A tailored internet-plus-email intervention for increasing physical activity among ethnically-diverse women

A tailored internet-plus-email intervention for increasing physical activity among ethnically-diverse women

Preventive Medicine 47 (2008) 605–611 Contents lists available at ScienceDirect Preventive Medicine j o u r n a l h o m e p a g e : w w w. e l s e v...

542KB Sizes 0 Downloads 118 Views

Preventive Medicine 47 (2008) 605–611

Contents lists available at ScienceDirect

Preventive Medicine j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / y p m e d

A tailored internet-plus-email intervention for increasing physical activity among ethnically-diverse women Genevieve Fridlund Dunton a, Trina P. Robertson b,⁎ a Health Promotion Research Branch, Behavioral Research Program, Division of Cancer Control and Population Sciences, National Cancer Institute, National Institutes of Health, Bethesda, MD, 20852, USA b Dairy Council of California, 2151 Michelson Drive, Ste. 235 Irvine, CA 92612-1339, USA

a r t i c l e

i n f o

Available online 15 October 2008 Keywords: Exercise Walking Intervention Internet Women Compliance

a b s t r a c t Objective. To evaluate the feasibility and efficacy of an individually tailored, Internet-plus-email physical activity intervention designed for adult women. Method. Healthy and ethnically-diverse adult females (N = 156) (mean age = 42.8 years, 65% Caucasian) from California were randomly assigned to an intervention (access to a tailored website and weekly emails) or waitlist control group. Participants completed web-based assessments of physical activity, stage of behavior change, and psychosocial variables at baseline, one month, two months, and three months. Data were collected during 2006–2007. Multilevel random coefficient modeling examined group differences in rates of change. Results. As compared to the control condition, the intervention group increased walking (+ 69 versus +32 min per week) and total moderate-to-vigorous physical activity (+23 versus −25 min per week) after three months. The intervention did not impact stage of behavior change or any of the other psychosocial variables. Conclusion. A tailored, Internet-based intervention for adult women had a positive effect on walking and moderate-to-vigorous physical activity in an ethnically-diverse sample. However, given the lack of comparable research contact in the control group, these findings should be taken cautiously. © 2008 Elsevier Inc. All rights reserved.

Introduction Participating in regular physical activity has been associated with reduced risk of type 2 diabetes, heart disease and cancer (Hu et al., 2007; Physical Activity Guidelines Advisory Committee, 2008; Thompson et al., 2003). To achieve these health benefits, it is recommended that all adults engage in 30 min or more of moderate-intensity physical activity on most days of the week (Haskell et al., 2007). Adult women are less likely to meet these physical activity recommendations than men. Recent estimates suggest that only 47.7% of adult women (versus 50.5% of men) report participation in 30+ min of moderate physical activity on five or more days per week or vigorous physical activity for 20+ min on three or more days per week (CDC, 2007). Given this gender gap, there is growing interest in developing interventions that effectively promote physical activity among adult women. Due to child and family obligations, women may face a number of challenges to participating in face-to-face physical activity programs (e.g., support groups, counseling), including a lack of time, child-care needs, limited transportation and an inability to cover the financial costs (Sørenson and Gill, 2007; Wilcox et al., 2003). In recent years,

⁎ Corresponding author. E-mail address: [email protected] (T.P. Robertson). 0091-7435/$ – see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.ypmed.2008.10.004

there has been growing interest in Internet-based physical activity interventions (Norman et al., 2007), which could be especially helpful to women by allowing them to participate from their own home at a convenient time and for a low cost. Studies examining the efficacy of Internet-delivered programs have demonstrated small (e.g., Palmer et al., 2005; PLotnikoff et al., 2005) to medium (Napolitano et al., 2003) effect sizes for physical activity. However, little is known about webbased physical activity interventions specifically designed for women. There is some evidence to suggest that women may be less interested in using interactive media than men. Women generally use the Internet less frequently, visit fewer web sites, and stay on the Internet for shorter periods of time (Pew Internet, 2005). On the other hand, participation on Internet-delivered health websites emphasizing support and community has been found to be greater among women than men (Ginossar, 2008). Although women are less attracted to interactive media, they may be more likely to consume and utilize Internetdelivered health information that takes a gender-based approach. Research on the efficacy of Internet-delivered physical activity interventions among women is generally lacking. Of the thirteen studies identified by the review conducted by Norman et al. (2007), only three Internet-based interventions specifically targeted adult women. An intervention providing brief weekly emails to women increased total walking minutes, but it did not include a control group (Dinger et al., 2004). In the two other studies, the effects on physical activity were

