Perceptions of Child Weight Status by Parents of Children on Medicaid

Background: Parental perception of child’s size has been evaluated in previous studies, but the degree of inaccuracy has been overlooked. In addition, parents of children on Medicaid may be more likely to have inaccurate perception of their child’s size. The objectives of this study were to assess the rate of overweight children, document the degree of discrepancy between parents’ perceptions and their children’s actual weight status, and identify factors related to inaccurate parental perception in a population predominantly insured by Medicaid. Methods: Participants in the cross-sectional survey included 241 parents of children age three to 12 years, who were English or Spanish speaking, and at least 18 years of age. Surveys asked parents to identify their child’s size based on a 5-point likert-type scale. Results: A Body Mass Index (BMI) at the 95 th percentile or above was found for 30.3% of children. Parents were only correct 39.8% of the time when describing their child’s weight status. In fact, 39.4% underestimated their child’s weight by one BMI category, 17.4% by 2 categories, and 1.7% by 3 categories. Parental accuracy decreased as child’s weight status increased. No other measured characteristics significantly impacted parental accuracy. Conclusions: Children on Medicaid have high levels of overweight and nearly 20% of parents underestimated their child’s size by at least 2 BMI categories. Parental perception needs improvement before interventions are likely to be effective. KJM 2009; 2(4):78-86. Introduction One of the major health concerns facing children today is the rising epidemic of obesity. Overweight children face a plethora of physical issues that can affect them for the rest of their lives, 1,2 including type 2 diabetes, hypertension, dyslipidemia, hepatic steatosis, cholelithiasis, sleep apnea, menstrual abnormalities, impaired balance, and orthopedic problems. 2 In addition, a social stigma is associated with larger body mass that may result in teasing, stereotyping, discrimination, and social marginalization. 2,3 These social issues may increase depression and anxiety, while decreasing esteem 2 and perceived cognitive and physical ability, 4 further impairing overall quality of life. Institutions such as government, public health programs, and school systems, as well as individual health care workers continue to combat the growing prevalence of childhood obesity, but parent’s involvement is instrumental in helping children stay healthy. Children whose mothers perceived them as overweight as infants or toddlers lost more weight by the time they were seven than children whose mothers perceived them as just right or underweight. 5 Also, parents were more likely to actively address their child’s excess weight if they perceive their child’s weight as a health problem. Studies in the United States 6-12 , Australia 13,14 , Argentina 15 , Canada 16 , Germany 5 , Italy 17 , New Zealand 18 , and the United Kingdom 19,20 have assessed parents’ perceptions of their child’s weight and found between 38% 5 and 94% 19 of parents were unable to assess their child’s weight Kansas Journal of Medicine 2009 Perceptions of Child Weight Status 79 accurately. Several predictors of accuracy have been identified, including child’s age 1,5,11,14 , gender 11,16,20 , ethnicity 16 , and Body Mass Index (BMI) 7,18,19 . While many studies have looked at predictors of inaccuracy, few have measured the degree of inaccuracy. It can be extrapolated from Eckstein et al. 6 that 2% of respondents whose children were a little overweight to overweight classified their child as underweight to a little underweight. Miller and colleagues 18 found that 6% of parents under-classified their child’s weight by two BMI categories. However, neither of these studies actually examined how far off parental perception was from reality. The only identified study that looked at degree of inaccuracy based the comparison on a visual analog scale of measurement. 7 If the child’s actual BMI was greater than or equal to the 95th percentile, 89.5% of parents were inaccurate by 30 percentage points or more, suggesting at least a two category discrepancy. However the visual analog scale only included the anchors of “Extremely underweight” and “Extremely overweight” so parents may not have been aware of which category their response would fall. Degree of inaccuracy is an important component in understanding parental perception of child weight and additional research is warranted. Greater variance between parental perception and actual child size may indicate a need for a paradigm shift to improve parental understanding of child weight issues. Helping parents accurately understand their child’s size may improve adherence to and efficacy of interventions to decrease child weight and improve health. However, before parental inaccuracy can be addressed, the degree to which parents misidentify their child’s size must be understood. Parents of children who receive Medicaid may be at increased risk for inaccurate perception of child weight. Children with Medicaid or no insurance have greater morbidity and mortality than children on private health insurance 21 and are more likely to be diagnosed overweight. 