Randomized trial of a neighborhood environment-focused physical activity website intervention

Preventive Medicine 48 (2009) 144–150

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

Randomized trial of a neighborhood environment-focused physical activity website intervention Shannon L. Ferney a,⁎, Alison L. Marshall b, Elizabeth G. Eakin c, Neville Owen c a b c

School of Human Movement Studies, The University of Queensland, St Lucia, 4072, Australia School of Public Health, Queensland University of Technology, Australia Cancer Prevention Research Centre, School of Population Health, The University of Queensland, Australia

a r t i c l e

i n f o

Available online 6 November 2008 Keywords: Physical activity promotion Walking Internet Website Randomized trial

a b s t r a c t Objective. To evaluate the use of a local neighborhood environment-focused physical activity website and its effects on walking and overall physical activity in middle-aged adults. Method. One-hundred and six (72% women) inactive adults aged 52 ± 4.6 years were randomly allocated to receive access to a neighborhood environment-focused website, (Neighborhood group, n = 52) or a motivational-information website (Comparison group n = 54). Participants also received eleven emails over the 26 weeks. Study outcomes were objectively-monitored website use, and self-reported total walking (min/ wk), total physical activity (min/wk) and neighborhood walking (min/wk) collected at baseline, 12 and 26 weeks. The study was conducted between August 2005 and February 2006 in Brisbane, Australia. Results. Website use was significantly greater among Neighborhood participants (p = 0.01). Statistically significant increases in walking and total physical activity were observed in both groups. There was also a statistically significant interaction effect for total physical activity, with Neighborhood group participants maintaining more of their initial increase in physical activity at week-26 (p b 0.05). Further, those in the Neighborhood group who used the website more often reported significantly more walking along the community trail at week-26 (p = 0.05) compared with those who did not. Conclusions. A local neighborhood-environment focused physical activity website was more effective at engaging participants than a motivational-information website. Moreover, its use resulted in meaningful increases in physical activity relative to the comparison website. © 2008 Elsevier Inc. All rights reserved.

Introduction Physical inactivity is a behavioral risk factor for chronic diseases (Bauman et al., 2000). Over 50% of adults in developed countries are not sufficiently active (World Health Organization, 2007). Mass-reach interventions capable of being delivered to large numbers of people at low cost are required. The Internet (websites and email) offers one method of delivering mass-reach interventions. Previous trials have explored the potential of using websites to deliver physical activity (PA) interventions with modest increases in PA reported in most studies (Carr et al., 2008; Marcus et al., 2007; McKay et al., 2002; Napolitano et al., 2003; Spittaels et al., 2007; Steele et al., 2007). In these studies participants who accessed the websites more frequently were more likely to increase their PA (Marcus et al., 2007; McKay et al., 2002). Overall, however, use of these interventions has been low and has declined

⁎ Corresponding author. Fax: +61 7 3365 6877. E-mail address: [email protected] (S.L. Ferney). 0091-7435/$ – see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.ypmed.2008.10.022

substantially over time (McKay et al., 2002; Napolitano et al., 2003; Leslie et al., 2005). Website interventions that can encourage and maintain participant engagement need to be developed (Norman et al., 2007). Participant engagement in website interventions may be enhanced by focusing more explicitly on user expectations and needs (Adelhard and Obst, 1999). This can be done through a user-centred design approach that involves potential users in the design process (Corry et al., 1997; Kinzie et al., 2002). The extent to which user-centered approaches have been used in the past is unknown (Vandelanotte et al., 2007). Most previous website-delivered PA interventions were developed using individually-focused theories of behavior change (e.g., Transtheoretical Model, Marcus and Simkin, 1994 and Social Cognitive Theory, Bandura, 2004). Recent research highlights the potential of developing more environmentally-focused approaches; in particular ‘behavior-setting’ interventions that focus attention on the environment and the context in which specific activities take place (Owen et al., 2004; Humpel et al., 2002; Sallis and Owen, 2002). For example recent research has shown that perceptions of the neighborhood

S.L. Ferney et al. / Preventive Medicine 48 (2009) 144–150

environment can affect walking (Humpel et al., 2004). Therefore, it may be relevant and useful for website interventions to adopt an ‘ecological’ approach to promote PA (walking) within defined settings. The aim of this pilot study was to evaluate the use and efficacy of a PA website that was developed specifically for residents of a local neighborhood community. Website design and content were developed following a user-centered design approach and focused on promoting local opportunities for walking and other activities. Website use was objectively monitored during the trial, and changes in PA assessed pre- and post-intervention. Focus group discussions were conducted following the trial to further elucidate the usefulness of the program. It was hypothesized that participants who received access to the neighborhood environment-focused website would exhibit higher levels of website use and report significantly more walking and overall PA post-intervention, than those who received access to a motivational-information website.

