Approximately 56% of Canadians are

Transcription

1 Characteristics of Participants Visiting the Canada on the Move Website Ronald C. Plotnikoff, PhD 1 Adrian Bauman, MBBS, PhD 5 John C. Spence, PhD 2 Scott A. Lear, PhD 6 Leonor S. Tavares, MSc 3 Linda McCargar, RD, PhD 7 Liza S. Rovniak, PhD 4 ABSTRACT Background: The Internet is a potential medium to attain large groups of Canadians for physical activity monitoring and interventions, however, little is known about the characteristics of participants who could be recruited on a national level. Our objectives were to determine: 1) the extent to which our sample was representative of the Canadian population; 2) the demographic, social-cognitive and physical activity behaviour, and environmental characteristics of one-time and multi-time users of the Canada on the Move website; and 3) whether certain recruitment strategies were more likely to reach specific demographic subgroups. Methods: Web-based self-report measures were collected on demographic, socialcognitive, physical activity behaviour, and environmental factors. Descriptive statistics were employed to address the study s research objectives. Results: Significant differences (p<0.001) were found between our total sample and 2001 Canadian census/canadian Community Health Survey (CCHS) data on all demographic and physical activity behaviour measures. One-time and multi-time users significantly differed on age (p<0.005) and being responsible for children under age 13 (p=0.01). No differences existed between the groups on any of the social-cognitive, behavioural or environmental variables. Source of knowledge about the Canada on the Move website between one-time and multi-time users was not significantly different. Conclusion: The Canada on the Move website presents a viable approach to reaching diverse demographic groups. Further work needs to be undertaken to: 1) develop engaging websites; 2) detail the monitoring of the web-based access information; 3) integrate the website with other organizations promoting physical activity; and 4) expand the number of recruitment sources. MeSH terms: Demographics; environment; social-cognitive; steps; physical activity; webbased interventions 1. Professor, Centre for Health Promotion Studies, Alberta Centre for Active Living & Faculty of Physical Education and Recreation, University of Alberta, Edmonton, AB 2. Associate Professor, Faculty of Physical Education and Recreation, University of Alberta 3. Research Assistant, Centre for Health Promotion Studies, University of Alberta 4. Adjunct Research Assistant Professor, Center for Behavioral Epidemiology & Community Health, Graduate School of Public Health, San Diego State University, San Diego, CA, USA 5. Sesquicentenary Professor of Public Health, School of Public Health, University of Sydney, Sydney, New South Wales, Australia 6. Assistant Professor, School of Kinesiology, Simon Fraser University, Burnaby, BC 7. Professor, Department of Agriculture, Food and Nutrition Science, University of Alberta Correspondence: Dr. Ronald Plotnikoff, Centre for Health Promotion Studies, University of Alberta, 5-10 University Extension Centre, Street, Edmonton, AB T6G 2T4, ron.plotnikoff@ualberta.ca Acknowledgements: The authors thank Stephanie Brunet and Sonia Lippke. Sources of support: The research was funded by the Institute of Nutrition, Metabolism, and Diabetes Canadian Institutes of Health Research and the University of Alberta. Dr. Plotnikoff is supported by Salary Awards from the Alberta Heritage Foundation for Medical Research and the Canadian Institutes of Health Research. Approximately 56% of Canadians are physically inactive and 48% are overweight or obese. 1 Developing strategies to monitor and increase Canadians physical activity levels is therefore a public health priority. The Internet has great potential as a cost-effective method for large-scale physical activity monitoring and interventions. 2 In 2003, 55% of Canadian households had home Internet access, with a steady annual increase of approximately 7%. 