Parental Perceptions of Children’s Weight Status in 22 Countries: The WHO European Childhood Obesity Surveillance Initiative: COSI 2015/2017

Childhood overweight and obesity are increasing worldwide, with the most rapid rises in low- and middle-income countries [1-3]. According to the World Health Organization (WHO) European Childhood Obesity Surveillance Initiative (COSI), the prevalence of overweight and obesity among 6- to 9-year-old children in Europe ranged from 9 to 43% in boys and 5–43% in girls in 2015–2017 [4]. This is of concern since childhood obesity is associated with insulin resistance, increased mechanical stress on joints, obstructive sleep apnea, impaired cardiovascular function and social functioning, stigma, and reduced quality of life and mental health [5-7]. Having obesity during childhood can also predispose individuals to many other chronic diseases in the long term, such as type 2 diabetes, hypertension, and cardiovascular disease [8-10].

Although identified as a global health priority, there are widening inequalities in the prevalence of childhood obesity and preventative progress has been poor [11, 12]. Addressing childhood obesity in an equitable manner requires a comprehensive approach that takes into consideration its complex causes. Root causes of childhood obesity include biological (e.g., genetics and epigenetic variants, intrauterine exposures, early nutrition, and microbiome) and environmental (both physical and social) factors [13]. Notwithstanding this scientific understanding of obesity, public and political discourse often simplifies obesity to, primarily, a consequence of unhealthy eating and physical inactivity [14]. Although global nutritional deficiencies and physical inactivity contribute to the development of many noncommunicable diseases including obesity, the oversimplification of obesity as a nutrition and physical activity problem may contribute to children and parents experiencing stigma, making it challenging for them to access evidence-based care [14, 15]. Stigma can also impact children’s health independently of any obesity-related complications and contribute to social inequalities at the population level [16, 17].

Addressing childhood obesity requires a life-course perspective while also addressing the root causes of obesity, well beyond nutrition and physical activity [12]. To date, however, interventions to prevent childhood obesity have been mainly focused on individual approaches to promote healthy eating and increase physical activity [18]. Policy interventions to prevent childhood obesity have also been limited due to a number of factors, including a lack of public funding and resources, commercial marketing of highly palatable and high-calorie foods, industry lobbying, and a lack of public understanding of obesity-related health and social outcomes [19].

At the health care system level, failure to recognize the seriousness of childhood obesity, particularly through a life-course lens of chronic disease prevention, has resulted in limited access to evidence-based treatment strategies [20]. Many healthcare professionals lack knowledge and skills to assess and diagnose childhood obesity, as evident by the low levels of routine diagnosis and treatment for childhood obesity in pediatric health care settings [21]. A recent study found that clinicians working across primary, secondary, and tertiary pediatric care levels had received only moderate training in obesity (48%), did not routinely measure height and weight (80%), and infrequently referred children to obesity services (25%) [22].

Parents can play an important part in the prevention and treatment of childhood obesity [19, 23] and family-based interventions to treat childhood obesity may be effective even if long-term effects remain unknown [24]. However, studies have found that parents may not be able to accurately perceive their children’s weight status [25-28]. Some factors that may influence parents’ perceptions of their child’s weight include child’s age, population prevalence of obesity, parental weight status, and internalized stigma. A systematic review, which pooled results from 51 studies published between 1990 and 2011, showed that 63.4% of parents were unable to recognize excess weight in their child; however, this number was higher (86%) for parents who had younger children (2–6 years) [28]. An Italian study also found that maternal perception of their child’s weight appears to be affected by the prevalence of obesity in their community [29]. Indeed, it has been suggested that parents may not be able to recognize obesity in their child because obesity has been normalized [30]. Considering that the social narrative of childhood obesity attributes blame to parents for their children’s food and exercise choices, internalized stigma may also be a factor driving parental perceptions of their children’s weight [31].

Parental inability to recognize children’s excess weight has been proposed as a contributing factor to the rising levels of childhood obesity [32]. The premise of this argument is that if parents were able to recognize their child’s obesity, they would be motivated to change their family’s behaviors and seek obesity treatment. However, while this contention may have face validity, there is an absence of evidence in its favor [33], and some to the contrary [34, 35]. A focus on parental recognition of their child’s weight status risks placing undue focus on families and individuals for a problem that is primarily driven by biological, environmental, commercial, social, and economic factors. Nevertheless, parents can have an important role as agents of change in health-promoting behaviors, and it is imperative that they are engaged in health promotion programs. To engage parents, we need to understand their perspectives [36]. At the population level, this knowledge can help us to identify opportunities for parental engagement to develop childhood obesity prevention programs and policies that are supportive, nonstigmatizing, and effective. Ultimately, health professionals and policy makers have a responsibility to support parents in accessing appropriate preventive and management care that can improve their child’s health and well-being.

