Worksite Nutrition

Guide to Community Preventive Services

T iteA onOALa arenLe igh RSa L. KSe

Ab matipro

menrventd pe

This review found that worksite nutrition and physical activity programs achieve modestimprovements in employee weight status at the 612-month follow-up. A pooled effect

Int

Ointpo

FrotionsonActandPubSciAntionDuverRes

AsonRo

34estimate of 2.8 pounds (95% CI4.6, 1.0) was found based on nine RCTs, and adecrease in BMI of0.5 (95% CI0.8,0.2) was found based on six RCTs. The findingsappear to be applicable to both male and female employees, across a range of worksitesettings.

Most of the studies combined informational and behavioral strategies to influence diet andphysical activity; fewer studies modified the work environment (e.g., cafeteria, exercisefacilities) to promote healthy choices. Information about other effects, barriers toimplementation, cost and cost effectiveness of interventions, and research gaps are alsopresented in this article. The findings of this systematic review can help inform decisionsof employers, planners, researchers, and other public health decision makers.(Am J Prev Med 2009;37(4):340357) Published by Elsevier Inc. on behalf of American Journal ofPreventive Medicine

roduction

besity is a major health problem in both devel-oped and developing countries. Many factorsgenetic, behavioral, social, and economic

eract to influence the development of obesity inpulations. People in societies with ample access to

energy-rich foods and low physical activity levels are atincreased risk of becoming overweight or obese. Inoccupational settings, economic and industrial innova-tion has resulted in far fewer workers in primaryindustries (e.g., agriculture, fishing, mining, or for-estry); more automation and labor-saving devices inproduction industries; and large increases in the pro-portion of people engaged in sedentary industries.Workplaces are a sedentary setting for many workersand also a place where access to energy-dense food andbeverages is common. Epidemiologic studies of char-acteristics of working conditions and worker over-weight or obesity have shown associations betweengreater BMI and long work hours, shift work, and jobstress.1 Schulte et al. recently described the associationbetween excess body weight and risk for a range ofoccupational conditions, including injury, asthma, mus-culoskeletal disorders, immune response, neurotoxic-ity, stress, cardiovascular disease, and cancer.1

m the Community Guide Branch, Division of Health Communica-and Marketing, National Center for Health Marketing (Ander-

, Quinn, Chattopadhyay, Kalra), Division of Nutrition, Physicalivity, and Obesity, National Center for Chronic Disease Prevention

Health Promotion (Buchanan, Archer), CDC; Rollins School oflic Health (Glanz), Emory University, Atlanta, Georgia; College of

ence and Health (Ramirez), Charles R. Drew University, Losgeles, California; Department of Veterans Affairs (Kahwati), Na-al Center for Health Promotion and Disease Prevention,

rham, North Carolina; School of Public Health (Johnson), Uni-sity of Washington, Seattle, Washington; and Yale Preventionearch Center (Katz), New Haven, Connecticutddress correspondence and reprint requests to: Laurie M. Ander-, PhD, MPH, Guide to Community Preventive Services, 1600 Clifton

ad, Mailstop E-69, Atlanta GA 30333. E-mail: [email protected] Effectiveness of Worksctivity Interventions for Cverweight and ObesitySystematic Review

urie M. Anderson, PhD, MPH, Toby A. Quinn, MPA, Kila C. Kahwati, MD, MPH, Donna B. Johnson, PhD, Lejal Chattopadhyay, PhD, Geetika P. Kalra, MPA, Davidrvices

stract: This report presents the results of a systenutrition and physical activity programs toThese results form the basis for the recomPreventive Services on the use of these inteweight in pounds or kilograms, BMI, aneffectiveness of these programs.0 Am J Prev Med 2009;37(4)Published by Elsevier Inc. on behalf of American Journal of PNutrition and Physicaltrolling Employee

Glanz, PhD, MPH, Gilbert Ramirez, DrPH,amsey Buchanan, PhD, W. Roodly Archer, PhD,

atz, MD, Task Force on Community Preventive

c review of the effectiveness of worksitemote healthy weight among employees.

dation by the Task Force on Communityions. Weight-related outcomes, includingrcentage body fat were used to assess0749-3797/09/$see front matterreventive Medicine doi:10.1016/j.amepre.2009.07.003

poca2 dbeassobabcagraultboeqgrainvwegreffcriPrtivanpr

Th

ThofmuFotivUSpaFo

onimlittiotivan

alsvarscorevwhassTa

HeCoAd

Threatheforthiincweob

Re

Exbeacop

Table 1. Selected Healthy People 201045 objectives related to adult overweight and obesity

Ob opul

Incwpc

eoplyea

Incw

eoplyea

Re

eoplyea

Ret

eoplyea

Incp3

eoplyea

aAg

OcConsidering that more than 30% of the U.S. adultpulation is obese and that a link between obesity andrdiovascular disease, hypertension, dyslipidemia, typeiabetes, stroke, osteoarthritis, and some cancers has

en established, concern about the economic burdenociated with obesity is growing.2,3 In the workplace,esity is an important driver of costs associated withsenteeism, sick leave, disability, injuries, and health-re claims.4 Employers are keenly interested in pro-ms and policies that improve worker health andimately reduce healthcare costs.5 Extant reviews,th qualitative69 and quantitative,1012 have yieldeduivocal results on the effectiveness of worksite pro-ms for controlling workers weight. These reviewsestigated multiple health risk outcomes besidesight status and did not attempt to quantify pro-am impact on weight as a summary measure ofect across the bodies of evidence reviewed. Theteria developed by the Task Force on Communityeventive Services13 were used to evaluate the effec-eness of worksite interventions targeting nutritiond physical activity behaviors among employees toomote healthy weight.

e Guide to Community Preventive Services

e systematic review in this report represents the workthe independent, nonfederal Task Force on Com-nity Preventive Services (the Task Force). The Task

rce oversees work on the Guide to Community Preven-e Services (Community Guide)14 with the support of theDHHS in collaboration with public and privatertners. The CDC provides staff support to the Taskrce for development of the Community Guide.Task Force recommendations are based primarily

the effectiveness of an intervention in improvingportant outcomes as determined by the systematicerature review process. In making its recommenda-ns, the Task Force balances information about effec-eness with information about other potential benefitsd harms of the intervention itself. The Task Force

jective P

rease the proportion of adults with high blood pressureho are taking action (e.g., losing weight, increasinghysical activity, and reducing sodium intake) to helpontrol their blood pressure (Objective 12-11)

P

rease the proportion of adults who are at a healthyeight (Objective 19-1)

P

duce the proportion of adults who are obese (BMI30) (Objective 19-2)

P

duce the proportion of adults who engage in no leisure-ime physical activity (Objective 22-1)

P

rease the proportion of adults who engage regularly,referably daily, in moderate physical activity for at least0 minutes per day (Objective 22-2)

P

e adjusted for the year 2000 standard populationtober 2009o considers the applicability of the intervention toious settings and populations in determining thepe of the recommendation. Finally, the Task Forceiews economic analyses of effective interventions,ere available. Economic information is provided toist with decision making but generally does not affectsk Force recommendations.

althy People 2010 Goals and Objectives forntrol of Overweight and Obesity Amongults

e interventions reviewed here may be useful inching objectives specified in Healthy People 2010,15

disease prevention and health promotion agendathe U.S. (Table 1). The interventions included in

s review focus on these objectives and the goal ofreasing the proportion of adults who are at a healthyight and reducing the proportion of adults who areese.

commendations from Other Advisory Groups

isting guidelines on the effects of counseling andhavioral strategies in improving diet and physicaltivity among overweight and obese adults were devel-ed for physicians, dieticians, and auxiliary personnelprimary care settings.1619 Counseling to increaseysical activity and improve diet, and behavioralategies to support and maintain these changes, areo relevant to worksite wellness programs. Recom-ndations of the U.S. Preventive Services Task ForceSPSTF) on such strategies are presented in Table 2.ere is mixed evidence on the effectiveness of behav-al counseling in the primary care setting in increas-

physical activity and limited evidence of the effec-eness of counseling in promoting a healthy dietong those not classified as having specific risk fac-s. However, among adults with hyperlipidemia ander known risk factors for cardiovascular and diet-

ated chronic disease, there is good evidence that

ation Baseline 2010 objective

e aged 18rs

82% of adult population(1998a)

95%

e aged 20rs


60%

e aged 20rs


15%

e aged 18rs


20%

e aged 18rs


30%inphstralsme(UThioringtivamtorothrelAm J Prev Med 2009;37(4) 341

intwittionicofhevegevicowitmemoyeaeffamcoselad

Methods

Covieelsthenitreptinaptioingouevistrlatsumwhpobariz

TmeBrObvarobRerevYal

Co

WosigcomproworanmeincfraprepoThindturaccriaoponrulbeshiInfknInfadcluvidtioofatt

Table 2. U.S. Preventive Services Task Forcerecact

Be

T

[Be

T

[T

[Scr

T

[T

[T

[

US

34ensive counseling (combined nutrition educationh behavioral dietary counseling provided by a nutri-nist, dietician, or specially trained primary care cli-ian) can produce meaningful change in daily intakeappropriate amounts of the core components of aalthy diet, including saturated fat, fiber, fruit, andetables. Further, the USPSTF found fair to gooddence that, in primary care settings, high-intensityunselingabout diet, exercise, or bothtogetherh behavioral interventions aimed at skill develop-nt, motivation, and support strategies producesdest, sustained weight loss (typically 35 kg for 1r or more) in adults who are obese (BMI 30). The

ectiveness of moderate- or low-intensity counselingong obese adults could not be determined, nor

uld the USPSTF determine the effectiveness of coun-ing to promote sustained weight loss in overweightults (BMI 25.029.9).

ommendations on behavioral counseling for physicalivity and diet, and screening for obesity among adults

havioral counseling to increase physical activity(www.ahrq.gov/clinic/uspstf/uspsphys.htm)he USPSTF concludes that the evidence is insufficientto recommend for or against behavioral counseling inprimary care settings to promote physical activity.

I recommendation]havioral counseling to promote a healthy diet

(www.ahrq.gov/clinic/uspstf/uspsdiet.htm)he USPSTF concludes that the evidence is insufficientto recommend for or against routine behavioralcounseling to promote a healthy diet in unselectedpatients in primary care settings.

I recommendation]he USPSTF recommends intensive behavioral dietarycounseling for adult patients with hyperlipidemia andother known risk factors for cardiovascular and diet-related chronic disease. Intensive counseling can bedelivered by primary care clinicians or by referral toother specialists, such as nutritionists or dieticians.

B recommendation]eening for obesity among adults (www.ahrq.gov/clinic/uspstf/uspsobes.htm)he USPSTF recommends that clinicians screen all adultpatients for obesity and offer intensive counseling andbehavioral interventions to promote sustained weightloss for obese adults.

B recommendation]he USPSTF concludes that the evidence is insufficientto recommend for or against the use of moderate- orlow-intensity counseling together with behavioralinterventions to promote sustained weight loss in obeseadults.

I recommendation]he USPSTF concludes that the evidence is insufficientto recommend for or against the use of counseling ofany intensity and behavioral interventions to promotesustained weight loss in overweight adults.

I recommendation]

PSTF, U.S. Preventive Services Task Force2 American Journal of Preventive Medicine, Volume 37, Numbemmunity Guide methods for conducting systematic re-ws and linking evidence to effectiveness are describedewhere13 and on the Community Guide website (www.communityguide.org/methods). In brief, for each Commu-y Guide review topic, a systematic review development teamresenting diverse disciplines, backgrounds, and work set-gs conducts a review by (1) developing a conceptualproach to identify, organize, group, and select interven-ns for review; (2) developing an analytic framework depict-

interrelationships among interventions, populations, andtcomes; (3) systematically searching for and retrievingdence; (4) assessing and summarizing the quality andength of the body of evidence of effectiveness; (5) trans-ing evidence of effectiveness into recommendations; (6)

marizing data about applicability (i.e., the extent toich available effectiveness data might apply to diversepulation segments and settings), economic impact, andrriers to implementation; and (7) identifying and summa-ing research gaps.

his review was conducted by a systematic review develop-nt team composed of CDC staff from the Community Guide

anch and the Division of Nutrition, Physical Activity andesity, along with a multidisciplinary team representing aiety of perspectives on worksite health promotion andesity prevention and control (Obesity Worksite Systematicview Development Team; see Acknowledgments). Theiew builds on an earlier systematic review conducted at thee Prevention Research Center.20

nceptual Model

rksite health promotion refers to strategies that are de-ned to improve health-related behaviors and health out-

es of workers. Worksite nutrition and physical activitygrams may occur separately or as part of a comprehensive

rksite health promotion program addressing a broaderge of objectives (e.g., smoking cessation, stress manage-nt, lipid reduction). These programs may or may notlude weight control as a primary objective. The analyticmework (Figure 1) depicts the components of such com-hensive programs. Worksite environmental change and

licy strategies are designed to make healthy choices easier.ey target the whole workforce or population, rather thanividuals, by modifying physical or organizational struc-es. Changes to the environment may include enhancingess to healthy foods (e.g., through modification of cafete-offerings or vending machine content) or enhancing

portunities to engage in physical activity (e.g., by providingsite facilities for exercise). Policy strategies may changees and procedures for employees, such as health insurancenefits or costs, reimbursement for health club member-p, or time allotted for breaks or meals at the worksite.ormational and educational strategies attempt to build theowledge base necessary to inform optimal health practices.ormation and learning experiences facilitate voluntaryaptations of behavior conducive to health. Examples in-de didactic instruction, health-related information pro-ed on the company intranet, posters or pamphlets, nutri-n education software, and information about the benefitshealthy diet and exercise. Behavioral and social strategiesempt to influence behaviors indirectly by targeting individ-r 4 www.ajpm-online.net

uaintstrprowegrocuof

Se

UsoccphyexesionbehhavbasDetive198anAbstrawoPreJou20refto

int

Fig al act

Ocl cognition (awareness, self-efficacy, perceived support,entions) believed to mediate behavior changes. Theseategies can include structuring the social environment tovide support for people trying to initiate or maintain

ight change. Such interventions may involve individual orup behavioral counseling, skill-building activities such as

e control, use of rewards or reinforcement, and inclusioncoworker or family members for support.

arch for Evidence

ing MeSH terms and text words (workplace or worksite/upational health services or occupational health/obesity or obese/sical activity or motor activity/weight loss/physical fitness/rcise/cardiovascular diseases/cholesterol/hyperlipidemia/hyperten-/nutrition/diet/body mass index/primary prevention/risk reductionavior/risk management/health promotion/health education/health be-ior/intervention studies/program evaluation) the following data-es were searched for studies between the date indicated andcember 2005: MEDLINE (1966); EMBASE (1980); Cumula-

Index to Nursing and Allied Health Literature (CINAHL,2); PsycINFO (1967); SPORTDiscus (1966); Latin American

d Caribbean Health Sciences Literature (1996), Dissertationstracts (1980), and the Cochrane Library (2005). Searchtegies are available at www.thecommunityguide.org/obesity/

rkprograms.html. Hand searches of the American Journal ofventive Medicine, Occupational Medicine, and the Internationalrnal of Obesity were conducted for the years 2004 through05. The reference lists of prior literature reviews, as well aserence lists from studies included in this review, were usedidentify relevant articles. Experts in obesity or worksite

ure 1. Analytic framework for worksite nutrition and physictober 2009erventions were consulted for additional citations.rches were limited to literature published in English.

dy Selection

be considered for inclusion in this review, studies had toevaluate a worksite health promotion program that in-

ded strategies involving diet, physical activity, or both;target current adult employees aged 18 years (not includ-retirees); and (3) provide data on at least one weight-

ated outcome measured at least 6 months from the start ofintervention program. Eligible interventions could be

geted at employees of any weight status (i.e., normalight, overweight, or obese), with or without identified risktors (e.g., elevated blood lipids, sedentary behavior) orditions (e.g., diabetes, hypertension).his review included worksite interventions with weighttrol or weight loss as the primary focus and also worksite

erventions aimed at general health promotion and riskuction (e.g., cardiovascular disease [CVD] risks, diabetes

ks) with a primary focus on healthy eating, physical activity,both. Intervention studies that evaluated commercial

ight-loss programs or products (e.g., Weight Watchers,m Fast) or diets 1000 calories per day were excluded.ration of intervention was defined in terms of months andluded both long-term (several sessions over severalnths) and short-term (e.g., one session) programs, as longthey provided a weight-related outcome measured at leastonths from the start of the intervention. No limitation wason worksite characteristics, including size of worksite,

mber of employees, nature of work (e.g., manufacturing,

ivity interventions to improve weight statusSea

Stu

To(1)clu(2)ingrelthetarwefaccon

TconintredrisorweSliDuincmoas6 msetnuAm J Prev Med 2009;37(4) 343

service industry), or location. Studies were eligible for inclu-sion if their intervention arm(s) was compared to an un-tretoeliadpu

Da

Stutortiodesturevaccprothethocolabpreponomeexeposurofarefoumoestexe

MeEf

Stumewedeob(an

1.

2.

3.

Waggmo

statistic.24 Statistical pooling of effects was done only whenthe studies and effect sizes were sufficiently similar to justifyintlev

Ostrpemekilcouanarmaggarm

Sduacrsamlensofeff

ReDe

Thvanstueriofthestuthetheals

Figstu

34ated comparison group or if an intervention was comparedanother intervention arm(s). Timeseries studies were alsogible. Studies were included whether or not they providedequate outcome variance statistics (e.g., SD, SE) to com-te CIs for their effect estimates.

ta Extraction and Quality Assessment

dy characteristics were coded by two independent abstrac-s using a web-based version of the Community Guide abstrac-n form.21 Each study was assessed for suitability of studysign and quality of study execution. Disagreements amongdy abstractors were reconciled by consensus among theiew team members. Classification of study designs is inord with the standards of the Community Guide reviewcess and sometimes differs from the classification used inoriginal studies. Studies with greatest design suitability arese in which data on exposed and control populations arelected prospectively; in studies with moderate design suit-ility, data are collected retrospectively or there are multiple- or post-measurements but no concurrent comparison

pulation; and in studies with least suitable designs, there iscomparison population and only a single pre- and post-asurement in the intervention population. Quality of studycution includes six categories of threats to validity (study

pulation and intervention descriptions, sampling, expo-e and outcome measurement, data analysis, interpretationresults, and other biases). Studies with no or one limitation

categorized as having good execution; those with two tor limitations have fair execution; and those with five orre limitations have limited execution.21 Studies with great-or moderate design suitability and good or fair quality ofcution were included in this review.

thods for Summarizing the Body of Evidence onfectiveness

dies typically reported means and SDs on continuousasures (e.g., mean number of pounds or kilograms ofight change or BMI change). Net program effects wererived by calculating the difference between the changesserved in the intervention (I) and comparison groupC) relative to their respective baseline levels. A three-partalytic strategy was used:

The effects (IC) on key outcomes of all studies (n31)in which an intervention was compared to an untreatedcontrol were examined, and pooled effects were estimatedusing study sample size as weights;Among studies (n21) in which an intervention wascompared to an untreated control, and variance data wereadequately reported, meta-analysis for pooling of effects(IC) on key outcomes was done using the inversevariance as the weight22;Among studies (n16) that compared an intervention toother intervention arms, change from baseline (I) foreach arm was summarized narratively to see how effectsvary according to intervention characteristics.

hen effect estimates with variance data were pooled, theregation of effect sizes was based on a random-effectsdel.23 Homogeneity of effects was tested using the Q4 American Journal of Preventive Medicine, Volume 37, Numbeegration (i.e., the Q statistic was nonsignificant at the .10el).

utcome measurements were related to the same con-uct: weight change (in pounds or kilograms, BMI, orrcentage body fat). These decision rules were used in theta-analyses: (1) where studies reported weight change in

ograms, these were converted to pounds; (2) a single studyld provide data for more than one of the outcomes above;

d (3) where studies compared two or more interventions to a common untreated control arm, the effects wereregated using comparisons of all applicable interventions.25

tratified analyses using random-effects models were con-cted to assess whether intervention effectiveness differedoss subgroups of studies based on characteristics of theple population, intervention features, study design, and

gth of follow-up assessment. Comprehensive Meta-Analysistware, version 2.2, was used for calculating pooledects.26

sultsscription of Included Studies

e literature search results and identification of rele-t studies is shown in Figure 2. A total of 54 candidatedies, reported in 78 papers, met inclusion crit-a.2780 Of these, seven studies7480 were consideredlimited quality of execution and were excluded from

review. A summary evidence table describing eachdy is available at the Community Guide website: www.communityguide.org/obesity/workprograms.html. Half

studies were conducted in the U.S.; studies wereo conducted in Europe, Australia, New Zealand,

ure 2. Flow diagram showing reasons for exclusion ofdies and number of qualifying studiesr 4 www.ajpm-online.net

Japco

tromuterstu(ratrirep(rastaoftphstuiorphcopoinfnepoallpastu(5intalwtheduintmestumostupolim

bluanbepafouAmmope

wewerepemrepstusta

In

Anusofcoweupamagthe(omatiosm(ramoonstustuagextenatpoda00stususu

Table 3. Study design and reporting characteristics of the 47 studies included in review

Clusrand

UnA 437,4

L 230,7

Mu 143

Ocan, Canada, India, and Iceland. Study design andmparison conditions are shown in Table 3.Studies comparing an intervention to untreated con-ls were analyzed separately from studies in whichltiple treatment arms were compared. Study charac-istics are summarized in Table 4. Among 46 of the 47dies in the analysis, the median sample size was 141nge 293728). The final study, a WHO multicountry

al, had a sample size of 63,732. Of the 39 studiesorting attrition, the median attrition rate was 17%nge: 0%82%). The purpose of the intervention, asted by the study author, was CVD risk reduction mosten (34%); followed by weight control (26%); andysical fitness (19%). The behavioral focus of 27dies (57%) was on diet and physical activity behav-s; 10 (21%) were diet only; and 10 (21%) wereysical activity only. Of the three types of interventionsded (1, informational; 2, behavioral skills; and 3,licy or environmental), 32 studies (69%) had bothormational and behavioral skills program compo-nts; among these, four included an environmental orlicy component. Intensity of the program, operation-y defined as the number of contacts with programrticipants, was reported as two to five contacts in 20dies (43%); more than five contacts in 26 studies

5%); and was not reported in one study. Ascertainingervention duration from the primary studies was notays possible. If the authors did not specifically state

duration of the program in weeks or months,ration was coded as the time from initiation of theervention to the time point when the first outcomeasurement occurred. Duration was 6 months in 19dies (40%); 69 months in 14 studies (30%); 1218nths in 8 studies (17%); and 18 months in 6dies (13%). Reporting on mean age of employeepulation, ethnicity, and urban or rural location wasited in the primary studies.

Study participants were coded as white-collar ore-collar workers based on descriptions in the studies

d nature of the company. In 19 studies, this could notdetermined; in 25 studies, the majority of partici-

nts (80%) held white-collar jobs; in the remainingr studies, the participants held blue-collar jobs.ong studies that reported gender of participants,st included both men and women (64%). Sixty-six

rcent of the studies indicated that some of the workers

RCTs

treated control group (n31)dequate variance data(n21)

1338,40,45,47,49,53,5862,65,66

imited variance data (n10) 428,32,35,55

ltiple intervention arms (n16) 727,34,36,39,44,57,63tober 2009re overweight, although the proportion that was over-ight was not generally reported. Half of the studiesorted the presence of chronic disease risks among theployee population. Only two of the randomized trialsorted intention-to-treat analysis,35,48 and many of thedies reported insufficient variance data to allow fortistical pooling with CIs (Table 3).

tervention Effectiveness

alysis, Part I. To broadly assess program effectsing all study designs and including studies regardlessavailability of variance statistics, the review team

mputed a difference between groups (IC) andighted each effect by study sample size at the follow-measurement interval. Three outcomes were ex-

ined: weight in pounds, BMI, and change in percent-e body fat. The results are presented in Table 5. Among

15 studies reporting a weight outcome in poundsr kilograms converted to pounds), the pooled sum-ry effect on change in weight favored the interven-n population, with few exceptions, and suggests aall difference of about 3 pounds at 612 monthsnge: 3.6 to 14.77; 21 data points). At 18 and 30nths, the effect is similar, but the result is based on

ly one study for each time period. At 60 months, onedy reported a 7-pound weight loss. Among the 15dies reporting a change in BMI, the results were

ain modest and favored the intervention groups,cept at 48 months. The BMI effects were less consis-t: 0.4 BMI (range: 0.3 to 1.57; 12 data points)

6 months; 0.02 BMI (range: 0.5 to 0.9; 12 dataints) at 12 months; 0.34 (range: 0.1 to 0.58; 4ta points) at 1824 months; 0.27 (range 0.2 to.75; 2 data points) at 36 months; and 0.08 (range:.4 to 1.7; 6 data points) at 48 months. Twelvedies reported change in percentage body fat, mea-

red most often as change in skinfold thickness. Themmary effect suggests a 1% decrease at 12 months.ble 6 provides information about all other outcomesorted in the studies with untreated comparisonups.

alysis, Part II. In the second stage of analysis, studyects were examined separately by study design: RCTs,ster RCTs, and nonrandomized designs. In 31 stud-, an intervention was compared to an untreated

teromized trials

Nonrandomizedtrials Cohort study

Timeseries

2,48,52 446,64,69,72

3 331,54,56 1 retrospective68 529,50,51,67,71 1 prospective70

1 retrospective40133Tarepgro

AneffcluiesAm J Prev Med 2009;37(4) 345

346A

merican

Journal

ofPreven

tiveM

edicine,

Volum

e37,

Num

ber4

ww

w.ajpm

-onlin

e.netTable 4. Characteristics of worksite diet and physical activity interventions reporting weight outcomesStudy and country N Study purposeOverwt (%)other risks (%)

Focus(diet,exercise,or both)

Components:informationalbehavioralenv & policy

Intensity:1: 1 contact2: 253: >5(duration inwk)

Pounds (mo f/u)IC

BMI(mo f/u)IC

Percent bodyfat(mo f/u)IC

Attrit(%)

RCTs with untreated comparison group (n17)

Aldana (2005)27 U.S. 145* CVD risk reduction NRNR

Both YYN

3 (6) 7.48 1.57 2 6

Barratt (1994)31Australia

683* CVD risk reduction NRNR

Diet YNN

3 (12) 62

Bruno (1983)34 U.S. 145* CVD risk reduction NRNR

Diet YYN

3 (8) (only reported idealwt)

33

Crouch (1986)37 U.S. 109 CVD risk reduction NRNR

Both YYN

2 (14) Arm A 3.97Arm B 1.10Arm C 0.88 (12 mo)

12

Drummond (1998)39Scotland

93 Weight loss NRNR

Diet YYN

2 (6) Arm A 1.1Arm B 2.65 (6 mo)

0.10.4

0.40.1

20

Fukahori (1999)44 Japan 108 Physical fitness NRCVD risk (NR)

Exercise NYY

3 (12) 0.5 (6 mo) 7

Gerdle (1995)46 Sweden 97 Physical fitness NRNR

Exercise NYN

3 (48) 2.2 (12 mo) 20

Grandjean (1996)48 U.S. 37 CVD risk reduction Yes (%NR)No

Exercise NYN

2 (24) 5.95 (6 mo) 2 NR

Juneau (1987)52 U.S. 120 Physical fitness Yes (25%)Sedentary

(100%)

Exercise YYN

2 (24) Men 2.43Women 1.1 (6 mo)

0.81.2

0

Krishnan (2004)54 India 100* Diabetes control Overwt (58%)Diabetic

(100%)

Both YNN

2 (4) Narrative group% changein BMI

17

Muto (2001)57 Japan 326 CVD risk reduction Overwt (65%)CVD risk

(% NR)

Both YYY

2 (4) 3.75 (6 mo)3.31 (18 mo)

0.5 (6 mo)0.5 (18 mo)

7

Nilsson (2001)58 Sweden 128 CVD risk reduction NRCVD risk (%

NR)

Both YYN

3 (72) 0.8 (12 mo)0.5 (18 mo)

31

Nisbeth (2000)59Denmark

85 CVD risk reduction Overwt (3.6%)CVD risk

(% NR)

Both YYN

2 (20) 3.53 (12 mo) 0.48 (12 mo) 29

Oden (1989)60 U.S. 45 Physical Fitness NRSmoke (13%)

Exercise NYN

3 (24) 2.56% NR

Okayama (2004)61 Japan 191 CVD risk reduction NR1 chol

Both YNN

2 (24) 1.1 (6 mo) 2

(continued on next page)

October

2009A

mJ

PrevM

ed2009;37(4)

347Table 4. (continued)Study and country N Study purposeOverwt (%)other risks (%)




Pounds (mo f/u)IC

BMI(mo f/u)IC


Attrit(%)

Pritchard(2002)64Australia

66 Weight loss BMI25(100%)1 B/P (% NR)

Both NYN

3 (48) Arm A 14.77Arm B 6.39 (12 mo)

12

Proper (2003)65Netherlands

299 Increase physicalactivity

NRNo

Both YYN

3 (36) 0.2 (9 mo) 0.8 (9 mo) 36

Cluster RCTs with untreated comparison group (n6)Anderson

(1999)29 U.S.234* CVD risk reduction NR

1 chol (100%)Diet Y

YN

2 (12) Arm A 3.6Arm B 5.4 (12 mo)

0.20.9

48

Cook (2001)36New Zealand

253 Healthy lifestyle Overwt: (36% Igp 40% Cgp)

hypertens (%NR)

Both YNY

3 (24) 0.66 (6 mo)0 (12 mo)

0.1 (6 mo)0 (12 mo)

6

Elliot (2004)41U.S.

33 Healthy lifestyle Overwt(% NR)

No

Both YYN

3 (24) Arm A 0.3Arm B 0.3

(6 mo)

0.30.4

0

Gomel (1993)47Australia

431 CVD risk reduction NRCVD risk

(%NR)

Both YYN

3 (24) Arm A 0.35Arm B 0.45Arm C 0.25

1.03.40.3

15

Jeffery (1993)51U.S.

32 sites(400900 empleach site)

Weight loss Overweight(36%)

CVD risk(% NR)

Both YYY

3 (96) 0.1 (24 mo) NR

WHO (1989)72UK, Belgium,Spain, Italy,Poland

63,732* CVD risk reduction x BMI 25.5CVD risk

(% NR)

Both YYN

2 (288) 0.7 (24 mo)0.4 (48 mo)0.6 (72 mo)

NR

Non-randomized studies with untreated comparison group (n7)Baer (1993)30

U.S.70* CVD risk reduction NR

1 chol (%NR)

Diet YYN

3 (48) 13.2 (12 mo) 4.0 (12 mo) 9

Furuki (1998)45Japan

1014 Health promot Overwt (14%)1 B/P & 1

chol (% NR)

Both YYN

NR4 yr f/u of

attenders& control

Men 0.4Women 0.0Overwt M 0.5Overwt W 0.1

51

Karlehagen(2003)53Sweden

169* CVD risk reduction Overwt (% NR)1 chol (%

NR)

Both YYN

2 (32) 0.58 (12 mo) 11

Linenger (1991)55U.S.

3728* Increase physicalactivity

NRNR

Both NYY

2 (48) 1.0 (12 mo) 50

Pohjonen(2001)63Finland

87 Physical activityeffects

NRNR

Exercise YYN

3 (36) 5.95 (12 mo)7.06 (60 mo)

1.3 (12 mo)0.7 (60 mo)

20

Talvi (1999)68Finland

885 Healthy lifestyle NR1 B/P (% NR)

Both YNN

3 (20) Men 0.20Women 0.75

10


348A

merican

Journal

ofPreven

tiveM

edicine,

Volum

e37,

Num

ber4

ww

w.ajpm

-onlin

e.netTable 4. Characteristics of worksite diet and physical activity interventions reporting weight outcomes (continued)Study and country N Study purposeOverwt (%)other risks (%)




Pounds (mo f/u)IC

BMI(mo f/u)IC


Attrit(%)

Weir (1989)79U.S.

258 Physical activitybenefits

NRNR

Exercise YNN

3 (12) W Arm A 2.6M Arm A 7.2W Arm B 2.6M Arm B 1.1

25.410.1

36

RCTs with different treatment arms (no untreated comparison) (n7)Abrams (1983)26

U.S.133* Wt loss (A) plus

maint (B)Overwt (% NR)No

Both YYY

3 (10) Arm A 9Arm B 3.3 (6 mo)

82

Brownell (1985)33U.S.

172 Weight loss Overwt (% NR)No

Diet YYN

3 (16) Study 3Arm A 5.9Arm B 5.5 (12 mo)

42

Cockcroft(1994)35England

297* Healthy lifestyle NRNR

Both YYN

2 (4) 0.55 (6mo) 72

DeLucia (1989)38U.S.

29 Weight loss 10 lbs. overwt(100%)

No

Diet YYN

3 (10) Arm A 6.5Arm B 4.21Arm C 5.49 (6 mo)

10

Forster (1985)43U.S.

131 Weight loss NRNR

Diet YYN

2 (24) Arm A (F) 11.0Arm A (M) 12.7Arm B (F) 11.4Arm B (M) 19.9

(6 mo)

21

Lovibond(1985)56Australia

75* CVD risk reduction Overwt (80%)CVD risks

(100%)

Both YYN

2 (8) Arm A 22.8Arm B 17.6Arm C 11.3 (6 mo)Arm A 21.1Arm B 18.3Arm C 12.1 (12 mo)

12

Peterson (1985)62U.S.

63* Weight loss Overwt (%NR)No

Both NYN

3 (8) Arm A 12.8Arm B 13.9 (6 mo)

30

Group RCTs with different treatment arms (no untreated control) (n1)Erfurt (1991)42

U.S.4 sites

(500600site)*

Health promot Overwt (30%)CVD risk

(1845%)

Both YYY

2 (144) High risk gpSite 1 3.1Site 2 0.6Site 3 1.2Site 4 4.7Over-wt gpSite 1 4.2Site 2 2.4Site 3 5.0Site 4 6.4 (8 mo)

NR

Cohort designs (n3)Elberson (2001)40

U.S.374* CVD risk reduction NR

1 chol (%NR)

Exercise YYN

3 (48) Arm A 0.57Arm B 0.3

N/A

Shimizu (2004)67Japan

629* CVD risk reduction Overwt (32%)1 chol1 B/P (% NR)

Both YYN

3 (192) Older 0.07Younger 0.30

N/A


October

2009A

mJ

PrevM

ed2009;37(4)

349Table 4. (continued)Study and country N Study purposeOverwt (%)other risks (%)




Pounds (mo f/u)IC

BMI(mo f/u)IC


Attrit(%)

Thorsteinsson(1994)69Iceland

155 Reduce CVD risk NR1 chol (% NR)

Diet YNY

2 (96) Arm A 0.1Arm B 0.3Arm C 0.1Arm D 0.4

N/A

Time series (n1)Briley (1992)32

U.S.40 Weight loss Overwt (100%)

NoDiet Y

YN

3 (16) 5 (12 mo) 30

Non-randomized studies with multiple treatment arms (no untreated control) (n5)Anderson

(1993)28 U.S.173* Weight loss x BMI 29

NRBoth Y

YN

2 (24) Arm A 3.9Arm B 12.9 (6 mo)

9

Harvey (1998)49U.S.

136* Healthy lifestyle NRCVD risk

(15%)

Exercise YNN

2 (48) Arm A 0.46Arm B 0.41

(12 mo)

0

Hedberg (1998)50Sweden

102* CVD risk reduction NR1 chol (%

NR)

Both YYN

2 (72) Arm A 0.3Arm B 0.2

(18 mo)

14

Robinson(1992)66 U.S.

137* Physical activitybenefits

NRNR

Exercise YYN

3 (24) Arm A 3.5Arm B 3.5

Arm A 1.6Arm B 0.3

32

Trent (1995)70U.S.

624* Weight loss Overwt (99%)No

Both YYN

3 (36) Arm A 1.7Arm B 4

(6 mo)Arm A 2.1Arm B 3.4

(12 mo)

41

*Study reported limited variance data1 elevated; attrit, attrition; BMI, body mass index; B/P, blood pressure; C, comparison; chol, cholesterol; CVD, cardiovascular disease; empl, employees; env, environmental; gp, group; hypertens,hypertension; I, intervention; M, men; maint, maintenance; mo, month; NR, not reported; overwt, overweight; promot, promotion; W, women; wk, week; wt, weight

covarOfincofsurepchcomo6estCITh

foc2wiRCdieThwe(9Litbetoanfivcestueff12venwitvenaddie

fatviswitpowa

poBMintch0

begrafitnbustuco

An

Table 5. Mean changes attributable to the intervention, weighted by sample size

Ou

Po

BM

Bo(

aSo e smabI pariso

35ntrol group, but only 21 of these reported adequateiance data to be included in this analysis (Table 3).these, four nonrandomized trials46,64,69,72 were notluded in these analyses because of the small numberstudies for each outcome measure (i.e., two mea-

red BMI and two weight in pounds). Nine RCTsorted weight change outcomes and six reported

ange in BMI. Four cluster RCTs reported BMI out-mes. Meta-analytic results using a random effectsdel for RCTs reporting weight change in pounds at

12 months are shown in Figure 3. The pooledimate indicates a change of 2.8 pounds (95%4.63, 0.96) in favor of the intervention group.e Q test for heterogeneity was nonsignificant.The pooled effect from three RCTs47,49,53 thatused on physical activity behaviors alone was.24 pounds (95% CI6.49, 2.00), compared

th 3.18 pounds (95% CI5.88, 0.50) in fiveTs38,58,60,62,65 in which both physical activity andtary behaviors were the focus of the intervention.e one study40 that focused on diet alone reportedight loss of1.71 pounds5% CI8.38, 4.95).tle difference was foundtween RCTs in which twofive contacts were maded those with more thane contacts, with the ex-ption of Prichard.65 Thisdy, which reported largerects than the others, was a-month weight-loss inter-tion for middle-aged men

h a BMI 25. One inter-tion group (n18) was

vised to follow a low-fatt (25% caloric intake as

Table 6. Impact on other ou

Study Ou

Pritchard (2002)65 Ab

Krishnan (2004)55 Gr

Nilsson (2001)59 Wa

Drummond (1998)40 Wa

Fukahori (1999)45 WaCook (2001)37 Wa

Bruno (1983)35 %

IC, difference between the c

tcome measured No. of study armsa No. of month

unds (n15) 928,37,40,49,53,58,62 61230,31,37,38,47,60,64,65 12158 18472 30164 60

I (n16) 1228,37,40,42,45,48,58,66 691230,37,46,48,59,60 12452,54,58,59 1824269 36646,68 48

dy fat (%)n12)

1328,40,42,48,49,53,61,66 69631,48,56,64 12472 30164 60

me studies had multiple arms. Number of citations may therefore b C, difference between the changes in the intervention and com0 American Journal of Preventive Medicine, Volume 37, Numbe); the other (n21) was advised to engage in unsuper-ed moderate exercise three times a week. Comparedh the control group (n19), a pooled effect of10.33unds was found. The benefit to the low-fatdiet groups twice that to the exercise group.Figure 4 shows changes in BMI among RCTs; theoled effect at the 612-month follow-up was 0.47I units (95% CI0.75, 0.19) in favor of the

ervention group. Among cluster RCTs reportinganges in BMI at 6 months,37,42,48 a pooled effect of.25 (95% CI0.64, 0.14) was found.

In subgroups analyses, no association was foundtween program effectiveness and focus of the pro-m (e.g., CVD risk reduction, weight loss, physicaless) or behavioral focus (diet or physical activity),

t this analysis was limited by a small number ofdies in each category that examined a similar out-me (i.e., weight, BMI, body fat).

alysis, Part III. Sixteen studies reported compari-s between different intervention arms (Table 3).

es reported (n7)

e Arm Months IC

nal fat (pounds) A 12 2.35B 12 1.25

change in BMI BMI 25 12 4.9BMI 2529 12 16.8BMI 30 12 11.9

-hip ratio 12 0.0118 0.01

-hip ratio A 6 0.01B 6 0.01

-hip ratio 6 0.05rcumference 6 0.7

12 0.5al body weight 8 3.5

s in the intervention and comparison group.

IC,b range No. of subjects Summary effect

0.66, 7.48 1104 2.8214.77, 3.6 905 3.153.31 302 3.311.10, 7.20 548 3.147.04 87 7.04

1.57, 0.3 1708 0.400.9, 0.5 2310 0.020.58, 0.1 946 0.340.2, 0.75 798 0.271.7, 0.4 1784 0.08

2.56, 1.2 1299 1.034, 0.1 1629 0.935.4, 0.1 548 1.880.7 87 0.70

ller than the number of study arms.n groupssson

tcom

tcom

domi

oup %

ist-to

ist-to

ist-toist ci

of ide

hanger 4 www.ajpm-online.net

Elegro(1prgrafivrizarecoresschdethatiothalaybe

exwo

Su

InweaimPrpomoRC1Simincdra(n

Fig Ts

Fig

Ocven of these studies could be divided into twoups based on similarity of the research questions:

) impact of programs with more or more intensiveogram components, or both and (2) impact of pro-ms offered by professional versus lay leaders. The

e remaining studies33,39,50,70,71 could not be summa-ed under a common research question. The results

summarized in Table 7. Offering multiple programmponents typically resulted in greater weight loss, but theults are not entirely consistent. Structured programs (i.e.,eduled individual or group sessions) for behavioral skills

velopment or physical activity conferred greater benefitn unstructured (i.e., self-directed) approaches. Informa-nal or educational approaches alone were less effectiven those that added behavioral counseling. With respect toversus professional group leaders, there appeared tolittle difference in program effects. Few studies

ure 3. Impact on weight in pounds at 612 months in 9 RC

ure 4. Impact on BMI at 612 months in 6 RCTstober 2009amined environmental approaches to improvingrker weight status.27,43,70

mmary of Effectiveness

conclusion, there is evidence of a modest reduction inight as a result of worksite health promotion programsed at improving nutrition, physical activity, or both.

ogram effects are consistent, with a net loss of 2.8unds (95% CI4.63,0.96) among workers at 612-nth follow-up, based on the meta-analysis of nineTs. In terms of BMI, a net loss of 0.47 BMI (95% CI.02, 0.2) at 612 months was observed in six RCTs.ilar results were observed for studies that could not be

luded in the meta-analysis. There was limited evidence tow conclusions about differential effects by program focus

utrition and physical activity) or program componentAm J Prev Med 2009;37(4) 351

(inicyarmintpaFosesproferwaon

Ap

Idprsurep

Table 7. Studies comparing multiple component approaches with fewer program components and studies with lay versuspro

Com E

Mup

BGEAA

E

SL

E

EEFABCD

en G

Layl

aINA

35formation, behavioral skills, or environmental and pol-). Among the group of studies comparing an intervention

with other intervention arms, when more or more-ensive modes of intervention were provided to partici-nts there appeared to be an increase in program impact.r example, offering structured programs (i.e., scheduledsions) appears more effective than unstructured ap-aches, and information plus behavioral counseling con-

s more benefit than providing information alone. Theres no apparent difference in program effectiveness basedlay versus professional group leaders.

plicability

entification of the effectiveness of worksite healthomotion programs on weight outcomes for specificbgroups of the population was constrained by limited

orting of important study population characteris-

fessional leaders

parison characteristic Study Study design Arm

ltiple versus fewerrogram components

Anderson (1993)29 RCT AB

Lovibond (1986)57 RCT ABC

Robinson (1992)67 Nonrandomized A

BAbrams (1983)27 RCT A

B

Erfurt (1991)43 RCT A

BCDABCD

Forster (1985)44 RCT A wom

B womC womD womA men

B menC menD men

Cockcroft (1994)36 RCT AB

Elberson (2001)41 Retrospective cohort AB

Hedberg (1998)51 Nonrandomized A

Bversus professional

eaderPeterson (1985)63 RCT A

BBrownell (1985)34 RCT A

B

, change in intervention group, not applicable2 American Journal of Preventive Medicine, Volume 37, Numbes. Ethnicity data were not reported, age was notorted in 70% of studies, and socioeconomic data

d information on blue- versus white-collar jobs weret reported in 40% of the studies. Although theesence of some overweight workers was indicated in% of studies, prevalence rates were not provided.Intervention programs take place in settings that mayve consequences for their effectiveness. In this re-w, the size of the worksite was not reported in 64% of

studies. Based on the data available, the results ofs review may be generalized to a white-collar work-ce where both overweight and other chronic diseasek conditions exist. Some studies27,37,43,46,52,56,58,70

amined policy and environmental changes in con-ction with instructional or behavioral approaches,

t it was difficult to summarize the contribution ofe environmental and policy component due to

lements NMonths frombaseline

I (poundsor BMI)a

ehavior skill development program 78 6 3.90oal setting only 95 12.90ducation, goal setting, training 25 22.82s programless intensity 25 17.64s programgreater loss ofintensity

25 11.36

ducation, goal setting, incentiveseducation

117 3.53

20 3.53tructured program, therapist 84 10.5 9.04ess-structured program, notherapist

49 3.31

ducation, counseling, socialsupport

783 36 4.7

ducation, counseling 1.2ducation 0.6itness center 3.1weight-loss program 6.4weight-loss program 5weight-loss program 2.4weight-loss program 4.2roup instruction, weigh-inrequired

19 6 11.3

roup instruction, weigh-in optional 21 10.7elf-instruction, weigh-in required 26 10.9elf-instruction, weigh-in optional 18 12roup instruction, weigh-inrequired

4 19.4

roup instruction, weigh-in optional 5 7.3elf-instruction, weigh-in required 2 24.5elf-instruction, weigh-in optional 8 18.8ealth screeningcounseling 297 6 0.54 BMIealth screening 0.1 BMI

tructured exercise 54 12 0.57 BMInstructured exercise 320 0.30 BMIealth assessmentgroupinstruction

102 18 0.3 BMI

ealth assessmentself instruction 0.2 BMIrofessional leader 30 6 12.79ay leader 33 13.89rofessional leader NA 12 5.9ay leader 5.5ticrepannopr66

haviethethiforrisexjunbuth

en Gen Sen S

G

GSSHHSUH

HPLPLr 4 www.ajpm-online.net

differences in comparison conditions and outcomesre

Ec

FodedeasEseliwhpuaBaseacoswitadexLitwethe

iesrewiliesforrisbinprseqthaba7.8anexcopewemathecluprTwpedisexpoweprincquGucojus

Price Index or the Medical Care component of the CPI(retheme

acowhmathrsptimorpo$4thra 1pecotioexanthebolostoiniveincdintimanseswepecastuAnlosgivwemaint$3$8ratwehoreq

ana wexclafirlabco

Ocported.

onomic Efficiency

r the systematic review of economic efficiency, thefinition and characteristics of the intervention, asfined by the effectiveness review team, were adoptedprimary inclusion and exclusion criteria for studies.tablished Community Guide standards to determinegibility for economic review are discussed else-ere.81 Broadly speaking, these require that studies beblished in English, be implemented in a country withhigh-income economy as defined by the Worldnk,82 and use an economic evaluation method. Therch strategy combined key economic terms such ast, costbenefit, costutility, and cost-effectiveness analysesh the terms used in the effectiveness review and, indition to the databases for the effectiveness review,amined economic-specific databases including Econ-

and the Social Science Citation Index. The datare abstracted following the procedures outlined in

economic abstraction form elsewhere.80

The search identified eight economic evaluation stud-falling within the scope of the current effectiveness

view: five cost-effectiveness studies,84,85,87,89,91 onelingness-to-pay study,85 and two cost-benefit stud-.83,86 One study90 that reported cost per employeeeach additional percentage point of cardiovascular

k reduced or relapse prevented, based on a com-ed measure of risk including weight loss, blood

essure, and smoking cessation components, was sub-uently excluded from the review. Another study90

t estimated an optimum number of fitness classessed on a participants willingness to pay for a loss ofkg in body weight, 8.8% loss in percentage body fat,

d 2.4 mmol/L loss in total cholesterol was alsocluded as the economic summary measure did notnsider actual expenditure per unit of weight loss orrcentage loss in body fat. One study87 looked at aight-loss outcome after 5 weeks, a follow-up periodrkedly shorter than the benchmark period set for

effectiveness review. This study, however, was in-ded in the current economic review because it

ovided important cost information for the program.o studies84,86 were identified that reported a returnr dollar invested in the program based on estimatedability and medical care costs averted; one86 wascluded from review because the outcome was re-rted as kilocalories expended per kilogram of bodyight per week for employees participating in a fitnessogram, rather than the actual loss in body weight. Allluded studies were rated as good following theality assessment criteria described in the Communityide economic abstraction form. For convenience ofmparability, economic summary measures were ad-ted to 2005 U.S. dollars using the all-item Consumertober 2009trieved from www.bls.gov), depending on whethermajority of cost items could be attributed to non-

dical or medical care goods and services.One study84 conducted a cost-effectiveness analysis ofworksite weight-loss program consisting of threempetitions held in business/industrial settings inich participants received a behavioral treatmentnual at each weigh-in. Intervention costs for theee competitions were $6149, $1377, and $762, re-

ectively, and included material costs and personnele for management, employees, and program staff to

ganize and supervise the program. The cost perund of weight lost in these three competitions was.16 for a 12-week program involving employees ofee banks, $2.19 for a 13-week program, and $1.60 for5-week program respectively, with the last two com-titions involving employees of two manufacturingmpanies. The lower cost per pound lost in Competi-ns 2 and 3 was due to decreased organizationalpenses compared to Competition 1. Another study91

alyzed the cost effectiveness of team competitions atworksite and estimated a cost of $1.45 to lose 1% of

dy weight. Effectiveness was measured by percentages in body weight rather than by amount of weight losspermit comparison of programs with widely varyingtial weights of participants. Although specific inter-ntion costs were not detailed in the study, costsluded time of the committee that planned, coor-ated, and administered the program, employeee, and minimal costs of photocopying manuals

d material costs for posters. A third study89 as-sed the cost effectiveness of a 3-month worksiteight-loss program that included concepts of com-tition and self-responsibility in an education-basedmpaign. The campaign cost was $25,376 and thedy reported a cost of $1.77 per pound of weight lost.other study87 reported costs for a self-help weight-s awareness campaign where each participant wasen a kit with information on how to start a safeight-loss program. Costs included personnel time andterials for typesetting, printing flyers, and posters. Theervention cost was $2634 excluding volunteer time and966 including 166 hours of volunteer time valued at.04 per hour. The study also reported cost-effectivenessios of $2.17 and $1.44, respectively, per pound ofight lost with or without the cost of volunteer time;wever, the follow-up period was much shorter than thatuired for the effectiveness review.

Finally, one study83 reported a reduction in disabilityd major medical costs of $1022.96 per participant atorksite physical fitness program for a 1-year period,

cluding costs of maternity or obstetrics-relatedims. The intervention costs of $75,750 included thest-year budget for operating expenses, annual cost oforatory tests and physical examinations, and annual

st of capital investment in equipment amortizedAm J Prev Med 2009;37(4) 353

more than 20 years. Although the study did not reportanderetres

stuboofabheincleawofavbatiomeem$7theweretpr

siteimbuniemeforoftrefocwhun

Ba

Notho

Ot

ThMacativmefou

Re

Alttarpoquwa

tions benefit the most from worksite health promotionintcoindlinheinmeagatmoemem

neincwofecmepewerepobintvenintpo

dewhtioers

revingprho64prCVbematiocoremwhprmoicainsfee

cafecRedetri

35y change in employee weight, there was a significantcline in the percentage of body fat. The interventionurned $1.59 for every dollar invested in the programulting in a net saving of $0.59 per $1.00.The range of cost-effectiveness estimates from threedies varies from $1.44 to $4.16 per pound of loss indy weight. Interpretation of the economic efficiencythese interventions is difficult without further detailsout how a pound of weight loss translates into a finalalth outcome of reduced incidence of a disease andreased quantity and quality of life-years. However, atst one study91 found that the cost effectiveness of arksite team competition to reduce weight comparedorably with commercial weight-loss programs. Also,sed on previous findings that a sustained 10% reduc-n in body weight would decrease expected lifetimedical care costs of hypertension, hypercholesterol-ia, type 2 diabetes, CHD, and stroke by $3258 (to

504) for men and by $3116 (to $7365) for women,88

intervention cost of $1.45 per 1% change in bodyight from this study appears to yield a high rate ofurn. The return would be even higher if savings fromoductivity loss averted are duly accounted for.In general, obesity prevention programs at the work-s may not only enhance employee self-confidence and

prove the relationship between management and labort also have the potential to boost the profits of compa-s by increasing employee productivity and reducingdical care and disability costs. More studies are neededdefinitive conclusions about the economic efficiency

such programs. However, recent worksite studies rarelyat obesity prevention as a single strategy and insteadus on comprehensive health promotion programsere obesity is but one of many components in theiverse of health risk factors considered.

rriers to Implementation

barriers to implementation were reported by au-rs of the studies reviewed.

her Benefits or Harms

is review looked at weight-related outcomes only.ny other physical and mental health effects were not

ptured, nor were possible benefits related to produc-ity or absenteeism. All studies were reviewed forntion of adverse effects, and no negative effects werend related to the interventions reviewed.

search Issues

hough evidence was found that worksite programsgeting nutrition and physical behaviors confer modest,sitive, weight-related benefits, important researchestions remain. One of the initial review questionss answered only partially: Which employee popula-4 American Journal of Preventive Medicine, Volume 37, Numbeerventions targeted at weight? Weight status variesnsiderably among employee populations. Reportingividual weight measures for employees from base-

e to follow-up is not feasible in large occupationalalth studies. Instead, measurement of weight changethe studies reviewed was usually presented as groupan change in BMI, pounds, kilograms, or percent-

e body fat. Thus, it could not be determined if thosegreatest risk (i.e., overweight or obese) benefitedre or less. Nor could it be determined if a fewployees lost a large amount of weight or if manyployees lost small amounts.

Growing numbers of employers have adopted well-ss programs as a means to lower health costs andrease productivity of workers. These employersuld benefit from studies that quantify program ef-ts at the population level. In addition to measuringan weight change, it would be useful to learn what

rcentage of participants had clinically meaningfulight loss (i.e.,5% or10% body weight loss). Also,orting changes in the prevalence of overweight and

esity in the employee population as a result of theervention would provide information about inter-tion effects at the population level. Highly effective

erventions that reach only a small percentage of thepulation will likely not affect the prevalence.Forty percent of the studies lacked information totermine differential effects according to blue- orite-collar job status. Those that did report occupa-nal status included predominantly white-collar work-. Race and ethnicity data were also limited.

A variety of worksite settings were represented in thisiew, which adds to the generalizability of the find-s. Information on the feasibility of implementing

ograms across small to very large worksite settings,wever, was limited by missing workplace size data in% of the studies. No association was found betweenogram effectiveness and focus of the program (e.g.,D risk reduction, weight loss, physical fitness) orhavioral focus (diet or physical activity). Because thejority of programs used behavioral plus informa-nal strategies, it was difficult to contrast programmponents with respect to effectiveness. Questions

ain about the effect on employee weight statusen implementation of environmental change (e.g.,

oviding easy access to affordable, healthy foods, ordifying the physical environment to encourage phys-l activity) and employer policy strategies (healthurance incentives, contribution to gym memberships) is included.

One third of the RCTs provided insufficient statisti-l information to allow meta-analytic pooling of ef-ts. Only a few reported intention-to-treat analysis.porting on intervention intensity, duration, and fi-lity was often ambiguous. Future studies will con-bute more to the empirical knowledge base if theyr 4 www.ajpm-online.net

follow the CONSORT guidelines for reporting RCTsanize

Dis

ThancostrThranshthicotimausuwewhberevtioturmaassimtivtieCogroRacakeaclittclu

mePrploseaevweanareexchpositemcli

CoYalGoWa

nity Preventive Services; and William Dietz, MD, PhD, Divi-sion of Nutrition, Physical Activity, and Obesity, NationalCetio

TanauO

TfolstuAnNiWaSerNeGin

Tiansea

TnitnoVe

Tlist

TtheThOR

TauRe

Re1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

Ocd TREND guidelines for reporting nonrandom-d studies.92,93

cussion

is review addressed the effects of worksite nutritiond physical interventions on employee weight out-mes. According to Community Guide rules,13 there isong evidence of a consistent, albeit modest, effect.e findings are applicable to men and women in age of worksite settings. However, some limitations

ould be considered when interpreting the results ofs review. Although several outcome measures may bellected in worksite health promotion studies, some-es not all are reported. The outcomes reported by

thors can be influenced by the significance of re-lts.94 This review included only studies that reportedight outcomes and thus may have omitted studies inich weight measures were collected but not reportedcause of nonsignificant effects. The findings of thisiew must be viewed within the context of the limita-ns of the available evidence in the published litera-e. Incomplete and less-than-transparent reportingde it difficult to use the full body of evidence toess program effectiveness. Finally, although it isportant for systematic reviews to report on the effec-eness of interventions to reduce health inequali-s,95 it was not possible in this review. The jointchrane and Campbell Collaborations health equityup recommends assessment of Place of residence,

ce or ethnicity, Occupation, Gender, Religion, Edu-tional level, SES, and Social capital (PROGRESS) asy indicators of how program effects are distributedross populations.96 With the exception of gender,le of this information was reported in studies in-ded in this systematic review.

This review, along with the accompanying recom-ndations from the Task Force on Community

eventive Services,97 should prove useful for em-yers, program planners, implementers, and re-rchers. It can support decisions to implement and

aluate worksite programs that promote healthyights through changes in diet and physical activity,d provide direction for further research in thisa. Although this intervention approach may be

pected to have only a modest effect on weightange, viewed from the population level it cantentially prevent and control overweight and obe-y when applied to a substantial proportion of theployee population and used in concert with other

nical and community approaches.

nsultants for the systematic review were David L. Katz, MD,e Prevention Research Center, New Haven CT; Ron Z.etzel, PhD, Institute for Health and Productivity Studies,shington DC, and member of the Task Force on Commu-tober 2009nter for Chronic Disease Prevention and Health Promo-n, CDC, Atlanta GA.his review began as a collaborative effort between CDC

d the Yale-Griffin Prevention Research Center, and thethors particularly want to thank David L. Katz, MD, MeganConnell, MPH, and Valentine Njike, MD.

he authors gratefully acknowledge the contribution of thelowing individuals who abstracted and coded data from thedies in this review: Carolynne Shinn, Stacie Anne Yung,ne Lund, Claire Roach, Angela Boardman, Maya Tholandi,cole Campbell, Susan Vendeland, Susan Anderson, Gregrd, Angela Thompson, Patrice Brooks, Parul Dube, Jannavi, Theresa Monica Dancel, Davida Carr, Andrea James,ha Desai, Jessica Marcinkevage, Michele Emory, and Sarahtel McGrath.he authors thank William Thomas, MLIS, research librar-with the CDC, for valuable assistance with literature

rches.he points of view are those of the Task Force on Commu-

y Preventive Services and of the respective authors and dot necessarily reflect those of the CDC or the Department ofterans Affairs.he names and affiliations of the Task Force members are

ed at www.thecommunityguide.org.he work of Toby Quinn was supported with funds fromOak Ridge Institute for Scientific Education (ORISE).

e work of W. Roodly Archer was supported in part byISE funds.he work of Donna Johnson was conducted under the

spices of the University of Washington Health Promotionsearch Center.

ferencesSchulte PA, Wagner GR, Ostry A, Blanciforti LA, Cutlip RG, Krajnak KM.Work, obesity, and occupational safety and health. Am J Public Health2004;97(3):42836.Popkin BM, Kim S, Rusev ER, Du S, Zizza S. Measuring the full economiccost of diet, physical activity and obesity-related chronic diseases. Obes Rev2006;7:27193.Kopelman P. Health risks associated with overweight and obesity. Obes Rev2007;8(1S):S137.Ostbye T, Dement JM, Krause MA. Obesity and workers compensation.Arch Intern Med 2007;167:76673.Goetzel RZ, Ozminkowski RJ. Whats holding you back: why should (orshouldnt) employers invest in health promotion programs for theirworkers? NC Med J 2006;67(6):42830.Heaney CA, Goetzel RZ. A review of health-related outcomes of multi-component worksite health promotion programs. Am J Health Promot1997;11(4):290308.Pelletier KR. A review and analysis of the health and cost-effective outcomestudies of comprehensive health promotion and disease prevention pro-grams at the worksite: 19931995 update. Am J Health Promot 1996;10(5):3808.Wilson MG, Holman PB, Hammock A. A comprehensive review of theeffects of worksite health promotion on health-related outcomes. Am JHealth Promot 1995;10(6):42936.Janer G, Sala M, Kogevinas J. Health promotion trials at worksites and riskfactors for cancer. Scand J Work Environ Health 2001;28(3):14157.Harden A, Peersman G, Oliver S, Mauthner M, Oakley A. A systematicreview of the effectiveness of health promotion interventions in theworkplace. Occup Med (Lond) 1999;49:5408.Proper KI, Koning M, van der Beek AJ, Hildebrandt VH. The effectivenessof worksite physical activity programs on physical activity, physical fitness,and heath. Clin J Sport Med 2003;13:10617.Am J Prev Med 2009;37(4) 355

12. Engbers LH, van Poppel MNM, Paw MJMC, van Mechelen W. Worksitehealth promotion programs with environmental changes. Am J Prev Med

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40. Drummond S, Kirk T. The effect of different types of dietary advice onbody composition in a group of Scottish men. J Hum Nutr Diet 1998;11:

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

53.

54.

55.

56.

57.

58.

59.

60.

61.

62.

63.

64.

352005;29(1):6170.Briss P, Zaza S, Pappaioanou M. Developing an evidence-based Guide toCommunity Preventive Servicesmethods. Am J Prev Med 2000;18(Suppl1):3543.Zaza S, Briss P, Harris KW. The Guide to Community Preventive Services.What works to promote health? New York: Oxford University Press, 2005.USDHHS. Healthy People 2010. www.healthypeople.gov/Document/HTML/Volume2/19Nutrition.htm.National Institutes of Health. Clinical guidelines on the identification,evaluation, and treatment of overweight and obesity in adults: the evidencereport. Obes Res 1998;6(2S):S51209.U.S. Preventive Services Task Force. Behavioral counseling in primary careto promote physical activity: recommendation and rationale. Arch InternMed 2002;137(3):2057.U.S. Preventive Services Task Force. Screening for obesity in adults:recommendations and rationale. Arch Intern Med 2003;139(11):9392.U.S. Preventive Services Task Force. Behavioral counseling in primary careto promote a healthy diet: recommendations and rationale. Am J Prev Med2003;21(1):93100.Centers for Disease Control and Prevention. Public health strategies forpreventing and controlling overweight and obesity in school and worksitesettings: a report on recommendations of the Task Force on CommunityPreventive Services. MMWR Morb Mortal Wkly Rep 2005;54(RR10):112.Zaza S, Wright de Aguero L, Briss P. Data collection instrument andprocedure for systematic review in the Guide to Community PreventiveServices. Am J Prev Med 2000;18(1S):S4474.Lipsey MW, Wilson DB. Practical meta-analysis. Thousand Oaks CA: Sage,2001.DerSimonian R, Laird N. Meta-analysis in clincial trials. Control Clin Trials2007;7:17788.Cooper H, Hedges L. The handbook of research synthesis. New York:Russell Sage, 1994.Hedges LV. Meta-analysis. In: Rao CR, ed. Handbook of statistics, vol. 26.Amsterdam: Elsevier, 2007:91953.Borenstein M, Hedges L, Higgins J, Rothstein H. Comprehensive metaanalysis [computer program]. Version 2.2. Englewood NJ: Biostat, 2006.Abrams DB, Follick MJ. Behavioral weight-loss intervention at the worksite:feasibility and maintenance. J Consult Clin Psychol 1983;51(2):22633.Aldana SG, Greenlaw RL, Diehl HA, Salberg A, Merrill RM, Ohmine S. Theeffects of a worksite chronic disease prevention program. J Occup EnvironMed 2005;47(6):55864.Anderson JV, Mavis BE, Robison JI, Stoffelmayr BE. A work-site weightmanagement program to reinforce behavior. J Occup Med 1993;35(8):8004.Anderson J, Dusenbury L. Worksite cholesterol and nutrition: an interven-tion project in Colorado. AAOHN J 1999;47(3):99106.Baer JT. Improved plasma cholesterol levels in men after a nutritioneducation program at the worksite. J Am Diet Assoc 1993;93(6):65863.Barratt A, Reznik R, Irwig L, et al. Work-site cholesterol screening anddietary intervention: the Staff Healthy Heart Project. Steering Committee.Am J Public Health 1994;84(5):77982.Briley ME, Montgomery DH, Blewett J. Worksite nutrition education canlower total cholesterol levels and promote weight loss among policedepartment employees. J Am Diet Assoc 1992;92(11):13824.Brownell KD, Stunkard AJ, McKeon PE. Weight reduction at the work site:a promise partially fulfilled. Am J Psychiatry 1985;142:4752.Bruno R, Arnold C, Jacobson L, Winick M, Wynder E. Randomizedcontrolled trial of a nonpharmacologic cholesterol reduction program atthe worksite. Prev Med 1983;12(4):52332.Cockcroft A, Gooch C, Ellinghouse C, Johnston M, Michie S. Evaluation ofa programme of health measurements and advice among hospital staff.Occup Med (Lond) 1994;44(2):706.Cook C, Simmons G, Swinburn B, Stewart J. Changing risk behaviours fornon-communicable disease in New Zealand working menis workplaceintervention effective? N Z Med J 2001;114(1130):1758.Crouch M, Sallis JF, Farquhar JW, et al. Personal and mediated healthcounseling for sustained dietary reduction of hypercholesterolemia. PrevMed 1986;15:28291.DeLucia J, Kalodner C, Horan J. The effect of two nutritional softwareprograms used as adjuncts to the behavioral treatment of obesity. J SubstAbuse 1989;1:2038.6 American Journal of Preventive Medicine, Volume 37, Numbe47385.Elberson KL, Daniels KK, Miller PM. Structured and nonstructured exer-cise in a corporate wellness program: a comparison of physiologicaloutcomes. Outcomes Manag Nurs Pract 2001;5(2):826.Elliot DL, Goldberg L, Duncan TE, et al. The PHLAME firefighters study:feasibility and findings. Am J Health Behav 2004;28(1):1323.Erfurt JC, Foote A, Heirich MA. Worksite wellness programs: incrementalcomparison of screening and referral alone, health education, follow-upcounseling, and plant organization. Am J Health Promot 1991;5(6):43848.Forster JL, Jeffery RW, Sullivan S, Snell MK. A work-site weight controlprogram using financial incentives collected through payroll deduction. JOccup Med 1985;27(11):80848.Fukahori M, Aono H, Saito I, Ikebe T, Ozawa H. Program of exercisetraining as Total Health Promotion Plan and its evaluation. J Occup Health1999;41(2):7682.Furuki K, Honda S, Jahng D, et al. The effects of a health promotionprogram on body mass index. J Occup Health 1999;41:1926.Gerdle B, Brulin C, Elert J, Eliasson P, Granlund B. Effect of a generalfitness program on musculoskeletal symptoms, clinical status, physiologicalcapacity, and perceived work environment among home care servicepersonnel. J Occup Rehabil 1995;5(1):116.Gomel M, Oldenburg B, Simpson JM, Owen N. Work-site cardiovascularrisk reduction: A randomized trial of health risk assessment, education,counseling, and incentives. Am J Public Health 1993;83(9):12318.Grandjean PW, Oden GL, Crouse SF, Brown JA, Green JS. Lipid andlipoprotein changes in women following 6 months of exercise training in aworksite fitness program. J Sports Med Phys Fitness 1996;36(1):549.Harvey HL. An evaluation of RMH Health Club: worksite wellness. DissertAbstr Int B: Sci Eng 1998;60(2B).Hedberg GE, Wikstrom-Frisen L, Janlert U. Comparison between twoprogrammes for reducing the levels of risk indicators of heart diseasesamong male professional drivers. Occup Environ Med 1998;55(8):55461.Jeffery RW, Forster JL, French SA, et al. The Healthy Worker Project: awork-site intervention for weight control and smoking cessation. Am JPublic Health 1993;83(3):395401.Juneau M, Rogers F, De Santso V, et al. Effectiveness of self-monitored,home-based, moderate intensity exercise training in middle-aged men andwomen. Am J Cardiol 1987;60:6677.Karlehagen S, Ohlson CG. Primary prevention of cardiovascular disease byan occupational health service. Prev Med 2003;37(3):21925.Krishnan N, Varman M, Roberts M. Diabetes control: role of healtheducation and other factors. A study in a newspaper industry. Indian J IndMed 2004;8(2):2933.Linenger JM, Chesson CV, Nice DS. Physical fitness gains following simpleenvironmental change. Am J Prev Med 1991;7(5):298310.Lovibond SH, Birrell PC, Langeluddecke P. Changing coronary heartdisease risk-factor status: the effects of three behavioral programs. J BehavMed 1986;9:41537.Muto T, Yamauchi K. Evaluation of a multicomponent workplace healthpromotion program conducted in Japan for improving employees cardio-vascular disease risk factors. Prev Med 2001;33(6):5717.Nilsson PM, Klasson EB, Nyberg P. Life-style intervention at the worksitereduction of cardiovascular risk factors in a randomized study. Scand JWork Environ Health 2001;27(1):5762.Nisbeth O, Klausen K, Andersen LB. Effectiveness of counseling over 1 yearon changes in lifestyle and coronary heart disease risk factors. Patient EducCouns 2000;40(2):12131.Oden G, Crouse SF, Reynolds C. Worker productivity, job satisfaction andwork-related stress: the influence of an employee fitness program. FitnessBus 1989;4:198204.Okayama A, Chiba N, Ueshima H. Non-pharmacological intervention studyof hypercholesterolemia among middle-aged people. Environ Health PrevMed 2004;9(4):1659.Peterson G, Abrams DB, Elder JP, Beaudin PA. Professional versus self-helpweight loss at the worksite: the challenge of making a public health impact.Behav Ther 1985;16(A2):21322.Pohjonen T, Ranta R. Effects of a worksite physical exercise interventionon physical fitness, perceived health status, and work ability amonghome care workers: five-year follow-up. Prev Med 2001 Jun;32(6):46575.r 4 www.ajpm-online.net

65. Pritchard JE, Nowson CA, Billington T, Wark JD. Benefits of a year-longworkplace weight loss program on cardiovascular risk factors. Nutr Diet2002 Jun;59(2):8796.

66. Proper KI, Hildebrandt VH, Van der Beek AJ, Twisk JWR, Van Mechelen W.Effect of individual counseling on physical activity fitness and health: arandomized controlled trial in a workplace setting. Am J Prev Med2003;24(3):21826.

67. Robison JI, Rogers MA, Carlson JJ, et al. Effects of a 6-month incentive-based exercise program on adherence and work capacity. Med Sci SportsExerc 1992;24(1):8593.

68. Shimizu T, Horiguchi I, Kato T, Nagata S. Relationship between aninterview-based health promotion program and cardiovascular risk factorsat Japanese companies. J Occup Health 2004;46(3):20512.

69. Talvi AI, Jarvisalo JO, Knuts LR. A health promotion programme for oilrefinery employees: changes of health promotion needs observed at threeyears. Occup Med (Lond) 1999 Feb;49(2):93101.

70. Thorsteinsson R, Johannesson A, Jonsson H, Thorhallsson T, SigurdssonJA. Effects of dietary intervention on serum lipids in factory workers. ScandJ Prim Health Care 1994;12(2):939.

71. Trent LK, Stevens L. Evaluation of the Navys obesity treatment program.Mil Med 1995;160:32630.

72. Wier LT, Jackson AS, Pinkerton MB. Evaluation of the NASA/JSC healthrelated fitness program. Aviat Space Environ Med 1989;60:43844.

73. World Health Organization European Collaborative Group. WHO Euro-pean collaborative trial in the multifactorial prevention of coronary heartdisease. Copenhagen: WHO, 1989.

74. Cox MH, Shephard RJ, Corey P. Influence of an employee fitness pro-gramme upon fitness, productivity, and absenteeism. Ergonomics 1981;24:795806.

75. Pescatello LS, Murphy D, Vollono J, Lynch E, Bernene J, Costanzo D. Thecardiovascular health impact of an incentive worksite health promotionprogram. Am J Health Promot 2001;16(1):1620.

76. Pingle SR, Deshpande AK, Malik JS. Impact of intervention strategies for

77.

78.

79.

80.

81. Carande-Kulis VG, Maciosek MV, Briss PA, et al. Methods for systematicreviews of economic evaluations for the Guide to Community PreventiveServices. Am J Prev Med 2000; 18(1S):7591.

82. World Bank. 2005 world development indicators. Available at: http://devdata.worldbank.org/wdi2005/home.htm. 2005.

83. Bowne D, Russell M, Morgan J, Optenberg S, Clarke A. Reduced disabilityand health care costs in an industrial fitness program. J Occup Med1984;26(11):80916.

84. Brownell K, Cohen R, Stunkard A, Felix M, Cooley N. Weight loss competitionsat the work site: impact on weight, morale and cost-effectiveness. Am J PubHealth 1984;74(11):12835.

85. Erfurt J, Foote A, Heirich M. The cost-effectiveness of worksite wellnessprograms for hypertension, control, weight loss, and smoking cessation. JOccup Med 1991;33(9):96270.

86. Gettman L. Cost/benefit analysis of a corporate fitness program. Fitness inBusiness 1986:117.

87. Nelson D, Sennett L, Lefebvre RC, Loiselle L, McClements L, Carleton R. Acampaign strategy for weight loss at worksites. Health Ed Res 1987:2(1):2731.

88. Osters G, Thompson D, Edelsberg J, Bird AP, Colditz GA. Lifetime healthand economic benefits of weight loss among obese persons. Am J PublicHealth 1999;89(10):153642.

89. Seidman L, Sevelius G, Ewald P. A cost-effective weight loss program at theworksite. J Occup Med 1984;26(10):72530.

90. Shirasaya K, Miyakawa M, Katsumi Y, Chieko T, Shimada N, Kondo T. Newapproach in the evaluation of a fitness program at a worksite. J OccupEnviron Med 1999;41(3):195201.

91. Stunkard A, Cohen R, Felix M. Weight loss competitions at the worksite:how they work and how well. Prev Med 1989;18:46074.

92. Moher D, Schulz KF, Altman DG. The CONSORT statement: revisedrecommendations for improving the quality of reports of parallel-grouprandomised trials. Lancet 2001;357:11914.

93. Des Jarlais DC, Lyles C, Crepaz N. Improving the reporting quality of

94.

95.

96.

97.

Ocrisk factor modification. Indian J Ind Med 2001;5(2).Rhodes EC, Dunwoddy D. Physiological and attitudinal changes in thoseinvolved in an employee fitness program. Can J Pub Health 1980;71(5):3316.Shi L. The impact of increasing intensity of health promotion interventionon risk reduction. Eval Health Prof 1992;15(1):325.Steinhardt M, Bezner JR, Adams TB. Outcomes of a traditional weightcontrol program and a nondiet alternative: a one-year comparison. J Psy-chol 1999;133(5):495513.Wier LT, Ayers GW, Jackson AS, Rossum AC, Poston WSC, Foreyt JP.Determining the amount of physical activity needed for long-term weightcontrol. Int J Obes 2001;25(5).tober 2009nonrandomized evaluation of behavioral and public health interventions:the TREND statement. Am J Public Health 2004;94(3):3616.Williamson PR, Gamble C. Identification and impact of outcome selectionbias in meta-analysis. Stat Med 2005;24:154761.Jackson N, Waters E. The challenges of systematically reviewing publichealth interventions. J Public Health 2004;26(3):3037.Tugwell P, Petticrew M, Robinson V, Kristjansson EML. Cochraneand Campbell Collaborations, and health equity. Lancet 2006;367:112830.Task Force on Community Preventive Services. A recommendation toimprove employee weight status through worksite health promotionprograms targeting nutrition, physical activity, or both. Am J Prev Med2010;In press.Am J Prev Med 2009;37(4) 357

The Effectiveness of Worksite Nutrition and Physical Activity Interventions for Controlling Employee Overweight and ObesityIntroductionThe Guide to Community Preventive ServicesHealthy People 2010 Goals and Objectives for Control of Overweight and Obesity Among AdultsRecommendations from Other Advisory Groups

MethodsConceptual ModelSearch for EvidenceStudy SelectionData Extraction and Quality AssessmentMethods for Summarizing the Body of Evidence on Effectiveness

ResultsDescription of Included StudiesIntervention EffectivenessAnalysis, Part IAnalysis, Part IIAnalysis, Part III

Summary of EffectivenessApplicabilityEconomic EfficiencyBarriers to ImplementationOther Benefits or HarmsResearch Issues

DiscussionAcknowledgmentReferences

Documents

Worksite Nutrition