606

G.F. Dunton, T.P. Robertson / Preventive Medicine 47 (2008) 605–611

marginal (Rovniak et al., 2005) and nonsignificant (Hageman et al., 2005). All of these studies were also limited by the small proportion of racial/ethnic minority participants (b19%). To address these gaps, the present study served as an evaluation of the feasibility and efficacy of a gender-targeted and individually tailored Internet-plus-email physical activity intervention as compared to a non-technology wait-list control in a group of racially/ethnically-diverse women. Methods Study design A total of 156 women were randomized to either the tailored Internet-plus-email physical activity intervention (n = 85) or wait-list control conditions (n = 71). Participants assigned to the intervention condition received ten weekly emails containing links to a webpage with an interactive information tailoring tool to promote physical activity. In contrast, women in the control condition did not have access to the webpage until after the study was completed. They were not provided with any health-related information during the study period. Participants in both study arms completed web-based assessments of physical activity and psychosocial variables four times throughout the study (baseline, one-month, two-months, and three-months). Participants were sent a check for $25 after completing all of the surveys. Prior to randomization, women (1) participated in a telephone screening to assess eligibility, (2) provided informed consent to participate in the study and (3) completed the baseline set of online questionnaires. Random allocation to groups was conducted through a coin toss. An email from the researchers then informed participants of whether they had been assigned to the intervention or control group. Due to the nature of the intervention, participants were not blinded to their study condition. Participants Healthy, middle-aged women (ages 21–65 years) living in Northern and Southern California were recruited for the study. Participants were initially contacted through posters and fliers posted at local health centers and clinics. Wellness coordinators and California dietitians also forwarded recruitment emails to their employees and clients. Interested individuals contacted the study coordinator by phone to obtain information about the study and to be screened for eligibility. Exclusion criteria included the following: (1) a history of major health problems including high blood pressure, diabetes, excessive overweight (i.e., body mass index of ≥40), joint problems, pregnancy and other issues; (2) inability to read English at the seventh-grade or higher level; (3) lack of regular access to the Internet; and (4) lack of an email account that can be accessed regularly over the next 12 weeks. Recruitment, enrollment, and data collection occurred on a rolling basis between March 2006 and January 2007. All study procedures were approved by the Institutional Review Board at Independent Review Consulting, Inc. Intervention webpage and emails The intervention webpage, called “Women's Fitness Planner” (WFP), was a component of a larger nutrition and physical activity website (www.MealsMatter.org). The WFP contained an interactive computer program that produced individualized physical activity feedback on the basis of information provided in an online assessment. The WFP was developed in 2000 by Dairy Council of California as a way to offer tailored output direct to consumers at a low cost. The webpage targeted female users specifically by displaying pictures of age and activity specific photos of women exercising (e.g., pregnant women walking, young women stretching, middle-aged women jogging, and mature women gardening). Upon being directed to the WFP, an online assessment device measured physical activity level (including fre-

