In the context of the community health strategy “Health in the Neighbourhoods”13 in Barcelona (Spain), parenting problems were identified as a community problem. After confirming that no parent-training interventions based on a theoretical framework had been previously used in the city, the Parenting Skills Program for families (PSP) was designed and implemented. PSP is based on the “Program Guide to Development of Emotional, Educational, and Parental Competencies”,14 which includes components associated with more effective parent training programs and, in addition, incorporates scientific recommendations.15 The program consists of 11 weekly sessions of 90minutes in which, following written guidelines, 10-14 parents with children 2-17 years old participate. Children only participate in two sessions. The program suggests some different strategies depending on the children's age, but the core components are the same. Parents are grouped according to their children age. The main contents of the program are development stages understanding; attention, respect, recognition and warmth needs; self-esteem and assertiveness; active listening and empathy; effective communication; problem-solving; negotiation and agreements; everyday family organization and boundaries and norms for children's behaviour regulation. The PSP was piloted between 2011 and 2013.11

This is a quasi-experimental study with measures before, immediately after the-intervention and after a follow-up of six months (T0, T1 and T2, respectively), with no control group. It combines quantitative and qualitative methods. The quantitative study aimed to identify the magnitude of changes from the baseline situation, immediately and six months later in the four outcomes examined. The objectives of the qualitative study were descriptive and explanatory. Specifically, the aim was to identify mechanisms related to changes in the four outcomes examined, increasing breadth of understanding of quantitative outcomes.

Professionals involved in the intervention: 1) worked in social, health or educational services or in socio-educational associations related to family's issues; 2) received training on PSP; and 3) implemented the interventions in Catalonia, primarily in Barcelona. Parents selected to participate: 1) had children 2 to 12 years old; 2) spoke Catalan or Spanish; 3) had not participated in other parent-training program; and 4) were recruited by education, health or social services or socio-educational associations. Professionals used different types of recruitment strategies (poster, personal interviews, phone calls…).

A questionnaire with five sections was used in order to collect data from parents about: 1) parental stress, 2) social support, 3) parenting skills, 4) children's negative behaviours, and 5) sociodemographic profile. The four initial sections were based on previously validated scales in similar populations. The first section included the Spanish version of the Parental Stress Scale16 consisting of 12 items distributed in two subscales (baby's rewards [α=0.77] and stressors [α=0.76]) scored on a five point Likert scale (degrees of agreement/disagreement) (points scored between 1-5). Section two included the Duke-UNC Functional Social Support Questionnaire17 consisting of eight items grouped into two subscales (confident [α=0,92] and emotional [α=0,81] social support) scored with a five point Likert temporal scale (points scored between 1-5). Section three used a scale that included 19 items distributed into six dimensions of parental skills, scored using a four point Likert temporal scale (points scored between 0-3), which has previously been translated into Spanish and validated.18 Its subscales were: 1) children's understanding, 2) emotional self-regulation, 3) parenting self-esteem, 4) empathetic and assertive communication, 5) agreements, and 6) behavioural regulation (Cronbach alpha ranged between 0,69 for subscale 6 and 0,89 for subscale 4). The fourth section used the five-item problems behaviours subscale of the Strength and Difficulties Questionnaire (α=0,83)19 where items are scored on a three point Likert scale (points scored between 0-2). The fifth section collected sociodemographic data about sex, age, country of birth, marital status, educational level and employment status.

Twenty-two groups were carried out between 2013 and 2015, 257 parents started the intervention and 216 completed the questionnaire both at baseline and immediately after the intervention (84% retention rate), and 130 of these completed it six months after the intervention (51% retention rate).

The qualitative study included 39 professionals and 34 parents, who participated in seventeen semi-structured interviews and five discussion groups. These instruments were used to collect information to describe the main effects of the PSP and the processes explaining these effects. They were conducted immediately after finishing the PSP sessions by an external researcher in an isolated, comfortable setting. Depending on the number of people, we used focus group scripts (for parents) and group or individual interview scripts (for professionals). Sessions duration were variable, depending on the number of participants (0.5-1.5hours). Scripts were prepared based on a literature review of the subject in which key elements to be examined were identified and were related to quantitative instruments. The questions sought to provide opportunities for participants to present their own experiences (parents) or parents’ experiences (professionals) in processes of participation in PSP. Scripts had seven questions piloted before starting the evaluation study. They were similar for professionals and parents, but worded differently. They focus on: a) the main effects of PSP; b) the mechanisms of changes associated to the four outcomes analysed; and c) the individual or social variables related to those effects.

We analysed information from 12 secondary documents with facilitators notes about the program effects. The parents who participate in the qualitative study were a subsample of those included in the quantitative sub-study. Some of them deny their participation in the qualitative sub-study because they did not have time to attend the focus groups or interviews or because they did not want to share their experiences verbally with an external researcher. Data collection ended when data saturation was reached.

Quantitative and qualitative data were collected and analysed separately and they were triangulated at the discussion phase.

