12 months, and 92.8 (SD = 6.5) for children 12 to 18 months. Additionally, Camp & Finkelstein (1995) found that 97% of children participating in an integrated program had development scores above a desired cut‐off. In Jansson and colleagues’ (1996) study, findings revealed that at 6 months of age, infants had a mean Mental Development Inventory (MDI) score of 100 (ranging from 69 to 129) and a mean Psychomotor Developmental Index score of 110 (ranging from 80 to 133). At the 12‐month timepoint, the mean MDI score was 107 (ranging from 66 to 134) while the mean PDI score was 107 (ranging from 80 to 122). At the 24‐month timepoint, the mean MDI score was 98 (ranging from 87 to 114) and the mean PDI score was 119 (ranging from 99 to 145). Despite the small sample size, these findings suggest that when receiving an integrated intervention, children of substance‐involved mothers with high levels of adversity still had development scores generally in the average range. It is important to note that large ranges are reported in the child outcomes across these studies, which may be related to varying exposure of substance type, timing, and duration, which can have an impact on child development.

Finally, in Pajulo and colleagues’ 2011 study, every infant scored within the normal range on the MDI, whereas 93% of the sample fell within normal ranges on the PDI. Furthermore, in their 2012 study, Pajulo and colleagues found that all infants had a development score within normal ranges at 4 months (BSID‐II MDI mean = 97.5, SD = 7.0; BSID‐II PDI mean = 95.3, SD = 8.4).

4. DISCUSSION

Interventions supporting mothers, their children, and the mother‐child relationship, often called integrated interventions, have shown promising outcomes for families grappling with substance use issues. Understanding the effectiveness of these interventions and determining who benefits most from them can inform program planning and implementation. This systematic review and meta‐analysis sought to provide an update and expansion on a previous review (Niccols et al., 2012), with a specific focus on child outcomes during infancy and early childhood (ages 0 to 6 years). Additionally, it examined several potential moderators that might influence the effect of integrated interventions on the outcomes of children of substance‐involved mothers. This synthesis found a moderate effect size for the effect of integrated interventions on child outcomes and a trend toward treatment duration being a moderator, suggesting that longer treatments were associated with larger effect sizes. Findings from Niccols et al.’s 2012 paper revealed a wider range of effect sizes, with larger effects observed for emotional and behavioral functioning outcomes. Overall, the results of the present study support the findings of Niccols and colleagues.

This review's first key finding is that integrated interventions, which encompass both addiction services and pregnancy‐, parenting‐ or child‐related services, have demonstrated moderate and positive effects on children's developmental outcomes. There are several potential mechanisms by which integrated interventions lead to enhanced child development outcomes. First, an important initial step within integrated interventions involves addressing and reducing maternal substance use difficulties (Neo et al., 2021). This not only decreases children's exposure to harmful substances in utero but also has a cascade of beneficial effects on their health and well‐being. Indeed, treating maternal substance use is associated with significant reductions in psychosocial risks, including homelessness and low income (Conners et al., 2004), as well as food and housing insecurity, family conflicts, and chaotic home environments (Moreland et al., 2023). These factors, especially when combined, present serious threats to children's well‐being. Moreover, reducing maternal substance use correlates with improved maternal mental health outcomes, particularly regarding depression (Chapman & Wu, 2013). Indeed, studies included in the current review showed that, at posttreatment specifically, mothers in the “Mothers and Toddlers Program” had significant reductions in substance use, as well as fewer reported symptoms of depression and psychiatric distress (Suchman et al., 2010; 2011). Maternal mental health is associated with better development outcomes, carrying long‐term implications for children's health, cognitive abilities, and socio‐emotional skills (von Hinke et al., 2022).