145

Methods Study design This study was a 26-week randomized trial. Outcome data were collected via telephone interviews at baseline (week-0) and again at 12 and 26 weeks post baseline. Trained interviewers were blind to participants' group allocation. Following baseline data collection, participants were randomized to either receive access to the local neighborhood environment-focused website intervention (Neighborhood group; ‘Get Up & Go’) or to a comparison motivational-information website intervention (Comparison group; ‘Active Living Online’) using a single-blind computer generated randomization sequence. Having a comparison website as the ‘control’ is consistent with previous PA intervention studies (King et al.,1998; Dunn et al.,1999). The comparison website was evaluated in a workplace setting, where no significant impact on PA was observed (Marshall et al., 2003a,b). The study was

Fig. 1. Flow diagram of participant recruitment and progress in a randomized trial (n = 106) of a neighborhood environment-focused physical activity website conducted in Australia between August, 2005 and February, 2006.

146

S.L. Ferney et al. / Preventive Medicine 48 (2009) 144–150

The neighborhood environment-focused website was developed for residents of a specific neighborhood. Six suburbs in Brisbane's inner north (Brisbane is Australia's third largest city, with a population of approximately 1.77 million) were selected. These suburbs were chosen as, based on observation, they were deemed relatively ‘high walkable’ in terms of their aesthetics, street connectivity and access to services (Saelens et al., 2003). The area includes mostly detached single-family residences and has a grid-like street pattern with few cul-de-sacs, which is generally indicative of greater street connectivity (Saelens et al., 2003). Further, the selected suburbs all border the same walking trail, which provided a focus for the content of the website.

intensity PA over at least five days were classified as sufficiently active, and those who did not were classified as insufficiently active (Haskell et al., 2007). Consistent with the case for using context- and behavior-specific measures of PA, a specific measure of neighborhood walking was included (Humpel et al., 2004; Owen et al., 2004; Sallis and Owen, 2002). Participants were asked ‘how many times a week do you go for a walk for any reason in and around your neighborhood’ and ‘how much time do you usually spend when you do go for a walk in and around your neighborhood’. This item has acceptable reliability and has been used in a previous trial of a print-delivered walking intervention (Humpel et al., 2005). An additional item asked ‘how many times a week do you walk along the community walking trail’ and ‘how much time do you usually spend when you do go for a walk along the community walking trail’. For both items, a measure of total walking (min/wk) was calculated.

Participant recruitment and eligibility

Intervention procedures

Study setting residents were invited to participate via advertisements in the community newspaper (n = 3) and a letterbox drop (n = 7000), between July and August, 2005. One hundred and fiftyeight residents who responded were screened for eligibility over the telephone (see Fig. 1). Eligible participants (n = 113) were aged between 45 and 60 years, had home Internet access, were able to speak and read English and were not meeting the current PA guidelines (Haskell et al., 2007). Participant's current PA was screened using a brief validated measure of PA (Marshall et al., 2005). Those not meeting the current PA guidelines were considered inactive and invited to participate. Seven participants withdrew before baseline assessment (week-0), leaving 106 participants (see Fig. 1). Despite the initial screening process, 27% of participants were accumulating sufficient PA when they completed the more comprehensive assessment at baseline (Australian Institute of Health and Welfare, 2003).

All participants received a personal email, with a unique username and password, which invited them to access their website using embedded hyperlinks. Both interventions were delivered over three distinct phases characterized by the frequency of email contact. During the ‘intensive’ phase (weeks one to four) participants were sent weekly emails and during the ‘tapered’ phase (weeks five to 12) emails were sent bi-weekly. During the ‘maintenance’ phase emails were sent monthly.

approved by a University Human Research Ethics Committee and was conducted between August 2005 and February 2006. Study setting

Measures Program use and receipt Background programming objectively monitored website use. The database recorded the date and time each participant accessed the website. From these data, total weekly login rates were calculated. Self-report data on recall and use of the respective website programs were collected at week-12. This included ‘unprompted recall’ which asked participants whether they had ‘seen any websites about PA’ and, if yes, ‘to name the website’, and ‘prompted recall’ which asked whether they had seen either the ‘Get Up & Go’ or ‘Active Living Online’ websites. Those who recalled a study website were asked to rate it in terms of layout, information and usefulness on a five-point Likert scale. Unprompted recall of study emails was also assessed. Those who recalled receiving the emails were asked to recall how many they had read and whether they used the hyperlinks to access the website. Physical activity Self-reported walking and PA data were collected via telephone interview at week-0, -12 and -26, using the Active Australia Questionnaire (AAQ) (Australian Institute of Health and Welfare, 2003). The AAQ assesses the total weekly number of sessions and time spent walking for transport and leisure and in moderate- and vigorous-intensity PA and shows good reliability and validity (Brown et al., 2004; Timperio et al., 2004). Data from the AAQ were treated according to standard protocols, which include weighting time reported in vigorous PA by two to account for its greater intensity (Australian Institute of Health and Welfare, 2003). A dichotomous variable based on meeting the criterion of sufficient PA was then created (Haskell et al., 2007). Participants who accumulated at least 150 min of at least moderate-to-vigorous

