Online positive parenting programme for promoting parenting competencies and skills: randomised controlled trial

Positive parenting programmes (PPP), albeit effective, are not readily accessible to the general public, particularly during the COVID-19 pandemic. In 103 healthy caregiver-child dyads, we investigated the effectiveness of online PPP on parenting sense of competencies (primary outcome), parenting styles and behavioural concerns of children aged 3–6 years (secondary outcomes) between 2 blinded, parallel groups. After block of 4 randomisations, intervention group (n = 52) attended live, group-based, internet delivered PPP while both intervention and active control group (n = 51) received weekly general education via communication application. Outcomes were measured at baseline, 8 and 14 weeks. Most parents from both groups had high education and household income. From the intervention group, 87.5% of the parents attended live sessions while 8.6% subsequently watched recorded sessions. At 14 weeks, the intervention group reported higher sense of competence (Wald 9.63, p = 0.008); both groups reported using more authoritative parenting style (Wald 15.52, p ≤ 0.001) from Generalised Estimating Equations model. Compared to baseline, both groups had significant reduction of children’s emotional problems at 14 weeks (mean change: Intervention = − 0.44, p = 0.033; Control = − 0.30, p = 0.046) and behavioural problems over time (Wald 7.07, p = 0.029). Online PPP offered an easily accessible, primary preventive measure to mitigate behavioural concerns and improve parental competency. Clinical trial registration Thai Clinical Trials Registry; https://www.thaiclinicaltrials.org/; TCTR20201030001; October 30, 2020.

www.nature.com/scientificreports/ and Difficulties Questionnaire (SDQ) defined as having total difficulties score of ≤ 18 11 and no mental health problems, and (6) the primary caregiver had access to a telephone, computer, and an internet connection in their home. Participants were excluded if: (1) their child had history of developmental delay or severe behavioural problems, or (2) their child was not able to attend in-person developmental evaluation at King Chulalongkorn Memorial Hospital. Study procedures. Eligible participants were randomised into two groups: the Online PPP intervention group and the active comparator control group; separate LINE (Line Corporation, Tokyo, Japan) OpenChat groups were created for each group whereby participants could not directly contact other participants. The participant flowchart (CONSORT diagram) is presented in Fig. 1.
At 8 and 14 weeks after the start of the intervention, both groups of participants received a message through a communication application (LINE OpenChat) to complete the follow-up assessments. A reminder message was sent to alert participants if they had not completed the assessments. A follow-up telephone call was made to the participant if the assessment was not completed after 2 reminders. Project participants were given 500 Baht (equivalent to £10.96) per person for participating in the project and completing the assessment at every time point.
Interventions. Between October to December 2020, participants in both the Online PPP intervention group and the active comparator group received similar weekly general parenting education via group communication application (LINE OpenChat groups) for 8 weeks in the forms of text articles or videos.
In the Online PPP intervention group, intensive video conferencing via Zoom platform (Zoom Video Communications, Inc., San Jose, California, USA) was conducted once a week, with the scheduled date, time, and link, and the reminder for each session's attendance sent via the LINE OpenChat. Each video conference session took 90-120 min and utilised transformative learning theory. As the sessions were held in many formats, some sessions were very interactive with many small breakout groups with trained facilitators where the parents practiced communication skills in different case scenarios. Some sessions started with a designated speaker discussing storytelling or planning daily routines and ended with a large group discussion. The entire intervention sample attended the same sequential group session spanning 8 weeks. The content included empowering parents, shifting the focus back to building a good parent-child relationship, fostering positive and effective communication, mastering behaviour modification techniques, organising daily life routine, supporting self-care of parents, promoting development through play and storytelling (Supplementary Table 1). Recorded videos of the live session were edited and available for parents who were unable to attend each week's activity. Furthermore, participants in the intervention group were able to ask specific questions about learning contents and their concerns regarding their children's behaviours via Line OpenChat during the intervention period, whereas such concerns raised by participants in the control group were not responded to until study completion at 14 weeks. In addition, the intervention group also received weekly video clips on general education, which were also provided to the active comparator group. www.nature.com/scientificreports/ Active comparator group. The active comparator group received succinct, weekly video clips on general education. General education topics consisted of bilingualism, appropriate sleep for pre-schoolers, appropriate learning environment, fine motor skills, self-help skills, feeding difficulty, electronic media use in children and bullying. All these video clips were sent via Line communication application. Such above weekly video clips on general education were also sent to the intervention group.