22 In addition, children on Medicaid exhibit lower literacy levels than the general population and may have more difficulty understanding health-related materials. 23 In previous studies, maternal perception of child weight was assessed in an urban Women, Infants and Children (WIC) program 8 and in a Head Start program 9 . Both populations often rely on Medicaid for their health care needs, 24,25 however, both studies were limited to Hispanic populations. Wald et al. 10 looked at parental perceptions in a pediatric primary care setting, but only a third utilized Medical Assistance. Further investigation is needed to determine accuracy of parental perception of weight status in a Medicaid population. The purpose of this study was to assess the rate of overweight children, the level of discrepancy between parents’ perceptions and their children’s actual weight status, and factors related to inaccurate parental perception in a population predominantly insured by Medicaid. Methods Data instrument. A one page, multiplechoice survey was developed regarding health and safety of children. The primary question of interest addressed parental perception of child’s weight status. Previous studies used several methods to assess parental perception of weight including a visual analog scale 7 , pictures of children 9 , drawings of children 6 , yes/no questions 8,11 , and four-point likert-type questions 10,15,18 . However, the most commonly used format involved a five-point likert-type question 5,6,12,14,16,17,19,20,26,27 which was chosen for this study. Response options Kansas Journal of Medicine 2009 Perceptions of Child Weight Status 80 included: “too skinny”, “thin”, “just right”, “somewhat overweight”, and “overweight”. The secondary set of questions included topics such as nutritional habits (servings of fruit and vegetables, frequency of fast food use), helmet use, hours of television viewed, hours of sleep per night, time spent in exercise, neighborhood safety, and proximity to closest park. These questions were included to allow parents to answer the survey without bias towards weight issues. 7,28 The survey was available in English and Spanish and written below the 6 th grade language level, as recommended for comprehension in at-risk populations. 29 Surveys were collected from March to August of 2007. Institutional Review Board approval was obtained from the medical school and the hospital where the ambulatory pediatric services were provided. Data collection. Parents had to be English or Spanish speaking, at least 18 years old, and the parent of a child age three to 12 years with an appointment at the Wesley Pediatrics Clinic (WPC). WPC provides services to many of the underserved children in Sedgwick County. Approximately 95% of patients at WPC have Medicaid benefits. Physicians staffing the clinic were pediatric and internal medicine/pediatric residents working under the supervision of pediatric faculty from the University of Kansas School of MedicineWichita. The survey was distributed to parents upon arrival for their child’s scheduled appointment. An information sheet attached to the survey explained the study goals and consent for responses to be used for research purposes was implied by survey completion. Parents were blinded to the exact topic of study to avoid bias. Surveys were completed in the waiting room and given to the child’s nurse when the patient was called to an exam room. Patients were included in the study regardless of the reason for seeking medical care. WPC staff performed routine height and weight measures per clinic procedures. Physical measurements were collected the same day the survey was completed. Following weight and height measurement, WPC staff placed the survey in a closed collection box. Variables and definitions. Body Mass Index (BMI) was calculated for the patients by taking the weight in kilograms divided by the square of height in meters (kg/m 2 ). Patient BMI results were classified into groups as follows: underweight (BMI less than 5 th percentile), normal (BMI 6 th through 84 th percentile), at risk of overweight (AROW; BMI 85 th through 94 th percentile), overweight (BMI 95 th percentile and above). Growth charts used to identify BMI percentiles were the May 2000 revised charts by the National Center for Health Statistics (NCHS). 30 Due to low response rates in the categories of “too skinny” and “thin”, parent perceptions of children’s weight were truncated from a five-point scale to a fourpoint scale concurrent with the BMI categories reported above. The categories “too skinny” and “thin” were combined to form an underweight group, “just right” became the normal group, “somewhat overweight” became the AROW group, a


Introduction
One of the major health concerns facing children today is the rising epidemic of obesity.Overweight children face a plethora of physical issues that can affect them for the rest of their lives, 1,2 including type 2 diabetes, hypertension, dyslipidemia, hepatic steatosis, cholelithiasis, sleep apnea, menstrual abnormalities, impaired balance, and orthopedic problems. 2In addition, a social stigma is associated with larger body mass that may result in teasing, stereotyping, discrimination, and social marginalization. 2,3hese social issues may increase depression and anxiety, while decreasing esteem 2 and perceived cognitive and physical ability, 4 further impairing overall quality of life.