Neighborhood-environment website Website development Social Cognitive Theory (Bandura, 2004) constructs were integrated into an ecological model (Owen et al., 2000; Sallis and Owen, 2002) to help focus the website on known individual (e.g., selfefficacy, social support) and environmental (e.g., perceived access, convenience and aesthetics) influences on walking and overall PA (Humpel et al., 2002; Owen et al., 2004). To assist with maximizing participant engagement, an iterative user-centered approach guided website development (Corry et al., 1997; Kinzie et al., 2002). This involved focus group discussions and two phases of user testing. Focus group discussions highlighted the need to include interactive features as well as relevant information on local opportunities for PA (Ferney and Marshall, 2006). Twenty-four participants (phase 1: n = 8; phase 2: n = 16) were involved in the user testing. These participants were not involved in the efficacy trial but had similar demographic profiles. Consistent with standard procedures (Dumas and Redish, 1992; Rubin, 1994) participants were monitored while conducting a series of tasks, and then interviewed on the websites usability. The website was revised following each phase of testing. Website content Eleven content areas were created: 1) ‘fact sheets’, provided information on cognitive and behavioral strategies (e.g., goal-setting, rewards, social support); 2) ‘activities’, included examples of stretching exercises, a target heart rate calculator and a progressive walking program; 3) links to relevant websites; 4) an interactive goal-setting and self-monitoring tool; 5) a searchable database of local opportunities for PA; 6) a calendar of events; 7) a map of the walking trail; 8) PA profiles of each suburb; 9) access to individualized email advice; 10) a bulletin-board, and 11) ‘news’ items updated bi-weekly (e.g., reviews of PA opportunities in the neighborhood). The inclusion of the news items was based on findings from previous trials and focus group research which highlighted the importance of continually updating a website to foster engagement (Ferney and Marshall, 2006). Although participants were encouraged to use all features of the website, they were not explicitly directed to do so.

S.L. Ferney et al. / Preventive Medicine 48 (2009) 144–150

The emails sent to participants focused on cognitive and behavioral strategies for increasing PA as well as on opportunities for PA in the local area. The emails encouraged ongoing use of the website content through embedded hyperlinks. All emails sent between weeks one to 12 included a separate four-item quiz assessing a theory-based PA construct (e.g., self-efficacy; barriers; awareness of opportunities) and those who completed a quiz received a tailored email response. Comparison website Comparison group participants received access to ‘Active Living Online’, a PA website with minimal interactivity. This website had four main sections corresponding to the motivational stages of change, and directed sequential movement through these sections (Marshall et al., 2003a,b). These participants also received the same number of emails as the Neighborhood group, but they included non-tailored information on cognitive and behavioral strategies as used in the previous trial (Marshall et al., 2003a,b). Statistical analyses Data analyses were guided by Intention-to-Treat (ITT) principles. Participants' data lost to follow-up were conservatively substituted with baseline values (Newell, 1992). As done in previous trials, Treatment Received (TR) analyses were conducted based on participant recall of the interventions (Marshall et al., 2003a,b; Marshall et al., 2004). To more closely explore the impact of the Neighborhood website, ‘Impact-of-Use’ analyses of the Neighborhood groups' data were also conducted, using a dichotomous category of website use, created using the median split of participant logins (McKay et al., 2002). All quantitative data were entered into and analyzed using SPSS v.11 for Windows. Categorical variables were compared between groups using chi-square tests for independence. For continuous data, differences in mean scores between and within groups were assessed using 3 (time) × 2 (group) ANCOVAS, controlling for BMI. One-way

Table 1 Characteristics of participants (n = 106) in a randomized trial of a neighborhood environment-focused physical activity website conducted in Australia between August, 2005 and February, 2006

Age

Gender Female Occupation Manager/administrator Other Education High school N High school Employment status Full time Other Retirement status Retired Marital status Married/living with partner Single/widowed/divorced Children at home Yes BMIb Normal Overweight/obese a b

χ

2

p

Neighborhood n = 52

Comparison n = 54

Mean ± SD

Mean ± SD

51.7 ± 4.1 n (%)

52.2 ± 5.0 n (%)

