Abstract
Few studies have used nationally representative data to focus specifically on gender differences in weight-related outcomes. This article examines gender differences in weight-related outcomes across the body mass index (BMI) spectrum in overweight and obese adults. Data from the National Health and Nutrition Examination Survey 2009–2010 was analyzed. Weight-related outcomes were accurate weight perception, weight dissatisfaction, attempted weight loss, successful weight loss, and weight loss strategies. Compared with women, overweight and obese men were less likely to have accurate weight perception (odds ratio [OR] = 0.36; 95% confidence interval [CI] = 0.30–0.44), weight dissatisfaction (OR = 0.39; 95% CI = 0.32–0.47), and attempted weight loss (OR = 0.55; 95% CI = 0.48–0.63). The modifying effect of gender on these associations decreased as BMI increased. By BMI 35, the mean probability of women and men to have accurate weight perception and weight dissatisfaction was 90%; attempted weight loss was 60% (women) and 50% (men). At lower BMIs, men had up to 40% less probability than women for these weight loss outcomes. Men who attempted weight loss were more likely than women to lose and maintain ≥10 lb over 1 year (OR = 1.41; 95% CI = 1.20–1.65) and increase exercise and eat less fat as weight loss strategies; women were more likely to join weight loss programs, take prescription diet pills, and follow special diets. A need exists for male-specific interventions to improve overweight and obese men’s likelihood for accurate weight perception, attempted weight loss, and ultimately, successful weight loss.
Keywords
Introduction
The overweight and obesity epidemic in the United States continues to be a significant public health issue with a multitude of health consequences (U.S. Department of Health and Human Services, 2010). Nearly 70% of U.S. adults aged 20 years and older are either overweight or obese (National Center for Health Statistics [NCHS], 2014). According to the National Health and Nutrition Examination Survey (NHANES) 2009–2010, the prevalence of overweight and obesity among men is higher than among women at 74% versus 64%, respectively (Flegal, Carroll, Kit, & Ogden, 2012). Specifically for obesity, between 1999–2000 and 2009–2010, rates increased from 27.5% to 35.5% in men but remained largely steady from 33.4% to 35.8% in women (Flegal et al.). The risk of type 2 diabetes increases with increasing prevalence of overweight and obesity; consequently, there are 2.1 million more men than women with diabetes (Centers for Disease Control and Prevention [CDC], 2014).
An important need exists for gender-specific weight loss interventions to more effectively engage men, thereby reducing gender disparities in overweight and obesity and its health consequences. Tailoring weight loss interventions to be gender specific requires continued understanding of gender differences in weight-related attitudes, such as accurate weight perception and weight dissatisfaction, and weight-related behaviors, such as attempting weight loss and weight loss strategies. Existing weight loss intervention studies have consistently underrepresented men or excluded them entirely (Barte, Veldwijk, Teixeira, Sacks, & Bemelmans, 2014; Tang, Abraham, Greaves, & Yates, 2014; Young, Morgan, Plotnikoff, Callister, & Collins, 2012). When men are included, the enrollment level is lower than for women (Pagoto et al., 2012). As a result, knowledge gained from women-dominated interventions may not be generalizable to men.
Existing evidence demonstrates the relationship between weight-related attitudes and behaviors, but less research is focused specifically on gender differences in weight-related outcomes on a population level. Many of the studies that did report gender differences have used select populations, self-reported weight data, or a predominately female cohort (Lemon, Rosal, Zapka, Borg, & Andersen, 2009; Millstein et al., 2008). Using NHANES data from 2003-2006 to examine associations between weight misperception and several weight-related attitudes and behaviors, Duncan et al. (2011) reported that weight misperception was associated with fewer weight loss attempts in overweight and obese adults. The current study extends the work of Duncan et al. by using the most recent NHANES data (2009–2010) to specifically examine gender differences in multiple weight-related outcomes, which include attitudes (accurate weight perception and weight dissatisfaction) and behaviors (attempted weight loss and successful weight loss), across the entire BMI spectrum of overweight and obese adults. The hypothesis of the current study was as follows: regardless of BMI, overweight and obese men are less likely to have accurate weight perception and, therefore, less likely to have weight dissatisfaction, attempted weight loss, and successful weight loss compared with women. Gender differences in weight loss strategies were also explored.
