The prevalence of obesity in the U.S.

Transcription

1 JOEM Volume 47, Number 12, December CME Available for this Article at ACOEM.org Lost Productive Time Associated With Excess Weight in the U.S. Workforce Judith A. Ricci, ScD, MS Elsbeth Chee, ScD Learning Objectives Recall the prevalence of obesity in this population of predominantly white-collar workers, and the ways in which obese individuals (body mass index [BMI] of 30 kg/m 2 or higher) differed from their overweight (BMI of kg/m 2 ) and normal-weight (BMI of kg/m 2 counterparts. Describe whether and to what degree obese workers differed from others in absenteeism and presenteeism as reflected by loss of productive time (LPT) in the past 2 weeks, and clarify relationships among obesity, LPT, and overall health status. Estimate the added dollar costs incurred by employers because of obese and overweight employees. Abstract Objective: The objective of this study was to examine health-related lost productive time (LPT) in overweight and obese workers. Methods: Cross-sectional study using data from a national telephone survey of the U.S. workforce. Body mass index defined normal-weight, overweight, and obese workers. LPT in hours and dollars was compared among the three groups. Results: Obese workers (42.3%) were significantly (P ) more likely to report LPT in the previous 2 weeks than normal-weight (36.4%) or overweight workers (34.7%). Health status mediated the relation between obesity and LPT. Obese workers cost an estimated $42.29 billion in LPT, an excess of $11.70 billion compared with normal-weight workers. Presenteeism accounted for 67.8% of the cost. Comparatively, overweight workers were not a significant source of excess LPT. Conclusions: Reducing excess weight in the workforce and improving the health of obese workers could positively impact U.S. workforce productivity. (J Occup Environ Med. 2005;47: ) From Caremark, Hunt Valley, Maryland. Address correspondence to: Judith A. Ricci, ScD, MS, Caremark, McCormick Road, Suite 230, Hunt Valley, MD 21031; judi.ricci@caremark.com. Copyright by American College of Occupational and Environmental Medicine DOI: /01.jom c3 The prevalence of obesity in the U.S. workforce has increased by an estimated 44% during the past decade. 1 Nearly 30% of all U.S. workers are obese as defined by a body mass index (BMI) of 30 kg/m 2 or greater. 1 Obesity, now considered an epidemic in the United States, 2 is associated with serious health and economic consequences. It is an established risk factor for numerous chronic conditions such as type 2 diabetes, hypertension, and coronary heart disease. 3,4 It is also linked to increased mortality 5,6 and decreased life expectancy. 7 Direct medical costs of disease attributable to obesity were estimated at $51.6 billion in 1995, which represents 5.7% of all U.S. health expenditures that year. 8 Obesity also impacts work. Obese workers are significantly more likely than their normal-weight counterparts to report poor work ability 9 or a limitation in the amount, type, or quality of work. 1,10 They also experience more disability and lost workdays than normal-weight adults. 8,11,12 In 1995, obese adults in the United States lost an estimated 39.2 million workdays, at a total annual cost of approximately $3.93 billion to employers. 8 The cost of work loss resulting from obesity in the U.S. workforce, including both presenteeism (ie, reduced performance while at work) and absenteeism (ie, time absent from work), has not been quantified previously. In addition, little information exists on the work impact of being overweight but not obese. The Caremark American Productivity Audit provides a unique opportunity to examine the cost of lost produc-

2 1228 LPT and Excess Weight in the U.S. Workforce Ricci and Chee tive time (LPT) attributed to excess weight in the U.S. workforce. We describe the results of a study to quantify health-related LPT, including both absenteeism and presenteeism in normal-weight, overweight but not obese, and obese U.S. workers, and to examine factors mediating the relationship between obesity and LPT. Materials and Methods The Caremark American Productivity Audit is a national populationbased, random-digit-dial telephone survey of the noninstitutionalized U.S. population that measures the relationship between health and work productivity. 14 The Caremark Work and Health Interview (WHI) is the validated data collection instrument administered in the audit survey. Both the audit 14 and the WHI 15,16 have been described in detail elsewhere and are summarized briefly here. Work and Health Interview The WHI is a computer-assisted telephone data collection instrument that measures LPT and its healthrelated causes in the 2 weeks before interview. 