The purpose of this study was to measure updated, global parental perceptions of their child’s weight status and to identify possible predictors of parental misperceptions.

This study uses data from the COSI. COSI is a unique system that has measured trends in overweight and obesity among primary school-aged children since 2007. COSI involves taking standardized height and weight measurements from over 300,000 children across the WHO European Region, providing nationally representative data for participating countries, as well as a large region-wide dataset for analysis of the determinants of childhood overweight and obesity.

In 2015–2017, the fourth round of COSI took place in 36 countries of the WHO European Region, providing national representative data in all countries expect Russian Federation where data collection was implemented only in the city of Moscow. Among these countries, 22 collected information on parents’ perceptions of their child’s weight status through COSI Family Record Form, namely Albania (ALB), Bulgaria (BUL), Croatia (CRO), Czechia (CZH), Denmark (DEN), Georgia (GEO), Italy (ITA), Kazakhstan (KAZ), Kyrgyzstan (KGZ), Latvia (LVA), Lithuania (LTU), Malta (MAT), Montenegro (MNE), Poland (POL), Portugal (POR), Romania (ROM), Russian Federation (only Moscow) (RUS), San Marino (SMR), Spain (SPA), Tajikistan (TJK), Turkey (TKM), and Turkmenistan (TUR) [37]. Data were gathered following a common protocol devised in 2007 by the WHO Regional Office for Europe and Member States [38] and minimally amended for COSI 2, 3, and 4 [39-41]. The COSI protocol is in accordance with the International Ethical Guidelines for Biomedical Research Involving Human Subjects [42]. Local Ethical Committees also approved the study.

A sampling approach was adopted for all countries except in Malta and San Marino, where the entire population of interest (i.e., all children in third-grade primary-school classes) was included. According to the COSI protocol, participating countries could select one or more of the following 4 age groups: 6.0–6.9, 7.0–7.9, 8.0–8.9, or 9.0–9.9 years. Out of the 22 countries, 12 countries only included 7 year olds (Bulgaria, Czechia, Denmark, Georgia, Lithuania, Malta, Montenegro, Portugal, Russian Federation – city of Moscow, Tajikistan, Turkey, and Turkmenistan); 5 only included 8 year olds (Albania, Croatia, Poland, Romania, and San Marino); 1 only included 9 year olds (Kazakhstan); and 4 included >1 age group (7.0–7.9 and 8.0–8.9 in Kyrgyzstan; 7.0–7.9 and 8.0–8.9 in Latvia; 8.0–8.9 and 9.0–9.9 in Italy; and all age groups in Spain). The effective sample size (i.e., the number of measured children per targeted age group) was mostly equal to or above the minimum suggested by the protocol [41] (i.e., 2,800 children per age group and sex). In 5 countries, the effective sample size was considerably lower (Czechia) or slightly lower (Denmark, Montenegro, Albania, Croatia, San Marino, and the Russian Federation). More information on the study design in each country is provided elsewhere [43].

Children’s measurements were conducted by staff trained to measure height and weight using the WHO standardized technique. The family questionnaires were completed by parents or caregivers. The form included questions about parent’s perception of their child’s weight status, as well as other simple indicators of children’s dietary intake, physical activity or inactivity patterns, the family’s socioeconomic characteristics, and the presence of comorbid conditions associated with obesity. Further details on the fourth round of COSI data and methodology are available elsewhere [41, 43].

We only included children belonging to COSI target age groups (i.e., children between the ages of 6 and 9 years) with complete information on data needed to assess their weight status (i.e., sex, age, height, and weight). The WHO COSI study evaluates parents’ perceptions about their child’s weight status through a question in the Family Record Form. The person completing the form indicated their relationship to the child. We included children who had at least one parent or caregiver that completed the form.