Second, the integration of parenting components into interventions bolsters the parent‐child relationship (Jeong et al., 2021), which ultimately influences development. Indeed, interventions aimed at increasing positive parent‐child interactions and reducing less adaptive parenting may help prevent behavioral problems and enhance developmental outcomes in children (Niccols et al., 2012). An evaluation of the “New Choices” intervention for substance‐using mothers and their children demonstrated significant improvements in mothers’ parenting attitudes. Specifically, after 3 months of program participation, mothers showed greater empathy toward their children (Niccols & Sword, 2005). Importantly, through improved parenting practices and behaviors, integrated interventions for maternal substance use have the potential to reduce foster care placements, emergency room visits, medical and psychiatric admissions, as well as enhance health and social care broadly (Niccols et al., 2012). As is the case with many interventions, early engagement of mothers is crucial, as it can help prevent the establishment of maladaptive parenting practices (Suchman et al., 2017), which in turn, profoundly impact children's developmental outcomes.

Women who are substance involved often face psychosocial risks and inequalities that can make it more challenging to access and engage in treatment compared to women who are not substance involved, especially in early phases (Pepler et al., 2002). Once engaged, significant trust building, and safety must be established prior to engaging in therapeutic work (Kruk & Sandberg, 2013). Thus, engaging substance involved mothers and their children can take time and longer integrated intervention durations (i.e., a greater number of sessions) may confer optimal outcomes. This trend was observed in this current meta‐analysis, whereby integrated intervention duration was found to be associated with greater effect sizes. Indeed, the previously mentioned “New Choices” intervention showed greater improvements after 6 months compared to 3 months (Niccols & Sword, 2005). Additionally, longer treatment durations have been linked with various positive outcomes, including abstinence from drug use, employment, higher income, reduced risk of arrest, improved maternal mental health, and more positive parenting attitudes (Conners et al., 2006), as well as enhanced pregnancy and neonatal outcomes (Milligan et al., 2011)–all factors directly contributing to children's developmental well‐being. Additionally, a mixed‐methods study also demonstrated that children who receive earlier and longer duration of integrated interventions have more optimal neurodevelopmental outcomes (Bondi et al., 2021), suggesting the important advantages of interventions during the sensitive period of early childhood (0 to 6 years).

Another potential explanation for this finding could be that significant time and support are required to stabilize mothers experiencing substance use issues and establish trust with clinicians while addressing deep‐rooted, co‐occurring trauma. Indeed, program evaluation findings have indicated that mothers are more likely to have improved outcomes with regards to their substance use, housing, and to keep their children in their care when they were engaged in an intervention program for a longer period (Rutman et al., 2021). Alternatively, Conners and colleagues (2006) have hypothesized that mothers who engage in treatment for longer durations may be more strongly motivated to change their life trajectories, leading to better outcomes for themselves and their children. Nevertheless, the multidimensional and relational aspects of integrated interventions are likely to contribute both directly and indirectly to mothers’ engagement. Key features that have shown success in increasing mothers’ participation duration include creating an environment centered on women's health and social issues and providing on‐site prenatal care, childcare, and child development services (Milligan et al., 2011).

Lastly, it is important to address the finding that maternal age, racial/ethnic minority status, single parent status, publication year, and study quality did not significantly moderate the relationship between the effect of integrated interventions on child outcomes. There are several potential explanations for this finding. First, given that participants in included studies shared similar demographics (e.g., most were under 30 years old, and many identified as racial/ethnic minorities), it is possible that there was insufficient variability within the sample. Additionally, the lack of statistical significance may be attributed to missing data concerning these variables. In fact, only nine of the included studies reported data on single‐parent status. It is also possible that there is no moderation effect, that the integrated mother‐child support is what is foundational to women's and children's outcomes for all populations.