0.6

0.56

40 (77)

36 (66)

1.5

0.22

29 (56) 24 (44)

24 (46) 30 (44)

1.1

0.34

14 (27) 39 (75)

18 (33) 36 (67)

0.6

0.53

27 (52) 25 (48)

29 (54) 25 (46)

1.0

0.93

5 (10)

5 (9)

1.0

1.0

36 (69) 17 (33)

39 (72) 15 (26)

0.2

0.68

32 (62)

30 (55)

0.8

0.44

14 (27) 36 (69)

28 (53) 25 (47)

6.6

0.02

Missing data (n = 2). Missing data for self-reported weight (n = 4).

a

147

Table 2 Participant recall and use of the study materials at week-12 in a randomized trial (n = 106) of a neighborhood environment-focused physical activity website conducted in Australia between August 2005 and February 2006 Comparison Neighborhood χ2 n = 49 n = 48 Website Unprompted recall Prompted recall Of those who recalled the website Useful (agree/strongly agree) Layout/presentation appealing (agree/strongly agree) Information interesting (agree/strongly agree) Added website to ‘Favorites’ folder (yes) I did not think the website would be enough to help me become more active (agree/strongly agree) Printed any information (yes) Email Unprompted recall Prompted recall Of those who recalled the emails Useful (agree/strongly agree) Number of emails received (≥ 7) Number of emails read (most/all of them) Followed links to website (often/always)

p

N (%)

N (%)

19 (39) 39 (80)

30 (63) 42 (88)

4.6 0.03 1.0 0.31

10 (21) 27 (56)

20 (41) 38 (78)

2.7 0.07 3.8 0.09

26 (38)

39 (80)

6.3 0.01

21 (44) 21 (44)

23 (47) 12 (24)

6.4 0.09 6.3 0.01

17(35)

25(51)

5.2 0.16

31 (63) 44 (90)

37 (77) 44 (92)

1.6 0.21 0.0 1.00

12 (25) 6 (13) 32 (67) 12 (25)

15 (31) 15 (31) 33 (67) 15 (31)

0.6 5.4 0.2 0.6

0.49 0.03 0.81 0.49

repeated measures ANOVAS with post hoc Scheffe tests were used to determine changes in PA within groups over time. Statistical significance was set at an alpha level of 0.05. Post trial focus group discussions Following the trial, two focus group discussions were conducted with ten Neighborhood participants to explore their experiences with the website. Each discussion was audio recorded and transcribed verbatim. Data were manually analyzed using a thematic analysis technique (Stewart and Shamdasani, 1990). Results Participant characteristics and retention Mean age of the participant sample (n = 106) was 52 (± 4.6) years and most were women (Table 1). More Neighborhood group participants were overweight or obese (69%) compared with Comparison group participants (47%; p = 0.02). There were no other statistically significant baseline differences between the groups (see Table 1). At week-12 and -26, 97 and 93 participants respectively completed the telephone interviews, resulting in an overall retention of 87% (Fig. 1). Participants without follow-up data did not differ from participants who completed the study protocol (data not shown). Receipt and use of program materials Compared with the Comparison group, significantly more Neighborhood group participants: recalled their website, unprompted, (63% vs. 39%; p = 0.03); reported that the website was interesting (80% vs. 38%; p = 0.01); felt that the website was sufficient to help them become active (44% vs. 24%; p = 0.001); and recalled that they received at least seven emails (31% vs. 13%; p = 0.03; see Table 2). Objective website-use data showed that on average, Neighborhood participants logged in to their website nearly three times more than the Comparison participants (8.2 logins ± 9.0 vs. 2.8 logins ± 2.4; p = 0.01). Most of these logins occurred during weeks one to 12.

148

S.L. Ferney et al. / Preventive Medicine 48 (2009) 144–150

Changes in walking and in total physical activity

Fig. 2. Total number of weekly logins to the Neighborhood and Comparison websites during weeks 1 to 12 in a randomized trial (n = 106) of a neighborhood environmentfocused physical activity website conducted in Australia between August, 2005 and February, 2006.

ITT analyses revealed statistically significant time effects for all walking measures and total PA (see Table 3). There was a significant interaction effect for total PA (p b 0.05) with the Neighborhood group maintaining their increase at week-26 (+57.8 min/wk) relative to the Comparison group (+13 min/wk). Pairwise comparisons showed a statistically significant increase in total PA from week-0 to week-26 (p b 0.05) in the Neighborhood group but no significant changes in the Comparison group. Similar findings were observed in the TR analyses thus the data are not shown. At baseline, 35% of participants in the Comparison group were categorized as sufficiently active compared with 19% of Neighborhood group participants: this difference was not statistically significant (p = 0.10). At week-26 there was an absolute increase of 25% (to 55%) in the proportion of sufficiently active Neighborhood group participants compared with a 6% increase (to 41%) in the Comparison group. There were no statistically significant differences between groups at either week-12 or week-26, however McNemar tests showed statistically significant increases in the proportion of Neighborhood group participants being categorized as sufficiently active at week-26 (p = b 0.05) compared with baseline (week-0). This was not observed in the Comparison group (p = 0.45).

