Our study used a mixed methods approach, to assess the feasibility of a brief universal intervention. The intervention targeted parents of infants and focused on mealtime communication, for the prevention of disordered eating. This approach is a first step in beginning to answer the call to action from numerous experts in the eating disorder field, who advocate for the use of knowledge translation research within healthcare environments (27, 28). Our results provide important insights into parent and CHN perspectives that can be used to inform future intervention research.
Acceptability
Three-quarters of the parents provided qualitative data about the MCM information session, reflecting their desire to provide additional comments. The two main themes generated from the data were “informative and relatable content” and the “engaging presentation”. This positive data from parents, may in part be attributed to the use of the Knowledge to Action Framework (KTA) (29). Using theory to guide prevention intervention and partnering with communities are argued as two ways to improve the effectiveness of interventions (27, 30). We used the KTA framework to guide the larger knowledge translation project, with this current study representing the feasibility evaluation phase. The framework suggests first using a knowledge creation cycle to develop an evidence base and then adapting this knowledge for use in specific environments. We chose to use a co-design approach working with CHNs for this step, asking what strategies they felt correlated with the evidence-based strategies provided [article currently under peer review]. Thus, ensuring the intervention was tailored and relevant for those that would ultimately use it.
Our study found CHNs highly valued the MCM content and resources, with both the qualitative and quantitative data reflecting this result. Our findings are consistent with previous research that implemented an evidenced-based program providing preventive strategies for child health nurses to use with parents of children 2–6 years old (31). This study revealed CHNs valued the content and believed early intervention and prevention of eating problems was a core part of their work in a community child health service (31). Interestingly, one of the recommendations from CHNs in this study was to start having conversations with parents when their children were younger; when they were transitioning to solids was deemed an appropriate time. One potential reason the content of the MCM education session was likely accepted by the CHNs in our study was they perceived it as valuable and worthwhile. In accordance with Rogers’ theory on the diffusion of innovations, this concept is termed “compatibility” and refers to how closely the innovation aligns with user values, past experiences and needs (32). This is a critical concept when it comes to busy clinicians, working in demanding environments. For a new innovations to be used in practice, its integration into the system for sustainable uptake is essential. Supporting families with nutrition and eating advice is part of the role of CHNs, therefore having access to evidence-based tools is vital. Our findings suggest that the MCM program was highly acceptable to parents and nurses for delivery in the Child Health Service setting.
Adequacy
The mean scores for the items related to the MCM Handout acceptability were very high, indicating parents found the handout fit for purpose. Interestingly, the qualitative data revealed some reasons as to “why” they valued the MCM handout. The theme of clear and easy to understand information was highlighted by many parents. The handout was deliberately designed to be a single A4 page of information, a decision adopted directly from insights gained through consulting with CHNs in the co-design of MCM education session [article currently under peer review]. During the co-design, CHNs unanimously expressed their view that a take home resource was required to summarise the MCM session and believed it needed to be “hard copy”, rather than digital. This is in keeping with the literature, which reveals parents are overwhelmed by the vast amounts of child feeding education on the internet (33). Our study complements these findings as the majority of CHNs reported the parent handout being completely appropriate and no mention of digital options was expressed.
When parents were asked which of the four MCM key messages they valued as the most important to use with their child and why, 37% of parents reported “Talk”. Parents reported it encompassed the other messages and conveyed their desire to create mealtime environments that are calm and joyful. It appears from the qualitative data parents are keen to engage and communicate at mealtimes. They reported feeling validated to connect and talk with their infants during mealtimes, rather than “forcing” them to eat more food. The concept of improved self-efficacy also resonated in the responses to “Eat”. The data associated with “Eat” centered on the importance of letting infants decide on how much to eat and many parents highlighted this strategy contrasted with how they were raised. This represents a potentially important shift in breaking the cycle intergenerationally, as direct associations between parents with disordered eating and their use of unhelpful feeding behaviors have been found in the literature (34). This finding underscores the opportunity to lay a positive foundation for lifelong habits around food and eating in the early years (35). Our findings suggest that the parents and nurses found the messages and resources associated with the MCM intervention to be adequate and suited to their needs.
Impact
Our findings related to the impact of the MCM intervention were much less clear than findings related to acceptability and adequacy. Only one of the four sub-scales of the FPSQ-M parent feeding measure had a statistically significant difference across pre- and post-intervention scores, and this was in the unexpected direction. The “Feeding on demand” subscale of the FPSQ-M scores after the intervention were lower than prior, indicating less “Feeding on demand”. We are unsure why this finding was observed but offer some possible explanations. First, the FPSQ-M questionnaire is relatively new and has not been comprehensively validated. It has, for example, not undergone test-retesting or previously been used in an intervention study. It is therefore possible that the decrease in scores we found over time is an effect of the properties of the measure, rather than the impact of the intervention. However, without a control group, we can only speculate about this.
Another explanation for the result may be in the construct of the items. In the mentioned subscale, half of the four items were reverse coded and perhaps the parents may have responded inappropriately to the change in scale structure across items. Further validation studies, especially with mothers of infants in a busy setting as our administration was, is important to rule this out and to ensure the quality of responding among participants. Two further explanations are also plausible. First, it is possible that the measure did not tap into the constructs MCM was designed to change most appropriately. Our intervention was developed with a focus on mealtime communication, while the FPSQ-M responsiveness subscale was designed to measure the “structure” of infant feeding. Thus, perhaps the measure was not an appropriate fit for our intervention. Alternatively, it is possible that MCM had the undesired effect of reducing responsive feeding. We believe this outcome, however, is unlikely, given the high acceptability and adequacy of the intervention outlined above, and the focus of the program on positive mealtime communication, which pre-supposes responsive feeding.
