Evaluation of the International Child Development Programme (ICDP) as a community-wide parenting programme

Background Many parenting programmes lack proper evaluation, especially under community-wide implementation. Objective Examining the effectiveness of the eight-week International Child Development Programme (ICDP), implemented as a general programme. Methodology Non-clinical caregivers attending ICDP (N = 141) and a non-attending community comparison group (N = 79) completed questionnaires on parenting, psychosocial functioning, and child difficulties before and after ICDP course. Analyses compare changes in scores for both groups over time. Results The ICDP group showed more positive attitudes towards child management and reported better child management, improved parental strategies and less impact of child difficulties. Caregivers with low initial scores benefited most. The comparison group showed little change with a significant decrease in scores on the caregiver–child activity scale. Discussion The results suggest that caregivers in the community who do not show clinical signs or have children with behaviour or other disorders, may benefit from participating in parent training based on ICDP.

Challengese xperienced in raising children are common and the positive results of parenting programmes with clinical groups may benefit abroader group of caregivers (Rodrigo, Almeida,Spiel, &Koops, 2012)toimprove all aspects of caregiving with possibleo nward benefitso nc hild behaviour and caregiver mental health (Long, 2007;Sanders&Morawska, 2010). One of the few efficacy studies of ag enerally implemented programme found increased positive and reduced negative parenting behaviour (Hahlweg,Heinrichs, Kuschel,Bertram, & Naumann, 2010).
Yet there is al acko fe fficacy research evidence (Rodrigoe ta l., 2012; Scott, 2010). In ar eview of 46 general and targeted interventions,S andler et al. (2011)f ound long-term effects of parenting interventions on child development and behaviour,b ut al ack of explanation of the processes mediating these effects. The effects of programmes designed for specific clinical groups may be easier to explain but may not apply to caregivers with more ordinary challenges.
In order to develop optimal programmes for community-widei mplementation, therei saneed to investigate programmes based on different theoretical foundations and which comprise av ariety of elements. Theoretically, most current parenting programmesare based on social learning theory and behaviour change,e .g., TheI ncredible Years (Webster-Stratton et al., 1988), Parental Management Training (Patterson, 2002)a nd the Triple P-programme ( Sanders, 2008). Few evaluations are conducted in community settings (Hutchings et al., 2011). ICDP is anon-instructivepsychosocial intervention programme directed towards parents and othercaregivers.The programme is well recognized and is used in about 35 countries, including both socially well-functioning societies and societies where political unresta nd war may make parenting especially challenging, and in collaboration with organizations such as Save the Children, Unicef, Care,and WHO. However, there is no evaluation of ICDP implemented as ac ommunity-widep rogramme in as ocially well-functioning society. The present study examined the impact of ICDP courses on ag eneral community sample of caregivers in Norway. The main research questionw as whether participation in an ICDP course would have ap ositive impact on parenting strategies and on how caregivers perceived their children and themselves. The moderating effectso fs elf-efficacy, depressiona nd social support were also investigated.

METHOD
The study used atwo-group design with anatural intervention group ( N ¼ 141) and acomparison group ( N ¼ 79) who both completed questionnaires before and after the intervention group's ICDP course.

The ICDP programme: Content and implementation
The theoretical foundation of ICDP is derived from developmental and humanistic psychology with focus on sensitive adulta djustmenta nd empathy (Hundeide &R ye, 2010). It is non-instructive and aims to guide carers' understandingo ft heir children and interaction with them. Thep hilosophyi s formulated in three dialogues containing eightguidelines:the emotional dialogue (e.g., showing loving feelings, praising and acknowledging the child),t he comprehension dialogue (e.g., supporting the child'sm eaning-making and showing enthusiasm for the child'sexperiences), and the regulativedialogue (e. g., regulating the child's actions step-by-step; Hundeide, 2001;s ee http://icdp. info for details on ICDP).
ICDP courses are offered nationally in Norway by the Ministry of Children, Equality and Social Inclusion through "The Parental GuidanceP rogramme". A filter-down approach is applied where the facilitators become qualified to run caregiver groups,and somefacilitators proceed to become qualified to train new facilitators. Mothers and fathers of children at all ages may participate, but ICDP groups tend to contain parents with children in the pre-school age, and are usually delivered through ICDP educated staff at kindergartens and child health centres. The groups usually consist of 5-10caregivers attending eight weekly two-hour sessions,o ne meeting for each guideline. Caregivers take an active role, participate in group discussions, rolep lay the guidelines,a nd do home assignments, like "Try to follow your child'sl ead. What happens?" The facilitators give positive comments, encourage active involvement and facilitate discussions. Further details at http://www.bufetat.no/foreldrerettleiing/.

