Explaining adherence to contrasted physical activity and nutrition scenarios in post-treatment childhood cancer patients: A cross-sectional study using variables from the Theory of Planned Behavior

Abstract

Children diagnosed with cancer are vulnerable to long-term health issues. Engaging in physical activity (PA) and adopting a healthy diet could mitigate these risks. This study aimed to understand the role of variables from the Theory of Planned Behavior (TPB) in adherence to healthy/unhealthy PA and nutrition scenarios. Through convenience sampling, four ad hoc questionnaires measuring variables from the TPB were completed by 96 parents of children diagnosed with cancer in paper format or via a secure online platform to assess attitude, perceived behavioral control (PBC), subjective norms (SN), and intention. We performed a MANOVA and multiple linear regressions. We found an effect of behavior domain (F(3, 4828.66) = 6.467, p < 0.001, ηp2 = 0.004), and scenario (F(3, 152.86) = 76.495, p < 0.001, ηp2 = 0.600). Intention was a complete intermediary variable between attitude/SN and healthy nutrition. Attitude, PBC, and intention are promising targets for PA and nutrition behaviors. SN should also be targeted for nutrition behaviors.

Keywords

healthy lifestyle cancer family physical activity nutrition psychosocial support

Background

Each year, approximately 400,000 children are diagnosed with cancer worldwide (American Childhood Cancer Organization 2024). In Canada, this yearly incidence ranges from 925 to 1,025 diagnoses for those under 15 years of age (Government of Canada, 2024). Although children diagnosed with cancer are now living longer, with 5-year survival rates of 83% in Canada (Lin et al., 2018), they still face high risks of complications and late effects resulting from their cancer and its treatment (National Research Council, 2003). One interventional avenue to target these cancer-related adverse events is the promotion of favorable health behaviors, such as engaging in physical activity and adopting a healthy diet (National Cancer Institute, 2022). Both of these health behaviors have been demonstrated to have the potential to limit the occurrence of late effects (Beulertz et al., 2016; Rogers and Barr, 2020). Indeed, in childhood cancer survivors, engaging in physical activity has been shown to limit the development of depression, obesity, low bone density, cardiovascular disease, osteoporosis, diabetes, and hypertension (Lemay et al., 2019; Stolley et al., 2010). Having an adequate nutritional status during treatments for childhood cancer has also been associated with increased treatment tolerance, lower susceptibility to infection, and better quality of life (Joffe and Ladas, 2020). The promotion of physical activity and healthy nutritional practices is all the more important when considering that children diagnosed with cancer are at a high risk of physical deconditioning and of developing nutritional deficiencies both in the short- and long-term (Ho et al., 2021; Joffe and Ladas, 2020).

To effectively promote favorable health behaviors, it is necessary to understand how and why children diagnosed with cancer adhere to these behaviors. To do so, many theories can be used, including for example, Albert Bandura’s Social Cognitive Theory, which places an emphasis on the importance of the interaction between the individual, their environment, and their behavior (Bandura, 2002); or Prochaska and DiClemente’s Stages of Change Theory, which focuses on decision-making to change a behavior based on different stages, ranging from pre-contemplation, to action and maintenance (Prochaska and Velicer, 1997). Although these models are well-established and can be useful for understanding the principles, theories and methods of behavioral science, given the scarcity of studies trying to model healthy behavioral factors in the context of pediatric cancer, the Theory of Planned Behavior (TPB) was chosen as the theoretical anchor of this study. This theory was selected since it allows to distinguish between the different factors that can contribute to the adoption or rejection of health behaviors. The TPB also emerges as one of the most commonly used models in behavioral medicine (Ajzen, 1985). The application of the TPB to understand physical activity and nutrition behaviors has been synthesized in reviews in other populations (Godin and Kok, 1996; Kirk and Haegele, 2019; Riebl et al., 2015). Although no review on the subject exists in children diagnosed with cancer to date, the TPB has been used in cross-sectional studies to describe health behavior changes related to physical activity in children with cancer and health behavior changes related to nutritional practices in healthy adults (Brouwer and Mosack, 2015; Caru et al., 2021). According to this model, three types of beliefs guide an individual’s intention to engage in a given behavior (Ajzen, 2019). First, behavioral beliefs refer to an individual’s subjective evaluation of the probability that the behavior will result in the desired outcome. These beliefs impact the individual’s attitude towards the behavior. Second, normative beliefs refer to an individual’s appreciation of whether people who are important to them will encourage the behavior. These beliefs impact the individual’s subjective norms (SN) (i.e., the social pressure they feel to engage in the behavior or not). Third, control beliefs refer to an individual’s perception of the existence of factors that can facilitate or limit the adoption of the behavior. These beliefs impact the individual’s perceived behavioral control (PBC) (i.e., their perception of their ability to engage in the behavior) (Ajzen, 2019). Together, these three beliefs guide intention to engage in a behavior, which is defined as, “an indication of a person’s readiness to perform a given behavior”. Intention then guides behavior (Ajzen, 2019).

Although the benefits of engaging in healthy PA and nutrition practices have been demonstrated in the context of childhood cancer (Beulertz et al., 2016; Rogers and Barr, 2020), research has been somewhat impeded by a lack of understanding of the specific factors that contribute to adhering to these behaviors. Identifying whether some factors from the TPB contribute to these behaviors more than others would allow to refine existing interventions and to develop more targeted interventions in this context. If e.g., attitude emerges as the single most important contributor to a given health behavior (i.e., there is little contribution of PBC or SN), interventions should target behavioral beliefs to elicit a change in intention. This change in intention would then theoretically affect the behavior. If, however, two or three of the constructs are found to contribute to the behavior, the relative weight of each construct should be considered and appropriately targeted. Indeed, if one or more of the constructs has greater relative weight, it is theorized that intervening upon this construct will have a greater influence on an individual’s intentions and their behavior (Ajzen, 2006a). Identifying these factors is essential as it will help formulate recommendations on targets to select in behavioral interventions, therefore allowing the development of more focused approaches and the refinement of existing interventions.

