Introduction
Maternal Substance Use Disorder (SUD) constitutes a major public health problem that affects both individuals and their families, especially their young children (Messina et al., Reference Messina, Calhoun, Conner and Miller2015). Offspring of mothers with SUD are at higher risk to experience poor quality parent-child relationships, characterized by maladaptive parenting practices that may result in the extremes of abuse and neglect (Peleg-Oren & Teichman, Reference Peleg-Oren and Teichman2006), and by developmental, emotional and behavioral difficulties subsequent to substance consumption that occurs prenatally (Beeghly & Tronick, Reference Beeghly and Tronick1994; Behnke & Smith, Reference Behnke and Smith2013) or in the home (Spehr et al., Reference Smith, Cumming and Xeros-Constantinides2017). Hurdles in the parent-offspring relationship often begin during pregnancy, where drug use represents one of the first stages of maladaptive child care (Eyler & Behnke, Reference Eyler and Behnke1999; Hagan et al., Reference Hagan, Balachova, Bertrand, Chasnoff, Dang, Fernandez-Baca, Kable, Kosofsky, Senturias, Singh, Sloane, Weitzman and Zubler2016; van IJzendoorn et al., Reference Van IJzendoorn and Bakermans-Kranenburg2020). Prenatal substance use impedes central caregiving tasks, that include the promotion of the unborn baby’s welfare, reducing risky behaviors, and ensuring a safe prenatal journey for the expected child (Curet & Hsi, Reference Curet and Hsi2002; Rubin, Reference Rodning, Beckwith and Howard1976). Pregnant women with SUD often show poor maternal-fetal bonding (Mikhail et al., Reference Mikhail, Youchah, DeVore, Ho and Anyaegbunam1995), characterized by low cognitive awareness of fetal characteristics and by avoidance of required medical screenings (Mayes, Reference Mayes2002) as a result of the continuous consumption of substances (Shieh & Kravitz, Reference Shaffer, Obradovic and Obradović2002). At the same time, newborns that have been exposed to substances in utero are at higher risk to present unfavorable physical and autonomic characteristics at delivery, such as reduced growth, higher activity and irritability, low levels of arousal as well as disturbances in sleep and feeding patterns (Behnke & Smith, Reference Behnke and Smith2013; LaGasse, Reference LaGasse2003). Although it is still not clear to which extent they represent a direct effect of in utero exposure to a specific substance or rather a cumulative effect of diverse risk factors (e.g. consumption of multiple substances and quality of the child’s environment), these behavioral manifestations are likely to promote additional relational difficulties in the perinatal period, affecting the comprehensibility of infant signals and constituting an adjunct source of stress for parents, especially mothers (Frank et al., Reference Frank, Augustyn, Knight, Pell and Zuckerman2001; Liles et al., Reference Liles, Newman, LaGasse, Derauf, Shah, Smith, Arria, Huestis, Haning, Strauss, DellaGrotta, Dansereau, Neal and Lester2012; Messina et al., Reference Messina, Calhoun, Conner and Miller2015).
The literature on maternal SUD highlights that poor quality parent-child interactions are likely to persist during infancy (Frigerio et al., Reference Frigerio, Porreca, Simonelli and Nazzari2019; Tronick et al., Reference Thompson Gershoff2005), toddlerhood (Eiden et al., Reference Eiden, Schuetze, Colder and Veira2011), preschool (Salo et al., Reference Salloum and Brown2009), and school age, with challenges resulting from both parent and child characteristics (Tronick et al., Reference Thompson Gershoff2005). On the parent side, mothers with SUD tend to be less sensitive and responsive to children’s emotional signals both at a neural (Kim et al., Reference Kim, Iyengar, Mayes, Potenza, Rutherford and Strathearn2017; Lowell et al., Reference Lowell, Maupin, Landi, Potenza, Mayes and Rutherford2020; Rutherford et al., Reference Rutherford, Byrne, Crowley, Bornstein, Bridgett and Mayes2021) and a behavioral level (Eiden, Reference Eiden2001; Flykt et al., Reference Flykt, Punamäki, Belt, Biringen, Salo, Posa and Pajulo2012; Frigerio et al., Reference Frigerio, Porreca, Simonelli and Nazzari2019; Porreca et al., Reference Porreca, De Palo, Simonelli and Capra2016; Salo et al., Reference Salloum and Brown2009, Reference Salo, Kivistö, Korja, Biringen, Tupola, Kahila and Kivitie-Kallio2010). During interactive play they are more inclined to be directive and interfering with offspring activities (Bauman & Dougherty, Reference Bauman and Dougherty1983; Bauman & Levine, Reference Bauman and Levine1986; Rodning et al., Reference Robinson, Delgadillo and Kellett1991; Johnson et al., Reference Johnson, Morrow, Accornero, Xue, Anthony and Bandstra2002; Pajulo et al., Reference Pajulo, Savonlahti, Sourander, Ahlqvist, Helenius and Piha2001), whereas during disciplinary situations they are more prone to adopt either harsh limit-setting strategies or laxness (or both) in response to their child’s challenging behaviors (Miner & Clarke-Stewart, Reference Miner and Clarke-Stewart2008; Thompson Gershoff, Reference Teichner, Horner, Roitzsch, Herron and Thevos2000). Negative parenting behaviors have also been observed during feeding interactions, resulting in less engagement and flexibility (LaGasse, Reference LaGasse2003; Minnes et al., Reference Minnes, Singer, Arendt and Satayathum2005). Globally, these strategies tend to be ineffective and related to negative developmental sequelae in children, including insecure and disorganized attachments (Swanson et al., Reference Suchman, Pajulo, DeCoste and Mayes2000) as well as internalizing and externalizing problems (Bada et al., Reference Bada, Das, Bauer, Shankaran, Lester, LaGasse, Hammond, Wright and Higgins2007; D’Onofrio et al., Reference D’Onofrio, Van Hulle, Waldman, Rodgers, Rathouz and Lahey2007; Misri et al., Reference Misri, Reebye, Kendrick, Carter, Ryan, Grunau and Oberlander2006). In their extreme forms, parental maladaptive practices can assume the shape of severe abuse or neglect, where the parent fails to take care of the child’s basic needs (Barnard & McKeganey, Reference Barnard and McKeganey2004), increasing the risk of involvement with child welfare services, and the subsequent child removal from home (Dakof et al., Reference Dakof, Cohen, Henderson, Duarte, Boustani, Blackburn, Venzer and Hawes2010; Matos et al., Reference Matos, Moleiro and Dias2014).
