AbstractSection Background

The adoption of protective health behaviors, such as physical activity (PA) and healthy nutritional practices during childhood cancer treatments should be encouraged. This study aimed to document PA levels and diet quality before and after a multidisciplinary family lifestyle intervention, and to assess whether differences in PA levels and diet quality were associated with the degree of participation in the intervention.

AbstractSection Methods

A multidisciplinary intervention (exercise intervention, nutritional, psychosocial support) was offered to families of children receiving treatments for childhood cancer. PA levels were assessed by self-reported total, moderate, and vigorous intensity minutes of physical activity. Diet quality was assessed using a dietary quality indicator. Participation in the intervention was measured by overall and domain-specific points of contact. We used non-parametric tests to evaluate changes over time and differences with an ad hoc comparison group.

AbstractSection Results

In 38 participants, we found that there was a significant difference in total minutes of PA pre- and post-intervention (p = 0.022, r = 0.27), but no difference in diet quality (p = 0.136, r = 0.19). We found that minutes of vigorous intensity PA improved rather than deteriorated (OR = 2.19, 95%CI: 1.13–4.25). Improvements in minutes of vigorous intensity PA were associated with participation in the intervention.

AbstractSection Conclusion

In its current form, this multidisciplinary lifestyle intervention is associated with limited improvements in PA. Before testing its effects in studies with adequate statistical power, the program should be refined to limit heterogeneity in levels of participation and optimize its active ingredients.

Improvements in childhood cancer treatments in recent decades have allowed for 5-year overall survival rates to surpass 80% [1]. Although children diagnosed with cancer live longer, they have to deal with a host of complications and chronic conditions resulting from cancer and its treatment [2]. The prevalence of these late effects is high, with as many as 95.5% of childhood cancer survivors living with at least one chronic condition [3], making the prevention of late effects a research topic of priority. Research in this field suggests that the severity of late effects could be mitigated by adopting protective health behaviors, such as engaging in regular physical activity and having a healthy diet [4]. When adopted during cancer treatments, these protective health behaviors might even limit the occurrence of late effects, while improving quality of life in children diagnosed with cancer [5, 6].

Indeed, children who engage in PA following their cancer diagnosis have a better physical function than children who receive standard care and do not engage in PA [7]. Exercise interventions have been found to have a positive effect on body composition, cardiorespiratory fitness, muscle strength and quality of life [8]. Furthermore, research has shown that exercise interventions are safe and feasible during treatments for childhood cancer [9]. According to the recent International Pediatric Oncology Exercise Guidelines (iPOEG), in-treatment children diagnosed with cancer should move whenever they can, regardless of their age, abilities, diagnosis, cancer stage, and cancer phase [10]. Hence, exercise interventions during and beyond cancer treatments in pediatric oncology emerge as a key protective health behavior to improve quality of life and limit the severity and occurrence of late effects [5, 11].

Beyond preventing late effects [11], favorable diet quality during treatments has been associated with improvements in quality of life, treatment tolerance, chances of remission, and survival, and a reduction in cancer recurrence, the development of infections, and the long-term development of obesity during survivorship [6, 12, 13]. Adopting favorable nutritional practices during treatments, however, can be a challenge since patients often report nausea and altered taste because of the medications that they receive [14, 15]. Nevertheless, promoting diet quality is of utmost importance because a poor diet quality can lead to being underweight or overweight, regardless of the cancer stage [16,17,18,19].

The exercise and nutritional interventions assessed in this study were offered to children and adolescents diagnosed with cancer and their family as part of a multidisciplinary family lifestyle intervention that also provided psychosocial support for parents. Although associations between the intervention and some outcomes have been explored previously [20, 21], the role of participation in the intervention has not been systematically documented.

Hence, the first aim of this exploratory study was to compare PA levels and diet quality at pre- and post-intervention, and to compare these outcomes between the intervention group (IG) and a comparison group (CG) recruited in the same setting. The second aim of this study was to assess whether the following outcomes were associated with the degree of participation in the intervention: changes in PA levels and diet quality from pre- to post-intervention in the IG, the change in status from pre- to post-intervention (i.e., improved and deteriorated) in the IG, and the difference in outcomes between the IG post-intervention and the CG.

In line with guidelines on the optimization of behavioral interventions, this study is positioned within phase 1b of the ORBIT model (i.e., Refine phase) since it aims to inform on some practical aspects of the intervention (i.e., the degree of participation) and to evaluate different components of the intervention to illustrate their possible effects on the targeted outcomes [22]. In line with this evaluation phase, no hypotheses were formulated.