Children’s weight status classification was based on the 2007 WHO-recommended growth references for school-aged children and adolescents [44, 45], and cutoffs were used to compute BMI-for-age Z-scores and to estimate prevalence of overweight/obesity. Each child was classified in the underweight, normal weight, overweight (excluding obesity), or obesity category. A child was classified in the overweight (excluding obesity) category if she/he had overweight but not obesity according to the WHO definitions [46]. That is, thinness is defined as a BMI-for-age value of <−2 Z-score, while overweight and obesity are defined as a BMI-for-age value of > +1 Z-score and > +2 Z-scores, respectively. Children for which a biologically implausible (or extreme) BMI-for-age value was estimated (i.e., values below −5 or above +5 Z-scores relative to the 2007 WHO growth reference median) were excluded from the analysis [45]. BMI was calculated as weight (kg) divided by height squared (m²). We used the category “underweight” to indicate “thinness” as per the WHO classification system and definition.

The Family Record Form asks the question “In your opinion, is your child: underweight – normal weight – a little overweight – extremely overweight.” Since the Family Record form was completed by different family members, we differentiated between responses from mothers and fathers. If the Family Record Form was completed by the mother, data were referred to as “maternal perception,” whereas if the form was completed by the father, data were referred to as “paternal perception.”

We constructed a measure of concordance to indicate whether a parent accurately perceived their child’s weight status relative to the WHO obesity categories. The perceived and actual classification of children’s weight status was considered consistent in the following cases: (i) children with thinness were correctly classified as “underweight”; (ii) normal weight children as “normal weight”; (iii) children with overweight (excluding obesity) as “a little overweight”; and (iv) children with obesity as “extremely overweight.”

Parental underestimation of their child’s weight status was observed if (i) children with normal weight were perceived as “underweight”; (ii) children with overweight (excluding obesity) were perceived as “underweight” or “normal weight”; and (iii) children with obesity were perceived as “underweight,” “normal weight,” or “a little overweight.” Parental overestimation of their child’s weight status was observed if (i) children with thinness were perceived as “normal weight,” “a little overweight,” or “extremely overweight”; (ii) children with normal weight were perceived as “a little overweight” or “extremely overweight”; and (iii) children with overweight (excluding obesity) were perceived as “extremely overweight.”

The accuracy of parents’ perceptions of their child’s weight status was also assessed by estimating underweight, normal weight, overweight (excluding obesity), or obesity using International Obesity Task Force (IOTF) cutoff points, as these are widely used in the WHO European Region. According to the IOTF, these points are age (in months) and sex specific and correspond to a BMI >25 at the age of 18 years [47].

All country datasets were reviewed for inconsistencies and completeness in a standard manner at the WHO Regional Office for Europe before they were aggregated for the international data analysis.

For each country included in the analysis, the distribution of children according to the WHO and IOTF weight status classification was estimated along with the distribution by parental perception of child’s weight status (percentage values and 95% confidence intervals [CIs]). Percentages of accurate, overestimating, or underestimating perceptions were estimated for each mother or father and child pair. Misclassification levels were also calculated for children with normal weight, overweight (excluding obesity), and obesity separately. For children with obesity, parental underestimation was estimated separating children perceived as “a little overweight” from those seen as “normal weight.” All above-mentioned estimates were produced at the country level and by pooling together all countries. Misclassification levels were not estimated in 3 countries (Denmark, San Marino, and Tajikistan) because the number of children with obesity in the sample was too low.

In order to balance the contribution of each country to the pooled estimates and to limit the differences in children’s age as much as possible, pooled estimates were calculated only for children belonging to one target age group for each country: 7 year olds for 15 countries (Bulgaria, Czechia, Denmark, Kyrgyzstan, Georgia, Latvia, Lithuania, Malta, Montenegro, and Russian Federation – city of Moscow, Portugal, Spain, Tajikistan, Turkey, and Turkmenistan), 8 year olds for 6 countries (Albania, Croatia, Italy, Poland, Romania, and San Marino), and 9 year olds for 1 country (Kazakhstan).

To examine factors associated with parents’ misclassification of their child’s weight status, we conducted a multilevel logistic regression analysis that included only children with overweight (including obesity). We estimated the adjusted odds ratios (adjORs) and confidence limits (95% CI) of a parent incorrectly classifying their child’s weight status (compared to correct classification). The parental perception was considered incorrect when the mother or the father classified her or his child who had overweight (including obesity) as normal weight or underweight.

Predictors of parental perceptions of their child’s weight status included age in months, sex, BMI, urbanization grade of child’s residence place, and respondents’ characteristics (i.e., relationship with the child – mother or father, and parental educational attainment and weight status). For BMI, its squared value was also included. For educational attainment, 2 categories were considered: low-medium level (i.e., “primary school or less,” “secondary or high school,” and “vocational school”) and high level (“undergraduate or bachelor’s degree” and “master’s degree or higher”). Parental weight status was estimated using maternal/paternal BMI, which was calculated based on self-reported height and weight. We used WHO definitions to classify mothers/fathers in the normal weight, overweight, and obesity categories.

