Parent-Child Relationship in the Clinical Context
In 1994, the Diagnostic Classification of Mental Health and Developmental Disorders of Infancy and Early Childhood (DC:0–3) (1) was introduced to provide an alternative and probably more adequate classification of mental diseases for children between zero and three years of age (2). This manual was later revised (DC:0-3R) (3) and extended to preschool age with the DC:0–5 (4,5). In the DC:0-3R manual, a diagnostic framework was applied to assess the disorder in a parent-child relationship (PCR) (DC 0–3: Axis II), and two measures were introduced: the Parent-Infant Relationship Global Assessment Scale (PIRGAS) and the Relationship Problems Checklist (RPCL) (1,3). The PIRGAS is a global one-item measure with a scale from 0 to 100 where scores, e.g., between 91–100 label a relationship as adapted, between 41–50 as disturbed and between 1–10 as documented maltreatment (3). The RPCL classifies relevant parental behaviour for relationship problems using seven global one-item measures, labelling parenting behaviour as overinvolved, underinvolved, anxious/tense, angry/hostile, verbally abusive, physically abusive, or sexually abusive (3).
The PIRGAS and RPCL are potential candidates for assessing important theoretical and practical information to explain and treat child psychopathology (6), even though they are more subjective and therefore susceptible to the biases inherent to all clinical assessments (7), because they can be applied easily and quickly; this is particularly important in times of increasing economic pressure. This makes global assessment approaches more attractive than multiple-item questionnaires (for an overview see Pritchett et al., 2011 (8)), observational instruments (for a review see Lotzin et al., 2015 (9)) that require training, or assessment approaches that use questionnaires with likely invalid parental reports (10). Also, from a validity perspective, a global measure may document clinical impressions better by taking the unique circumstances of a parent-child relationship into account.
However, a disadvantage of these measures is that the assessment approach in PIRGAS and RPCL is not standardized in terms of recommendations on training or regarding how long or in what setting a parent-child dyad should be observed to obtain reliable clinical information (11–13). It does, however, provide vague diagnostic guidelines and names aspects that should be included in a full diagnostic evaluation (1,3). Furthermore, the manual does not provide any references to a clear theoretical background or related empirical studies. A potential further limitation is that the RPCL provides only a dichotomous classification instead of a graded dimensional score. We are also aware of the problem that the RPCL domains represent less a direct measure of relationship quality but more the parental precondition for an adapted parent-child interaction, where these interactions represent the experience background for the mother and the child, which result in an adapted relationship quality. Furthermore, parenting behaviour is reciprocally associated with parental depression, the parent-child relationship(14,15) and the child’s externalizing and internalizing behaviour (16–20).
Given the widespread use and attractiveness of these measures, we want to know whether these global measures are useful in the context of daily routine diagnostics. In our study, useful means that the measures show covariation and concordant changes to child and maternal psychopathology. Moreover, we aim to provide more descriptive context to compare with the scale outcomes; these descriptives include the actual parenting behaviours observed (and their frequency), which behaviours influence the parent-child relationship, the child’s internalizing and externalizing behaviour, and the parental psychopathology. Thus, our paper is not intended to add to the lack of scientific basis for these measures, but rather we attempt to link the content in the following paragraphs and discussion to possibly related constructs and studies. We provide an overview of studies examining the usefulness of the measures, and our focus is on the diagnostic utility of using these global measures in clinical contexts, considering that other more scientifically based assessment methods are available. Further, note that many treatment concepts incorporate the parent-child interaction and parent-child relationship, like Parent-Child Interaction Therapy (21,22), Incredible Years (23,24), and the Positive Parenting Program (Triple P) (25–27). All these approaches have proven to be effective as well as treatment at our Family Day Hospital, which focusses especially on an interaction coaching approach (28) that has shown improvement in child (29) and parental psychopathology (30).
PIRGAS and the RPCL
In her review in 2010, Skovgaard reported a prevalence range of relationship disorders for two- to three-year-old children (DC:0–3, Axis II) between 7.3% up to 26.5% in community or general population-based samples (31). In the same review, Skovgaard described an even wider range of 7.8–72.6% for children aged zero to six years in clinical samples (31). In a subsequently published study, Akca et al. reported a prevalence of 74.8% for a relationship disorder according to DC 0–3: Axis II for an outpatient sample of 457 (32). Note that this large variation in prevalence rates occurs not only for PIRGAS-based classification of relationship disorders but also for the frequency distributions of RPCL subgroups. For example, the frequency of underinvolved parents has been found to range from below 5% (12) up to over 50% (32). A summary of published frequencies of the RPCL categories is given in Table 1. Skovgaard et al. indicated that the methodological diversity between the studies may explain the large variation in frequencies (31), which may be due to differences in the assessed population of children and their parents as well as the age of the children; it may also be due to the largely unstandardized assessment conditions regarding observed interaction settings, raters, and duration of the observation as well as the absence of clear criteria for assigning the diagnosis on Axis II (13,33). However, while studies about the reliability of PIRGAS indicate moderate to good inter-rater reliability (34,35), to our knowledge no inter-rater reliability have been reported for the RPCL measure.