Participants
All ICDP facilitators were contacted and loggedforthcoming groups for potential inclusion (Sherr, Skar, Clucas, von Tetzchner,&Hundeide, 2009). Atotal of 75 ICDP groups were approached during the data collection period.The groups were run at kindergartens and health centres with carers recruited through open billboard information, staff advertisemento ri nvitation. At the first meeting, caregivers were informed about the project verbally and in writing.
Ag roup of 269 caregivers completed the first set of questionnaires and 141 (52.4%)c ompleted follow-upq uestionnaires. Ac omparison group ( N ¼ 157) to control for passage of time was recruited from child health centres and kindergartens in areas where the ICDP programme was not implemented, of whom 79 (50.3%) returned the secondq uestionnaire. Thed ata were collected from October2008 to March 2010. TheICDP caregivers had an average age of 36.6 (range 23-60), 2.0 children (range 1-6) and 3.6 peopleinthe home (range 1-6). The focus child was 4.0 yearso na verage ( SD ¼ 2.64, range 0.5-16), 66 girls and 59 boys (16 caregivers did not provide information about gender). Caregiversi nt he comparison group had an average age of 34.2 ( SD ¼ 1.83, range 24 -47). They had an average of 1.8 children (range 1-4)and 3.5 peoplein the home (range 1-6). Thef ocus child was 3.3 years ( SD ¼ 1.83, range 0.25-11), 35 were girls and 26 boys (18 caregivers did not provide information about gender). Caregivers in the two groups did not differ significantly on these variables. Caregiversi nt he comparison group were significantly more likely to be married or live with ap artner and to have higher education than the ICDP group but the groups didn ot differ on gender,b eing born in Norway or employment (see Table 1).

Materials
All participants completed aquestionnaire designed to gather information about demographics, social relationships,a nd emotionala nd parenting issues. Measures include self-efficacy, depression, social support, parenting (activities with the child, discipline, household commotion, happiness with partner, parenting strategy, engagement with the child and child management). Child measures include the Strengths and Difficulties questionnaire (see Table 2).

Procedure
Caregiverscompleted the questionnaires at the first meeting. They were asked to complete asecondquestionnaire after the last group meeting or returned it by mail.
Caregiversi nt he comparison group received the first questionnaires at time of consent and the follow-upquestionnaire by mail after the same number of weeks as the ICDP group. If the questionnaire was not returned, one reminder was sent.

Plan of analyses
Chi-squared tests and t -testsw ere used to compare the groups on demographic variables, and to compare thoselost to follow-up on demographic variables and scale scores to examine factorsa ssociated with no follow-up. As ar esult of differencesbetweenthe ICDP group and the comparison in education, the study used 2(Group: ICDP/comparison) £ 2(Education: higher education/not higher education) £ 2(Time of Measurement: before/after) mixed analysis of variance (ANOVA) with repeated measures on time of measurement. Interactions between Group and Timeo fM easurement are reporteda st hese indicate differential changesf or the intervention group and the comparison group and suggesta n effect of the intervention on the outcome. Civilstatus differed betweenthe groups but it was not possiblet oe nter this factor in the main analysis because of the small number of caregivers who were notmarried or with ap artner.
Two 2( Group: ICDP/comparison) £ 2( Education: highere ducation/not higher education) £ 2( Time of Measurement: before/after) multivariate analyses of variance (MANOVAs) were used to study the effect of the intervention on the SDQ subscales (SDQ total difficulties and SDQ prosocial) and the subscales of the engagement scale (emotional engagement and strategic engagement).