During this time Neighborhood participants made an average of 6.3 (±8.2) logins to their website compared with an average of 2.3 (±2.3) made by Comparison participants (p b 0.01). Weekly login rates decreased in both groups, but were higher in the Neighborhood group for all weeks except week-12 (Fig. 2). During the maintenance phase (weeks 13 to 26) use of both websites was low. Neighborhood participants made an average of 1.9 logins to their website during this time compared with 0.2 logins made by Comparison group participants. Due to the low use no further analyses were conducted. Use of the interactive features of the Neighborhood website was low. Thirteen percent of participants (n = 5) regularly used the selfmonitoring tool and 25% emailed the activity counselor. Only one participant posted a message on the bulletin board. Fifty-two percent (n = 22) of participants completed at least one quiz. These participants completed an average of 2.2 (±1.4) quizzes.

Neighborhood group participants who accessed the website less than two times (‘non-users’; n = 31) were compared with those who accessed the website at least two times or more (‘users’; n = 21). There were no differences between ‘users’ and ‘non-users’ in terms of demographic characteristics (data not shown). There were significant time effects for neighborhood walking, total walking and total PA but no between group or interaction effects. There was however, a significant interaction effect for time reported walking along the community walking trail (p = 0.05; Table 4). Post-hoc analysis revealed those classified as non-users reported significantly less walking along the community trail between week-12 to week-26 (p b 0.05), but no change in website ‘users’. This suggests that website ‘users’

Table 3 Walking and physical activity in the Neighborhood (n = 52) and Comparison (n = 54) groups at baseline and 12 and 26 weeks post baseline in a randomized trial (n = 106) of a neighborhood environment-focused physical activity website conducted in Australia between August, 2005 and February, 2006

Table 4 Walking and physical activity in Neighborhood website ‘users’ (n = 21) and ‘non-users’ (n = 31) based on Impact-of-Use analyses in a randomized trial (n = 106) of a neighborhood environment-focused physical activity website conducted in Australia between August, 2005 and February, 2006

Wk-0

Wk-12

Wk-26

Mean (SD)

Mean (SD)

Mean (SD)

Walking anywhere in the neighborhood (min/wk) Comparison 56.7 (54.0) 80.5 (82.8)

72.4 (81.1)

Neighborhood

76.6 (81.0)

59.2 (76.9)

102.5 (106.6)

Walking along the community walking path (min/wk) Comparison 23.6 (39.5) 41.2 (66.5)

35.6 (66.9)

Neighborhood

58.0 (88.1)

45.4 (68.5)

Total walking (min/wk) Comparison 103.8 (116.5)

150.0 (101.0)

108.6 (99.0)

Neighborhood

133.9 (125.9)

108.5 (96.4)

Total physical activity (min/wk) Comparison 194.8 (184.3)

250.0 (173.5)

207.5 (197.3)

Neighborhood

225.0 (240.9)

218.1 (175.7)

a b c

Time. Group. Group × time.

38.5 (72.6)

81.6 (77.9)

160.3 (167.3)

Impact-of-use analyses on changes in walking and in physical activity

Wk-0

Wk-12

Wk-26

Mean (SD)

Mean (SD)

Mean (SD)

Neighborhood walking (min/week) Non-user 63.3 (90.9)

75.5 (67.5)

57.4 (73.9)

User

120.8 (124.2)

89.7 (84.1)

Community walking path (min/week) Non-user 41.2 (90.9)

52.9 (76.3)

25.0 (44.9)

User

61.5 (96.4)

59.2 (78.4)

Total walking (min/week) Non-user 82.9 (88.8)

136.2 (131.9)

95.2 (100.1)

User

132.4 (124.0)

117.4 (94.4)

Total physical activity (min/week) Non-user 142.4 (142.9)

239.1 (241.6)

205.5 (187.8)

User

215.5 (244.0)

226.6 (169.7)

p

0.001a 0.44b 0.73c

0.01a 0.23b 0.88 c

0.001a 0.39b 0.37c

0.04a 0.32b 0.04c

a b c

Time. Group. Group × time.