To overcome the limitations of the FSPQM, four items that directly mapped onto knowledge and behaviours the MCM intervention was trying to change, were implemented. The results for these four items again produced unclear evidence of impact. Of the four items, two showed no statistically significant difference over time. Of the remaining two, one showed a significant improvement in the expected direction, but the final item showed a significant difference in the unexpected direction.
Our results showed an unexpected change in enjoyment of meals from pre to post intervention for the parents, with less enjoyment reported post intervention. Again, it is unclear why this result was found and it appears to contrast with the positive findings in the acceptability and adequacy data. We suggest four possible explanations. First, the four MCM items were constructed with two being reverse-coded, hence the result may be erroneous. Second, it is possible that the content of the intervention brought parents’ attention to their engagement or satisfaction with mealtimes in a way that was not salient before the intervention. With time to reflect on the nature of mealtimes, perhaps parents provided a more reflective score than at baseline. If this were the case, previous research supports the notion that mealtimes with infants can be stressful for parents (36), and mealtime enjoyment may be affected by a wide variety of environmental factors, which were not measured in this study (e.g. sleep deprivation, infant tiredness/illness, parental stress, child rejection of foods, etc). Third, recent research has shown parents use a range of feeding behaviours, driven by a variety of parental factors, including mood and these are not static over time (37). It is possible that as infants age and come closer to weaning age, that mealtimes become more difficult or stressful and thus less enjoyable for parents. It is also possible that daily variability in mealtime enjoyment may be very high, and thus ecological momentary assessment with averaging over more administrations would have been a more accurate measure of enjoyment than our pre/post design. Last, it is possible the intervention functioned to decrease parents’ enjoyment of mealtimes, perhaps through perceived pressure for mealtimes to be satisfying and joyful. Perhaps the MCM program increased parental expectations of mealtimes being positive, when the reality is often infants can be fussy, messy and unpredictable.
Our findings suggest that the impact of the MCM intervention on parents feeding behaviours remains unclear and further testing of the resource with robust validated measures is required to better understand this aspect of the program’s feasibility.
Implications for future research
Given the lack of validated tools available to adequately measure food communication between parents and their children, further research is required. Perhaps extending existing child feeding questionnaires to include specific questions relating to the language parents use to describe foods (e.g., “good” vs “bad”) would be advantageous. Such measures would then enable a more robust measure of impact, along with the use of a control group. Additionally, ethnography studies may be useful in broadening our understanding about how parents are communicating about food, within a family mealtime context. There is limited research specifically focused on the content of family mealtime conversations. However, a qualitative study of 150 family groups, conducted by Thomas and colleagues (38), highlighted parents believed it was their job to tell children about the dangers of “fatness” and frequently used negatively framed messages and scare tactics during mealtimes. Hence, the importance of continuing to explore and expand this area of research, as many opportunities exist to further develop interventions to support parents with this essential part of daily life.
Strengths and limitations
The strengths of this study include the triangulation of data through collection of quantitative and qualitative data to assess feasibility of the intervention. Developing interventions with end users and subsequently piloting them in real world conditions is essential for improving effectiveness. The temptation for researchers is to design complex and costly RCT trials only to discover no health service has the capacity or interest in their ongoing delivery. Hence, a strength of our pilot study was the collection of data from parents and CHNs within a health service setting. We acknowledge this is a brief intervention and not aimed at parents with an eating disorder, however it is a cost effective and practical way to provide universal prevention messages. A feasibility study by Sadeh-Sharvit and colleagues (39), examined an intervention aimed at mothers with eating disorders and their spouses, targeted behavioral change in feeding practices, in a small sample (n = 16). Findings revealed improved feeding practices; however, the investment was very extensive, as the program required attendance at 12 group sessions (90minutes each) followed by a further 12 family sessions (1 hour in duration). Given the mean age of the children was 19.6 months, perhaps many of the feeding practices were already established. A further strength of our study was the use of the same facilitator for all groups, increasing the fidelity of the content presented. Additionally, the facilitator was an experienced Paediatric Dietitian able to succinctly provide real life case examples, highlighting the importance of preventive interventions. Despite these strengths, there were several limitations which need to be acknowledged. There was a lack of a control group, therefore causality cannot be assumed from the results of the pre/post data. However, the qualitative data provided consistent themes relating to specific MCM strategies and parents’ willingness to implement the recommended strategies. An additional limitation was the lack of a set progression criteria determined prior to the study. Such criteria would have been beneficial to include, enabling set boundaries for moving to a larger trial or modifying the intervention or measures. Examples include, mean acceptability rating of greater than 80% for content and resources, from both CHNs and parents and less than 10% missing data in surveys.
Another limitation was the use of the FPSQ-M questionnaire (25). While a validated instrument, the items were not directly related to our core content of “mealtime communication” which made it more difficult to draw meaningful outcome conclusions. However, the data provided information on parental feeding behaviors (e.g., feeding on demand, using food to calm) not previously examined in an intervention study targeting infants.