Application of an integrated behaviour-change model on grandparental adherence towards childhood domestic injury prevention in Hong Kong: a longitudinal study

Introduction

Unintentional injury is one of the leading causes of childhood mortality and disability. The annual global estimate of unintentional injury mortality rate is 80.5 per 100 000 children aged under 4, accounting for 522 167 immature deaths.1 Most young children are injured in the home environment.2 Likewise, surveillance reports in Hong Kong showed that 64.5% of unintentional injuries that required medical attention in children aged less than 4 occurred at home3 and that 45.9% of childhood fatal incidents occurred at home.4 Injury risks are largely mitigated by the consistent application of safety precautions such as active supervision and home adaptation.5 It is estimated that at least one-third of childhood mortality due to unintentional injury could be prevented.3 However, these precautions are seldom universally implemented. In Hong Kong, only 66.7% of caregivers adopt general domestic safety precautions for young children.6 The lack of caregivers’ commitment towards childhood domestic injury prevention may be due to their underestimation of injury risks and severity,7 false perception of domestic accidents as an inevitable and normal part of upbringing and unfavourable weighing of potential risk against inconvenience.8 Therefore, this study employed a theory-based approach to investigate the motivational and decision-making processes underlying domestic injury prevention behaviours within a significant population group of caregivers, grandparent caregivers.

The integrated model unites concepts from self-determination theory (SDT) and the theory of planned behaviour (TPB) to formulate an effective explanation of behavioural variances9 and details the psychosocial sequence of behaviour from the distal motivational antecedents to the proximal decision-making process. The distal self-regulatory process within the integrated model is described in SDT, which highlights that social environments that satisfy fundamental psychological needs nurture a more adaptive form of motivation, autonomous motivation.10 In contrast to individuals with controlled motivation, who perform due to external pressure and desire to maintain internal ego (eg, pride or guilt avoidance), those with autonomous motivation engage in behaviour that is aligned with their own interests and goals, or of great personal importance and often obtain more favourable outcomes such as enhanced performance, long-term behavioural adherence and maintenance and better psychological well-being.11 The gap between motivational antecedents and behaviour can be bridged by the social-cognitive constructs and behavioural intentions from TPB. TPB posits that intention formulation is determined by three social-cognitive constructs: (1) attitude (ie, degree of positive or negative appraisal of the behaviour), (2) subjective norms (ie, perceived appropriateness and social pressure to perform the behaviour) and (3) perceived behavioural control (PBC, perception and beliefs in one’s own ability and efficacy to perform the behaviour).12 Individuals who view the behaviour as positive, socially acceptable and easy to perform are more likely to intend to perform the behaviour of interest.

The integrated model has also been used to explain preventive behaviours against various types of unintentional injury, including the prevention of sports13 and occupational14 15 injuries and parental adoption of sun protective behaviours for their children.16 Aside from SDT and TPB models, other health and behavioural models have been used to account for the childhood domestic injury prevention, including the health belief model (which focuses on factors such as perceived severity and susceptibility)17 18 and Hadden model (which focuses on environmental factors),19 and the knowledge, awareness and practice model (which consider knowledge acquisition, awareness of risks and the translation of knowledge into preventive practices).20–22 However, these models do not explicitly outline how caregivers’ motivation and beliefs are related directly and indirectly to their behavioural adherence to domestic injury prevention. The integrated model of SDT and TPB may offer a plausible theoretical explanation in this regard, but no studies so far have investigated the application of this model specifically in the context of domestic injury prevention among infants and toddlers. Therefore, our aim is to address this literature gap and seek a more comprehensive understanding of the underpinning psychosocial and social-cognitive factors that contribute to individuals’ adherence towards domestic injury prevention in early childhood.

The current study examines whether the tenets of the integrated model of SDT and TPB are applicable to explaining grandparents’ injury prevention behaviours in protecting their 0–2-year-old grandchildren from domestic injury. We narrowed the age group of participants’ grandchildren to 0–2 years, as infants and toddlers in this age group have the highest domestic injury incidence and severity compared with older children.2 23 24 Grandparents were chosen as our target participants, as a growing number of grandparents now play an active role as their grandchildren’s main caregiver when both parents join the workforce,25 26 and little research has shed light on the grandparents’ perspectives. As previous studies focusing solely on SDT or the TPB generally supported the universalities of these theories across age groups,27 28 we expected the integrated model could proficiently predicted behavioural adherence in grandparent caregivers. Specifically, the relationships among psychological need support, autonomous motivation and positive outcomes would remain consistent in studies with older adults.27 Additionally, although not all socio-cognitive variables have been found significant in the studies regarding TPB in older adults, PBC appeared to be a consistent factor in the prediction of intention towards health behaviours across these studies.28–30 Therefore, it is plausible that PBC might hold greater weight in intention formation in older population.

Our study employs a two-wave longitudinal design to allow a test of the pathways of the integrated model at the change-score level. Rooted in the framework of the integrated model, the following five hypotheses were formulated:

(H1) Grandparents’ perception of psychological need support provided by their family members would be directly and positively associated with grandparents’ autonomous motivation in childhood domestic injury prevention.