quency, intensity and duration) and behavior change process variables such as barriers, motivators and stages of change. After this information was entered, the computer program generated a graph displaying each respondent's self-reported current level of activity compared to the 2005 USDA Dietary Guidelines for physical activity (USDHHS, 2005). The intervention utilized theoretical framework provided by the Health Belief Model (HBM) (Rosenstock, 1974; Rosenstock et al, 1988) and the Transtheoretical Model (TTM) (Prochaska, 1995). Theory-based constructs were operationalized through the generation of tailored messages. Following the HBM, the intervention sought to reduce barriers to physical activity by providing suggestions for overcoming them. For example, participants who reported “too busy to exercise” as a barrier, received a tailored message that provided tips for overcoming this obstacle (e.g., “Get your family to exercise with you. That way you're spending quality time together and being active at the same time!”). The TTM was used to identify readiness to change and tailor output accordingly. For example, if participants reported they did not want to change their activity level (precontemplation or maintenance), a tailored message provided either the consequences of inactivity or reinforcement for meeting activity levels based on selfreported number of minutes of physical activity. Intervention participants also received 10 weekly follow-up email newsletters to support the messages and information provided on the website and to encourage further learning. The weekly email newsletter addressed topics such as how to measure activity intensity, how to keep an activity journal, goal setting, strength training, a review of the number of minutes recommended for activity and a link to downloadable log. Tips for weight management such as reminders to eat breakfast, information about appropriate portion sizes, and osteoporosis prevention were also included. Measurements All outcome and process measures were completed online. Instruments used to assess the outcome measures were the same ones employed to generate the tailored Internet-based messages in the intervention. Outcome measures Physical activity Using a standardized activity inventory format (Hopkins et al., 1991), participants reported the frequency and duration (in minutes) of sport or exercise activities undertaken in the past two weeks. The survey instrument used a predetermined list of a possible 29 activities (e.g., walking, running, gardening or yard work, dancing, yoga, volleyball, weight lifting or training). If respondents participated in activities that were not on the list, they could add that activity through a write-in option. Physical activity estimates generated from this type of inventory format have been shown to correspond moderately with the Stanford 7-day recall instrument and other objective measures (e.g., maximum load, vital capacity, and body mass index) (Hopkins et al., 1991). Researchers converted each activity into Metabolic Equivalents (METs) using the compendium developed by Ainsworth et al. (2000). Minutes-per-week in activities greater than or equal to 3.5 METS were summed in order to develop an indicator of moderate-tovigorous physical activity (MVPA). Minutes per week reported in the brisk walking (greater than 3.5 mi/h) category were examined separately. Stage of change Stage of change for physical activity was assessed through the Stages of Exercise Change Questionnaire (SECQ) (Reed et al, 1997). The SECQ is a four-item measure with a yes/no response format. Items assessed past and current participation in physical activity, and intentions to engage in physical activity in the future. A scoring algorithm

G.F. Dunton, T.P. Robertson / Preventive Medicine 47 (2008) 605–611

was used to categorize all response combinations as one of the five primary stages of the TTM (Prochaska, 1979): precontemplation (do not intend to be active in the next six months), contemplation (intend to be active in the next six months), preparation (intend to be active in the next 30 days), action (have been active for less than six months) and maintenance (have been active for more than six months). These stages have been found to significantly differ by self-reported strenuous and moderate exercise (Hausenblas et al, 2003) and aerobic fitness (Dannecker et al., 2003). Psychosocial variables The following psychosocial factors related to physical activity were assessed: self-efficacy, perceived barriers and perceived benefits. Selfefficacy for exercise was measured using a scale developed by Sallis et al. (1998). The current study also assessed perceived barriers and benefits to physical activity using instruments developed by Sallis et al. (1989). Research supports the validity of these instruments, showing the following correlations with vigorous exercise: self-efficacy (r = .48), perceived barriers (r = −.22), perceived benefits (r = .24) (Sallis et al., 1989). Missing responses for any individual items on the psychosocial instruments resulted in missing summary scores for that variable.

607

Results Descriptive statistics Of the 231 adult women who expressed interest in the WFP evaluation study, 178 (77%) eligible women completed the baseline survey and 156 (67.5%) women were subsequently randomized to either the intervention (n = 85) or control (n = 71) group (see Fig. 1). Descriptive statistics for demographic, psychosocial and physical activity variables at all three time points for the intervention and control groups are displayed in Tables 1 and 2. There were no significant baseline differences between the intervention and control groups on any of the demographic, psychosocial or physical activity variables. Approximately 75% of participants (n = 117) completed surveys at all four time points. A chi-square test found that a significantly greater proportion of intervention as compared to control participants failed to complete surveys at all four time points (35.2% versus 16.9%) (χ2 (df = 1) = 6.65, p = .01). When examining individual time points, there were no group differences in survey completion rates at baseline,