Table 1
Frequency distribution of the RPCL subgroups from six studies
Study | N | no rel. | over- | under- | anxious | angrily | abusive |
---|
| | diagnosis | involved | involved | tensed | hostile | |
Cordeiro et al., 2003* (36) | 343 | 13.12% | 4.37% | 29.45% | 9.62% | 2.92% | 3.21% |
Keren et al., 2003* (12,37) | 414 | 48.0% | 11.1% | 4.7% | 12.1% | 3.0% | 0.7% |
Minde & Tidmarsh, 1997* (37) | 57 | 47.37% | 10.53% | 35.09% | 5.36% | 1.75% | |
Maldonado-Durán et al., 2003* (38) | 167 | 62.8% | 7.18% | 22.75% | 2.3% | 2.3% | 0.5% |
Akca et al., 2012* (32) | 457 | 25.2% | 3.5% | 52.1% | 4.8% | 3.5% | 12.0% |
Skovgaard et al., 2007** (39) | 211 | | | 5.2% | 0.5% | 0.5% | |
*clinical sample, **random sample
Probably as a result of the limited reliability of the PIRGAS and RPCL measures and their limited ability to predict specific diagnoses according to Axis I of the DC 0–3/0-3R, an association between these measures could not be found (38); an association could only be found on a more global level of having or not having any mental health diagnosis according to ICD 10 (39,40) and on the association with the mother’s primary diagnosis (7). However, for dimensional domain-specific measures of mental health problems like increased aggressiveness measured by the Child Behaviour Checklist (CBCL), only a marginal spearman correlation of r = − .23 (41) was found with the PIRGAS, while Aoki et al. described that low PIRGAS scores predict increased internalizing behaviour in the child but not externalizing behaviour (42). Esins et al. (43) reported several associations of the PIRGAS to other interaction-related measures like child-related behaviour, with r = .47 for the observational tool SIRS (Caregiver-Child Socioemotional and Relationship Rating Scale) and r = .49 for parenting behaviour by Play-PAB (Play-Parenting Assessment Battery). However, Albers et al. did not report a significant association between PIRGAS and a questionnaire-based measure for parent-child relationship or clinical interview data with the children (44). The use of PIRGAS for treatment evaluation showed an effect size by Cohens d = .58 (45), and it was correlated with a clinician-reported routine outcome measure for infants from zero to 47 months old with r = − .73 (Health of the Nation Outcome Scales for Infants (HoNOSI)) score; see Brann et al., 2021) (46). Regarding the RPCL measure, to our knowledge no data are available on the validity of the RPCL categories or changes associated with treatment effects. In sum, empirical data are needed to evaluate the usefulness of PIRGAS and RPCL to indicate therapy success or failure and especially in concurrent changes in child and maternal psychopathology, as this is a central psychological mechanism underlying treatments focused on the parent-child relationship.
The clinical value of the RPCL categories lies in their ability to point to parental behaviours that may explain disruptions in the parent-child relationship; these behaviours may lead to specific treatment goals and therapeutic foci. Further, using a review of parental behaviour and its relevance to child behaviour by Achtergarde et al. (47), which was based more on pedagogical rather than clinical literature, and another review by Skinner et al (48), we attempted to link the RPCL categories to their short list of six basic ‘dimensions’ of parenting, namely warmth, rejection, structure, chaos, autonomy support, and coercion (48). These six core parenting dimensions are based on Skinner et al.’s extensive review and list of concepts and terms related to describing parenting behaviours(48).
First, we expect that parental behaviour in the RPCL category overinvolved is related to increased intrusive parenting behaviour, which is associated with increased internalizing problems. Moreover, overinvolved behaviour may lead to conflicts with the child’s autonomy and may lead to an increase in externalizing problem behaviour. Second, we link the RPCL category underinvolved to behaviour with less parental warmth and less structure, which could be interpreted from a child’s perspective as disinterest or rejection, which have been shown to be associated with increased internalizing problems (49). Third, we expect that a parent with high levels of anxious/tense behaviour may be associated with overprotective parenting behaviours that reinforce internalizing problems. Fourth, we expect that an angry/hostile parenting behaviour is connected to power-assertive parenting behaviour, which is associated with externalizing problem behaviour.
From the perspective of the child, children generally use two main coping strategies to display avoidance behaviour. The first is flight and/or freeze (50,51), which tend to be part of an internalizing behaviour. The second strategy, fight, refers to more active behaviour in challenging and frightening situations and is part of externalizing behaviour (50,51). Finally, Bowlby´s attachment theory (52) described a further coping behaviour as a mixture or inconsistent style with the use of both strategies (disorganized Type D behaviour), which is probably the result of verbal, physical, or sexual abusive parenting behaviour. This coping strategy is expected to show higher scores in internalizing and externalizing behaviour (53).
Research Question
Our first objective was to describe the distribution and frequencies of the parent-child relationship quality and parenting behaviour by PIRGAS and RPCL at admission (Table 3) and compare it with frequency data in other studies (Table 1), and then describe how PIRGAS and RPCL categories are associated to each other (Table 4), as such information has not been previously reported. This serves to help describe which parenting behaviours described by the RPCL categories may particularly trigger a disturbed relationship. Moreover, this association helps give a descriptive impression of how strongly different parenting behaviours are associated to each other and whether, for example, overinvolvement often co-occurs with angry/hostile parenting, and this information can be used by the treatment team to include these issues as a distinct therapeutic focus. This information can also be provided at discharge and could provide insight into whether our treatment involves a possible co-occurrence of parenting behaviours, e.g., if a reduction in overinvolvement may lead to a reduction in angry/hostile parenting.
Second, we described the level of child and maternal psychopathology and changes from admission to discharge (Table 3), how strong both psychopathologies were correlated to each other at admission and discharge (Table 5), and whether improvement in child psychopathology was associated with improvement in maternal psychopathology or vice versa. Note that we only wanted to describe the associations without interpreting them as causally directed changes.
In a third and final step, we asked whether improvement in the parent-child relationship and parenting behaviour were associated with a decrease in child or maternal psychopathology (Table 6). The results may indicate which specific parenting behaviours are associated with a greater reduction in child internalizing and externalizing behaviours, respectively.