Families whose child had been diagnosed and treated at the Sainte-Justine University Health Center in Montreal (Quebec), Canada were eligible to participate in this study. For the IG, participants were recruited between 02/2018 and 12/2019 to be part of the VIE (Valorization, Implication, Education) program, a multidisciplinary family lifestyle intervention with an exercise intervention, nutritional, and psychological support [23]. To be eligible for the IG, families had to have a child meeting the following criteria: [1] be under 21 years of age at diagnosis; [2] be treated with chemotherapy/radiotherapy; [3] be able to provide informed consent (by parents or legal guardians); [4] be a minimum of 4 weeks post-diagnosis. The participants also had to be able to read, speak, and understand French. Participants 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 twelve months.

The CG had the same inclusion criteria as the IG, with the exception that participants were 1.5 to 3.5 years post-diagnosis (and therefore not exposed to the multidisciplinary family lifestyle program). Participants were recruited between 06/2017 and 10/2019 from patients who received a cancer diagnosis at Sainte-Justine University Health Center between 2013 and 2015. Data from the CG was collected independently and consecutively.

We obtained written informed consent from participants (≥ 18 years old) or from parents or legal guardians. We obtained assent from participants < 18 years old. This study was conducted in accordance with the Declaration of Helsinki and the protocol was approved by the Ethics Review Committee of Sainte-Justine University Health Center (#2017 − 1413).

IG participants were approached 1.58 ± 1.62 months after their cancer diagnosis, with the approval of their medical oncologist. IG participants were informed that the study had three intervention domains: exercise, nutritional, and psychosocial support. IG participants were encouraged to participate in all interventions domains, but they could also choose to participate in any domain(s) of their choice. Participants did not receive any compensation.

The detailed procedures of all intervention domains have been published previously [24,25,26,27]. Briefly, for the exercise intervention, an individualized exercise program tailored to the participants’ needs, abilities, health and functional status, and goals was developed by a kinesiologist. IG participants had follow-ups every two months over a two-year period. For the nutritional intervention, registered dieticians provided nutritional counselling and encouraged the adoption and refinement of nutritional goals every two months for a one-year period. For the psychosocial intervention, a 6-session problem-solving skills training intervention was offered to support parents over a 6 to 8-week period (Fig. 1). Participants in the CG received standard treatment and completed the VIE measures after the end of their cancer treatments. Collecting data at this time point in the CG allowed us to assess whether there was a difference in the outcomes between participants who had followed the intervention and those who had not followed the intervention when they reached or neared the end of treatments.

Timeline of participation in the VIE project for the intervention group

Full size image

In the IG, pre-intervention measures were taken at study entry/at the beginning of treatments; and post-intervention measures were taken when the VIE study ended/when participants were nearing the end of treatments. The median time elapsed between the pre- and post-intervention measures in the IG was 15 months. In the CG, since participants were recruited when treatments had already been completed, measures were only collected at one time point (i.e., in the early aftercare period). For ethical and methodological constraints, we could not include a baseline measure in the comparison group.

Daily minutes of physical activity at moderate and/or vigorous intensity. Participants self-reported the frequency, duration and intensity of each activity. This allowed us to calculate their total daily minutes of PA, and number of minutes of moderate and vigorous intensity PA were calculated. This is an age-specific adapted version of the Minnesota Leisure Time Physical Activity Questionnaire [28, 29] and of the Tecumseh Self-Administered Occupational Physical Activity Questionnaire [30], which are strongly associated with accelerometer data [31,32,33] and have been broadly used in pediatric oncology [20, 34, 35]. The metabolic equivalent value (MET) from the Compendiums of Physical Activity for Youth [36] was used to quantify the intensity of each activity.

A recent review on randomized controlled trials evaluating exercise interventions in pediatric oncology highlighted that the lowest reported duration of an exercise session was 40 min [37]. Based on this finding and on our data distribution, we dichotomized the total minutes of PA and the minutes of PA at different intensities in order to investigate improvements and deteriorations in minutes of PA. More specifically, we dichotomized below-threshold and above-threshold minutes of PA, as follows:

We defined improvement as participants who went from below-threshold at pre-intervention to above-threshold at post-intervention (e.g., went from a total number of minutes of PA below 40 at pre-intervention to a total number of minutes of PA above or equal to 41 at post-intervention). We defined deterioration as the opposite (going from above-threshold at pre-intervention to below-threshold at post-intervention).