Institutions such as government, public health programs, and school systems, as well as individual health care workers continue to combat the growing prevalence of childhood obesity, but parent's involvement is instrumental in helping children stay healthy.Children whose mothers perceived them as overweight as infants or toddlers lost more weight by the time they were seven than children whose mothers perceived them as just right or underweight. 5lso, parents were more likely to actively address their child's excess weight if they perceive their child's weight as a health problem.
While many studies have looked at predictors of inaccuracy, few have measured the degree of inaccuracy.It can be extrapolated from Eckstein et al. 6 that 2% of respondents whose children were a little overweight to overweight classified their child as underweight to a little underweight.Miller and colleagues 18 found that 6% of parents under-classified their child's weight by two BMI categories.However, neither of these studies actually examined how far off parental perception was from reality.The only identified study that looked at degree of inaccuracy based the comparison on a visual analog scale of measurement. 7If the child's actual BMI was greater than or equal to the 95th percentile, 89.5% of parents were inaccurate by 30 percentage points or more, suggesting at least a two category discrepancy.However the visual analog scale only included the anchors of "Extremely underweight" and "Extremely overweight" so parents may not have been aware of which category their response would fall.
Degree of inaccuracy is an important component in understanding parental perception of child weight and additional research is warranted.Greater variance between parental perception and actual child size may indicate a need for a paradigm shift to improve parental understanding of child weight issues.Helping parents accurately understand their child's size may improve adherence to and efficacy of interventions to decrease child weight and improve health.However, before parental inaccuracy can be addressed, the degree to which parents misidentify their child's size must be understood.
Parents of children who receive Medicaid may be at increased risk for inaccurate perception of child weight.Children with Medicaid or no insurance have greater morbidity and mortality than children on private health insurance 21 and are more likely to be diagnosed overweight. 22In addition, children on Medicaid exhibit lower literacy levels than the general population and may have more difficulty understanding health-related materials. 23n previous studies, maternal perception of child weight was assessed in an urban Women, Infants and Children (WIC) program 8 and in a Head Start program 9 .Both populations often rely on Medicaid for their health care needs, 24,25 however, both studies were limited to Hispanic populations.Wald et al. 10 looked at parental perceptions in a pediatric primary care setting, but only a third utilized Medical Assistance.Further investigation is needed to determine accuracy of parental perception of weight status in a Medicaid population.
The purpose of this study was to assess the rate of overweight children, the level of discrepancy between parents' perceptions and their children's actual weight status, and factors related to inaccurate parental perception in a population predominantly insured by Medicaid.

Methods
Data instrument.A one page, multiplechoice survey was developed regarding health and safety of children.The primary question of interest addressed parental perception of child's weight status.Previous studies used several methods to assess parental perception of weight including a visual analog scale 7 , pictures of children 9 , drawings of children 6 , yes/no questions 8,11 , and four-point likert-type questions 10,15,18 .
The secondary set of questions included topics such as nutritional habits (servings of fruit and vegetables, frequency of fast food use), helmet use, hours of television viewed, hours of sleep per night, time spent in exercise, neighborhood safety, and proximity to closest park.These questions were included to allow parents to answer the survey without bias towards weight issues. 7,28he survey was available in English and Spanish and written below the 6 th grade language level, as recommended for comprehension in at-risk populations. 29urveys were collected from March to August of 2007.Institutional Review Board approval was obtained from the medical school and the hospital where the ambulatory pediatric services were provided.
Data collection.Parents had to be English or Spanish speaking, at least 18 years old, and the parent of a child age three to 12 years with an appointment at the Wesley Pediatrics Clinic (WPC).WPC provides services to many of the underserved children in Sedgwick County.Approximately 95% of patients at WPC have Medicaid benefits.Physicians staffing the clinic were pediatric and internal medicine/pediatric residents working under the supervision of pediatric faculty from the University of Kansas School of Medicine-Wichita.
The survey was distributed to parents upon arrival for their child's scheduled appointment.An information sheet attached to the survey explained the study goals and consent for responses to be used for research purposes was implied by survey completion.Parents were blinded to the exact topic of study to avoid bias.Surveys were completed in the waiting room and given to the child's nurse when the patient was called to an exam room.Patients were included in the study regardless of the reason for seeking medical care.