On the child’s side, the persistence of the neurodevelopmental and emotional characteristics associated with prenatal exposure are likely to further exacerbate poor quality parent-child interactions (Coles & Platzman, Reference Coles and Platzman1993; Eiden et al., Reference Eiden, Schuetze and Coles2011; Mattson et al., Reference Mattson, Schoenfeld and Riley2001; O’Connor et al., Reference O’Connor, Kogan and Findlay2002), undermining the readability of child signals and, consequently, adequate parental responses (Behnke & Smith, Reference Behnke and Smith2013; O’Connor et al., Reference O’Connor, Kogan and Findlay2002). Moreover, children of parents with SUD are at increased risk of cognitive, linguistic, and physical delays (Behnke & Smith, Reference Behnke and Smith2013; Lewis et al., Reference Lewis, Singer, Short, Minnes, Arendt, Weishampel, Klein and Min2004), as well as the manifestations of emotional (e.g. social withdrawal, depression, and anxiety), or behavioral problems (e.g. aggression, delinquency, and substance use) throughout childhood and adolescence (Bada et al., Reference Bada, Das, Bauer, Shankaran, Lester, LaGasse, Hammond, Wright and Higgins2007; Kilpatrick et al., Reference Kilpatrick, Acierno, Saunders, Resnick, Best and Schnurr2000; Minnes et al., Reference Minnes, Singer, Min, Wu, Lang and Yoon2014). As a consequence of negative transactional patterns resulting from the encounter between parents’ and children’s difficulties, offspring of caregivers with SUD are at higher risk of being exposed to low quality attachment relationships across different situations and ages (Frigerio et al., Reference Frigerio, Porreca, Simonelli and Nazzari2019; LaGasse, Reference LaGasse2003; Thompson Gershoff, Reference Teichner, Horner, Roitzsch, Herron and Thevos2000), which could play a role in the intergenerational transmission of substance use, increasing offspring’s susceptibility to addiction later in life (Strathearn et al., Reference Strantz and Welch2019).
Information about the “normative” development of mother-child relationships over time in the context of SUD is almost absent, with the majority of extant work presenting cross-sectional data or data on treatment effects. A previous longitudinal study showed that, without the support of appropriate and focused interventions, difficulties in mother-child exchanges tend to persist over the course of the first years of life (Blackwell et al., Reference Blackwell, Lockman and Kaiser1999). Specifically, only children, but not parents, seemed to show gains in their emotional competencies, albeit their scores remained low compared to low-risk and other high-risk groups (Blackwell et al., Reference Blackwell, Lockman and Kaiser1999).
Nevertheless, a growing body of literature points out that, when appropriately addressed through targeted interventions, quality of parent-child relationships in groups with SUD can be effectively improved (Belt et al., Reference Belt, Flykt, Punamäki, Pajulo, Posa and Tamminen2012; Suchman et al., Reference Suchman, Decoste, Mcmahon, Rounsaville and Mayes2012). Early and preventive interventions are identified as a window for providing effective treatment in the condition of SUD, improving parent-child relationship quality and preventing infants’ physical and emotional problems (Belt et al., Reference Belt, Flykt, Punamäki, Pajulo, Posa and Tamminen2012; Howell et al., Reference Howell, Heiser and Harrington1999; Niccols et al., Reference Niccols, Milligan, Smith, Sword, Thabane and Henderson2012; Suchman et al., Reference Suchman, Decoste, McMahon, Dalton, Mayes and Borelli2010). For example, psychodynamic group therapy was found effective in reducing hostile and intrusive maternal behaviors towards children (Belt et al., Reference Belt, Flykt, Punamäki, Pajulo, Posa and Tamminen2012), whereas attachment and mentalization-based interventions enhancing maternal mentalizing abilities supported the improvement of relationship quality (Pajulo et al., Reference Pajulo, Pyykkönen, Kalland, Sinkkonen, Helenius, Punamäki and Suchman2012; Suchman et al., Reference Suchman, DeCoste, Castiglioni, Legow and Mayes2018; Suchman et al., Reference Suchman, DeCoste, Borelli and McMahon2010). These interventions were administered both in outpatient (e.g. Suchman et al., Reference Suchman, DeCoste, Castiglioni, Legow and Mayes2018) or residential treatment (e.g. Berlin et al., Reference Berlin, Shanahan and Appleyard Carmody2014), two of the most commonly adopted settings for the treatment of women with substance use disorder (Daley et al., Reference Daley, Argeriou, McCarty, Callahan, Shepard and Williams2000). The results of these studies suggest the persistence of positive outcomes over time (Smith et al., Reference Shieh and Kravitz2010; Suchman et al., Reference Suchman, Decoste, McMahon, Dalton, Mayes and Borelli2010).