Diet Quality Index (DQI). We assessed nutritional intake pre- and post-intervention using the Diet Quality Index (DQI). The DQI is a dietary quality indicator that is calculated using the scores of four dietary quality components: variety (in terms of food groups and protein sources), adequacy, moderation, and balance. A score between 0 and 15 is calculated for the variety of food groups component, 0–5 for the variety of protein sources component, 0–40 for adequacy, 0–30 for moderation and 0–10 for balance. The DQI is calculated by adding the scores of each component. The DQI varies between 0 and 100, with a higher score indicating better dietary quality [38]. As there are no established thresholds in our population to classify individuals according to DQI results, the DQI scores were divided into quartiles. A dichotomous variable was then created for diet quality, we used thresholds determined according to Q1 and Q3 quartiles. In the present sample, Q1 and Q3 values were of 44.54 and 52.00 respectively [39]. We defined improvement as going from below-threshold at pre-intervention to above-threshold at post-intervention. We defined deterioration as the opposite (going from above- to below-threshold).

Exposure to the VIE program. We derived two exposure variables: overall and domain-specific number of points of contact. In the exercise intervention, the points of contact were defined as the initial and follow-up sessions, and the supervised exercise sessions between follows-ups. In the nutrition intervention, the points of contact were defined as the initial and follow-up sessions, and in the psychosocial intervention, they were defined as all sessions of the problem-solving skills intervention. The overall number of points of contact referred to the total number of points of contact for all intervention domains (i.e., exercise, nutrition, psychosocial). The domain-specific number of points of contact referred to the number of points of contact for a single intervention domain (e.g., for nutritional intervention only) (Supplementary Table S1).

We used the MatchIt package in R software to perform propensity score matching (PSM) and IBM SPSS Statistics software, version 27.0 for all other statistical analyses.

Propensity score matching. In line with the recommendations of Kazdin [d40], since our sample sizes differed between the IG and CG, we performed PSM to produce two groups of equal size with similar clinical and sociodemographic characteristics (Fig. 2). We matched participants based on their sex, age at recruitment, and whether they received radiotherapy. We could not match participants on time since diagnosis given the inclusion criteria. After checking the initial balance and confirming the necessity to perform PSM since one variable (i.e., age at recruitment) was imbalanced [41], we performed radius matching. In line with established recommendations, we set our caliber to 0.2 of the standard deviation of the propensity scores [42]. We followed the assumption that data from PSM is not independent [43].

Flowchart for the VIE intervention group and the comparison group

Full size image

Missing data. For all statistically significant results, we performed best-worst and worst-best sensitivity analyses to assess the range of uncertainty of our results (0.08% missing data). For pre-intervention data in the best-case scenario, we considered that those with missing data had the same percentage increase in outcomes as their counterparts. For pre-intervention data in the worst-case scenario, we considered that participants with missing data did not change (i.e., same value at pre-intervention as the value at post-intervention). There was no missing data at post-intervention.

Aim 1– comparison of outcomes. We performed Wilcoxon signed-rank tests to assess whether there was a difference in the outcomes from pre- and post-intervention and between the IG post-intervention and the CG. We performed McNemar’s tests to compare outcomes according to threshold. We calculated odds ratios with their 95%CI to compare the proportion of participants who improved to those who deteriorated according to these thresholds.

Aim 2– association with exposure variables. We performed bivariate correlations between the change in outcomes from pre- to post-intervention and each exposure variable. We also compared exposure across status (improved and deteriorated) for the PA and nutritional outcomes with non-parametric Mann-Whitney tests. To limit Type I error due to multiple testing, we corrected p-values by the Benjamini-Hochberg false discovery rate (FDR) method. The same strategy was applied when comparing the IG and CG.

A total of 62 families participated in the IG of the VIE program, and a total of 84 families were included in the CG. Pre- and post-intervention exercise data was available for 48 participants in the IG and 82 participants in the CG. Pre- and post-intervention nutritional data was available for 43 participants in the IG and 81 participants in the CG. After performing PSM, 38 participants from each group were matched and included in subsequent analyses (Fig. 2). At their enrollment, participants’ children from the matched sample had a mean age of 8.11 years (SD = 4.62) and their mean time since diagnosis was 1.58 months (SD = 1.62). Participants’ characteristics are presented in Table 1.

Following radius matching, we found that the balance for sex (standardized mean difference (SMD) = 0.10, variance ratio (VARR) = 1.01, eCDF = 0.03), age at recruitment (SMD = -0.15, VARR = 1.04, eCDF = 0.07), and radiotherapy (SMD = 0.10, VARR = 1.8, eCDF = 0.02) was good.