In order to examine the constructs from the TPB, Ajzen, 2006(b) recommends formulating a questionnaire that measures adherence to each TPB construct for a given behavior (Ajzen, 2006b). Based on this recommendation, the first aim of this study was to document whether there were differences in adherence and rejection of typical contrasting scenarios in levels of behavior, attitude, intention, PBC, and SN according to the behavioral domain (i.e., physical activity and nutrition) and the nature of the scenario (healthy vs unhealthy). The second aim of this study was to: (a) Examine the relative contribution of attitude, PBC, and SN to behavior across healthy versus unhealthy scenarios and behavioral domain (i.e., PA and nutrition), and (b) Explore the role of intention as an intermediary variable in the relationship between the sociocognitive contributors and healthy PA and nutrition behavior.

Methods

Participants

This was an ancillary analysis of data collected as part of an interventional study called the VIE (valorisation, implication, education) program, a multidisciplinary family lifestyle intervention with nutrition, physical activity, and psychological support components (Bélanger et al., 2022; Caru et al., 2020; Levesque et al., 2023). Parents of post-treatment children diagnosed with cancer who had been diagnosed and treated at Sainte-Justine University Health Center were recruited to participate in this program between June 2017 and October 2019. To be eligible to participate, parents had to have a child meeting the following criteria: (1) be under 21 years of age at diagnosis; (2) be/have been treated with chemotherapy or radiotherapy; (3) be able to provide informed consent. Families also had to be able to read, speak, and understand French. Families were excluded if the child: (1) had not received chemotherapy or radiotherapy (e.g., surgery only) and (2) had advanced cancer with a prognosis of less than 12 months (Bélanger et al., 2022; Caru et al., 2020; Levesque et al., 2023). The sample description is available in Table 1.

Table 1.

Participant characteristics.

	M (SD) or n (%) n = 96
Sex
Male	45 (46.9)
Female	51 (53.1)
Ethnicity
White/Caucasian	81 (84.4)
Black, Latin/Hispanic, middle east, other	9 (9.4)
Missing	6 (6.3)
Age at time of study (years)	9.98 (5.31)
Clinical characteristics
Age at diagnosis (years)	8.01 (5.14)
Time since diagnosis (months)	33.76 (7.91)
Diagnosis
Leukemia/lymphoma	69 (71.9)
Brain tumor	6 (6.3)
Other	21 (21.9)
Cancer stage
1	2 (2.1)
2	8 (8.3)
3	16 (16.7)
4	11 (11.5)
Not determined	57 (59.4)
Radiotherapy treatment
Yes	26 (27.1)
No	70 (72.9)

Note. The patients’ characteristics are described in this table.

We obtained written informed consent from each participant in this study. The project was conducted in accordance with the Declaration of Helsinki and was approved by the Ethics Review Committee of Sainte-Justine University Health Center (#2017-1413).

Procedure

The participants of this study were parents of post-treatment children diagnosed with cancer. All parents who participated in the VIE program were asked to complete the measures in paper format or via a secure online platform (i.e., LimeSurvey®) at the end of their child’s cancer treatments or in the following months (Mean time since diagnosis = 33.76 ± 7.91 months, Table 1).

Assessment tools

Questionnaire evaluating TPB measures for physical activity by parent proxy

Following the guidelines presented in Fishbein and Ajzen (2011), our team developed a questionnaire made up of 26 items, evaluated on a Likert scale from 1 (“disagree”) to 7 (“agree”) (Supplemental Table 1) (Fishbein and Ajzen, 2011). This questionnaire presents two contrasting scenarios (healthy vs unhealthy) related to physical activity with items measuring the different components of the TPB (i.e., behavior, attitude, intention, perceived behavioral control, subjective norms). The questionnaire was developed to inquire about adherence and rejection of Canadian PA recommendations (Canadian Society of Exercise Physiology, 2021). The “healthy” and “unhealthy” sections of the questionnaire are each made up of 13 items. The questionnaire is designed for parents to assess their child’s adherence to either of the contrasting scenarios. The contrasting physical activity scenarios are available in Supplemental Table 2. For the healthy scenario TPB questionnaires, high behavior scores, with scores ranging from 1 to 7, indicate that the described scenario closely resembles the child’s actual behavior. High attitude scores, with scores ranging from 5 to 35, indicate that the described scenario is perceived as good for the child’s health, important, and easy to put in place. High intention scores, with scores ranging from 2 to 14, indicate a high intention to adopt the described scenario. High PBC scores, with scores ranging from 3 to 21, indicate that the parents perceive that they have control over the behavior, and they feel able to put it in place and maintain it. High SN scores, with scores ranging from 0.5 to 24.5, indicate the extent to which close ones approve of the described scenario, controlling for whether their opinion is important for the parent. To note that behavior, attitude, intention, and PBC scores are calculated by summing the scores of each item assessing the construct of interest. SN scores are calculated by averaging the scores of the items assessing SN because of the complementarity of these items (Supplemental Table 3). We reverse-coded the unhealthy scenario questionnaires so that high scores in each component of the TPB would indicate a rejection of the described behavior (i.e., parents perceive their child’s behavior as being different from the described scenario), a rejection of positive attitude (i.e., a negative attitude) towards the described scenario, a rejection of the intention to adopt the described scenario, a reduced sense of PBC over the unhealthy scenario described, and a reduced impact of SN.

