Prevention of Divorce-Related Problems in Dutch 4- to 8-Year-Olds

Family patterns have changed substantially in Europe over the past 50 years, with one of the markedly changes being the increases in rates of parental divorce (Oláh, 2015). These rates have increased steeply in Western Europe and other parts of Europe (e.g., Scandinavia) from the 1970s onward. Oláh (2015) forecasted that about half of marriages may end in divorce if early 21st-century trends will hold. In 2009, about 1 million marriages ended in divorce in the whole European Union (EU-27 (European Commission, 2013), and to this number add a similarly high number of terminations of live-in partnerships. To compare with, 2.2 million marriages took place in 2010. In the United States, divorce rates are just as high (National Centre for Health Statistics, 2010).

Children are frequently involved in divorce and separation. For example, in England and Wales, half of the couples divorcing had at least one child under the age of 16 in 2011 (Office for National Statistics, 2012). In the Netherlands, about 70,000 at-home-living children are involved in parental divorce or separation each year (Spruijt & Kormos, 2014). This regards 57,000 at-home-living children aged below 18 and 13,000 aged 18 years and over and includes termination of both common law relationships and long-standing live-in partnerships. For comparison, a Dutch birth cohort comprises about 180,000 children per year.

Research has shown that children encounter notably negative consequences of their parents’ divorce or separation (from hereof “divorce” refers to both divorce and separation). Children of divorce achieve worse in school, have more behavioral problems, score lower on psychological and emotional well-being, have a lower self-esteem, and display more problems in social relationships than children growing up in intact families (Amato & James, 2010). Parental divorce or separation is regarded as one of the 10 major adverse childhood experiences (ACEs; including various types of child maltreatment and childhood adversities rooted in household dysfunctions) studied by Felitti and Anda (2010) in their ACE study. Felitti and Anda illustrated the long-lasting, strongly proportionate, and often profound relationship between ACE and important categories of emotional state, health risks, disease burden, sexual behavior, disability, and health-care costs, decades later (Felitti & Anda, 2010).

Consequences of parental divorce for children are rather generalizable over countries. A large body of research is based on populations in the United States, but other studies have confirmed similar results in Canada, Australia, New Zealand, and Europe, including the Netherlands (e.g., Amato, 2010; Amato & James, 2010; Pedro-Carroll & Klein Velderman, 2015; Spruijt & Kormos, 2014). In general, these studies demonstrate that parental divorce is a major stressor for children across countries and cultures and from families of varying wealth. They indicate that a breakup of the family structure puts high strains on the psychological adjustment of children.

Lack of Preventive Support for Young Children of Divorce in the Netherlands

Prevention of problems for children of divorce and helping children adapt to the divorce are major priorities, given the large number of children involved and the high, negative impact of divorce on children. However, research-based interventions aiming at these topics are rare, and those that are available have almost exclusively been based on research in the U.S. setting. The effectiveness of such interventions still needs to be further confirmed in other cultural settings. An example of such a setting is the Dutch one. In the Netherlands, half a decade ago, no effective research-based prevention program was available to support children dealing with divorce (Cloostermans, Klein Velderman, & Pannebakker, 2013; Klein Velderman et al., 2011; Vermeij, Van der Wel, & Krooneman, 2005). The available support for children only concerned rather local or isolated initiatives, in particular support groups, websites, brochures, and opportunities for individual (telephone) consultation. These initiatives were primarily aimed at children aged at least 8 years, and none of these had been shown to be effective in a well-designed experimental study. Accordingly, there was a need for an evidence-based prevention program for children of divorce in the Netherlands, first and foremost for children under 9 years of age.

The Children of Divorce Intervention Program

The Children of Divorce Intervention Program (CODIP), developed in the United States to prevent divorce-related problems in children, might provide a response to this need for preventive support in the Netherlands. CODIP is a selective preventive intervention program, built on the assumption that timely intervention for children of divorce can offer important short-term and long-term benefits. It is based on theories of play therapy, developmental psychology, stress and coping, resilience promotion, and theories of prevention that emphasize the importance of timely supportive outreach to reduce risk across systems that impact children. CODIP aims at creating a supportive group environment in which children can freely share experiences, establish common bonds, clarify misconceptions, and acquire skills that enhance their capacity to cope with the stressful changes that divorce often poses (Pedro-Carroll, 2005).

CODIP is based on research on risk and protective factors that shape children’s adjustment to parental divorce. Studies on risk and resilience in divorce research provide an important understanding of why some children get back on track after an initial period of adjustment while others struggle with long-term problems. This includes family factors, such as parental conflict, psychological well-being of parents, quality of parenting, social support and family stability; and individual child factors, such as effective coping skills, emotion regulation, age-appropriate understanding of family changes, a realistic sense of control, and positive outlook and hope for the future (Pedro-Carroll, 2010). Many of these factors are incorporated in the CODIP, which has served over 20,000 children since its inception in 1982.

Four CODIP modules exist: one for kindergarten and first-grade youngsters, one for early latency children (second and third grades), one for older latency children, and one for early adolescents. Although the goals and objectives embedded in the CODIP curricula remain relatively constant, these versions contain varying topics and techniques according to the developmental differences of the four age-groups. For instance, groups for younger children are smaller and take 45 minutes contrasting to sessions of 1 hour for older children.

Experimental studies in the United States by Pedro-Carroll and Cowen (1985), Alpert-Gillis and colleagues (1989), Pedro-Carroll, Alpert-Gillis, and Cowen (1992), and Pedro-Carroll, Sutton, and Wyman (1999) have shown positive effects of CODIP on self-, parent-, and teacher-reported internalizing and externalizing problems, competence, adjustment of participating children to the new family situation, and divorce-related cognitions for children of different ages and demographic backgrounds. Effects persisted over a 2-year follow-up. CODIP has been translated into French, German, and Turkish, but evidence on the effectiveness of adapted CODIP modules as established in well-designed experimental studies outside the United States was lacking.