3 Although Internet-based physical activity interventions could potentially reach large groups of inactive and active Canadians, little is known about the characteristics of participants who could be recruited on a national level. Most Internet-based, physical activity or weightloss interventions have been conducted with small (<100), self-selected, highly educated samples of participants 4-13 with the exception of two recent intervention studies in Australia and Canada 2 that included approximately 300 and 2,500 adults, respectively. No published evaluations exist on the capacity of nationwide Internet-based physical activity web platforms to recruit and retain population samples. Moreover, the feasibility of Internet surveys for assessing nationwide physical activity levels and related socialcognitive variables has not been explored. Internet surveys, however, have been successfully used in other national data collection efforts 14 and recent studies have reported that this mode of data collection can provide reliable data. 15 This article is based on a national initiative aimed at increasing the Canadian population s awareness and participation in regular physical activity using pedometers through a web-based platform. The unique natural experiment included, in part, more than 800,000 pedometers initially distributed between December 2003 and early 2004 in cereal boxes (Kellogg s* Special K*) for the purposes of monitoring and reinforcing individuals efforts to increase their amount of daily activity. A further 1.2 million pedometers were distributed in cereal boxes commencing in June The cereal boxes, along with other public-reach modalities, provided a written message about the Canada on the Move (COTM) website and invited participants to donate their step counts to science. The aim was to raise positive atti- S28 CANADIAN JOURNAL OF PUBLIC HEALTH VOLUME 97, SUPPLEMENT 1

2 tudes towards physical activity and to support individuals in increasing their daily steps. The purpose of the present study was to evaluate the extent to which the COTM initiative recruited and retained website users. Specifically, the objectives were to determine: 1) the extent to which our sample was representative of the Canadian population; 2) the demographic, socialcognitive, behavioural and environmental characteristics of one-time and multitime users of the COTM website; and 3) whether certain recruitment sources were more likely to reach specific demographic subgroups. METHODS Developmental work A pilot study was conducted with convenience samples of 98 community adults and 80 undergraduate students to: 1) assess the validity and reliability of the COTM social-cognitive constructs; and 2) determine the average length of time required to complete the questionnaire in a pen and paper format. Participants from both samples were also asked to identify questionnaire items requiring improvement in content and/or clarity. To assist the validation process, the student sample was provided with pedometers and asked to wear them for a five-day period before being administered the questionnaire. The pilot study results demonstrated the measures validity and reliability and allowed the researchers to identify and improve item content and clarity. A final convenience sample of 18 adults (including 2 study-team members) completed an instrument assessing: 1) the clarity of the website content question items and instruction sets; and 2) the userfriendliness of the website and time required to read and complete its various components. Results indicated the website was easy to use and provided adequate information on pedometers. Several structural and content changes were suggested and subsequently made to the website. Website structure and access/navigation protocols An independent agency and select members of the study team developed the website for participant access and datacollection protocols. The site was divided into two sections: 1) a public area containing information about the study (i.e., project goals, information about pedometers, and frequently asked questions), with a tab for participants to sign up and for registered participants to log-in; and 2) a secure area where registered participants could complete questionnaires, record their steps and review their individual progress toward step goals. Participants were asked to agree to the conditions specified in a consent form before registering in the study. Individuals had to be 14 years of age or older to sign in and complete the survey. Following consent, participants were required to indicate a username, password, postal code, province and whether they owned a pedometer. Ownership of a pedometer from the cereal box distribution was not required to take part in the study; however, participants needed a pedometer to enter and track their steps on the website. Registered participants could edit personal information at any