Statistical Analysis
Sample size was determined by the accrual time period. Due to our relatively small sample size (N=23), robust nonparametric statistical methods were used for data analysis. The continuous variables were summarized by median, minimum, and maximum and the categorical variables were summarized by frequency and rates. At each visit timepoint, boxplot was used to illustrate the distribution of the median BMI z-score overall and within subgroups. The Friedman’s test was used to examine if there is any significant difference in median BMI z-score across the three timepoints, the Fisher’s exact test was used to examine the difference in BMI categories for the three visits, and the Wilcoxon rank sum test was used for comparison of BMI z-score between groups at each visit time point. The Friedman’s test was also used to examine the change of dietary intake and diet quality over time. The Wilcoxon signed rank test was performed to investigate the change of dietary intake and diet quality between study entry and end of month six. Median regression analysis was used to evaluate the relationship between change in dietary factors and change in BMI z-score from study entry to end of induction or end of month six after adjustment of age, gender, ethnicity, and BMI z-score at study of entry. The R 4.0 and SAS 9.4 (Cary, NC) were used for data analysis.
RESULTS
Twenty-eight participants enrolled in the study between June 2017 through August 2019. Two participants withdrew due to feeling too overwhelmed by the diagnosis of ALL and three withdrew due to reluctance to complete study assessments; all withdrew within the first two months. Two additional participants were removed from the study prior to end of month six due to change in institution and clinical course, and thus did not receive the full intervention. The analysis represents 23 (82.1%) evaluable participants.
The demographic and clinical characteristics are presented in Table 1. The cohort reflected the demographics of childhood ALL with slightly more males (52.2%) than females (47.8%); 43.5% were Hispanic and 52.2% were white. Nearly 22% of participants were classified as obese at study entry; higher than previously reported studies in ALL.6,22 Most participants (82.6%) selected to initiate nutrition counseling at study entry. Of the four participants who elected to initiate nutrition counseling (N=4) later in the induction phase of treatment, the majority were female (75%), Non-Hispanic and white (75%), and younger (less than 10 years old) (75%).