Study Objectives

The effectiveness of CODIP as shown in the United States supports the introduction of CODIP in other settings such as Dutch preventive practice. However, fundamental differences between countries may hamper its generalizability. For example, applied to the Dutch setting, a program from elsewhere might require an implementation structure that does not fit into the Dutch context or “care-as-usual.” Therefore, the introduction of CODIP in the Netherlands was given careful consideration.

This study encompasses the adaptation of the 12-session CODIP module for kindergarten and first-grade children of divorce and the 15-session CODIP module for second-grade and third-grade children of divorce to support 4- to 6-year-old and 6- to 8-year-old children in the Dutch setting and the assessment of the feasibility of these adapted versions (i.e., “CODIP-Netherlands [CODIP-NL]”). It therewith assesses preconditions to fill the lacuna regarding research-based preventive support for Dutch 4- to 8-year-old children of divorce in the Netherlands. Its objectives are:

Translation and initial adaptation of CODIP materials for school-going kindergarten and first-, second-, and third-grade children of divorce to the Dutch context;

Testing the feasibility of implementing CODIP for school-going children of divorce aged 4–8 years in a pilot study in the Dutch setting; and

Determining the feasibility of replicating positive and desired effects of CODIP for Dutch 4- to 8-year-olds as shown in the United States.

This study is the first one to provide information about the preconditions for introducing CODIP for young school-going children in the Netherlands and about the feasibility of these CODIP-NL modules in the Dutch setting.

Method

In order not to lose the core principles of the original program, the process of adapting and testing CODIP in the Netherlands took place in two stages. Stage 1 related to the first study goal. The CODIP modules were translated and adapted for 4- to 6-year-old and 6- to 8-year-old children of divorce. Stage 2 concerned an empirical stage: A preexperimental pilot study was conducted in order to provide answers to the second and third study goals, regarding the feasibility of implementing the CODIP-NL modules for 4- to 8-year-olds and the possibility of replicating positive intervention effects of CODIP as proven in the U.S. studies.

Stage 1: Translation and Initial Adaptation

Cultural adaptation adhering to a conceptual outline

We have started our work with CODIP-NL for 6- to 8-year-old children of divorce. Next, we extended this to 4- to 6-year-olds, following the same procedure. The rationale behind the choice for this particular age-group (early latency children) as a first step in this trajectory of introducing CODIP in the Netherlands was as follows: Together with a CODIP module for older latency children, the CODIP module for early latency children was the first developed by Pedro-Carroll and colleagues in the 1980s. Therefore, we could build on extensive practical and evaluative experience in regard to this specific module. Moreover, children in Grades 2 and 3 have started to learn how to read and write. We assumed that this enables them to join in some of the CODIP games and activities more easily than younger children.

Thorough translations of the CODIP program materials (Children’s Institute, Rochester, NY) were the basis of first Dutch versions of the procedure manual and other written materials per age-specific module. First, we translated materials from English into Dutch as literally as possible. Some expressions were reconsidered or altered if a literal translation was not appropriate. Next, the cultural adaptation of CODIP and the design of the initial Dutch version were done. Whenever possible, adapted or designed materials and alternatives (e.g., for U.S. reading books, posters, etc.) were kept alike for the two age-specific modules.

To prevent that translated and adapted CODIP materials deviated from the core elements that constitute the effectiveness of the original program, the U.S. version of CODIP was adapted to the Dutch user population based on the conceptual outline (i.e., depicting underlying theoretical principles of the program from which concrete program activities can be derived) and consultation of Dr. Pedro-Carroll, PhD, the founder and developer of CODIP in the United States. For example, the adaptations to meet the characteristics of the Dutch user population were made only as far as they stayed in line with the theories (for change) that constituted the original program. Program materials were designed to be attractive to intermediate users (organizations and CODIP-NL group leaders) as well as end users (participating children of divorce and their parents). The principal group leader from the pilot organization participating in the current study (i.e., this article’s third author) commented on the draft versions of the Dutch procedure manual and materials.

The current study was preceded by a pilot study in four groups using the first version of CODIP-NL for 6- to 8-year-olds (Klein Velderman et al., 2011). This led to an important reduction in the number of intervention sessions in the current second version of CODIP-NL for 6- to 8-year-olds. Half of the group leaders in that study had suggested such a reduction. For practical and financial reasons, this suggestion was considered, in consultation with group leaders and Dr. Pedro-Carroll, leading to a reduction from 15 to 12, 45-minute sessions. The resulting second version of CODIP-NL for 6- to 8-year-old children (Klein Velderman et al., 2011) was assessed in the present pilot study.

Examples of adaptations

In general, adaptations in CODIP-NL compared to CODIP were limited to wordings, layout, and mode. To give an example, per age-specific CODIP module, Children’s Institute (Rochester, NY) has developed a procedure manual for conducting the support groups, tailored to the developmental needs of children. These manuals offer group leaders clear guidelines for their work and are well structured. First, an introduction is given about the intervention program and its goals. Next, the authors give a module overview and information about program implementation (including group facilitation techniques and group process issues). The main part of the manual consists of information about the module sessions. For each session, goals, procedure, and needed materials are listed and subsequently illuminated. The program procedure mentions concrete acts, giving sample quotations, and providing review questions, closing remarks, and/or notes to the group leaders. In the CODIP-NL procedure manuals, this structure was maintained. Some sample quotations were expanded to give the group leaders even more examples during structured group activities. In addition, first names of children in given examples were changed into Dutch names (including immigrants’ names).

For reading or picture books used in the original U.S. version of CODIP, possible alternatives were sought or de novo designed. For example, for the 6- to 8-year-olds, we found an existing alternative reading book that well suited the intervention: Julia Has Two Houses (Wisse Smit, 2008). It has chapters about living in two homes, difficult choices (e.g., missing mum or dad); visiting dad by train (moving back and forth); dad in love; and having a new baby brother. The group leaders can select sections from the book that are most relevant for their group to generate discussion.