15,16 Health-related reasons for LPT are attributed directly by the respondent. The interview also captures information on selfreported employment status, occupational characteristics, health conditions and symptoms, lifestyle factors, and demographic characteristics, including annual salary. LPT is measured as the sum of selfreported hours per week absent from work for a health-related reason (ie, absenteeism) and the hourequivalent per week of self-reported health-related reduced performance while at work (ie, presenteeism). Presenteeism is quantified by measuring the average amount of time between arriving at work and starting work on days when a worker is not feeling well and the average frequency of engaging in five specific work behaviors (ie, losing concentration, repeating a job, working more slowly than usual, feeling fatigued at work, and doing nothing at work). Sample Selection and Data Collection A random sample of residences with telephones in the 48 contiguous states and the District of Columbia were selected as audit households. Residents were eligible to participate if they were 18 to 65 years of age, responded yes to the Current Population Survey (CPS) question on employment status (ie, Last week, did you do any work for either pay or profit? ), 17 and were a permanent member of the household contacted. Up to two eligible respondents were interviewed per household. The Essex Institutional Review Board (Lebanon, NJ) approved the research protocol and data collection instrument. Oral informed consent was obtained from each participant before initiating the interview. Audit data collection began on August 1, 2001, and continued through May 31, Beginning August 26, 2002, the number of interviews completed per month was reduced from 2500 to 1250 by randomly selecting 50% of eligible respondents to complete the interview. Anthropometric data collection (ie, self-reported height and weight) began September 4, 2002, and continued through May 31, Estimated audit participation during this period was 68%. The prevalence of a number of diseases and conditions was determined at the time of the interview. The total number of the following conditions was used as a proxy measure of overall health status: headache, musculoskeletal pain, digestive problems, feeling sad or blue, asthma or other chronic breathing problems, cancer, heart disease, hypertension, diabetes, fatigue, and urinary urgency or incontinence. Analysis Analyses were completed to estimate the cost of health-related LPT in normal-weight, overweight, and obese U.S. workers. The sampling frame for this study included the 7472 employed adults 18 to 65 years of age who participated in the audit survey between September 4, 2002, and May 31, BMI was calculated as weight in kilograms divided by height in meters squared for the 7092 respondents with valid self-reported height and weight information. Based on BMI, respondents were categorized as underweight ( 18.5 kg/m 2 ), normal weight ( kg/m 2 ), overweight ( kg/m 2 ), or obese ( 30.0 kg/m 2 ) according to federal guidelines. 18 Pregnant women (n 73) and underweight respondents (n 125) were excluded. The final analytic sample of 6894 respondents comprised 2868 normal-weight, 2490 overweight, and 1536 obese workers. Health-related LPT including absenteeism and presenteeism was derived from the WHI as described previously. 14 LPT was examined in two ways: as the percent of workers with 0 LPT hours in the previous 2 weeks and as the mean number of LPT hours per week in workers with 0 LPT hours in the previous 2 weeks. Logistic regression models were built in two steps to investigate factors mediating the relation between obesity and health-related LPT. In particular, the analysis focused on determining if obesity (present vs absent) operated through overall health status to affect LPT (0 LPT hours vs 0 LPT hours). The variable, number of cooccurring health conditions, was selected as a proxy for overall health status. In the first step, the relationship between obesity and the likelihood of losing productive work time for a healthrelated reason was examined adjusting for covariates that demonstrated a statistically significant (P 0.05) association with a report of 0 health-related LPT hours in bivariate comparisons. Covariates included gender, age, education, annual salary, geographic region, cigarette use, and alcohol consumption. In the sec-

3 JOEM Volume 47, Number 12, December ond step, number of cooccurring health conditions was added to the model to evaluate the inclusion of this variable on the obesity coefficient and measure the significance of overall health status in predicting 0 hours of health-related LPT. Lost labor costs, expressed in 2002 dollars, were estimated by converting hours of LPT into lost dollars using self-reported annual salary or wage. The costs of LPT resulting from obesity and overweight were defined as the excess health-related LPT costs in obese and overweight workers compared with normalweight workers. Excess LPT cost was calculated as the difference in the annual cost of LPT for any health-related reason between obese or overweight workers and normalweight workers. Sample estimates were projected to the U.S. workforce 18 to 65 years of age using a two-step weighting method to account for selective participation (ie, noncoverage and nonresponse). 