WPC staff performed routine height and weight measures per clinic procedures.Physical measurements were collected the same day the survey was completed.Following weight and height measurement, WPC staff placed the survey in a closed collection box.
Variables and definitions.Body Mass Index (BMI) was calculated for the patients by taking the weight in kilograms divided by the square of height in meters (kg/m 2 ).Patient BMI results were classified into groups as follows: underweight (BMI less than 5 th percentile), normal (BMI 6 th through 84 th percentile), at risk of overweight (AROW; BMI 85 th through 94 th percentile), overweight (BMI 95 th percentile and above).Growth charts used to identify BMI percentiles were the May 2000 revised charts by the National Center for Health Statistics (NCHS). 30ue to low response rates in the categories of "too skinny" and "thin", parent perceptions of children's weight were truncated from a five-point scale to a fourpoint scale concurrent with the BMI categories reported above.The categories "too skinny" and "thin" were combined to form an underweight group, "just right" became the normal group, "somewhat overweight" became the AROW group, and "overweight" remained the overweight group.
Data analyses.Analyses were conducted using SPSS 15.0 for Windows. 31requencies, means, and standard deviations were computed for demographic information, while chi-square statistics were computed for categorical and ordinal data.An independent t-test was run to determine the effect of the continuous variable of child age on parental accuracy.

Results
Participants.A total of 290 surveys were collected.Only 241, however, included the information necessary to study parent perception of child's weight status, child's weight, and height.Children were age three to 12 years with an average age of 6.38 (SD = 2.781).Slightly more male (122; 50.6%) than female children were evaluated.Equal numbers of participants (31%) were Hispanic or African American; 21% were white, 8% "other", and the remaining 8% chose not to identify their ethnicity.The majority of surveys completed were in English (83.4%).
BMI.The rate of overweight children (BMI 95 th percentile or above) at the WPC was 30.3% and the highest percentage of overweight children were African-American (Table 1).However, chi-square analysis found no significant difference between ethnic/racial groups (χ 2 (9) = 9.082, p = 0.430).Similar distributions of BMI were found based on gender; again, no significant difference was found (χ 2 (3) = .991,p = 0.803).Table 1.Percent (and number) of children in each BMI group by race and gender.
* Numbers missing from the total indicate missing data.
Parental perception.Parents of pediatric patients at WPC were only correct 39.8% of the time when describing perception of their child's weight status (Table 2).In fact, 39.4% underestimated their child's weight by one BMI category, 17.5% underestimated by two categories, and 1.7% underestimated by three categories; in contrast, only 1.6% overestimated by one category.
To determine whether differences existed in perception based on child characteristics, chi-square analyses were run to compare parental accuracy to child's race, gender, and BMI group, respectively.To avoid expected cell counts less than five, race was limited to Hispanic, white, and African-American.The results by race were not significant (χ 2 (2) = 5.483, p = 0.064).Gender also did not significantly affect accuracy (χ 2 (1) = 0.467, p = 0.494).However, a comparison of normal, AROW, and overweight children found that BMI group was significantly related to parental accuracy (χ 2 (2) = 66.527, p < 0.001).

All * White
African-American Hispanic Other Male Female Less than normal 3.3% Finally, an independent t-test was computed to identify whether the child's age affected parental accuracy.No difference was found between the age of children whose parents accurately classified their weight (M = 6.50, SD = 0.286) and the age of children whose parents were inaccurate (M = 6.30,SD = 0.230; t(239) = 0.555, p = 0.579).

Discussion
BMI.In 2007, the prevalence of overweight was 14.0% for youths in Kansas and the highest rate belonged to Washington, D.C. at 22.8%. 32For our sample, the rate of overweight was 30.3%, suggesting a much higher rate than the Kansas average and greatly exceeding the highest national rate.
Parental perception.Accuracy of parental perception for a sample of children on Medicaid was extremely poor at less than 40%, however, accuracy was not lower than previously published studies on more general populations.Parents were accurate in identifying their child's size 62% of the time if the child's BMI was below the 85 th percentile, compared with an accuracy of only 9% when the child's BMI was greater than or equal to the 85 th percentile.This supported previous research findings suggesting parents are less accurate as child's weight increases. 18,19creased accuracy identifying normal weight children may be inflated due to the wider range of weights (6 th to 84 th percentile) found within this category.