5. STRENGTHS AND LIMITATIONS

This meta‐analysis provides an update on pooled effects for integrated interventions for mothers and children who are substance‐involved, as well as potential moderators of effect. Although previous research has demonstrated the positive effects of integrated interventions for child development (Niccols et al., 2012), this review includes contemporary studies and points to the importance of treatment duration when providing services to this population. There are some limitations that must be considered. First, there were insufficient studies to examine different study designs separately (e.g., pre‐post designs and intervention/control designs). Some study designs allow us to make firmer conclusions about the cause of effects, and our ability to do this in the current review was limited. Another limitation is that more than half of the studies did not report on race or ethnicity data. As such, there may be gaps in understanding how interventions or treatments affect different racial and ethnic groups. Understanding the demographic characteristics of study participants, including race and ethnicity, is crucial for assessing the generalizability and potential disparities in the findings (Canadian Institute for Health Information, 2022). There was also a paucity of sociodemographic information across studies, making it difficult to extract moderators. Sociodemographic factors such as age, gender, education, income, and employment status can influence how individuals respond to interventions and may interact with other variables to affect outcomes. It is also important to acknowledge that prenatal substance exposure—both the type of substance used and the timing—are key confounding factors that may influence child outcomes and the effectiveness of interventions; however, few studies in our review examined this issue. Lastly, the search strategy for the current review did not include relational terms (e.g., mother‐child, parent‐infant) to identify relevant studies, which may have resulted in the omission of some studies. To conclude, these limitations highlight the need for larger and higher‐quality studies that follow recommended guidelines for reporting with regard to the sociodemographic characteristics of participants.

6. CLINICAL IMPLICATIONS

From a practice perspective, maternal substance use interventions are often provided in the health sector, as opposed to the child development or children's mental health sectors. Health and addictions services that are provided to mothers are often divorced from services that are provided to children and are funded through different mechanisms. Indeed, substance‐involved mothers are often vilified and not deemed eligible to parent or supported to participate in general services (Greaves & Poole, 2004; Kruk & Banga, 2011). Integrated interventions that provide wraparound support to mothers allow them to address the simultaneous goals of decreasing their substance use, enhancing their mental health, improving their parenting, and maintaining custody of their children. Previous research has shown that mothers are able to meet these goals in tandem when being supported by integrated programs (Rutman et al., 2021).

The development and implementation of integrated programs to address maternal substance use and child development outcomes requires using frameworks that not only consider the needs of the mother but specifically target the mother‐child relationship and child development as part of a family service plan. For example, in addition to identifying goals for enhancing maternal health and decreasing substance use, routine assessment of child development and specific interventions that target the child, such as enriched childcare, speech‐language services, occupational therapy support, and enhancement of the attachment relationship through evidence‐based parenting interventions and home‐visiting, are needed. Collaboration across sectors and the inclusion of child development specialists, early childhood educators, child psychologists, and parent‐infant therapists are needed to ensure that service goals for the mother, the child, and the dyad are being met as part of their program involvement. Growing research evidence suggests that integrated programs need to employ a relational approach to service provision that promotes the relationship between the mother and child, the mothers and service provider, and among service providers and community partners, are critical to the support of women and children who are substance involved (Andrews et al., 2018).

7. DIRECTIONS FOR FUTURE RESEARCH

This meta‐analysis provided an updated synthesis of the current literature examining the effects of integrated interventions on the developmental functioning of children of substance‐involved mothers. Findings from the current review support integrated interventions as best practice for promoting the development of children of mothers who are substance‐involved. Specifically, findings suggest that when substance use programs for women consider their parenting role and their children, this increases the likelihood that their children will have the opportunity to thrive. In the current meta‐analysis, eight studies compared the child development outcomes of children whose mothers were in integrated interventions versus those in nonintegrated interventions or a control (Cheng et al., 2022; Field et al., 1998; Nair et al., 2003; Schuler et al., 2003; Suchman et al., 2010, 2011, 2017; Whiteside‐Mansell et al., 1999). Across the four studies (Field et al., 1998; Nair et al., 2003; Schuler et al., 2003; Whiteside‐Mansell et al., 1999) that compared integrated interventions to a control, child development outcomes were enhanced in the intervention as compared to children whose mothers did not receive an intervention. It is possible that the developmental benefits of integrated interventions may be demonstrated later as increased neurocognitive capacities develop and come online. As such, it is important for future research to conduct follow‐up studies to identify the long‐term child development benefits of integrated programs.

