REVIEW CLINICIAN’S CORNER

Using Pedometers to Increase Physical Activity

and Improve Health
A Systematic Review
Dena M. Bravata, MD, MS Context Without detailed evidence of their effectiveness, pedometers have re-
Crystal Smith-Spangler, MD cently become popular as a tool for motivating physical activity.
Vandana Sundaram, MPH Objective To evaluate the association of pedometer use with physical activity and
Allison L. Gienger, BA health outcomes among outpatient adults.

Nancy Lin, ScD Data Sources English-language articles from MEDLINE, EMBASE, Sport Discus,
PsychINFO, Cochrane Library, Thompson Scientific (formerly known as Thompson ISI),
Robyn Lewis, MA and ERIC (1966-2007); bibliographies of retrieved articles; and conference proceedings.
Christopher D. Stave, MLS Study Selection Studies were eligible for inclusion if they reported an assessment
Ingram Olkin, PhD of pedometer use among adult outpatients, reported a change in steps per day, and
included more than 5 participants.
John R. Sirard, PhD
Data Extraction and Data Synthesis Two investigators independently ab-

I
NCREASED PHYSICAL ACTIVITY IS AS- stracted data about the intervention; participants; number of steps per day; and pres-
sociated with improvements in nu- ence or absence of obesity, diabetes, hypertension, or hyperlipidemia. Data were pooled
merous health conditions, includ- using random-effects calculations, and meta-regression was performed.
ing coronary artery disease, Results Our searches identified 2246 citations; 26 studies with a total of 2767 par-
hypertension, stroke, insulin sensitiv- ticipants met inclusion criteria (8 randomized controlled trials [RCTs] and 18 observa-
ity, osteoporosis, and depression.1-4 Be- tional studies). The participants’ mean (SD) age was 49 (9) years and 85% were women.
cause of these extensive health ben- The mean intervention duration was 18 weeks. In the RCTs, pedometer users signifi-
efits, the Department of Health and cantly increased their physical activity by 2491 steps per day more than control par-
ticipants (95% confidence interval [CI], 1098-3885 steps per day, P⬍.001). Among
Human Services recommends “physi-
the observational studies, pedometer users significantly increased their physical activ-
cal activity most days of the week for ity by 2183 steps per day over baseline (95% CI, 1571-2796 steps per day, P⬍.0001).
at least 30 minutes for adults.”5 De- Overall, pedometer users increased their physical activity by 26.9% over baseline. An
spite these recommendations and the important predictor of increased physical activity was having a step goal such as 10 000
well-documented evidence that physi- steps per day (P=.001). When data from all studies were combined, pedometer users
cal activity is beneficial, more than half significantly decreased their body mass index by 0.38 (95% CI, 0.05-0.72; P=.03).
of all adults in the United States do not This decrease was associated with older age (P=.001) and having a step goal (P=.04).
get adequate physical activity and ap- Intervention participants significantly decreased their systolic blood pressure by 3.8
proximately one quarter do not get any mm Hg (95% CI, 1.7-5.9 mm Hg, P⬍.001). This decrease was associated with greater
baseline systolic blood pressure (P=.009) and change in steps per day (P=.08).
leisure time physical activity.6
The costs associated with physical Conclusions The results suggest that the use of a pedometer is associated with sig-
inactivity are high. For example, if nificant increases in physical activity and significant decreases in body mass index and
blood pressure. Whether these changes are durable over the long term is undetermined.
10% of adults in the United States
JAMA. 2007;298(19):2296-2304 www.jama.com
began a regular walking program, an
estimated $5.6 billion in heart dis-
Author Affiliations: Center for Primary Care and Out- (Dr Bravata); Veterans Affairs Palo Alto Health Care
ease costs could be saved.6 Pedom- comes Research (Drs Bravata and Lin and Mss System, Palo Alto, California (Ms Sundaram); and
eters are small, relatively inexpensive Sundaram, Gienger, and Lewis), Department of In- School of Public Health, University of Minnesota, Min-
ternal Medicine (Dr Smith-Spangler), Stanford Uni- neapolis (Dr Sirard).
versity School of Medicine, Lane Medical Library (Mr Corresponding Author: Dena M. Bravata, MD, MS, Pri-
CME available online at Stave), and Department of Statistics (Dr Olkin), Stan- mary Care and Outcomes Research, 117 Encina Com-
www.jama.com ford University, Stanford; Department of Internal Medi- mons, Stanford, CA 94305-6019 (dbravata@stanford
cine, California Pacific Medical Center, San Francisco .edu).