Given the complexity of SUD, various authors point to the need to identify mechanisms that support or hamper intervention effects, especially in the context of parenting (Belt et al., Reference Belt, Flykt, Punamäki, Pajulo, Posa and Tamminen2012; Bosk et al., Reference Bosk, Paris, Hanson, Ruisard and Suchman2019; Suchman et al., Reference Suchman, Decoste, Mcmahon, Rounsaville and Mayes2012). For example, the presence of additional symptoms of psychopathology co-occurring with substance use could constitute an adjunct risk factor for caregiving practices, further exacerbating difficulties in parental responses and quality of parent-child relationships (De Palo et al., Reference De Palo, Capra, Simonelli, Salcuni and Di Riso2014; Dixon et al., Reference Dixon, Lee, Gratz and Tull2018; Hans, Reference Hans1999; Porreca et al., Reference Porreca, De Carli, Filippi, Parolin and Simonelli2020; Salloum & Brown, Reference Rutherford, Kim, Yip, Potenza, Mayes and Strathearn2017; Yilmaz & Dilbaz, Reference Williams and Roberts2016; Zuckerman & Brown, Reference Zahn-Waxler, Duggal, Gruber and Bornstein1993). The comorbidity of substance use and severe psychological distress measured at treatment intake has been extensively reported to have a negative impact on program response, with higher risks of poor outcomes and dropouts in adults with SUD with (Haller et al., Reference Haller, Miles and Dawson2002; Haller & Miles, Reference Haller and Miles2008; Howell et al., Reference Howell, Heiser and Harrington1999) and without children (Herbeck et al., Reference Herbeck, Fitek, Svikis, Montoya, Marcus and West2005). This association was found for several psychiatric disorders (Elmquist et al., Reference Elmquist, Shorey, Anderson and Stuart2016; Herbeck et al., Reference Herbeck, Fitek, Svikis, Montoya, Marcus and West2005; Petry & Bickel, Reference Petry and Bickel2000; Strantz & Welch, Reference Stocco, Simonelli, Capra and De Palo2012; Williams & Roberts, Reference Wickham2009; Zweben et al., Reference Zuckerman, Brown and Zeanah2012) and dysfunctional personality traits (Fernandez-Montalvo & López-Goñi, Reference Fernandez-Montalvo and López-Goñi2010). Although co-morbid psychological distress tend to generally decrease during and after program completion, both in residential (Schinka et al., Reference Schiepek, Eckert and Weihrauch1999) and outpatient services (Suchman et al., Reference Strathearn, Mertens, Mayes, Rutherford, Rajhans, Xu, Potenza and Kim2008, Reference Suchman, DeCoste, Borelli and McMahon2018), the effect of its changes may vary according to the specific type of treatment and to the outcomes considered. For example, some studies reported that decreases in depressive symptoms were specifically related to improvements in maternal caregiving behaviors (Suchman et al., Reference Suchman, Decoste, Mcmahon, Rounsaville and Mayes2012), yet other work highlighted that mothers with SUD presenting more severe clinical profiles tended to benefit more from treatment in terms of reflective functioning although they displayed a slower alleviation in their symptoms over time (Suchman et al., Reference Suchman, DeCoste, Castiglioni, McMahon, Rounsaville and Mayes2017).
Other studies examined the role of cognitive functioning, given the strong associations between chronic use of drugs and structural and functional brain changes, especially in the prefrontal area (Ceceli et al., Reference Ceceli, Bradberry and Goldstein2021; Goldstein & Volkow, Reference Goldstein and Volkow2011). Impairments in executive functioning have been linked to treatment abandonment in adults with SUD (Brorson et al., Reference Brorson, Ajo Arnevik, Rand-Hendriksen and Duckert2013; Teichner et al., Reference Taylor Tavares, Clark, Cannon, Erickson, Drevets and Sahakian2002), distinguishing between completers and dropouts (Aharonovich et al., Reference Aharonovich, Hasin, Brooks, Liu, Bisaga and Nunes2006). Moreover, cognitive functioning significantly improved in individuals completing treatment (Vergara-Moragues et al., Reference van IJzendoorn, Bakermans-Kranenburg, Coughlan and Reijman2017). Additionally, cognitive functioning presents relevant links with psychopathology, being negatively associated with symptoms of depression (Taylor Tavares et al., Reference Swendsen2007), obsessive-compulsive disorders (Penadés et al., Reference Penadés, Catalán, Rubia, Andrés, Salamero and Gastó2007), and psychotic disorders (MacDonald et al., Reference MacDonald, Carter, Kerns, Ursu, Barch, Holmes, Stenger and Cohen2005). Notably, recent research provided evidence for the role of parental cognitive functioning in adequate caregiving practices, supporting the development of favorable parent-child relationships (Bridgett et al., Reference Bridgett, Kanya, Rutherford and Mayes2017; Martin et al., Reference Martin, Bridgett, Mayes and Rutherford2020; Rutherford et al., Reference Rubin2018). Cognitive resources are often impaired in parents with SUD, contributing to poor quality parent-child relationships (Porreca et al., Reference Porreca, Biringen, Parolin, Saunders, Ballarotto and Simonelli2018). Thus, in the condition of SUD psychopathology and cognitive impairments might coexist, with the risk to aggravate parenting quality and negatively affect the course of treatment.
A better comprehension of the mechanisms underlying maladaptive parenting practices and affecting treatments provided to parents with SUD is necessary, in order to be able to timely intervene and to mitigate the negative impact of the side effects of substance use on child developmental outcomes. Although some studies focused on the impact of psychopathology (Herbeck et al., Reference Herbeck, Fitek, Svikis, Montoya, Marcus and West2005) and cognitive functioning (Brorson et al., Reference Brorson, Ajo Arnevik, Rand-Hendriksen and Duckert2013; Teichner et al., Reference Taylor Tavares, Clark, Cannon, Erickson, Drevets and Sahakian2002) in SUD treatment, little is known about these mechanisms in the context of parenting. Some studies provided evidence for the role of psychopathological symptoms in interventions provided to mothers with SUD (Haller & Miles, Reference Haller and Miles2008; Howell et al., Reference Howell, Heiser and Harrington1999; Suchman et al., Reference Suchman, Decoste, Mcmahon, Rounsaville and Mayes2012). As far as we know, no study considered the role of cognitive functioning in intervention effects on mothers with SUD. This aspect could be of particular importance given that impaired cognitive processes could reduce the resources dedicated to both parental demands and treatment requirements.