Different use of emotion words

In some cases, differences between the United States and the Netherlands necessitated more adaptation. An example of program materials that resulted from a necessary cultural adaptation concerns the new dinosaur feeling faces posters (i.e., depicting basic and more complex emotions). These were redesigned to substitute the original “feeling faces” poster. The reason for this adaptation was that consultations with social workers who led the CODIP-NL pilot groups yielded that some of the emotions depicted in CODIP are not commonly used by children in Dutch spoken language (e.g., loving). Also, the social workers requested to include some extra emotions that were not shown on the original poster but were frequently used by young children during group or individual support sessions (e.g., calm, worried, enthusiastic, content, disappointed, relieved, unhappy, relaxed). Thus, a new selection of emotions was needed for the Dutch version of the intervention.

We chose the form of dinosaur faces to match the Dutch name of the CODIP-NL module for the 6- to 8-year-olds, “Dappere Dino’s” (Dutch for “courageous dinosaurs”), named after the group puppet “Dinosaur Rex” and the “Daring Dinosaurs” board game (see below). The feeling faces are also available as separate cards, to be used, for example, in the “feelings charades” game. Children take turns choosing cards depicting the various emotions and silently act out the emotion, while group members guess the feelings and name a time when they too felt that way. For the youngest age-group, the four basic emotions happy, sad, angry, and scared are displayed on a separate poster, being the first to be used during emotion games.

Implementation requirements: The Dutch primary school system

Although CODIP can be implemented in a variety of settings, such as mental health centers, community centers, private practitioners’ offices, after-school care programs, and court-connected service groups (Pedro-Carroll, 2005), in the current study, we focused on implementation of CODIP-NL in the school setting by school social workers. The U.S. program developers regard schools as a natural setting for CODIP groups, because of the accessibility of large numbers of children sharing similar experiences who can continue in their supportive relationships after the groups end (Pedro-Carroll, 2005). Furthermore, school-based professionals are ideal candidates for group leaders because of the potential continuity of their contacts with program children and their families.

In the Netherlands, children enter primary school when reaching the age of 4, the first 2 years being kindergarten. A child proceeds to Dutch third grade when 6 years old. The Dutch primary school system thus differs from that in the United States, where a child is eligible for admission to kindergarten if the child is 5 years of age and continues to first grade after a year in kindergarten. In line with this, the original U.S. CODIP module targeting the youngest children was developed for 5- and 6-year-old kindergarten and first-grade children. However, preventive support in the Netherlands was primarily lacking for young children of divorce, in their first 4 years of primary school: 4- to 6-year-olds (Dutch Kindergarten) and 6- to 8-year-olds (Dutch third and fourth grades). Therefore, the U.S. kindergarten and first-grade children module was the basis for CODIP-NL, targeting the younger children aged 4 and over. To adapt more to this young children’s developmental stage, some additional adaptations (i.e., simplifications) were made. A good example of this concern is the adaptations made in the intervention’s board game for this age-group (e.g., adding emotions and feelings cards).

Adaptations of the intervention’s board game

As in U.S. CODIP, play, including puppet play, is a significant element of the CODIP-NL program. The procedure manuals are consequently accompanied by, for example, the intervention’s board game. For CODIP-NL for 6- to 8-year-olds, the game is named “The Daring Dinosaurs game.” This game was specially designed by U.S. CODIP developers to foster children’s sense of self-efficacy and assess children’s progress in understanding divorce-related issues and developing resilience skills. The game is one of several therapeutic techniques used in the program, designed to address emotion regulation, family and divorce-related issues, social problem-solving, communication, and anger control skills and to promote self-awareness and self-esteem. Game cards incorporated in the board game ask questions about children’s thoughts (e.g., “Do you believe you can make your parents get back together?”), their feelings (e.g., “How do kids feel when their parents fight?”), and ways to self-soothe when feeling upset (e.g., “Act like you are feeling lonely. Name two things that you could do to feel better.”). If a child cannot answer a question, other children and the group puppet are invited to help. The content of the cards covers most of the topics explored in the previous sessions. Blank cards are included with the game so that group leaders can write individualized cards to reflect problems, situations, or feelings specific to their group. Group leaders are encouraged to “stack the deck” so that the most relevant cards for the children in their group are placed on top (cf. Pedro-Carroll & Klein Velderman, 2015).

For the 6- to 8-year-olds, the Daring Dinosaurs board game’s game cards in CODIP-NL were to a large extent kept similar to those in U.S. CODIP. However, it was structured a bit more than in the original version. That is, participants play the game twice, and the second time, the game focuses on the parents and the parent–child relationship. Therefore, two sets of cards were made (compared to one set in U.S. CODIP) denoted by color (green for parent–child game cards, blue for other game cards). Cards were numbered, and in the attachment of the manual, the numbered list of the available cards was provided to assist leaders in stacking the deck (see above). Both sets (parent–child and other) contain basic cards (i.e., cards suggested to use in all CODIP-NL groups), the so-called teasers (e.g., “Your dinosaur falls down, go back 3 places”; “You are lucky, you can have another go”; and “Run around the table”), and additional cards (i.e., cards to be selected by the group leaders; fitting the relevant themes in their group).

In CODIP-NL for 4- to 6-year-olds, the game is named the “Tough Turtles board game” and played during one session only (as in U.S. CODIP). For this age-group, the Dutch set of play cards has four categories: basic cards, teasers, emotions, and feeling cards. The basic cards and teasers are based on those in CODIP-NL for 6- to 8-year-olds, although formulation is made more simple. For these young children, the hand puppet serves as quiz master, reading out basic or teaser cards. The emotions and feeling cards are new additions. The emotions cards contain the dinosaur faces of the four basic emotions. When a child gets such a card, a group leader can welcome the child to act out this emotion or name an occasion the child had felt like that. The feeling cards are 14 icons specifically drawn for CODIP-NL for 4- to 6-year-olds. These icons depict daily experiences (e.g., celebrating a birthday, falling from your bike, waking up during a thunderstorm) and are used as examples during other intervention activities. When a child gets such an icon card, it can be asked how the child shown on the icon feels, and in case this is not a happy feeling, what the child could do to feel better (i.e., practicing problem-solving skills).