2296 JAMA, November 21, 2007—Vol 298, No. 19 (Reprinted) ©2007 American Medical Association. All rights reserved.

Downloaded From: by a University of the Sciences User on 02/28/2018

 PEDOMETERS AND INCREASED PHYSICAL ACTIVITY

devices worn at the hip to count the change in number of steps walked per Data Synthesis
number of steps walked per day. day. We excluded studies that required For each of the included studies, we cal-
Although there is not detailed evi- participants to be hospitalized or con- culated 2 effect sizes for each of the out-
dence of their effectiveness, they fined to a research center, sealed the pe- comes of interest: the mean difference
have recently experienced a surge in dometer so that intervention partici- (postintervention steps per day − pre-
popularity as a tool for motivating pants could not see the number of steps intervention steps per day) and stan-
and monitoring physical activity. 7 walked per day (often the control sub- dardized mean differences ([postinter-
Additionally, some guidelines specifi- jects wore sealed pedometers), or used vention steps per day − preintervention
cally recommend taking 10 000 steps a pedometer to measure the effects of a steps per day]/pooled standard devia-
per day.8 However, it is not known drug on an individual’s ability to be tion). The standardized mean differ-
whether encouraging adults to walk physically active. ence lacks units, which limits its inter-
10 000 steps per day is associated pretability, whereas the mean difference
with any significant improvement in Data Extraction retains its units, which facilitates clini-
health outcomes compared with not Two authors independently abstracted cal interpretation. For randomized
setting a goal or to setting an alterna- 4 categories of variables from each of the controlled trials (RCTs), we also cal-
tive activity goal. included studies: intervention variables culated the difference in the pre-
The primary purpose of this study (eg, intervention duration, whether intervention and the postintervention
was to evaluate the association be- counseling was included, and whether changes in outcomes between the in-
tween pedometer use and physical ac- participants were asked to achieve a par- tervention and control participants. Be-
tivity among adults in the outpatient ticular activity goal); participant vari- cause we found no significant differ-
setting. Additionally, we sought to de- ables (demographics; baseline activity; ences in summary results between these
termine the association between pe- and the presence or absence of obesity, 2 outcome metrics, we present only the
dometer use and changes in body diabetes, hyperlipidemia, or hyperten- mean differences. We calculated sum-
weight, serum lipid levels, fasting se- sion); outcome variables (number of mary outcomes using both random-
rum glucose and insulin, and blood steps per day, measures of body mass, effects and fixed-effects calculations and
pressure. Finally, we sought to evalu- glycemic control, serum lipid levels and found no significant differences be-
ate the association between setting a blood pressure); and quality variables tween the 2, thus present only the ran-
daily step goal and improvements in (method of blinding control partici- dom effects estimates.
health outcomes. pants to step counts, the extent to which Because the participant, physical
participants participated fully in the ac- activity, and outcome variables evalu-
METHODS tivity program, methods used to deter- ated are correlated, the corresponding
Data Sources and Search mine baseline physical activity, com- effect sizes for these measures are cor-
Strategies pleteness of follow-up and use of related. 9 We used meta-regression
In collaboration with a professional li- intention to treat analysis, the use of va- weighted by the sample size to calcu-
brarian, we developed individualized lidity- and reliability-tested pedom- late the summary effect of the phy-
search strategies for 7 databases: eters, and the extent to which cointer- sical activity and participant char-
MEDLINE (January 1966 to February ventions may have affected physical acteristic variables on the outcome
2007); and EMBASE, Sport Discus, activity). If a study reported both imme- variables.10
PsychINFO, Cochrane Library, Thomp- diate postintervention and longer-term We performed sensitivity analyses
son Scientific (formerly known as follow-up data, we used the immediate and assessed heterogeneity to evalu-
Thompson ISI), and ERIC ( January postintervention data in our primary ate the robustness of our results. We re-
1966 to May 2006). We used search analyses. moved each study individually to evalu-
terms such as pedometer, activity moni- We resolved discrepancies by re- ate that study’s effect on the summary
tor, and step counter. We also reviewed peated review and discussion between estimates. We assessed publication bias
the bibliographies of retrieved articles abstractors. If 2 or more studies pre- by visual inspection of funnel plots
and relevant conference proceedings and sented the same data from a single pa- comparing physical activity (x-axis) to
contacted experts in exercise physiol- tient population, we included these data sample size (y-axis) and calculated
ogy for additional studies. only once in our analyses. If a study pre- the fail-safe N (the number of missing
sented data on 2 types of activity pro- studies that would be required to
Study Selection grams and if 1 of the programs did not change a significant summary effect
We considered English-language stud- meet our inclusion criteria (eg, 1 pro- to one that was not statistically signifi-
ies eligible for inclusion if they re- gram without a pedometer), then we ab- cant).11 For each summary effect size,
ported an assessment of pedometer use stracted data for only those partici- we assessed statistical heterogeneity by
among adult outpatients, included more pants receiving the intervention that calculating the Q statistic (considered
than 5 participants, and reported a met our inclusion criteria. Q statistics with P ⬍ .05 as heteroge-
©2007 American Medical Association. All rights reserved. (Reprinted) JAMA, November 21, 2007—Vol 298, No. 19 2297

Downloaded From: by a University of the Sciences User on 02/28/2018

 PEDOMETERS AND INCREASED PHYSICAL ACTIVITY

ies that met inclusion criteria but re- included physical activity counseling
Figure 1. Study Flow Diagram
ported insufficient data to be included with a mean (SD) number of 7 (19)
2246 Potentially relevant articles
in our analyses—3 provided sufficient counseling sessions (range, 0-104 ses-
identified through literature search data to be included in our study.13-15 Af- sions). Only 3 studies included di-
2087 in MEDLINE, Cochrane,
ERIC, PsychINFO, and ter synthesizing the data from mul- etary counseling: 1 study prescribed a
Thompson Scientific tiple reports on the same set of partici- diet,15 and the other 2 gave advice on a
117 in SPORTDISCUS
41 in EMBASE pants, 26 studies met our inclusion healthful eating habits.26,36 Twenty stud-
1 in bibliographic search criteria (TABLE 1).13-39 ies were from the United States or
Canada, 2 were from Japan, 2 were from
1823 Excluded based on review of Study Characteristics Europe, and 2 were from Australia.
titles and abstracts
912 No pedometer use The designs of the included studies were
408 Not conducted in humans
highly heterogeneous. Eight of the in- Participant Characteristics
144 Cross-sectional studies
110 Review articles cluded studies were RCTs in which the The included studies evaluated 2767
86 Validity studies
22 Non-English language intervention participants wore pedom- participants of physical activity pro-
6 Pedometer not part of eters and were encouraged to view and grams (TABLE 2). Their mean (SD) age
intervention
5 Drug intervention trials record their daily step counts, whereas was 49 (9) years, and only 5 studies had
130 Other
the control participants wore pedom- participants whose mean age was more
eters that were sealed so that they could than 60 years. Nine studies exclu-
423 Full-text articles reviewed
not see their own step counts.14,16-23 sively enrolled women and overall, only
Six additional RCTs used pedometers 15% of the participants were men. Seven
396 Excluded
122 Cross-sectional studies
with visible step counts in both trial studies reported participants’ race/
76 No pedometer use cohorts, so we treated each of these co- ethnicity—the mean (SD) proportion
36 Validity studies
17 Drug intervention trials horts as separate observational stud- of white participants was 93% (7.5%).
11 Review articles
10 Pedometer not part of
ies.25,30-32,35,37 Twelve studies were single- Most participants were overweight, nor-
intervention group observational studies.* motensive, and had relatively well con-
6 Non-English language
1 Not conducted in humans
Overall, the quality of the reporting trolled serum lipid levels. Most partici-
117 Other of the included studies was relatively pants were relatively inactive at baseline
good. Only 4 studies did not specify the with a mean (SD) of 7473 (1385) steps
27 Articles (26 unique studies) included method by which participants’ base- per day (range, 2140-12 371 steps per
in analysis (2767 study participants)
8 RCTs line physical activity was determined day).
18 Observational studies including 6 (most asked participants not to change
RCTs treated as observational
studiesa their usual activity and to wear a sealed Pedometer Use
pedometer for 3 to 7 days prior to the and Physical Activity
CI indicates confidence interval; RCT, randomized con- start of the intervention to determine RCT Results. FIGURE 2 shows the dif-
trolled trial. a Six RCTs that used visible step counts in
both trial cohorts were each treated as separate ob- baseline activity). Nine studies had ference between the increase in physi-
servational studies. 100% of participants complete the in- cal activity among the participants ran-
tervention, and the average dropout rate domly assigned to pedometer use and
neous) and I2 statistic (considered I2 among the other studies was 20%—a control participants in the 8 RCTs.
statistics greater than 50% as heter- rate that is somewhat higher than the Figure 2 shows that the 155 interven-
ogeneous). 9,12 We considered and 4% to16% dropout rate reported by tion participants significantly in-
evaluated heterogeneity through pre- other physical activity interven- creased their physical activity by 2491
determined subgroup analyses (eg, de- tions.40 Sixteen studies used the Ya- steps per day more than the 122 con-
mographics, body mass index, which max pedometer (Yamax Corp, Tokyo, trol participants (95% confidence in-
is calculated as weight in kilograms di- Japan)—a model that has been well vali- terval [CI], 1098-3885 steps per day,
vided by height in meters squared), dated for accuracy and reliability and P⬍.001). However, this result was sta-
baseline activity, intervention type, in- is frequently used in physical activity tistically heterogeneous (Q = 74.9,
tervention setting, study design, etc. We research.41-44 P⬍.001; I2 =91). When we removed the
performed analyses using Comprehen- The physical activity interventions study by Moreau et al,20 a 24-week ex-
sive Meta-Analysis v.2 software (Bio- evaluated in the included studies var- ercise intervention involving postmeno-
stat, Englewood, New Jersey). ied considerably: mean (SD) duration pausal hypertensive women, which re-
was 18 (24) weeks (range, 3-104 ported a much higher increase in
RESULTS weeks), 5 took place in the work- physical activity than any of the other
Our searches identified 2246 poten- place, 23 included a step diary, and 17 trials, the summary increase in physi-
tially relevant articles (FIGURE 1). We cal activity among the remaining inter-
e-mailed the authors of 13 of the stud- *References 13, 15, 24, 26-29, 33, 36, 38, 39. vention participants was 2004 steps per
2298 JAMA, November 21, 2007—Vol 298, No. 19 (Reprinted) ©2007 American Medical Association. All rights reserved.