Sociodemographic variables were sex (woman vs. man), age (<30 years old, 30-35 years old, 36-40 years old or>40 years old), origin (born in Spain vs born in other country), marital status (married or cohabiting vs non-cohabiting), educational level (no studies or primary studies, secondary studies or university studies finished) and employment status (unemployed, homemakers, employed or others). Four outcomes were examined: parenting skills, parental stress, social support and children's negative behaviours. For parenting skill variables, results were examined for each dimension separately. Regarding the scales, pre-, immediate-post and 6-month post-intervention item scores for each variable were summed and medians calculated. The bivariate association between sociodemographic and outcome variables were tested with the chi square test. Differences between T0, T1 and T2 were analysed. The percentage of parents that showed any improvement between baseline and immediately after the intervention, and between baseline and six months after the intervention were calculated for the four outcomes. Since the Kolgomorov-Smirnov test confirmed a non-normal distribution of data, medians were compared through the non-parametric Friedman test, carried out with SPSS.20.

The qualitative information was audio recorded and then transcribed. After that, a content analysis performed by an external researcher using a system of categories previously created by one researcher and checked by the other two, based on the quantitative sub-study and on theoretical considerations. The categories were: 1) general effects, 2) positive factors, 3) barriers, 4) parenting skills effects, 5) children's behaviour effects, 6) parental stress effects, 7) social support effects, 8) outcome variables relationships, 9) personal determinants of change, and 10) contextual determinants of change. These categories were the same for each of the informants’ transcriptions. The first step of the analysis was textual, selecting quotes from transcripts of the interviews and discussion groups. The second step was conceptual in order to identify emergent subcategories and the interrelations between them. These were defined when the data had reached saturation. For the qualitative analysis, the Atlas.ti 6.2 program was used. Some quotes were selected and organized according to the emergent subcategories in order to show the main effects of the PSP as well as the processes that could explain them.

Ethical considerations were taken into account in the study. Participants received a document where the research process was explained and granted their authorization by signing an informed consent. Two datasets were created, one with personal information and a code assigned to each participant, and another without personal information and with the code and the different variables. Participants were informed about the study during the implementation of the qualitative study. The research procedure adhered to the Barcelona Public Health Agency and University of Barcelona protocols of ethical considerations for good research practices of the ethical committee of each institution. It followed all legal requirements.

Data came from 22 groups of 10-14 parents with a child 2-12 years old. Most parents were married or cohabiting, almost one third were 36-40 years old; 40% were immigrants, mostly from Latin-American countries; one quarter had primary studies or less and 31% were unemployed. The main reasons for not finalizing the intervention were work-related factors (e.g. change in work schedule or finding a job), family reasons (having a new child or caring for a dependant relative) and change of house. Difficulties to contact participants after six months (because they changed mobile phone number, they didn’t answer to phone call or they move to another city) were the main reason for losses to follow-up and most of them where men and people with secondary educational level. Statistically significant differences between participants in the pre-intervention and in the immediate post-intervention were observed for educational level with people with secondary studies being more likely to drop out. When participants who started the intervention were compared with those followed up at six months, the same pattern for educational level was found and additionally the proportion of men at the follow-up significantly decreased (Table 1). Most participants had one child (between 34%) or two children (between 43%), most of whom were aged between 3 and 5 years; 50% were girls. Professionals (n=39) were 95% women and 51% were community nurses. All of them implemented PSP and had previous experience as facilitators of family groups.

Table 2 presents medians of the evolution of the 216 participants who completed the intervention. Compared with T0, all outcomes significantly improved at T1. Specifically, 76% of the participants improved their parenting skills, 56% reduced children's negative behaviours, 66% decreased parental stress and 61% reported increased social support.

Table 3 presents the results of the follow-up six months after the intervention. All the outcomes significantly improved and the magnitude of the percentages of improvement was similar to that seen at T1 for most outcomes.

Parents and professionals described different types of changes in parents’ parenting skills, stress and social support and in the children's behaviours after the participation in the PSP. Table 4 shows emergent subcategories resulting from parents and professionals’ discourses. These allow us to identify the main changes for each outcome. Different quotes are included to illustrate each main effect of participation in the PSP.

This study has several limitations. The response rate immediately after the intervention (84%) was higher than in some previous studies7 and similar to other studies that have been highlighted as successful because of their high participation.33 However, the higher dropout rate at the six-month follow-up was a limitation. Almost half of the sample was lost to follow-up after six months, again similar to other studies.34 The causes of this dropout were the difficulties to contact or meet with families 6 months later of the intervention (telephone number or living place changes, or they did not answer phone calls). In the future, specific recruitment strategies should be implemented in order to reduce the dropout in this phase of data collection (i.e. incentives or reminder content session). This was an important limitation and the follow-up results should be viewed with caution, because the trend observed in post-intervention results is less robust because of the sample reduction.

Another limitation was that a randomized control trial design with a control group was not used. Therefore, the quantitative outcomes identified should be viewed with caution and considered as an association and not a direct effect caused by the intervention. Although randomized control trial is the most recommended design for effectiveness evaluation, some authors35 consider that the evaluations with no control group should be accepted when an intervention is well developed, meets generally accepted standards, and qualitative methods show the involved mechanisms. Therefore, our qualitative findings helped to reduce this limitation, and contributed to confirm effects caused by PSP.