Measure Detail
Activities with the child: The Parent-Child Activity Scale (Bigner, 1977) Twenty-five items scored on aLikert scale 1(Never)to5( Always), total scores ranging from 25 to 125. Cronbach's ( a ¼ .88 at first and .92 at second completion Positive discipline: Conflict Tactics Scale (Straus, 1979) Seven items on positive discipline created, e.g., "Praised them for achieving something on their own" and "Told them that you were proud of them". Caregivers indicated how frequently they engaged in the behaviours (0, 1-2,3-10ormore than 10 times). Asummed score was created by adding midpoints for the response categories, ranging from 0to105. Cronbach's ( a ¼ .68 at first and .37 at second completion. (The full Conflict Tactics Scale was used, but the results for the other subscales are not reported due to poor inter-item reliability; Cronbach's ( , .5) Household commotion : The Confusion, Hubbub, and Order Scale (Matheny et al., 1995) Fifteen items, which are scored as true or false, with summed scores ranging from 0to15. Ahigher score represents amore chaotic, disorganized and hurried household. Cronbach's ( a ¼ .73 at first and .73 at second completion Parenting strategy: "Parenting strategies"measured the parental strategies with afocus on the comprehensive dialogue in the ICDP components. The five items loaded on one factor at first completion. The summed score for "Parentingstrategies" ranged from 5to3 0(a ¼ .72 at first and .76 at second completion). Negatively phrased items were reverse coded such that ahigher score was always better Child management: "Child Management" measured child management strategies with a focus on the emotional and regulative dialogue in the ICDP. The scale consists of 22 items scored on aLikert scale from 1(Agree completely )t o5( Completely disagree). Average scores range from 1to5( a ¼ .77 at first and .69 at second completion). Negativelyp hrased items reverse coded, so alower score was always better Engagement with the child: Ten bipolar items to measure key ICDP components, scored in counterbalanced order from 1to7 .Three scales were created: "engagement scale" eight items (e.g., sensitive-insensitive), loading on one factor at first completion in aprincipal components analysis (PCA); "emotional engagement scale" six items (e.g., loving-unloving), loading on one factor at second completion, and a"strategic engagement scale" three items (e.g., rewarding-punitive), loading on one factor at second completion. One item (strict-lenient) did not load on any of the two factors at second completion. Three mean scores generate dranging from 1to7 .For "engagement scale" Cronbach's ( continued) In order to studywhether the intervention had adifferentialeffect accordingto carer self-efficacy, depression or social support, moderation analyses were conducted. Amedian split was used to categorize carers into low and high selfefficacy, low and high depression and low and high satisfaction with social support. This factor was added to the 2( Group: ICDP/comparison) £ 2 (Education: highereducation/not highereducation) £ 2(Time of Measurement: before/after) mixed ANOVA (or MANOVA).

Ethical considerations
The study was approved by the Regional Committee for Medical and Health Research Ethics and the NorwegianS ocial Science Data Services. Therapist referral was available but no participants expressedaneed for such ac ontact.  (Spanier, 1976) Avisual analogue scale scored from 0( Extremely unhappy)t o6 ( Perfectlyhappy)w as utilized Strength and Difficulties Questionnaire: (SDQ; Goodman, 1999)F ive subscales (emotional symptoms, conduct problems, hyperactivity, peer problems, prosocial) generating three scores (1) total difficulties( 0to40; ( a ¼ .73 at first and .74 at second completion), (2) prosocial scale score (0 to 10; ( a ¼ .75 at first and .80 at second completion), and (3) an impact score derived from questions on overall distress and social impairmentranging from 0to1 0 Depression: The Hospital Anxiety and Depression Scale (HADS; Zigmond &S naith, 1983) Consists of seven anxiety and seven depression items, which are scored from 0(Not at all)to3( Very often, most of the time, definitely, very much), giving asummeds core for depression ranging from 0to21( a ¼ .69 at first completion) Social support: The Social Support Questionnaire -S hort Form (SSQ6; Sarason et al., 1987) Two scores were generated from this scale: atotal number of social supports score ranging from 0to9( a ¼ .93 at first completion) and asatisfaction with social supports score ranging from 1to6 ( a ¼ .91 at first completion)
Linear regression analyses showed no significant relationship between the number of sessions attended and change scores between first and second completion of questionnaires, whena djusting and whenn ot adjusting for the number of meetings held.

