Currently Viewing:
The American Journal of Accountable Care September 2016
Early Provider Perspectives Within an Accountable Care Organization
Jill A. Marsteller, PhD, MPP; J. Hunter Young, MD, MHS; Oludolapo A. Fakeye, MA; Yea-Jen Hsu, PhD, MHA; Maura McGuire, MD; Matthew G. Poffenroth, MD, MBA; and Scott A. Berkowitz, MD, MBA
Cost Savings From Avoidance of Early Elective Deliveries
Charles Carlini, MD, JD, and Teresa Forth, RN, BSN
Currently Reading
Is Occupational Type Related to Obesity Risk?
Jodi H. Leslie, DrPH, RDN, LDN; Eric L. Hurwitz, PhD; Rachel Novotny, PhD; and Deborah A. Taira, ScD
In Louisiana, Necessity Breeds Innovation to Bring Medicaid Expansion
Mary K. Caffrey
Quality Improvement and Leadership Capacity Development Through Lean Methodology
Lisa M. Nicolaou, RN, MSNI
Using a Modified Next Generation ACO Benchmark Can Improve the MSSP
Lulu Liu, MA; Richard Svoboda, MA; and Yuting Zhang, PhD

Is Occupational Type Related to Obesity Risk?

Jodi H. Leslie, DrPH, RDN, LDN; Eric L. Hurwitz, PhD; Rachel Novotny, PhD; and Deborah A. Taira, ScD
Sociodemographic characteristics of blue-collar workers may be attributed to the higher rates of obesity and chronic disease seen among them compared with white-collar workers.

Objectives: To examine the association between occupational type and obesity among adult workers in Hawai`i, using both international and suggested ethnic-specific body mass index (BMI) cut-points. 

Study Design: Cross-sectional study. 

Methods: The study population included 22,340 adult subscribers of a large health plan who completed the Succeed Health Risk Assessment (Succeed) questionnaire from July 1, 2008, to June 31, 2009. Logistic regression analysis was performed to examine the association between occupational type (blue-collar workers [BCWs] vs white-collar workers [WCWs]) and obesity, controlling for age, gender, and other demographic, health, and behavioral variables known to be related to obesity. 

Results: Being a BCW and specific occupational categories were associated with obesity—an effect that remained, although attenuated, after controlling for ethnicity and using ethnic-specific BMI cut-points. After controlling for demographic, health, and behavioral variables, an inverse association was seen between obesity and BCWs and most occupational categories relative to WCWs. Ethnicity, education, smoking status, stress, and depression were found to be associated with obesity. Specific food groups were positively associated with obesity to different degrees, with greater odds ratios seen for the protein and saturated fat groups. A series of regression models, examining the effects of selected variables (ie, age, gender, ethnicity, and education) on the relationship between BCWs and obesity, demonstrated an inverse association only after education was added to the model. 

Conclusions: Researchers seeking to reduce obesity among employees should target ethnic groups at highest risk for obesity (eg, Native Hawaiians) and those with lower education levels, rather than occupational type. 
Obesity is increasing at an alarming rate and has more than doubled since 1980.1 Worldwide, an estimated 600 million individuals are obese. In the United States alone, more than 2 in 3 adults are considered overweight (body mass index [BMI] = 25-29.9) or obese (BMI ≥30).2 In 2010, overweight and obesity combined contributed to an estimated 3.4 million deaths globally. In Hawai`i, more than 2 in 4 adults are overweight or obese, with differing rates seen between ethnic groups3; for example, among Native Hawaiians, almost 3 in 4 are considered overweight or obese.