56.5 (67.2)

36.6 (58.6)

80.8 (71.2)

172.4 (183.3)

p

0.04a 0.30b 0.09c

0.12a 0.49b 0.05c

0.01a 0.80b 0.70c

0.03a 0.67b 0.69c

S.L. Ferney et al. / Preventive Medicine 48 (2009) 144–150

maintained their initial increases in PA up to 26 weeks, while nonusers decreased. Post trial focus group discussions Focus group participants rated the Neighborhood website favorably. Participants reported liking the neighborhood specific content, in particular the updated News items; “I loved them and the fact that you [the website moderator] had walked them and could say there was shade and there were toilets — you knew what to expect before you got there”. Many reported that the neighborhood specific content motivated them to be active; “I did one of the Forest Walks — that is a really nice walk”, and “well I am starting one of those yoga classes on Monday and I have been thinking about Yoga for months and until this [study] have never got around to doing it”. All reported that they read and enjoyed receiving the emails. They liked that the emails provided practical information on opportunities for PA and reported that they were a source of motivation; “there were a couple of times when I was getting slack and it made you think — ‘ok let's go’, I better do something now”, and “just to remind you because life gets too hectic”. Discussion This trial examined the efficacy of a neighborhood environmentfocused PA website in terms of its use, acceptability and ability to encourage residents of a neighborhood community to be more physically active, relative to a motivational-information website. Total PA, total walking and neighborhood walking all increased over time, and there was a statistically significant interaction effect for total PA. This suggests that the Neighborhood environment-focused website was more effective at promoting PA than the motivational-information website. Neighborhood group participants also demonstrated significantly greater use of their website, accessing it nearly three times more, than the Comparison group participants accessed their website. Neighborhood group participants also rated their website more favorably on a number of measures. A specific focus of this study was to develop a more engaging website, since previous evaluations of website-delivered interventions have reported low participant engagement (Marcus et al., 2007; McKay et al., 2002; Napolitano et al., 2003). This involved adopting a user-centered design approach to develop the website that focused on user needs and preferences whilst also promoting local opportunities for PA. It is encouraging therefore that participant's randomized to the new Neighborhood website demonstrated greater use of the program and reported that the environmental focus was appealing. Despite the fact that participants accessed the Neighborhood website more, overall use of its interactive features was lower than anticipated, particularly when compared with other, albeit more intensive, PA websites (McKay et al., 2002; Tate et al., 2001; Tate et al., 2006). This may be explained, in part, by the small number of participants involved in this pilot trial. The relatively lower engagement may also be due to the fact that participants in this study were otherwise healthy and therefore may not have been as motivated to access a website. Consistent with previous trials, use of the Neighborhood website decreased substantially over time (Leslie et al., 2005; Marcus et al., 2007; McKay et al., 2002; Tate et al., 2001). The website built on previous research by providing regular updates; however, it appears this was insufficient to foster continued use. Additional strategies or supplementary media (i.e., print or telecommunications) may be required to enhance participant engagement. Indeed, a recent evaluation of a combined website plus mobile phone PA intervention suggests that the use of supplementary media may hold promise (Hurling et al., 2007). The findings from this study are comparable with earlier trials of website PA interventions that have reported within group increases in

149

PA at 26 weeks, in the absence of between group differences (Marcus et al., 2007; Spittaels et al., 2007; Steele et al., 2007). One of those studies reported that participants who received access to a website plus email program increased their PA by +77 min/wk, compared with +37 min/wk in participants who received access to the website only and +25 min/wk increase in control group participants at 6 months (Spittaels et al., 2007). A well documented advantage of website interventions is their potential to be delivered to large numbers of people at a relatively low cost. The neighborhood environment-focused website examined in this trial was no exception, costing approximately $12,500 AUD to be developed and maintained. Use of the website resulted in a 25% increase in the proportion of participants categorized as sufficiently active compared with an increase of only 6% in the Comparison group at week-26. Another important finding was that not only did Neighborhood group participants increase their PA at week-12 but they also maintained this increase at week-26, relative to Comparison group participants. This increase represented a 40% increase in PA at week-26, which is considered large for a mass-reach intervention (Marcus and Forsyth, 1999). With research indicating that an increase in as little as 30 min of PA a week can have considerable health benefits (Pate et al., 1995), a nearly 60 min/week increase in PA is meaningful from a public health perspective. The lack of a no-treatment control group inhibits the ability to assess the potential impact of other external factors on intervention engagement or PA. For example, the increases in PA seen over time may have been due to seasonal variation. Further, although not statistically significant, the Comparison group participants did report higher levels of PA at baseline. This may have negatively impacted on their potential to increase their PA relative to Neighborhood group participants. The findings are also limited in terms of their applicability to other settings given that the study setting was chosen as it was regarded as a ‘high walkable’ setting. Finally, self-report measures of PA were used. This limitation was partially addressed by using standard tools with established measurement properties. This trial has shown that those who used the neighborhood environment-focused website more frequently also reported more time walking along the community walking trail. This is consistent with previous research that has also demonstrated associations with website use and PA (Marcus et al., 2007; McKay et al., 2002; Tate et al., 2001) and suggests that if use of a website intervention can be maintained, then greater increases are likely. It is possible, however, that those who are already motivated to change their behavior are also more motivated to access a website, and this should be explored in future trials. Conclusions The findings suggest that a neighborhood environment-focused PA website is more engaging than a motivational-information website, and can promote greater increases in PA. As this is the first website intervention to adopt a neighborhood environment approach, further research to explore the effectiveness of this approach in a larger sample is warranted. Conflict of interest The authors declare that there are no conflicts of interest.