Methods
Data Source
The NHANES is a national survey involving household interviews and medical examinations conducted by the NCHS. Since 1999, the continuous NHANES survey data have been released in 2-year increments for public use. The NHANES uses a stratified, multistage probability cluster sampling design and weighting methodology that allows unbiased national estimates to be produced for the civilian, noninstitutionalized U.S. population. NHANES sample weights adjust for unequal probabilities of selection, nonresponse, and planned oversampling (of low-income persons, persons ≥60 years of age, African Americans, and Hispanic Americans). The NCHS Ethics Review Board approved all NHANES survey protocols, consent forms, and information materials. Detailed documentation of the NHANES survey and public use data files can be found at http://www.cdc.gov/nchs/nhanes.htm.
Study Population
A total of 5,225 males and 5,312 females of all ages were included in NHANES 2009–2010 (the latest release at the time of study). From this population, survey participants who were younger than 20 years (n = 4,319) or pregnant (n = 68) or whose BMI was missing (n = 224) or less than 25 kg/m2 (n = 1,668) based on measured height and weight at the time of the NHANES medical examination were excluded. As a result, the analysis sample included 2,133 men and 2,125 nonpregnant women aged 20 years or older who were overweight or obese.
Measurements
Height and weight were measured by trained technicians using standardized protocols and calibrated equipment. BMI was calculated as weight (kg) divided by height (m2) and rounded to the nearest 0.1 kg/m2. BMI categories were defined using widely accepted cut points, that is, 25.0–29.9 kg/m2 for overweight and ≥30.0 kg/m2 for obesity.(27) The obese category was further subdivided into obese class I (BMI 30.0–34.9 kg/m2), obese class II (35.0–39.9 kg/m2), and obese class III (≥40.0 kg/m2).
Weight-related outcomes
Accurate weight perception
Participants were asked whether they perceived their weight to be “overweight,” “underweight,” or “about the right weight.” Only individuals with a measured BMI ≥25 were included; thus, an accurate weight perception was definehd as a participant self-identifying as “overweight,” whereas an inaccurate weight perception was when the participant answered “about the right weight” or ”underweight.”
Weight dissatisfaction
Weight dissatisfaction was defined by a desire to lose weight. Participants were asked if they would like to “weigh more,” “weigh less,” or “stay about the same.” If the participant’s answer was “weigh less,” they were regarded as expressing dissatisfaction with their current body weight. Otherwise, “weigh more” or “stay about the same” were regarded as participants expressing satisfaction with their current body weight.
Attempted weight loss
Attempted weight loss was defined by two separate questions that participants answered: “During the past 12 months, have you tried to lose weight?” and “Was the change between current weight and your weight a year ago because you tried to lose weight?” The latter question was asked if participants had lost 10 lb or more within the past year. If participant’s answer was “Yes” to either question, they were regarded as having attempted weight loss. Those reporting attempted weight loss were asked to select all that applied from a list of weight loss strategies. Individual strategies were coded as a “yes/no” variable.
Successful weight loss
For those who reported attempted weight loss, a participant was classified as achieving successful weight loss if the measured weight at the time of the 2009–2010 NHANES medical examination was at least 10 lb less than the person’s self-reported weight 12 months earlier.
Covariates
Country of birth
This was categorized as either U.S. or foreign born. The latter encompassed all participants who answered as born in Mexico or elsewhere (unspecified) outside the United States.
Sociodemographic characteristics
Characteristics included age (20–24, 25–34, 35–44, 45–64, and ≥65 years) and ethnicity/race (Hispanic, non-Hispanic White, non-Hispanic Black, and other race including multirace). Participants were categorized by annual family income (<$20,000, $20,000–<$45,000, $45,000–<$75,000, and ≥$75,000), highest achieved education level (less than high school, high school or general equivalency diploma, or college or higher), and marital status (never married, married or live with partner, and separated/divorced/widowed).