Poor diet and physical inactivity have been considered the main causes of obesity, which is also a major risk factor for leading chronic diseases and conditions.4 Considering the relationship of these diseases with obesity, it is not surprising that, compared with the state of Hawai`i, Native Hawaiians specifically experience a high prevalence of these chronic diseases, including coronary heart disease (4.2% vs 2.8%), stroke (4.1% vs 2.9%), diabetes (9.5% vs 7.8%), and hypertension (31.3% vs 28.7%).3

BMI is calculated from one’s weight and height.4 According to the World Health Organization (WHO), the BMI weight categorical cut-points demonstrate a “continuum of increasing risk with increasing BMI,” based on “statistical data from reference populations” examining “excess morbidity and mortality associated with increasing body fat content.”5 Considering that these international BMI cut-points are based on research among individuals who are white and that a variety of research has demonstrated differences in body composition, fat distribution, skeletal dimension, and genetic susceptibility for disease between whites and different Asian and Pacific Island population groups, ethnic-specific BMI cut-points for these groups have been proposed.6-8 However, due to observed variation in health risks reported at different BMI levels, the WHO recommended that the current BMI cut-points should be retained as the international classification pending further review and assessment. 

Given that working adults spend approximately 30% of their waking hours at their jobs, worksites are a sensible place to promote healthy lifestyle behaviors aimed at preventing and reducing obesity.9 In planning worksite health programs, consideration must be taken of the higher rates of obesity and chronic diseases in blue-collar workers (BCWs), who make up approximately 61% and 65% of the working population aged under 65 years in the United States and Hawai`i, respectively, compared with white-collar workers (WCWs).10-13

A clear understanding of the factors related to the high prevalence of obesity among BCWs is needed in order to effectively target them in worksite health programs. This is particularly necessary when examining those relationships among Hawai`i’s BCWs, as most obesity-related worksite studies comprised those ethnic groups typified of the United States (ie, whites, blacks, and Hispanics), as opposed to being inclusive of groups largely representative of Hawai`i’s population (ie, Native Hawaiians and Asians).



This study is a quantitative, exploratory study examining secondary data of 22,340 adult subscribers of one of Hawai`i’s largest medical insurers, who also completed the Succeed Health Risk Assessment (Succeed) questionnaire, from July 1, 2008, to June 31, 2009. The Succeed questionnaire was developed by HealthMedia, Inc, and is used by health plans, employers, and pharmaceutical firms nationally to evaluate health behavior and history of participants.14 About 85% of HealthMedia Succeed participants rated the questionnaire as “excellent,” “very good,” or “good.” 

The Succeed questionnaire was available at no cost to statewide subscribers of eligible plans, who were 18 years or older.15 Subscribers were encouraged to participate in this self-administered survey through the insurer’s website, their employer groups, and via mail-outs. A limited dataset was provided by the health insurer, with all direct identifiers removed for use in this study. This study was reviewed and approved by the University of Hawai`i (UH) Committee on Human Studies, the Institutional Review Board for the UH system.

Variables Examined

Ethnicity. Survey respondents were asked to select 1 ethnic group with whom they identified, from choices pre-determined by HealthMedia. (see eAppendix A [eAppendices available at] for Succeed questions used).

Obesity status. For this study, only the validated BMIs (ie, those obtained through office visits vs self-reports) were used. Obesity associations were estimated using both the international BMI definition for obesity (BMI ≥30), as well as ethnic-specific BMI cut-points for obesity suggested for Asians (BMI ≥25) and Pacific Islanders (BMI ≥32).5 Analysis examining sociodemographic and behavioral factors of obesity used international BMI cut-points only.

Occupational types and categories. Occupational categories were classified as BCW or WCW, using definitions by the US Department of Labor for each. BCWs are those individuals working in production, maintenance, cleaning, and other manual labor occupations who are paid by the hour or on an incentive basis.16 WCWs are those who work in office, clerical, administrative, sales, professional, and technical occupations. Those participants “not currently working outside the home” were omitted from the study population.

Physical activity level. The activity levels of each participant were categorized into 1 of 3 levels: sedentary (<150 minutes of physical activity per week), moderately active (150-300 minutes per week), and active (>300 minutes per week).17

Nutritional habits. Each participant’s intake was compared with nutrition recommendations based on a 2000-calorie reference diet (adopted as such by the FDA to provide a general assessment on whether recommended food intake patterns were being met or not).18  In this paper, intake of different foods (total grains, protein, oils) is reported as units of measurement versus number of servings. An individual was determined to meet the sodium intake recommendation if they “rarely” consumed salt/sodium foods, the whole-grain intake recommendation if at least half or more of their total grain servings were whole grains, and the saturated/trans fats intake recommendation if they ate 0 servings.