Questionnaire evaluating TPB measures for nutrition by parent proxy

The same questionnaire was used for two contrasting scenarios in nutrition. The questionnaire is designed for parents to assess their child’s adherence to either of the contrasting scenarios. The questionnaire was developed to inquire about adherence and rejection of the Mediterranean diet (Dieticians of Canada, 2017). The contrasting nutrition scenarios are available in Supplemental Table 2. The full questionnaire is available in Supplemental Table 1. The scores were interpreted in the same way as for the TPB measures for PA questionnaire.

Data analysis

We used IBM SPSS Statistics software, version 27.0, and RStudio for all statistical analyses.

Psychometric analyses

To confirm that the items were consistent across constructs, we examined inter-item and item-total correlations. These analyses revealed that item six should not be included as a measure of attitude because it had a low correlation with the other attitude items (Median inter-item r = −0.09) (Streiner, 2003). The Cronbach’s alpha was therefore satisfactory for attitude (α = 0.828 with item six removed), intention (α = 0.940), and PBC (α = 0.742). The Cronbach’s alpha was not calculated for SN because the items measuring this construct are complementary and were hence multiplied. A classification of the items that measured each construct is available in Supplemental Table 3.

Multiple imputation

In accordance with recommendations from Jakobsen et al. (2017), we performed a multiple imputation. Our analysis of missing data revealed that there was a random arrangement of missing values across all of our variables. Hence, we generated five datasets using a linear regression model. The multiple imputation database was used for all subsequent analyses. The reported results represent the pooled results using Rubin’s rules and the miceadds package in RStudio (Heymans and Eekhout, 2019; Robitzsch et al., 2017).

Aim 1 – Difference in outcomes

We performed a two-way MANOVA on behavior, attitude, PBC, and SN, controlling for sex, age at time of study, and time since diagnosis, to assess whether there were differences in levels of adherence to the healthy scenario and levels of rejection of the unhealthy scenario across behavior domain (i.e., PA and nutrition). Wilk’s lambda values were reported. For intention, because there was insufficient variability in both unhealthy scenarios (69.8% and 76% of participants responded “1” for unfavorable PA and unfavorable nutrition respectively), we performed an ANCOVA to detect whether there were differences between PA and nutrition in the healthy scenario while controlling for the same variables as previously.

Aim 2 – Assessment of contributors to behavior

To account for correlations between the sociocognitive contributors (Supplemental Table 4), we adjusted each contributor (e.g., attitude) for the level of the other two contributors (e.g., PBC and SN) by performing multiple regressions and saving the residuals.

For aim 2a, we performed multiple regressions with 5000 bootstrap samples to assess the contribution of attitude, PBC, and SN to healthy and unhealthy PA and nutrition behaviors. Models were adjusted for sex, age at time of study, and time since diagnosis. We assessed equality of the regression coefficients using Paternoster et al.’s methodology (Paternoster et al., 1998). We performed aim 2a in order to interpret the differences in the sociocognitive contributors based on both behavioral domain (PA and nutrition) and scenario (healthy and unhealthy). This choice was justified by the lack of variability in the intention variable in the unhealthy scenarios, which did not allow us to include this variable as an intermediary variable in aim 2b.

For aim 2b, using model four in the PROCESS macro for SPSS (Hayes, 2023), we performed analyses with attitude, PBC, and SN as contributors, intention as the intermediary variable, and behavior as the outcome. The analyses were performed with 5000 bootstrap samples and controlled for sex, age at time of study, and time since diagnosis. We assessed equality of the regression coefficients using Paternoster et al.’s methodology (Paternoster et al., 1998). Because of insufficient variability in intention scores across both unhealthy scenarios, the model placing intention as an intermediary variable was only performed in the healthy scenarios.

Results

Sample characteristics

A total of 96 families (53.1% with a female child) completed the measures and were included in this study. Children’s mean age at the time of data collection was 9.98 ± 5.31 years old, and their mean time since diagnosis was 33.76 ± 7.91 months. A full description of the sample is available in Table 1.

Aim 1 – Difference in outcomes

The MANOVA yielded a main effect of behavior domain (F(3, 4828.66) = 6.467, p < 0.001, ηp2 = 0.004), and a main effect of healthy/unhealthy scenario (F(3, 152.86) = 76.495, p < 0.001, ηp2 = 0.600). Univariate test analyses for the type of behavior factor revealed that behavior (Mnutrition = 5.631, Mpa = 4.937, F(3, 28.91) = 10.436, p < 0.001 ηp2 = 0.520) and PBC (Mnutrition = 5.494, Mpa = 5.152, F(3, 559.34) = 5.767, p < 0.001, ηp2 = 0.030) were significantly higher for nutrition than for physical activity (Figure 1). We also found that for the scenario factor, rejection of the unhealthy scenario was significantly higher than adherence to the healthy scenario for behavior (Mhealthy = 4.913, Munhealthy = 5.655, F(3, 130.51) = 14.533, p < 0.001, ηp2 = 0.250) and SN (Mhealthy = 3.991, Munhealthy = 6.556, F(3, 1770.12) = 201.946, p < 0.001, ηp2 = 0.255). The opposite finding emerged for attitude (Mhealthy = 5.953, Munhealthy = 5.553, F(3, 61.07) = 10.89, p < 0.001, ηp2 = 0.349) and PBC (Mhealthy = 5.536, Munhealthy = 5.110, F(3, 678.83) = 9.103, p < 0.001, ηp2 = 0.039), with adherence to the healthy scenario being significantly higher than rejection of the unhealthy scenario (Figure 2). The ANCOVA revealed that there was no difference in intention between physical activity and nutrition in the healthy scenario (F(1,7460.12) = 0.018, p = 0.892, ηp2 = 0).