Statistical Analysis
Chi-square tests (PROC SURVEYFREQ) examined unadjusted associations of gender with the four weight-related outcomes and covariates. Adjusted odds ratios (ORs) with 95% confidence intervals (CIs) from the logistic regression procedure (PROC SURVEYLOGISTIC) were used to determine the significance of the independent association of an outcome with gender controlling for all the covariates. Additionally, the gender by BMI interaction was investigated and the estimated probability of each weight-related outcome versus BMI was plotted. A nonparametric method of locally weighted scatterplot smoothing (PROC LOESS) was overlaid in each plot to demonstrate the trend. Stepwise survey logistic regression was used to examine gender differences in the use of weight loss strategies. All the weight loss strategies were included in the initial model. The stepwise procedure selected the combination of weight loss strategies that significantly differentiated by gender.
All analyses were conducted in SAS, version 9.3 (SAS Institute Inc., Cary, North Carolina) and took account of the complex sampling design and sample weights of the NHANES. Statistical significance was set at p <.05 (two-sided).
Results
Study Population Characteristics
This study sample included 2,133 males and 2,125 females, representative of 77.3 million men and 70.8 million women in the general U.S. population (Table 1). Most participants regardless of gender were born in the United States and slightly more than half were at least college educated. But compared with women, men were younger (mean, 47.3 vs. 49.9 years) and less obese (mean BMI, 30.9 vs. 32.5 kg/m2) and had a higher percentage of non-Hispanic whites (69.1% vs. 64.5%), persons married or living with a partner (70.6% vs. 58.5%), and those with annual family income ≥$75,000 (37.7% vs. 26.6%).
Sample Characteristics by Gender.
Weight-Related Outcomes by Gender
Compared with women, men were less likely to have accurate weight perception (OR = 0.36, 95% CI = 0.30–0.44), weight dissatisfaction (OR = 0.39, 95% CI = 0.32–0.47), and attempted weight loss (OR = 0.55, 95% CI = 0.48–0.63). However, men reporting attempted weight loss were more likely than women to have successful weight loss (OR = 1.41, 95% CI = 1.20–1.65). These significant associations persisted after controlling for all the covariates (Table 2).
Unadjusted and Adjusted Associations of Weight-Related Outcomes With Gender.
Note: OR = odds ratio; CI = confidence interval.
Adjusted for the following variables: birth of origin, age category, race/ethnicity, education, family income category, marital status, and obesity category.
An accurate weight perception was defined as participants self-identifying as “overweight” when asked whether they perceived their weight to be “overweight,” “underweight,” or “about the right weight.”
Weight dissatisfaction was defined by a desire to “weigh less” when participants were asked whether they would like to “weigh more,” “weight less,” or “stay about the same.”
Attempted weight loss was defined by two separate questions to which participants answered that they had attempted weight loss within the past year whether or not they actually lost any weight or that they had lost weight ≥10 lb within the past year and that the loss was intentional.
Successful weight loss among those who attempted weight loss. Male: n = 883, N = 33.1 million; female: n = 1121, N = 40.5 million. Successful weight loss was defined as participants reporting weight loss attempt and achieving a weight loss of at least 10 lb within the past year.