14 In the first step, a weight was applied to individuals as the inverse of the number of phone numbers available for incoming calls to account for the unequal probability of selecting households. In the second step, a population-weighting (ie, benchmarking) adjustment accounted for selection bias as a result of incomplete coverage of the U.S. population and ensured that estimates of certain sample demographic subgroups totals conformed to the CPS, an external database providing high-quality data on a nationally representative sample of the U.S. workforce. The raking method used for the population weighting adjustment benchmarked to four variables common to both the productivity audit and the CPS. Benchmarking and weighting variables with missing data (ie, 21.2% with 93.5% of these attributed to missing number of phone lines) were imputed using the age- and gender-specific mode for categorical variables and the ageand gender-specific median for continuous variables. If one of the five variables used in the calculation of presenteeism was missing, the mean value of the remaining four variables was substituted, reducing the proportion of respondents with missing presenteeism estimates from 2.6% to 2.1%. Linear regression modeling, which included gender, age, race, education, region of residence, job code, and duration in job, was used to reduce missing salary information from 17.7% to 3.7%. SAS version 8.2 (SAS Institute Inc., Cary, NC) was used for data management and logistic regression modeling. Wesvar version 4 statistical software (Westat, Rockville, MD) was used to perform the raking adjustments and for all other analyses. P 0.05 was used to determine statistical significance. Results The majority of individuals in the audit sample were female (55.5%), between 40 and 65 years of age (57.3%), white (81.8%), and formally educated beyond high school (69.0%). They worked full-time (80.9%) in a white collar job (67.0%) in which they earned less than $50,000 annually (71.3%). Approximately half never smoked cigarettes (52.0%) and two thirds consumed less than one alcoholic drink per week (66.7%). Approximately 50% reported two or more health conditions or symptoms in the previous 2 weeks. Musculoskeletal pain (43.3%) and fatigue (36.1%) were reported most commonly. A total of 22.3% were defined as obese, and 36.1% were defined as overweight but not obese. Normal-weight, overweight, and obese U.S. workers differed significantly on almost all demographic, employment, and health characteristics examined (Table 1). Although the majority of normal-weight workers were female, the majority of overweight and obese workers were male. Males comprised 42.8%, 68.3%, and 58.3% of normal-weight, overweight, and obese workers, respectively (P ). Compared with their normal-weight and overweight counterparts, obese workers were more likely to be older, black, and working in a blue collar occupation. They were less likely to have earned a bachelor s degree. Of the three groups, obese workers were most likely to be former smokers, consume less than one alcoholic drink per week, and report four or more health conditions in the previous 2 weeks. They were also the most likely to report the presence of eight of 10 individual health conditions in the previous 2 weeks (ie, headache, musculoskeletal pain, digestive problems, asthma or chronic breathing problems, heart disease, hypertension, diabetes, and fatigue). Overweight workers were more similar to normal-weight workers than to obese workers in health characteristics. Overweight workers differed significantly from normalweight workers in only four of 13 health characteristics (ie, cigarette use and presence of self-reported digestive problems, hypertension, and diabetes) (Table 1). However, compared with obese workers, they differed significantly in 12 of 13 characteristics (ie, cigarette use, alcohol consumption, number of cooccurring health conditions, and presence of self-reported headache, musculoskeletal pain, digestive problems, feeling sad/blue, asthma/ chronic breathing problems, heart disease, hypertension, diabetes, and fatigue). Overweight workers differed significantly from normalweight workers and obese workers in six and seven, respectively, of the eight demographic and employment characteristics examined (Table 1). Lost Productive Time Obese workers were significantly (P ) more likely to report 0 hours of health-related LPT in the previous 2 weeks (42.3%) than either normal-weight (36.4%) or overweight (34.7%) workers (Table 2). Overweight and normal-weight workers did not differ significantly in the percent reporting 0 LPT hours in the previous 2 weeks. Although almost 10% (9.3%; 95%

4 1230 LPT and Excess Weight in the U.S. Workforce Ricci and Chee TABLE 1 Percent Distribution of Normal-Weight, Overweight, and Obese Workers by Selected Demographic, Employment, and Health Characteristics* Characteristic Categories Normal Weight Overweight Obese P Value Gender Male Female Age yr yr yr yr Race White Black Other Education High school diploma High school graduate or GED Some college or Associate s degree Bachelor s degree Graduate degree Annual salary $10, $10,000 19, $20,000 29, $30,000 39, $40,000 49, $50, Geographic region Northeast Midwest South West Cigarette use Never smoked Former smoker Smokes 1 pack/d Smokes 1 pack/d Alcohol consumption Does not drink drink/wk drinks/wk drinks/wk No. of cooccurring health conditions Presence of cooccurring conditions Headache Musculoskeletal pain Digestive problems Feeling sad/blue Asthma/chronic breathing problems Cancer Heart disease Hypertension Diabetes Fatigue Urinary urgency/incontinence Employment status Full-time Part-time Type of occupation White collar Blue collar *Estimates benchmarked to the U.S. workforce using an iterative proportional fitting procedure described in the Methods section. Chi-squared statistic measures differences among normal-weight, overweight, and obese workers. Percents will not total 100 because categories are not mutually exclusive. Full-time was defined by 35 hr per week; part-time was defined by 35 hr per week. White collar jobs included professional, administrative, or support-type occupations; blue-collar jobs included trade or labor occupations. 27