Figure 1.

Difference in Theory of Planned Behavior variables across behavior domain assessed, physical activity versus nutrition (N = 96). Note. **p < 0.01. Error bars represent pooled standard error. Analysis controlled for sex, age at time of study, and time since diagnosis. Comparison focusing on differences between physical activity and nutrition, regardless of the scenario (the unhealthy scenario was reverse coded). Figure 1 Alt Text. A bar graph illustrating participants’ endorsement of variables from the Theory of Planned Behavior for physical activity and nutrition scenarios. Endorsement is significantly higher in the nutrition scenario than in the physical activity scenario for behavior and perceived behavioral control.

Figure 2.

Difference in Theory of Planned Behavior variables across scenario assessed, healthy versus unhealthy (N = 96). Note. **p<0.01. Error bars represent pooled standard error. Analysis controlled for sex, age at time of study, and time since diagnosis. Comparison focusing on differences between adherence to the healthy scenario and rejection of the unhealthy scenario, regardless of the behavior domain assessed. Figure 2 Alt Text. A bar graph illustrating participants’ endorsement of variables from the Theory of Planned Behavior for contrasting healthy/unhealthy scenarios. Endorsement is significantly higher in the healthy scenario for attitude and perceived behavioral control. Endorsement is significantly higher in the unhealthy scenario for behavior and subjective norms.

Aim 2a – Assessment of contributors to healthy and unhealthy scenarios

For physical activity, adherence to the healthy scenario was associated with a more positive attitude (c = 0.995, p < 0.001, 95%CI [0.562, 1.428]) and higher PBC (c = 0.814, p < 0.001, 95%CI [0.476, 1.153]). There was no difference between the regression coefficients of attitude and PBC (Z = 0.645, p = 0.741). There was no association between PA in the healthy scenario and SN. Similar results were found for nutrition in the healthy scenario, where behavior was associated with a more positive attitude (c = 0.899, p < 0.001, 95%CI [0.406, 1.392]) and higher PBC (c = 1.004, p < 0.001, 95%CI [0.513, 1.496]). There was no difference between the regression coefficients of attitude and PBC (Z = 0.295, p = 0.616). There was no association between nutrition in the healthy scenario and SN.

For physical activity, rejection of the unhealthy scenario was associated with a higher rejection of positive attitude (i.e., a higher negative attitude) (c = 1.000, p < 0.001, 95%CI [0.679, 1.321]). There was no association between PA in the unhealthy scenario and PBC or SN. Similar results were found for nutrition in the unhealthy scenario, where behavior was associated with a higher rejection of positive attitude (i.e., a higher negative attitude) (c = 0.778, p = 0.024, 95%CI [0.147, 1.408]). There was no association between nutrition in the unhealthy scenario and PBC or SN (Figure 3, Supplemental Table 5).

Figure 3.

Factors contributing to behavior across behavioral domain (physical activity/nutrition) and scenario (healthy/unhealthy) (N = 96). Note. PA, physical activity. Significant p-values are in bold and color, unsignificant relationships are in gray and dotted. Results are the pooled multiple imputation results using Rubin’s rules (https://doi.org/10.1186/1471-2288-9-57). Analyses controlled for sex, age at time of study, and time since diagnosis. Analyses performed on 5000 bootstrap samples and each predictor was adjusted for the level of the other two. Panel A illustrates results for healthy scenario physical activity, Panel B illustrates results for healthy scenario nutrition, Panel C illustrates results for unhealthy scenario physical activity, and Panel D illustrates results for unhealthy scenario nutrition. Figure 3. Alt Text. A four-panel figure illustrating the association between attitude/perceived behavioral control/subjective norms and behavior for healthy/unhealthy physical activity and nutrition. The association between attitude/perceived behavioral control and behavior is significant for healthy physical activity and nutrition (Panels A and B). For unhealthy physical activity and nutrition, the only significant association is between attitude and behavior (Panels C and D).

Aim 2b – Assessment of contributors to behavior, with intention as an intermediary variable (healthy scenario only)

As illustrated in Figure 4, when focusing on PA behavior in the healthy scenario, we found that intention was a partial intermediary variable in the relationship between attitude/PBC and healthy PA behavior. Indeed, both the direct (Attitude: c’ = 0.567, p = 0.006, 95%CI [0.161, 0.973], PBC: c’ = 0.405, p-0.021, 95%CI [0.062, 0.747]) and indirect (Attitude: ab = 0.428, p < 0.001, 95%CI [0.146, 0.710], PBC: ab = 0.410, p < 0.001, 95%CI [0.163, 0.657]) effects were significant for these sociocognitive contributors. We found that intention accounted for 43% of the total effect of attitude on behavior, and 50% of the total effect of PBC on behavior. There was no difference in the strength of the indirect effects between attitude and PBC (Z = 0.094, p = 0.537). We found no effect of intention as an intermediary variable in the relationship between SN and healthy PA behavior.

Figure 4.

Models of factors contributing to physical activity and nutrition in the healthy scenario with intention as an intermediary variable (N = 96). Note. Significant associations in color, unsignificant associations in gray. Analyses controlled for sex, age at time of study, and time since diagnosis. Analyses performed on 5000 bootstrap samples and each predictor was adjusted for the level of the other two. “a” refers to the regression of intention on the predictors, “b” refers to the regression of behavior on intention, “c’” refers to the direct effect of the predictors on behavior, and “ab” refers to the indirect of the predictors on behavior, through intention. The total effect coefficient from aim 2a is called “c” in a mediation model and corresponds to the sum of the direct (ab) and indirect (c’) effect from the mediation model. Figure 4. Alt Text. A two-panel figure depicting a model with attitude, perceived behavioral control, and subjective norms as the independent variables, behavior as the dependent variable, and intention as the intermediary variable for healthy physical activity (Panel A) and nutrition (Panel B). Panel A shows that intention is a partial intermediary variable for attitude and perceived behavioral control. Panel B shows intention as a partial intermediary variable for perceived behavior control, and a complete intermediary variable for attitude and subjective norms.