Pre- and post-intervention in the IG. In participants in the IG, we found that there was a significant difference in total minutes of PA pre- and post-intervention (Wilcoxon Z=-2.29, p = 0.022, r = 0.27). Best-worst and worst-best sensitivity analyses revealed that the results remained significant (r = 0.20–0.27). There was no increase, however, in minutes of moderate intensity PA (Z=-1.43, p = 0.154, r = 0.17), minutes of vigorous intensity PA (Z=-1.26, p = 0.207, r = 0.15), and diet quality (Z=-1.49, p = 0.136, r = 0.19) (Fig. 3). To note that in the full unmatched sample (N = 39–45), we observed significant increases in all PA outcomes following the program, but no difference in DQI score (Supplementary Table S2). These results were confirmed when considering the outcomes according to thresholds (Table 2).

Difference in minutes of physical activity and diet quality pre- and post- intervention (matched sample)

Note. *p < 0.05 on Wilcoxon signed rank test. DQI, Diet Quality Index; ns, not significant; PA, Physical Activity. Panel A depicts differences in minutes of PA. Panel B depicts difference in diet quality (DQI) score. Values on graphs represent the median difference in outcomes from pre- to post-intervention (n = 38). For total minutes of PA, the median percent change was to + 65.89% [-2.40, 131.73]. For moderate intensity PA, the median percent change was + 27.09% [-19.04, 189.99]. For vigorous intensity PA, the median percent change was + 21.31% [-47.09, 268.51]. For diet quality, the median percent change was + 2.69% [-2.52, 10.62]

Full size image

When considering improvement and deterioration status in total minutes of PA following the program, we found no difference in the likelihood to improve (73%) versus deteriorate (38%) (OR = 1.89, 95%CI: 0.95–3.75).

For minutes of moderate intensity PA (Table 2b), we found no difference in the likelihood to improve (55%) versus deteriorate (58%) (OR = 0.94, 95%CI: 0.46–1.89).

For minutes of vigorous intensity PA (Table 2c), we found a difference in the likelihood to improve (68%) versus deteriorate (31%). This translated to participants being 2.19 times more likely to have their minutes of vigorous intensity PA improve rather than deteriorate following the program (OR = 2.19, 95%CI: 1.13–4.25). Sensitivity analyses confirmed this result with worst case OR = 2.10 (95%CI: 1.11–3.99) and best-case OR = 2.04 (95%CI: 1.07–3.88).

When diet quality was dichotomized according to the Q1 threshold (i.e., lowest score), we found no difference in the likelihood to improve (30%) versus deteriorate (23%) (OR = 1.32, 95%CI: 0.63–2.79). The same result was found when the Q3 threshold (i.e., highest score) was used (40% improved, 25% deteriorated, OR = 1.69, 95%CI: 0.78–3.27) (Table 2).

IG Post-intervention and CG. We found no difference between the IG and the CG for total minutes of PA (Wilcoxon Z=-0.24, p = 0.811, r = 0.03), minutes of moderate intensity PA (Z=-0.47, p = 0.637, r = 0.05), minutes of vigorous intensity PA (Z=-0.82, p = 0.413, r = 0.09), and diet quality (Z=-1.27, p = 0.206, r = 0.15) (Fig. 4). The results were similar in the full unmatched sample (Supplementary Table S3). These results were also confirmed when comparing frequencies in the IG and CG according to the thresholds described in the Assessment tools section (Supplementary Fig. S1).

Difference in minutes of physical activity and diet quality between the intervention group and the comparison group (matched sample)

Note. *p < 0.05 on Wilcoxon signed rank test. DQI, Diet Quality Index; ns, not significant; PA, Physical Activity. Panel A depicts differences in minutes of PA. Panel B depicts difference in diet quality score. Values on graphs represent the median difference in outcomes between the intervention group post-intervention and the comparison group. For total minutes of PA, the median percent change was to + 3.35% [-34.86, 93.38]. For moderate intensity PA, the median percent change was + 9.97% [-27.90, 215.79]. For vigorous intensity PA, the median percent change was − 17.31% [-51.80, 73.08]. For diet quality, the median percent change was − 6.61% [-12.60, 0.12]

Full size image

Difference pre- and post-intervention in the IG. A description of all exposure variables is available in Supplementary Table S1. When exploring correlations between the exposure variables (overall and domain-specific number of points of contact) and the difference in outcomes from pre- to post-intervention in the IG, we found no significant association (Supplementary Table S4).