Measures. At baseline, we collected demographic data including parents' sex, age, marital status, and education level. Data on household income, child's age, and sex were also recorded. Primary outcome was parental competence and secondary outcomes were parenting skills, parental stress and child's behavioural problems assessed at baseline, 8 weeks (end of intervention phase), and 14 weeks after the intervention began.
Outcome measures. Parenting sense of competence. The Parenting Sense of Competence Scale Thai Version (PSOC), a validated questionnaire, consisted of 17 self-report and self-administered items assessing parent's perception of their own parenting performance 16,36 . To date, the PSOC tool was designed for parents of children aged 0-17 years old 36 13 . Possible scores ranged between 18 and 90 with higher scores indicating higher parenting stress level. The Thai version of the Parenting Stress Index-Short Form (PSI-SF) was a self-report, 36-item questionnaire assessing both parenting stress and particularly the stress in child-parent relationship with Cronbach's alpha coefficient of 0.94 34 . The questionnaire was divided into 3 areas as follows: (1) Parental Distress (PD): 12 items, (2) Parent-Child Dysfunctional Interaction (P-DI): 12 items, and (3) Difficult Child (DC): 12 items. Responses were based on a 5-point Likert scale with total score higher than the 85th percentile or a raw score ≥ 86 points representing clinically significant parenting stress level.
Child's behavioural concerns. Strengths and Difficulties Questionnaire (SDQ) was a highly reliable, 25-item, 3-point Likert scale questionnaire assessing behaviours of children aged 4-16 years 11 . Higher scores for conduct problems, hyperactivity, emotional problems, and peer problems, and total difficulties, sum of the 4 aforementioned domains, indicated greater behavioural problems.

Statistical analyses.
Baseline data from all randomised participants were summarised by treatment group.
Continuous variables are presented as means and standard deviations and categorical variables as frequency and percentage. Primary and secondary outcomes were summarised descriptively by treatment group at each time point. The primary and secondary outcomes analyses used the intention-to-treat (ITT) population that included all randomised participants. All study outcomes within groups at both 8 and 14 weeks were analysed using Paired t-test. Analysis of Variance (ANOVA) was performed to analyse outcomes between groups at 14 weeks. For the outcomes that were not normally distributed and measured at various time points, we used Generalised Estimating Equations (GEE) model to estimate the parameters of a generalised linear regression and determine the interaction between 2 groups over 3 time points. We adjusted potential confounding factors including household income, caregiver's age, education, number of child's siblings, child's age, and gender. Moreover, we applied natural log transformation to adjust the non-normally distributed data for GEE analysis. GEE was computed with mean values, standard error (SE), and Wald test for each independent variable. Statistical significance was defined as p values < 0.05. Statistical analyses were performed using SPSS statistical software version 22 (IBM Inc., Bangkok, Thailand) for windows with the support from Chulalongkorn University.
Ethical consideration. This study was approved by the Institutional Review Boards of the Faculty of Medicine, Chulalongkorn University (IRB 310/63). All procedures were performed in accordance with the International Council on Harmonisation (ICH) Guidance for Industry, E6 Good Clinical Practice: Consolidated Guidance (ICH-E6) to ensure research integrity and protect human subjects.

Results
From 711 interested participants, 314 participants did not meet initial screening inclusion criteria. On further contact of 397 participants via telephone, 211 were excluded due to not meeting eligibility criteria, inability to attend on-site developmental evaluation, or participants declining to participate when more information about the programme was shared over the phone. Then, we sent the final screening form via electronic mail, which included the child development and behavioural screening, and parental stress screening forms to interested participants. Of 108 who passed the final screening, 105 attended the in-person developmental evaluation for their child at King Chulalongkorn Memorial Hospital. Finally, 103 participants met all eligibility criteria and were finally randomised into intervention (n = 52) and control (n = 51) groups (Fig. 1). Detailed demographics were collected at this time as baseline data.
Four participants from the intervention group dropped out of the study. At 8-week follow-up, there were 3 participants from the intervention group, and at 14-week follow-up, there were 2 participants from the intervention group and 1 participant from the control group who did not complete the questionnaire.