Sociodemographic variables Sociodemographic variables including age, race/ethnicity, highest level of education, annual household income and marital status were assessed through self-report on the baseline survey. Process measures Participants' satisfaction with the intervention was assessed on the three-month survey. Respondents reported the extent to which the website and emails were helpful and interesting and whether they would recommend them to a friend. Intervention compliance (i.e., number of times the website was visited, number of emails opened, and number of email-embedded website links opened) was measured using self-report and email-tracking technology. Statistical analyses A target sample size of N = 200 for the present study was on the basis of detecting a small to medium effect with power (β = .80) and a one-tailed error probability (α/2 = .05). Demographic characteristics and baseline values for physical activity and psychosocial variables were compared between the intervention and control groups using chi-square for categorical variables and Analysis of Variance (ANOVA) for continuous variables. The extent to which the WFP impacted knowledge, perceived barriers, perceived benefits, self-efficacy, and physical activity behavior (MVPA and walking) was evaluated using multilevel random coefficient modeling (HLM version 6.0, Scientific Software International, Lincolnwood, IL; Bryk and Raudenbush, 1992). This statistical technique determined group differences (i.e., intervention versus control) in the rates of change in the outcome variables. The level-1 model estimated separate slope coefficients for each participant, which were subsequently entered as dependent variables in the level-2 model predicted by group assignment. Multilevel random coefficient modeling is especially well suited for the analysis of repeated measures data clustered within the individual that violate standard statistical assumptions of non-independence. It also can handle missing data (Bryk and Raudenbush, 1992). The extent to which the intervention influenced the likelihood of being in the action or maintenance stage of behavior change was tested using the Hierarchical Generalized Linear Model (HGLM) function, a non-linear analysis for binary outcomes using the Bernoulli distribution. Robust standard errors were utilized for all analyses due to the positive skew observed for physical activity data. An intent-to-treat analysis was conducted for all outcomes.

Fig. 1. Flow diagram of study participants. The study was conducted in California in 2006–2007.

608

G.F. Dunton, T.P. Robertson / Preventive Medicine 47 (2008) 605–611

Table 1 Baseline demographic characteristics (mean and standard deviation or %) by study group Variable

p-valuea

Group

Age (years) BMI

Race/Ethnicity African-American Asian Hispanic Caucasian Other Education High school graduate Some college Vocational/technical degree College graduate Post-graduate degree Annual household income Less than $25,000 $25,000–$49,999 $50,000–$99,999 $100,000–$149,000 $150,000–$199,999 $200,000–$249,999 $250,000–$299,999 $300,000–$499,999 Marital status Married Divorced Never married Separated Children at home (yes) General health Excellent Very good Good Fair Poor

Intervention

Control

n

Mean (S.D.)

n

Mean (S.D.)

51 79

42.8 (12.8) 25.8 (5.5)

52 65

42.8 (10.5) 26.2 (6.4)

n

%

n

%

85

69

.53

15.3 10.6 8.2 60.0 5.9

8.7 11.6 5.8 71.0 2.9

85

69

.50

3.5 14.1 7.1 38.8 32.5

1.4 15.9 4.3 50.7 27.5

82

67

.23

7.3 18.3 41.5 26.8 6.1 0.0 0.0 0.0

1.5 26.9 35.8 23.9 7.5 0.0 3.0 1.5

85

85 85

.99 .66

69 44.7 18.8 34.1 2.4 42.4

68 71

16.5 54.1 27.1 2.4 0.0

Mean values for minutes per week of walking for each group at baseline, one-, two-, and three-months are shown in Fig. 2. Across the whole intervention, walking increased at a faster rate in the intervention group than the control group at three months, β = 15.04 (SE = 8.38), p = .035 (one-tailed). After three months, the intervention group increased walking by 69 min per week, as compared to the increase by 32 min per week observed in the control group. Fig. 3 shows the mean values for minutes per week of MVPA by group at each time point. Multilevel modeling analyses found that there was a significant group difference in the rate of change in MVPA β =17.02 (SE = 10.11), p = .045 (one-tailed). Between baseline and the three months assessment, minutes per week of MVPA increased to a greater extent in the intervention group (mean increase of 23 min per week) as compared to the control group (mean decrease of 25 min per week). The proportion of participants in action or maintenance stage by group at each time point is displayed in Fig. 4. After three months, the proportion of participants in action or maintenance significantly increased across both the intervention and control group (OR = 1.31 (95% CI = 1.16–1.48) (14% for the control group and 18% for the intervention group). The rate of change in the likelihood of being in action or maintenance did not significantly differ between the two groups OR = 1.16 (95% C.I. = 0.93–1.4) across the three-month time period. Effect of the intervention on psychosocial variables

.20 59.4 8.7 30.4 1.4 42.6

Effect of the intervention on physical activity

There were no significant group differences in the rate of change in perceived physical activity barriers and benefits (p's N .05). Likewise, rates of change in self-efficacy did not differ between the intervention and control group (p's N .05).