DISCUSSION
The results suggestt hat the ICDP may be effective in promoting positive parenting in ag eneral community sample.T here wereac onsistent number of positivea nd significant effects of the programme including parenting strategies and attitudest owards child management and perceived ability to manage the child, and social impairment resulting from child difficulties. The lackofsimilar changesinthe comparison group indicates that the results cannot be explained by the passage of time. Several measures converged, suggesting that ICDP also may have ap ositive effect on the caregivers' evaluation of child difficulties, use of positived iscipline, and engagement with the child, but these trends were not significant in the interaction analysis. The comparison group scored lower on activities with the child at follow-up, whereas the ICDP group did not change. Participation in the study resulted in increased awareness of parentinginfluences among the participants. Both groups may have becomem ore concerned about their parenting but only the caregivers in the comparison group scored lower because they did not receive the support that the ICDP group benefited from through the course. Another hypothesis is that participation in the ICDP course prevented some of the usual stressors on family life that may imply fewer caregiver -child interactions.
It is noteworthy that participants with lowinitial scores in particular seemed to benefit from participation in thep rogramme.C aregiversm ay have different reasonsf or taking part in thep rogramme.S omec aregiversm ay look fora parentingprogramme becausetheyare struggling with everyday childrearing and arelacking useful strategies,because it hasbeensuggestedbyafriend or ateacher in thekindergarten,ortheymay be referred to theprogramme by social services. Thef actt hatt here were more single parentsi nt he ICDP groupm ay reflect the increasedstrainofchildrearingwithone rather than twoparents.Other studiesof parentingp rogrammesa lsor eportarelatively largep roportiono fs inglep arents (Almeida et al., 2012). It is in line with theaimsofthe programmethatparents who seem to struggle most,showthe greatest positive change.Other caregivers might have good self-confidence ande xperiencef ew problems with theirc hildren, and attend thep rogramme becauset heyw anta ll thek nowledge they cang et.T his groupshows less change becauseparticipation in theprogramme is consolidating theirexistingattitudes anduse of strategies rather than initiating change.
One result of the broad focus of the ICDP approach is differential effect on various subgroups. There was ad ecrease in commotion in the home following ICDP participation but this was only significant for caregivers with higher education. Other studies of parenting programmes have alsof ound that parents with lower education may benefit less than more educated parents (Almeida et al., 2012;F ossum,D rugli, Handegå rd, &M ørch, 2010), but in this study the moderating effect of education applied only to this domain. Participation in the programme seemed to lead to alarger positive change in attitudestowards child management and perceived ability to manage the child in caregivers with relatively higher depression scores on HADS.C aregiversw ho expressed less satisfaction with the social supportthey receivedshowed greater improvement in parenting strategies and lower commotion in the home than caregivers who were more satisfiedwith their social support. Several studies of parenting programmes have focused on caregivers who are clinically depressedo ra ts evere social risk (Boyd &G illham,2 009;L aw, Plunkett, Taylor, &G unning, 2009). Acommunity-wideprogramme may contribute to positive parenting in mothers and fatherswith subclinical depression who are not usually referred to the mental health services. Several authors have pointed to an eed for generally available parenting programmes (Sanders &Morawska, 2010;Shapiro, Prinz, &Sanders, 2008)s uch as this programme, which appears to reachc aregivers in need of supporta nd advice who may experience the task of raising children as manageable but challenging, but have no diagnosis or clear clinical challenge.
Thereare some limitationswiththe presentstudy.All children in both groups were of preschoolage,yet therewas avariation in theages. Therewereinsufficient subgroupstobreak down theresults by age, butthismay be an importantfactorin subsequent studies. Anotherlimitationwas thefactthatthere were some significant differencesbetween theICDPand comparison groupatbaselineand hencetheydo notnecessarily representthe same population.The comparison groupwas recruited from health centresa nd kindergartensw here ICDP coursesw eren ot available, whilet he ICDP groups eems to have made some self-initiated or other-initiated effortst ofollowthe programme. This meansthatthe ICDP grouptosomeextent mighth aveb eenb iaseda nd have hadm orem otivationa nd potentialf or change than thecomparisongroup.However,thisisnot only aweakness. Most evaluations areofprogrammesimplemented by research institutions(Rodrigoetal.,2012).The presents tudy givese vidence that ICDP (oro ther generalp arenting programmes) may reach the group it is aiming at and function in community-wide implementation.O ther limitationsa re attendance andl osst of ollow-up by approximatelyh alfo ft he participants,w hich mayh aves kewedt he resultsf or successful participants;t hose with more time,s upport or less depression being overrepresenteda tf ollow-up.B aselinec omparisons showed that thosew ith partners andl ower depression scores were more likely to respond. Question omission indicatedbythe variationinNmayhaveresultedinreduced numbersand diminished power. Finally, cautionshouldalsobeexerted as aresultofthe multiple F-testsconducted (Bakan, 1969)and self-reportedmeasures.
Ap re-investigation was conducted to address the quality of implementation, and the relationship between number of meetings attendedand change scores was addressed. Futurei nvestigationss houlda ddress the relationship between implementation quality and programme effects more specifically. There is still al ack of knowledge about mediating processesa nd the interaction between programme features and child and parentc haracteristics ( Deković ,S toltz, Schuiringa, Manders, &A sscher,2 012;L aw et al., 2009). Some programme features may benefit mostc aregivers while other features may be beneficial for parents and children with particularc haracteristics, even if clinical groups are excluded. Mechanisms accounting for change are not fully understood and may include group conversation, regulation without strictc ontrol or new skills. A larger randomized controlled trial could shed morel ight on this, as would additional observations (Davé ,Nazareth, Senior, &Sherr, 2008), and longerterm outcomes (Sandler et al., 2011).
The results of this and otherstudies indicatethat caregivers may benefit from participating in parenting programmes, including caregivers (and children) without clinical conditions. The results supportt he call for community-wide implementations of ICDP and other parenting programmesi ns pite of the heterogeneous nature of the population (Rodrigo et al., 2012). The basic philosophy of ICDP, with afocus on positiveemotion and regulation, rather than on control, which is more apparent in parenting programmes for parents who have children with behaviour disorders, may resonatewell in manyparents who experience the ordinarychallenges of everyday child-rearing, and who may not need or feel comfortable with am ore controlling approach.