Participant characteristics. Of 103 caregiver-child dyads randomised (101 [98.1%] female), 100 main caregivers (97.1%) were mothers, one participant (0.97%) was a father, one (0.97%) was an uncle, and another one (0.97%) main caregiver was a child's cousin. Caregivers' mean age (SD) was approximately 36.3 (4.3) years in the intervention group and 37.2 (3.4) years in the control group. Most parents from both groups had at least a bachelor's degree and had monthly household income of more than 50,000 Baht (equivalent to £1,146.00). The average child age (SD) of intervention and control groups were 54.3 (12.6) and 52.4 (10.7) months, respectively. Other demographic and baseline characteristics of study participants are shown in Table 2.
Programme attendance. From   www.nature.com/scientificreports/ significant baseline differences were found in the primary outcome variables at both 8 and 14 weeks after intervention by using Paired t-test and ANOVA, total PSOC scores significantly increased in GEE model (Wald = 9.63; p = 0.008), indicating that the total PSOC score of intervention group significantly improved compared to the control group over time. Notably, PSOC skill and knowledge subscale score almost demonstrated statistically significant improvement over time (time variable, Wald = 5.99; p = 0.05); however, this subscale did not significantly differ between groups (group variable) and when considering group by time interaction (group × time in the GEE model). At 8 weeks, all PSOC subscale scores had decreased in the intervention group but had increased in the control group but not statistically significant. While PSOC skill/knowledge subscale and total scores at 14 weeks had an increasing trend in the intervention group more than the control group, albeit not statistically significant difference within and between groups as shown in Table 3. Other subscales of PSOC did not significantly improve (Tables 3 and 4).

Secondary outcomes.
Parental outcomes: PSI-SF, PSS and PSDQ. PSI, PSS, and PSDQ total scores at baseline were not statistically different between two groups. Both parental and parenting stress level were relatively low in both groups. Parenting styles in both groups were predominantly authoritative with mean authoritative score (SD) of 4.42 (0.30) in intervention group vs. 4.36 (0.43) in control group. Parenting stress had a slight increasing trend in both intervention and control groups over time, although not statistically significant in Paired t-test, ANOVA, and GEE analysis (Table 3 and Supplementary Table 2). Although permissive subscale and positive parenting score on the PSDQ at 14 weeks demonstrated significant changes from baseline in the control group compared with the intervention group by using Paired t-test, ANOVA analysis showed no significant difference in mean changes between groups in all PSDQ subscales ( Children's outcomes. Compared to mean total difficulties score (SD) of 8.35 (2.56) for the intervention group, total difficulties score for the control group was similar at 8.57 (3.83) at baseline. This demonstrated minimal behavioural concerns for children from both groups. Within groups, mean differences of scores at 8 and 14 weeks compared to baseline showed that there were no significant changes in child's behaviours at 8 weeks. At 14 weeks compared with baseline, there was a significant improvement in children's emotional problems in both groups, with a mean difference of 0.44 (p = 0.033) in the intervention group and 0.30 (p = 0.046) in the control group. Moreover, children from the control group were found to have more attention problems (mean difference 0.40, p = 0.024). There were no significant differences in behavioural problems, as reflected by total difficulties scores, at 14 weeks compared with baseline both within and between groups (Table 3). However, by 14 weeks after the intervention, children's behavioural problems overall decreased in both groups (Wald 7.07, p = 0.029) ( Table 4). More detail of SDQ subscales in GEE model are illustrated in Supplementary Table 3.