Smoking status. Consistent with Prochaska’s Stages of Change model, responses were placed into 1 of 3 categories for this study: never smoked, current smoker (pre-contemplation, contemplation, and preparation), and past smoker (action and maintenance).19

Other health behaviors.   Responses from questions regarding strength exercises, stress, depression, and sleeping hours were also examined in this study. 

Statistical Analysis

Descriptive statistics were used to characterize the study population. Logistic regression modeling was employed to estimate age- and gender-adjusted associations (odds ratios [ORs] and 95% CIs) between occupational type and categories with obesity, using international BMI and ethnic-specific BMI cut-points. Models with (n = 12,101) and without (n = 9503) nutritional habits were specified, due to possible inaccuracies in estimating portion sizes.20 A series of 4 regression models were specified, in which, selected variables (age, gender, ethnicity, and education) were chosen—based on their positive associations seen with obesity—and individually added into a model to determine which may be potential confounders affecting the relationship between being a BCW and obesity. Sets of 1 or more binary (dummy) variables were used to model the effects of all predictors, with the exception of age, which was modeled as a continuous predictor. Observations with missing variables were omitted from analysis. Those who were pregnant were excluded from the analysis. Data were analyzed using SAS version 9.2 (SAS Institute, Cary, North Carolina).


The descriptive characteristics of participants are listed in eAppendices B and C. Approximately 18.5% of respondents were BCWs, the majority being male, Asian, and with less than a college degree. Most WCWs were female, Asian, and college graduates. Twice as many Native Hawaiians (10% vs 5.9%) and Pacific Islanders (5.8% vs 3.2%) were BCWs than WCWs.

More BCWs (41.9% vs 23.3%) and WCWs (32.9% vs 16.2%) were classified as obese when the ethnic-specific cut-points were applied versus international ones. BCWs were more likely to be classified as physically active; however, they were also more likely to smoke and be depressed at least some of the time compared with WCWs, although WCWs were more likely to be stressed at least some of the time. Compared with WCWs, more BCWs exceeded the calcium/dairy, protein, oils, and saturated/trans fat recommendations,

After adjusting for age and gender only and using the international BMI cut-points, obesity was found to be associated with being a BCW (OR, 1.4; 95% CI, 1.3-1.5); this positive association remained, although attenuated, when ethnicity was controlled for (Table 1). When ethnic-specific BMI cut-points were used, obesity continued to be associated with being a BCW (OR, 1.2; 95% CI, 1.1-1.3), with odds unchanged before and after adjustment for ethnicity, and similar to the ethnicity-adjusted model using international BMI cut-points.

After controlling for all sociodemographic and behavioral characteristics, obesity was found to be inversely associated with being a BCW (OR, 0.8; 95% CI, 0.7-1.0) and either inversely or not associated with most occupational categories (Table 2). Native Hawaiians (OR, 3.3; 95% CI, 2.6-4.1) and those with less than a college education were most likely to be obese compared with their referents. Being a current or former smoker and/or having stress or depression often or most of the time was associated with obesity; higher levels of physical activity, strength exercises, and sleep were associated with decreased odds of obesity. 

Positive associations between obesity and whole grain intake were seen in those who had less than half of their total grain intake as whole grains, and considerably greater for those consuming at least 6.5 ounces of protein (compared with the referent of 5.5 ounces) (Table 3). A dose-response relationship was demonstrated between obesity and calcium/dairy foods and saturated/trans fats, with increasing amounts of each associated with increasing odds of obesity.    

Table 4 lists odds of obesity among BCWs using 4 different models. Obesity was positively associated with being a BCW (OR, 1.6; 95% CI, 1.4-1.7) in the unadjusted model; however, the odds of obesity steadily decreased as additional variables were included (ie, age/gender, ethnicity). An inverse relationship was seen only after education was included.


Copyright AJMC 2006-2019 Clinical Care Targeted Communications Group, LLC. All Rights Reserved.
Welcome the the new and improved, the premier managed market network. Tell us about yourself so that we can serve you better.
Sign Up