Prevent Language Delays and Disorders Among Children of Caregivers With Substance Use Disorders With Community Groups and Individualized Services

The investigators will recruit and randomize a total of N=244 parent-child dyads to one of four treatment conditions using a 2x2 factorial design. Specifically, all participants will receive the group-based intervention from "Learn and Play Every Day" parent-implemented naturalistic communication intervention (PI-NCI) and will be randomized to receive the additional intervention modalities of (1) interaction review, (2) practical supports, (3) both, or (4) neither. The investigators will recruit participants from community agencies that provide support to adults with substance use disorders. Participants will be parents of a child between 18 and 42 months old and who have either a history of substance use disorder or active use within the last year (see Protection of Human Subjects for details). Interaction review and practical support intervention modalities will be conducted via telehealth. The investigators will randomize parents to intervention modality at the participant level through permuted blocks. The investigators will stratify randomization by the parent's primary language (English vs. Spanish) to ensure balance on this factor. Participant randomization will be masked to the interventionists delivering the group-based intervention. Interventionists delivering the interaction review and practical support modalities will not be aware of the presence or absence of the other intervention component. The investigators will collect information about parent responsivity and child language at baseline (T1, prior to receiving the intervention, immediately following the intervention (T2, 2 months), and at two follow-up assessments (T3, 4 months; T4, 6 months). Our primary analysis will compare pre-post gains in parent responsivity and child language comparisons between baseline and follow-up assessments. Parent-child interactions will be video recorded and then outcomes will be coded by trained staff, whereas child performance on the dynamic assessment task and parent performance on the functional cognitive-communication assessment will be scored in real-time by the data collector. Other self-reported measures such as demographics and substance use history will be collected through electronic survey software and stored on a secure server. The primary outcome of interest is parental responsivity, which is measured by the number of parental responses to intentional communicative acts from their child per minute. The secondary outcome of interest is child vocabulary (receptive and expressive). Additional outcomes include parenting knowledge and self-efficacy as well as secondary child outcomes such as rate of communication, sentence diversity, and language learning skill on the dynamic assessment task. Our primary interest (Aim 1a) is to compare the effectiveness of intervention modalities. The investigators will model changes in responsivity at two months post-baseline, as well as other outcomes, as a function of the intervention modalities and their interaction using generalized linear models adjusting for outcome values at baseline. Hypothesis tests will be conducted on the difference in outcome of each intervention modality from the mean of group means. Our overall judgment of the component will be primarily dependent on the component's effect on parental responsivity, but investigators will also take into account the effects on other parent and child outcomes. The investigators will conduct an intent-to-treat analysis, grouping participants by their treatment condition randomizations regardless of the fidelity of the intervention delivered. The effectiveness of an intervention can be undermined if it is difficult to administer or unacceptable to potential recipients. As a result, the investigators will collect feedback from parents, interventionists, and community partners through surveys and interviews. The questions posed will focus on evaluating the reach, acceptability, appropriateness, and fidelity of the intervention delivery (Aim 1b). The investigators will summarize and review the responses and use it to inform the interpretation of the results across Aim 1a and 1b. For example, if the interventions are found to be arduous to implement or unappealing to participants, this will be used to interpret the efficacy findings in context. In addition to assessing the efficacy and implementation of PI-NCI modalities, the investigators will assess differential treatment response and the mechanisms through which the intervention modalities function (Aim 2). The intervention modalities may not be equally effective across various psycho-social dimensions which may give opportunities for personalized treatment. Thus, investigators will assess the interactions (effect moderation) between the interventions and parental psychopathology, cognitive communicative impairment, social determinants of health on parental outcomes as well as prenatal exposure to substances on child outcomes. It is also important to understand how our interventions improve parental responsivity and other parent/child outcomes. Indeed, the investigators hypothesize that these interventions first improve parenting self-efficacy and knowledge, which in turn improves outcomes. Our analyses of mediators will use full information structural equation models to adjusted for baseline values and unbalanced covariates and the investigators will model moderators using similarly specified linear models. Our moderation analyses will assess the interaction terms between intervention components and our mediation analysis will examine the indirect effect of treatment modalities on outcomes through parenting efficacy and knowledge.