Discussion
In this randomised controlled clinical trial of an online positive parenting training programme as a potential preventive intervention for previously healthy parent-child dyads, Online PPP significantly improved the parenting sense of competence, the primary outcome, compared to the active comparator control group over time. For secondary outcomes, both groups showed significantly increased authoritative parenting at 14 weeks and decreased permissive parenting styles over time. Moreover, both groups had significantly reduced children's emotional problems at 14 weeks and overall behavioural problems over time, as reflected by total difficulties scores on the SDQ. The traditional positive parenting programmes such as the Triple P-Positive Parenting Programme 23 and the Korean Parent Training Programme 40 have been shown to have a wide range of desirable child, parent, and family outcomes in both short-and long-term. For example, the programmes reduced emotional and behavioural problems of children, promoted effective parenting, improved parenting satisfaction and efficacy in addition to strengthened parental relationship similar to previous online positive parenting programmes. There is evidence to demonstrate that online positive parenting programmes improved parenting sense of competence and reduced parental mental health problems and also ameliorated children's behavioural and emotional problems 3,20,22,41 . According to a meta-analysis, components of highly effective positive parenting programmes included teaching parental self-control, guiding parents to foster problem-solving skills and self-control in their children, promoting child's development through play, and increasing the quality of time spent together 22 . Consistent with previous traditional and online parenting programmes, Online PPP for the intervention group in this present study focused on building and strengthening parent-child relationship, supporting caregiver's mental health and well-being, promoting effective communication, fostering daily routine and activity schedule, and managing both desirable and undesirable behaviours through live instruction via video conferencing, discussion and exchange of opinions, in addition to role play for some sessions. Small and whole group discussions in our study utilised transformative learning theory. If caregivers were unable to participate in live sessions, they were able to acquire the basic knowledge via an edited video recording. This modality provided the flexibility for parents to access parenting techniques via video recordings even when they may be busy on the specified day of the live sessions. www.nature.com/scientificreports/ In the active comparator control group, general knowledge which did not overlap with the specified Online PPP curriculum was provided via communication application. Furthermore, Online PPP was suitable for facilitating and supporting parents in dealing with their children during the COVID-19 pandemic. Due to the nature of the online platform and the flexibility of the programme, parents could access sessions that they missed via video recordings. This led to the high "attendance" and completion of all sessions (median = 8 sessions/person). Attendance of previous parenting programmes were often reported as 35-50% of all lessons, and only 60% of parents fully attended the programme 22 . Furthermore, the attrition rate was about 33% 20 as compared to our study's attrition rate of 7.69% for the intervention group. Consequently, various online parent training programmes have been developed to address accessibility concerns such as multi-point videoconferencing 42 , web-based online parent training sessions and telephone-based programmes 3 , and tailored programmes utilising text messages 20 . These online positive parenting programmes have high rates of programme satisfaction due to saved cost and travel time in addition to effectiveness as rated by the parents 3,20,[41][42][43] . Programme attendance rate was higher in online parent training programmes with up to 74% of participants who attended all sessions 44 .
In the intervention group, the parenting sense of competency minimally decreased initially at 8 weeks but later greatly improved at 14 weeks compared to the control group over time. We suspected that this was likely due to the nature of transformative learning strategy which may take some time to change behaviours, skills, and attitudes. As the intervention period coincided with the new wave of the COVID-19 outbreak in Thailand leading to lockdown restrictions including public orders to work from home and implementing online schooling to minimise the spread of infection, parents had to spend more time with their children and inevitably had to manage children's problematic behaviours more frequently. Moreover, some families had increased exposure to economic and social stressors that resulted from stay-at-home orders [30][31][32] . These external stressors may, in turn, affect parents' ability to apply lessons learnt for further self-improvement. Thus, this may partially explain the attenuated effect of the intervention at 8 weeks and the lack of change in stress level despite improved parents' sense of competence at 14 weeks. This finding differed from a previous study which showed decreased parental distress after parenting intervention 20 . At 14 weeks, both groups reported more authoritative parenting style and less permissive parenting style over time, possibly due to relatively high positive parenting skills at baseline and participants' high education, and household income which may lend to more support and means to positive child-rearing practices.
This research found that emotional concerns within each group significantly improved at 14 weeks for both intervention and control groups. Furthermore, over time, children's behavioural concerns decreased in both groups. This may in part be explained by improved parenting sense of competence over time in intervention group and increased positive parenting, defined as more authoritative and less permissive parenting, in both groups at 14 weeks in addition to relatively high levels of positive parenting reported at baseline. The decrease in emotional and behavioural concerns as caregivers became more responsive and competent in their own parenting skills exemplified the transactional model and bidirectionality in parent-child relationships. Positive parenting has been found to be associated with desirable behaviours such as lower screen time in children 37 . Additionally, searching for positive parenting information online these days has never been easier, which may mean that the control group may attain such information from other sources as it would be unethical to restrict the control group from navigating other online resources for parenting support. Moreover, those in the control group still received weekly general parenting education via group communication application for 8 weeks in the forms of text articles or videos, which may have increased parental awareness and practice in such topics. As a result, the efficacy of the intervention in reducing behavioural concerns did not differ significantly between the intervention and control groups. This study, along with previous work by Sanders and team, found that both online parent training and self-help workbooks were effective in reducing disruptive behaviours 23 .