Differences in common practice: The use of puppet play

As in the original version of CODIP, puppet play is used extensively in CODIP-NL. The CODIP-NL module for 4- to 6-year-olds was called “Stoere Schildpadden” (Dutch for Tough Turtles) and the group hand puppet is a turtle named Sam, whereas the group hand puppet in the CODIP-NL module for 6- to 8-year-olds Dappere Dino’s (Courageous Dinosaurs) is a young dinosaur named Rex. The use of hand puppets is not very common in the Netherlands. This therefore receives a lot of focus during the training of professionals. Additionally, more than in the original procedure manual, the use of the hand puppet in activities is explicitly requested, and sample quotations of group leader hand puppet role-plays are provided. To remind group leaders of hand puppet usage, every time a group leader is supposed to use the hand puppet, small paws are printed in the margin of the manual. This is an example of how differences in common practice have led to cultural adaptations of the intervention.

Stage 2: Preexperimental Pilot Study

Sample and intervention

Participants of the current study evaluating the feasibility of the adapted CODIP-NL modules were 43 Dutch children (17 participating in CODIP-NL for 4- to 6-year-olds and 26 in CODIP-NL for 6- to 8-year-olds), supported in nine pilot intervention groups (2–6 children each). Eight CODIP-NL groups took place in the urban region of The Hague (primarily in The Hague itself; the third largest city in the Netherlands) and one in Almere (the fifth largest city in the Netherlands). Children participating in CODIP-NL for 4- to 6-year-olds were eight boys and nine girls, aged 4 (50 months)–7 years (85 months; pretest group leader report; M = 5.7 years, SD = 0.7). Participants in the 6- to 8-year-olds’ curriculum were 13 boys and 13 girls, aged 5 (68 months)–8 years (106 months; pretest parent report; M = 7.5 years, SD = 0.8). The length of time parents had been separated at the time of data collection differed: from 3 months up to over 4 years for the parents of the 4- to 6-year-olds and 4 months up to 5 years for the parents of the 6- to 8-year-olds. In some cases, common law relationships had not yet officially been terminated (i.e., juridical procedures were still ongoing).

Criteria for inclusion in the program were similar to those in the U.S. studies (Alpert-Gillis, Pedro-Carroll, & Cowen, 1989): (a) parent separation or divorce, (b) children not currently in therapy or receiving other mental health services, and (c) permission by both parents who have custody rights. A two-step procedure for inclusion was applied in this study. First, schools were invited to participate in the project. Next, children from these schools who met the criteria were invited to participate in the intervention. Questionnaires were handed out by the group leaders and returned to the researchers by post in sealed envelopes.

The CODIP-NL interventions for children aged 4–8 years entailed 12 weekly sessions (45 minutes each) of CODIP-NL’s modules Stoere Schildpadden™ (Tough Turtles; 4- to 6-year-olds) or Dappere Dino’s™ (Courageous Dinosaurs; 6- to 8-year-olds), provided by two experienced professionals (i.e., social workers with a college or university of applied sciences degree). These professionals have all participated in a face-to-face training and sessions of telephone supervision to conduct groups according to the adapted CODIP-NL manual. The study was approved by the medical ethical board of the Leiden University Medical Center (P09.083).

Procedure and measures

The intervention process and user satisfaction were evaluated based on group leader (intermediate users) and mother and child reports (end users). Mothers, group leaders, and school teachers completed questionnaires to evaluate program outcomes on child functioning. Each child’s primary school teacher and mother received a questionnaire before the start (pre) and after the final (12th) intervention session (post). Group leader ratings were completed after Sessions 4 (pre) and 12 (post). This resembles the procedure in U.S. research (e.g., Alpert-Gillis et al., 1989).

Process evaluation and user satisfaction measures

Mothers filled out a questionnaire (by Pedro-Carroll & Cowen, 1985) on how their child might have changed since the program began. The questionnaire starts with 6 close-ended items (e.g., since the program began, my child “talks about his/her feelings” or “is able to handle problem situations”) to be rated on a 5-point scale ranging from 1 = much less to 5 = much more. Additionally, three open-ended questions ask (1) “In what ways have your child’s feelings and behavior changed since the program began?” (2) “What were your child’s reactions to his/her ‘special’ group?” And (3) “What were your reactions to the program? To the parent newsletter?” In addition to this U.S.-developed evaluative questionnaire, mothers were asked some evaluative questions, specifically addressing several intervention aspects (e.g., duration, group size, materials, content).

Group leaders filled out a semistructured logbook after each intervention session, assessing program quality and fidelity. It focused on the group leaders’ usage and perception of each session’s goals, materials, and content. That is, for each session, group leaders first registered the date of the session, possible absence of children, and the amount of time they spent on preparation. Second, an open-ended question asked them to give an overall description of this session. Third, the objectives of this session were listed. For each objective, group leaders scored if they felt they had reached that objective: yes, partly, or no. Similarly, they were asked to state for each session activity if these were put into practice: yes as described, yes but deviated from manual instruction, or no. When deviated from instruction, they were asked to explain how and why. Finally, group leaders were asked to reflect on session materials: were materials used and, if so, were these regarded as positive, neutral, or negative. Again, group leaders were asked for an explanation. In addition to the logbooks, at the posttest, group leaders were asked some evaluative questions, specifically addressing intervention aspects (e.g., duration, group size, materials, and content), training, and supervision.

In the CODIP-NL for 6- to 8-year-old groups, a short evaluation with participating children took place. To do so, a group interview was held at the start of the last intervention session. During this last session, children celebrated a small party to wrap up the intervention. One researcher was present during this evaluation, but a group leader led the discussion because of familiarity to the children. The evaluation was recorded and written out afterward. Responses of children participating in the CODIP-NL for 4- to 6-year-old groups were not formally assessed as part of the study. However, group leaders normally do a group evaluation with the children as part of the last intervention session and have shared child experiences from these evaluations with the researchers.