5 JOEM Volume 47, Number 12, December TABLE 2 Percentage of the U.S. Workforce with 0 Health-Related Lost Productive Time (LPT) and Mean LPT per Week in Workers with 0 LPT by Body Mass Index Category* BMI Category Percent with >0 LPT % (95% CI) confidence interval [CI] %) of obese workers reported a healthrelated work absence in the previous 2 weeks, over 40% (41.2%; 95% CI %) reported healthrelated reduced performance while at work during the same period. Workers with LPT in the 2 weeks before interview lost an average of 4.3 hours per week (95% CI hours). The mean amount of LPT did not differ significantly among normal-weight, overweight, and obese workers. Workers in the three weight classes also did not differ in the health conditions cited most often as the source of LPT in the previous 2 weeks. Of the top five reported reasons for LPT, the cold or flu was mentioned most often, followed by musculoskeletal pain, headache, fatigue, and stomach or digestive problems. Logistic regression analysis to examine factors mediating the relationship between obesity and LPT is summarized in the two models presented in Table 3. The first model demonstrates that obesity (odds ratio [OR], 1.4), female sex (OR, 1.8), and cigarette use (former smoker OR, 1.3; smokes 1 pack per day OR, 1.6; smokes 1 pack per day OR, 1.7) were associated with significantly elevated odds of losing productive time for a health-related reason. The second model shows that, with the addition of number of cooccurring health conditions to the model, the OR of any LPT associated with obesity is no longer statistically significant (OR, 1.1; 95% CI ) Mean LPT (95% CI) (hours/wk) Normal weight 36.4 ( ) 4.2 ( ) Overweight 34.7 ( ) 4.2 ( ) Obese 42.3 ( ) 4.8 ( ) *Estimates benchmarked to the U.S. workforce using an iterative proportional fitting procedure described in the Methods section. BMI indicates body mass index; CI, confidence interval. and the OR of any LPT in the previous 2 weeks increases as the number of cooccurring conditions increases (one condition OR, 4.7; two to three conditions OR, 10.2; four or more conditions OR, 21.3). These findings suggest that the relation between obesity and LPT is mediated by workers overall health status. National Lost Productive Time Cost Estimates The estimated total annual cost of health-related LPT in obese U.S. workers was $42.29 billion (Table 4), or $1627 (95% CI $ ), on average, per obese worker per year. Two thirds of the total cost (ie, $28.69 billion) was explained by reduced performance while at work, not work absence. Compared with normal-weight workers, obese workers cost U.S. employers an estimated $11.70 billion per year in excess health-related LPT. Overweight workers cost U.S. employers an estimated $55.36 billion per year in health-related LPT. The expected LPT cost falls within the 95% confidence interval of this estimate and indicates that overweight, unlike obesity, is not associated with excess health-related LPT compared with normal-weight workers (Table 4). It was also noted that the mean annual cost of LPT in overweight workers (ie, $1250; 95% CI $ ) was significantly (P ) less than that of obese workers, but did not differ significantly from that of normal-weight workers (ie, $1201; 95% CI $ ). These cost results are consistent with the findings from the LPT analysis. Discussion This study provides new information on the total cost of health-related work loss (ie, absenteeism and presenteeism) in obese U.S. workers. It also provides, to the best of our knowledge, the first U.S. national estimate of the annual cost of presenteeism attributable to obesity (ie, $7.84 billion). Previously, Wolf and Colditz estimated that obesity-related absenteeism cost employers $3.93 billion per year. 8 Our estimate of the annual cost of absenteeism attributable to obesity (ie, $3.86 billion) is consistent with this finding. Moreover, our estimate of the total annual cost of LPT attributable to obesity (ie, $11.70 billion per year) demonstrates that absenteeism comprises only 33% of the total cost of obesityrelated LPT and that measuring only missed workdays significantly underestimates the total cost of LPT attributed to obesity. Hertz and colleagues recently estimated that 29.4% of the U.S. workforce was obese. 1 In our sample, 22.3% of workers met criteria for obesity. Although the obesity case definitions used in their study and ours were identical, the lower prevalence of obesity in our study could be explained, in large part, by the difference in method of collecting height and weight data. Hertz used data from National Center for Health Statistics surveys where height and weight were measured. In our study, workers self-reported height and weight. Previous studies have examined the validity of self-reported height and weight. Palta and colleagues demonstrated that self-reported weight can be underreported, particularly in women, and that self-reported height can be overreported, particularly in men. 19 Both types of errors lead to the underestimation of BMI and, thus, the prevalence of obesity. Given our lower prevalence estimate, it is likely that the cost of LPT attributable to obesity in the U.S. workforce is conservative.