Downloaded From: by a University of the Sciences User on 02/28/2018

 PEDOMETERS AND INCREASED PHYSICAL ACTIVITY

Table 1. Study Characteristics: Randomized Controlled Trials and Observational Studies

Intervention No. of Intervention
Source Population Duration, wk Participants Intervention and Control Description Details
Randomized Controlled Trials
Araiza et al,16 Adults with diabetes 6 15 Intervention: given a pedometer; instructed to walk 10 000 steps per day, 5 Diary
2006 times/wk
15 Control: given a pedometer, but instructed to maintain normal activity
during intervention
Butler and Sedentary 4 17 Intervention: modified version of Just Walk It Program; given pedometers Other physical
Dwyer,17 normotensive and a daily goal (first 2 wk: 3000 steps/d or 30 min/d, following 2 wk: activity goal
2004 adults, aged 4000 steps/d or 40 min/d)
45-65 y
16 Control: wore a pedometer with an obscured monitor and asked to walk Other physical
30 min/d for the first 2 wk, 40 min/d for the following 2 wk activity goal
de Blok et al,18 Adults with COPD 9 10 Intervention: COPD rehabilitation (including exercise training and dietary Individualized
2006 intervention), given a pedometer and 4 lifestyle physical activity goals, diary
counseling and goal-setting sessions to increase steps/d
11 Control: received COPD rehabilitation only
Hultquist et Sedentary, 4 31 Intervention: directed to walk 10 000 steps/d, wearing pedometer for Diary
al,19 2005 nonsmoking feedback
women, aged
33-55 y 31 Control group: wore a sealed pedometer and asked to walk 30 min/d
on most days
Izawa et al,14 Adult MI patients 24 25 Intervention: received supervised cardiac rehabilitation and performed Goal not
2005 completing 6 pedometer self-monitoring specified,
mo of cardiac diary
rehabilitation
25 Control: received supervised cardiac rehabilitation
Moreau et al,20 Sedentary, 24 15 Intervention: given pedometer and step goal that would be equal to 3 km/d Individualized
2001 hypertensive, above baseline by end of intervention; wk 1, 1.4 km/d increase, then goals, diary
postmeno- increased by 0.5 km/d per wk for 3 wk until the 3 km target
pausal women was reached
9 Control: wore pedometer 1 wk/mo and asked to not make any changes to Diary
daily physical activity
Ransdell et Daughter, mother, 24 30 Intervention: home-based physical activity intervention with 2 classroom Other activity
al,21 2004; maternal sessions of 2 h; given home-based physical activity packet and monthly goal, diary
Ornes et grandmother reminder telephone calls; asked to exercise 3 times/wk with the triad and
al,22 2005 triads with ⱖ increase walking time and intensity in a step-wise fashion.
1 sedentary
member 21 Control: asked not to increase physical activites; attended a 2-h info session Diary
about using pedometer; recorded baseline and 6-mo steps over 3 d
Talbot et al,23 Adults ⬎60 y with 12 17 Intervention: received pedometer walking program (30% increase in baseline Individualized
2003 arthritis over 12 wk) and 12-h arthritis self-management program goals, diary
17 Control: attended a 12-h arthritis self-management course
Observational Studies
Chan et al,13 Sedentary adults 12 177 During the 4-wk adoption phase, met weekly for 30-60 min with program Individualized
2004 facilitators (RNs) to discuss fitness and set individual goals; given goals, diary,
pedometers to wear daily; recorded steps in a log or online. During the in workplace
adherence phase (8 weeks), monitored themselves and continued to set
their own step goals
Croteau,24 Adult volunteer 8 37 “Healthy Steps” program: 8-week physical activity intervention included a Other physical
2004 employees counseling session with goal setting and strategy selection, activity goal,
self-monitoring, and weekly e-mails. Step goal: if ⬍8000 at baseline, then diary, in
10% increase over baseline every 2 weeks until ⬎10 000/d; if workplace
8000-10 000, then increase by 5% every 2 weeks until ⬎10 000/d; if
baseline ⱖ10 000, then maintain
Eastep et al,25 Healthy adults 8 14 Group A: had visible pedometer counts for first 3 wk but was blinded for the Goal not
2004 a last 3; attended “Walking for Fitness” classes met 2 times/wk specified,
diary, in
12 Group B: blind to the pedometer first 3 wk, but allowed to see step counts workplace
for the last 3; attended “Walking for Fitness” classes met 2 times/wk
Jensen et al,26 Women ⱖ60 y with 12 26 Attended 8 counseling sessions (30 min each) with a dietitian and 1 physician Other physical
2004 BMI ⬎30 visit educating participant on diet, physical activity, and behavior activity goal,
modification; pedometer given to each participant with a goal of at least diary
5000 steps/d, advanced as indicated
Kilmer et al,15 Adults with 26 20 Dietary and activity intervention prescribed based on each participant’s Individualized
2005 neuromuscular baseline activity (25% increase in physical activity) and given during a goals, diary
disease 1-on-1 counseling session
Koulouri et Healthy adults 3 12 Instructed to increase step counts by 2000 steps/d; physical activity Individualized
al,27 2006 assessed using diary, active Pal activity monitor, and pedometer; food goals, diary
and drink intake recorded
Lindberg, 28 Adults 35-50 y, 8 92 Mail-based program using a pedometer, personal action planner, step log, 10 000-step
2000 contemplating biweekly motivational cards for 8 weeks, and bimonthly for the following goal, diary
becoming 6 mo
active
(continued)