Modification Effect of Gender on the Associations Between BMI and Weight-Related Outcomes
Gender modified the associations between BMI with accurate weight perception, weight dissatisfaction, and attempted weight loss (p < .0001 for Gender × BMI interaction for all three outcomes; Figure 1). At a BMI of 25, the mean probability of women to have accurate weight perception, weight dissatisfaction, and attempted weight loss was about 60%, 70%, and 50%, respectively, whereas the mean probability of men was only about 20%, 40%, and 30%, respectively. At a BMI of 30, the mean probability of women to have accurate weight perception, weight dissatisfaction, and attempted weight loss was about 80%, 80%, and 50%, respectively, and the mean probability of men had increased to about 60%, 70%, and 40%, respectively. The modifying effect of gender on these associations decreased when BMI approached class II obesity. By BMI 35, the mean probability was 90% for accurate weight perception and for weight dissatisfaction in both genders; for attempted weight loss, it was 60% in women and 50% in men. The adjusted ORs of all these three weight-related outcomes were significantly greater than 1 per unit increase in BMI. The associations between BMI with successful weight loss were not different between men and women, with mean probability ranging from 15% to 35% at a BMI of 25–75 (p = .17 for gender × BMI interaction). Men had modestly reduced odds of achieving significant weight loss (adjusted OR = 0.97, 95% CI = 0.93–1.00) per unit increase in BMI, whereas the odds remained constant in women.
Scatter plots of probabilities of accurate weight perception, weight dissatisfaction, attempted weight loss, and successful weight loss with body mass index (BMI) separated by gender.
Weight Loss Strategies by Gender
Weight loss strategies chosen by men and women who attempted to lose weight are shown in Table 3. About two-thirds of both men and women chose healthy options: to eat less or exercise. Few of them chose to use a liquid diet formula, take prescription diet pills, take nonprescription supplements, take laxatives or vomit, follow a special diet, or smoke. However, significant gender differences still emerged showing that a combination of exercising, eating less fat, joining weight loss program, taking prescription diet pills, following a special diet, and eating more fruits, vegetables, and salads could differentiate men versus women who attempted weight loss. Among these six strategies, men were more likely to eat less fat (OR = 1.23, 95% CI = 1.01–1.49) and exercise (OR = 1.45, 95% CI = 1.18–1.77), but less likely to join a weight loss program (OR = 0.16, 95% CI = 0.10–0.26), take prescription diet pills (OR = 0.17, 95% CI = 0.08–0.35), follow a special diet (OR = 0.58, 95% CI = 0.35–0.95), and eat more fruits, vegetables, and salads (OR = 0.73, 95% CI = 0.57–0.92) than women.
Weight Loss Strategies by Gender Among Those Who Have Intention to Lose Weight.
Odds ratios (ORs) with 95% confidence intervals (CIs) were derived from the final model determined by stepwise survey logistic regression analysis. Percentages and p value for each weight loss strategy were generated using bivariable analysis (PROC SURVEYFREQ).
Discussion
This study identified several important gender differences in weight-related attitudes and behaviors among overweight and obese U.S. adults, and the modifying effect of gender on the association between BMI and these weight-related attitudes and behaviors. Overweight men (BMI 25–29.9) were less likely than overweight women to have accurate weight perception, weight dissatisfaction, and attempted weight loss. However, this gender difference decreased as men became obese. Consequently, men and women with class II or III obesity (BMI ≥35) shared a similar likelihood to have accurate weight perception, weight dissatisfaction, and attempted weight loss. Men who attempted weight loss were 41% more likely than women who attempted weight loss to have successful weight loss. Men who attempted weight loss were more likely to exercise and eat less fat as weight loss strategies whereas women were more likely to join a weight loss program, take prescription diet pills, follow a special diet, and eat more vegetables and fruits.
The finding that men and women have similar likelihood for accurate weight perception, weight dissatisfaction, and attempted weight loss at class II and class III obesity adds to previous evidence on gender differences in these weight-related outcomes. Previous studies reported that men were less likely than women to have accurate weight perception and weight dissatis-faction in all BMI categories (Lemon et al., 2009; Millstein et al., 2008). In their study of hospital employees from a health care system in Massachusetts, Lemon et al. (2009) reported men were less likely than women to perceive they were overweight at each BMI category. Even men with class III obesity had less accurate weight perception than women, with 25.0% of men versus 5.5% of women reporting that their weight was just right or underweight. In their study using a population-based telephone survey of U.S. adults, Millstein et al. (2008) concluded that men were less likely than women to have weight dissatisfaction at each BMI category. Most studies combine obesity classes into a BMI ≥30 category and subsequently do not report differences between the obesity classes. Although Lemon et al. (2009) examined weight perception by obesity class, the sample was small and may not have been representative of the U.S. general adult population. In contrast, the current study specifically investigated gender differences in weight-related outcomes across the entire overweight and obese BMI spectrum in a large national sample and showed less modifying effect of gender at higher BMIs for weight perception, weight dissatisfaction, and attempted weight loss. Moreover, the probability of successfully sustaining 10 lb or more weight loss over 1 year was similarly low, <40%, at all BMI levels among both women and men who were overweight or obese and attempted weight loss.