Child functioning measures

Mother-reported child functioning concerned the Parent Evaluation Form (PEF; Alpert-Gillis et al., 1989) and the Strengths and Difficulties Questionnaire Parent Form (SDQ-PF; Goodman, 1997). The PEF contains 22 items on parents’ views of children’s feelings (e.g., “Feels responsible for family problems if they occur”), behavior (e.g., “Talks with me about how he/she feels”), and problem-solving skills (e.g., “Tries to solve own problems”). Items are rated on a 4-point Likert scale (1 = very true to 4 = not true at all). High PEF sum scores indicate better adjustment. The αs in the current study were (pretest/posttest) .86/.77 at 4–6 years and .88/.82 at 6–8 years.

The SDQ-PF (for aged 4–16) is a brief behavioral screening questionnaire consisting of 25 items, assessing psychological adaptation of children (e.g., “Considerate of other people’s feelings”; “Restless, overactive, cannot stay still for long”; and “Generally liked by other children”). Each item has to be scored on a 3-point scale (0 = not true, 1 = somewhat true, and 2 = certainly true). It generates scores for conduct problems, hyperactivity, emotional symptoms, peer problems, and prosocial behavior. All but the last are summed to generate a total difficulties score. Cronbach’s αs for total difficulties in the current study held (pretest/posttest): .79/.79 at 4–6 years and .84/.62 at 6–8 years.

Group leader–reported measures concerned the Group Leader Evaluation Form (GLEF; Alpert-Gillis et al., 1989) and the SDQ Teacher Form (SDQ-TF for ages 4–16; see SDQ-PF above). The 23-item GLEF assesses children’s strengths or competences (e.g., “Expresses feelings appropriately” and “Recognizes differences between problems he/she can and cannot solve”), as well as problems that children of divorce experience (e.g., “Believes he/she can bring parents together”), leading to an overall adjustment score. Group leaders rate each item on a 4-point Likert scale, ranging from 1 = very true to 4 = not true at all. High scores indicate better adjustment. Cronbach’s αs were (pretest/posttest): .88/.94 at 4–6 years and .89/.77 at 6–8 years.

Group leaders also filled out the SDQ-TF. Cronbach’s αs for total difficulties were (pretest/posttest): .60/.78 at 4–6 years and .89/.81 at 6–8 years. Each participating child’s primary school teacher filled out the SDQ-TF as well. The αs for total difficulties were (pretest/posttest): .77/.67 at 4–6 years and .84/.87 at 6–8 years.

Analyses

Process evaluation and user satisfaction measures

To analyze the results of our pilot study, we chose a simple (statistical) design. The process of implementing CODIP-NL in the pilot organizations was described qualitatively by discussing the open-ended questions and quantitatively by frequency counts. For interpretation purposes, the group leaders as well as the program developer in the United States were asked to reflect on findings, prior to publication.

Impact of CODIP-NL on child functioning measures

For the two age-groups separately, we assessed differences in child functioning between pretest and posttest with paired t-tests. Statistical significance concerned the two-tailed 5% level. In addition, (standardized) mean difference scores for pretest and posttest outcomes were computed as measures of effect sizes over time (Cohen’s d; Cohen, 1969).

Comparison of child functioning results in the Netherlands and the United States

We compared the (standardized) effect sizes over time of the PEF and GLEF with those reported for the United States (i.e., the U.S. results reflected differences between pretest and posttest mean scores; Alpert-Gillis et al., 1989; Pedro-Carroll & Alpert-Gillis, 1997). In Pedro-Carroll and Alpert Gillis (1997) and Alpert-Gillis, Pedro-Carroll, and Cowen (1989), CODIP participants (“U.S. program,” N _5–6 = 37, N _7–8 = 50) were recruited through program announcements in school newsletters or contacts with families by school professionals. Control (“U.S. divorce controls,” N _5–6 = 26, N _7–8 = 47; i.e., children with separated parents but not participating in CODIP) and comparison subjects (“U.S. intact controls,” N _5–6 = 39, N _7–8 = 75; i.e., children from intact families) were recruited by a letter describing a study of child development and family life. The three samples in the U.S. studies did not differ in sex, grade, or racial composition.

Results

Process Evaluation and User Satisfaction

4- to 6-year-olds

Pilot study results showed that mothers were enthusiastic about the adapted CODIP-NL intervention. All mothers of participating 4- to 6-year-olds (available on n = 11) had perceived positive responses of their child to the intervention, and the majority (n = 9) found that their child was positively changed by the intervention. A mother stated about her son: “It has become more quiet in his head. He is calmer now, and understands things better in regard to the divorce.” And another mother stated about her daughter: “She is more eager to initiate the expression of her feelings or problems; talks more about what she does when she has visited her father, seems to have become more open about this.” The evaluative questionnaires furthermore revealed that most mothers (80%) were satisfied about the amount of sessions that constituted CODIP-NL. Similarly, the majority of mothers thought the length of the sessions to be good. Within the sessions, mothers appreciated most the fact that their child had peers to talk to, that their child had more understanding of the situation, the playful way of teaching by the trainers, and the feedback to them as a parent.

The primary group leader of the first four pilot CODIP-NL for 4- to 6-year-old groups liked working with the program. She appreciated the materials (“practical,” “SUPER”) and felt that it was possible to reach program goals by means of the instructions provided. She was positive about the amount and length of sessions. Logbooks show that most of the specific goals of every session were accomplished. Also, the group leader reported that almost all exercises were executed as described in the manual. In some cases, the trainer executed exercises in a different way (e.g., not doing a stand-alone puppet play but doing a conversation with the puppet and the children). Reported reasons were a need for repetition after school holidays or lack of children’s concentration. In her logbooks, the group leader provided us with some specific suggestions for further development of the intervention. For instance, the group leader suggested to increase the amount of “movement exercises,” because of the attention span of this age-group. As an example, in the first session, a ball game could be used for introduction instead of getting to know each other while sitting on chairs. The group leader’s suggestions were incorporated in the program materials that are currently used in practice.