When focusing on nutrition behavior in the healthy scenario, we found that intention was a complete intermediary variable in the relationship between attitude/SN and healthy nutrition behavior. Indeed, although there was no direct effect of these contributors on healthy nutrition behavior, there was a significant indirect effect (Attitude: ab = 0.497, p < 0.001, 95%CI [0.174, 0.820], SN: ab = 0.101, p < 0.001, 95%CI [0.002, 0.199]). There was no difference in the strength of the indirect effects between attitude and SN (Z = 2.297, p = 0.989). For PBC, intention emerged as a partial intermediary variable since both the direct (c’ = 0.711, p < 0.001, 95%CI [0.426, 0.996]) and indirect (ab = 0.294, p < 0.001, 95%CI [0.067, 0.521]) effects of PBC on healthy nutrition behavior were significant. We found that intention accounted for 29% of the total effect of PBC on behavior (Figure 4, Supplemental Table 6).

Discussion

This study aimed to document the extent to which sociocognitive variables from the TPB contribute to adherence/rejection of healthy and unhealthy scenarios of PA and nutrition behavior in post-treatment children diagnosed with cancer. We found that levels of the TPB variables and their contribution to behavior differed based on both the behavioral domain (i.e., PA and nutrition) and the scenario (i.e., healthy and unhealthy) assessed. We also found that intention was a partial or complete intermediary variable in the relationship between some of the sociocognitive variables and healthy behaviors.

Physical activity versus nutrition behavior: Differences in behavior and perceived control

One key finding that emerged from this study was that parents perceived that the healthy nutrition behaviors more closely resembled their child’s actual behavior than the healthy PA behaviors. This finding was pleasantly surprising considering that previous research has found that post-treatment children with cancer tend to have poor dietary habits (Cohen et al., 2012). The parent-child relationship and the home environment have an impact on nutrition and PA behavior. A child’s dietary intake is largely influenced by the parents’ dietary choices, their behavior, and their feeding styles. Parental PA levels have also been correlated with the child’s PA levels (Tomayko et al., 2021). It could be that in our sample and in the context of childhood cancer, parents themselves had healthier eating patterns than PA behaviors, which may have in turn have influenced their evaluation of their child’s behaviors. Another explanation for this result could be that the scenarios described in the nutrition questionnaires provided more concrete examples of the behavior than the PA questionnaires. For example, since the PA questionnaires did not specify that active play can be considered PA (Brockman et al., 2010), it is possible that parents underestimated their child’s actual healthy PA behavior.

We also found that parents perceived that they had more control over their child’s nutrition practices than their PA practices. This finding is in line with previous research that found that parents of post-treatment children with cancer are likely to control their child’s food intake (Cohen et al., 2012). One qualitative study in parents of in-treatment children found that although parents reported being motivated to overcome the barriers related to their child’s limited PA practices, they showed a low capacity of being able to do so (Grimshaw et al., 2020), which could be a reflection of a reduced sense of control over the behavior.

These interesting differences in TPB outcomes between PA and nutrition did not extend to attitude and subjective norms, since we did not evidence a difference in these constructs. This could be because attitude toward PA and nutrition tend to go hand-in-hand. Indeed, Vaughan et al. (2018) identified that individuals typically fit into four profiles (e.g., they have moderate diet and negative exercise attitudes, or they have positive overall attitudes). In each of these profiles, participants’ attitude toward both PA and nutrition tended to go in the same direction (e.g., positive attitudes towards both health behaviors) (Vaughan et al., 2018). Similarly, for subjective norms, it could be that the importance of the opinions of close ones play a similar role for both PA and nutrition. This finding has been evidenced in other populations. For example, Ball et al. (2010) found that in healthy socioeconomically disadvantaged women, social norms for both PA and nutrition behaviors was predictive of an engagement in these behaviors (Ball et al., 2010).

Adhering to healthy behaviors and rejecting unhealthy behaviors: Difference in behavior, attitude, perceived control, and subjective norms

Our study also highlighted that parents report that both they and their close ones tend to reject unhealthy behaviors for the child more strongly than they adhere to healthy behaviors. This finding should be interpreted in the context of previous research, which has found that 3 to 5 years after their diagnosis, only 25% of childhood cancer survivors engage in adequate levels of low-intensity PA and certain favorable nutrition practices (i.e., consuming fruits and vegetables, 14%, and dairy, 28%). The researchers in this study also evidenced that only 52% of survivors engaged in high-intensity PA and 57% limited their fast-food intake (i.e., engaged in healthy behaviors) (Fisher et al., 2019). Taken together, these findings show that even though parents tend to strongly reject unhealthy behaviors early in the post-treatment phase, as shown in the current study, this does not necessarily translate to a sustained engagement in healthy PA and nutrition practices in nearly half of survivors. The social desirability of rejecting unhealthy behaviors also needs to be considered. Indeed, social desirability has been associated with overreporting physical activity and with eating behavior (i.e., lower reports of uncontrolled and emotional eating) (Adams et al., 2005; Kowalkowska and Poínhos, 2021). Although these findings were evidenced in healthy adults, they remain relevant since our measures were completed by parents and should be considered when documenting adherence and rejection of healthy/unhealthy behaviors. Along with social desirability, the perception of their child’s vulnerability might also motivate parents to reject unhealthy behaviors. Even after their child is in the post-treatment stage, 28% of parents still perceive their child to be vulnerable to illness (Staba Hogan et al., 2018), which could explain their tendency to reject unhealthy behaviors that might be perceived as a threat.