Although not entirely clear why overweight and moderately obese men are less likely than women to have accurate weight perception, the reason may include men internalizing an ideal male body type that is more muscular and athletic (Christensen, 2011), and societal norms among men to prefer heavier body weight (Duncan et al., 2011). Thus, men may not perceive a problem until they are severely obese, reaching a BMI of 35 or higher. Evidence suggests accurate weight perception is a necessary prerequisite to weight dissatisfaction and attempted weight loss. Lemon et al. (2009) reported that overweight and obese men who perceived themselves to be mildly or moderately overweight, compared with men who perceived their weights to be just right, were 14.4 and 13.8 times more likely to attempt weight loss. Duncan et al. (2011), using NHANES 2003-2006, reported that overweight and obese men who inaccurately perceived their weight, compared with men who accurately perceived their weight, were 71% less likely to be dissatisfied with their weight, and 60% less likely to attempt weight loss. Our findings that men’s likelihood to have accurate weight perception only reaches 90% when their BMI is categorized as class II obesity underscores the critical need to improve overweight and moderately obese men’s accurate weight perception. Improving men’s awareness of the increased risk of chronic health problems such as diabetes, cardiovascular disease, and sleep apnea, as well as the associated decrements in quality of life and longevity, even with overweight or moderate obesity, may be an important strategy. At the same time, person-centered research is needed to develop, evaluate, and disseminate innovative lifestyle interventions that are specifically targeted at and tailored to overweight and obese men.
This study showed that when adults attempted weight loss, gender differences in preference of weight loss strategies existed. Men were 84% less likely than women to join a weight loss program. This gender disparity deserves attention because men appear to benefit from weight loss programs especially if tailored to them. Bye, Avery, and Lavin (2005) studied men-only groups in a commercial weight loss program and showed that by 8 weeks, 90% of men had achieved at least 5% weight loss. Similar to findings from Kruger, Galuska, Serdula, and Jones (2004), who used 1998 National Health Interview Survey data, the current study identified that most men who attempted weight loss preferred healthy options for weight loss including eating fewer calories and exercise, with a smaller portion of men eating less fat. In addition, this study identified that a small portion (about one third) of men who attempted weight loss chose to eat more fruits, vegetables, and salads; however, men were less likely than women to do this. Other studies have reported that men eat less healthily than women, and by childhood and adolescence, girls already consume more fruits and vegetables than boys (Kiefer, Rathmanner, & Kunze, 2005). Thus, addressing a decreased preference for fruits, vegetables, and salads should begin in childhood.
Men who attempted weight loss in the current study were more likely than women to exercise. Kruger, Blanck, and Gillespie (2006) reported that successful weight losers were more likely than unsuccessful weight losers to exercise ≥30 minutes/day and add physical activity to daily life. Men’s increased likelihood to exercise as a weight loss strategy may partly explain why men who attempted weight loss in the current study were more likely than their women counterparts to sustain 10 lb or more weight loss in 1 year. Within men, however, the likelihood of successfully sustained weight loss decreased modestly with increasing BMI. More importantly, the probability of success was low at all overweight and obese BMI levels in both genders. A gender-specific, comprehensive intervention approach to weight loss should take into account the weight loss strategies men prefer, yet still encourage men to increase their fruit and vegetable intake and decrease their intake of saturated and trans fats as well as added sugars consistent with recommended guidelines for a healthy diet (Jensen et al., 2014). For both genders, innovative research is imperatively needed to promote long-term maintenance of clinically significant weight loss.