point during the study. Drop-down tabs (i.e., pre-selected options) were formatted for participants entering their demographic, socialcognitive, behavioural and environmental support information. Radio buttons (i.e., clicking a circle to place a mark for the correct response) were used to identify sex, while text boxes (i.e., manually typing in information) were used to enter height and weight. A step calendar was also formatted for participants to enter in their steps. The calendar was displayed in weeks according to the calendar months. Midway through the study, participants were able to include steps for each day of the month. Personal goals and strategies (e.g., joining walking groups) could also be included in the step calendar. Visual graphs of steps entered were a further option for participants to help monitor their progress. The COTM website ( was officially opened on December 15, This article reports the characteristics of individuals accessing the site up until October 16, MEASURES Demographic characteristics Demographic data were collected on sex, age, education, ethnicity and provincial location based on questions and categories from Statistics Canada. 16,17 Participants were asked to indicate their height (in centimeters, feet or inches) and weight (in kilograms or pounds) to assess body mass index (BMI). Cut-points for underweight, healthy weight and overweight were then categorized based on the Canadian Community Health Survey (CCHS). 18 Participants indicated whether they were responsible for the care of children under 13 years of age who live with them with either a yes or no response. If yes, a tab was included to specify how many. Social-cognitive characteristics Participants were asked to identify their confidence in accumulating an additional 2,000 steps per day for the next 3 months (i.e., general self-efficacy), using a 1-item 5-point Likert-type scale with response options ranging from no confidence (1) to completely confident (5). 19 Participants were also asked to reflect on their confidence in accumulating these additional steps when tired, in a bad mood, and when there is bad weather (i.e., threeitem, coping efficacy measure). 19 Similarly, participants identified their confidence level in incorporating these additional steps into a daily routine and into a schedule (i.e., two-item, scheduling efficacy measure). 19 The intention to increase the number of steps taken per day by 2,000 steps over the next 3 months was measured using a 1-item 5-point Likert-type scale ranging from definitely not (1) to definitely yes (5). 20 Receiving support for walking on a regular basis from the people closest to the participant was measured using a 1-item 5-point Likert-type scale ranging from no support (1) to a great deal of support (5). 21 Three pairs of parallel items assessing the likelihood and desirability of completing an additional 2,000 steps daily for the next 3 months to increase fitness (i.e., fitness expectancy), improve health (i.e., health expectancy), and lose and/or control weight (i.e., weight expectancy) were measured using 5-point Likert-type scales based on measures developed by Sallis et al. 22 The product for the 3 sets of parallel items were calculated to create a scale mean (a range of 1 to 25). The response options for the respective scales ranged from not at all likely (1) to extremely MARCH APRIL 2006 CANADIAN JOURNAL OF PUBLIC HEALTH S29

3 likely (5), and not at all desirable (1) to extremely desirable (5). Cronbach alpha coefficients for the multi-item, social-cognitive scales were: coping efficacy (α=0.83), likelihood (α=0.86) and desirability (α=0.79). The correlation for the two scheduling efficacy items was r=0.81. Physical activity Physical activity behaviour was measured with the short format of the International Physical Activity Questionnaire (IPAQ). 