(H2) Grandparents’ autonomous motivation in childhood domestic injury prevention would be directly and positively related to the social-cognitive constructs from TPB.

(H3) Grandparents’ social-cognitive beliefs would form a direct positive association with their intention to prevent childhood domestic injury.

(H4) Grandparents’ intention would be directly and positively associated with their behavioural adherence to prevent childhood domestic injury.

(H5) A positive indirect effect of psychological need support on behavioural adherence to prevent childhood domestic injury would be observed. In other words, social-cognitive constructs and intention serve as mediators in the pathways between psychological need support and behavioural adherence.

Method

Results

All measures exhibited satisfactory internal consistency across two time points (Cronbach’s alpha: 0.90−0.96, Raykov’s rho: 0.87−0.95). The linearity of the regression paths was confirmed with Ramsey’s Regression Equation Specification Error Test (RESET) and fitted versus residuals plots. Some demographic variables were found to be significantly correlated with the study variables. These include the number of days grandparents spent with children per week, number of hours grandparents spent with children per day, grandparents’ educational level, household income, the number of children in the household and the gender of the grandchildren. As such, they were added into the model as covariates to statistically account for their potential confounding effects on the model pathways. Table 1 displays the descriptive statistics, score reliability coefficients and zero-order correlations of the study variables.

The proposed model exhibited an acceptable fit with the data: χ²=905.09 (df=531), CFI=0.94, TLI=0.93, RMSEA=0.05 (90% CI=0.05 to.06) and SRMR=0.08. For alternative models, although the SDT-only model, TPB-only model and reverse-effect (TPB→SDT) model were found to have satisfactory fit (CFI>0.90, TLI>0.90, RMSEA<0.08 and SRMR<0.08; see table 2), these alternative models fitted significantly better than the proposed model (ΔCFI>0.01). However, the indirect effects of the mediation pathways were mostly supported in our proposed model (p<0.05 for 14 out of 18 mediation pathways), but not in the reversed-effect model (p<0.05 for 6 out of 18 mediation pathways, see the indirect effects of the reversed-effect model in online supplemental appendix C). Therefore, our proposed integrated model was chosen as the final model since it offered a more robust and comprehensive explanation, encompassing the motivational antecedents to behavioural actualisation. Figure 1 presents the standardised parameter estimates (β) of the direct pathways in the proposed integrated model (see online supplemental appendix B for unadjusted estimates).

Request permission

Path estimates fully supported our hypotheses. As proposed, we found a positive and significant effect of psychological need support on autonomous motivation. In turn, autonomous motivation was positively and significantly associated with subjective norms, PBC and attitude. Consistent with our hypothesis, these three social-cognitive beliefs were found to be positively and significantly associated with intention. Finally, in support of H4, there was a positive and significant relationship between intention and behavioural adherence to childhood domestic injury prevention.

The parameter estimates for the indirect pathways partially supported our hypothesis (H5, summarised in table 3). The positive and significant indirect effects of psychological need support and autonomous motivation on behavioural adherence were found to be mediated by subjective norms, PBC and intention. Contrary to our expectation, attitude did not serve as a significant mediator in the indirect paths.

Discussion

This study was the first to examine the motivational and reasoning processes underlying grandparents’ adoption of domestic injury prevention for their grandchildren. The results demonstrated significant direct and indirect effects of psychological need support, autonomous motivation, and social-cognitive factors on grandparents’ intentions and behavioural adherence to childhood domestic injury prevention and mostly supported the hypotheses derived from the integrated model.39 Interestingly, our findings showed that the constructs of SDT and TPB may have varying impacts on childhood domestic injury preventive behaviour. Identifying these key factors might provide useful insights for building behaviour-change interventions for grandparent caregivers.

Consistent with the distal motivational regulation of behaviour specified by SDT,10 we found that psychological need support from other family members and grandparents’ autonomous motivation in childhood domestic injury prevention were positively and significantly associated. The results confirm our hypothesis (H1) and are comparable to the findings in sports and occupational injury prevention contexts.13 15 The large amount of variance in autonomous motivation (60%) explained by psychological need support underscores the importance of providing grandparents with a need-supportive climate within the childcare setting. Unlike parental caregivers, grandparental caregivers often navigate a delicate balance between maintaining their own autonomy and satisfying the preferences of the parents.40 Therefore, need-supportive climate may help to mitigate conflicts and avoid compromises on the part of grandparents. In this context, need-supportive behaviours consist of including grandparents in the discussion of required safety measures, providing rationales behind the necessary precautions and acknowledging and valuing grandparents’ experience and effort in domestic injury prevention.