6 1232 LPT and Excess Weight in the U.S. Workforce Ricci and Chee TABLE 3 Adjusted Odds Ratios (ORs) for Any Health-Related Lost Productive Time in the Previous 2 Weeks by Demographic and Health Covariates in Logistic Regression Models Excluding and Including Number of Cooccurring Health Conditions Covariate Category Model 1 Model 2 Excludes Number of Cooccurring Conditions Adjusted OR (95% CI) Includes Number of Cooccurring Conditions Adjusted OR (95% CI) Obese Yes 1.4 ( ) 1.1 ( ) No* No. of cooccurring health conditions 0* Excluded from model ( ) ( ) ( ) Gender Female 1.8 ( ) 1.3 ( ) Male* Age Modeled as continuous variable 1.0 ( ) 1.0 ( ) Education High school diploma 0.9 ( ) 0.8 ( ) High school graduate or GED 0.9 ( ) 0.9 ( ) Some college or Associate s degree 1.0 ( ) 0.9 ( ) Bachelor s degree 0.9 ( ) 1.0 ( ) Graduate degree* Annual salary $10, ( ) 0.9 ( ) $10,000 19, ( ) 1.2 ( ) $20,000 29, ( ) 1.4 ( ) $30,000 39, ( ) 1.1 ( ) $40,000 49, ( ) 1.1 ( ) $50,000* Geographic region Northeast 1.1 ( ) 1.1 ( ) South 1.3 ( ) 1.3 ( ) West 1.2 ( ) 1.1 ( ) Midwest* Cigarette use Never smoked* Former smoker 1.3 ( ) 1.1 ( ) Smokes 1 pack/d 1.6 ( ) 1.3 ( ) Smokes 1 pack/d 1.7 ( ) 1.3 ( ) Alcohol consumption Does not drink* 1 drink/wk 1.2 ( ) 1.3 ( ) 1 6 drinks/wk 1.0 ( ) 1.1 ( ) 7 drinks/wk 0.9 ( ) 0.9 ( ) *Reference category. OR indicates odds ratio; CI, confidence interval. TABLE 4 Total Cost of Health-Related Lost Productive Time in Billions of Dollars per Year (2002 Dollars) in Obese and Overweight U.S. Workers and Expected Costs in the Absence of Excess Weight by Type of Lost Productive Time (LPT)* Obese Overweight Type of LPT LPT Cost (95% CI) Expected LPT Cost LPT Cost (95% CI) Expected LPT Cost Absenteeism ( ) ( ) Presenteeism ( ) ( ) Total LPT ( ) ( ) *Estimates benchmarked to the U.S. workforce using an iterative proportional fitting procedure described in the Methods section; expected costs were based on the estimated health-related LPT of normal-weight workers. Unable to estimate variance as a result of adjustment to reflect a population equivalent to obese workers in number of individuals. Unable to estimate variance as a result of adjustment to reflect a population equivalent to overweight workers in number of individuals. CI indicates confidence interval.