Acknowledgments Stuart Fraser programmed the Neighborhood website. Dr Alison Marshall, Professor Neville Owen, and Associate Professor Eva Leslie developed the content of the comparison website. The research was supported by funding from National Health and Medical Research Council of Australia (NHMRC) program grant (#301200) and by a research Infrastructure Grant from Queensland Health. The first

150

S.L. Ferney et al. / Preventive Medicine 48 (2009) 144–150

author was supported by a doctoral scholarship provided by The University of Queensland. References Adelhard, K., Obst, O., 1999. Evaluation of medical Internet sites. Methods Inf. Med. 38, 75–79. Australian Institute of Health and Welfare, 2003. The Active Australia Survey: a guide and manual for implementation, analysis and reporting. Australian Institute of Health and Welfare, Canberra. Bandura, A., 2004. Health promotion by social cognitive means. Health Educ. Behav. 31, 143–164. Bauman, A., Owen, N., Leslie, E., 2000. Physical activity and health outcomes: epidemiological evidence, national guidelines and public health initiatives. Aust. J. Nutr. Diet. 57, 229–232. Brown, W.J., Bauman, A.E., Chey, T., Trost, S.G., Mummery, K., 2004. Comparison of surveys used to measure physical activity. Aus. NZ J. Public Health 28, 128–134. Carr, L.J., Bartee, R.T., Dorozynski, C., Broomfield, J.F., Smith, M.L., Smith, D.T., 2008. Internet-delivered behaviour change program increases physical activity and improves cardiometabolic disease risk factors in sedentary adults: results of a randomized controlled trial. Prev. Med. 46, 431–438. Corry, M.D., Frick, T.W., Hansen, L., 1997. User-centered design and usability testing of a web site: an illustrative case study. Educ. Tech. Res. Dev. 45, 65–76. Dumas, J., Redish, J., 1992. A Practical Guide to Usability Testing. Intellect, Exeter. Dunn, A.L., Marcus, B.H., Kampert, J.B., Garcia, M.E., Kohl, H.W., Blair, S.N., 1999. Comparison of lifestyle and structured interventions to increase physical activity and cardiorespiratory fitness: randomized trial. JAMA 281, 327–334. Ferney, S.L., Marshall, A.L., 2006. Website physical activity interventions: preferences of potential users. Health Educ. Res. 21, 560–566. Haskell, W.L., Lee, I.M., Pate, R.R., et al., 2007. Physical activity and public health: updated recommendation for adults from the American College of Sports Medicine and the American Heart Association. Med. Sci. Sports Exerc. 39, 1423–1434. Humpel, N., Owen, N., Leslie, E., 2002. Environmental factors associated with adults' participation in physical activity. Am. J. Prev. Med. 26, 119–125. Humpel, N., Marshall, A.L., Leslie, E., Bauman, A., Owen, N., 2004. Changes in neighborhood walking are related to changes in perceptions of environmental attributes. Ann. Behav. Med. 27, 60–67. Humpel, N., Owen, N., Iverson, D., Leslie, E., Bauman, A., 2005. Perceived environment attributes, residential location, and walking for particular purposes. Am. J. Prev. Med. 27, 60–67. Hurling, R., Catt, M., De Boni, M., William Fairley, B., et al., 2007. Using internet and mobile phone technology to deliver an automated physical activity program: randomized controlled trial. J. Med. Internet Res. 9, e7. King, A., Rejeski, D., Buchner, 1998. Physical activity interventions targeting older adults. A critical review and recommendations. Am. J. Prev. Med. 15, 316–333. Kinzie, M.B., Cohn, W.F., Julian, M.F., Knaus, W.F., 2002. A user-centered model for web site design. J. Am. Med. Inform. Assoc. 9, 320–330. Leslie, E., Marsall, A.L., Owen, N., Bauman, A.E., 2005. Engagement and retention of participants in a physical activity website. Prev. Med. 40, 54–59. Marcus, B.H., Forsyth, L.H., 1999. How are we doing with physical activity? Am. J. Health Promot. 14, 118–124. Marcus, B.H., Simkin, L.R., 1994. The Transtheoretical Model: application to exercise behavior. Med. Sci. Sport Exer. 26, 1400–1404. Marcus, B.H., Lewis, B., Williams, D.M., et al., 2007. A comparison of Internet and printbased physical activity interventions. Arch. Intern. Med. 167, 944–949.