.97 .68

Intervention compliance

13.9 50.0 30.6 5.6 0.0

a For the difference between groups by t-test or chi-square (two-tailed). The study was conducted in California in 2006–2007.

1-month, and 3-months. However, a greater proportion of control than intervention participants completed the 2-month survey (p = .01). Also, individuals who did not complete surveys at all four time points reported lower levels of MVPA than individuals who completed all four surveys at baseline (181.0 versus 249.8 min per week) (t(146.997) = 2.10, p = .028). In particular, baseline MVPA was higher among individuals completing versus not completing the survey at 1-month and 2-months (p's b .05). There were no differences in baseline MVPA by survey completion status at 3-months. Participants completing as compared to not completing all four surveys or any individual survey did not differ on any other demographic or physical activity variables.

Overall, 6% of intervention participants reported that they did not receive any weekly newsletters via email, and 11% of control participants reported that they received the weekly newsletters. Among intervention participants, the frequency of visiting the website was distributed as follows: 1–2 times (21%), 3–5 times (37%), 6–10 times (29%), more than 10 times (8%), and 5% never visited it. Email compliance data was available for n = 78 of the intervention participants. Out of ten emails that were sent, participants opened an average of 7.44 (SD = 4.06). Approximately 23% opened all of the weekly emails and 8% did not open any of the emails. On average, participants opened 6.65 (SD = 6.33) out of 25 web links (27%) that were embedded in the emails. Four percent of intervention participants did not open any links. Intervention satisfaction Overall, results suggest that participants were fairly satisfied with components of the intervention. Approximately 46% of intervention participants reported that the intervention website was moderately or

Table 2 Mean scores for psychosocial and physical activity variables at Baseline, 1-month, 2-months, and 3-months Variable

Barriersa Benefitsa Self-efficacya Walking (min/week) MVPA (min/week) a

Intervention

Control

Baseline

1-month

2-months

3-months

Baseline

1-month

2-months

3-months

n

M (SD)

n

M (SD)

n

M (SD)

n

M (SD)

n

M (SD)

n

n

M (SD)

n

M (SD)

71 82 78 84

2.2 (0.6) 4.4 (0.8) 3.4 (0.7) 147.2 (183.9)

78 85 81 75

2.2 (0.6) 4.4 (0.7) 3.4 (0.7) 217.9 (272.3)

79 85 82 63

2.1 (0.6) 4.4 (0.8) 3.4 (0.8) 173.0 (158.7)

82 85 82 66

2.1 (0.6) 59 4.5 (0.7) 69 3.4 (0.8) 64 215.8 (225.4) 71

2.3 4.5 3.4 125.9

67 2.3 (0.6) 70 4.6 (0.4) 69 3.4 (0.8) 66 164.4 (246.0)

69 70 70 64

2.2 (0.6) 4.5 (0.6) 3.3 (0.8) 127.5 (122.7)

69 70 69 62

2.2 4.5 3.3 158.3

84 234.0 (234.1) 75 282.6 (265.7) 63 263.0 (266.9) 66 275.9 (281.7)

(0.6) (0.6) (0.7) (188.4)

M (SD)

71 230.0 (269.3) 66 267.5 (248.8)

(0.6) (0.6) (0.9) (252.2)

64 232.3 (195.6) 62 223.4 (198.7)

Variables use a 5-point response scale. MVPA = moderate-to-vigorous physical activity. The study was conducted in California in 2006–2007.

G.F. Dunton, T.P. Robertson / Preventive Medicine 47 (2008) 605–611

609

Fig. 4. Proportion of participants in action or maintenance stage of behavior change at baseline, 1-month, 2-months, and 3-months by intervention versus control group. The study was conducted in California in 2006–2007. Fig. 2. Mean minutes of walking at baseline, 1-month, 2-months, and 3-months by intervention versus control group. The study was conducted in California in 2006–2007.

extremely useful, and 63% reported that the website was either extremely or moderately interesting. Furthermore, 66% reported that they would recommend the website to a friend. About 55% of intervention participants reported that the weekly emails were either extremely or moderately useful, and 58% reported that the weekly emails were extremely or moderately interesting.