7 JOEM Volume 47, Number 12, December Furthermore, because this study focused only on LPT in actively employed workers, our obesity work-loss estimate does not fully account for other potential obesity-related costs such as unemployment, disability, and quality-adjusted life-years lost as a result of premature mortality. The relationship between obesity and health-related LPT is mediated by workers overall health status as measured in our study by number of health conditions present in the previous 2 weeks. It is well-documented that overweight and obese individuals experience increased morbidity associated with hypertension, type 2 diabetes, coronary heart disease, stroke, gallbladder disease, osteoarthritis, sleep apnea, respiratory problems, and some types of cancer (endometrial, breast, prostate, and colon). 18 Obesity also has been shown to exacerbate many chronic conditions such as hypertension and dyslipidemia. 8 In our study, we observed that obese workers exhibited a higher prevalence of self-reported hypertension, diabetes, heart disease, digestive problems, musculoskeletal pain, and respiratory problems than their nonobese counterparts, which is consistent with these previously documented relationships. Furthermore, we demonstrated that some of these health conditions such as musculoskeletal pain and digestive problems are important sources of LPT in obese workers. It is possible that intervention efforts targeting obese workers for health promotion and disease management programs may improve not only workers health status and quality of life, but also work productivity by decreasing the prevalence and impact of obesityrelated comorbidity. Overweight workers share some characteristics with both their normalweight and obese counterparts. On one hand, overweight workers were not significantly different from normal-weight workers in productive time lost for a health reason in the previous 2 weeks. On the other hand, they were significantly more likely than normal-weight workers to report several health conditions considered obesity-related risk factors such as hypertension and diabetes. Given that overweight is an intermediate state between normal weight and obesity, employers and overweight workers could potentially benefit from health and/or lifestyle interventions to lower overweight workers risk of additional weight gain and obesity-related comorbidities. This hypothesis requires further investigation. In addition to weight and height described previously, we relied on respondent self-report for other data used in our analysis including LPT in the previous 2 weeks, annual salary, and health status. The implications of using self-reported information in our study vary, depending on the measure. For example, LPT was measured using the Caremark Work and Health Interview, a selfreported data collection instrument, which was validated in a workplace setting using hourly electronic diary data over 10 workdays as the gold standard. The study demonstrated statistically significant correlations between the questionnaire-based and electronic diary-based LPT estimates and concluded that the questionnaire provided an accurate estimate of LPT at the population level. 15 Others have also demonstrated that one component of LPT (ie, absenteeism measured by episodes of sick leave) can be accurately reported by respondents. 20,21 On the other hand, self-reported salary is prone to high rates of nonresponse. Our use of linear regression analysis to impute salary for respondents with missing data may have introduced error into our LPT cost estimates. Given that individuals who refuse to provide salary information are more likely to be older, self-employed, or in professional occupations 22 (ie, more likely to be higher earners, on average, than those who report salary), we may have underestimated annual salary, and thus, LPT costs, in this group. Finally, subjective information on health status could also affect the observed relationship between obesity and LPT. A study comparing self-reported health status with physician assessments in subjects from the Third National Health and Nutrition Examination Survey found that subjects reporting excellent health decreased with increasing levels of obesity and overall, physicians rated health status more favorably than did subjects. 23 However, as obesity levels increased, the discrepancy between subjects and physicians assessments decreased; obese subjects were more accurate in reporting health status. If this occurred in our study, the effect would have been to decrease observed differences in health status between BMI groups because subjects with lower BMI would have overestimated health problems compared with those with higher BMI. This study has some potential limitations. First, we used BMI to categorize individuals as normal weight, overweight, and obese. BMI, an imperfect indicator of body fatness, can, for example, overestimate body fat in individuals with well-developed musculature and underestimate body fat in older individuals as the fat-to-lean body mass ratio increases progressively with age. 24 Second, our LPT estimates do not fully account for all lost labor costs associated with obesity. They did not include the potential costs of hiring and training replacement workers, impact on coworkers productivity, and employees forfeited leisure time. 25,26 It is unclear what net impact this combination of factors would have on our health-related LPT cost estimates. In conclusion, obesity in the U.S. workforce is associated with substantial LPT costs to employers primarily resulting from presenteeism. Workers overall health status mediates the relationship between obesity and LPT. Overweight workers are more similar to normal-weight workers than to obese workers in health characteristics and LPT profile. Health promotion and disease manage-