Children liked the program (e.g., “Stoere Schildpadden is fun!”) and appreciated the group as a place to discuss their feelings. The children reported to have learned the difference between childhood and grown-up problems and made new friends.

6- to 8-year-olds

For the 6- to 8-year-olds, 80% of all mothers (n = 20) perceived positive responses and 85% saw positive changes in their child since the start of CODIP-NL. In their opinions, CODIP-NL had provided clarification, children became more open (more expression of feelings) or were calmer after participating. For example, “My son can now handle the divorce more and knows that his father and mother will always be there for him.” And “My daughter talks more, is spontaneous again like she always was before and increasingly expresses her feelings.” In response to the question on what they appreciated most, 11 mothers named the support of the group leaders (for their child and themselves; e.g., “The personal attention and profundity and openness to parental feedback.”). Mothers (n = 9) also mentioned the content of the intervention: “A lot of attention for the situation, problem-solving, expressing, and sharing feelings.” And “My child received support to be able to understand and accept the divorce.” Eight mothers related to peer contact in a safe environment as something they appreciated most: “She could freely give her opinion without hurting anyone (me or her dad).” Or “The support my daughter experienced and the feeling not to be the only child of divorced parents.”

We also asked mothers what they had missed. Fourteen (of the 20) mothers had not missed anything. Two mothers would have liked to receive more information about their child (e.g., a personal child report). Three mothers had other ideas about parent involvement: “Maybe another meeting with parents to teach us how to talk about the divorce with our children.” Or “A one-to-one conversation with the group leaders (without my ex-partner present).” One mother had missed more profundity. Furthermore, the evaluative questionnaires revealed that most mothers (95%) were satisfied about the amount and length of sessions that constituted CODIP-NL.

The group leaders of CODIP 6–8 years liked to work with the intervention. They were positive about the intervention course and were satisfied by the variety of activities within sessions. Group leaders were also positive about the number of sessions and duration. Their logbooks showed that most sessions’ specific goals were accomplished. Two group leaders had not accomplished one of Session 5’s specific goals, namely, reflection of children’s at-home experiences with the problem-solving cartoon. Children in these two groups may have not practiced this at home. Most session exercises were executed as described. However, one trainer reported that sometimes lack of time led to shortened or simplified execution of exercises as compared to description in the procedures manual. This trainer also sometimes refrained from doing the session evaluation (or executed this evaluation without using hand puppet Rex). For planning reasons, one group leader (after discussion with the project team and Dr. Pedro-Carroll) decided to cluster Sessions 1 and 2 and Sessions 11 and 12. This worked well and was appreciated by the group leaders involved. Group leaders reported that they would have liked more suggestions for movement exercises in the curriculum. The logbooks furthermore showed that most program materials were used and appreciated. In some occasions, group leaders had added materials (e.g., child characteristics from the child qualities game, dinosaurs as game pawns, stickers).

All children participating in CODIP-NL for 6- to 8-year-olds liked the intervention. The intervention’s board game was appreciated best, alongside with the noughts and crosses problem solutions game, the problem-solving cartoon, and hand puppet Rex. All children expressed to have learned something. As examples, learning ways to solve problems and to talk about their feelings were named most (e.g., “I know more to tell my stepmother now. This Thursday, they come together to my presentation at school about rabbits.” “That I’m not that sad anymore.”). Participating children enjoyed the small-sized groups: “That way I can work more peacefully.” And “I prefer a small group. Because I found out that in a larger group I sometimes want to say something, and then I interfere with others. I don’t like that. I think a small group is better.”

A group leader’s quote about a CODIP-NL for 6- to 8-year-olds’ participant:

A boy had problems concentrating in his school class. Staff even planned to do a diagnostic evaluation in regard to conduct problems. At “Dappere Dino’s,” he discovered, from the first session, that more children experienced similar things with their divorced parents. From that first moment, things went better at school. Apparently, he felt lonely and misunderstood. Recognition is the added value.

Impact of CODIP-NL on Child Functioning Measures

4- to 6-year-olds

Figure 1 presents the standardized mean difference scores (Cohen’s d) for pretest and posttest outcomes of 4- to 6-year-old CODIP-NL participants, on the SDQ-PF and TF scales, the PEF, and the GLEF. As Figure 1 shows, mothers (n = 11) reported statistically nonsignificant pretest to posttest increases in PEF scores ( p = .18, d = 0.43) and psychological adaptation of the 4- to 6-year-old children as measured using the SDQ-PF (e.g., SDQ total difficulties, p = .70, d = 0.12). Posttest group leader–reported participants’ adjustment (overall GLEF; n = 17) scores statistically significantly exceeded pretest scores. After participation, children tended to show less problems measured using the GLEF ( p = .09, d = 0.44) and significantly had more competencies ( p < .05, d = 0.68) than before participation. After participation in CODIP-NL, no statistically significant decreases in group leader–reported problems on the SDQ-TF were found. The teacher-reported conduct problems of children (n = 17) had significantly lowered after participation in CODIP-NL (p < .05, d = 0.67). Other pretest to posttest SDQ-TF differences as reported by the teachers did not reach significance.

OPEN IN VIEWER

Figure 1.

Standardized mean difference scores (Cohen’s d) for pretest and posttest outcomes of 4- to 6-year-old CODIP-Netherlands participants, on the SDQ-PF and TF scales, the PEF, and the GLEF, by the mothers (n = 11), group leader (n = 17), and teachers (n = 17). SDQ-PF = Strengths and Difficulties Questionnaire Parent Form (Goodman, 1997); SDQ-TF = Strengths and Difficulties Questionnaire Teacher Form (Goodman, 1997); PEF = Parent Evaluation Form (Alpert-Gillis et al., 1989); GLEF = Group Leader Evaluation Form (Alpert-Gillis et al., 1989); CODIP = Children of Divorce Intervention Program. *p < .05 (two-sided).