other types of programs. The present study tested the efficacy of an individually tailored Internet-plus-email physical activity intervention versus wait-list control in an ethnically diverse sample of adult women. After three months, the intervention had a positive impact on walking and moderate-to-vigorous physical activity. Stage of change and potential psychosocial mediators were not significantly influenced by the intervention. Given the lack of comparable research contact in the control group, these findings should be taken with caution. Effects on physical activity

Discussion Internet-based interventions provide an opportunity to promote physical activity among women who may face barriers to participate in

The favorable impact on walking is consistent with other Internetdelivered physical activity programs in women, which have focused on walking as a key outcome (Dinger et al., 2004; Hageman et al., 2005; Rovniak et al 2005). The fact that the Internet-based intervention influenced walking could reflect women's interest in walking as a form of exercise (Humphreys and Ruseski, 2007). Although women tend to report lower levels of total physical activity than men (NCHS, 2007b), research suggests that a greater percentage of women engage in walking for leisure-time physical activity than men (Rafferty et al., 2002; Simpson et al., 2002). Significant group differences in walking and MVPA were not observed until after three months―suggesting that extended exposure to internet-based interventions may be necessary to sufficiently impact behavior. Overall, the beneficial effect on minutes of walking and moderate-to-vigorous physical activity per week observed in the current and other related studies suggests that using the Internet may be an effective strategy to promote healthful levels of physical activity among women. Effects on psychosocial variables

Fig. 3. Mean minutes of moderate-to-vigorous physical activity (MVPA) at baseline, 1-month, 2-months, and 3-months by intervention versus control group. The study was conducted in California in 2006–2007.

Consistent with other recent research on Internet-based physical activity interventions (Cook et al., 2007), support was not found for the impact of the intervention on psychosocial mediators of behavior change. One explanation for this finding is that the intervention components did not impact the hypothesized psychosocial variables. It is possible that these intervention components were not received in the manner that was intended. Also, the materials could have influenced different psychosocial variables altogether. It may be that web-delivered physical activity interventions impact a different set of psychosocial factors that were not assessed in the current study such as knowledge and social support (Barrera et al., 2002; Kalichman et al., 2006).

610

G.F. Dunton, T.P. Robertson / Preventive Medicine 47 (2008) 605–611

Process outcomes Although a majority of participants reported that the intervention website and weekly emails were interesting, levels of engagement with these two components were moderate to low. These results suggest that compliance with e-health interventions poses a similar challenge among women as it does in other populations (Leslie et al., 2005; Spittaels et al., 2007). Future research should seek to identify the unique barriers to Internet- and email-based interventions that are faced by women, such as lack of time and lack of familiarity with technology. Study limitations Despite the randomized-controlled design and ethnic diversity of the sample, this study had some limitations. First, the wait-list control condition did not receive comparable contact with research staff (other than the monthly web-based assessments). Thus, it is possible that the observed effects were due to non-specific factors. A second limitation is the differential attrition in the intervention and control groups. However, using a multilevel modeling approach to analyzing the data, which can handle missing data within individuals, avoids biases that would be introduced through the otherwise case-wise deletion of participants with missing values. Third, some of the observed differences between groups were small in size (e.g., change in the proportion in action and maintenance), possibly due to the moderate to low levels of intervention compliance (e.g., website usage and email opening) and/or initial desire for behavior change among members of the control group (i.e., overall study selection bias). Fourth, using a selfreport instrument to assess physical activity, the main outcome measure, could introduce recall and reporting biases. Lastly, findings may lack generalizability. Participants appeared more physically active and less overweight than national averages (CDC, 2007; NCHS, 2007a). Women in the study also had higher educational attainment and income than the general U.S. population. Conclusion Internet-based programs may help to overcome many of the barriers that women face to participate in physical activity programs. Results of the current study found that an individually tailored, Internet-plus-email physical activity intervention had a positive effect on walking and Moderate-to-vigorous physical activity in an ethnically-diverse group of women. The findings of this pilot study help to address a gap in the literature regarding the utility of Internet-delivered programs targeting women. However, given the lack of research comparable contact in the control group, these findings should be taken cautiously. Conflict of interest statement Dairy Council of California developed the internet-based intervention and funded its evaluation. Genevieve Dunton (of the National Cancer Institute) was responsible for the study design, analysis and interpretation of data, writing the manuscript, and the decision to submit the manuscript for publication. Dairy Council of California had minimal input in these activities.