Future research should apply various methods to advance this field of study and ensure that future findings are applicable and practical. First, future research must prioritize high‐quality studies that comprehensively evaluate the impact of integrative interventions on infant and child outcomes. Ideally, this would entail comparing infant and child outcomes across mothers engaged in integrative intervention programs and those in nonintegrative programs. By designing studies that compare integrated and nonintegrated programs, meaningful conclusions could be drawn about the empirical validity of integrated interventions (Tolin et al., 2015). Moreover, conducting multi‐site studies could further help compare the impact of integrated and nonintegrated interventions on child outcomes. Multi‐site studies may allow for the recruitment of more extensive and more diverse samples. As a result, study findings may be more generalizable. Although randomized control trials would allow for more robust conclusions to be drawn, given that it is the gold standard for intervention evaluations (Piantadosi, 2005), it is imperative to consider the feasibility of such a design when working in community settings.

Second, future research should strive to comprehensively collect information on variables that may impact the relationship between integrative interventions and infant and child outcomes. The effectiveness of an intervention may vary considerably depending on individual characteristics, such as socioeconomic status, maternal and child age, mental health challenges, and racial/ethnic identity (Acevedo et al., 2012; Lui et al., 2017; Mudiyanselage et al., 2024). Yet, as demonstrated by the review, many studies fail to collect such variables. For example, in this review, there was an insufficient number of studies reporting on maternal mental health status to examine moderation. Future research must prioritize investigating the impact of individual characteristics on the relationship between integrative interventions and child outcomes to identify subgroups that may benefit most from specialized interventions.

Third, as this review found larger effect sizes in studies with longer treatment duration, it is crucial to ensure that future studies implement long‐term follow‐ups. While the included studies assessed short‐term outcomes of integrated interventions, longer‐term follow‐ups would be beneficial as they could determine if these outcomes are maintained over time and evaluate their impact in later development stages. Additionally, it is important to consider nuanced measurement of treatment engagement given that treatment duration might not entirely capture the nature of the treatment women receive. For example, due to the chaos and challenges in their lives, many substance‐involved mothers have difficulty attending services regularly, which may extend their treatment duration. Consideration of depth and range of services (e.g., Andrews et al., 2018) received may provide additional insight.

As highlighted by the current meta‐analysis, fewer than half of the included articles reported the racial and ethnic breakdown of their samples. This information is essential for gaining a deeper understanding of how race and ethnicity intersect with other forms of marginalization within this vulnerable population. Moreover, it is critical to determine whether there is a need for integrated interventions specifically tailored to different cultural and ethnic groups. This information remains unknown to date, leaving an important gap in knowledge that future research must address.

8. CONCLUSIONS

Findings from the current systematic review and meta‐analysis provide evidence that integrated interventions for maternal substance use have moderate effects on children's developmental outcomes. Moreover, our findings indicate a trend where longer treatments are associated with even greater outcomes. These results are encouraging in reducing addiction and associated psychosocial risks, as well as improving maternal mental health and parenting practices—critical factors directly shaping child welfare. Taken together, our results represent an important step in addressing this pressing public health concern, laying the groundwork for further evaluations and implementation of integrated interventions more broadly. As this work progresses, we anticipate positive changes in the lives of affected children, as well as benefits that extend well into society.

CONFLICT OF INTEREST STATEMENT

The authors have no conflict of interest to declare. Nicole Racine holds a Chair in Child and Youth Mental Health from the Faculty of Social Sciences at the University of Ottawa and the Children's Hospital of Eastern Ontario.

Supporting information

Supporting Information

Barriault, S. , Motz, M. , Firasta, L. , McDowell, H. , Labelle, P. R. , Poole, N. , & Racine, N. (2025). Effects of integrated programs for substance‐involved mothers on infant and child development outcomes: A systematic review and meta‐analysis. Infant Mental Health Journal, 46, 653–674. 10.1002/imhj.70029

DATA AVAILABILITY STATEMENT

All data used in the current review were obtained from published studies. The datasets used for analysis can be made available by the senior author upon request.