We conducted a multilevel logistic regression analysis at the country level as well as by pooling countries. All models included random effects for the primary sampling units. Primary sampling units were replaced by primary schools to account for random effects specified in the Poland model. Children with a missing value for any of the covariates were excluded from the regression analysis. The regression analysis was not carried out for Denmark, San Marino, and Tajikistan due to the limited number of available observations. The regression model for Czechia and Malta did not include information about the relationship of the respondent with the child because it was not collected. Data collected in Moscow city were not included in the analysis of factors associated with parents’ misclassification of their child’s weight because of the high level of missing data on parents’ characteristics.

The regression analysis on pooled dataset was only carried out with countries in which it was possible to estimate a specific-country model and with data available for all covariates (16 countries – all but Czechia, Denmark, Malta, and Russian Federation – city of Moscow, San Marino, and Tajikistan). The pooled model was estimated only for children belonging to one target age group for each country: 7 year olds for 10 countries (Bulgaria, Kyrgyzstan, Georgia, Latvia, Lithuania, Montenegro, Portugal, Spain, Turkey, and Turkmenistan), 8 year olds for 5 countries (Albania, Croatia, Italy, Poland, and Romania), and 9 year olds for 1 country (Kazakhstan). Pooled models were also estimated for boys and girls separately.

Poststratification weights to adjust for the sampling design, oversampling, and nonresponse were available for all countries that applied a sampling approach in round 4 of COSI (except for Lithuania) and were used in all analyses to infer the results from the sample to the population. For Lithuania, unweighted analysis was carried out. All analyses accounted for the cluster sample design. In the pooled analysis, an adjusting factor was applied to the poststratification weights to take into consideration the differences in the population size of the countries involved. The adjusting factor was calculated based on the number of children belonging to the target age group according to Eurostat figures or national official statistics for 2016. All statistical analyses were performed using the statistical software package Stata version 15.1.

Twenty-two countries from the WHO European Region were included in this study. A total of 124,296 children out of over 174,000 invited to participate in COSI Round 4 were eligible for inclusion (Table 1).

The total number of children included varied among countries, ranging from below 1,000 children in San Marino and Denmark to close to 10,000 in Spain and Turkey and over 40,000 in Italy. Multiple factors affected these figures. Country sample size was affected by the number of targeted age groups and the characteristics of the national school systems through which the children were enrolled [43]. In Italy, the sample size was considerably larger because regional and national estimates were included. The level of children’s participation in the survey also affected the number of available data. The proportions of children who participated in measurements also varied from country to country, ranging between 70 and 80% in Croatia, Russian Federation (city of Moscow), Latvia, and Spain to over 90% in Albania, Kazakhstan, Kyrgyzstan, Malta, Portugal, San Marino, Tajikistan, and Turkmenistan.

The lowest level of proportion with a completed family form was registered in Denmark (29.9%), Albania (36.2%), and Moscow city (52.6%). Finally, in most of the countries for almost all measured 6- to 9-year-old children with a completed family form, the information about mother/father perception of their child’s weight status was available (Table 1).

On average, children were 7.9 years old, with the majority being male (51.3%) and residing in urban areas (71.0%). Parents on average had a low-medium education level (64.8%) and were classified as normal weight (61.7%). Mothers completed the family form most often (85.9%). Details by the country can be found in Table 2.

Table 3 highlights children’s weight status classification based on the WHO definitions and parental perceptions. Overall, there were important differences between these categorizations, wherein 9.6% (95% CI: 9.1–10.1), 79.1% (95% CI: 78.5–79.8), 10.5% (95% CI: 10.0–10.9), and 0.8% (95% CI: 0.7–1.0 CI) of parents perceived their child’s weight as “Underweight,” “Normal weight,” “A little overweight,” and “Extremely overweight,” respectively, compared to 2.3% (95% CI: 2.0–2.5), 68.8% (95% CI: 67.9–69.7), 17.2% (95% CI: 16.6–17.8), and 11.7% (95% CI: 11.1–12.3) being classified as such, respectively, based on WHO standards. In all countries, parents more frequently classified their children as with underweight or normal weight than WHO classification (Table 3). This finding is confirmed also considering the IOTF classification of children’s weight status, even if the discrepancy with parents’ perception is less pronounced (online suppl. Table 1; for all online suppl. material, see www.karger.com/doi/10.1159/000517586).