Acknowledgments We would like to acknowledge Kathryn Juhl for proofreading the manuscript. Genevieve F. Dunton was supported by the Cancer Prevention Fellowship Program, National Cancer Institute during the preparation of this paper. References Ainsworth, B.E., Haskell, W.L., Whitt, M.C., et al., 2000. Compendium of physical activities: an update of activity codes and MET intensities. Med. Sci. Sports Exerc. 32, S498–S504.

Barrera, M., Glasgow, R.E., McKay, H.G., Boles, S.M., Feil, E.G., 2002. Do Internet-based support interventions change perceptions of social support? An experimental trial of approaches for supporting diabetes self-management. Am. J. Commun. Psychol. 30, 637–654. Bryk, A.S., Raudenbush, S.W., 1992. Hierarchical Linear Models: Applications and Data Analysis Methods. Newbury Park, CA: Sage. Centers for Disease Control and Prevention (CDC), 2007. Behavioral Risk Factor Surveillance System Survey Data. U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, Atlanta, Georgia. Cook, R.F., Billings, D.W., Hersch, R.K., Back, A.S., Hendrickson, A., 2007. A field test of a web-based workplace health promotion program to improve dietary practices, reduce stress, and increase physical activity: randomized controlled trial. J. Med. Internet Res. 19, e17. Dannecker, E.A., Hausenblas, H.A., Connaughton, D.P., Lovins, T.R., 2003. Validation of a stages of exercise change questionnaire. Res. Q. Exerc. Sport 74, 236–247. Dinger, M.K., Heesch, K.C., McClary, K.R., 2004. Feasibility of a minimal contact intervention to promote walking among insufficiently active women. Am. J. Health Promot. 20, 2–6. Ginossar, T., 2008. Online participation: a content analysis of differences in utilization of two online cancer communities by men and women, patients and family members. Health Commun. 23, 1–12. Hageman, P.A., Walker, S.N., Pullen, C.H., 2005. Tailored versus standard internetdelivered interventions to promote physical activity in older women. Geriatr. Phys. Ther. 28, 28–33. Haskell, W.L., Lee, I.M., Pate, R.R., et al., American College of Sports Medicine, American Heart Association, 2007. Physical activity and public health: updated recommendation for adults from the American College of Sports Medicine and the American Heart Association. Circulation 116, 1081–1093. Hausenblas, H.A., Dannecker, E.A., Downs, D.S., 2003. Examination of the validity of a stages of exercise change algorithm. J. Appl. Soc. Psychol. 33, 1179–1189. Hopkins, W.G., Wilson, N.C., Russell, D.G., 1991. Validation of the physical activity instrument for the life in New Zealand National Survey. Am. J. Epidemiol. 133, 73–82. Hu, G., Lakka, T.A., Kilpeläinen, T.O., Tuomilehto, J., 2007. Epidemiological studies of exercise in diabetes prevention. Appl. Physiol. Nutr. Metab. 32, 583–595. Humphreys, B.R., Ruseski, J.E., 2007. Participation in physical activity and government spending on parks and recreation. Cont. Econ. Policy 25, 538–552. Kalichman, S.C., Cherry, C., Cain, D., et al., 2006. Internet-based health information consumer skills intervention for people living with HIV/AIDS. J. Consult. Clin. Psychol. 74, 545–554. Leslie, E., Marshall, A.L., Owen, N., Bauman, A., 2005. Engagement and retention of participants in a physical activity website. Prev. Med. 40, 54–59. Napolitano, M.A., Fotheringham, M., Tate, D., et al., 2003. Evaluation of an internetbased physical activity intervention: a preliminary investigation. Ann. Behav. Med. 25, 92–99. National Center for Health Statistics, 2007a. Prevalence of overweight and obesity among adults in the United States. Retrieved on Feb. 15, 2007 from http://www.cdc. gov/nchs/products/pubs/pubd/hestats/3and4/overweight.htm. National Center for Health Statistics, 2007b. Health, United States, 2007 with Chartbook on Trends in the Health of Americans. Hyatsville, MD. Norman, G.J., Zabinski, M.F., Adams, M.A., Rosenberg, D.E., Yaroch, A.L., Atienza, A.A., 2007. A review of ehealth interventions for physical activity and dietary behavior change. Am. J. Prev. Med. 33, 336–345. Palmer, S., Graham, G., Elliott, E., 2005. Effects of a web-based health program on fifth grade children's physical activity knowledge, attitudes, and behavior. Am. J. Health Educ. 36, 86–93. Pew Internet, 2005. Reports: Demographics. Retrieved on July 1, 2008 from http:// www.pewinternet.org/PPF/r/171/report_display.asp. Physical Activity Guidelines Advisory Committee, 2008. Physical Activity Guidelines Advisory Committee Report, 2008. U.S. Department of Health and Human Services, Washington, DC. PLotnikoff, R.C., McGargar, L.J., Wilson, P.m., Loucaides, C.A., 2005. Efficacy of an e-mail intervention for the promotion of physical activity and nutrition behavior in the workplace context. Am. J. Health Promot. 19, 422–429. Prochaska, J.O., 1979. Systems of Psychotherapy: A Transtheoretical Analysis. Dorsey, Oxford, England. Prochaska, J.O., 1995. Changing for Good. Avon Books, Pennsylvania. Rafferty, A.P., Reeves, M.J., McGee, H.B., Pivarnik, J.M., 2002. Physical activity patterns among walkers and compliance with public health recommendations. Med. Sci. Sports Exerc. 34, 1255–1261. Reed, G.R., Velicer, W.F., Prochaska, J.O., Rossi, J.S., Marcus, B.H., 1997. What makes a good algorithm: examples from regular exercise. Am. J. Health Promot. 12, 57–66. Rosenstock, I.M., 1974. Historical origins of the Health Belief Model. Health Educ. Monogr. 2, 328–335. Rosenstock, I.M., Strecher, V.J., Becker, M.H., 1988. Social learning theory and the Health Belief Model. Health Educ. Q. 15, 175–183. Rovniak, L.S., Hovell, M.F., Wojcik, J.R., Winett, R.A., Martinez-Donate, A.P., 2005. Enhancing theoretical fidelity: an e-mail-based walking program demonstration. Am. J. Health Promot. 20, 85–95. Sallis, J.F., Hovell, M.F., Hofstetter, C.R., et al., 1989. A multivariate study of exercise determinants in a community sample. Prev. Med. 18, 20–34. Sallis, J.F., Pinski, R.B., Grossman, R.M., Patterson, T.L., Nader, P.R., 1998. The development of self-efficacy scales for health-related diet and exercise behaviors. Health Educ. Res. 3, 283–292. Simpson, M.E., Serdula, M., Galuska, D.A., et al., 2002. Walking trends among U.S. adults: the Behavioral Risk Factor Surveillance System, 1987–2000. Am. J. Prev. Med. 25, 95–100.