6- to 8-year-olds

As Figure 2 shows, mothers (n = 20) reported statistically significant increases in PEF scores of 6- to 8-year-olds (p < .05, d = 0.56). In addition, small- to medium-sized standardized pretest to posttest decreases on the mother-reported SDQ-PF scales (n = 19) were found. These were statistically significant on mother-reported emotionality ( p < .05, d = 0.56) and total difficulties ( p < .05, d = 0.76). Moreover, participation in CODIP tended to lead to decreased mother-reported conduct problems ( p < .10, d = 0.50). Posttest group leader–reported child adjustment scores (overall GLEF; n = 26) were higher than at the pretest (after Session 4; p < .01, d = 0.61). More in detail, GLEF problem scores at the posttest exceeded pretest scores ( p < .01, d = 0.69). Participants’ GLEF competencies measured also tended to have increased after participation ( p = .07, d = 0.37). In addition, emotionality had significantly decreased after participation ( p < .05, d = 0.49), and conduct problems tended to show a decline after participation as well ( p = .09, d = 0.44). There were no statistically significant pretest to posttest differences on the SDQ-TF as reported by teachers (n = 18). However, participating children tended to show less teacher-reported emotionality after participation than at the pretest ( p = .097, d = 0.41).

OPEN IN VIEWER

Figure 2.

Standardized mean difference scores (Cohen’s d) for pretest and posttest outcomes of 6- to 8-year-old CODIP-Netherlands participants, on the SDQ-PF and TF scales, the PEF, and the GLEF, by the mothers (n = 19), group leaders (n = 26), and teachers (n = 18). SDQ-PF = Strengths and Difficulties Questionnaire Parent Form (Goodman, 1997); SDQ-TF = Strengths and Difficulties Questionnaire Teacher Form (Goodman, 1997); PEF = Parent Evaluation Form (Alpert-Gillis et al., 1989); GLEF = Group Leader Evaluation Form (Alpert-Gillis et al., 1989); CODIP = Children of Divorce Intervention Program. *p < .05. **p < .01 (two-sided).

Comparison of Child Functioning Results in the Netherlands and the United States

Comparison of the Dutch intervention participants’ PEF functioning scores (mother reports) with results of children in previous U.S. research (Alpert-Gillis et al., 1989; Pedro-Carroll & Alpert-Gillis, 1997) indicated that pretest as well as posttest results of the 4- to 6-year-olds in the Dutch sample were statistically lower than that of children participating in CODIP in the United States (U.S. program; p < .05). This is shown in Figure 3. In the United States, CODIP had resulted in a medium-sized pretest to posttest difference in PEF scores (d = 0.61). The magnitude of the increase in PEF scores between pretest and posttest as found for 4- to 6-year-olds in the current study (d = 0.43) is smaller than that previously found in the U.S. CODIP participants but outweighs that of children in the two other U.S. comparison groups (i.e., intact controls, d = −0.01; divorce controls, d = −0.14). For 6- to 8-year-old children, the noteworthy (d = 0.56) positive difference in mother-reported PEF child functioning between pretest and posttest found in the current study is—in terms of magnitude—comparable to that found in previous U.S. research (d = 0.64; compared to d = 0.06 and 0.12 for U.S. intact and divorce controls, respectively).

OPEN IN VIEWER

Figure 3.

Average pretest and posttest Parent Evaluation Form scores of CODIP-Netherlands pilot study participants (mother reports) and U.S. Program (CODIP) participants (Alpert-Gillis et al., 1989; Pedro-Carroll & Alpert-Gillis, 1997). CODIP = Children of Divorce Intervention Program.

Comparison of group leader–reported child functioning results (the GLEF overall scores) of the current study with those of previous U.S. research (Alpert-Gillis et al., 1989; Pedro-Carroll & Alpert-Gillis, 1997; see Figure 4) shows that children in our study had a significantly higher baseline level than American children participating in CODIP ( p < .05). However, at the posttest, the scores of children in the current study resembled the U.S. scores (4- to 6-year-olds, p = .46; 6- to 8-year-olds, p = .33). So despite the (medium-sized) increase in adjustment scores in the current study (4- to 6-year-olds, d = 0.61; 6- to 8-year-olds, d = 0.58), group leader–reported progress of the American children participating in CODIP was larger (4- to 6-year-olds, d = 1.53; 6- to 8-year-olds, d = 1.44).

OPEN IN VIEWER

Figure 4.

Average pretest and posttest Group Leader Evaluation Form scores of CODIP-Netherlands pilot study participants and U.S. Program (CODIP) participants (Alpert-Gillis et al., 1989; Pedro-Carroll & Alpert-Gillis, 1997). CODIP = Children of Divorce Intervention Program.

Discussion and Applications to Practice

The aim of this study was (1) to translate and culturally adapt CODIP for 4- to 8-year-old children of divorce in the Dutch setting, (2) to test the feasibility of these adapted CODIP-NL modules in Dutch practice, and (3) to compare the effects of CODIP-NL to those found in the United States. This pilot study showed that the intervention could indeed be adapted to the Dutch setting in an acceptable format. Intermediate (i.e., group leaders) as well as end users (mothers and children) were satisfied with the intervention program. In addition, we found modest effects based on all informants (mother, group leader, and teacher reports), showing that participation in this intervention leads to higher child adjustment scores and lower child behavioral problems for children dealing with divorce or separation. Effects were more modest than those previously found in U.S. research.

In this study, it appeared to be useful to start an intervention translation and adaptation process with the literal translations of written intervention materials. In case of some more autonomous intervention materials or activities, such as copyrighted reading books, it seemed to be a suitable option to search for readily available alternatives in the new context. The usage of the book Julia Has Two Houses in the current study is a good example of such an alternative.

Based on our results, we suggest to always develop a conceptual outline with the core elements that constitute the effectiveness of the original program, before culturally adapting the translated materials with adherence to this outline. This article describes sample adaptations to detail nature and extent of changes made to ensure relevance of the program to the Dutch context. It was always guaranteed that these adhered to the conceptual outline of CODIP. Intermediate users were found to be capable of accomplishing most of the intervention’s goals and sessions’ specific goals, using the adapted versions.