Improvement status. When exploring the number of points of contact according to improvement/deterioration status, we observed that participants in the IG whose minutes of vigorous intensity PA improved (N = 13, Mdn = 16 points of contact) had significantly more points of contact in the exercise intervention than participants whose minutes of vigorous intensity PA deteriorated (N = 5, Mdn = 4) (U = 59.50, FDR-corrected p = 0.020). Those with improved minutes of vigorous intensity PA (N = 13, Mdn = 25) had also received more points of contact across all intervention domains than participants whose minutes of vigorous intensity PA deteriorated (N = 5, Mdn = 13) (U = 58.00, FDR-corrected p = 0.050).

We also found that when considering the raw association, participants in the IG whose diet quality improved (N = 8, Mdn = 5.5 points of contact) when dichotomized according to the Q3 threshold had significantly more points of contact in the nutritional intervention than participants in the IG who deteriorated (N = 3, Mdn = 4.0) (U = 1, raw p = 0.024). This association was not significant, however, when using the FDR corrected p-value (FDR-corrected p = 0.060). Other associations were not significant (Supplementary Table S5).

Difference between IG post-intervention and CG. We found no association between the exposure variables and the difference in outcomes between the IG and CG. We found no correlation between exposure variables and the difference in outcomes (Supplementary Table S6). This was also the case when considering thresholds in outcome measures.

The present results suggest that the multidisciplinary lifestyle VIE intervention is associated with limited follow-up improvements in total and vigorous intensity minutes of PA but stability in diet quality. Nevertheless, some results pointed to an effect of the program, notably that improvements were associated with the intensity of participation in the program. Beyond mere statistical significance, we observed that for most outcomes, except for minutes of moderate intensity PA, participants’ behavior tended to improve, rather than deteriorate (median OR = 1.79).

Although this study lacked the statistical power necessary to conclude whether the intervention had a positive impact, it allowed to detect whether the results indicated a promising tendency. Our results showed that participants’ minutes of PA and diet quality score, with the exception of minutes of moderate intensity PA, neared or surpassed a two-fold likelihood to improve rather than deteriorate following the intervention, though this result was not always statistically significant (likely as a result of the small sample size). This tendency for PA and diet quality to improve following a supportive intervention is aligned with previous studies in the field. Indeed, in an evidence synthesis used to develop the international Pediatric Oncology Exercise Guidelines, Wurz et al. highlighted that 42% of included studies found increased PA behavior following an intervention [44]. Similarly, diet quality has been associated with improvements following a lifestyle intervention in children with abdominal obesity [45]. It should be noted that diet quality assessments in children with cancer are still limited due to the emerging nature of this field. However, a detailed account of changes in dietary intakes following the nutritional support intervention detailed in this study has been conducted by Delorme et al., who evidenced that participants in the intervention group had a lower caloric intake and higher calcium intake than participants in the comparison group [46]. In regard to PA, Rapti et al., 2023 evidenced in an umbrella review that systematic reviews on exercise interventions in the field of pediatric oncology tend to lack information on exercise intensity, exercise interventions generally tend to improve targeted outcomes such as cancer-related fatigue, muscle strength, and cancer-related pain [47]. Doing so likely requires favorable changes in the total number of minutes of PA. Further studies are needed to confirm the minimum amount and intensity of exercise necessary to induce changes in commonly targeted outcomes. Regarding diet quality, even though improvements were not significant, it is essential for future iterations of the VIE project to continue targeting diet quality and building on the favorable trends that we detected in this preliminary analysis. Indeed, a recent review investigating the state of the evidence on nutritional assessments and interventions in childhood cancer survivors highlighted that survivors tend to exhibit a poor diet quality, which can include insufficient consumption of fruits, vegetables, protein and fiber, and an excess consumption of carbohydrates, fats and ultra-processed foods [48]. Poor diet quality can lead to deleterious effects such as abnormal lipid marker levels, elevated triglyceride or inflammatory marker levels [48], making it of utmost importance to continue targeting diet quality in supportive care interventions.

