Psychosocial, pharmacological, and legal interventions for improving the psychosocial outcomes of children with substance misusing parents: A systematic review

Background Parental substance misuse is a pervasive risk factor for a range of detrimental outcomes for children across the life course. While a variety of interventions have been developed for this population, the existing evidence‐base requires consolidation and consideration of the comparative effectiveness of different interventions to facilitate evidence‐informed decisions between different intervention approaches. Objectives This review aimed to use network meta‐analysis to synthesise the comparative effectiveness of psychosocial, legal, and pharmacological interventions for improving outcomes for children with substance misusing parents. Network meta‐analysis was not possible; however, we synthesised the effects of a broad range of interventions on child psychosocial outcomes. Another aim was to examine potential moderators of the effects, yet this was also not possible due to data limitations. A secondary objective was to qualitatively synthesise economic, treatment completion, and treatment acceptability information for included studies. Search Methods Searches were performed in November 2020 and again in April 2021. Encompassing multiple disciplines, we searched 34 databases, 58 grey literature repositories, and 10 trial registers. Supplementary hand searches were conducted on 11 journals, along with harvesting the references of all included studies and existing reviews, and forward citation searching each report of all included studies. Study authors were contacted to obtain missing data. Selection Criteria Eligible studies included randomised and quasi‐experimental evaluations of psychosocial, pharmacological, and/or legal interventions using either a placebo, no treatment, waitlist control, treatment‐as‐usual, or alternative treatment as a comparison condition. Study participants needed to be comprised of families with children under the age of 18 with one or more currently substance‐misusing parents (or caregivers). Studies were required to evaluate the eligible intervention using a child‐focused psychosocial outcome. If reported in eligible studies, the following secondary outcomes were also synthesised in the review: cost‐effectiveness, treatment completion, length of time in treatment and acceptability of treatment (e.g., participant perspectives of the intervention). There were no restrictions placed on publication status or geographic location, however only research written in English was included. Data Collection and Analysis Standard methodological procedures were followed across all stages of the review, as guided by the published protocol for the review (Eggins et al., 2020). Due to the inability to conduct network meta‐analyses, random effects pairwise meta‐analyses with inverse variance were used to synthesise effects when two or more studies with conceptually similar interventions and outcomes were available. Results of the meta‐analyses are displayed in forest plots, and separate analyses are provided for conceptually distinct outcomes and time‐points of measurement. Sensitivity analyses are used to explore possible sources of heterogeneity in the absence of sufficient studies to conduct subgroup analyses. Main Results 99 studies (reported in 231 documents) met review inclusion criteria, encompassing 22,213 participants. Most studies were conducted in the United States (k = 76), almost half were randomised controlled trials (k = 46), and the most common comparator was treatment‐as‐usual (k = 50). Interventions were evaluated using a large range of child psychosocial outcomes which broadly fell under: (a) child welfare; (b) child development; (c) child emotional and behavioural; and (d) educational domains. Intervention models were rarely only legal or pharmacological in nature, with most studies evaluating integrated psychosocial treatments with either pharmacology, coordinated health care, case‐management, and/or judicial or child welfare oversight and coordination. Thirty‐six meta‐analyses and 227 single effect sizes were used to appraise the effectiveness of included interventions, based on 68 studies with sufficient data for effect size calculation. The size and direction of the effects varied across interventions, type of outcomes, and time‐point of measurement. Twenty‐seven meta‐analyses and 186 single effect sizes suggested null effects. Only five single effect size estimates based on three studies indicated negative effects including: higher depressive and somatisation symptoms (parent‐report), hopelessness (child‐report), educational achievement difficulties (parent‐report), and substantiated child protection reports for those engaged with interventions versus treatment‐as‐usual or no treatment. Nine meta‐analyses and 36 single study effect estimates suggest that psychosocial, pharmacological and/or legal interventions have a positive effect on a range of specific child welfare, developmental, and emotional/behavioural outcomes for children. The risk of bias varied across domains and studies, which further lowers confidence in the results. Based on a subset of included studies, treatment completion tends to vary, yet cost‐benefits can be achieved when intervening with children whose parents misuse substances. Authors' Conclusions Despite a large body of evaluation evidence, disparate outcomes, and missing data precluded analyses to formally examine the comparative effectiveness of psychosocial, legal, and pharmacological interventions for improving outcomes for children with substance misusing parents. The large amount of unreported (missing) data meant that many effect estimates were underpowered due to single studies and small sample sizes. The review findings suggest that interventions for families affected by parental substance misuse can be effective when they holistically address multiple domains such as parent wellbeing/mental health, parenting, children's wellbeing, and/or other factors impacting family wellbeing (e.g., housing).