We stress two main points to be addressed by researchers interested in culturally adapting these type of research-based interventions to another cultural context. First, one should think about the most suitable intermediate users in the new context. In our case, these were school social workers. Do these intermediate users have the skills to apply the intervention’s main techniques, and are they familiar with the target population? To come back to this study’s example, these professionals were already experienced in working with structured intervention programs within groups of young children. However, it appeared to be needed to pay extra attention to puppet play in training and procedure manuals, as the usage of puppet play is not very common for these Dutch professionals.

A second point of attention concerns the implementation requirements, focusing on how the intervention will fit in daily practice. For example, it appeared to be not feasible to fit a 15-session CODIP 6- to 8-year curriculum into Dutch daily practice. The original U.S. CODIP module for early latency children was therefore condensed from 15 to 12 sessions. Adherence to the conceptual outline and consultation of the intervention developer, in order not to lose essential intervention elements, was of utmost importance in this adaptation process.

Regarding the feasibility of finding positive and desired effects of CODIP-NL, outcomes of children participating in the adapted CODIP-NL modules indeed improved regarding functioning based on mother, teacher, and group leader reports. However, our results show differences in the outcomes across informants. That is, pretest and posttest child adjustment score, predominantly for the 4- to 6-year-old age-group, differed between informants, despite that the evaluations of process and user satisfaction were very much alike. It is not fully clear how to interpret these differences. Other studies have also shown that perspective on child functioning may differ across informants (e.g., Miller, Martinez, Shumka, & Baker, 2014). Moreover, Goodman, Ford, Simmons, Gatward, and Meltzer (2000) concluded that when using the SDQ to screen for child mental disorder, parents and school teachers provide information of roughly equal predictive value, although their relative value depends on the type of disorder. More specifically, they concluded the information from parents to be slightly more useful for detecting emotional disorders, while information from teachers was so for detecting conduct and hyperactivity disorders (Goodman, Ford, Simmons, Gatward, & Meltzer, 2000).

However, it must be noted that complete mother-reported data in our study were only available on 11 (65%) of the children participating in CODIP 4–6 years. This may have biased our findings. A closer look at children for whom pretest data were available, but posttest data lacked (n = 3), showed that these children had all scored in the clinical range of the SDQ. It could be that children dealing with the most (severe) problems in their daily lives were underrepresented in the final results based on maternal reports of 4- to 6-year-old participants.

In this study, we compared our results with previous findings in the United States. The standardized effect estimates found for the Dutch participants were smaller than those found in the U.S. research. As no Dutch comparison group was available, it remains unclear if smaller pretest to posttest differences in child adjustment scores found in the Netherlands as compared to the United States reflect reduced effectiveness of the adapted curriculum or, for example, application of the intervention in a different context. Were the Dutch and U.S. target groups comparable at the pretest? Comparison showed that pretest mother-reported 4- to 6-year-olds’ functioning scores were significantly lower than those in the United States. Mother-reported child functioning scores for 6- to 8-year-olds also differed from U.S. scores. Instead, these were very much alike the Dutch scores in the younger age-group. The same held true for group leader–reported functioning scores. Dutch children’s pretest scores were very much alike for the two age-groups. However, these were significantly higher than U.S. pretest scores. Dutch posttest group leader–reported child functioning scores were statistically similar to those in the United States. We hypothesize that the U.S. intervention was possibly implemented in a sample of children that could profit more from the intervention due to different levels of preintervention problems.

Our precise approach in this pilot with an explicit focus on translation and adaption of the initial program is a strength of this study. We purposely chose this design for this first feasibility study. And we chose to invest in a feasibility study before investing in further research and dissemination of what is later found to be a badly fitting program. This pays back the investment once it comes to the actual implementation in Dutch practice. Randomized controlled trials may be premature and thus unnecessary when conducted on interventions that are not yet fully developed or interventions that have yet to be accepted by the actual practice of child welfare services.

This study could provide guidance for researchers interested in adapting other research-based interventions to a different culture. A pilot study as the present one can provide information on intervention process, as well as program satisfaction, attractiveness, and acceptability as experienced by intermediate and end users. Final changes made to the CODIP-NL intervention modules, as are currently implemented in Dutch practice, may have more optimally anticipated possible implementation issues at a later stage.

However, some study limitations should also be noted. The most severe limitations of this study were the fact that a Dutch care as usual control group was lacking, and no independent informants were consulted. As informants in this study knew that these children were participating in an intervention group, they may have been biased toward the expectation that this was helpful to the children. Our study can thus have overestimated the changes in these children. Moreover, we cannot rule out that time alone might have accounted for observed changes. However, some effectiveness seems likely.

Our pilot study was furthermore limited by its low power due to the small sample size. For instance, we did find small-sized differences between pretest and posttest scores, but power was lacking to reach statistical significance. Still, we found significant positive outcomes based on mother, group leader, and teacher reports.

This study has shown that the further introduction and implementation of CODIP in the Netherlands are feasible and that we can replicate some of the positive effects of the module as found in the United States. We found significant positive outcomes based on group leader reports and converging outcomes for mother and teacher reports.

In the Netherlands, a gap still exists regarding research-based interventions for young children. This study shows CODIP-NL to be a promising and much appreciated direction for future intervention in this domain. Based on our findings, we advise social workers in the Netherlands to apply CODIP-NL as a favorable method to support school-going young children of divorce with the aim of contributing to improved child adjustment and lower child behavioral problems. More in general, this study has shown that implementation of CODIP is feasible and promising in settings outside the United States. Adopting the CODIP model may therewith be valuable for social workers outside the United States and in the Netherlands.

Given the promising nature of the current results, evaluating the impact of CODIP-NL in a future controlled trial in the Dutch setting is warranted. Future large-scale controlled research in the Dutch context may not only shed light on intervention effects compared to Dutch care as usual but possibly also on subgroups of children for whom the intervention is most effective and therewith inform participant selection in social work practice. In 2015, a quasi-experimental study into the effects of CODIP-NL for 6- to 8-year-olds was started. Results are to be expected in 2017.