©2007 American Medical Association. All rights reserved. (Reprinted) JAMA, November 21, 2007—Vol 298, No. 19 2299

Downloaded From: by a University of the Sciences User on 02/28/2018

 PEDOMETERS AND INCREASED PHYSICAL ACTIVITY

day more than the control partici- Overall, pedometer users increased Predictors of Improvements in Physi-
pants (95% CI, 878-3129 steps per day, their physical activity by 26.9% over cal Activity. We used meta-regression
P ⬍.001). baseline. We did not find evidence of to evaluate the participant and inter-
Observational Study Results. Among significant publication bias (eg, fail- vention characteristics associated with
the observational studies, the pedom- safe N was 127). However, this result increased physical activity among pe-
eter users significantly increased their was statistically heterogeneous (Q=212, dometer users in RCTs and observa-
physical activity by 2183 steps per day P ⬍.001; I2 =89), which is not surpris- tional studies. Among the participant
over baseline (95% CI, 1571-2796 steps ing given the differences in the physi- characteristics, there was a trend for
per day, P⬍.001). cal activity interventions. studies of younger pedometer users and

Table 1. Study Characteristics: Randomized Controlled Trials and Observational Studies (cont)
Intervention No. of Intervention
Source Population Duration, wk Participants Intervention and Control Description Details
Observational Studies
Schneider Obese and 36 56 Given a physical activity prescription that built up to 10 000 steps/d by week 10 000-step goal,
et al,29 overweight 4, asked to attend biweekly information sessions monthly for 2 mo diary
2006 sedentary adults afterward
Sidman Sedentary women, 3 45 Group A: asked to walk 10 000 steps/d 10 000 steps or
et al,30 20-65 y individualized
2004 a goals, diary
47 Group B: asked to increase walking by 1000 to 3000 steps/d over baseline Diary,
individualized
goals
Stovitz Patients presenting 9 50 Group A: brief physician endorsement of physical activity; given brochure, Individualized
et al,31 at a family pedometer, and 3 follow-up telephone calls from a health educator; goals, diary
2005 a,b medicine clinic each week, told to increase steps/d by 400 (approximately 10% above
baseline)

44 Group B: same as intervention, but not given pedometer Diary

Sugiura Sedentary, 104 32 Group A: participated in a 90-min weekly exercise class and wore a Individualized
et al,32 menopausal pedometer daily with goal setting and a daily log goals
2002 a women
16 Group B: wore pedometer only; goal to increase steps by 2000-3000 Other goals
⬎baseline
Swartz et al,33 Overweight and 12 19 0-4 wks: wore pedometers; 4-12 weeks, wore pedometers with a goal of 10 000-step goal,
2003 obese women 10 000 steps/d diary
with family
history of type 2
diabetes
Thomas and Adult volunteers 4 927 One information session explaining the importance of physical activity and 10 000-step goal,
Williams,34 describing program; wore pedometer and recorded steps daily; given diary, in
2006 booklet about the program and walking; received encouraging e-mails workplace
each wk
Tudor-Locke Inactive diabetic 16 30 Group A: First Step behavioral modification program; 4-wk adoption phase Other physical
et al,35 adults, 40-60 y included weekly group meetings, pedometers, a program manual with activity goal,
2004 a goals and problem-solving exercises, and a calendar for self-monitoring diary
of steps; daily step goals were increased each week; 12-wk adherence
phase included postcard mailings
30 Group B: pedometer and 2 motivational postcards
VanWormer Adults with CAD 17 22 Adults with CAD referred by their medical provider to the intervention “Living Goal not specified
et al,36 Well with Heart Disease”; received a course manual and ⱕ12
2004 counseling sessions by telephone, which lasted about 20 min each
(7 sessions were required).
Williams Inactive, 3 43 All participants met with the primary investigator for goal-setting and Individualized
et al,37 postmenopausal discussion of physical activity goals, diary
2005 a black women
35 Group B, the walking contract group, also wrote and signed an agreement; 10 000-step goal,
goals were negotiated based on current baseline, recommendation of diary
10 000 steps or 30 min of walking most days
Wilson et al,38 Black women who 8 24 Pedometers given to participants and 8 weekly counseling sessions were Other physical
2005 survived breast held at a community center/church to help motivate participants using activity goal,
cancer the Health Belief Model as theoretical framework; progressive step goals diary
were provided
Wyatt et al,39 Adult volunteers at 14 503, work- Starter kit given to interested worksites and churches that included Individualized
2004 worksites and site; information about Colorado on the Move, instructions on how to obtain goals, diary,
churches 232, church pedometers, ideas for initiatives, competitions to increase physical in workplace
site activity, and instructions for recording data; goal to increase by 2000
steps/d
Abbreviations: BMI, body mass index, which is calculated as weight in kilograms divided by height in meters squared; CAD, coronary artery disease; COPD, chronic obstructive pulmo-
nary disease; MI, myocardial infarction; RN, registered nurse.
a A randomized controlled trial that included pedometers in both study groups and was thus treated as 2 separate observational groups.
b Data were not reported for group B in this trial; therefore, trial was treated as a single-group observational study.