23 Participants were asked to record the amount of days and time spent walking during the last 7 days, for at least 10 minutes at a time. Both vigorous (e.g., lifting, digging, aerobics or fast bicycling) and moderate activities (e.g., light loads, bicycling at a regular pace, doubles tennis) were also assessed for frequency and duration. A total physical activity score was calculated by combining walking with vigorous and moderate activity metabolic equivalent (MET) minutes. Physical activity categories of active, minimally active and inactive were subsequently formed using a combination of calculations as specified in the short version of the IPAQ. 23 Environmental support Environmental support for walking was measured with four-point Likert-type scales 24 ranging from strongly disagree to strongly agree. These measures are detailed in this issue by Spence et al. 25 RESULTS Descriptive statistics and chi-square analyses were conducted to answer the study s research questions. Participants who visited the website and registered were considered one-time users (n=2,709). Multi-time users (n=466) were those who visited the website more than once (i.e., logged on to the website, completed their initial questionnaire information and/or follow-up questionnaires on social-cognitive and behavioural measures prompted at one, three and six months). Eighty-two percent of these multi-time users completed at least one of the follow-up questionnaires. (For the purposes of this article, participants step information was not considered in determining user status.) Table I presents the characteristics of the sample who completed the survey (N=3,175). Census data and/or the CCHS statistics are presented for existing, comparable measures. Significant differences (p<0.001) were found between our total sample and 2001 Canadian census/cchs data on all demographic characteristics and on physical-activity behaviour. Notably, 77% of the sample were females, versus 50.5% in the Canadian population (χ 2 =742.50, p<0.001). Twenty-seven percent in our sample, versus 63.5% of the census population, cared for children under 13 years of age (χ 2 = , p<0.001). In comparison to the CCHS, there was a greater proportion of overweight people in our study (53% versus 31.9%; χ 2 =334.67, p<0.001). Further, physical activity levels were higher in our sample (38.6% were considered active versus 21% in the CCHS sample: χ 2 = , p<0.001). Also presented in Table I are the characteristics of one-time and multi-time users in our sample. These groups significantly differed in age (χ 2 =17.09, p<0.005) and being responsible for children under 13 years of age (χ 2 =6.20, p=0.01). No differences existed between one-time and multitime users on any of the social-cognitive, behavioural or environmental variables. Table II provides a demographic breakdown of the source of learning about the website s existence. Significant differences were found for age (χ 2 =58.69, p<0.001), gender (χ 2 =45.58, p<0.001), education (χ 2 =63.58, p<0.001) and BMI (χ 2 =26.55, p<0.01). Source of website knowledge between one-time and multi-time users, however, was not significantly different (χ 2 =2.59, p=0.46). Table III provides the demographic comparison between the one-time and multi-time users for each of the four listed sources. No differences existed for age, sex or education between the two groups for any of the four sources (although p=0.05 was reported for age in the media source). Significant differences were reported for two of the sources for BMI: a greater proportion of one-time users were overweight who surfed/linked in from another site (χ 2 =15.76, p<0.01); and the multi-time users in the word of mouth source appeared to have greater BMI levels (χ 2 =8.67, p=0.03). DISCUSSION The present study is the first to examine demographic characteristics of participants visiting a physical activity website promoted on a national level. We aimed to determine the characteristics (demographic, social-cognitive, physical activity behaviour and environmental) of one-time and multi-time users of the COTM website. Overall, 3,173 people were recruited in the first 10 months. Although this is an extremely small proportion of the Canadian population, this initiative demonstrates the promise of a national website system for collecting demographic and health-related data from a large number of individuals. Not surprisingly, this initiative mainly attracted overweight, middle-aged women (the primary target and purchasing group of Special K* cereal). Although this is one of the at-risk groups for physical inactivity, 1 it is important to widen recruitment strategies to include greater proportions of other population subgroups for future population-based website interventions. Although our sample was biased toward females (77%) and overweight people (53%), this may point to the capacity of website systems to reach those most in need of physical activity interventions. 26 Craig et al. report that between 1995 and 2000, the proportion of physically active women did not increase, despite increases in physical activity among men during the same time period in Canada. 