There were differences in the accuracy of parents’ perceptions of their child’s weight status based on the actual WHO weight status categories (Table 4). In the overall population, which included all of the WHO weight status categories (children with normal weight, children with overweight and obesity, children with overweight, and children with obesity categories), 64.1% of parents classified their child’s weight status accurately relative to the WHO classification. Similarly, most parents (86.5%) were able to accurately assess their child’s weight if the child’s weight fell in the normal weight category. However, for children who were classified in the overweight category (excluding obesity), only 17.4% of parents accurately assessed their child’s weight status, while 82.3% underestimated their child’s weight status. Last, for children who fell in the obesity category, while 6.2% of parents accurately perceived their child’s weight status, 56.2% perceived their child’s as “a little overweight” and 37.2% as “normal weight.”

In regard to the IOTF classification of child weight status, parents’ perception was slightly more accurate (relative to the WHO classification), with 72% of parents classifying their child’s weight status accurately. However, there was still a high proportion of parents who underestimated their child’s weight status – 68.1% among children with overweight (excluding obesity) and 90.7% among children with obesity (online suppl. Table 2).

Parents’ underestimating perceptions of their child’s weight status by country are shown in Figure 1 and Table 4. Normal weight children were mostly accurately perceived by parents in all countries – accurate perception accounted for <80% of children only in 2 countries (Kyrgyzstan and Turkey). As for children with overweight or obesity, data showed a higher level of heterogeneity among countries. The percentage of children classified as with overweight (excluding obesity) according to WHO’s definition, and whose weight status was underestimated, varied from <70% in Latvia, Romania, and San Marino to >90% in Portugal, Tajikistan, Turkmenistan, and Turkey. This percentage increased to over 50% in 3 countries, with the highest values recorded in Turkmenistan (75.5%). In 7 countries, <25% of children in the obesity category were seen as “normal weight” by their parents.

We assessed predictors of parents’ perceptions about their child’s overweight (including obesity) weight status by estimating country-specific and pooled multilevel logistic models that included child’s age in months, sex, BMI, and urbanization grade of child’s residence place, as well as respondent’s characteristics (i.e., relationship with the child – mother or father, and parental educational attainment and weight status) (Table 5). Based on pooled regression analyses, parents were more likely to underestimate their child’s weight status if the child was male (adjOR: 1.41; 95% CI: 1.28–1.55), the mother/father had a low-medium level of educational attainment (ajdOR: 1.41; 95% CI: 1.26–1.57) and, to a less extent, if the father completed the form (adjOR: 1.14; 95% CI: 0.98–1.33) or the family lived in rural areas (adjOR: 1.10; 05% CI: 0.99–1.24). There were also important differences at the country level. First, boys with overweight (including obesity) in Georgia were 2.57 times more likely to be perceived as having normal weight or underweight. Second, parents with low-to-moderate educational attainment in Croatia, Kazakhstan, Montenegro, and Portugal were >2 times more likely to underestimate their child’s weight status. Third, fathers from Albania underestimated their child’s weight status more than mothers (adjOR estimate over 2). Finally, in Bulgaria, Turkmenistan, and Turkey, parents from rural areas were 2 times more likely to underestimate their child’s weight status.

There was an association between a child’s BMI and their parents’ perceptions about their own weight status. Although not entirely linear, the odds of parents underestimating their own weight status decreased as the child’s BMI increased. However, parents’ BMI was not strongly associated with underestimation of child’s weight status. However, in some countries, there was a stronger association between parents’ BMI and their ability to accurately perceive their child’s weight status. In Croatia, Italy, Lithuania, and Spain, for example, parents with overweight or obesity were more likely to underestimate their child’s weight status.

The pooled data were also analyzed separately for boys and girls (Table 5). The adjORs estimated through the regression analysis were similar for boys and girls. Nevertheless, the less accurate perception of fathers compared to mothers was more evident for boys (adjOR: 1.29; 95% CI: 1.04–1.60) than girls (adjOR: 1.05; 95% CI: 0.84–1.17).

The percentage of children with overweight (including obesity) who were perceived as “underweight” or “normal weight” by their parents is shown in Figure 2. These values varied from 49.4% in Italy to 89.4% in Turkmenistan, with other countries’ values ranging between 50–60% and 70–75%. After adjusting for the child’s age in months, sex, BMI, and urbanization grade of child’s residence place, as well as respondent’s characteristics, differences among countries in terms of parental underestimation were mostly confirmed.