8 1234 LPT and Excess Weight in the U.S. Workforce Ricci and Chee ment programs, aiming to reduce the prevalence of excess weight in the workforce, and improve the health of obese workers, in particular, could have a positive impact on the productivity of the U.S. workforce. Acknowledgments This research was financially supported by Caremark. The authors thank Joshua Liberman at Caremark for comments on an earlier draft of the manuscript. References 1. Hertz RP, Unger AN, McDonald M, Lustik MB, Biddulph-Krentar J. The impact of obesity on work limitations and cardiovascular risk factors in the US workforce. J Occup Environ Med. 2004;46: National Institutes of Health (NIH). NIH releases research strategy to fight obesity epidemic. NIH News. August 24, Available at: pr/aug2004/niddk-24.htm. Accessed June 17, Van Itallie TB. Obesity: adverse effects on health and longevity. Am J Clin Nutr. 1979;32(suppl): Pi-Sunyer FX. Medical hazards of obesity. Ann Intern Med. 1993;119: Manson JE, Willett WC, Stampfer MJ, et al. Body weight and mortality among women. N Engl J Med. 1995;333: Flegal KM, Graubard BI, Williamson DF, Gail MH. Excess deaths associated with underweight, overweight, and obesity. JAMA. 2005;293: Fontaine KR, Redden DT, Wang C, Westfall AO, Allison DB. Years of life lost due to obesity. JAMA. 2003;289: Wolf AM, Colditz GA. Current estimates of the economic cost of obesity in the United States. Obes Res. 1998;6: Laitinen J, Nayha S, Kujala V. Body mass index and weight change from adolescence into adulthood, waist-to-hip ratio and perceived work ability among young adults. Int J Obes Relat Metab Disord. 2005;29: Pronk NP, Martinson B, Kessler RC, Beck AL, Simon GE, Wang P. The association between work performance and physical activity, cardiorespiratory fitness, and obesity. J Occup Environ Med. 2004;46: Narbro K, Jonsson E, Larsson B, Waaler H., Wedel H, Sjostrom L. Economic consequences of sick-leave and early retirement in obese Swedish women. Int J Obes. 1996;20: Burton WN, Chen C-Y, Schultz AB, Edington DW. The economic costs associated with body mass index in a workplace. J Occup Environ Med. 1998;40: Visscher TL, Rissanen A, Sidell JC, et al. Obesity and unhealthy life-years in adult Finns: an empirical approach. Arch Intern Med. 2004;164: Stewart WF, Ricci JA, Chee E, Morganstein D. Lost productive work time costs from health conditions in the United States: results from the American Productivity Audit. J Occup Environ Med. 2003;45: Stewart WF, Ricci JA, Leotta C, Chee E. Validation of the work and health phone interview. Pharmacoeconomics. 2004;22: Stewart WF, Ricci JA, Leotta C. Healthrelated lost productive time: recall interval and bias in cost estimates. J Occup Environ Med. 2004;46(suppl):S Bureau of Labor Statistics. Bureau of the Census. Current Population Survey. Available at: cps/cpsmain.htm. Accessed June 17, National Heart, Lung and Blood Institute (NHLBI) Panel on the Identification, Evaluation, and Treatment of Overweight and Obesity in Adults. NIH Publication No National Institutes of Health; September Palta M, Prineas RJ, Berman R, Hannan P. Comparison of self-reported and measured height and weight. Am J Epidemiol. 1982;115: Fredriksson K, Toomingas A, Torgen M, Thorbjornsson CB, Kilbom A. Validity and reliability of self-reported retrospectively collected data on sick leave related to musculoskeletal diseases. Scand J Work Environ Health. 1998;24: Severens JL, Mulder J, Laheij RJ, Verbeek AL. Precision and accuracy in measuring absence from work as a basis for calculating productivity costs in The Netherlands. Soc Sci Med. 2000;51: Turrell G. Income non-reporting: implications for health inequalities research. J Epidemiol Community Health. 2000;54: Okosun IS, Choi S, Matamoros T, Dever GEA. Obesity is associated with reduced self-rated general health status: evidence from a representative sample of white, black and Hispanic Americans. Prev Med. 2001;32: Prentice AM, Jebb SA. Beyond body mass index. Obes Rev. 2001;2: Berger ML, Murray JF, Xu J, Pauly M. Alternative valuations of work loss and productivity. J Occup Environ Med. 2001;43: Sach TH, Whynes DK. Measuring indirect costs: is there a problem? Appl Health Econ Health Policy. 2003;2: US Office of Personal Management (OPM). Federal Civilian Workforce Statistics: Occupations of Federal White- Collar and Blue-Collar Workers, as of September 39, Available at: Accessed June 17, 2005.