2300 JAMA, November 21, 2007—Vol 298, No. 19 (Reprinted) ©2007 American Medical Association. All rights reserved.

Downloaded From: by a University of the Sciences User on 02/28/2018

 PEDOMETERS AND INCREASED PHYSICAL ACTIVITY

those with less baseline activity to have participants with relatively high baseline provements in health outcomes. For
the greatest increases in physical activ- physicalactivity(FIGURE 3). Intervention these analyses, we included the change
ity, albeit not statistically significantly duration and physical activity counsel- in activity and outcomes from base-
(P = .06 and P = .09 respectively). Sex, ing were not significant predictors of in- line among all participants using pe-
BMI, and race/ethnicity were not sig- creased steps per day. There was no sta- dometers (from both the RCTs and ob-
nificant predictors of increased activity. tistically significant difference in effect servational studies).
Among the intervention characteris- sizes between the interventions that used Change in BMI. Intervention par-
tics, having a step goal was the key pre- a Yamax brand pedometer vs another ticipants significantly decreased their
dictor of increased physical activity pedometer. BMI by 0.38 from baseline (P = .03,
(P=.001). Indeed, the 3 studies that did Table 2). This was a statistically homo-
not include a step goal14, 21,22,36 had no sig- Pedometer Use geneous result. This decrease was as-
nificant improvement in physical activ- and Health Outcomes sociated with older age (P=.001), in-
ity with pedometer use in contrast to in- We used regression, weighted by the c re a s i n g p e rc e n t a g e o f w h i t e
creases of more than 2000 steps per day sample size, to evaluate the associa- participants (P = .009), having a step
with the use of the 10 000-step-per-day tion between steps per day and im- goal (P = .04), and interventions of
goal or other goal (TABLE 3). Only 2 stud-
ies reported the number of participants Table 2. Baseline Participant Characteristics a
who achieved their step goal, limiting our No. of Studies Change Postintervention
abilitytostratifyouranalysisbythisfactor. Reporting
This Characteristic Preintervention, Mean Change (95% P
Notably, participants in the studies Characteristic (No. of Participants) Mean (SD) Confidence Interval) b Value
that did not require the use of a step di- BMI 18 (562) 30 (3.4) −0.38 (−0.05 to −0.72) .03
ary17,34,42 did not significantly increase Blood pressure, mm Hg
their activity over baseline (mean Systolic 12 (468) 129 (7.5) −3.8 (−1.7 to −5.9) ⬍.001
change, 832; 95% CI: −258 to 1922 Diastolic 12 (468) 79 (4.5) −0.3 (0.02 to −0.46) .001
steps per day; P=.10), whereas partici- Cholesterol, mmol/L
Total 7 (192) 5.14 (0.3) −0.09 (−0.32 to 0.15) .50
pants in interventions that required the
HDL 7 (192) 1.34 (0.20) 0.06 (−0.012 to 0.14) .10
use of a diary did (mean change, 2649;
LDL 7 (192) 2.93 (0.01) −0.06 (−0.25 to 0.13) .50
95% CI, 2032 to 3266 steps per day,
Triglycerides, mmol/L 7 (192) 2.19 (0.85) −0.26 (−0.56 to 0.04) .09
P⬍ .001). Five studies measured par-
Fasting glucose, 7 (211) 7.09 (2.09) −0.03 (−0.11 to 0.11) .70
ticipants’ adherence with keeping a step mmol/L
diary (mean [SD] 83% (20%) adherent). Abbreviations: BMI, body mass index, which is calculated as weight in kilograms divided by height in meters squared; HDL,
Having the intervention in a setting high density lipoprotein; LDL, low-density lipoprotein.
SI conversion factors: To convert total, high-density lipoprotein, and high-density lipoprotein cholesterol from mmol/L to
other than the workplace also predicted mg/dL divide by 0.0259; triglycerides to mg/dL, divide by 0.0112; and fasting glucose to mg/dL, divide by 0.0555.
a For this analysis, data from all participants who wore pedometers (ie, participants in the intervention groups of the ran-
increased physical activity (P=.02). This domized controlled groups and all participants in the observational studies) were included and the changes in physical
may be explained by the finding that the activity and health outcomes were calculated as the change from baseline.
b A negative value indicates that the parameter fell after the invention, whereas a positive value indicates that the parameter
workplaceinterventionstendedtoinclude rose after the intervention.

Figure 2. Increase in Physical Activity Among Participants Randomly Assigned to Pedometer Interventions vs Control Participants

Sample Size
Difference in Change in
Source Intervention Control Steps/d, Mean (95% CI) P Value
Butler and Dwyer,17 2004 17 16 395 (–118 to 908) .13
Hultquist et al,19 2005 31 27 2226 (1488 to 2964) <.001
Araiza et al,16 2006 15 15 3189 (905 to 5473) .006
de Blok et al,18 2006 8 8 567 (–1872 to 3006) .65
Talbot et al,23 2003 17 17 1498 (–300 to 3296) .10
Moreau et al,20 2001 15 9 5066 (4003 to 6129) <.001
Izawa et al,14 2005 24 21 3254 (1851 to 4657) <.001
Ransdell et al,21 2004 and 28 9 3994 (1050 to 6938) .008
Ormes et al,22 2005

Summary difference 155 122 2491 (1098 to 3885) <.001

–2500 0 2500 5000 7500

Difference in Change in Steps/d, Mean (95% CI)

Presents the difference in the change in steps per day before and after the intervention between the participants in the experimental and control arms of the random-
ized controlled trials. The size of the data markers are proportional to the sample size, which represents the number of individuals who completed the trials.

©2007 American Medical Association. All rights reserved. (Reprinted) JAMA, November 21, 2007—Vol 298, No. 19 2301