27 These findings highlight the need for interventions that reach women. Furthermore, Katzmarzyk and Ardern 28 found that overweight, obesity and deaths attributable to overweight and obesity increased in Canada between 1985 and 2000, highlighting the need for interventions that reach those in the population who are overweight. It is also possible that there was some perception of the website being an aid for weight loss. Since mostly women report an ongoing desire to lose weight, this may further explain the large number of women in our sample. While most of our sample was overweight, the majority were also more physically active than the rest of the Canadian population (81%). This finding parallels the study conducted by Craig et al., which reported that 80% of the study s interna- S30 CANADIAN JOURNAL OF PUBLIC HEALTH VOLUME 97, SUPPLEMENT 1

4 TABLE I Characteristics of the Sample Completing the Web-based Survey* Characteristics One-time Users Multi-time Users χ 2 (p) Total (N=3175) Census 2001 χ 2 (p)** CCHS 2001 χ 2 (p) n=2709 (85%) n=466 (15%) n (%) n (%) n (%) n (%) Demographics Age, (0.002) (<0.001)*** (1.9) 6 (1.3) 47 (1.8) (9.0) 22 (5.0) 212 (8.2) 2,223,200 (7.0) (47.6) 191 (43) 1197 (46.7) 9,535,400 (29.9) (37.1) 204 (45.8) 988 (38.5) 8,192,600 (25.6) (4.5) 22 (4.9) 119 (4.6) 4,141,000 (13.1) Total 2118 (100) 445 (100) 2563 (100) Sex 0.01 (0.91) (<0.001) Male 505 (23.1) 102 (22.6) 607 (23) 15,816,500 (49.5) Female 1684 (76.9) 349 (77.4) 2033 (77) 16,129,800 (50.5) Total 2189 (100) 451 (100) 2640 (100) Education (0.55) (<0.001) Less than high school 45 (2.1) 8 (1.8) 53 (2.0) 7,935,075 (33.2) High school/trades certificate or diploma/college 1072 (48.8) 210 (46.4) 1282 (48.4) 11,677,210 (48.9) University certificate/ diploma/degree 1078 (49.1) 235 (51.9) 1313 (49.6) 4,289,070 (17.9) Total 2195 (100) 453 (100) 2648 (100) Ethnicity,, Aboriginal 23 (1.1) 9 (2.0) 32 (1.2) 976,305 (3.3) African 4 (0.2) 0 (0.0) 4 (0.2) 137,315 (0.5)**** Asian 86 (4.0) 13 (2.9) 99 (3.8) 2,931,865 (9.9) Black 6 (0.3) 3 (0.7) 9 (0.3) 662,210 (2.2) Canadian 1047 (48.9) 212 (47.9) 1259 (48.8) 11,682,680 (39.4) European 85 (4.0) 17 (3.8) 102 (3.9) 3,742,890 (13.1)**** Latin American 10 (0.5) 2 (0.5) 12 (0.5) 216,980 (0.7) White 828 (38.7) 176 (39.7) 1004 (38.8) Other 50 (2.3) 11 (2.5) 61 (2.4) Total 2139 (100) 443 (100) 2582 (100) BMI 0.86 (0.83) (<0.001) <20 (underweight) 51 (4.0) 13 (5.2) 64 (4.2) 1,510,478 (8.1) (acceptable weight) 343 (26.7) 65 (25.9) 408 (26.5) 8,008,426 (42.9) (some excess weight) 200 (15.6) 41 (16.3) 241 (15.7) 2,908,975 (15.6) >27 (overweight) 682 (53.0) 132 (52.6) 814 (53.0) 5,953,296 (31.9) Total 1286 (100) 251 (100) 1527 (100) Province,, ***** (0.14) (0.001) (<0.001) Alberta 306 (11.3) 63 (13.7) 369 (11.6) 3,201,900 (10.0) 14,200 (10.7) British Columbia 406 (15.0) 87 (18.9) 493 (15.6) 4,196,400 (13.1) 18,090 (13.7) Manitoba 86 (3.2) 18 (3.9) 104 (3.3) 1,170,300 (3.7) 8000 (6.0) New Brunswick 78 (2.9) 11 (2.4) 89 (2.8) 751,400 (2.4) 5150 (3.9) Newfoundland and Labrador 37 (1.4) 5 (1.1) 42 (1.3) 517,000 (1.6) 4010 (3.0) Northwest Territories 9 (0.3) 9 (0.3) 42,800 (0.1) 900 (0.7) Nova Scotia 157 (5.8) 31 (6.7) 188 (5.9) 937,000 (2.9) 5040 (3.8) Ontario 1155 (42.7) 189 (41.0) 1344 (42.3) 12,392,700 (38.8) 42,260 (31.9) Prince Edward Island 21 (0.8) 3 (0.7) 24 (0.8) 137,900 (0.4) 2000 (1.5) Quebec 371 (13.7) 42 (9.1) 413 (13.0) 7,542,800 (23.6) 24,280 (18.3) Saskatchewan 74 (2.7) 11 (2.4) 85 (2.7) 995,400 (3.1) 7720 (5.8) Yukon 7 (0.3) 1 (0.2) 8 (0.3) 31,200 (0.1) 850 (0.6) Total 2707 (100) 461 (100) 3168 (100) Responsible for the Care of Children Under (0.01) (0.001) Yes 605 (27.6) 99 (22.0) 704 (26.6) 5,311,795 (63.5) No 1589 (72.4) 352 (78.0) 1941 (73.4) 3,059,225 (36.5) Total 2194 (100) 451 (100) 2645 (100) MARCH APRIL 2006 CANADIAN JOURNAL OF PUBLIC HEALTH S31