Although the current study was not designed to identify whether future interventions should target the rejection of unhealthy scenarios or an adherence to healthy scenarios, it did allow to highlight which TPB constructs should be targeted in both of these cases. Indeed, we evidenced that when promoting adherence to healthy PA and healthy nutrition, attitude and PBC should be targeted. If an intervention is developed to target the rejection of unhealthy behaviors, attitude should be favored above the other two constructs (i.e., PBC and SN). Healthy and unhealthy scenarios can be used in different ways when promoting changes in behavior. For example, much literature opposes the use of gain framed and loss based messaging, though the efficacy of the different types of messaging might depend on the characteristics of the message, the context, and the population (Wansink and Pope, 2015). However, if we specifically want to act upon an individuals’ motivation to engage in a behavior, healthy scenarios (and therefore gain-framed messaging) might be better suited. One scoping review on the messaging used in existing research to promote PA (not limited to the context of pediatric cancer) concluded that the following three key characteristics of effective messaging should be present: (1) Messages should be positively framed and highlight short-term social and mental health outcomes; (2) The content of the message should be adapted to the targeted population; and (3) Psychological theories should be employed to develop the messages (Williamson et al., 2020). For nutrition behavior specifically, messages should be aligned with the population’s motivations/intentions, but should also incorporate factors such as: Talking about food instead of nutrients, segmenting the message to ensure that it resonates with the targeted demographic, and addressing mixed messages (e.g., promoting a reduction of foods that are cheap and widely available would likely be ineffective) (Ruxton et al., 2023). Recommendations on the most optimal type of messaging to promote healthy PA and nutrition behavior should be taken together since healthy PA and nutrition practices may go hand-in-hand (Buja et al., 2020).

Beyond sociocognitive predictors of behavior: The key role of intention

Beyond attitude, PBC, and SN, the TPB also places an emphasis on the key role of intention to modify a given behavior (Ajzen, 2006a). When considering the role of intention in this study, we found that intention accounted for 43% of the total effect of attitude on healthy PA, 50% of the total effect of PBC on healthy PA, and 29% of the total effect of PBC on healthy nutrition. We also found that intention was a complete intermediary variable in the relationship between both attitude/SN and healthy nutrition. These results show that although targeting the sociocognitive contributors from the TPB (i.e., attitude, PBC, SN) remains of utmost importance, the role of intention is non-negligible, especially when considering that 30%–40% of the variation in health behaviors can be predicted by intention (Faries, 2016). The percentage of variation that is not predicted by intention is called the intention-behavior gap, a process through which individuals fail to translate their intentions into concrete actions. This gap has been reported in relation to both physical activity and nutrition (Faries, 2016). A key aspect that plays a role in the intention-behavior gap is the strength of the initial intention. Initially strong intentions have been described as potentially being more durable, stable, and having a bigger impact on behavior and the processing of intention-relevant information. Because strong intentions are considered to be more predictive of behavior, they can tend to reduce the intention-behavior gap. Researchers have highlighted that the stability, pliability, and effects of the strength of intention on information processing could be key moderators in the relationship between intention and behavior (Conner and Norman, 2022). Nevertheless, having strong initial intentions does not provide a guarantee that the goal behavior will be reached. Gollwitzer and Sheeran (2006), describe that this could be because individuals face issues with self-regulation when trying to reach their goals. Developing a strategy to implement intentions by detailing the “when, where, and how” a priori has been shown to have a medium-to-large positive effect on goal attainment (Gollwitzer and Sheeran, 2006). Although intention could remain an intervention target (medium-to-large changes in intention have been found to lead to small-to-medium changes in behavior (Webb and Sheeran, 2006)), future studies should focus on assessing the extent to which initial intention is a reliable and predictive marker of engagement in healthy behaviors.

Study limitations

We should recognize the limitations of the present study. First, the data was partially collected during the COVID-19 pandemic, where disruptions to daily life and established routines likely impacted the well-being of families, and the results reported in this study (Bates et al., 2021). Second, some participants followed a lifestyle-promotion intervention, and others did not. Nevertheless, preliminary analyses revealed that there were no differences in the outcomes between those who participated in the intervention and those who did not, which allowed us to include participants regardless of their participation in the intervention. Third, the results highlighted in this study are limited to the constructs included in the TPB. Fourth, there was a lack of variability in the intention variable in the unhealthy scenarios, which did not allow us to explore its role as an intermediary variable in these scenarios.

Conclusion

This study illustrates that sociocognitive contributors (attitude, PBC, SN) from the TPB should be targeted in PA and nutrition interventions in the context of childhood cancer. Intention emerged as a partial/complete intermediary variable in the relationship between the sociocognitive contributors and healthy PA/nutrition. This study allowed to better understand the factors that play a role in adherence to healthy scenarios and rejection of unhealthy scenarios following childhood cancer. Future studies should aim to reproduce this methodology during the in-treatment phase to confirm whether the same sociocognitive contributors from the TPB should be targeted in this context.

Supplemental Material

Supplemental Material - Explaining adherence to contrasted physical activity and nutrition scenarios in post-treatment childhood cancer patients: A cross-sectional study using variables from the theory of planned behavior

Supplemental Material for Explaining adherence to contrasted physical activity and nutrition scenarios in post-treatment childhood cancer patients: A cross-sectional study using variables from the theory of planned behavior by Ariane Levesque, Daniel Curnier, Valérie Marcil, Maxime Caru, Caroline Laverdière, Émélie Rondeau, Caroline Meloche, Véronique Bélanger, Isabelle Bouchard, Daniel Sinnett, and Serge Sultan in Health Psychology Open

Footnotes

Acknowledgements

We would like to thank all the families and the oncology unit clinical team of the Sainte-Justine University Health Center.

