Elsevier

Nutrition

Volume 53, September 2018, Pages 54-58
Nutrition

Applied nutritional investigation
Self-reported weight and predictors of missing responses in youth

https://doi.org/10.1016/j.nut.2018.01.003Get rights and content

Highlights

  • Missing self-reported weight information is a common problem in youth studies but such data are rarely analyzed.

  • There was high percentage of missing self-reported weight information in the 2014 Scottish Health Behaviour in School-aged Children study.

  • Self-reported weight non-responses in this population were missing due to a systemic error.

  • Self-reported weight non-response was associated with sociodemographic factors.

  • Self-reported weight non-response was associated with behavioral factors.

Abstract

Objective

The aims of the present manuscript are to analyse self-reported data on weight, including the missing data, from the 2014 Scottish Health Behaviour in School-Aged Children (HBSC) Study, and to investigate whether behavioural factors related with overweight and obesity, namely dietary habits, physical activity and sedentary behaviour, are associated with weight non-response. Research Methods & Procedures: 10839 11-, 13- and 15-year-olds participated in the cross-national 2014 Scottish HBSC Study. Weight missing data was evaluated using Little's Missing Completely at Random (MCAR) test. Afterwards, a fitted multivariate logistic regression model was used to determine all possible multivariate associations between weight response and each of the behavioural factors related with obesity.

Results

58.9% of self-reported weight was missing, not at random (MCAR p < 0.001). Weight was self-reported less frequently by girls (19.2%) than by boys (21.9%). Participants who reported low physical activity (OR 1.2, p < 0.001), low vegetable consumption (OR 1.24, p < 0.001) and high computer gaming on weekdays (OR 1.18, p = 0.003) were more likely to not report their weight.

Conclusions

There are groups of young people in Scotland who are less likely to report their weight. Their weight status may be of the greatest concern because of their poorer health profile, based on key behaviours associated with their non-response. Furthermore, knowing the value of a healthy weight and reinforcing healthy lifestyle messages may help raise youth awareness of how diet, physical activity and sedentary behaviours can influence weight.

Introduction

Childhood obesity is one of the greatest worldwide public health challenges of the 21st century. One in 10 young people ages 5 to 17 y are estimated to be overweight or obese and these levels have increased rapidly in recent decades [1]. Research has identified dietary practices, decreased physical activity, and increased sedentary lifestyle as the main factors that contribute to this epidemic [2], [3], [4].

Recently, the Health Behaviour in School-aged Children (HBSC) collaborative, cross-national study by the World Health Organization released its report with regard to adolescent obesity trends and related behaviors in the World Health Organization European Region from 2002 to 2014 [1]. This study collected international data for more than 25 y on the health behaviors, health outcomes, and social environments of young people ages 11, 13, and 15 to allow for a comparison of data across countries and over time [5].

The HBSC study aimed to analyze behaviors that are established during adolescence and can continue into adulthood and that affect issues such as mental health, development of health complaints, tobacco use, diet, physical activity levels, and/or alcohol use. The study measured obesity using body mass index (BMI) based on self-reported weight and height information. This measurement is commonly used in population-based surveys and has been shown to be a useful and cost-effective tool to estimate weight status (e.g., overweight and obesity) in large epidemiologic studies [6], [7], [8], [9], [10], [11]. However, measurements of this type may be subject to recall or bias and potentially lead to either an over- or underestimation of the results [12], [13]. Among young people, self-reported weight measurements can be influenced by sex, pubertal status, adiposity, body image concerns, food choices, sociocultural environment influences, awareness of social ideals with regard to slimness, and attitudes toward obesity among others [7], [14], [15].

Previous studies have reported a considerable proportion of missing data in population-level surveys of adolescent health and have indicated that weight data are more likely to be missing than height data, particularly for girls [7], [10]. The likelihood of missing self-reported weight data may increase as the actual BMI increases, which could indicate that the non-response may be intentional among young people [7]. Although there is no established cutoff in the literature with regard to an acceptable percentage of missing information in a data set for valid statistical inferences, the quality is directly related to the missing data proportion [11], [16]. Consequently, in the latest HBSC obesity report, data from nine countries were excluded from the analysis due to their high levels (>30%) of missing data: Belgium, Ireland, Israel, Lithuania, Malta, Romania, England, Wales, and Scotland [1].

With the purpose of understanding the rationale behind the shortage of weight responses and prevent this in future studies, the aims of the present study are (1) to analyze if the weight non-response in the 2014 Scottish HBSC study is missing either at random or due to a systemic error and (2) to investigate whether the non-response with regard to weight data can be associated with behavioral factors that are related to obesity (i.e., dietary habits, physical activity, and sedentary behaviors).

Section snippets

Setting and participants

The analysis of missing weight response data was taken from the Scottish sample of the 2014 HBSC study [17], [18], which was designed to be nationally representative of 11, 13, and 15 y old children in Scotland. The survey was conducted in schools using classrooms as the primary sampling unit and all students in the selected classrooms were asked to complete a questionnaire anonymously. The target population was school children in the final year of primary school (average age: 11.5 y) and in

Results

A total of 10839 school children ages 11, 13, and 15 y participated in the Scottish 2014 HBSC study. A low percentage of self-reported weight data (41.1%) was present in this sample, which resulted in 58.9% of missing data (Table 1). Weight was self-reported less frequently by girls than boys (19.2% and 21.9%, respectively) and despite variations in the rates of self-reported weight responses in the different age groups, the rates remained lower for girls than for boys for all age groups.

Discussion

In the Scottish 2014 HBSC study, a high percentage of missing data was observed for self-reported weight among adolescents age 11, 13, and 15 y. The present study shows that weight non-response in this population was missing due to a systemic error. Moreover, this weight non-response was associated with predictors such as sociodemographic determinants (i.e., age, sex, and family affluence) and behavioral factors such as vegetable consumption, physical activity practice, and computer usage for

Conclusions

This study found that groups of young people in Scotland are less likely to report their weight information. These young people's weight status may be of the greatest concern because of their poorer health profile on the basis of key behaviors that are associated with their non-response. The increasing numbers of children and adolescents who are overweight and obese in recent years have important implications for both current and future health outcomes. Monitoring the overweight and obese

References (30)

  • E.A. Spencer et al.

    Validity of self-reported height and weight in 4808 EPIC-Oxford participants

    Public Health Nutr

    (2002)
  • H. Fonseca et al.

    Validity of BMI based on self-reported weight and height in adolescents

    Acta Paediatr

    (2010)
  • L. Béghin et al.

    Nutritional and pubertal status influences accuracy of self-reported weight and height in adolescents: the HELENA Study

    Ann Nutr Metab

    (2013)
  • H. Fonseca et al.

    Perception of overweight and obesity among Portuguese adolescents: an overview of associated factors

    Eur J Pub Health

    (2005)
  • F.J. Elgar et al.

    Validity of self-report screening for overweight and obesity. Evidence from the Canadian Community Health Survey

    Can J Pub Health

    (2008)
  • Cited by (14)

    View all citing articles on Scopus

    The authors have no conflicts of interest to declare.

    View full text