Overall, parents in Croatia, Kazakhstan, Kyrgyzstan, Montenegro, Portugal, Spain, Turkey, and Turkmenistan were more likely to underestimate their children’s weight status (compared to Italy). On the other hand, parents in Latvia and Georgia were less likely to underestimate their child’s weight status (compared to Italy).

Consistent with previous studies, our results indicate that parents tend to underestimate their child’s weight status [26-28]. In our study, 82.3% and 93.8% of parents underestimated their child’s weight status in the overweight (excluding obesity) and obesity categories, respectively. Further, we found that fathers were more likely to underestimate weight status in boys with overweight (excluding obesity), and that parents with low-medium educational attainment underestimated boys’ more than girls’ weight status. Finally, our findings also show differences between countries. Cultural and gender differences such as social expectations of boys’ and girls’ body weight, shape, and size may influence parental perceptions; however, more studies are needed to explore how fathers in countries such as Turkmenistan, Kazakhstan, and Montenegro perceive obesity in their children (specifically boys).

As other parental perception studies have done, we used anthropometric measures (i.e., height, weight, and BMI) to classify children with overweight and obesity. However, parental perception studies have used different definitions to classify children with overweight and obesity. Some studies have used the US Centres for Disease Control and Prevention BMI categories, while others have used the IOTF standard or the WHO growth charts. The use of different cutoff values, standards, and growth charts can lead to different obesity classifications and may cause confusion [26]. Similar to other studies, we found that the level of underestimation of a child’s weight varies depending on the obesity classification system that is used. Specifically, our study found that using the IOTF standard, parents were less likely to underestimate their child’s weight status (online suppl. Table 2).

A lack of a clear definition and assessment protocol of childhood obesity limits the usability of parental perception studies. For example, a systematic review of studies conducted between 2006 and 2012 found that parents not only underestimated their child’s weight status, but they also did not express concern about their child’s excess weight [27]. Based on these studies, we simply do not know if parents understand the significance of weight status in relation to their child’s health risk, especially if the risk has not been assessed and communicated by a qualified healthcare provider [27, 48].

Some researchers argue that quantifying health in terms of numbers on a scale has resulted in changing body size ideals globally [49]. Changing body size ideals can be demonstrated through increasing weight bias and weight-based discrimination trends. As population levels of childhood obesity have increased, so have levels of weight bias and weight-based discrimination [50, 51]. Framing healthy eating and physical activity interventions based on weight outcomes can have unintended consequences such as children only exercising and eating healthy to lose weight or engaging in unhealthy weight control strategies as opposed to engaging in these healthy living activities for health improvement and overall well-being and quality of life [52, 53]. Indeed, a limitation of parental perceptions studies to date is that weight and BMI alone are not sufficient measures to diagnose clinical obesity because they simply do not provide enough information about health risk [54, 55]. Although BMI is used to track obesity rates at the population level, at the individual level it only serves as a screening tool. Obesity should be diagnosed by a qualified healthcare professional, conducting a full medical assessment to determine if the child’s weight is impairing their health. If a parent perceives their child to have health risks related to their excess weight, then it is essential for a qualified healthcare professional to conduct further clinical investigation and for healthcare services to be accessible and affordable for families [36, 56].

Various public health interventions have been developed to correct parents’ misperceptions about their child’s weight. In the USA and UK, schools have established initiatives to send BMI report cards to inform parents that their child may be at risk for or has developed obesity. However, there is no robust evidence that providing parents with this information leads to change in behavior. On the contrary, there is some evidence that it can be perceived as stigmatizing [57]. Specifically, children who were perceived by parents to have excess weight were at a greater risk for weight gain regardless of their actual weight and despite being more likely to be actively try to lose weight [57]. Children whose parents have identified them as having overweight also report viewing their body size negatively [57]. A possible explanation may be that obesity is highly stigmatized in society and identifying oneself or a child as having obesity could lead to unhealthy coping strategies, including unhealthy weight loss approaches, which may actually lead to more weight gain in the long run [58-60].