ORCID iD

Serge Sultan

Author contributions

Ariane Levesque substantially contributed to the conceptualization and design of the article, performed the data analysis, drafted the initial manuscript, and critically revised the manuscript for important intellectual content. Daniel Curnier, Valérie Marcil, Maxime Caru, Caroline Laverdière, Émélie Rondeau, Caroline Meloche, Véronique Bélanger, Isabelle Bouchard, and Daniel Sinnett substantially helped analyse and interpret the data, and critically revised the manuscript for important intellectual content. Serge Sultan substantially contributed to the conceptualization and design of the study, substantially helped analyse and interpret the data, and critically revised the manuscript for important intellectual content. All authors approved the final manuscript as submitted.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was funded by the Charles-Bruneau Cancer Care Foundation and the Sainte-Justine University Health Center Foundation. Ariane Levesque is a recipient of the Doctoral Research Award from the Canadian Institutes of Health Research.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data Availability Statement

Data is available at the following link: https://doi.org/10.5683/SP3/ULY7QJ (Sultan and Levesque, 2024)

Supplemental Material

Supplemental material for this article is available online.

References

Adams

Matthews

Ebbeling

, et al. (2005) The effect of social desirability and social approval on self-reports of physical activity. American Journal of Epidemiology 161(4): 389–398.

Ajzen

(1985) From intentions to actions: a theory of planned behavior. Action Control. Springer, 11–39.

Ajzen

(2006a). Behavioral interventions based on the theory of planned behavior.

Ajzen

(2006b) Constructing a theory of planned behavior questionnaire.

Ajzen

(2019) Theory of planned behavior diagram. Retrieved January 24, 2022 from. https://people.umass.edu/aizen/tpb.diag.html#null-link

American Childhood Cancer Organization (2024). Childhood Cancer Satatistics, Retrieved 12-05-2024 from. https://www.acco.org/childhood-cancer-statistics/#:∼:text=Globally%2C_approximately_400%2C000_children_are,their_second_decade_of_life

Ball

Jeffery

Abbott

, et al. (2010) Is healthy behavior contagious: associations of social norms with physical activity and healthy eating. International Journal of Behavioral Nutrition and Physical Activity 7: 1–9.

Bandura

(2002) Social cognitive theory in cultural context. Applied Psychology 51(2): 269–290.

Bates

Nicholson

Rea

, et al. (2021) Life interrupted: family routines buffer stress during the COVID-19 pandemic. Journal of Child and Family Studies 30(11): 2641–2651.

10.

Bélanger

Delorme

Napartuk

, et al. (2022) Early nutritional intervention to promote healthy eating habits in pediatric oncology: a feasibility study. Nutrients 14(5): 1024.

11.

Beulertz

Prokop

Rustler

, et al. (2016) Effects of a 6‐month, group‐based, therapeutic exercise program for childhood cancer outpatients on motor performance, level of activity, and quality of life. Pediatric Blood and Cancer 63(1): 127–132.

12.

Brockman

Jago

Fox

(2010) The contribution of active play to the physical activity of primary school children. Preventive Medicine 51(2): 144–147.

13.

Brouwer

Mosack

(2015) Expanding the theory of planned behavior to predict healthy eating behaviors: exploring a healthy eater identity. Nutrition & Food Science 45(1): 39–53.

14.

Buja

Grotto

Brocadello

, et al. (2020) Primary school children and nutrition: lifestyles and behavioral traits associated with a poor-to-moderate adherence to the Mediterranean diet. A cross-sectional study. European Journal of Pediatrics 179: 827–834.

15.

Canadian Society of Exercise Physiology (2021). Canadian 24-hour movement guidelines: An integration of physical activity, sendentary behavior, and sleep. https://csepguidelines.ca/

16.

Caru

Duhamel

Marcil

, et al. (2020) The VIE study: feasibility of a physical activity intervention in a multidisciplinary program in children with cancer. Supportive Care in Cancer: Official Journal of the Multinational Association of Supportive Care in Cancer 28(6): 2627–2636.

17.

Caru

Curnier

Levesque

, et al. (2021) The impact of cancer on theory of planned behavior measures and physical activity levels during the first weeks following cancer diagnosis in children. Supportive Care in Cancer: Official Journal of the Multinational Association of Supportive Care in Cancer 29: 823–831.

18.

Cohen

Wakefield

Fleming

, et al. (2012) Dietary intake after treatment in child cancer survivors. Pediatric Blood and Cancer 58(5): 752–757.

19.

Conner

Norman

(2022) Understanding the intention-behavior gap: the role of intention strength. Frontiers in Psychology 13: 923464.

20.

Dieticians of Canada (2017) The mediterranean diet: a guide to healthy eating. https://www.dietitians.ca/

21.

Faries

(2016) Why we don’t “just do it” understanding the intention-behavior gap in lifestyle medicine. American Journal of Lifestyle Medicine 10(5): 322–329.

22.

Fishbein

Ajzen

(2011) Predicting and Changing Behavior: The Reasoned Action Approach. Psychology Press.

23.

Fisher

Rausch

Ferrante

, et al. (2019) Trajectories of health behaviors across early childhood cancer survivorship. Psycho-Oncology 28(1): 68–75.

24.

Godin

Kok

(1996) The theory of planned behavior: a review of its applications to health-related behaviors. American Journal of Health Promotion: AJHP 11(2): 87–98.

25.

Gollwitzer

Sheeran

(2006) Implementation intentions and goal achievement: a meta‐analysis of effects and processes. Advances in Experimental Social Psychology 38: 69–119.

26.