To fill this gap in the literature, the present study aims to investigate the role of psychopathological distress (Objective 1) and of cognitive functioning (Objective 2) measured at treatment intake on the quality of the parent-child relationship during a residential treatment program for mothers with SUD and their children, a setting usually characterized by structured daily care and a focus on intensive recovery activities (Reif et al., 2014). Specifically, we conducted a naturalistic study of longitudinal outcomes, with observation of the degree and direction of change in parent-child relationship over time (Caruana et al., Reference Caruana, Roman, Hernández-Sánchez and Solli2015). Naturalistic studies are often carried out in clinical settings in order to study prediction of treatment outcomes and trajectories in complex contexts where more structured research protocols (e.g. RCTs) might not yet be feasible (Philips & Falkenström, Reference Philips and Falkenström2021). Naturalistic longitudinal studies can provide important hypotheses about treatment processes and are characterized by high ecological validity (Borckardt et al., Reference Borckardt, Nash, Murphy, Moore, Shaw and O’Neil2008; Leichsenring, Reference Leichsenring2004). In line with previous studies (Brorson et al., Reference Brorson, Ajo Arnevik, Rand-Hendriksen and Duckert2013; Suchman et al., Reference Suchman, Decoste, Mcmahon, Rounsaville and Mayes2012; Teichner et al., Reference Taylor Tavares, Clark, Cannon, Erickson, Drevets and Sahakian2002), we expected to find significant associations between parental psychopathological distress (H1), cognitive functioning (H2) and quality of the parent-child relationship at admission and with changes in the parent-child relationship over time. Specifically, we expected greater psychopathological distress and lower cognitive functioning measured at intake to be associated with lower parent-child relationship quality at the beginning of treatment, and more modest improvements in parent-child relationship quality over time. Moreover, we expected to find a significant interaction between parental psychopathological distress and cognitive functioning in predicting the quality of the parent-child relationship (H3), with mothers with higher psychopathology and lower cognitive functioning exhibiting worse parent-child relationship outcomes during treatment.
Methods
Participants and procedure
The study involved 60 mothers with SUD (mean age = 30.13 years; SD = 6.79) and their children (48% male, mean age = 17.17 months; SD = 23.60) attending a residential treatment for substance use. Inclusion criteria were: Italian speaking and reading, recent enrollment in the residential setting (within 1–2 weeks), presence of SUD, with caregiving responsibilities for at least 6 days per week of a child aged between 0 and 7 years old and co-residing with the child in the facility. Mothers with more than one child chose with which child they took part in the study. Exclusion criteria included diagnosed psychotic disorders in an active phase and significant cognitive impairment due to physical problems in the mother that prevented the execution of the tasks, or significant developmental delays or neurodiverse development in the child. The diagnosis of SUD was based on the mothers’ medical history and urine toxicology. Sample characteristics were collected through a structured interview. They are presented in Table 1. Most of the participants presented an early onset of drug use (M = 18.22; SD = 3.34) and a pattern of poly-drug use (86%), with heroin as primary substance of use (73%). Moreover, with respect to past events in the family of origin, more than a half of the participants reported having experienced the death of a significant person (60%) and some or multiple forms of maltreatment (68%) between childhood and early adolescence. Most of the children were exposed to substances in utero (83%). In most cases (78%), the mothers’ entrance to the facility was mandatory, after a Juvenile Court Decree. The participants were enrolled in the study after their admission to the facility. A research assistant met with all eligible participants individually and presented the project to them. Mothers interested in taking part in the study (92% of the eligible patients entering the treatment) received information and gave their written informed consent. Information on socio-demographic and clinical variables, maternal cognitive functioning, and psychopathological symptoms was collected at enrollment. Parent-child relationship quality was measured through observation of parent-child interactions at enrollment and every 3 months during the following 15 months of treatment. Figure 1 displays the phases of the study.
Table 1. Demographic characteristics of the sample
Figure 1. Phases of the study. AR = attrition rate since admission.
The residential treatment
The facility offers a psychodynamic-oriented comprehensive program which includes pharmacological, educational and psychological intervention, and parent counseling as described in Stocco et al. (Reference Stocco, Simonelli, Capra and De Palo2012). Educational intervention is provided on a daily basis by psychologists and professional educators and includes practical strategies, guidance and supervision targeted at: personal wellbeing (eating, hygiene, organization of spaces), handling of the SUD condition (abstinence, prevention of relapses, administration of substitutive or psychological treatment), management of personal relationships (with partner, friends, relatives and the family of origin), parenting (feeding, cleaning, changing, clothing), and adherence to residential tasks and guidelines (e.g. preparation of meals, cleaning of spaces). Psychological treatment is provided through individual psychotherapy and group psychotherapy. Both these treatments foresee one session per week and are carried out by expert clinicians trained in psychodynamic psychotherapy. The interventions are characterized by a focus on the clinical relationship with the aim of modifying the clients’ internal representations and defenses and promoting adaptive coping strategies. Parent counseling sessions were available on request. These sessions were carried out individually, without the child, by an expert clinician trained in psychodynamic psychotherapy, who used different approaches based on the client’s needs, ranging from supporting the mother in the achievement of the objectives of the educational intervention to approaches addressing mothers’ mental representations of themselves (both as individual and as parent) and of the relationship with their significant others. All the participants received at least 1 session of parental counseling, up to 3 consecutive sessions in case of particularly severe difficulties. The scheduling of the sessions differed according to the issues treated and in line with the mothers’ requests (from weekly to monthly sessions). The sessions could be requested at any point during the residential treatment. Consultations with psychiatrists and pediatricians external to the facility are scheduled when needed. The different types of intervention are coordinated in weekly staff meetings monitoring treatment of each mother-child pair.
