Data Analysis
Dietary data from the 24-hour recalls was analysed using the FoodWorks nutrient analysis software program (version 8, 2015; Xyris Software, Queensland, Australia). For food and drink items missing from the FoodWorks database, nutrient content was obtained from product nutrition panels and entered manually. Serves of food group intake (fruits, vegetables, grains, meat/alternatives and milk/alternatives) were automatically calculated by FoodWorks dietary analysis software (Xyris software, Version 8). Foods were then classified as core or discretionary foods by a researcher according to the 2013 Australian Dietary Guidelines 34 whereby discretionary foods are described as higher in energy density, saturated fat, sodium, sugars and or alcohol). Manual extractions of discretionary foods were required in some instances. For example, fruit juice is considered a discretionary item so was separated from fruit intake. Food intake data was compared to age and sex appropriate recommended serves for Australian children34. Dietary intake results for children in the general Australian population and were obtained from the 2011-12 Australian Health Survey 35,36 and used as a comparative norm. Food group intake was expressed as a percent of serves recommended by the Australian Dietary Guidelines 34 to allow for comparisons between age groups.
A nutrient was included for dietary assessment if an appropriate Nutrient Reference Value or Australian Dietary Guideline recommendation existed. Incorporating all nutrients possible contributes to a thorough assessment of dietary quality. Children’s nutrient intake was expressed as a percentage of their age and sex appropriate estimated average requirement (EAR) or adequate intake (AI) when EAR was not available37. The estimated energy requirement of each child was calculated using age appropriate Schofield equations38 multiplied by a physical activity level of 1.5 (sedentary)39. The Schofield equation calculates an estimation of a person’s basal metabolic rate based on their age, gender and weight 40. Total energy intake was expressed as a percentage of estimated energy intake (%EER) for each child which allowed mean %EER to be calculated. Intake of carbohydrate, protein and fat was also expressed as a %EER to determine macronutrient distribution ranges and compare these to the acceptable macronutrient distribution ranges for reduced risk of chronic disease34. Descriptive statistics for demographic and dietary intake data were calculated using SPSS statistics (version 21.0, 2012; IBM Corp., Armonk, NY).
Assumptions of normality did not appear to hold for some variables, therefore for consistency non-parametric tests were used. Non-parametric tests (Mann-Whitney and Kruskall-Wallis) were used to compare the difference in dietary intake between patient sex (male versus female), diagnosis (ALL versus other) and treatment intensity. As children treated for ALL are at a higher risk of obesity than other diagnoses21, a comparison of the dietary intake between children with ALL and other diagnoses was undertaken. Statistical significance was set at a level of 0.05 with no adjustment for multiple comparisons.