Marshall, A.L., Leslie, E.R., Bauman, A.E., Marcus, B.H., Owen, N., 2003a. Print versus website physical activity programs: a randomised trial. Am. J. Prev. Med. 25, 88–94. Marshall, A.L., Bauman, A.E., Owen, N., Booth, M.L., Crawford, D., Marcus, B.H., 2003b. Population-based randomized controlled trial of a stage-targeted physical activity intervention. Ann. Behav. Med. 25, 194–202. Marshall, A.L., Bauman, A.E., Owen, N., Booth, M.L., Crawford, D., Marcus, B.H., 2004. Reaching out to promote physical activity in Australia: a state-wide randomized controlled trial of a stage-targeted intervention. Am. J. Health Prom. 18, 283–287. Marshall, A.L., Smith, B.J., Bauman, A.E., Kaur, S., 2005. Reliability and validity of a brief assessment for use by family doctors. Br. Med. J. 39, 294–297. McKay, H.G., King, D., Eakin, E.G., Seeley, J.R., Glasgow, R.E., 2002. The diabetes network Internet-based physical activity intervention: a randomised pilot study. Diabetes Care 24, 1328–1334. Napolitano, M.A., Fotheringham, M.J., Tate, D., et al., 2003. Evaluation of an Internetbased physical activity intervention: a preliminary investigation. Ann. Behav. Med. 25, 92–99. Newell, D.J., 1992. Intention-to-treat analysis: implications for quantitative and qualitative research. Int. J. Epi. 21, 837–841. 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. Owen, N., Leslie, E., Salmon, J., Fotheringham, M., 2000. Environmental determinants of physical activity and sedentary behavior. Exerc. Sport Sci. Rev. 28, 153–158. Owen, N., Humpel, N., Leslie, E., Bauman, A.E., Sallis, J.F., 2004. Understanding environmental influences on walking: review and research agenda. Am. J. Prev. Med. 27, 67–76. Pate, R.P., Pratt, M., Blair, S.N., et al., 1995. Physical activity and public health: a recommendation from the Centers for Disease Control and Prevention and the American College of Sports Medicine. JAMA 273, 402–407. Rubin, J., 1994. Handbook of Usability Testing: How to Plan, Design and Conduct Effective Tests. John Wiley, New York. Saelens, B.E., Sallis, J.F., Frank, L.D., 2003. Environmental correlates of walking and cycling: findings from the transportation, urban design, and planning literatures. Ann. Behav. Med. 25, 80–91. Sallis, J.F., Owen, N., 2002. Ecological models of health behavior. In: Glanz, K., Rimer, B.K., Marcus Lewis, F. (Eds.), Health Behavior and Health Education: Theory, Research and Practice. Jossey-Bass, San Francisco, pp. 462–484. Spittaels, H., De Bourdeaudhuij, I., Brug, J., Vandelanotte, C., 2007. Effectiveness of an online computer tailored physical activity intervention in a real-life setting. Health Educ. Res. 22, 385–396. Steele, R., Mummery, K., Dwyer, T., 2007. Using the Internet to promote physical activity: a randomised trial of intervention delivery modes. J. Phys. Act. Health 4, 245–260. Timperio, A., Salmon, J., Rosenberg, M., Bull, F.C., 2004. Do logbooks influence recall of physical activity validation studies? Med. Sci. Sports Exerc. 36, 1181–1186. Stewart, D., Shamdasani, P., 1990. Focus Groups: Theory and Practice. Sage Publications, London. Tate, D.F., Wing, R.R., Winett, R.A., 2001. Using Internet technology to deliver a behavioral weight loss program. JAMA 285, 1171–1177. Tate, D.F., Jackvony, E.H., Wing, R.R., 2006. A randomized trial comparing human e-mail counseling, computer-automated tailored counseling and no counseling in an Internet weight loss program. Arch. Intern. Med. 166, 1620–1625. Vandelanotte, C., Spathonis, K.M., Eakin, E.G., Owen, N., 2007. Website-delivered physical activity interventions. A review of the literature. Am. J. Prev. Med. 33, 54–64. World Health Organisation, 2007. Move for Health Information Sheet. Sedentary lifestyle: a global problem. Available at: http://who.com/int. Accessed May, 2007.