Downloaded From: by a University of the Sciences User on 02/28/2018

 PEDOMETERS AND INCREASED PHYSICAL ACTIVITY

longer duration (P=.07 for trend). The poprotein levels, and 7 studies re- relevant reductions in weight and blood
decrease in BMI was not significantly ported change in serum glucose pressure.
associated with baseline steps per day, concentration. Intervention partici- We found that setting a step goal and
change in steps per day, sex, diet coun- pants did not significantly improve their the use of a step diary may be key mo-
seling, or BMI at the start of the inter- serum lipid levels or decrease their fast- tivational factors for increasing physi-
vention. ing serum glucose concentration cal activity. Pedometer users who were
Change in Blood Pressure. Interven- (Table 2)—not a surprising finding given a goal, whether the 10 000-step
tion participants significantly de- given that these values were fairly nor- goal or an alternative personalized step
creased their systolic blood pressure by mal for participants at baseline. goal, significantly increased their physi-
3.8 mm Hg (P ⬍ .001) and their dia- cal activity over baseline, whereas pe-
stolic blood pressure by 0.3 mm Hg COMMENT dometer users who were not given a
(P=.001) (Table 2). These were statisti- The results of this meta-analysis, which goal did not increase their physical ac-
cally heterogeneous results. This de- is to our knowledge, the first pub- tivity. The study by Sidman et al30 spe-
crease was associated with greater sys- lished quantitative synthesis of the lit- cifically compared alternative goals in
tolic blood pressure at baseline (P=.009) erature on the effectiveness of pedom- an RCT. In their intervention involv-
and change in steps per day (P=.08 for eters, suggest that pedometer use is ing sedentary women aged 20 to 65
trend) but not significantly associated associated with significant increases in years, they found that although par-
with age, change in BMI, setting a step physical activity—a magnitude of about ticipants with low levels of baseline ac-
goal, or intervention duration. 2000 steps or about 1 mile of walking tivity rarely reached their goal of 10 000
Other Health Outcomes. Six stud- per day. Moreover, the use of pedom- steps per day, they increased their steps
ies reported change in low-density li- eters may be associated with clinically as much as those asked to achieve a
more modest goal.30 Given the rela-
tively similar increases in physical ac-
Table 3. Use of a Step Goal tivity among those pedometer users
Mean Change in Physical Activity P given the 10 000-step goal and users
From Baseline, Steps/d Value given other goals, we conclude that the
Alternatives Sources a (95% Confidence Interval)
relative benefits of setting different goals
No step goal 14, 21, 22, 36 686 (−1621 to 2994) .60
remains unclear.
10 000 step/d goal 16, 19, 28-30, 33, 34, 37 2998 (1646 to 4350) ⬍.001 We found that workplace interven-
tions were associated with relatively
Other step goal b 13, 15, 17, 18, 20-24, 26, 2363 (1789 to 2936) ⬍.001 small increases in physical activity.
27, 30-32, 35, 38, 39
a Studies are included in more than 1 category because they compared 2 or more study groups that had different goals.
Workplace exercise programs have
b Typically, these were based on incremental increases in daily steps over baseline. been criticized for attracting staff who
are already active34—our results cor-
roborate this observation. Thus, for
Figure 3. Association of Baseline Physical Activity With Change in Physical Activity After the workplace interventions to have a
Intervention
broader health benefit, they might need
5000
to specifically target sedentary employ-
4500
Observational studies not in workplace ees who are not currently engaged in a
After the Intervention (Steps per Day)

Observational studies in workplace

4000 Randomized controlled trials
walking or other exercise program.
Change in Physical Activity

3500 Overall mean (95% confidence intervals) We did not find that physical activ-
3000 ity counseling increased steps walked
2500 per day. This may have been because
2000 of the heterogeneity of the counseling
1500
provided by the included studies (with
1000
some providing several weekly ses-
500
0
sions to motivate walking and give in-
–500 dividualized feedback, whereas others
–1000 provided only a brief general physical
0 2000 4000 6000 8000 10 000 12 000 14 000
activity lecture). Additionally, some
Baseline Physical Activity (Steps per Day)
studies that provided some counsel-
Presents the association of baseline physical activity in steps per day (x-axis) with the change in physical ac-
tivity in steps per day (y-axis). The Figure includes both the RCTs and the observational studies. The data mark-
ing may not have specifically reported
ers representing the workplace interventions include all the study groups in each trial: Butler and Dwyer17 and doing so. Our results are in keeping
Croteau et al28 each had 3 study groups; Eastep et al,25 Thomas et al,34 and Wyatt et al39 each had 2 study with a recent systematic review that
groups. The mean change in steps per day was 1964 over baseline (P=.01).
found mixed results of the effects of
2302 JAMA, November 21, 2007—Vol 298, No. 19 (Reprinted) ©2007 American Medical Association. All rights reserved.