Promoting the idea that parents should be educated to recognize whether their child has obesity using anthropometric measures or visual cues about their child’s weight or size could have harmful unintended consequences. Health messages that imply parents are unaware or simply in denial of their child’s overweight or obesity may add to the pervasive stigma that children and parents experience. Internalized blame and social judgment related to parenting can reduce the quality of life and impede parent-child interactions, which may ultimately impact children’s well-being [31]. Furthermore, it is important to recognize that parents’ ability and motivation to take action for their child’s health vary from family to family. It is the responsibility of health care professionals to be aware of any assumptions they make about parents’ personal responsibility for their child’s weight and health as this can contribute to further weight bias and stigma. Specifically, healthcare professionals should avoid making assumptions about parents’ motivations to promote healthy behaviors in their children and explore the family’s understanding of health risks, their values, and inherent socioeconomic realities [36]. Importantly, although vital signs such as height, weight, and BMI are important surveillance tools, healthcare professionals should always conduct a full medical assessment to determine a child’s health risk, while also educating and supporting parents and children about the complexity of childhood obesity [36].

Our study has a number of limitations to acknowledge. First, missing or insufficient data from some countries may have resulted in inaccurate calculations of weight status perceptions and predictors. Second, results may be related to the wording of the question and variables on the record form. For example, “A little overweight” and “Extremely overweight” may have led some parents to misperceive their child’s weight status due to subjective interpretations and internalized weight bias. Third, completion of this question may have been subject to disclosure error in which parents purposefully did not accurately document their perceived classification of their child’s weight status. Despite these limitations, our findings underscore the importance of revisiting the way in which interventions for obesity are communicated and delivered, and the need to take stronger actions to include parents’ perspectives. Considering that many interventions to date have been criticized for casting blame and shame on parents for misperceiving their child’s weight status or for not being concerned about their child’s weight, special attention should be given to the way information about obesity risk is provided to parents. It is also critical for healthcare professionals to systematically assess children’s weight during growth and consider an ecological approach in addressing childhood obesity while practicing a nonjudgmental clinical environment to strengthen open communication and rapport with families.

Our findings also have a number of policy implications, including supporting parents and further research. Healthcare professionals should always conduct a full medical assessment to determine if a child’s growth (height and weight) may present a health risk, while also educating and supporting parents and children about the complexity of childhood obesity. Special attention should be given to the way information about healthy childhood growth and development is given to parents in order to avoid unintended consequences such as weight bias and stigma. Healthcare professionals may benefit from training on the assessment and holistic treatment of overweight and obesity supported by investments in primary care. In addition, quantitative and qualitative research at the country level is needed to understand how gender, parental education, and sociocultural factors impact perception about childhood growth including weight status, and how this can impact children. Further research is needed to understand the dynamics of communicating childhood growth, including weight, information, and health risks to parents.

Our study supplements the current literature on factors that influence parental perceptions of their child’s weight status. Public health interventions aimed at promoting healthy childhood growth and development should consider parents’ knowledge and perceptions, as well as the sociocultural contexts in which children and families live. Further research is needed to understand the dynamics of communicating with parents, in particular with fathers and parents of boys, to help ensure that families have an accurate understanding of obesity risks and access to resources and support to prevent and manage childhood obesity. Continued awareness and implications of childhood obesity are deserving of increased attention by stakeholders, including parents, school staff, healthcare professionals, and policy advisors.

We gratefully acknowledge all participating children, their parents, teachers, and principals for kindly volunteering to participate in the study. We also thank the examiners and regional and local supervisors/coordinators who collected data in each country. Furthermore, we would like to acknowledge the contribution of other researchers and/or principal investigators in this study.

The WHO COSI study protocol was approved by the International Ethical Guidelines for Biomedical Research Involving Human Subjects and all procedures were also approved by Local Ethical Committees in each country (ALB: Scientific Committee of Institute of Public Health, Decision number: 953, July 13, 2015; BUL: Project identification code – 060 – МП 325 – 68 COSI, February 25, 2016; CRO: Ethical Committee of the Croatian Institute for Public Health, Registry number: 80-2660/1-15, September 25, 2015; CZH: Ethical Committee of the Institute of Endocrinology, AZV MZČR 17-31670 A, June 20, 2016; DEN: Research and Innovation Organization, SDU, 10.829, June 27, 2016; GEO: Bioethics Council at National Center for Disease Control and Public Health of Georgia, Project identification code: 2019-52, 4 November 2019; ITA: National Institute of Health, Prot. PRE–739/15, November 10, 2015; KAZ: Scientific and technical program “Development and implementation of modern technologies for healthy lifestyle promotion and prevention of diseases based on the study of non-medical determinants of health among children,” 2015; KGZ: The Ethics Committee on compliance of research to ethical norms for medical research, Project identification code: No. 1/1, February 22, 2018; LVA: Central Medical Ethics Committee, Project identification code: 01-29.1/6, September 25, 2015; LTU: Lithuanian Bioethics Committee (Lietuvos bioetikos komitetas), Project identification code: 08-02-19, February 19, 2008 and bioethics authorization renewals in on January 04, 2010, January 09, 2013, and March 12, 2019; MAT: Malta did not go through Ethics Committee as all the work involved in COSI data collection and analysis is part and parcel of the existing School Health Service, which is an ongoing process; MNE: Ethical Committee of the Institute of Public Health of Montenegro, Project identification code: WHO 2016/627456-0, April 28, 2016; POL: Bioethics Committee of the Institute of Mother and Child, Project identification number: 22/2015, November 26, 2015; POR: National Commission of Data Protection; Aut nº5418/2016 for all rounds of COSI Portugal, June 07, 2016; ROM: Intern Ethical Committee of the National Institute of Public Health, Project identification code: WHO 2016/650301-0, April 06, 2016, RUS: National Institute of Health., Prot. PRE–739/15, November 2015; SMR: Ethical Committee of the National Institute of Health, Prot. PRE – 739/15, November 10, 2015; SPA: Data for COSI in Spain were collected as part of the ALADINO Spanish Study, which did not ask for an Ethics Committee approval, since it is not mandatory in Spain. However, the Principal Investigators confirm that the study was conducted in accordance with the Declaration of Helsinki and all parents/guardians of subjects participating in round 4 gave their informed consent for inclusion before they participated in the study; TJK: Ministry of Health and Social Protection of Tajikistan, Project identification code: #858, November 18, 2016; TUR: Keçiören Training and Research Hospital, Clinical Researches Ethics Committee, October 26, 2016; TKM: Ethics approval was granted from the Ministry of Healthcare and Medical Industry (MOHMI), Project identification code: Health System Strengthening and Support Project, LN: 8531-TR, L.2.12, Obesity Fighting Project, Subcomponent 1.1, L.2.11, 2015). Furthermore, children’s parents or guardians gave their written informed consent.

The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.

The authors gratefully acknowledge support from a grant from the Russian Government in the context of the WHO European Office for the Prevention and Control of NCDs. Data collection in the countries was made possible through funding by: Albania: World Health Organization through the Joint Programme on Children, Food Security and Nutrition “Reducing Malnutrition in Children,” funded by the Millennium Development Goals Achievement Fund, and the Institute of Public Health; Bulgaria: Ministry of Health, National Center of Public Health and Analyses, World Health Organization Regional Office for Europe; Croatia: Ministry of Health, Croatian Institute of Public Health and World Health Organization Regional Office for Europe; Czechia: Grants AZV MZČR 17-31670 A and MZČR – RVO EÚ 00023761; Denmark: Danish Ministry of Health; France: French Public Health Agency; Georgia: World Health Organization; Ireland: Health Service Executive; Italy: Ministry of Health; Istituto Superiore di sanità (National Institute of Health); Kazakhstan: Ministry of Health of the Republic of Kazakhstan and World Health Organization Country Office; Latvia: n/a; Lithuania: Science Foundation of Lithuanian University of Health Sciences and Lithuanian Science Council and World Health Organization; Malta: Ministry of Health; Montenegro: World Health Organization and Institute of Public Health of Montenegro; Poland: National Health Programme, Ministry of Health; Portugal: Ministry of Health Institutions, the National Institute of Health, Directorate General of Health, Regional Health Directorates and the kind technical support of Center for Studies and Research on Social Dynamics and Health (CEIDSS); Romania: Ministry of Health; Russia (Moscow): n/a; San Marino: Health Ministry; Educational Ministry; Social Security Institute; the Health Authority; Spain: Spanish Agency for Food Safety and Nutrition (AESAN); Tajikistan: World Health Organization Country Office in Tajikistan and Ministry of Health and Social Protection; and Turkmenistan: World Health Organization Country Office in Turkmenistan and Ministry of Health. The authors alone are responsible for the views expressed in this article and they do not necessarily represent the views, decisions, or policies of the institutions with which they are affiliated.

X.R.S., M.B., and J.W. conceptualized and designed the study and were responsible for the overall direction and planning. M.B. performed data analysis and drafted and designed figures and tables. X.R.S and M.K. wrote the manuscript with support from J.W., A.S., P.N., A.R., and J.B. All authors contributed to the design and implementation of the study, the analysis of the results, and the editing of the manuscript. All authors approve the final version of the manuscript.

The data that support the findings of this study are available from the World Health Organization.