The lack of change in stress level as evident on PSS and PSI may be due to the fact that Online PPP included intensive, weekly 1.5-h sessions lasting for 8 weeks with a gentle reminder a day before each scheduled session. As parents from both groups were from relatively high socioeconomic status and education, they may have high expectations on the parenting programme and their own child-rearing practices. As a result, having to discuss and share their own parenting issues, complete homework, be reminded to attend the live sessions, and listen to the sharing by other parents who may be more experienced, may increase parents' personal stress. Additionally, parents had to juggle with the first wave of COVID-19 pandemic in Thailand which may have led to more stress in both groups while the intervention group had to also manage their time to attend weekly 1.5-h sessions which may introduce more stress. There could be other factors affecting parenting stress that we were unable to account for in this study. Finally, PSS and PSI may not be sensitive enough to detect the differences in stress level, especially in the context of an unexpected pandemic. Therefore, we suggest that future studies should use questionnaires that specifically measure the desired outcome and preferably account for specific context to minimise contradictory findings between different measures. From these outcomes, we found that our online PPP benefited both parents and children.
There are some limitations in this study. First, as this research aimed to administer Online PPP as a modality for primary prevention in order to increase parents' sense of competence and minimise overall behavioural concerns, we specifically selected caregivers with low stress and caregivers and children who did not have any chronic physical or mental health illnesses. This may cause selection bias; however, given the aim of testing Online PPP as a primary preventive measure, we decided to define the inclusion criteria as such. Therefore, both primary and secondary outcomes at various time points may not significantly differ. Secondly, the intervention group required direct discussion with the facilitators based on transformative learning theory and some Thai parents might not feel readily comfortable talking and sharing among strangers. As a result, live, interactive sessions may not be suitable for all Thai parents. In addition, due to the virtual, interactive nature of our intervention, completely blinding participants was not possible. Nonetheless, all caregivers were informed that they will receive parent training via one of the two modes, either live video conference or one-way information-sharing www.nature.com/scientificreports/ via communication application. None of the participants knew whether they were in intervention or control group and the desired outcomes of the study. Participants received their personal identification code which served as blinding for the researchers analysing the outcomes. Thirdly, caregivers from our study had relatively high education, household income, and high positive parenting style at baseline. Consequently, this may limit the generalisability of our results. Fourthly, during the study, there was a new wave of COVID-19 outbreak in Thailand which may have led to the lack of apparent reduction of stress levels in parents. Noteworthily, the stress levels did not statistically significantly increase in both groups. Fifthly, children's behavioural problems measured by Strengths and Difficulties Questionnaire were completed by only the main caregiver, which may not be representative of behavioural problems from other perspectives or settings such as classroom teachers. Nevertheless, given that the main caregiver spent the majority of each day with the child during the COVID-19 pandemic, this measurement could serve as reflection of children's overall behaviours. Sixthly, as the participants were followed up to 14 weeks, this precluded us from determining the long-term effectiveness of Online PPP. Finally, even though our study did not include cost-effectiveness analysis as our main outcome, the total cost of hosting Online PPP was estimated to be 4,000 Baht (equivalent to £87.00) per family for the whole course as compared to the control group's cost estimated to be at 800 Baht (equivalent to £19.00) per family. In comparison to the long-term, high annual cost of behavioural and mental health disorders, preventive programmes such as Online PPP may lessen the overall economic burden and the national budget may be better spent investing in building parenting skills. The topics included in Supplementary Table 1 have proven to be effective targets for primary care providers to utilise in virtual clinical settings as tools in increasing parental sense of competence and positive parenting skills and ultimately reducing children's emotional and behavioural concerns. In spite of these limitations, to the best of our knowledge, this research is the first randomised controlled trial examining the effects of online positive parenting programme, designed with the integration of transformative learning theory, on healthy parents and children as compared to general parenting education via a communication application alone in Thailand. Our study further showed that parents had a higher parenting sense of competency after Online PPP and more positive parenting styles in both groups, which ultimately led to less emotional and overall behavioural concerns at 14 weeks.

Conclusion
Online positive parenting programme formats, either through live, interactive video conferencing or one-way communication application, enhanced the access of parenting training and led to more positive parenting and reduced children's behavioural problems. Online PPP specifically increased parents' sense of competence compared to active comparator control group. However, the intensity and duration of live, interactive video conference training sessions may need to be adjusted to better suit the Thai culture. Online PPP offers an innovative positive parenting training model suited for the "new normal" era limited by social distancing requirements and may provide some guidance on further development of online parent training modules aimed to increase access for parents near and far.

Data availability
The datasets generated and analysed during the current study are available from the corresponding author on reasonable request.