In our study, we were interested in measuring not only the rate of inaccuracy, but also the degree of inaccuracy.We found nearly 2% of parents were inaccurate by three categories (identifying overweight children as underweight) and over 17% of parents were inaccurate by two categories (identifying overweight children as normal or AROW children as underweight).These rates are higher than those previously reported in the literature. 6,7,18Degree of inaccuracy is important because the less able parents are to perceive their child's true size, the less likely they are to engage in or support interventions to decrease their child's weight.Contrary to previous findings our study did not find significant differences in parental perception by gender or age. 1,5,14,16One possible reason could be due to differences in children's ages included in the studies.However, differences in findings between our study and previously published studies may have more to do with the characteristics of the parents than the characteristics of the children.
Medicaid recipients throughout the United States tend to be poor, unemployed, and have low educational attainment. 23arents of overweight children with Medicaid may associate with other parents and children in the same socioeconomic group which could influence their perception of normal weight.Low income mothers' beliefs about overweight also appear to differ greatly from those of the medical and health care communities, 9,33 as growth charts, the most common tool for identifying overweight status, lack face validity for low-income parents. 33,34Lower income parents also tend to characterize overweight by functional impairment and felt a child should not be labeled overweight unless mobility is compromised. 9,33he only other issue reported as cause for concern about overweight was if a child was being teased and suffering from low self-esteem as a result. 9,33Dominant among low-income mothers' beliefs about their children's weight is that growth and weight are predestined, therefore, out of the mother's control. 33,35Inaccurate parental beliefs about child weight may lead parents to disregard medical advice as irrelevant if the health professional fails to take into account the individual child and family when making suggestions for remedy. 20,27,33owever, parents are more likely to identify weight issues in their child if their doctor had addressed the child's weight. 27ddressing parental perception of child weight and the ramifications of excess weight is the first step in addressing childhood overweight.It is clear from the current literature that parents do not accurately perceive their child's weight.Nurses, pediatricians, and family physicians must bear the burden of informing and influencing parents of the importance of maintaining a healthy weight, not just in childhood, but throughout adulthood.As researchers, we must develop effective and culturally sensitive scripts for educating parents about their child's weight and the potential health outcomes. 8,16he development of effective interventions also is imperative to decrease the growing problem of overweight in our children, including strategies for parents, schools, and public health workers.The results of the current study suggested that parents of children receiving Medicaid may be an ideal population to pilot interventions due to the higher levels of overweight, potential lower levels of literacy, and inaccurate parental perception of weight.
Limitations.There are several limitations to the current study.The crosssectional design does not lend itself to identification of a cause and effect relationship, therefore, we cannot be sure that having AROW or overweight children causes participants to be less accurate in classifying their child's weight.Other factors, such as number of generations in the United States or denial may play a role in decreasing parents' ability to perceive accurately.
While the survey instrument was based on questions from previous studies and pretested with a small group of participants, it was not examined for test-retest reliability.Therefore, we cannot assume that parents would answer the same if the survey was given again.The survey also was missing key demographic questions regarding respondents (i.g., age, relationship to child) due to a copying error.Further, the survey was administered at a single facility, which does not allow for generalization to other practices, settings, or populations.
The survey looked at general perceptions of weight and did not attempt to identify cognitive beliefs that might contribute to misperception of weight.Utilizing BMI as a classification tool may have affected results as it does not take into account muscle mass or bone density.Further, the use of arbitrary cut-offs for BMI ranges also may have affected results.
Conclusions.The identification of low rates of accurate parental perception of overweight in Medicaid insured children could help in the development of appropriate prevention strategies as well as treatment strategies for decreasing childhood overweight in this underserved population.When informed, many parents appear surprised to learn how overweight their child is compared to normal and do not identify tactics easily to improve their child's health.Pediatric healthcare providers need to develop successful communication strategies to inform families when their child is overweight and identify appropriate recourse to combat overweight.Future research should work to identify such strategies as well as determine parental characteristics related to inaccurate perception of child's weight.

Table 2 .
Percent (and number) of children in each BMI group compared to the parents' perceived BMI group.