It is compelling that significant results particularly emerged throughout our study in favor of the exercise intervention and counselling. This finding is not entirely surprising considering that in a review on theory-based PA and/or nutrition interventions, Rodrigues et al. (2023) highlighted that in interventions that target both PA and nutrition, nutritional outcomes tend to be less consistent, while most studies report improvements in PA outcomes. These authors hypothesized that this difference in PA and nutrition outcomes could be due to the primary focus of most multidisciplinary interventions not being changing diet, but rather improving other distal outcomes, such as quality of life [49]. It is also possible that significant results were observed for PA outcomes, but not for diet quality, because of the different techniques adopted by each intervention domain to induce change. Although all interventionists were trained in motivational communication [50] and both intervention domains used established behavior change strategies, the kinesiologists who administered the exercise intervention module used behavior change techniques based on a “repetition and substitution” model of action (instruction, demonstration, and practice of the behavior), whereas the nutritionists who administered the nutritional intervention module used a “goals and planning” model of action (goal setting, information about health consequences, highlighting discrepancies between the current behavior and the goal) [51]. It should be noted that previously published studies on the VIE program evidenced significant improvements in both PA and diet quality directly after the end of the intervention [20, 21]. The present study suggests that these short-term improvements were not sustained at the same level at follow-up.

As for participation, this study demonstrated that in such a complex intervention program involving vulnerable families and patients, real-life constraints can limit implementation. Although we had originally planned to have families engage in all three components of the program with a predetermined frequency/duration, the results of the participation measures show that there was variability between families in exposure to the program. Despite this variability, the multidisciplinary nature of our intervention remains a strength, since experts have pointed out that multiple health behavior change interventions (i.e., interventions that aim to improve two of more behaviors simultaneously) are: “A unique opportunity to improve health and well-being and reduce health care costs by maximizing intervention contacts” [52]. We nevertheless found little support for an association between exposure measures and differences in the outcomes. Yet the results suggest that participants whose minutes of vigorous intensity PA improved were more likely to have a higher number of overall and domain-specific points of contact. Although many factors could have contributed to response variance across participants, the results support the hypothesis that some response variance could be explained by more intense participation to some or all intervention domains [53].

There are certain key changes that would benefit the intervention prior to its formal pilot-testing. First, to limit exposure variability, we would advocate for a simplification of the program to limit its burden. This would be instrumental to encourage all participating families to systematically engage in the same sets of activities. Second, it would be useful to select behavior change techniques that hold promise in changing PA and nutritional outcomes. In the context of the VIE program, both behavioral modules could use behavior change techniques under the label “repetition and substitution” [51], since these behavior change techniques were associated with clearer changes in the current study. A third modification that could be made in future studies would be to adapt the inclusion criteria of the CG so that participants are in the same treatment phase as the IG.

We recognize the limitations of the present study. First, due to the nature of the VIE program, there was some variability in the nature and intensity of each intervention domain. Although the interventions were manualized and full descriptions are available [21, 24], treatment fidelity was not systematically analyzed. Thus, it is not possible to certify that the interventions were standardized. A second limitation is that the analyses of this study were conducted on a subsample of participants to compare the IG and CG. This methodological choice was in line with established recommendations regarding how to deal with missing data [54]. Fourth, there was some missing data for participants in the intervention group, at pre-intervention. To address this limitation, we conducted best-worst and worst-best sensitivity analyses to provide the range of uncertainty of our results due to missing data. Finally, the IG and CG were not in the same treatment phase. This choice was initially made for ethical concerns that prevented us from offering the program to some families and not others. It would be necessary for future studies to assess whether there is a difference between families who receive such an intervention or not at the same moment of the cancer trajectory.

In a proof-of-concept of a multidisciplinary lifestyle intervention (the VIE program) for children diagnosed with cancer and their family, we found some support for improvements in PA at follow-up, but not in diet quality. The association of significant change over time with participation testifies in favor of the intervention. Given the limited results and the high variability in participation, a simplification and systematization of the VIE program is warranted. A second proof-of-concept study should evaluate this refined version and explore whether it is associated with sustained changes in the outcomes.

This research was funded by the Charles-Bruneau Cancer Care Foundation and the Sainte-Justine University Health Center Foundation. Ariane Levesque is recipient of a Doctoral Research Award from the Canadian Institutes of Health Research (CIHR). We would like to thank all the families and the oncology unit clinical team of the Sainte-Justine University Health Center.

Download references

Author notes

A.L. 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. D.C., V.M., V.B., I.B., J.D., C.L., M.N., D.O., E.R., K.P., and D.S. substantially helped analyse and interpret the data, and critically revised the manuscript for important intellectual content. M.C. and S.S. 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.

Correspondence to Serge Sultan.

The authors declare no competing interests.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Below is the link to the electronic supplementary material.

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions