Interventions to increase physical activity · 2010.32 The two main foci of the Healthy People preven-tion objectives are to increase (1) the amount of moderate or vigorous physical

The Effectiveness of Interventionsto Increase Physical ActivityA Systematic Review

Emily B. Kahn, PhD, MPH, Leigh T. Ramsey, PhD, Ross C. Brownson, PhD, Gregory W. Heath, DHSc, MPH,Elizabeth H. Howze, ScD, Kenneth E. Powell, MD, MPH, Elaine J. Stone, PhD, MPH, Mummy W. Rajab, MS,Phaedra Corso, PhD, and the Task Force on Community Preventive Services

Overview: The Guide to Community Preventive Service’s methods for systematic reviews were used toevaluate the effectiveness of various approaches to increasing physical activity: informa-tional, behavioral and social, and environmental and policy approaches. Changes inphysical activity behavior and aerobic capacity were used to assess effectiveness. Twoinformational interventions (“point-of-decision” prompts to encourage stair use andcommunity-wide campaigns) were effective, as were three behavioral and social interven-tions (school-based physical education, social support in community settings, and individ-ually-adapted health behavior change) and one environmental and policy intervention(creation of or enhanced access to places for physical activity combined with informationaloutreach activities). Additional information about applicability, other effects, and barriersto implementation are provided for these interventions. Evidence is insufficient to assess anumber of interventions: classroom-based health education focused on informationprovision, and family-based social support (because of inconsistent findings); mass mediacampaigns and college-based health education and physical education (because of aninsufficient number of studies); and classroom-based health education focused on reduc-ing television viewing and video game playing (because of insufficient evidence of anincrease in physical activity). These recommendations should serve the needs of research-ers, planners, and other public health decision makers.

Medical Subject Headings (MeSH): exercise, leisure activities, physical fitness, physicalendurance, decision making, evidence-based medicine, economics, preventive health services,public health practice, meta-analysis, review literature (Am J Prev Med 2002;22(4S):73–107)

Introduction

Regular physical activity is associated with en-hanced health and reduced risk of all-causemortality.1–4 Beyond the effects on mortality,

physical activity has many health benefits, includingreduced risk of cardiovascular disease,5,6 ischemicstroke,7–9 non–insulin-dependent (type 2) diabe-tes,10–16 colon cancers,17–20 osteoporosis, 21–23 depres-sion,24–27 and fall-related injuries.28–31 Despite the ben-efits of regular physical activity, only 25% of adults inthe United States report engaging in the recommendedamounts of physical activity (i.e., 30 minutes of moder-ate-intensity activity on 5 or more days per week, or 20minutes of vigorous-intensity activity on 3 or more daysper week)32; 29% report no leisure-time regular physi-cal activity33; and only 27% of students (grades 9through 12) engage in moderate-intensity physical ac-tivity (30 minutes, 5 or more days per week).32

In Healthy People 2010,32 physical activity is ranked asa leading health indicator. Healthy People 2010 hasdeveloped goals to improve levels of physical activityamong adults, adolescents, and children and to reducesedentary behavior among adolescents (Table 1).

Recommendations to increase physical activity havebeen made for individuals and clinical settings but notfor community settings. Increased physical activity has

From the Division of Prevention Research and Analytic Methods,Epidemiology Program Office (Kahn, Ramsey, Rajab, Corso), Divi-sion of Nutrition and Physical Activity, National Center for ChronicDisease Prevention and Health Promotion, Centers for DiseaseControl and Prevention (Heath), Atlanta, Georgia; the Task Force onCommunity Preventive Services and St. Louis University, School ofPublic Health (Brownson), St. Louis, Missouri; Division of HealthEducation and Promotion, Agency for Toxic Substances and DiseaseRegistry (Howze), Atlanta, Georgia; Epidemiology and PreventionBranch, Georgia Department of Human Resources (Powell), Atlanta,Georgia; and National Heart, Lung, and Blood Institute, NationalInstitutes of Health (Stone), Bethesda, Maryland

Dr. Stone is currently affiliated with the Department of PhysicalPerformance and Development, University of New Mexico, Albuquer-que, New Mexico.

The names and affiliations of the Task Force members are listed inthe front of this supplement and at www.thecommunityguide.org.

Address correspondence and reprint requests to: Peter A. Briss,MD, Community Guide Branch, Centers for Disease Control andPrevention, 4770 Buford Highway, MS-K73, Atlanta, GA 30341.E-mail: [email protected].

73Am J Prev Med 2002;22(4S) 0749-3797/02/$–see front matterPublished by Elsevier Science Inc. PII S0749-3797(02)00434-8

been linked not only to behavioral and social correlatesbut also to physical and social environmental corre-lates. Therefore, the role of community-based interven-tions to promote physical activity has emerged as acritical piece of an overall strategy to increase physicalactivity behaviors among the people of the UnitedStates. In 1996, the American College of Sports Medi-cine and the Centers for Disease Control and Preven-tion (CDC) recommended that every adult in theUnited States accumulate 30 minutes or more of mod-erate-intensity physical activity on most, preferably all,days of the week.34 That same year, the U.S. PreventiveServices Task Force recommended that healthcare pro-viders counsel all patients on the importance of incor-porating physical activity into their daily routines.35 Todate, community-based interventions to increase phys-ical activity have not been summarized in an evidence-based process.

The Guide to Community Preventive Services

The systematic reviews in this report represent the workof the independent, nonfederal Task Force on Com-munity Preventive Services (the Task Force). The TaskForce is developing the Guide to Community Preventive

Services (the Community Guide) with the support of theU.S. Department of Health and Human Services(DHHS) in collaboration with public and private part-ners. The Centers for Disease Control and Prevention(CDC) provides staff support to the Task Force fordevelopment of the Community Guide. A special supple-ment to the American Journal of Preventive Medicine,“Introducing the Guide to Community Preventive Services:

Methods, First Recommendations, and Expert Com-mentary,” published in January 2000, presents thebackground and the methods used in developing theCommunity Guide.36

Healthy People 2010 Goals and Objectives for

Increasing Physical Activity

The interventions reviewed in this article should beuseful in reaching the objectives set in Healthy People

2010.32 The two main foci of the Healthy People preven-tion objectives are to increase (1) the amount ofmoderate or vigorous physical activity performed bypeople in all population subgroups and (2) opportuni-ties for physical activity through creating and enhanc-ing access to places and facilities where people can bephysically active. This article provides information oninterventions that relate to both of these foci, which canbe used by communities to help increase levels ofexercise and fitness. The specific objectives are listed inTable 1.

Recommendations from Other Advisory Groups

The U.S. Preventive Services Task Force

In 1996, in the Guide to Clinical Preventive Services,35 theU.S. Preventive Services Task Force (USPSTF) recom-mended that healthcare providers counsel their pa-tients to incorporate regular physical activity into theirdaily routines. This recommendation was based on theaccepted health benefits of such activity rather than theproven effectiveness of clinician counseling. The

Table 1. Selected objectives for increasing physical activity (PA), Healthy People 201032

Percentage of population

Objective Population Baselinea 2010 objective

No leisure-time PA Adult 40% (1997) Reduce to 20%At least 30 minutes of moderate physical activity

regularly, preferably dailyAdult 15% (1997) Increase to 30%

At least 30 minutes of moderate physical activityon $5 of previous 7 days

Adolescents 27% (1999) Increase to 35%

Vigorous PA that promotes the developmentand maintenance of cardiorespiratory fitness$3 days per week for 20 minutes/occasion

Adult 23% (1997) Increase to 30%

Vigorous PA that promotes the developmentand maintenance of cardiorespiratory fitness$3 days per week for 20 minutes/occasion

Adolescents 65% (1999) Increase to 85%

Daily school physical education Adolescents 29% (1999) Increase to 50%View television #2 hours on a school day Adolescents 57% (1999) Increase to 75%Trips of #1 mile made by walking Adults 17% (1995) Increase to 25%Trips to school of #1 mile made by walking Children and adolescents 31% (1995) Increase to 50%Trips of #5 miles made by bicycling Adults 0.6% (1995) Increase to 2%Trips to school of #2 miles made by bicycling Children and adolescents 2.4% (1995) Increase to 5%aYears indicate when the data were analyzed to establish baseline estimates. Some of the estimates are age adjusted to the year 2000 standardpopulation.Source: U.S. Department of Health and Human Services. Healthy people 2010: conference edition. Washington, DC: U.S. Department of Healthand Human Services, 2000.32

74 American Journal of Preventive Medicine, Volume 22, Number 4S

USPSTF found that problems associated with establish-ing the effectiveness of counseling included lack ofinformation about long-term compliance and the typi-cal problems of generalizing to a broad population.The USPSTF is revisiting the effectiveness of clinicalinterventions for promoting long-term changes in phys-ical activity, and a revised recommendation will bereleased later this year.

American College of Sports Medicine/CDC

A panel of scientists convened by the American Collegeof Sports Medicine and CDC developed a consensusstatement recommending that every adult in theUnited States accumulate 30 minutes or more of mod-erate-intensity physical activity on most, preferably all,days of the week.34 This recommendation was modifiedto emphasize that physical activity does not need to becontinuous or strenuous to produce health benefits.The Surgeon General’s Report37 and the NationalInstitutes of Health Consensus Statement38 support thisrecommendation.

Methods for Conducting the Review

The general methods used to conduct systematic re-views for the Community Guide have been described indetail elsewhere.39,40 The specific methods for conduct-ing this review, including intervention selection, out-come determination, and search strategy for interven-tions to increase physical activity, are presented inAppendix A. The conceptual approach to the review,critical both for describing the methods and for under-standing the results of the review, is described below.

Conceptual Model

The general conceptual model (also called the “logicframework”) used to evaluate the effectiveness of inter-ventions to increase physical activity is shown in Figure1. This framework illustrates the relationships betweenphysical activity, several indicators of physical fitness,and morbidity and mortality outcomes. For example, alarge body of literature shows that increasing physicalactivity results in physiologic improvements, affectingendurance, strength, body composition, insulin sensi-tivity, and lipid levels.37 In turn, these improvementshave been shown to result in improved health andquality of life across a variety of conditions. Those whoare physically active have a reduced risk of developingcardiovascular disease,41–44 type 2 diabetes (formerlycalled non–insulin-dependent diabetes),14–16 coloncancers,17–20 osteoporosis,21–23 and depression,25–27

and of having fall-related injuries.28–31

Selection of Interventions for Review

The relationships between increased physical activityand health were assumed by the coordination team tobe well established and, subsequently, were not thefocus of the systematic review. Instead, the coordina-tion team focused on interventions to increase physicalactivity behaviors. By using the categories of behavioralprecedents established in the logic framework, thecoordination team (defined in Appendix A, see authorlist) initially developed the following three categories ofinterventions:

● Informational approaches to change knowledge andattitudes about the benefits of and opportunities forphysical activity within a community;

● Behavioral and social approaches to teach people thebehavioral management skills necessary both for suc-cessful adoption and maintenance of behaviorchange and for creating social environments thatfacilitate and enhance behavioral change; and

● Environmental and policy approaches to change thestructure of physical and organizational environ-ments to provide safe, attractive, and convenientplaces for physical activity.

Within these three categories, the coordination teamgenerated a comprehensive list of candidate interven-tions (Appendix A, Table A-1) for inclusion that ad-dressed each of the modifiable determinants (i.e.,individual level factors, social environment, and physi-cal environment). This list was put in priority order forreview through a process of polling the coordinationteam, consultation team,a and other specialists in thefield about their perception of the public health impor-tance (number of people affected), the practicality ofapplication, and the need of those promoting physicalactivity for information on each intervention. Interven-

aConsultants for the systematic review on increasing physical activitywere Terry Bazzarre, PhD, Robert Wood Johnson Foundation, Prince-ton, NJ; Carl J. Caspersen, PhD, National Center for Chronic DiseasePrevention and Health Promotion, CDC, Atlanta, GA; Diana Cassady,DrPH, California Department of Health Services, Sacramento; CarlosJ. Crespo, DrPH, State University of New York School of Medicineand Biomedical Sciences, Buffalo; Steve Hooker, PhD, CaliforniaDepartment of Health Services, Sacramento; Jonathan Fielding, MD,MPH, MBA, University of California Los Angeles School of PublicHealth; Barbara Fraser, RD, MS, Nebraska Department of Health andHuman Services, Lincoln; George J. Isham, MD, HealthPartners,Minneapolis, MN; Abby C. King, PhD, Stanford University School ofMedicine, Stanford, CA; I-Min Lee, MD, ScD, Harvard MedicalSchool/Brigham and Women’s Hospital, Boston, MA; Denise G.Simons-Morton, MD, PhD, National Heart, Lung, and Blood Insti-tute, National Institutes of Health, Bethesda, MD; Reba A. Norman,MLM, National Center for Chronic Disease Prevention and HealthPromotion, CDC, Atlanta, GA; Cindy Porteous, MA, IndianapolisPark Foundation, Indianapolis, IN; Michael Pratt, MD, MPH, Na-tional Center for Chronic Disease Prevention and Health Promotion,CDC, Atlanta, GA; Thomas Schmid, PhD, National Center forChronic Disease Prevention and Health Promotion, CDC, Atlanta,GA; Christine G. Spain, MA, The President’s Council on PhysicalFitness and Sports, Washington, DC; Wendell C. Taylor, PhD, MPH,University of Texas Health Science Center at Houston.

Am J Prev Med 2002;22(4S) 75

tions reviewed were either single component (i.e.,using only one activity to achieve desired outcomes) ormulticomponent (i.e., using more than one relatedactivity). Some interventions fell into more than onecategory.

Selection of Outcomes for Review

Many of the studies included in the body of evidencetargeted other behavioral risk factors in addition tophysical inactivity, most often poor diet and smoking.Many of the physiologic bene®ts observed in thosestudies could be the result of improvements in theseother behaviors rather than improvements in physicalactivity. To be able to estimate the effectiveness of anintervention in increasing physical activity, the teamdecided to base recommendations on changes in phys-ical activity behavior. However, self-reported measure-ment of behavior change, particularly in children, isprone to signi®cant measurement error; objective mea-sures are needed (e.g., heart rate monitoring, acceler-ometers) to enhance reliability and validity.45 Addition-ally, many of the studies did not attempt to measure

behavior even when the intervention targeted behavior,measuring instead physiologic changes as an indicatorof behavior. The team chose to use measures of aerobiccapacity as well as behavioral measures as outcomes onwhich recommendations would be based. Aerobic ca-pacity is de®ned as the maximum amount of oxygenthat can be transported from the lungs to the tissuesduring exercise.46 Maximal oxygen uptake (VO2 max)or a measure of VO2 max is commonly used to measurechanges in ®tness level associated with increased activ-ity (behavior change).

Results: Part I. Informational Approaches toIncreasing Physical Activity

Informational approaches are designed to increasephysical activity by providing information necessary tomotivate and enable people to change their behavior,as well as to maintain that change over time. The focusis mainly on the cognitive skills thought to precedebehavior. The interventions use primarily educationalapproaches to present both general health informa-

Figure 1. Logic framework illustrating the conceptual approach used in this review.


tion, including information about cardiovascular dis-ease prevention and risk reduction, as well as speci®cinformation about physical activity and exercise. Theseprograms were originally developed to complement amedical model of disease management by involvingcommunities in understanding the cognitive anteced-ents of behavior.

The provision of information is intended to changeknowledge about the bene®ts of physical activity, in-crease awareness of opportunities within a communityfor increasing physical activity, explain methods forovercoming barriers and negative attitudes about phys-ical activity, and increase participation in community-based activities. Interventions reviewed here are(1) ªpoint-of-decisionº prompts to encourage use ofstairs as an alternative to elevators or escalators,(2) community-wide education campaigns, (3) massmedia campaigns, and (4) classroom-based health ed-ucation focused on information provision and skillsrelated to decision making.

Point-of-Decision Prompts

Point-of-decision prompts are signs placed by elevatorsand escalators to motivate people to use nearby stairs.Messages on the signs recommend stair use for healthbene®ts or weight loss. Signs are thought to be effectivein one of two ways: by reminding people alreadypredisposed to becoming more active, for health orother reasons, about an opportunity at hand to be moreactive or by informing them of a health bene ®t fromtaking the stairs. All interventions evaluated in thiscategory were single-component interventions, inwhich placement of the sign was the only interventionactivity.

Reviews of evidence

Effectiveness. Our search identi®ed six reports (onepaper reported two studies)47±51 on the effectiveness ofpoint-of-decision prompts. All studies were of moderatesuitability, using time-series designs. All were con-ducted between 1980 and 2000. Two of the studies(reported in one paper) 49 were of good execution; theremaining four 47,48,50,51 were rated as fair. All wereincluded in the body of evidence. Details of the sixqualifying studies are provided at the website(www.thecommunityguide.org).

Baseline rates of stair use were generally low, with allbut one under 12% (range, 4.8% to 39.6%). In ®vestudies47±49,51 the median increase in stair-climbing was53.9%. The remaining study50 showed an unspeci®edincrease in stair-climbing and also found that the signswere effective in getting those who were less active (asmeasured by responses to a brief survey) to take thestairs. The range of effect sizes varied from a 5.5% netincrease to 128.6%.

Applicability. The body of evidence used to evaluatethe applicability of this intervention was the same asthat used to evaluate effectiveness. Four studies wereconducted in the United States,47,49,51 and one studyeach was conducted in England50 and Scotland.48 Thestudies included in this review were implemented inshopping malls,47,49,50 train and bus stations,48,49 and auniversity library.51 Three studies47,48,51 measured ef-fectiveness separately among men and women andfound that signs were effective in both groups.

Two studies47,49 reported effectiveness separately forobese and nonobese people. Although the signs wereeffective in both groups, the median net increase in thepercentage of people taking the stairs was greateramong the obese group. Among obese people, a signthat linked stair use to the potential for weight lossshowed a higher increase in stair use than the signlinking stair use to general health bene®ts.

The same two studies47,49 reported results separatelyfor blacks and found mixed evidence of effectiveness.One study47 showed a decline in the percentage ofblacks taking the stairs when the sign contained ageneric message. A message speci®cally designed for ablack population, however, was effective in increasingthe percentage of stair users.

This type of intervention is likely to be effectiveacross diverse settings and population groups, providedthat appropriate care is taken to adapt the messages.

Other positive or negative effects. None of the studiesmeasured outcomes other than the percentage of peo-ple using the stairs.

Economic. No studies were found that met the require-ments for inclusion in a Community Guidereview.40

Barriers to intervention implementation. Stairways inmany buildings and facilities are often dif ®cult to ®ndand poorly lit, maintained, or secured. As a result theymay appear to be, or actually be, unsafe. Additionally,some stairwells are locked, preventing user access tothem.

Conclusion. According to Community Guiderules ofevidence,39 suf®cient evidence shows that point-of-de-cision prompts are effective in increasing levels ofphysical activity, as measured by an increase in thepercentage of people choosing to take the stairs ratherthan an elevator or escalator. Findings from several ofthe studies suggest that tailoring the prompts either byspecifying the bene®ts of stair use or by customizing thesign to appeal to speci®c populations may increaseintervention effectiveness.

Community-Wide Campaigns

Community-wide campaigns involve many communitysectors in highly visible, broad-based, multiple interven-tion approaches to increasing physical activity. In addi-

Am J Prev Med 2002;22(4S) 77

tion to addressing sedentary behavior, most of thestudies in this body of evidence also addressed othercardiovascular disease risk factors, particularly diet andsmoking. Communication techniques were a commonelement in all of the campaigns. Campaign messageswere directed to large and relatively undifferentiatedaudiences through diverse media, including television,radio, newspaper columns and inserts, direct mailings,billboards, advertisements in transit outlets, and trailersin movie theaters. Messages were communicated in theform of paid advertisements, donated public serviceannouncements, press releases, the creation of featureitems, or a combination of two or more of theseapproaches.

In addition to incorporating substantial communica-tion activities through mass media, interventions in thisreview typically included some combination of socialsupport, such as self-help groups; risk factor screening,counseling, and education about physical activity in avariety of settings, including worksites, schools, andcommunity events; and environmental or policychanges such as the creation of walking trails. Theseinterventions were evaluated as aªcombined packageºbecause it was impossible to distinguish the relativecontributions of each component.

Reviews of evidence

Effectiveness. Our search identi®ed ten reports52±61

on the effectiveness of community-wide campaigns. Allstudies were used in the body of evidence. Two stud-ies52,53 had the greatest suitability of study design andgood execution. The remaining eight studies had fairexecution. Of those, ®ve54±58 had the greatest suitabilityof study design and three59±61 were of moderate suitabil-ity. Details of the ten qualifying studies are provided atthe website (www.thecommunityguide.org).

Many of the interventions52±58 were designed todecrease levels of cardiovascular disease morbidity andmortality throughout a community over a period ofseveral years. Activities were directed both at increasinglevels of physical activity and improving dietary behav-iors. These interventions included many activities sup-ported by media efforts but conducted independently.In the remaining studies,59±61the duration of interven-tion activities ranged from 6 weeks to 6 months, andactivities were more circumscribed, although still pre-sented community-wide.

Six arms from ®ve studies52,53,55,56,60 measuredchange in the percentage of people being active, with amedian net increase of 4.2% (range, 2 2.9% to 9.4%).Three arms from two studies54,58 measured change inenergy expenditure with a median net increase of16.3% (range, 7.6% to 21.4%). Five arms from fourstudies53,58±60 reported other measures of physical ac-tivity; all but one showed increases in physical activity.

Applicability. The body of evidence used to evaluatethe applicability of this intervention was the same asthat used to evaluate the effectiveness. Five studies wereconducted in the United States,52,54,56,58,59 and onestudy each was conducted in Sweden,57 Denmark,55

Australia,60 Scotland,61 and Wales.53 Studies were con-ducted in rural, suburban, and urban areas and encom-passed all socioeconomic groups. Two studies reportedresults for black populations,56,59 and two studies re-ported results for Hispanic or Latino populations. 54,59

This type of intervention is likely to be effectiveacross diverse settings and population groups, providedthat appropriate care is taken to adapt the interven-tions to the target populations.

Other positive or negative effects. Seven arms from®ve studies52±54,56,59 examined weight change amongpeople in the communities receiving the interventions.Measures included body mass index (BMI), weight,relative weight (i.e., ideal weight/actual weight), andthe percentage of people who were overweight. Threearms from three separate studies52,56,59 showed weightlosses, two arms from two studies52,54 showed nochanges, and two arms from one study53 showed slightweight gains over the course of the intervention. Acrossthe body of evidence, the net decrease in weight was0.6%. Four arms from two studies54,58 reported positivechanges in knowledge about exercise, with a medianincrease of 19.9%. Three studies55,61,62with seven studyarms reported physical activity intentions, that is, thepercentage of people either intending to increaseactivity or actually doing so. Five of the arms reportedan increase in intention to increase physical activity,one showed no change, and one reported a decrease.

In addition to increased physical activity, otherhealth bene®ts may result from the interventions. Mostof the studies in this body of evidence also addressedother cardiovascular disease risk behaviors, particularlydiet and smoking. Although these results were notsystematically reviewed, improvements were evident insome studies. Community-wide education campaignsmay also produce other bene®ts that can improvehealth and build social capital in communities. Forexample, through working together communities maydevelop a greater sense of cohesion and collectiveself-ef®cacy. Social networks may also be developed orstrengthened to achieve intervention goals, and com-munity members may become involved in local govern-ment and civic organizations, thereby increasing socialcapital.


Barriers to intervention implementation. Community-wide campaigns require careful planning and coordi-nation, well-trained staff, and suf®cient resources tocarry out the campaign as planned. Success is greatly


enhanced by community buy-in, which can take a greatdeal of time and effort to achieve. Insuf®cient resourcesmay result in exposure to messages and other plannedcampaign interventions that is inadequate to achievethe ªdosesº necessary to change knowledge, attitudes,or behavior over time, especially among high-risk pop-ulations. Inadequate resources and lack of profession-ally trained staff may affect how completely and appro-priately interventions are implemented and evaluated.

Conclusion. According to Community Guiderules ofevidence,39 strong evidence exists that community-widecampaigns are effective in increasing levels of physicalactivity, as measured by an increase in the percentage ofpeople engaging in physical activity, energy expendi-ture, or other measure of physical activity.

Mass Media Campaigns

Mass media campaigns are interventions that addressmessages about physical activity to large and relativelyundifferentiated audiences. The campaigns are de-signed to increase knowledge, in̄ uence attitudes andbeliefs, and change behavior. Messages are transmittedby using channels such as newspapers, radio, television,and billboards singly or in combination. In this review,interventions that use mass media and include othercomponents (e.g., support groups, risk factor screeningand education, and community events) are classi®ed ascommunity-wide campaigns and are discussed else-where (see Community-Wide Campaigns section).

Reviews of evidence

Effectiveness. Our search identi®ed three stud-ies54,59,63 evaluating the effectiveness of mass mediacampaigns. All studies had fair execution. The studydesigns were a nonrandomized community trial,54 atime-series,59 and a before-and-after design with noconcurrent comparison group. 63 These designs wererated as having greatest, moderate, and least suitablestudy designs, respectively. Details of the three qualify-ing studies are provided at the website(www.thecommunityguide.org).

Effectiveness measures were (1) the change in thepercentage of people doing a speci®ed level of activity(i.e., walking, moderate- and vigorous-intensity physicalactivity, aerobic and nonaerobic activity),59,63

(2) change in energy expenditure, 54 and (3) the per-centage of the population categorized as sedentary.63

Overall, some but not all measures suggested a modesttrend toward increasing physical activity, especially atlower levels of physical activity. According toCommunityGuiderules of evidence,39 we concluded that, becauseof small numbers of available studies and limitations inthe designs and execution of the available studies,insuf®cient evidence was available to assess the effec-

tiveness of mass media campaigns, when used alone, toincrease physical activity or improve®tness.

Applicability. Evidence about applicability was not as-sessed for this intervention because effectiveness wasnot established.

Other positive or negative effects. Two studies54,59

examined the effects of mass media campaigns onadiposity: One59 showed a decrease in weight over the3-month intervention period and the other 54 found noeffect. Signi®cant and substantial improvements inknowledge and beliefs were seen in two studies.54,63

Mass media campaigns might play important roles inchanging awareness of opportunities for and bene®ts ofphysical activity, helping to build support for environ-mental and policy changes that improve physical activ-ity behavior and ®tness, or both. Our reviews, however,did not assess the effect of mass media campaigns onsuch outcomes.

Economic. Evidence about economic effectiveness wasnot collected because effectiveness of the interventionwas not established.

Barriers to intervention implementation. Evidenceabout barriers to implementation of this interventionwas not collected because effectiveness was notestablished.

Conclusion. Available studies provide insuf®cient evi-dence to assess the effectiveness of mass media cam-paigns, when used alone, to increase physical activity orimprove ®tness. Media campaigns are, however, a com-ponent of other effective interventions (see Communi-ty-Wide Campaigns section) and might provide addi-tional bene®ts. Until more and better informationbecomes available, communities might choose to makedecisions about the use of education-only programs ongrounds other than the evidence found in availablestudies.

Classroom-Based Health Education Focused onInformation Provision

Health education classes that provide information andskills related to decision making are usually multicom-ponent, with the curriculum typically addressing phys-ical activity, nutrition, smoking, and cardiovasculardisease. Health education classes, taught in elementary,middle, or high schools, are designed to effect behaviorchange through personal and behavioral factors thatprovide students with the skills they need for rationaldecision making. Many of the classes in the studiesreviewed had a behavioral skills component (e.g., role-play, goal-setting, contingency planning) but did notadd time spent in physical activity to the curriculum. Inmost cases, comparison groups received the standardhealth education curriculum.

Am J Prev Med 2002;22(4S) 79

Reviews of evidence

Effectiveness. Our search identi®ed 13 studies64±76

evaluating the effectiveness of classroom-based healtheducation focused on information provision. Of these,three studies had limited quality of execution and werenot included in our review. 74±76 The ten remainingstudies had greatest suitability of study design. One hadgood execution64 and nine had fair execution. 65±73

Details of the ten qualifying studies are provided at thewebsite (www.thecommunityguide.org).

Most of the interventions were designed to reducethe risk of developing chronic disease. Four interven-tions65,71±73 were designed with the use of the KnowYour Body curriculum. (The Know Your Body curricu-lum is designed to provide children with the skillsneeded to adopt behaviors that reduce the risk ofdeveloping cardiovascular disease. The classes focus onnutrition, physical ®tness, and preventing cigarettesmoking.) One intervention 68 focused on prevention oftype 2 diabetes by encouraging students to eat low-fatfoods and to exercise regularly. The duration of theintervention activities ranged from 3 months to 5 years.

The studies reviewed showed variable effects of theseinterventions on time spent in physical activity outsidethe school setting: three study arms from two studiesshowed increases in activity,66±68 and ®ve study armsfrom two studies showed decreases in self-reportedactivity.64,66 Other measures of physical activity werealso varied: ®ve study arms from one study foundpositive changes in self-reported behavior,66 and 11study arms from two studies found no change ornegative changes in self-reported behavior.64,66Aerobiccapacity was not measured in any of these studies.


Other positive or negative effects. Although thesestudies did not show changes in activity, four of ®vestudy arms from three studies showed increases ingeneral health knowledge, exercise-related knowledge,and self-ef®cacy about exercise.64,67,68 Effects on adi-posity were mixed,69 showing decreases in BMI amongboth boys and girls but decreases in skinfold measure-ments among boys only.

Economic. Evidence about economic effectiveness wasnot collected because effectiveness of the interventionwas not established.


Conclusion. According to Community Guiderules ofevidence,39 the studies reviewed provided insuf®cient

evidence to assess the effectiveness of classroom-basedhealth education focused on information provision inincreasing levels of physical activity or improving®tnessbecause of inconsistent results among studies.

Results. Part II. Behavioral and Social Approaches toIncreasing Physical Activity

Behavioral and social approaches focus on increasingphysical activity by teaching widely applicable behav-ioral management skills and by structuring the socialenvironment to provide support for people trying toinitiate or maintain behavior change. Interventionsoften involve individual or group behavioral counselingand typically include the friends or family membersthat constitute an individual 's social environment. Skillsfocus on recognizing cues and opportunities for phys-ical activity, ways to manage high-risk situations, andways to maintain behavior and prevent relapse. Inter-ventions also involve making changes in the home,family, school, and work environments.

Interventions reviewed here are (1) school-basedphysical education (PE), (2) college-based health edu-cation and PE, (3) classroom-based health educationfocused on reducing television viewing and video gameplaying, (4) family-based social support interventions,(5) social support interventions in community settings,and (6) individually-adapted health behavior changeprograms.

School-Based PE

Interventions that used this approach modi®ed curric-ula and policies to increase the amount of time stu-dents spend in moderate or vigorous activity while inPE classes. This can be done in a variety of ways,including (1) adding new (or additional) PE classes,(2) lengthening existing PE classes, or (3) increasingmoderate to vigorous physical activity (MVPA) of stu-dents during PE class without necessarily lengtheningclass time. Examples of the last approach includechanging the activities taught (e.g., substituting soccerfor softball) or modifying the rules of the game so thatstudents are more active (e.g., having the entire teamrun the bases together if the batter makes a hit). Manyof these interventions also included the presentation ofinformation on cardiovascular disease prevention, ren-dering it dif ®cult to separate the effects of healtheducation and modi®ed PE.

Reviews of evidence

Effectiveness. Our search identi®ed 16 articles (onestudy was reported in two papers)77±93 reporting on 17studies that evaluated the effectiveness of modi®edschool-based PE curricula and policies. Of these, fourstudies had limited quality of execution and were not


included in our review. 78,84,85,92Twelve of the remain-ing thirteen studies (one study reported in two papers)had greatest suitability of study design.77,79±83,86±91,93

The remaining study79 had a least suitable design. Fourhad good execution80,81,83,89,93and nine had fair exe-cution.77,79,82,86±88,90,91 Details of the 13 qualifyingstudies are provided in Appendix B and at www.thecommunityguide.org.

Reported behavioral outcomes were energy expendi-ture,80,81,90 percentage of class time spent inMVPA,80,81,90,91 minutes spent in MVPA,80,81,90,91 ob-served activity score,77,90 and self-reported type andfrequency of physical activities outside ofschool.80,81,86,89,90,94Eleven studies also reported aero-bic capacity as estimated maximal oxygen uptake (VO2max),77,79 results from timed runs,77,80,81,88,90,93or en-durance testing (step test or shuttle runs).82,89,93Thesestudies showed consistent increases in time spent inphysical activity at school. Five arms from four studiesshowed increases in the amount77,91,94 and percent-age80,81,90,91of time spent in MVPA in PE classes. Thenet increase in the amount of PE class time spent inMVPA was 50.3% (range, 6.0% to 125.3%; because onestudy91 reported an increase from 0, percentage in-crease from baseline could not be calculated). The netincrease in the percentage of class time in MVPA was10% (range, 3.3% to 15.7%), with an additional studyreporting a 762% increase from a very small baselinevalue. Three arms from two studies80,81,90 showed in-creases in energy expenditure as well. Fourteen armsfrom eleven studies77,79±84,88±90,93 showed increases inaerobic capacity with a median of 8.4% (interquartilerange, 3.1% to 18.9%).

Applicability. The body of evidence used to evaluatethe applicability of this type of intervention was thesame as that used to evaluate effectiveness. Nine studieswere conducted in the United States,77,82,83,86±88,90,91,94

one in Crete,89 and two in Australia.79,93 Ten studies(one study was reported in two papers)77,79±81,86±91,93

were conducted among elementary school students,and two studies were conducted among high-schoolstudents.82,83 The intervention was successful amongboth elementary and high school students.

Overall results indicate that this type of interventionis likely to be effective across diverse settings andpopulation groups, provided that appropriate care istaken to adapt the intervention to the targetpopulation.

Other positive or negative effects. Ten stud-ies77,79,82,83,88±90,93±95 examined weight change. Mea-sures included BMI, skinfold measurements, percent-age of body fat, and weight. BMI estimates mostlyshowed small decreases or no change; one study armshowed a small increase of 1.2%. Skinfold estimatesshowed a median change of 2 4.4% (interquartilerange, 2 6.0% to 2 0.7%). Changes in percentage of

body fat showed inconsistent results, with some in-creases and some decreases.

Indicators of physical ®tness other than aerobiccapacity also showed improvements. Flexibility, as mea-sured through sit-and-reach tests, showed improve-ments in two87,89 of three studies.87,89,90 Muscular en-durance also increased in most arms in the twostudies89,90 reporting this measurement; the exceptionwas for girls performing pull-ups in one study.90 Theseinterventions were also associated with increases inknowledge about exercise,86,87 ®tness,89 nutri-tion, 80,81,87,88 and general health,83,89 as well as self-ef®cacy about physical activity.80,81

Although there is a perception that the time spent inPE may harm academic performance, both an exami-nation of the literature included in this review and theresults of a systematic search of other studies on theeffects of PE on academic performance found noevidence of this harm.


Barriers to intervention implementation. The primarybarrier to implementation exists within the schoolsystems. PE is mandated in almost every state, butrequirements for the amount of PE instruction aregenerally low (e.g., four semesters, two to three timesper week or two semesters of daily PE). Few middle andhigh schools require daily PE, and schools face increas-ing pressure to eliminate PE to make more timeavailable for academic subjects.

Conclusion. According to Community Guiderules ofevidence,39 there is strong evidence that school-basedPE is effective in increasing levels of physical activityand improving physical ®tness.

College-Based Health Education and PE

College-based health education and PE interventionsaim to set long-term behavioral patterns during thetransition to adulthood. To this end, they use didacticand behavioral education efforts to increase physicalactivity levels among college students. The physicaleducation classes do not have to be offered by PE orwellness departments in college and university settings,but they do include supervised physical activity in theclass.

The studies in our review included lecture classesthat addressed the bene®ts and potential risks of phys-ical activity, the current recommendations about theamount and type of physical activity one should get,and behavioral management techniques. Students ap-plied these lessons in ªlaboratoryº-type sessions inwhich they engaged in supervised physical activity,developed goals and activity plans, and wrote termpapers based on their experiences. Students also re-ceived social support and phone calls from each other

Am J Prev Med 2002;22(4S) 81

and made behavioral contracts for an agreed-onamount of physical activity.

Reviews of evidence

Effectiveness. Our search identi®ed ®ve studies (onestudy was reported in two papers)96±101 evaluating theeffectiveness of college-based health education and PEclasses. Three96,98,101 of the ®ve studies were not in-cluded in the review of effectiveness because they hadlimited quality of execution. The designs of the tworemaining studies were categorized as greatest suitabil-ity. One study, reported in two papers,99,100 had goodexecution and an initial intervention period of 15weeks; participants were followed for an additional 2years to examine maintenance of effect after the inter-vention ended. The other study97 was fair in executionand measured a 5-week intervention period. Details ofthe two qualifying studies are provided at the website(www.thecommunityguide.org).

The studies generally showed consistent increases inphysical activity and aerobic capacity in the short term.The 2-year follow-up showed declines in activity back toprevious levels and did not ®nd the desired effects onthe proposed mediators of behavioral change. Accord-ing to Community Guiderules of evidence,39 because ofthe small number of qualifying studies, limitations insome of the studies' design and execution, and someinconsistency in the results (with positive results mostlylimited to very short follow-up times), insuf ®cient evi-dence was available to assess the effectiveness of col-lege-based health education and PE interventions toincrease physical activity behavior and®tness.


Other positive or negative effects. One study, reportedin two papers,99,100measured effects on several hypoth-esized mediators of behavioral change. Among men,there was some evidence of an increase in the socialsupport of friends and family for physical activity and ofself-ef®cacy for resisting relapse into physical inactivity.Some evidence was available of a decrease in theperceived barriers to being physically active. Amongwomen, increases were found in the use of experientialand behavioral processes of change and statisticallynonsigni®cant increases in enjoyment, social support,self-ef®cacy, and perceived bene®ts and barriers.

Economic. Evidence about economic effectiveness wasnot collected for this intervention because effectivenesswas not established.

Barriers to intervention implementation. Evidenceabout barriers to implementation was not collected forthis intervention because effectiveness was notestablished.

Conclusion. Available studies provided insuf®cient ev-idence to assess the effectiveness of college-basedhealth education and PE interventions to increasephysical activity and®tness.

Classroom-Based Health Education Focused onReducing Television Viewing and Video GamePlaying

In elementary school classrooms, as part of a generalhealth curriculum, regular classroom teachers taughtclasses that speci®cally emphasized decreasing theamount of time spent watching television and playingvideo games. Classes included instruction in behavioralmanagement techniques or strategies such as self-mon-itoring of viewing behavior, limiting access to televisionand video games, and budgeting time for television andvideo. All studies included a ªTV turnoff challenge º inwhich the students were encouraged not to watchtelevision for a speci®ed number of days. Activities thatrequired greater energy expenditure than watchingtelevision or playing video games were not speci®callyrecommended. Parental involvement was a prominentpart of the intervention, and all households were givenautomatic television use monitors.

Reviews of evidence

Effectiveness. Our search identi®ed three studies102±

104 evaluating health education focused on reducingtelevision viewing and video game playing. All studieshad fair execution and greatest suitability of design,incorporating a concurrent comparison group andtaking measurements before and after the intervention.Two studies102,103 occurred over a 2-year interventionperiod, and one study104 took place during a 6-monthintervention period. All studies measured the timespent watching television and playing video games asthe primary outcome of interest. One study104 alsomeasured time spent in other sedentary behavior. Themain effectiveness measure from these studies used inour review was the time spent in vigorous,103 moderateor vigorous,102 or unspeci®ed104 physical activity perday.

The studies showed a consistent and sizable decreasein television viewing and video game playing for bothboys and girls according to children's self-report102±104

and parental report. 104 Time spent in other sedentarybehaviors also decreased in a single study.104 Reduc-tions in television viewing and video game playing didnot, however, consistently correspond with increases inphysical activity. Six measures of physical activityshowed inconsistent results, with two measures showingincreases and four measures showing decreases.

According to Community Guiderules of evidence,39

available studies provide insuf®cient evidence to assessthe effectiveness of health education classes focused on


reducing television viewing and video game playing toincrease physical activity behavior, because of inconsis-tent results for that outcome.


Other positive or negative effects. Two of the stud-ies102,104 examined effects on adiposity. One study104

showed signi®cant decreases in both BMI (2.3%) andskinfold measurements (11.5%). One study102 showeda 24.2% reduction in the prevalence of obesity amonggirls and no change among boys, as well as higher ratesof obesity remission in the intervention group com-pared with the control group. (Obesity remission oc-curs when a child who is de®ned as obeseªgrows intoºhis or her weight: body weight does not decrease butthe child grows in height, thus reducing the BMI to alevel that is no longer considered obese.) Because therewas no evidence of effect on levels of physical activity, itis thought that reductions in adiposity might be theresult of lower levels of snacking while watching televi-sion. An alternative explanation is that the interven-tions caused an increase in light physical activity thatwas not captured by the methods used.



Conclusion. Available studies provide insuf®cient evi-dence to assess the effectiveness of classroom-basedhealth education focused on reducing television view-ing and video game playing to increase physical activity.Such classes do, however, have additional bene®ts interms of reducing television watching and may lowerlevels of adiposity. More research is needed into thelinks between reducing time spent watching televisionor playing video games and increasing physical activity.

Family-Based Social Support

Family-based interventions attempt to change healthbehavior through the use of techniques that increasethe support of family members for behavior change.The family is a major source of in¯ uence for childrenin the modeling of health behaviors and is, therefore,an appropriate target for intervention. Many diseaserisk factors, both behavioral and physiologic, aggregatewithin families. Moreover, a supportive social environ-ment has been shown to increase maintenance ofbehavior change.

These interventions target factors in the social envi-ronment and interpersonal and behavioral patterns

that are likely to in ¯ uence physical activity behaviors.Interventions may be targeted to families with childrenor to spouses or partners without children. Programstypically include joint or separate educational sessionson health, goal-setting, problem-solving, or family be-havioral management and will often incorporate somephysical activities.

Interventions in this category targeted to childrenand their families are often implemented as part of alarger strategy that includes other school-based inter-ventions, such as school-based PE or classroom-basedhealth education. In this setting, the family componentis often conceptualized as an adjunct home curriculumto the school activities, involving take-home packets,reward systems, and family record keeping. They mayalso include family-oriented special events (e.g., theCATCH [Child and Adolescent Trial for CardiovascularHealth] program has Family Fun Nights, which areªmini-health fairsº for family and peers that offergames, prizes, food, and beverages).

Reviews of evidence

Effectiveness. Our search identi®ed 12 stud-ies67,80,81,87±90,105±110 on the effectiveness of family-based social support interventions. Of these, one hadlimited quality of execution and was not included inour review. One of the reviewed studies109 had a leastsuitable study design. The remaining 11 stud-ies67,80,81,87±90,105±108,110generally indicated no change,with some studies showing increases in activity andothers showing decreases. This inconsistency of resultsacross the body of evidence can also be seen in thephysiologic measures. Results were considered incon-sistent because many of the studies had multiple armswith varying results. Four studies reported an increase inenergy expenditure, and four arms from one study re-ported a decrease in energy expenditure. Both increasesand decreases also were seen in aerobic capacity and¯ exibility. Details of the 11 qualifying studies are pro-vided at the website (www.thecommunityguide.org).

The 11 studies included in this review can be broadlydivided into those that were implemented as part of aschool-based program67,87±90,94,107and those that wereimplemented as independent studies in the communitysetting.105,106,108,110Studies that were implemented aspart of a school-based program had slightly greaterevidence of effectiveness, but the combination of tech-niques makes it impossible to attribute that success tothe family-based social support intervention. In addi-tion, studies that compared home and school interven-tions with school-only interventions showed no differ-ences in effectiveness.


Am J Prev Med 2002;22(4S) 83

Other positive or negative effects. Eight arms from sixstudies67,87±89,107,108 showed increases in knowledgeabout disease risk factors,®tness, exercise, and healthamong both children and adults. One study94 showedevidence of an increase in self-ef®cacy for physicalactivity. One study108 showed an increase in satisfactionwith the amount of family activity.



Conclusion. According to Community Guiderules ofevidence,39 available studies provide insuf®cient evi-dence to assess the effectiveness of family-based socialsupport interventions in increasing levels of physicalactivity or improving ®tness, because of inconsistentresults among the studies.

Social Support Interventions in CommunitySettings

These interventions focus on changing physical activitybehavior through building, strengthening, and main-taining social networks that provide supportive relation-ships for behavior change. This change can be achievedeither by creating new social networks or workingwithin pre-existing networks in a social setting outsidethe family, such as the workplace. Interventions typi-cally involved setting up a ªbuddyº system, making aªcontractº with others to achieve speci®ed levels ofphysical activity, or setting up walking or other groupsto provide companionship and support while beingphysically active.

Reviews of evidence

Effectiveness. Our search identi®ed nine reports59,111±

118 on the effectiveness of social support interventionsin community settings. All studies were used in the bodyof evidence. One study116 had the greatest suitability ofdesign and good execution. The remaining eight studieshad fair execution. Of those, seven111±115,117,118had great-est suitability of study design and one59 was of moderatesuitability. Details of the nine qualifying studies are pro-vided at the website (www.thecommunityguide.org).

The typical intervention reviewed involved recruitingpeople into voluntary groups in which members pro-vided companionship and support for attaining self-selected activity goals. Each study participant receivedphone calls from other participants and from study staffmembers to monitor progress and to encourage con-tinuation of activities. Some studies involved formaldiscussion groups in which barriers to exercise and

negative perceptions about activity were addressed. Themeasured outcomes were varied and included thefrequency and duration of exercise episodes (e.g.,blocks walked daily, ¯ ights of stairs climbed daily,frequency of attendance at exercise sessions, frequencyof jogging episodes, participation in exercise and orga-nized sports, or minutes spent in activity).

Five arms from four studies59,111,112,117measuredchange in time spent in physical activity with a mediannet increase of 44.2% (interquartile range, 19.9% to45.6%). Six arms from three studies111,114,118measuredchange in frequency of exercise or physical activity witha median net increase of 19.6% (interquartile range,14.6% to 57.6%). Fitness also improved:®ve arms fromthree studies111±113 showed a median net increase inaerobic capacity of 4.7% (interquartile range, 3.3% to6.1%).

One study116 found that those who received morefrequent support were more active than those whoreceived less frequent support, although both highlystructured and less formal support were equally effec-tive in getting people to be more active.

Applicability. The body of evidence used to evaluatethe applicability of this intervention was the same asthat used to evaluate effectiveness. Seven studies wereconducted in the United States,59,111±116 one in Cana-da,118 and one in Australia.117 Six stud-ies59,111,112,115,117,118were conducted in community set-tings, including community centers and churches. Onestudy was conducted at a worksite113 and two wereconducted in university settings.114,116

Six studies111,112,114±116,118 exclusively or primarilyreported results for women, and three studies59,113,117

also included men in their study populations. Studypopulations for most studies were middle-aged; onestudy111 included women aged 18 years or older, andone study115 focused on ªolderº women (aged 50 to 65years). Three studies111,112,114 restricted their popula-tions to those that were sedentary at the beginning ofthe study; the rest included people at any initial level ofactivity. Given the diversity of countries, settings, andpopulations included in this body of evidence, theseresults should be applicable to diverse settings andpopulations, provided appropriate attention is paid toadapting the intervention to the target population.

Other positive or negative effects. Five arms from fourstudies59,111,112,117 reported changes in adiposityamong those enrolled in social support interventions.BMI, waist-to-hip ratio, and percentage of body fat andweight were measured. Ten of eleven measurementsshowed decreases in adiposity, with a median netchange of 2 7.3% (interquartile range, 2 8.1% to2 6.8%). Four arms from three studies111,112,117showedincreases in con®dence about exercise and in knowl-edge of and social support for exercise.



Barriers to intervention implementation. No barrierswere identi®ed to the implementation of social supportin community settings to promote physical activity.

Conclusion. According to Community Guiderules ofevidence,39 there is strong evidence that social supportinterventions in community settings are effective inincreasing levels of physical activity, as measured by anincrease in the percentage of people engaging inphysical activity, energy expenditure, or other measureof physical activity.

Individually-Adapted Health Behavior ChangePrograms

Individually-adapted health behavior change programsare tailored to the individual 's readiness for change,speci®c interests, and preferences. These programsteach participants speci®c behavioral skills that enablethem to incorporate moderate-intensity physical activityinto daily routines. Behaviors may be planned (e.g., adaily scheduled walk) or unplanned (e.g., taking thestairs when the opportunity arises). Many or most ofthese interventions use constructs from one or moreestablished health behavior change models such asSocial Cognitive Theory,119 the Health Belief Model, 120

or the Transtheoretical Model of Change.121 All pro-grams incorporated the following behavioral approach-es: (1) setting goals for physical activity and self-moni-toring of progress toward goals, (2) building socialsupport for new behavioral patterns, (3) behavioralreinforcement through self-reward and positive self-talk, (4) structured problem-solving geared to mainte-nance of the behavior change, and (5) prevention ofrelapse into sedentary behaviors. All of the interven-tions evaluated were delivered to people either ingroup settings or by mail, telephone, or directed media.

Reviews of evidence

Effectiveness. Our search identi®ed 20 reports on theeffectiveness of individually-adapted health behaviorchange programs.60,122±140 Of these, two studies138,139

had limited quality of execution and were not included inour review. Seven studies had good execution; of these,four,128,130,132,134 two,125,126 and one129 had greatest,moderate, and least suitable study design, respectively.Eleven studies had fair execution.60,122±124,127,131,133,135±

137,140 Of those, seven,60,123,124,127,135±137 two,122,133 andtwo131,140 had greatest, moderate, and least suitable de-signs, respectively. Details of the 18 qualifying studies areprovided at the website (www.thecommunityguide.org).

The typical intervention reviewed involved recruitingpeople into voluntary groups working toward physicalactivity goals. Members provided companionship and

support for attaining self-selected activity goals. Studyparticipants received phone calls from each other andfrom study staff members to monitor progress and toencourage continuation of activities. Some studies in-volved formal discussion groups in which barriers toexercise and negative perceptions about activity wereaddressed. The measured outcomes were varied andincluded the frequency and duration of exercise epi-sodes (i.e., blocks walked daily,̄ ights of stairs climbeddaily, frequency of attendance at exercise sessions,frequency of jogging episodes, participation in exerciseand organized sports, and minutes spent in activity).

Twenty arms from ten studies60,124±128,131,134±136

measured change in the time spent in physical activitywith a median net increase of 35.4% (interquartilerange, 16.7% to 83.3%). Thirteen arms from fourstudies125,126,132,136measured change in VO2 max witha median increase of 6.3% (interquartile range, 5.1% to9.8%; Figure 2). Fifteen arms from four studies125±

127,129 measured change in energy expenditure with anet median increase of 64.3% (interquartile range,31.2% to 85.5%). Other measures of physical activity,including attendance at exercise sessions,135 the num-ber of prescribed exercise sessions completed,132 thepercentage of people starting exercise programs,122

and the frequency of physical activity60,136 increased aswell.

Although we did not attempt a single quantitativesummary across the diverse effect measures becausethey could not be transformed to the same scale, theresults of the various effect measures support a gener-ally similar narrative conclusion: the preponderance ofthe evidence suggests that this type of intervention iseffective in increasing physical activity.

Applicability. The body of evidence used to evaluatethe applicability of this intervention was the same asthat used to evaluate effectiveness. Seventeen studieswere conducted in the United States122±137,140and onein Australia.60 Fourteen studies were conducted incommunity settings.60,123±134,140 Four studies were con-ducted at worksites,59,112,122,125 two at schools or uni-versities,135,136 and one at a telecommunicationscompany.137

These studies typically included volunteer samples,which limit generalizing to the entire population.Three studies123,124,131 reported results only forwomen. The remainder provided combined results formen and women. Study populations for most studieswere middle-aged; four studies128,130,132,135focused onpeople age 50 years or older.

Given the diversity of settings and populations in-cluded in this body of evidence, these results should beapplicable to diverse settings and populations, providedappropriate attention is paid to adapting the interven-tions to the target populations.

Am J Prev Med 2002;22(4S) 85

Other positive or negative effects. Fourteen of sixteenmeasurements from six studies125,126,129,131,140,141 re-ported a decrease in body weight, with a median netchange of 23.9% (interquartile range, 27.2% to 20.7%).All six measurements from three studies125,126,141 re-ported a decrease in percentage of body fat, with amedian of 24.1% (interquartile range, 25.9% to22.3%). Seven of nine measurements from two stud-ies130,141 reported an increase in strength, with a medianof 7.8% (interquartile range, 4.1% to 11.0%). Two mea-surements from two studies130,141 reported an increase inflexibility.

Economic. Our search identified one economic evalu-ation of individually-adapted health behavior changeprograms. This 2-year study,142 conducted at a fitnessfacility in Dallas, Texas, evaluated the cost-effectivenessof two physical activity interventions (lifestyle and struc-tured interventions) provided to adults aged 35 to 60years. The lifestyle intervention consisted of behavioralskills training to integrate moderate-to-intense physicalactivity into the lives of participants. Behavior modifi-cation and cognitive-behavior modification techniqueswere used for behavior change. The structured exerciseintervention consisted of supervised center-based exer-cise, in which participants received an exercise intensityprescription of 50% to 85% of maximum aerobic powerand exercise of 20 to 60 minutes at each session.Outcome measures included energy expenditure, mod-

erate activity and hard activity (kilocalories per kilo-gram per day), sitting (hours per week), walking (min-utes per day), stair climbing (flights per day), VO2 max(milliliters per kilogram per minute), and treadmilltime (minutes). The effect size for these outcomemeasures ranged between 0.23 (additional minutes onthe treadmill) and 13.07 (walking minutes per day),and 0.33 (additional kilocalorie per kilogram expendi-ture of moderate activity) to 26.75 (walking minutesper day for lifetime and structured interventions).Program costs included personnel, capital equipment,facilities, and general supplies, but research costs, re-cruitment costs, and value of participants’ time werenot included. The adjusted cost-effectiveness ratio foreach intervention arm ranged between $0.05 to $3.94and $0.07 to $5.39 per average unit (as defined in theoutcomes measured above) of improvement for life-style and structured intervention, respectively. Thisstudy was classified as good, based on the qualityassessment criteria used in the Community Guide.40 Theeconomic summary table for the study is provided inAppendix A, Table A-3, and at the website(www.thecommunityguide.org).

Barriers to intervention implementation. Individually-adapted health behavior change programs requirecareful planning and coordination, well-trained staffmembers, and resources sufficient to carry out theprogram as planned. Inadequate resources and lack of

Figure 2. Individually-adapted health behavior change: net percentage change in maximal oxygen consumption (VO2 max)from baseline in individuals being taught health behavior change strategies.con, control; ex, exercise; inten, intensity; lo, low; v, versus.


professionally trained staff members may affect howcompletely and appropriately interventions are imple-mented and evaluated.

Conclusion. According to Community Guide rules ofevidence,39 there is strong evidence that individually-adapted health behavior change programs are effectivein increasing levels of physical activity, as measured byan increase in the percentage of people engaging inphysical activity, energy expenditure, or other measureof physical activity.

Results. Part III. Environmental and Policy

Approaches to Increasing Physical Activity

Environmental and policy approaches are designed toprovide environmental opportunities, support, andcues to help people develop healthier behaviors. Thecreation of healthful physical and organizational envi-ronments is attempted through development of policythat lends itself to creating supportive environmentsand strengthening community action. Correlationalstudies have shown that physical activity levels areassociated with factors such as the availability of exer-cise equipment in the home and the proximity anddensity of places for physical activity within neighbor-hoods. Other neighborhood and environmental char-acteristics such as safety lighting, weather, and airpollution also affect physical activity levels, regardless ofindividual motivation and knowledge.

To affect entire populations, interventions in thiscategory are not directed to individuals but rather tophysical and organizational structures. The interven-tions are implemented and evaluated over a longerperiod of time than more individually oriented inter-ventions. Interventions are conducted by traditionalhealth professionals, but they also involve many sectorsthat have not previously been associated with publichealth, such as community agencies and organizations,legislators, and the mass media. The goal is to increasephysical activity through changing social networks, or-ganizational norms and policies, the physical environ-ment, resources and facilities, and laws. In addition tothe intervention reviewed below, two more interven-tions to increase physical activity by using environmen-tal and policy approaches (i.e., [1] transportation pol-icies and infrastructure changes to promotenonmotorized transit and [2] urban planning ap-proaches—zoning and land use) are under way and willbe included in a subsequent report.

Creation of or Enhanced Access to Places forPhysical Activity Combined with InformationalOutreach Activities

These interventions involve the efforts of worksites,coalitions, agencies, and communities to create or

provide access to places and facilities where people canbe physically active. For example, interventions in thebody of evidence include providing access to weightand aerobic fitness equipment in fitness centers orcommunity centers, creating walking trails, and provid-ing access to nearby fitness centers.

In addition to promoting access, many of thesestudies incorporated components such as training onequipment, health behavior education and techniques,seminars, counseling, risk screening, health forums andworkshops, referrals to physicians or additional ser-vices, health and fitness programs, and support orbuddy systems. These multicomponent interventionswere evaluated together because it was not possible toseparate out the incremental benefits of eachcomponent.

Reviews of evidence

Effectiveness. Our search identified a total of 12 stud-ies143–154 evaluating the effectiveness of creation of orenhanced access to places for physical activity com-bined with informational outreach activities. Of thesestudies, two were excluded because of limited studydesign.146,147 The remaining ten studies all had fairquality of execution, seven had greatest suitability ofstudy design,143–145,148,150,153,154 and three had leastsuitability of study design.149,151,152 Details of the tenqualifying studies are provided at the website(www.thecommunityguide.org).

Eight arms from five studies144,150,151,153,154 mea-sured change in aerobic capacity: the median increasewas 5.1% (interquartile range, 2.8% to 9.6%). Threearms from two studies144,153 measured change in en-ergy expenditure: the median increase was 8.2%(range, 22.0% to 24.6%). Four arms from two stud-ies145,148 measured change in the percentage reportingsome leisure-time physical activity: the median increasewas 2.9% (interquartile range, 26.0 to 8.5%). Six armsfrom one study152 measured exercise score: the medianincrease was 13.7% (interquartile range, 21.8% to69.6%). Four arms from three studies143,148,151 mea-sured the percentage reporting three or more exercisesessions per week, and two studies149,150 measuredfrequency of physical activity (Figure 3); the medianincrease was 48.4% (interquartile range, 21.0% to83.8%). Although we did not attempt a single quanti-tative summary across the diverse effect measures be-cause they could not be transformed to the same scale,the results of the various effect measures support agenerally similar narrative conclusion: the preponder-ance of the evidence suggests that this type of interven-tion is effective in increasing physical activity.

Applicability. All of the studies were conducted in theUnited States. Eight studies were conducted at work-sites, which included industrial plants (automotive,brewing, printing), universities, and federal agen-

Am J Prev Med 2002;22(4S) 87

cies.143,144,148–151,153,154 Two studies were conducted inlow-income communities.145,152 One study includedonly men150 and two studies stratified for men andwomen.145,152 One study152 included only blacks, andone study reported results specific to blacks.145 Giventhe diversity of settings and populations included inthis body of evidence, these results should be applicableto diverse settings and populations, provided appropri-ate attention is paid to adapting the intervention to thetarget population.

Other positive or negative effects. Ten arms from sixstudies145,148–150,153,154 examined weight change. Mea-sures included percentage of body fat, weight, andpercentage of people who were overweight. Nine armsfrom the same six studies showed decreases in percent-age of body fat or weight losses.145,148–150 One armfrom one study148 showed a weight gain. One studyshowed improvements in strength and improvementson the Physical Readiness Test composite score.153 Onestudy154 showed improvements in flexibility. Onestudy150 showed increases in perceived energy andconfidence in the ability to exercise regularly.

In addition to the direct health benefits in terms ofphysical activity, other health benefits also may haveresulted from the intervention. Many of the studies inthis body of evidence also addressed cardiovascular

disease risk factors, particularly diet and smoking,through provision of information.

Economic. Our search identified two economic evalu-ations155,156 of interventions to create or enhance ac-cess to places for physical activity. One 4-year study155

conducted at fitness facility in Houston, Texas, foremployees of an insurance company conducted a cost–benefit analysis of a structured physical fitness program.The program included regularly scheduled classes inaerobic dancing, calisthenics, and jogging; seminars onobesity, smoking, alcohol abuse, and stress reductionalso were offered. Program benefits included savings inmajor medical costs, reduction in average number ofdisability days, and reduction in direct disability dollarcosts. Program costs included personnel, nonsalaryoperating expenses, and medical claims. The adjustedestimates for benefits and costs for 1 year of theprogram are $1106 and $451, respectively. On the basisof the quality assessment criteria used in the Community

Guide this study was classified as good.40 The economicsummary table for the study is provided at the website(www.thecommunityguide.org).

A 5-year study156 with projections for an additional 10years was conducted in a workplace setting among36,000 employees and retirees of an insurance com-pany. The researchers conducted a cost–benefit analy-

Figure 3. Creation of or enhanced access to places for physical activity combined with informational outreach: net percentagechange from baseline in frequency of physical activity.PA, physical activity.


sis of a company-sponsored health and fitness programthat used health promotion centers, newsletters, medi-cal reference texts, videotapes, and quarterly mediablitzes. Program benefits included cost savings fromhealthcare costs averted, absenteeism reduction, deathsaverted, and increased productivity. Program costs in-cluded personnel, overhead, capital equipment, mate-rials, and rent; employee time away from the job wasnot included as a program cost. The adjusted estimatesfor benefits and costs are $139 million and $43 million,respectively. On the basis of the quality assessmentcriteria used in the Community Guide, this study wasclassified as good.40

Barriers to intervention implementation. One poten-tial barrier to creation of or enhanced access to placesfor physical activity is that building new facilities is timeand resource intensive. In addition, creation of orenhanced access to facilities requires careful planningand coordination, as well as resources sufficient to carryout the construction. Success is greatly enhanced bycommunity buy-in, which can take a great deal of timeand effort to achieve. Inadequate resources and lack ofprofessionally trained staff members may affect howcompletely and appropriately interventions are imple-mented and evaluated.

Conclusion. According to Community Guide rules ofevidence,39 there is strong evidence that creation of orenhanced access to places for physical activity com-bined with informational outreach activities is effectivein increasing levels of physical activity, as measured byan increase in the percentage of people engaging inphysical activity or other measures of physical activity.

Research Issues

Informational Approaches to Increasing PhysicalActivity

Effectiveness. The effectiveness of recommended andstrongly recommended interventions in this section(i.e., community-wide campaigns and point-of-decisionprompts) is established. However, important researchissues about the effectiveness of these interventionsremain.

Point-of-decision prompts

● What is the sustained effect of placing signs near theelevator or escalator?

● What effect does varying the message or format of thesign have on providing a “booster” to stair climbingamong the targeted population?

● What type of sign is most effective? What effect doformat or size have, if any?

● Does effectiveness vary by setting and targetaudience?

● Is there a “critical distance” from the elevator or

escalator to the stairs in which the effect of signageon stair-climbing behavior is reduced?

Because the effectiveness of mass media campaigns andclassroom-based health education focused on informa-tion provision has not been established, basic researchquestions remain.

● Are these interventions effective in increasing physi-cal activity?

● Do these interventions promote positive or negativeattitudes toward physical activity?

● Do these interventions promote changes in physicalactivity mediators, such as stage of change or changesin policy, which may lead to population shifts?

Community-wide campaigns

● What characteristics and components of community-wide campaigns are most effective?

● How can community-wide efforts be institutionalized?● What are the most effective and efficient delivery

settings and channels (e.g., media, work settings)?● Do coalitions enhance the delivery and effectiveness

of interventions in community settings? If so, is theenhanced effect worth the potential added cost andburdens of implementation?

Behavioral and Social Approaches to IncreasingPhysical Activity

Effectiveness. The effectiveness of recommended andstrongly recommended interventions in this section(i.e., school-based PE, social support interventions incommunity settings, individually-adapted health behav-ior change programs) is established. However, researchissues about the effectiveness of these interventionsremain.

School-based PE

● Is school-based PE as effective for preschool, elemen-tary, and high school students as for middle schoolstudents?

● Is effectiveness of school-based PE different in coedclasses versus single-sex classes in junior high andhigh school?

● Are classroom teachers as effective as PE specialists?● What is the relationship between PE class and overall

daily physical activity? Is activity outside the schoolsetting reduced when activity in PE is increased?

● Are before-school and after-school PE programs ef-fective in increasing student’s total daily activity levelsor improving fitness?

● Does physical activity incorporated into regularclasses result in effects similar to physical activityincorporated in a dedicated PE class?

● Is the effectiveness or efficacy of school-based PEaffected by school setting (e.g., type of school, urban,

Am J Prev Med 2002;22(4S) 89

suburban, etc.) or by population served (e.g., lowersocioeconomic status, racial or cultural differences)?

Social support interventions in communitysettings

● What type of social support and what medium worksfor whom? Do intensity and structure of the supportmake a difference?

● How does effect size vary by frequency of socialinteraction?

● Does the effect of these interventions vary by gender?

Individually-adapted health behavior change

● What characteristics and components are mosteffective?

● What mode of delivery is most effective?● Does the effectiveness of behavioral change method

vary by type of physical activity?

Because the effectiveness of college-based health edu-cation and PE, classroom-based health education fo-cused on reducing television viewing and video gameplaying, and family-based social support has not beenestablished, basic research questions remain.

● Are these interventions effective in increasing physi-cal activity?

● Do these interventions promote positive or negativeattitudes toward physical activity?

Environmental and Policy Approaches toIncreasing Physical Activity

Effectiveness. The effectiveness of the strongly recom-mended intervention in this section (i.e., creation of orenhanced access to places for physical activity) is estab-lished. However, research issues about the effectivenessof these interventions remain.

● What characteristics of a community are necessary forthe optimal implementation of policy and environ-mental interventions?

● Does the effectiveness vary by type of access (e.g.,worksite facility or community facility) or socioeco-nomic group?

● How can the necessary political and societal supportfor this type of intervention be created or increased?

● Does creating or improving access motivate sedentarypeople to become more active, give those who arealready active an increased opportunity to be active,or both?

● If you build it, will they come? In other words, isenhanced access to places for activity sufficient tocreate higher physical activity levels, or are otherintervention activities also necessary?

● What are the effects of creating new places forphysical activity versus enhancing existing facilities?

● Which neighborhood features (e.g., sidewalks, parks,traffic flow, proximity to shopping) are the mostcrucial in influencing activity patterns?

● How does proximity of places such as trails or parksto residence affect ease and frequency of use?

General Research Issues

Effectiveness. Several crosscutting research issuesabout the effectiveness of all of the reviewed interven-tions remain.

● What behavioral changes that do not involve physicalactivity can be shown to be associated with changes inphysical activity? For example, does a decrease intime spent watching television mean an increase inphysical activity or will another sedentary activity besubstituted? Does an increase in the use of publictransportation mean an increase in physical activityor will users drive to the transit stop?

● Physical activity is difficult to measure consistentlyacross studies and populations. Although severalgood measures have been developed, several issuesremain to be addressed.

● Reliable and valid measures are needed for thespectrum of physical activity.

Rationale: Current measures are better for vigor-ous activity than for moderate or light activity.Sedentary people are more likely to begin activityat a light level; this activity is often not captured bycurrent measurement techniques.

● Increased consensus about “best measures” for phys-ical activity would help to increase comparabilitybetween studies and would facilitate assessment ofeffectiveness.

Note: This is not intended to preclude researchers’latitude in choosing what aspects of physical activityto measure and to decide which measures are mostappropriate for a particular study population. Per-haps a useful middle ground position would be theestablishment of selected core measures that mostresearchers should use which could then be supple-mented by additional measures. The duration of anintervention’s effect was often difficult to determine.Although some researchers did attempt long-termfollow-up and assessment, many questions remain.

● How long does the effect of an intervention endureafter intervention activities cease?

● Does the duration of an intervention affect themaintenance of activity? For example, does a 2-yearintervention show effectiveness for a longer periodafter the intervention ends than a 10-week interven-tion? What strategies can be used to maintain anintervention effect after the intervention ends? Areperiodic “boosters” necessary or helpful?

● What is the nature and role of program “champions”


in ensuring the successful implementation and adop-tion of an intervention?

Applicability. Each recommended and strongly recom-mended intervention should be applicable in mostrelevant target populations and settings, assuming thatappropriate attention is paid to tailoring. However,possible differences in the effectiveness of each inter-vention for specific subgroups of the population oftencould not be determined. Several questions about theapplicability of these interventions in settings and pop-ulations other than those studied remain.

● Are there significant differences in the effectivenessof these interventions, based on the level or scale ofan intervention?

● What are the effects of each intervention in varioussociodemographic subgroups, such as age, gender,race, or ethnicity?

Other positive or negative effects. The studies in-cluded in this review did not report on other positiveand negative effects of these interventions. Research onthe following questions would be useful:

● Do informational approaches to increasing physicalactivity help to increase health knowledge? Is itnecessary to increase knowledge or improve attitudestoward physical activity to increase physical activitylevels?

● Do these approaches to increasing physical activityincrease awareness of opportunities for and benefitsof physical activity?

● What are the most effective ways to maintain physicalactivity levels after the initial behavior change hasoccurred?

● Are there other benefits from an intervention thatmight enhance its acceptability? For example, doesincreasing social support for physical activity carryover into an overall greater sense of community?

● Are there any key harms?● Is anything known about whether or how approaches

to physical activity could reduce potential harms(e.g., injuries or other problems associated withdoing too much too fast)?

Economic evaluations. The available economic datawere limited. Therefore, considerable research is war-ranted on the following questions:

● What is the cost-effectiveness of each of these inter-ventions? or

● How can effectiveness in terms of health outcomes orquality-adjusted health outcomes be better mea-sured, estimated, or modeled? or

● How can the cost benefit of these programs beestimated?

● How do specific characteristics of each of theseapproaches contribute to economic efficiency?

● What combinations of components in multicompo-nent interventions are most cost-effective?

Barriers. Research questions generated in this reviewinclude the following:

● What are the physical or structural (environmental)barriers to implementing these interventions?

● What resource (time and money) constraints preventor hinder the implementation of these interventions?

Summary

The Community Guide’s physical activity recommenda-tions identify intervention tools that practitioners canuse to achieve the Healthy People 2010 Objectives forPhysical Activity and Fitness.32 The Task Force stronglyrecommends community-wide health education cam-paigns, school-based PE, and social support in commu-nity settings, highlighting the role of multisite, multi-component interventions in successfully increasingphysical activity behaviors. Two strongly recommendedapproaches—individually-adapted health behaviorchange and creation of and enhanced access to placesfor physical activity combined with informational out-reach activities—point out the roles that policy andenvironmental approaches and behavioral and socialapproaches to increasing physical activity can play incombating inactivity in our culture. The recommenda-tion for the use of point-of-decision prompts as astrategy to promote physical activity underlines therelative simplicity of many of the recommended strate-gies. These recommendations should serve well theneeds of researchers, planners, and other public healthdecision makers in shaping the future agenda forefforts to explore and promote physical activity andthereby improve the health of the nation.

We thank the following individuals for their contributionsto this review: William Callaghan, Preventive Medicine Resi-dent; Jonathan Stevens, Research Assistant; Benedict I. Tru-man, Stephanie Zaza, Maureen McGuire, and David R. Brownfor their early help in developing the review; Reba A. Normanand Delle B. Kelley, Research Librarians; Kate W. Harris,Editor; and for their ongoing technical support and exper-tise, David Buchner, William H. Dietz, Jeffrey R. Harris, andPeter Briss. We are also indebted to the investigators whocreated the body of literature we examined for this review.

Our Consultation Team: Terry Bazzarre, PhD, RobertWood Johnson Foundation, Princeton, NJ; Carl J. Caspersen,PhD, National Center for Chronic Disease Prevention andHealth Promotion, CDC, Atlanta, GA; Diana Cassady, DrPH,California Department of Health Services, Sacramento; Car-los J. Crespo, DrPH, State University of New York School ofMedicine and Biomedical Sciences, Buffalo; Steve Hooker,PhD, California Department of Health Services, Sacramento;Jonathan Fielding, MD, MPH, MBA, University of CaliforniaLos Angeles School of Public Health; Barbara Fraser, RD, MS,Nebraska Department of Health and Human Services, Lin-

Am J Prev Med 2002;22(4S) 91

Table A-1. Candidate interventions

Proposed intervention category Selected examples (for illustration only)

Interventions aimed at changing the physicalenvironment to increase exercise opportunities

● Creation of safe, lighted walking paths● Creation of worksite and school fitness facilities● Creation of safer pedestrian environments● Provision of signs, other prompts at choice points for

physical activity

Organization-based policy interventions to increaseexercise opportunities

● School-based skills-oriented interventions● Classroom curricula● Physical education curricula

● Johnson & Johnson Live for Life

Community-wide policy interventions to increaseexercise opportunities

● Policies or legislation establishing financial incentives fororganizations and communities to provide access to exerciseopportunities

Community coalition-building, partnerships,community organizations to promote physicalactivity

Health education classes to change knowledge andattitudes about benefits of exercise, ways toincrease exercise

Provider training to change knowledge, attitudes,beliefs, and behavior about physical activityscreening and promotion among clinicalhealthcare providers

Community-leader training, e.g., “Train-the-trainers”

Special support mechanisms to facilitate andmaintain behavior change

Skills-oriented health or fitness classes to identifyhigh-risk situations and exercise barriers, developphysical activity and exercise skills, self-management and monitoring skills, and relapseprevention skills

Use of behavioral reinforcements

Home-based exercise programs

Structured/supervised exercise programs

Public information and social marketing campaignsto change knowledge, attitudes, and beliefsabout benefits, opportunities for exercise;change social norms about desirability or needfor exercise; or create demand for increasedopportunities for exercise

● Planned Approach to Community Health (PATCH)● Bootheel Project (State of Missouri)

● School-based knowledge-oriented curriculum● Adult health education classes offered by health

maintenance organizations

● Physician-based Assessment and Counseling for Exercise(PACE)

● Physically Active for Life (PAL)

● Community Health Advisors● Physical Activity Risk Reduction (PARR)● Zuni study

● Telephone support, counseling● Physical activity and exercise clubs (e.g., walking, biking

clubs)● Designation of walking partners● Family-based programs● School-based social support● Faith-based social support

● Worksite-based physical activity courses● Healthcare-site health promotion classes

● Use of contingencies, incentives to change exercise andphysical activity behavior

● Contracting and/or goal-setting to develop intentions toexercise

● Reminder systems for exercise and physical activity

● Videotapes● Exercise “prescription” for home

● Hospital and cardiac rehabilitation programs● Health spa—proprietary programs

● Mass media campaigns● National health initiatives


Logic Framework

To develop the logic framework, the coordination team firstillustrated the relationships between physical activity, physicalfitness, morbidity, and mortality. For example, physical activ-ity has been shown to improve measures of fitness such asaerobic capacity, muscular strength and endurance, bodycomposition, agility, and coordination.4 Physical activity alsoimproves metabolic functioning, exemplified by improve-ments in bone density, lipid profiles, insulin levels, andimmune function. Regular physical activity is also associatedwith improved health and quality of life and a reduced risk forall-cause mortality.5–8 Those who are physically active have areduced risk of developing cardiovascular disease,9–12 non-insulin-dependent (type 2) diabetes,13–15 colon cancers,16– 19

osteoporosis,20–22 depression,23–25 and fall-related inju-ries.26–29

These relationships between increased physical activity andhealth were assumed by the team to already be well estab-lished and were not the focus of the systematic review.Instead, the team focused on whether particular interventionsincreased physical activity behaviors. The logic frameworkalso shows the means by which interventions are thought tobe effective. Modifiable determinants of behavior aregrouped into three categories: (1) information-based deter-minants such as knowledge and attitudes about physicalactivity and behaviors that precede physical activity, motiva-tion to be active, or intentions to engage in activity; (2) be-havioral and social skills that facilitate the adoption andmaintenance of behavioral change; and (3) characteristics ofthe physical environment that increase the possibility andlikelihood of physical activity occurring, such as safe andaccessible parks and recreation facilities.

Candidate Interventions

Using the logic framework, the coordination team initiallydeveloped three categories of interventions:

● Individual approaches: behavior modification, social learn-ing, and individually-tailored programs.

● Setting-specific interventions: school-based, worksite-based,health facility-based, and faith-based.

● Community-wide interventions: community organizing, co-alitions, and partnerships; mass media and social marketingapproaches; environmental change approaches; legislativeaction, policy change, and incentives.

From these three categories, the team generated a compre-hensive list of candidate interventions for inclusion thataddressed each of the modifiable determinants (i.e., individ-ual level factors, social and behavioral factors, and physicalenvironment). This list was put in priority order for reviewthrough a process of polling the coordination and consulta-tion teams and other specialists in the field. Factors forpriority setting were the perceived public health importance(e.g., the number of people affected and likely magnitude ofintervention effect), the practicality of application, and theneed of those promoting physical activity to have informationon each intervention. Interventions reviewed were eithersingle-component (i.e., using only one activity to achievedesired outcomes) or multicomponent (i.e., using more than

one related activity). The specific interventions chosen forreview are described in the main body of this article.

Selection of Interventions for Review

In this review, the coordination team decided to excludeinterventions that use physical activity solely as a therapeuticintervention or address a population group because themembers share a clinical condition. For example, we ex-cluded studies that examined the effect of exercise on reha-bilitation from myocardial infarction or stroke as well asstudies that focus on exercise as a treatment for conditionssuch as arthritis or depression. We also excluded trainingstudies, that is, efficacy studies of physical activity on healthparameters. These decisions were made for several reasons.Rehabilitative studies were excluded because the team wantedto focus on interventions aimed at changing the behavior ofthe general population; it was felt that populations withclinical conditions or those receiving therapy would havespecial motivating factors that might positively influence theapparent effectiveness of a given strategy. Training studieswere excluded because the health benefits of physical activityare already well established.4

Selection of Outcomes for Review

Many of the studies included in our body of evidence targetedbehavioral risk factors in addition to physical inactivity, mostoften poor diet and tobacco use. Many physiologic benefits ofincreased physical activity can also result from improvementsin these other behaviors.

The team’s primary outcome of interest was physical activ-ity behavior. We also used aerobic capacity, defined as themaximum amount of oxygen that can be transported fromthe lungs to the tissues during exercise.30 We used aerobiccapacity for two reasons: first, physical activity is difficult tomeasure, especially among children, making an alternativemeasure desirable. Second, changes in aerobic capacity andchanges in physical activity are inextricably linked, since theonly way to increase aerobic capacity is to increase physicalactivity. Therefore, an increase in aerobic capacity can beused as an indicator of (or marker for) increased physicalactivity.

Aerobic capacity is measured with greater reliability andvalidity than is behavior. Therefore, in most cases, measuresof aerobic capacity were considered superior to measures ofphysical activity if both were available for the same study. Theteam developed an algorithm to guide the Task Force whenreviewing conflicting results in behavioral (physical activity)and physiologic (aerobic capacity) measures (Table A-2).

Search Strategy

The reviews of interventions to increase physical activityreflect systematic searches of seven computerized databases(MEDLINE, Sportdiscus, Psychlnfo, Transportation ResearchInformation Services [TRIS], Enviroline, Sociological Ab-stracts, and Social SciSearch) as well as reviews of referencelists and consultations with experts in the field. These yielded6238 titles and abstracts for review. Studies were eligible forinclusion if they:

● were published in English during 1980-2000;

Am J Prev Med 2002;22(4S) 97

● were conducted in an Established Market Economy;a

● assessed a behavioral intervention primarily focused onphysical activity;

● were primary investigations of interventions selected forevaluation rather than, for example, guidelines or reviews;

● evaluated outcomes selected for review; and● compared outcomes among groups of persons exposed to

the intervention with outcomes among groups of personsnot exposed or less exposed to the intervention (whetherthe study design included a concurrent or before-and-aftercomparison).

After review of the abstracts and consultation with specialistsin the field, a total of 849 reports was retrieved. Of these, 253were retained for full review. On the basis of limitations inexecution or design or because they provided no additionalinformation on studies that were already included, 159 ofthese were excluded and were not considered further. Theremaining 94 studies were considered qualifying studies.

Individual studies were grouped together on the basis ofthe similarity of the interventions being evaluated, and ana-lyzed as a group. Some studies provided evidence for morethan one intervention. In these cases, the studies were dividedinto arms and reviewed for each applicable intervention,population, or outcome measure. Interventions and outcomemeasures were classified according to definitions developedas part of the review process. The classification and nomen-clature used in our systematic reviews sometimes differs fromthat used in the original studies.

Evaluating and Summarizing the Studies

Each study that met the inclusion criteria was evaluated usinga standardized abstraction form and was assessed for suitabil-ity of the study design and threats to validity.1,31 On the basisof the number of threats to validity, studies were character-ized as having good, fair, or limited execution.1,31 Studieswith limited execution were not included in the summary ofthe effect of the intervention. The remaining studies (i.e.,those with good or fair execution) were considered qualifyingstudies. Estimates of effectiveness are based on those studies.

In this review, effect sizes were calculated as the net percentchange from baseline. This was done by one of three different

methods, depending on study design. Our preferred measurewas calculated from studies with the greatest suitability ofdesign, that is, randomized or non-randomized clinical orcommunity trials, multiple measurement before-and-after de-signs with concurrent comparison groups, and prospectivecohort studies. These present both a control (C) group andmeasurements made before and after an intervention (I).The intervention effect was calculated according to thefollowing formula:

Ipost2Ipre

Ipre

2

Cpost2Cpre

Cpre

When studies did not include a control group, we assumedthat in the absence of an intervention, no change would haveoccurred; that is, Cpost 2 Cpre 5 0, and we calculated the netintervention effect using measurements from the interven-tion group:

Ipost2Ipre

Ipre

When studies had a control group but no baseline measure-ments, we assumed that the intervention and comparisongroups were equivalent at baseline, that is, Ipre 5 Cpre . Thenext intervention effect was calculated as:

Ipost2Cpost

Cpost

Net intervention effects were calculated for all reportedmeasurements of a given outcome. Often, different variableswere used within a study to assess changes affecting the sameoutcome (e.g., changes in physical activity might be calcu-lated by measuring times per week in physical activity, self-reported physical activity score, minutes per week in physicalactivity, or all three). Multiple measurements of the sameoutcome were examined for consistency. Medians were cal-culated as summary effect measures for each type of measure-ment and were compared across outcomes for consistency.

Recommendations were based on behavioral measure-ments or measurements of aerobic capacity or both, accord-ing to the algorithm outlined in Table A-2.

Bodies of evidence of effectiveness were characterized asstrong, sufficient, or insufficient on the basis of the number ofavailable studies, the suitability of study designs for evaluatingeffectiveness, the quality of execution of the studies, theconsistency of the results, and the effect size.1

Other Effects

The Community Guide’s systematic reviews of interventions toincrease physical activity also sought information on other

aEstablished Market Economies as defined by the World Bank areAndorra, Australia, Austria, Belgium, Bermuda, Canada, ChannelIslands, Denmark, Faeroe Islands, Finland, France, Germany,Gibraltar, Greece, Greenland, Holy See, Iceland, Ireland, Isle of Man,Italy, Japan, Liechtenstein, Luxembourg, Monaco, the Netherlands,New Zealand, Norway, Portugal, San Marino, Spain, St. Pierre andMiquelon, Sweden, Switzerland, the United Kingdom, and theUnited States.

Table A-2. Algorithm to aid in making recommendations when there are conflicting results in the two outcomes measured

Aerobic capacity(VO2 max)

Physical activity behavior

Increase No change Decrease

Increase Recommend for Recommend for Recommend fora

No change Recommend forb Insufficient evidence Recommend againstDecrease Insufficient evidencea Insufficient evidence Recommend againstaBodies of evidence (or studies) that fall into these categories should be examined to look at issues related to measurement error.bThis result may occur, depending on dose (frequency, intensity, and duration) of physical activity, duration of the intervention, or the initiallevel of fitness among the population being studied.


effects (i.e., positive and negative health or nonhealth ‘sideeffects’). In addition to physical activity and aerobic capacityoutcomes, we collected information about adiposity andmeasures of physical fitness, including flexibility, muscularstrength, agility, balance, and coordination. We also collectedinformation about changes in knowledge, attitudes, andintentions, which are outcomes that may precede and thusaffect physical activity. Information on other effects was onlyderived from studies that measured physical activity andaerobic capacity, as having information about these variableswas one of the inclusion criteria.

Evidence of potential harms of these interventions wassought if they were mentioned in the effectiveness literatureor if the team thought they were important. For example, inthe reviews of school-based physical education, the teamspecifically sought information about the effect of the inter-ventions reviewed on academic performance, because ofstated concerns about the potential negative effects of takingtime away from academic subjects.

Evaluating Economic Efficiency

For all interventions that were either recommended orstrongly recommended by the Task Force, the team con-ducted systematic reviews of the evidence of economic effi-ciency. These reviews are provided to help decision makerschoose among recommended interventions on the basis ofcomparing costs with intended outcomes.

The general methods for conducting systematic reviews ofeconomic efficiency have been previously reported2 and aresummarized here as they were adapted for the review ofinterventions to increase physical activity. The four basic stepsare:

● searching for and retrieving evidence;● abstracting and adjusting the economic data;● assessing the quality of the identified economic evidence;

and● summarizing and interpreting the evidence of economic

efficiency.

Searching for and retrieving economic evidence

The databases MEDLINE, Transportation Research Informa-tion Services (TRIS), Combined Health Information Data-base (CHID), ECONLIT, PsychInfo, Sociological Abstracts,Sociofile, Social SciSearch, and Enviroline were searched forthe period 1980–2000. In addition, the references listed in allretrieved articles were reviewed and experts were consulted.Most of the included studies were either government reportsor were published in journals. To be included in the review astudy had to:

● be a primary study rather than, for example, a guideline orreview;

● take place in an Established Market Economy;a

● be written in English;● meet the team’s definitions of the recommended and

strongly recommended interventions;● use economic analytical methods such as cost analysis,

cost-effectiveness analysis, cost-utility, or cost-benefit analy-sis; and

● itemize program costs and costs of illness or injury averted.

Abstraction and adjustment of economic dataTwo reviewers read each study meeting the inclusion criteria.Any disagreements between the reviewers were reconciled byconsensus of the team members. A standardized abstractionform (available at www.thecommunityguide.org) was used forabstracting data. For those studies conducting cost-effective-ness and cost-utility analysis, results were adjusted to approx-imate the analysis to the reference case suggested by thePanel on Cost-effectiveness in Health and Medicine.32 Resultsfrom cost-benefit analyses were adjusted for currency (to U.S.dollars) and base-year (to 1997 dollars) only. When feasible,results were recalculated if the discount rate used in the studywas other than 3%.

Assessing the quality of the evidenceQuality of study design and execution was systematicallyassessed across five categories: study design, cost data, out-come measure, effects, and analysis. By subtracting points foreach limitation from a perfect score of 100, study quality wascharacterized as very good (90–100), good (80–89), satisfac-tory (60–79), or unsatisfactory (less than 60). Results fromunsatisfactory studies were not presented.

Summarizing the body of evidenceThe findings regarding the economic efficiency of interven-tions are presented in summary tables. The summary tablesinclude information on seven aspects of each included study.Table A-3 is an example of a summary table.

Ratios or net present values (i.e., the summary measures foruse in cost-effectiveness or cost-utility analyses and cost-benefit analysis, respectively) are pooled in ranges in thosecases in which the intervention definition, population at risk,and comparator match across studies.

Barriers

Information regarding barriers to implementation of theinterventions was abstracted from reviewed studies, evaluatedon the suggestion of the team, or both. Information onbarriers did not affect the Task Force recommendations, butis provided to assist readers contemplating implementation ofthe interventions.

Translating Strength of Evidence intoRecommendations

The Task Force recommendations presented in the accom-panying article3 are based on the evidence gleaned from thesystematic reviews conducted in accordance with the methodspresented here. The strength of each recommendation isbased on the strength of the evidence of effectiveness (e.g., anintervention is strongly recommended when there is strongevidence of effectiveness, and recommended when there issufficient evidence).1 Other types of evidence can also affecta recommendation. For example, evidence of harms resultingfrom an intervention might lead to recommendation that theintervention not be used if adverse effects outweigh improvedoutcomes. In general, the Task Force does not use economicinformation to modify recommendations.

A finding of insufficient evidence of effectiveness shouldnot be seen as evidence of ineffectiveness, but rather reflectsthe fact that the systematic review did not identify enoughinformation for the Task Force to make a recommendation.

Am J Prev Med 2002;22(4S) 99

Tab

leA

-3.

Sam

ple

of

eco

no

mic

sum

mar

yta

ble

–A

uth

ors

–A

uth

ors

’affi

liati

on

–F

un

din

gso

urc

e–P

ub

lica

tio

nD

ate

–S

tud

yP

eri

od

–A

naly

tic

meth

od

–R

ep

ort

ed

or

calc

ula

ted

sum

mary

measu

re

–S

tud

ylo

cati

on

–S

ett

ing

typ

e–P

op

ula

tio

nd

esc

rip

tio

n–F

oll

ow

-up

peri

od

–In

terv

en

tio

ns

stu

die

d–C

om

pari

son

s

–R

ep

ort

ed

curr

en

cyan

db

ase

year

–C

ost

sin

clu

ded

–R

ep

ort

ed

sum

mary

measu

re–R

ep

ort

ed

eff

ect

size

–A

dju

sted

curr

en

cyan

db

ase

year

–A

dju

sted

valu

esu

mm

ary

measu

re–N

ote

s

–Q

uali

tyca

tego

ry–Q

uali

tysc

ore

–N

ote

s

Ind

ivid

uall

y-ad

ap

ted

Healt

hB

eh

avi

or

Ch

an

ge

Pro

gra

m

–Sev

ick

MA

,D

un

nA

L,

Mo

rro

wM

S,

Mar

cus

BH

,C

hen

GJ,

Bla

irSN

–Wak

eF

ore

stU

niv

ersi

ty;

Co

op

erIn

stit

ute

for

Aer

ob

ics

Res

earc

h–N

atio

nal

Inst

itu

tes

of

Hea

lth

,N

ord

icT

rack

,Sta

irm

aste

r,C

ybex

,P

reC

or

and

Yam

axC

orp

ora

tio

n.

–200

0–2

year

s

–Co

st-e

ffec

tive

nes

san

alys

is–P

rogra

mco

stp

erav

erag

eu

nit

of

imp

rove

men

tfo

rli

fest

yle

inte

rven

tio

n(C

1)

and

stru

ctu

red

inte

rven

tio

n(C

2)

–Dal

las,

Tex

as–F

itn

ess

faci

lity

–Ad

ult

sag

ed35–6

0ye

ars

–No

foll

ow

-up

–Cen

ter-

bas

edli

fest

yle

exer

cise

con

sist

ing

of

beh

avio

ral

skil

ls,

and

stru

ctu

red

exer

cise

con

sist

ing

of

sup

ervi

sed

cen

ter-

bas

edex

erci

se–B

asel

ine

–1994

U.S

.d

oll

ars

–Co

sts

incl

ud

edp

erso

nn

el,

cap

ital

equ

ipm

ent,

faci

liti

esan

dgen

eral

sup

pli

es–P

rogra

mco

stp

erav

erag

eu

nit

of

imp

rove

men

ta

C1:

$0.0

4–$

3.6

3/

aver

age

un

ito

fim

pro

vem

ent

C2:

$0.0

6–$

4.9

8/

aver

age

un

ito

fim

pro

vem

ent

–Eff

ect

size

incl

ud

edd

iffe

ren

tm

easu

res

of

ph

ysic

alac

tivi

tyan

dca

rdio

resp

irat

ory

fitn

ess

–1997

U.S

.d

oll

ars

–Pro

gra

mco

stp

erav

erag

eu

nit

of

imp

rove

men

tC

1:

$0.0

5–$

3.9

3/

aver

age

un

ito

fim

pro

vem

ent

C2:

$0.0

7–$

5.3

9/

aver

age

un

ito

fim

pro

vem

ent

–Go

od

–88

–Stu

dy

did

no

tu

seso

ciet

alp

ersp

ecti

ve,

did

no

td

efin

eth

ean

alyt

ich

ori

zon

,an

dd

idn

ot

rep

ort

bas

e-ye

arfo

rre

sou

rce

pri

ces

aT

he

ran

ge

valu

esar

efo

rse

vera

lo

utc

om

em

easu

res.


Further, it is important for identifying areas of uncertaintythat require additional research. In contrast, sufficient orstrong evidence of ineffectiveness leads to a recommendationthat the intervention not be used.

Summarizing Research Gaps

Systematic reviews in the Community Guide identify existinginformation on which to base public health decisions aboutimplementation of interventions. An important additionalbenefit of these reviews is the identification of areas in whichinformation is lacking or of poor quality. To summarize thesegaps, remaining research questions for each interventionevaluated were identified. Where evidence of effectiveness ofan intervention was sufficient or strong, remaining questionsregarding effectiveness, applicability, other effects, economicconsequences, and barriers were summarized.

Where evidence of effectiveness of an intervention wasinsufficient, remaining questions regarding only effectivenessand other effects were summarized. Applicability issues weresummarized only if they affected the assessment of effective-ness. The team decided that it would be premature to identifyresearch gaps in economic evaluations or barriers beforeeffectiveness was demonstrated.

For each category of evidence, issues that had emergedfrom the review were identified, based on the informedjudgment of the team. Several factors influenced that judg-ment. When a conclusion was drawn about evidence, theteam decided if additional issues remained.

● If effectiveness was demonstrated using some but not alloutcomes, all other possible outcomes were not necessarilylisted as research gaps.

● If the available evidence was thought to be generalizable, allsubpopulations or settings where studies had not beendone were not necessarily identified as research gaps.

● Within each body of evidence, the team consideredwhether there were general methodologic issues that wouldimprove future studies in that area.

The Reviews of Evidence

This appendix describes the general methodologic approachused in the systematic reviews of interventions to increasephysical activity. The main text of this article presents thesupporting evidence on which the Task Force based itsrecommendations about these interventions. It describes thescope and extent of the problem studied, discusses theconceptual approach to the review of evidence for theinterventions studied, and presents additional informationabout methodology specific to the review of those interven-tions, in addition to giving a detailed report on the findingsfor each intervention.

References1. Briss PA, Zaza S, Pappaioanou M, et al. Developing an evidence-based

Guide to Community Preventive Services—methods. The Task Force on

Community Preventive Services. Am J Prev Med 2000;18 (suppl 1):35–

43.

2. Carande-Kulis VG, Maciosek MV, Briss PA, et al. Methods for systematic

reviews of economic evaluations for the Guide to Community Preventive

Services. Task Force on Community Preventive Services. Am J Prev Med

2000;18 (suppl 1):75–91.

3. Task Force on Community Preventive Services. Recommendations to

increase physical activity in communities. Am J Prev Med 2002;22(suppl

4):67–72.

4. U.S. Department of Health and Human Services. Physical activity and

health: a report of the Surgeon General. Atlanta, GA: U.S. Department of

Health and Human Services, Centers for Disease Control and Prevention,

National Center for Chronic Disease Prevention and Health Promotion,

1996.

5. Lee IM, Hsieh CC, Paffenbarger RS Jr. Exercise intensity and longevity in

men. The Harvard Alumni Health Study. JAMA 1995;273:1179–84.

6. Paffenbarger RS Jr, Hyde RT, Wing AL, Lee IM, Jung DL, Kampert JB.

The association of changes in physical-activity level and other lifestyle

characteristics with mortality among men. N Engl J Med 1993;328:538 –

45.

7. Paffenbarger RS Jr, Kampert JB, Lee IM, Hyde RT, Leung RW, Wing AL.

Changes in physical activity and other lifeway patterns influencing longev-

ity. Med Sci Sports Exerc 1994;26:857–65.

8. Blair SN, Kohl HW III, Barlow CE, Paffenbarger RS Jr, Gibbons LW, Macera

CA. Changes in physical fitness and all-cause mortality. A prospective study

of healthy and unhealthy men. JAMA 1995;273:1093–8.

9. Paffenbarger RS Jr, Hyde RT, Wing AL, Steinmetz CH. A natural history of

athleticism and cardiovascular health. JAMA 1984;252:491–5.

10. Kannel WB, Sorlie P. Some health benefits of physical activity. The

Framingham Study. Arch Intern Med 1979;139:857–61.

11. Kannel WB, Belanger A, D’Agostino R, Israel I. Physical activity and

physical demand on the job and risk of cardiovascular disease and death:

the Framingham Study. Am Heart J 1986;112:820–5.

12. LaCroix AZ, Leveille SG, Hecht JA, Grothaus LC, Wagner EH. Does

walking decrease the risk of cardiovascular disease hospitalizations and

death in older adults? J Am Geriatr Soc 1996;44:113–20.

13. Helmrich SP, Ragland DR, Leung RW, Paffenbarger RS Jr. Physical activity

and reduced occurrence of non-insulin-dependent diabetes mellitus.

N Engl J Med 1991;325:147–52.

14. Kaye SA, Folsom AR, Sprafka JM, Prineas RJ, Wallace RB. Increased

incidence of diabetes mellitus in relation to abdominal adiposity in older

women. J Clin Epidemiol 1991;44:329–34.

15. Uusitupa M, Siitonen O, Pyorala K, et al. The relationship of cardiovas-

cular risk factors to the prevalence of coronary heart disease in newly

diagnosed type 2 (non-insulin-dependent) diabetes. Diabetologia 1985;

28:653–9.

16. Brownson RC, Zahm SH, Chang JC, Blair A. Occupational risk of colon

cancer. An analysis by anatomic subsite. Am J Epidemiol 1989;130:675–

87.

17. Brownson RC, Chang JC, Davis JR, Smith CA. Physical activity on the job

and cancer in Missouri. Am J Public Health 1991;81:639–42.

18. Dosemeci M, Hayes RB, Vetter R, et al. Occupational physical activity,

socioeconomic status, and risks of 15 cancer sites in Turkey. Cancer Causes

Control 1993;4:313–21.

19. Giovannucci E, Ascherio A, Rimm EB, Colditz GA, Stampfer MJ, Willett

WC. Physical activity, obesity, and risk for colon cancer and adenoma in

men. Ann Intern Med 1995;122:327–34.

20. Nichols DL, Sanborn CF, Bonnick SL, Ben Ezra V, Gench B, DiMarco NM.

The effects of gymnastics training on bone mineral density. Med Sci Sports

Exerc 1994;26:1220–5.

21. Rubin K, Schirduan V, Gendreau P, Sarfarazi M, Mendola R, Dalsky G.

Predictors of axial and peripheral bone mineral density in healthy children

and adolescents, with special attention to the role of puberty. J Pediatr

1993;123:863–70.

22. Kohrt WM, Snead DB, Slatopolsky E, Birge SJ Jr. Additive effects of

weight-bearing exercise and estrogen on bone mineral density in older

women. J Bone Miner Res 1995;10:1303–11.

23. Ross CE, Hayes D. Exercise and psychologic well-being in the community.

Am J Epidemiol 1988;127:762–71.

24. Camacho TC, Roberts RE, Lazarus NB, Kaplan GA, Cohen RD. Physical

activity and depression: evidence from the Alameda County Study. Am J

Epidemiol 1991;134:220–31.

25. Weyerer S. Physical inactivity and depression in the community. Evi-

dence from the Upper Bavarian Field Study. Int J Sports Med 1992;13:

492– 6.

26. Farmer ME, Harris T, Madans JH, Wallace RB, Cornoni-Huntley J, White

LR. Anthropometric indicators and hip fracture. The NHANES I epidemi-

ologic follow-up study. J Am Geriatr Soc 1989;37:9–16.

Am J Prev Med 2002;22(4S) 101

27. Meyer HE, Tverdal A, Falch JA. Risk factors for hip fracture in middle-aged

Norwegian women and men. Am J Epidemiol 1993;137:1203–11.

28. Cummings SR, Nevitt MC, Browner WS, et al. Risk factors for hip fracture

in white women. Study of Osteoporotic Fractures Research Group. N Engl

J Med 1995;332:767–73.

29. Jaglal SB, Kreiger N, Darlington G. Past and recent physical activity and risk

of hip fracture. Am J Epidemiol 1993;138:107–18.

30. U.S. Department of Health and Human Services, Public Health Service,

Centers for Disease Control and Prevention, National Center for Chronic

Disease Prevention and Health Promotion, Division of Nutrition and

Physical Activity. Promoting physical activity: a guide to community action.

Champaign, IL: Human Kinetics, 1999.

31. Zaza S, Wright-De Aguero LK, Briss PA, et al. Data collection instrument

and procedure for systematic reviews in the Guide to Community Preven-

tive Services. Task Force on Community Preventive Services. Am J Prev Med

2000;18(suppl 1):44–74.

32. Gold MR, Siegel JE, Russell LB, Weinstein MC. Cost-effectiveness in health

and medicine. New York: Oxford University Press, 1996.


Appendix

B.

Stu

die

sM

easu

ring

the

Eff

ecti

veness

of

School-

base

dP

hys

ical

Educati

on

(PE)

Am J Prev Med 2002;22(4S) 103


Am J Prev Med 2002;22(4S) 105


Am J Prev Med 2002;22(4S) 107

Reprinted by permission of Elsevier Science from:The effectiveness of interventions to increase physical activity. Kahn EB, Ramsey LT,

Brownson R, Heath GW, Howze EH, Powell KE, Stone EJ, Rajab MW, Corso P, Task Force on Community Preventive Services., American Journal of Prevention Medicine. Vol 22 No 4S, pp 73-107.

Documents

Interventions to increase physical activity · 2010.32 The two main foci of the Healthy People preven-tion objectives are to increase (1) the amount of moderate or vigorous physical