Measures
Parent-child relationship quality
Mothers and children were videotaped during 15-minute free-play episodes that were subsequently assessed by trained coders. The procedure was carried out at admission and repeated every 3 months, during the first 15 months of treatment, for a total of 6 measurements. Quality of parent-child relationship was assessed coding mother-child interactions using the Emotional Availability Scales — 4th ed. (EAS; Biringen, Reference Biringen2008). The coding system includes six scales assessing both parenting behaviors (sensitivity, structuring, non-intrusiveness, and non-hostility) and child interactive behaviors (responsiveness and involvement), providing a multifaceted measure of parent-child interaction. Each scale is rated on a 7-point Likert scale, with higher scores referring to more optimal behaviors. Scores between 5.5 and 7 indicate the presence of proper interactive features in the individual. Scores around 4 indicate inconsistency (i.e., behaviors that are appropriate in some way but that are not fully optimal). Scores of 3 or below point out less optimal interactions. Scores lower than 2 indicate low qualities on that dimension. For the present study the videos were coded by two independent raters trained in the system. To minimize the risks of halo effects and of biases linked to the child’s developing age, the videos were assigned to the coders randomizing dyads and time points of observation. Moreover, the raters were kept blind with respect to the objectives of the study and to the participants’ clinical and treatment condition. Inter-rater reliability was calculated using Intraclass Correlation Coefficients on a randomly selected subsample of 20% of the cases, with values ranging from 0.80 to 0.95. For the purposes of the present study, a composite measure (i.e. Parent-child interaction quality) including both parent and child scale scores was computed as an index of the global quality of parent-child relationship. Analysis yielded good internal consistency of the composite measure (α = .80).
Maternal psychopathological symptoms
Maternal psychopathological symptoms were measured at admission through the Symptom Checklist-90 Revised (SCL-90-R; Derogatis, Reference Derogatis1994; Sarno et al., Reference Salo, Politi, Tupola, Biringen, Kalland, Halmesmäki, Kahila and Kivitie-Kallio2011), a self-report questionnaire aimed at assessing the presence of psychopathological symptoms on nine primary symptom dimensions (somatization, obsessive-compulsive, interpersonal sensitivity, depression, anxiety, hostility, phobic anxiety, paranoid ideation, and psychoticism) and psychological distress according to three global distress indexes (Global Severity Index, Positive Symptom Total, and Positive Symptom Distress Index). Each item is rated on a 5-point Likert scale, ranging from 0 (not at all) to 4 (extremely). Raw scores are converted into T-scores and compared to normative values being classified as above/below clinical range. The SCL-90 is normed for both clinical and nonclinical populations. The Italian translation and adaptation of the SCL-90-R (Sarno et al., Reference Salo, Politi, Tupola, Biringen, Kalland, Halmesmäki, Kahila and Kivitie-Kallio2011) has been shown to have good internal consistency (α = 0.68–0.97) in adolescents and adults.
For the purposes of the present study, we considered only the Global Severity Index as a measure of overall parental psychopathological distress (α = 0.97).
Maternal cognitive functioning
Cognitive functioning was investigated at admission through the Esame Neuropsicologico Breve-2 (ENB-2; Mondini et al., Reference Mondini, Mapelli, Vestri and Bisiacchi2003, Reference Mondini, Mapelli, Vestri, Arcara and Bisiacchi2011), a comprehensive neuropsychological battery standardized for the Italian population. The battery includes 16 subtests: digit span, immediate and delayed recall prose memory, interference memory at 10 and 30 s, trail making test part A and B, token test, word phonemic fluency, abstract reasoning, cognitive estimation, overlapping figures, spontaneous drawing, copy drawing, clock drawing, and ideomotor praxis test. These subtests cover the cognitive domains of attention, memory, comprehension, executive functioning, perception, and praxis abilities. The scoring system yields a score for each subtest and a total score, the Global Cognitive Index, ranging from 0 to 100. The scores can be classified as average, borderline, and below average, according to established norms weighted for different age-ranges (M = 83.51, SD = 6.52 for 15–20-year-olds, M = 85.65, SD = 5.40 for 21–30-year-olds, M = 84.91, SD = 5.66 for 31–40-year-olds, M = 80.72, SD = 6.98 for41–50-years-olds). The battery has good psychometric characteristics, good differential validity in discriminating normative and clinical groups, and sufficient test-retest reliability (range from 0.57 to 0.97) (Mondini et al., Reference Mondini, Mapelli, Vestri and Bisiacchi2003, Reference Mondini, Mapelli, Vestri, Arcara and Bisiacchi2011). Moreover, the battery has already been successfully applied to adults with SUD (Parolin et al., Reference Parolin, Simonelli, Mapelli, Sacco and Cristofalo2016; Porreca et al., Reference Porreca, Biringen, Parolin, Saunders, Ballarotto and Simonelli2018). For the purposes of the present study, we considered only the Global Cognitive Index that summarizes the individual’s cognitive functioning.