Autonomous motivation promoted social-cognitive beliefs, which in turn resulted in increased domestic injury prevention intention and behaviour. This confirms our hypotheses (H2–H4) and fully supports the use of the integrated model for explaining childhood domestic injury prevention behaviour. Again, the results are consistent with the findings in prior injury prevention research.13 15 Interestingly, PBC had the largest effect on the explanation of intention. One possible explanation for this result is the influence of injury attribution on injury prevention behaviour. Given that PBC is conceptualised as a measure of perceived controllability and self-efficacy,41 our findings are in line with the findings of the previous correlational studies which suggested that individuals’ injury prevention practices, such as higher levels of supervision, fewer risky behaviours and safer home environments, have been shown to be associated with their perceived controllability and self-efficacy.42 43 Grandparents who are fatalistic and perceive that domestic injuries are uncontrollable are likely to possess a lower intention to prevent domestic injury. Additionally, the large effect of PBC on intention may be attributed to grandparents’ concerns about their physical ability to prevent such injuries. As functional ability was shown to be a determinant of effort-exerting activity in older adults,44 whether grandparents perceive limitations in their physical capabilities can affect their sense of control and self-efficacy in preventing injuries.

The full mediation effects of psychological need support and autonomous motivation on behavioural intention and adherence were found to be significant only through subjective norms and PBC. Although the results only partially supported our hypothesis (H5), they demonstrate that social (ie, subjective norms) and control (ie, PBC) beliefs are necessary for translating autonomous motivation into intention and behaviour. This finding corroborates the idea that autonomously motivated individuals align their beliefs to better form and carry out their intentions. Notably, the SDT constructs exerted their positive effects on grandparents’ engagement in childhood domestic injury prevention predominantly through subjective norms. Indeed, social norms have been shown to be a strong predictor of how caregivers make health-related decisions for their children.45 46 The findings could be attributed to the phenomenon of individuals internalising the beliefs and expectations of significant others. Individuals then conform to such social influences, as they perceived these social norms as beneficial to their own goals and autonomy support.39 In a need-supportive environment, grandparents are motivated not only by their own desire to safeguard their grandchildren but also by realising parents’ expectations of prioritising safety and upholding certain supervision practices. With these normative beliefs, grandparents are more likely to have stronger intentions and behavioural adherence towards domestic injury prevention. This is particularly relevant to Asian societies, where supportive grandparenting is a culturally valued intergenerational expectation. In contrast, grandparents in Western societies may harbour less familial and social expectations and responsibilities for the provision of childcare.47 The different caring roles may have motivational and practical implications for domestic injury prevention. Future studies can investigate the reciprocal relationships between autonomous motivation and the socio-cognitive beliefs between parents and grandparents to better understand the role of significant others on home safety for families with young children.

Altogether, our study offers original evidence for the application of the integrated model in predicting childhood domestic injury prevention provided by grandparents. However, some potential limitations should be noted. First, we contacted potential participants via social media platforms, such as parenting chat rooms. This online recruitment procedure might have introduced selection bias, as only parents active in parenting groups were exposed to the recruitment. This might explain why our sample was over-represented by female and younger parent participants and could potentially be more influenced by social norms.48 However, as we asked the parents to relay the questionnaire to the grandparent caregivers, we hoped to minimise the effect of this sampling bias. Second, while well-established scales with high reliability and validity were used to assess the primary constructs in SDT and TPB, self-reported measures are still subjected to social desirability and response bias.49 In particular, it was found that caregivers tended to over-report their safety practices towards their children.50 More objective approaches, for example, home visits and behavioural observation,51 can be used to evaluate grandparents’ behavioural adherence to domestic injury prevention. Nonetheless, such auditing would be difficult to execute due to privacy and health concerns. Third, the longitudinal design of the current study renders the data correlational in nature. Hence, we could not draw causal conclusions regarding the relationships between variables from the main effects. Future research might consider adopting an experimental or interventional design to generate higher levels of evidence. Fourth, although we controlled for some potential confounding demographic and caregiver-based variables, other child-based and situational determinants (eg, constraints in the home safety environment and children’s temperament) were not examined.5 Future studies might consider their incorporation into the behavioural model. Finally, it is crucial to acknowledge that our sample comprised homogenous Chinese grandparents from Hong Kong, China. Since different features of grandparenting can be found in different societies, our results may not be generalisable to other regions or cultures.52 Moreover, this study focused on supportive caregiving, but domestic safety is also highly important in settings involving other modes of grandparenting, such as skipped-generation or kinship caregiving.53 To determine the applicability of the integrated model in these settings, replication studies in other countries and target groups are highly warranted.

Conclusion

Despite the growing numbers of grandparent caregivers, there is a dearth of childhood injury prevention research that includes grandparents as target participants. Our results provide initial evidence that a psychological need-supportive climate is essential for promoting autonomous motivation, which in turn relates to better attitude, social and control beliefs that facilitate intention formation and behavioural adherence to childhood domestic injury prevention in grandparent caregivers. Given the importance of autonomous motivation, social norms and PBC in facilitating behavioural adherence, future interventions may focus on promoting familial support to cultivate autonomy in and foster appreciation towards grandparent caregivers, while providing them with necessary assistance in ensuring child safety. These findings support the use of the integrated model to explain domestic injury prevention behaviour and lay the foundation for future theory-based interventional research.