Downloaded From: by a University of the Sciences User on 02/28/2018

 PEDOMETERS AND INCREASED PHYSICAL ACTIVITY

physical activity counseling for adults est or provided detailed information Author Contributions: Dr Bravata had full access to
all of the data in the study and takes responsibility for
in the primary care setting.45 about their participants. Third, be- the integrity of the data and the accuracy of the data
Pedometer users had significant re- cause many interventions included the analysis.
Study concept and design: Bravata, Smith-Spangler,
ductions in BMI; however, their weight use of 2 or more components (eg, pe- Sundaram, Sirard.
loss was not a function of increase in dometers, step goals, diaries, counsel- Acquisition of data: Bravata, Smith-Spangler,
Sundaram, Gienger, Lin, Lewis, Stave, Sirard.
daily steps. This suggests that partici- ing), the independent contribution of Analysis and interpretation of data: Bravata, Smith-
pation in the intervention either in- any one of these components is diffi- Spangler, Gienger, Lin, Olkin, Sirard.
creased activity not measured by the cult to establish. Fourth, pedometers are Drafting of the manuscript: Bravata, Smith-Spangler,
Sirard.
pedometer or resulted in decreased ca- used in these studies both as an inter- Critical revision of the manuscript for important in-
loric consumption or both. Unfortu- vention to motivate physical activity tellectual content: Bravata, Smith-Spangler, Sundaram,
Gienger, Lin, Lewis, Stave, Olkin, Sirard.
nately, too few interventions specifi- and as a tool to measure steps per day Statistical analysis: Bravata, Olkin.
cally reported providing dietary and participants may have increased Obtained funding: Bravata.
Administrative, technical, or material support: Bra-
counseling for us to include this fac- their physical activity just by virtue of vata, Smith-Spangler, Sundaram, Gienger, Lin, Lewis,
tor in our analyses. knowing that they are being moni- Sirard.
Pedometer users also significantly de- tored. However, this type of Haw- Study supervision: Bravata, Sirard.
Library searches: Stave.
creased their systolic blood pressure by thorne effect is likely to affect both in- Financial Disclosures: None reported.
almost 4 mm Hg from baseline. The tervention and control groups similarly. Funding/Support: This project was supported by grant
AG017253-06 from the National Institute on Aging
magnitude of this finding is consistent Finally, because only 5 studies in- through The Stanford Center on the Demography and
with other published meta-analyses of volved participants with a mean age Economics of Health and Aging. Dr Olkin was funded
in part by grant DMS 9626265 from the National Sci-
the effects of physical activity on blood older than 60 years and only 15% of the ence Foundation .
pressure.46-50 Reducing systolic blood participants were men, the generaliz- Role of the Sponsor: The funding agencies had no role
pressure by 2 mm Hg is associated with ability of our results to older and male in the design and conduct of the study; collection, man-
agement, analysis, and interpretation of the data; and
a 10% reduction in stroke mortality and populations is limited. preparation, review, or approval of this manuscript.
a 7% reduction in mortality from vas- Given these limitations, to fully elu- Acknowledgment: We thank Dawn Bravata, MD, In-
diana University School of Medicine, for her thought-
cular causes in middle-aged popula- cidate the potential benefits of pedom- ful comments on a prior draft of the manuscript, for
tions51; thus, it is critical that the ef- eters, large, randomized controlled trials which she received no compensation.
fects of pedometer use on blood of men and women over a range of ages
pressure be examined closely in fu- in the outpatient setting is required. REFERENCES
ture studies. Because blood pressure re- Such trials should make the following 1. Alevizos A, Lentzas J, Kokkoris S, Mariolis A, Ko-
ductions were greatest among partici- comparisons: (1) pedometer use in rantzopoulos P. Physical activity and stroke risk. Int J
Clin Pract. 2005;59(8):922-930.
pants with the highest baseline blood which participants can see their daily 2. Brown DR. Physical activity, ageing, and psycho-
pressure, this result may in part be due step counts vs pedometer use in which logical well-being: an overview of the research. Can
to regression to the mean. However, the they are blinded to their daily step J Sport Sci. 1992;17(3):185-193.
3. Speck BJ, Looney SW. Effects of a minimal inter-
overall reduction in blood pressure in counts, (2) pedometer use with vs with- vention to increase physical activity in women: daily
the included studies is particularly in- out a step goal, (3) pedometer use with activity records. Nurs Res. 2001;50(6):374-378.
4. Miller TD, Balady GJ, Fletcher GF. Exercise and its
teresting given that most of partici- vs without physical activity counsel- role in the prevention and rehabilitation of cardiovas-
pants were normotensive at baseline— ing and feedback (including face-to- cular disease. Ann Behav Med. 1997;19(3):220-229.
5. Surgeon General’s healthy weight advice for
only 1 of our included studies targeted face sessions and electronic feed- consumers. http://www.surgeongeneral.gov/topics
hypertensive patients.20 Our finding that back), and (4) pedometer use with vs /obesity/calltoaction/fact_advice.htm. Accessed May
5, 2005.
the reduction in systolic blood pres- without the use of step diaries. Key out- 6. Preventing obesity and chronic diseases through
sure was independent of decreases in comes for such trials include both good nutrition and physical acivity. July 2005. http:
BMI was consistent with the results of physical activity as well as detailed as- //www.cdc.gov/nccdphp/publications/factsheets
/Prevention/obesity.htm. Accessed September 19,
Whelton et al.46 By highlighting the sessments of key health outcomes mea- 2007.
health benefits from physical activity sured both in the short and longer term. 7. Pedometers: walking by the numbers. Consum Rep.
October 2004:30-31.
exclusive of weight loss, health profes- Despite the abundance of lay litera- 8. Tudor-Locke C, Bassett DR Jr. How many steps/
sionals may encourage patients who are ture on the use of pedometers, our study day are enough? preliminary pedometer indices for
public health. Sports Med. 2004;34(1):1-8.
frustrated by an inability to lose weight is the first published synthesis of the 9. Hedges LV, Olkin I. Statistical Methods for
to engage in physical activity. evidence. Our results suggest that the Meta-Analysis. San Diego, CA: Academic Press; 1985.
Our analyses reflect some limita- use of these small, relatively inexpen- 10. Gleser L, Olkin I. Stochastically Dependent Effect
Sizes. In: Cooper J, Hedges L, eds. The Handbook of
tions of the included studies. First, sive devices is associated with signifi- Research Synthesis. New York, NY: Russel Sage Foun-
study sizes were relatively small and in- cant increases in physical activity and dation; 1994:339-355.
11. Cooper H, Hedges L. The Handbook of Re-
terventions were of relatively short du- improvements in some key health out- search Synthesis. New York, NY: Russell Sage Foun-
ration and heterogeneous in their de- comes, at least in the short term. The dation; 1994.
12. Higgins JP, Thompson SG. Quantifying hetero-
sign. Second, few studies evaluated extent to which these results are du- geneity in a meta-analysis. Stat Med. 2002;21(11):
more than 1 of the outcomes of inter- rable over the long term is unknown. 1539-1558.

©2007 American Medical Association. All rights reserved. (Reprinted) JAMA, November 21, 2007—Vol 298, No. 19 2303