Plan of analysis
As a preliminarily analysis, we performed a confirmatory factor analysis of the EAS scales at t1 and tested for measurement invariance in the factorial structure across the 6 assessments in order to check the hypothesized one factor solution. Then, in order to achieve the most unbiased analysis possible, we imputed the missing data in the EAS scores, taking into account the longitudinal nature of the data with observations nested in participants. We imputed the data producing 5 different datasets, and after performing the analysis in each of this dataset, we pooled the results (Honaker et al., Reference Honaker, King and Blackwell2011). Next, we implemented a model comparison approach to determine the best predictive mixed linear models of the EAS scores across time. In order to find the best fitting model, we started with a model including only the linear effect of time as predictor and then added step by step the quadratic effect of time, the main effects of global cognitive functioning, psychopathological symptoms, as well as the interaction between these parental factors. The choice of the best model was based on a procedure of model selection depending on AIC scores and AIC weights (Wagenmakers & Farrell, Reference Vergara-Moragues, Verdejo-García, Lozano, Santiago-Ramajo, González-Saiz, Betanzos Espinosa and Pérez García2004). The best model was repeated with non-imputed data as a sensitivity analysis, for which the results are presented in the Supplemental Materials. Concerning the variable participation to parental counseling, given the heterogeneity of the treatment provided due to the different number of sessions, different scheduling and different contents, all mothers were considered as a group in the sets of analyses. This is in line with dose-response studies on psychological treatment suggesting that at least 4 consecutive sessions are needed to observe clinically significant differences in improvements between groups (Robinson et al., Reference Reif, George, Braude, Dougherty, Daniels, Ghose and Delphin-Rittmon2020). All analyses were performed in R (R Development Core Team, 2021), using the “Amelia” package (Honaker et al., Reference Honaker, King and Blackwell2011) to impute missing data, the “lme4” package (Bates et al., Reference Bates, Mächler, Bolker and Walker2015) to perform the mixed models and the “effects” (Fox, Reference Fox2003) and “ggplot2” (Wickham, Reference Wetterling and Junghanns2009) packages to produce the plot. R-codes are available upon request.
Results
In Table 2 we present the descriptive analyses of the variables of the study.
Table 2. Descriptive analyses of the variables of the study
Means, standard deviations, and correlations of the variables of the study at admission. N = xx. * indicates p < .05. ** indicates p < .01.
A confirmatory factor analysis on the EAS scores at t1 was performed in order to test the one factor factorial structure hypothesized (Bornstein et al., Reference Bornstein, Suwalsky and Breakstone2012; Garvin et al., Reference Garvin, Tarullo, Van Ryzin and Gunnar2012). Results yielded a satisfactory solution (X 2(5) = 9.12, p = 0.10, CFI = 0.98, TLI = 0.94, RMSEA = 0.12 [0–0.24]) and requirements indicating weak measurement invariance (Δχ 2 = 121.17, df = 25, p = .15) were met for the six assessments. Therefore, we considered the one factor solution confirmed and the EAS scales were averaged into one score (i.e. Parent-child interaction quality). The description of the CFA model is presented in Figure S1 in the supplemental materials. Table 2 presents means and standard deviations of the variables of the study at admission (t1) and the bivariate correlations. In order to determine the best model to explain the variability in EAS scores, we provide AIC and AIC weights for each hypothesized model in Table 3.
Table 3. Models selection
The best fitting model was the last one, including both cognitive functioning and symptomatology, as well as the interactions between these factors and the quadratic and linear effects of time. The parameters of the selected model are presented in Table 4, showing a significant negative main effect of maternal self-reported psychopathological symptoms, and significant interactions between both the linear and quadratic effects of time and cognitive functioning. Notably, no interaction between time and symptomatology reached the statistical significance threshold. Table S1 in the supplemental materials shows the parameters of the selected model fitted on participants with complete data without imputation of the missing data, as a sensitivity analysis. Results remain essentially unaltered.
Table 4. Parameters of the selected model predicting parent-child interaction quality
In order to explore the interpretation of the moderating role of cognitive functioning on the quadratic effect of time, a simple slope analysis was implemented. Results show that for mothers with low levels of cognitive functioning (i.e., 1 SD below the mean), the quadratic effect of time was significant and negative (b = –0.95, SE = 0.47, t = –2.02, p = 04), while for mothers with high levels of cognitive functioning the quadratic effect of time was significant and positive (b = 1.09, SE = 0.46, t = 2.37, p = .02). The graphic representation of the quadratic effects for high and low levels of executive functioning is presented in Figure 2. For sake of completeness, we tested also a model with only linear effects of time (i.e., no quadratic main effect of time and no interactions between the quadratic terms and cognitive functioning and psychopathology). Results show a non-significant (even if close to the significance threshold, i.e., p = .07) interaction between the linear effect of time and cognitive functioning. The parameters of this model can be found in Table S2 in the supplemental materials.
Figure 2. Simple slope analysis of the moderating role of cognitive functioning on the association between time and parent-child interaction quality. The blue slope refers to parent-child relationship quality in the group of mothers with high cognitive functioning whereas the red slope indicates the trajectory in mothers with low cognitive functioning.
Discussion
The aim of the present study, using a naturalistic correlational design, was to investigate the role of psychopathological symptoms and cognitive functioning in the development of the quality of the parent-child relationship during a residential treatment program for mothers with SUD. In line with our expectations, we found a negative association between maternal self-reported psychopathological distress and observed parent-child interaction quality, suggesting that higher distress at admission in the facility was associated with poorer caregiving and relational quality with the offspring. Previous research highlighted that the co-occurring presence of psychopathological symptoms with the condition of substance use (Dixon et al., Reference Dixon, Lee, Gratz and Tull2018; Salloum & Brown, Reference Rutherford, Kim, Yip, Potenza, Mayes and Strathearn2017; Swendsen, Reference Swanson, Beckwith and Howard2000; Yilmaz & Dilbaz, Reference Williams and Roberts2016) represents a risk factor for poor caregiving practices (De Palo et al., Reference De Palo, Capra, Simonelli, Salcuni and Di Riso2014; Hans, Reference Hans1999; Porreca et al., Reference Porreca, De Carli, Filippi, Parolin and Simonelli2020; Zuckerman & Brown, Reference Zahn-Waxler, Duggal, Gruber and Bornstein1993). Attunement to child signals is reduced both by the parents’ efforts to attend psychological needs subsequent to psychopathology (Berg-Nielsen et al., Reference Berg-Nielsen, Vikan and Dahl2000; Zahn-Waxler et al., Reference Yilmaz and Dilbaz2002), and to manage psychophysical needs related to substance dependence, consumption, and abstinence (Rutherford & Mayes, Reference Rutherford and Mayes2017).