G.F. Dunton, T.P. Robertson / Preventive Medicine 47 (2008) 605–611 Sørensen, M., Gill, D.L., 2007. Perceived barriers to physical activity across Norwegian adult age groups, gender and stages of change. Scand. J. Med. Sci. Sports 18, 651–663. Spittaels, H., De Bourdeauhuij, I., Vandelanotte, C., 2007. Evaluation of a websitedelivered computer-tailored intervention for increasing physical activity in the general population. Prev. Med. 44, 209–217. Thompson, P.D., Buchner, D., Pina, I.L., et al., 2003. Exercise and physical activity in the prevention and treatment of atherosclerotic cardiovascular disease: a statement from the Council on Clinical Cardiology (Subcommittee on Exercise, Rehabilitation,

611

and Prevention) and the Council on Nutrition, Physical Activity, and Metabolism (Subcommittee on Physical Activity). Circulation 107, 3109–3116. U.S. Department of Health and Human Services and U.S. Department of Agriculture, 2005. Dietary Guidelines for Americans, 2005, 6th Edition. U.S. Government Printing Office, Washington, DC. Wilcox, S., Bopp, M., Oberrecht, L., Kammermann, S.K., McElmurray, C.T., 2003. Psychosocial and perceived environmental correlates of physical activity in rural and older African American and white women. J. Gerontol. B. Psychol. Sci. Soc. Sci. 58, P329–P337.