Downloaded From: by a University of the Sciences User on 02/28/2018

 PEDOMETERS AND INCREASED PHYSICAL ACTIVITY

13. Chan CB, Ryan DAJ, Tudor-Locke C. Health ben- Shultz B. Does augmented feedback from pedom- African American breast cancer survivors: a pilot study.
efits of a pedometer-based physical activity interven- eters increase adults’ walking behavior? Percept Mot Prev Chronic Dis. 2005;2(2):A16.
tion in sedentary workers. Prev Med. 2004;39(6): Skills. 2004;99(2):392-402. 39. Wyatt HR, Peters JC, Reed GW, et al. Using elec-
1215-1222. 26. Jensen GL, Roy MA, Buchanan AE, Berg MB. tronic step counters to increase lifestyle physical ac-
14. Izawa KP, Watanabe S, Onvya K, et al. Effect of Weight loss intervention for obese older women: im- tivity: Colorado on the Move. J Phys Act Health. 2004;
the self-monitoring approach on exercise mainte- provements in performance and function. Obes Res. 1(3) 181-191.
nance during cardiac rehabilitation—a randomized, 2004;12(11):1814-1820. 40. Hillsdon M, Foster C, Thorogood M. Interven-
controlled trial. Am J Phys Med Rehabil. 2005;84 27. Koulouri AA, Tigbe WW, Lean ME. The effect of tions for promoting physical activity. Cochrane Da-
(5):313-321. advice to walk 2000 extra steps daily on food intake. tabase Syst Rev. 2005;(1):CD003180.
15. Kilmer DD, Wright NC, Aitkens S. Impact of a J Hum Nutr Diet. 2006;19(4):263-266. 41. Schneider PL, Crouter SE, Lukajic O, Bassett DR Jr.
home-based activity and dietary intervention in people 28. Lindberg R. Active living: on the road with the Accuracy and reliability of 10 pedometers for mea-
with slowly progressive neuromuscular diseases. Arch 10,000 steps program. J Am Diet Assoc. 2000;100 suring steps over a 400-m walk. Med Sci Sports Exerc.
Phys Med Rehabil. 2005;86(11):2150-2156. (8):878-879. 2003;35(10):1779-1784.
16. Araiza P, Hewes H, Gashetewa C, Vella CA, Burge 29. Schneider PL, Bassett DR Jr, Thompson DL, Pronk 42. Crouter SE, Schneider PL, Karabulut M, Bassett
MR. Efficacy of a pedometer-based physical activity pro- NP, Bielak KM. Effects of a 10,000 steps per day goal DR Jr. Validity of 10 electronic pedometers for mea-
gram on parameters of diabetes control in type 2 dia- in overweight adults. Am J Health Promot. 2006; suring steps, distance, and energy cost. Med Sci Sports
betes mellitus. Metabolism. 2006;55(10):1382-1387. 21(2):85-89. Exerc. 2003;35(8):1455-1460.
17. Butler L, Dwyer D. Pedometers may not provide 30. Sidman CL, Corbin CB, Le Masurier G. Promot- 43. Le Masurier GC, Tudor-Locke C. Comparison of
a positive effect on walking activity. Health Promo J ing physical activity among sedentary women using pedometer and accelerometer accuracy under con-
Australia. 2004;15(2):134-136. pedometers. Res Q Exerc Sport. 2004;75(2):122- trolled conditions. Med Sci Sports Exerc. 2003;35
18. de Blok BMJ, de Greef MHG, ten Hacken NHT, 129. (5):867-871.
Sprenger SR, Postema K, Wempe JB. The effects of a 31. Stovitz SD, VanWormer JJ, Center BA, Bremer 44. Bassett DR Jr, Ainsworth BE, Leggett SR, et al. Ac-
lifestyle physical activity counseling program with feed- KL. Pedometers as a means to increase ambulatory ac- curacy of five electronic pedometers for measuring dis-
back of a pedometer during pulmonary rehabilita- tivity for patients seen at a family medicine clinic. tance walked. Med Sci Sports Exerc. 1996;28(8):
tion in patients with COPD: a pilot study. Patient Educ J Am Board Fam Pract. 2005;18(5):335-343. 1071-1077.
Couns. 2006;61(1):48-55. 32. Sugiura H, Kajima K, Mirbod SM, Iwata H, Mat- 45. Eden KB, Orleans CT, Mulrow CD, Pender NJ,
19. Hultquist CN, Albright C, Thompson DL. Com- suoka T. Effects of long-term moderate exercise and Teutsch SM. Does counseling by clinicians improve
parison of walking recommendations in previously in- increase in number of daily steps on serum lipids in physical activity? a summary of the evidence for the
active women. Med Sci Sports Exerc. 2005;37(4): women: randomised controlled trial US Preventive Services Task Force. Ann Intern Med.
676-683. [ISRCTN21921919]. BMC Womens Health. 2002; 2002;137(3):208-215.
20. Moreau KL, Degarmo R, Langley J, et al. Increas- 2(1):3. 46. Whelton SP, Chin A, Xin X, He J. Effect of aero-
ing daily walking lowers blood pressure in postmeno- 33. Swartz AM, Strath SJ, Bassett DR, et al. Increas- bic exercise on blood pressure: a meta-analysis of ran-
pausal women. Med Sci Sports Exerc. 2001;33(11): ing daily walking improves glucose tolerance in over- domized, controlled trials. Ann Intern Med. 2002;
1825-1831. weight women. Prev Med. 2003;37(4):356-362. 136(7):493-503.
21. Ransdell LB, Robertson L, Ornes L, Moyer- 34. Thomas L, Williams M. Promoting physical ac- 47. Murphy MH, Nevill AM, Murtagh EM, Holder
Mileur L. Generations exercising together to improve tivity in the workplace: using pedometers to increase RL. The effect of walking on fitness, fatness and rest-
fitness (GET FIT): a pilot study designed to increase daily activity levels. Health Promot J Aust. 2006; ing blood pressure: a meta-analysis of randomised, con-
physical activity and improve health-related fitness in 17(2):97-102. trolled trials. Prev Med. 2007;44(5):377-385.
three generations of women. Women Health. 2004; 35. Tudor-Locke C, Bell RC, Myers AM, et al. Con- 48. Kelley GA, Kelley KS, Tran ZV. Walking and rest-
40(3):77-94. trolled outcome evaluation of the First Step Program: ing blood pressure in adults: a meta-analysis. Prev Med.
22. Ornes LL, Ransdell LB, Robertson L, Trunnell E, a daily physical activity intervention for individuals with 2001;33(2 pt 1):120-127.
Moyer-Mileur L. A 6-month pilot study of effects of type II diabetes. Int J Obes Relat Metab Disord. 2004; 49. Halbert JA, Silagy CA, Finucane P, Withers RT,
a physical activity intervention on life satisfaction with 28(1):113-119. Hamdorf PA, Andrews GR. The effectiveness of ex-
a sample of three generations of women. Percept Mot 36. VanWormer JJ, Boucher JL, Pronk NP, Thoennes ercise training in lowering blood pressure: a meta-
Skills. 2005;100(3 pt 1):579-591. JJ. Lifestyle behavior change and coronary artery dis- analysis of randomised controlled trials of 4 weeks or
23. Talbot LA, Gaines JM, Huynh TN, Metter EJ. A ease: effectiveness of a telephone-based counseling longer. J Hum Hypertens. 1997;11(10):641-649.
home-based pedometer-driven walking program to program. J Nutr Educ Behav. 2004;36(6):333-324. 50. Kelley G, Tran ZV. Aerobic exercise and normo-
increase physical activity in older adults with osteo- 37. Williams BR, Bezner J, Chesbro SB, Leavitt R. The tensive adults: a meta-analysis. Med Sci Sports Exerc.
arthritis of the knee: a preliminary study. J Am Geri- effect of a behavioral contract on adherence to a 1995;27(10):1371-1377.
atr Soc. 2003;51(3):387-392. walking program in postmenopausal African Ameri- 51. Lewington S, Clarke R, Qizilbash N, Peto R, Col-
24. Croteau KA. A preliminary study on the impact can women. Top Geriatr Rehab. 2005;21(4):332- lins R. Age-specific relevance of usual blood pressure
of a pedometer-based intervention on daily steps. Am 342. to vascular mortality: a meta-analysis of individual data
J Health Promot. 2004;18(3):217-220. 38. Wilson DB, Porter JS, Parker G, Kilpatrick J. An- for one million adults in 61 prospective studies. Lancet.
25. Eastep E, Beveridge S, Eisenman P, Ransdell L, thropometric changes using a walking intervention in 2002;360(9349):1903-1913.

Author in the Room Teleconference

Join Dr Bravata, an author of this article, on Wednesday, December 19, 2007, from
2 to 3 PM eastern time for the Author in the Room teleconference.
To register for Author in the Room, please visit http://www.ihi.org/authorintheroom.
You can listen to past conferences or subscribe to the podcast at http://jama.ama-assn
.org/authorintheroom/authorindex.dtl.

2304 JAMA, November 21, 2007—Vol 298, No. 19 (Reprinted) ©2007 American Medical Association. All rights reserved.

Downloaded From: by a University of the Sciences User on 02/28/2018