Furthermore, previous research provided evidence that the co-occurrence of substance use and severe psychological distress measured at intake represents a negative prognostic factor for treatment response, with a higher risk of poor outcomes and treatment dropout (Haller et al., Reference Haller, Miles and Dawson2002; Herbeck et al., Reference Herbeck, Fitek, Svikis, Montoya, Marcus and West2005; Howell et al., Reference Howell, Heiser and Harrington1999). Notably, self-reported psychopathological symptoms at admission to the facility did not allow us to distinguish among different trajectories within our group of participants, suggesting a need to further investigate the impact of psychopathology on treatment outcome. The severity of symptoms might have changed during the 15 months of observation and variability in change may have prevented the identification of diverging pathways of parent-child interaction quality. This possibility is in line with data suggesting that much of the psychopathological distress experienced by individuals with SUD could be linked to withdrawal symptoms and is likely to decrease after a period of 2–3 weeks, even without specific treatment (Wetterling & Junghanns, Reference Wetterling and Junghanns2000). Moreover, this assumption is also consistent with studies indicating that integrated treatment programs providing parenting support can decrease maternal distress with different impact according to the type of treatments and to the outcome variable (Huebner, Reference Huebner2002; Lowell et al., Reference Lowell, DeCoste, Dalton, Dias, Borelli, Martino, McMahon and Suchman2023; Smith et al., Reference Shieh and Kravitz2010; Suchman et al., Reference Suchman, Decoste, McMahon, Dalton, Mayes and Borelli2010; Suchman et al., Reference Suchman, Decoste, Mcmahon, Rounsaville and Mayes2012).
In accordance with our second hypothesis, mothers with higher cognitive functioning displayed significant increase in quality of mother-child relationship, whereas mothers with lower cognitive resources presented a poorer treatment course. Parental cognitive resources have been found to be important indexes of appropriate parenting (Deater-Deckard et al., Reference Deater-Deckard, Wang, Chen and Bell2012), allowing caregivers to organize emotions, thoughts, and attitudes into integrated plans of action, both in clinical and nonclinical samples (Dix, Reference Dix2000; Shaffer et al., Reference Schinka, Hughes, Coletti, Hamilton, Renard, Urmann and Neri2017). The results of the present study suggest that cognitive functioning measured at intake differentiates not only parent-child relationship quality at admission but also its trajectory over time in mothers with SUD. In line with this finding, a growing body of research has identified cognitive functioning as an important prognostic factor for the treatment of substance use, with individuals with lower executive functioning being at higher risk to relapse or to abandon treatment (Aharonovich et al., Reference Aharonovich, Hasin, Brooks, Liu, Bisaga and Nunes2006; Brorson et al., Reference Brorson, Ajo Arnevik, Rand-Hendriksen and Duckert2013; Teichner et al., Reference Taylor Tavares, Clark, Cannon, Erickson, Drevets and Sahakian2002).
Notably, for both groups we found a quadratic effect of time, suggesting that change might be nonlinear and discontinuous in this clinical population (Hayes et al., Reference Hayes, Feldman, Beevers, Laurenceau, Cardaciotto and Lewis-Smith2007). Several authors suggested that the process of change in psychotherapy is characterized by destabilization before change (Hager, Reference Hager1992; Hayes et al., Reference Hayes, Feldman, Beevers, Laurenceau, Cardaciotto and Lewis-Smith2007; Hayes & Strauss, Reference Hayes and Strauss1998). Thus, before providing a stable environment and increasing patients’ readiness and resources for change, therapy may introduce a variety of experiences that interrupt, challenge, and destabilize old patterns (Hayes et al., Reference Hayes, Feldman, Beevers, Laurenceau, Cardaciotto and Lewis-Smith2007). From this perspective, treatment course could be characterized by discontinuous transitions, where significant changes are preceded by critical fluctuations and instabilities (Schiepek et al., Reference Sarno, Preti, Prunas and Madeddu2003). Specifically, in our group, mothers with higher cognitive functioning presented first a decrease in quality of parent-child interactions, followed by a gradual improvement, whereas mothers with low cognitive resources showed the opposite pattern, characterized by a decrease in relationship quality after an initial improvement.
We hypothesize that cognitive functioning in mothers with SUD might intervene as moderating factor for better or poorer treatment outcome. Of course, this moderator interacts with the specific treatment that in our case was characterized by a psychodynamic-oriented comprehensive approach, composed of both psychotherapeutic and educational sessions mostly focused on individual adult functioning (e.g. substance use, psychological disease, adjustment, and traumatic experiences), and to a lesser extent on quality of parenting. In the present study, dyads whose mothers presented higher cognitive resources seemed to benefit more from the treatment offered by the residential community. Anyway, the process of change appeared to be slow and results required a rather long time to become observable (Greenson, Reference Greenson2018; Knekt et al., Reference Knekt, Lindfors, Laaksonen, Renlund, Haaramo, Härkänen and Virtala2011), suggesting the need to integrate the treatments provided with modules specifically focused on parent-child relationship which could be more rapid and effective in supporting quality of care (Lowell et al., Reference Lowell, DeCoste, Dalton, Dias, Borelli, Martino, McMahon and Suchman2023; Mayes, Reference Mayes2023; Suchman et al., Reference Suchman, Decoste, Rosenberger and McMahon2004, Reference Suchman, Mayes, Conti, Slade and Rounsaville2006).