Government of Canada (2024) The cancer in young people in Canada program. Retrieved 12-05-2024 from. https://www.canada.ca/en/public-health/services/chronic-diseases/cancer/cancer-young-people-canada-program.html

27.

Grimshaw

Taylor

Mechinaud

, et al. (2020) Physical activity for children undergoing acute cancer treatment: a qualitative study of parental perspectives. Pediatric Blood and Cancer 67(6): e28264.

28.

Hayes

(2023) The Process macro for SPSS and SAS website. Retrieved 01/01/2024 from. https://www.processmacro.org/index.html

29.

Heymans

Eekhout

(2019) Applied missing data analysis with SPSS and (R)studio. https://bookdown.org/mwheymans/bookmi/rubins-rules.html

30.

Betz

Marchese

(2021) Exploring pulmonary function and physical function in childhood cancer: a systematic review. Critical Reviews in Oncology 160: 103279.

31.

Jakobsen

Gluud

Wetterslev

, et al. (2017) When and how should multiple imputation be used for handling missing data in randomised clinical trials–a practical guide with flowcharts. BMC Medical Research Methodology 17(1): 1–10.

32.

Joffe

Ladas

(2020) Nutrition during childhood cancer treatment: current understanding and a path for future research. The Lancet. Child & adolescent health 4(6): 465–475.

33.

Kirk

Haegele

(2019) Theory of planned behavior in research examining physical activity factors among individuals with disabilities: a review. Adapted Physical Activity Quarterly: Adapted Physical Activity Quarterly 36(1): 164–182.

34.

Kowalkowska

Poínhos

(2021) Eating behaviour among university students: relationships with age, socioeconomic status, physical activity, Body Mass Index, waist-to-height ratio and social desirability. Nutrients 13(10): 3622.

35.

Lemay

Caru

Samoilenko

, et al. (2019) Prevention of long-term adverse health outcomes with cardiorespiratory fitness and physical activity in childhood acute lymphoblastic leukemia survivors. Journal of Pediatric Hematology 41(7): e450–e458.

36.

Levesque

Ogez

Gravel

, et al. (2023) Feasibility study of taking back control together, an intervention to support parents of children with cancer. Journal of Clinical Psychology in Medical Settings 31: 1–11.

37.

Lin

Jay

Howard

(2018) Childhood cancer incidence in Canada: demographic and geographic variation of temporal trends (1992–2010). Health promotion and chronic disease prevention in Canada: Research, policy and practice 38(3): 79.

38.

National Cancer Institute (2022) Late effects of treatment for childhood cancer (PDQ) - patient version. Retrieved August 17 from. https://www.cancer.gov/types/childhood-cancers/late-effects-pdq#:∼:text=Health_habits_that_promote_a,A_heart%2Dhealthy_diet

39.

National Research Council . (2003). Childhood cancer survivorship: improving care and quality of life.

40.

Paternoster

Brame

Mazerolle

, et al. (1998) Using the correct statistical test for the equality of regression coefficients. Criminology 36(4): 859–866.

41.

Prochaska

Velicer

(1997) The transtheoretical model of health behavior change. American Journal of Health Promotion: AJHP 12(1): 38–48.

42.

Riebl

Estabrooks

Dunsmore

, et al. (2015) A systematic literature review and meta-analysis: the Theory of Planned Behavior's application to understand and predict nutrition-related behaviors in youth. Eating Behaviors 18: 160–178.

43.

Robitzsch

Grund

Henke

, et al. (2017) Package ‘miceadds’. R Foundation for Statistical Computing.

44.

Rogers

Barr

(2020) The relevance of nutrition to pediatric oncology: a cancer control perspective. Pediatric Blood and Cancer 67: e28213.

45.

Ruxton

Ruani

Evans

(2023) Promoting and disseminating consistent and effective nutrition messages: challenges and opportunities. Proceedings of the Nutrition Society 82(3): 394–405.

46.

Staba Hogan

Ross

Balsamo

, et al. (2018) Parental perception of child vulnerability in childhood cancer survivors. Pediatric Blood and Cancer 65(11): e27364.

47.

Stolley

Restrepo

Sharp

(2010) Diet and physical activity in childhood cancer survivors: a review of the literature. Annals of behavioral medicine: A Publication of the Society of Behavioral Medicine 39(3): 232–249.

48.

Streiner

(2003) Starting at the beginning: an introduction to coefficient alpha and internal consistency. Journal of Personality Assessment 80(1): 99–103.

49.

Sultan

Levesque

(2024) “Dataset for the study “Explaining adherence to physical activity and nutrition scenarios in post-treatment childhood cancer patients: A contrasted scenario-based approach using variables from the Theory of Planned Behavior””. doi: 10.5683/SP3/ULY7QJ.

50.

Tomayko

Tovar

Fitzgerald

, et al. (2021) Parent involvement in diet or physical activity interventions to treat or prevent childhood obesity: an umbrella review. Nutrients 13(9): 3227.

51.

Vaughan

Ghosh-Dastidar

Dubowitz

(2018) Attitudes and barriers to healthy diet and physical activity: a latent profile analysis. Health Education & Behavior: The Official Publication of the Society for Public Health Education 45(3): 381–393.

52.

Wansink

Pope

(2015) When do gain-framed health messages work better than fear appeals? Nutrition Reviews 73(1): 4–11.

53.

Webb

Sheeran

(2006) Does changing behavioral intentions engender behavior change? A meta-analysis of the experimental evidence. Psychological Bulletin 132(2): 249–268.

54.

Williamson

Baker

Mutrie

, et al. (2020) Get the message? A scoping review of physical activity messaging. International Journal of Behavioral Nutrition and Physical Activity 17: 1–15.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.36 MB