Improving overweight and obesity prevalence in men requires increased male participation in weight loss research studies. Unfortunately, men have been consistently underrepresented in weight loss research (Pagoto et al., 2012); consequently, most weight loss interventions are based on evidence gained from women-dominant weight loss trials. In a systematic review of 80 weight loss trials with a minimum 1-year follow-up, the average proportion of male participants was only 27% (Franz et al., 2007). However, men appear to benefit from weight loss interventions. Of the few male-only weight loss and weight maintenance studies with clear intent to change behavior or lifestyle, a recent systematic review reported that 74% of 23 interventions were effective, defined as a mean weight loss change of ≥5% at final assessment (Young et al., 2012). Most of the male-only studies implemented gender-neutral weight loss interventions; but men may prefer interventions that are tailored to men. Hunt, McCann, Gray, Mutrie, and Wyke (2013) reported positive evaluations from participants in a male-only walking program at a professional football club, which was tailored to men in context, content, and style of delivery. The persistent gender disparity of high male overweight and obesity prevalence requires a gender-specific approach, which may encourage more male participation in weight loss research studies and weight loss programs.
Limitations
This study has several limitations worth considering when interpreting the findings. First, NHANES is a cross-sectional survey, which prevents us from drawing causal conclusions from the data. However, the objective was to illustrate gender differences in weight-related outcomes using national level data for which this cross-sectional survey was ideal. Second, successful weight loss relied on a self-reported weight from the past, and was defined as the participant’s current measured weight being 10 lb or less than his/her self-reported weight from the prior 12 months. Self-reported weights may lead to misclassification because of reporting errors and biases, which may vary by age, gender, ethnicity, and overweight status (Ezzati, Martin, Skjold, Vander Hoorn, & Murray, 2006; Kuczmarski, Kuczmarski, & Majjar, 2001; Palta, Prineas, Berman, & Hannan, 1982). In this study, a comparison of measured weights with self-reported weights at the time of the interview (Lin’s [1989] concordance correlation coefficient was 0.96) and their corresponding BMI categories (weighted kappa was 0.804) indicated a substantial degree of concordance (Lin; Viera & Garrett, 2005). Third, all weight loss strategies were self-reported. There may be reliability and validity challenges with self-reported weight loss strategies if participants did not understand what the weight loss strategy suggested or if participants misreported the use of weight loss strategies owing to cultural, social, or psychological reasons.
Need for Gender-Specific Weight Loss Interventions
Addressing the gender disparity in overweight and obesity prevalence requires more gender-specific weight loss research and increased male participation in weight loss interventions. Weight loss interventions that are tailored to appeal to men will likely improve male participation. Gender differences in weight-related attitudes and behaviors should inform the development of weight loss interventions for men to increase the likelihood of effectiveness and sustainability. For example, male weight loss interventions need to address overweight and mildly obese men’s decreased likelihood to have accurate weight perception. Increasing their awareness of the connection between overweight and obesity with poor health outcomes and the benefits of healthy diet and exercise may increase their ability to have accurate weight perception at lower BMI categories. In addition, weight loss interventions that use new intervention delivery strategies may appeal to men. In their study using a Diabetes Prevention Program (DPP) lifestyle intervention in a primary care setting, Ma et al. (2013) integrated health information technology tools (the AHA free Heart360 web portal for weight and physical activity goal setting and self-monitoring, and electronic health record-embed secure messaging) into their intervention arms and reported that men (53.5% of the sample) had statistically significant weight loss compared with usual care, similar to the female study participants. In their study using web-based group video conferencing, Azar et al. (2015) created men-only virtual small groups to effectively deliver an evidence-based weight management program. As illustrated by the effectiveness and positive evaluation of male-only weight loss interventions, more male-only weight loss programs should be developed. Rates of successful weight loss remain low for men and women, which underscores the need to increase gender-specific research in weight loss interventions, which may then increase the likelihood of successful weight loss for men and women.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Internal funding from the Palo Alto Medical Foundation Research Institute supported this research.