Mothers with lower cognitive resources presented fewer benefits during this type of intervention, showing an asymptotic curve of growth that quickly reached its upper limit and arrested during the course of treatment. This result is in line with studies indicating that individuals with SUD exhibiting low executive functioning benefit less from complex treatments, due to difficulties in learning and retaining new material (Aharonovich et al., Reference Aharonovich, Hasin, Brooks, Liu, Bisaga and Nunes2006; Alterman et al., Reference Alterman, Bridges and Tarter1986), and in the acquisition of different coping behaviors (McCrady & Smith, Reference McCrady and Smith1986). Notably, mothers with lower cognitive functioning presented an initial increase in quality of the parent-child relationship, suggesting that shorter interventions repeated over time could be more effective in targeting quality of parenting in this population. This hypothesis seems in line with meta-analytic work providing evidence that families with multiple problems benefit more from interventions structured in a moderate number of sessions and with easily accessible themes (Bakermans-Kranenburg et al., Reference Bakermans-Kranenburg, Van IJzendoorn and Juffer2003; Juffer et al., Reference Juffer, Bakermans-Kranenburg and van IJzendoorn2017).
Finally, concerning our third hypothesis, we did not find an interaction between adult psychopathology and cognitive functioning in shaping the development of the parent-child relationship. This finding could be related to low statistical power linked to the small sample of participants in our study and should be further investigated in the future (Nosek et al., Reference Nosek, Hardwicke, Moshontz, Allard, Corker, Dreber, Fidler, Hilgard, Kline Struhl, Nuijten, Rohrer, Romero, Scheel, Scherer, Schönbrodt and Vazire2022; van IJzendoorn & Bakermans-Kranenburg, Reference Tronick, Messinger, Weinberg, Lester, LaGasse, Seifer, Bauer, Shankaran, Bada, Wright, Poole and Liu2021).
Taken together, the results of the present study suggest that both parental psychopathology and cognitive functioning at intake may be linked to the ongoing of parent-child relationship quality in the context of maternal SUD, although with different trajectories. While psychopathological symptoms did not differentiate amongst diverse patterns of change during treatment, the quality of cognitive functioning discriminates between different trajectories of change.
The current study has some limitations. The first limitation concerns the small sample with limited statistical power for multivariate analyses such as the ones conducted in the current exploratory study. The inclusion of a larger group could provide more valid information on the associations of psychopathology and cognitive functioning with parent-child relationship quality in mothers with SUD. A second limitation of the study is linked to the adoption of a naturalistic study design characterized by the absence of a comparison to a randomized group without intervention or administered a different treatment, which could help to examine whether the treatment is really effective in stimulating change in the SUD parents. It might be that shorter interventions, characterized by a more specific focus on parenting and on relationship quality are more effective in enhancing quality of parent-child interactions (Bakermans-Kranenburg et al., Reference Bakermans-Kranenburg, Van IJzendoorn and Juffer2003; Suchman et al., Reference Suchman, Decoste, Rosenberger and McMahon2004). Moreover, it might be that an intervention characterized by more easily accessible contents which requires less cognitive efforts might be profitable for parents with low cognitive functioning (Aharonovich et al., Reference Aharonovich, Hasin, Brooks, Liu, Bisaga and Nunes2006; Bakermans-Kranenburg et al., Reference Bakermans-Kranenburg, Van IJzendoorn and Juffer2003; Porreca et al., Reference Porreca, Simonelli, De Carli, Barone, Filippi, Rigo, van IJzendoorn and Bakermans-Kranenburg2022). A third limitation concerns the lack of repeated measurements of the independent variables over time. Psychopathology and cognitive functioning may be modified by the residential program and these changes could partly explain any effects of treatment on parenting quality. Finally, no measures of maternal reflective functioning were included, which could be particularly relevant with respect to parenting quality in the context of SUD, given its associations with parenting behaviors and treatment outcomes in this clinical group (Pajulo et al., Reference Pajulo, Pyykkönen, Kalland, Sinkkonen, Helenius, Punamäki and Suchman2012; Suchman et al., Reference Suchman, DeCoste, Leigh and Borelli2011). In our study this aspect was captured only to some extent through the observational coding system adopted and, thus, should be further investigated in future research, also in light of studies identifying associations between parental executive functioning and reflective abilities (Håkansson et al., Reference Håkansson, Söderström, Watten, Skårderud and Øie2018).
Conclusions
In conclusion, this is the first study providing evidence that both parental psychopathology and cognitive functioning are associated with changes in the quality of parent-child relationships in the context of a residential treatment for maternal SUD, although with different sequelae. While psychopathological symptoms were associated in a global way with quality of mother-child relationship, cognitive performances seemed to differentiate between diverging patterns of change during treatment.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S0954579424000026.
Acknowledgments
The authors thank Casa Aurora (Comunità di Venezia scs), the director of the facility (Dr Nicoletta Capra), the coordinator (Dr Cappelletto) and the Psychotherapists (Dr De Palo, Dr Prandini, Dr Dalla Cia), for giving the permission to conduct the study and for helping in recruitment and data collection. We thank Micol Parolin for providing supervision in the administration of neuropsychological testing. Special thanks go to the families that took part in the research and to Zeynep Biringen and Marjo Flykt that supervised the first draft of the manuscript providing insightful suggestions for its development.
Funding statement
This work was supported by Comunità di Venezia scs
Competing interests
None.
Open access