5 Characteristics One-time users Multi-time users χ 2 (p) Total (N=3175) Census 2001 χ 2 (p) CCHS 2001 χ 2 (p) n=2709 (85%) n=466 (15%) n (%) n (%) n (%) n (%) Behaviours Leisure-time physical activity 0.17 (0.92) (<0.001) Active 810 (38.6) 171 (39.1) 982 (38.6) 5,423,162 (21.0) Minimally active 895 (42.7) 188 (43.0) 1089 (42.8) 5,577,001 (21.6) Physically inactive 392 (18.7) 78 (17.8) 472 (18.6) 12,661,729 (49.1) Mean (SD) Mean (SD) t (p) Mean (SD) Socio-cognitions General self-efficacy 3.89 (1.02) 3.89 (1.04) (0.93) 3.89 (1.02) Coping efficacy 3.22 (1.03) 3.27 (1.04) (0.37) 3.23 (1.03) Scheduling efficacy 3.64 (1.01) 3.66 (1.03) (0.63) 3.64 (1.01) Fitness expectations (5.60) (5.66) (0.88) (5.61) Health expectations (5.70) (5.46) 0.44 (0.66) (5.65) Weight expectations (6.50) (6.48) 1.51 (0.13) (6.51) Intention 4.27 (0.76) 4.33 (0.71) (0.12) 4.28 (0.75) Environment/Neighbourhood Shops close by 2.74 (1.27) 2.79 (1.25) (0.48) 2.75 (1.26) Transit close by 2.88 (1.33) 2.84 (1.35) 0.61 (0.54) 2.87 (1.33) Sidewalks close by 3.25 (1.24) 3.21 (1.22) 0.61 (0.54) 3.24 (1.23) Crime rate 1.78 (0.96) 1.81 (0.95) (0.43) 1.78 (0.96) Recreation facilities available 3.01 (1.04) 3.04 (1.01) (0.48) 3.01 (1.04) Interesting scenery 3.02 (0.89) 3.07 (0.87) (0.23) 3.03 (0.89) Many four-way intersections 2.74 (1.17) 2.75 (1.12) (0.84) 2.74 (1.16) Many places to go that are easy walking distance 2.78 (1.05) 2.87 (1.00) (0.07) 2.79 (1.04) * Note: percent of missing data for the total sample ranged between 16.2% and 33.4% for all variables excluding province (0.2%), physical activity (5.2%) and BMI (51%). Multi-time users reported less missing cases across all variables than one-time users (a range of 1.0%-20.7%). Participants who had visited the website only once by at least registering their login ID name and password were considered one-time users. Those who had registered and visited the website more than once were considered multi-time users. Chi-square test/t-test compares one-time users to multi-time users of the website. ** Chi-square test compares the total sample to Census data. Chi-square test compares the total sample to CCHS data. Numbers are represented by those who reported both age and gender for census data. Census data proportions reflect Canadians status only on the selected categories. *** Ages were not included in analysis. Data represent ages Ethnic categories identified on the website were collapsed for this article as follows: Arab, South Asian, South East Asian, Japanese, Korean, Chinese, Filipino were collapsed into the Asian category, and South American was collapsed into the Latin American category. Could not be computed due to a low cell frequency. **** 1996 Census data. Categories were not possible to collapse due to the ambiguity of ethnic origin from participants who identified themselves as either Canadian or white. BMI categories are based on Canadian standard; CCHS numbers reflect individuals between ages Census and CCHS proportions are different due to CCHS over-sampling smaller provinces. ***** Chi-square analysis between one-time and multi-time users did not include: Northwest Territories, Prince Edward Island and Yukon. Physical-activity level data for the IPAQ and the CCHS reflect different measures. CCHS follows Health Canada guidelines. S32 CANADIAN JOURNAL OF PUBLIC HEALTH VOLUME 97, SUPPLEMENT 1

6 TABLE II Demographic Characteristics of Sample and Source of Hearing about the Website tional population were categorized as active in their development of the IPAQ surveillance measure. 23 Three potential issues can be considered here: 1) the IPAQ measures multiple domains of activity such as walking, household chores and leisure time, which can lead to higher prevalence rates when compared to other measures; 2) although individuals may be overweight, they may also be achieving recommended physical activity levels; and 3) we expected those who were currently active to selfselect themselves into the study since these people were already interested in physical activity. To date, the literature examining multiand single-time users in this electronic mode of communication has been scant. Overall, we found little difference between one-time and multi-time users. Both groups were very similar on demographic, attitudinal, behavioural and environmental characteristics. One notable finding was that multi-time users had considerably less missing data than one-time users (see footnote for Table I), reflecting a self-selection and interest bias for the multi-time user group. The majority of people used the COTM website once; only 15% were multi-time users. Such limited website engagement is similar to findings reported in a recent study where website use declined over time MARCH APRIL 2006 Back of Surfed/Linked Word Media: Television, a Cereal Box from Another of Mouth Magazine, N=627 (24%) Website N=559 Newspaper, Radio N=410 (16%) (21%) N=1052 (40%) n (%) Age (<0.001) (2.5) 5 (1.3) 11 (2.0) 13 (1.3) (7.4) 47 (11.8) 44 (8.1) 73 (7.4) (51.2) 202 (50.8) 270 (50.0) 399 (40.4) (34.2) 133 (33.4) 203 (37.6) 435 (44.0) (4.6) 11 (2.8) 12 (2.2) 68 (6.9) Total 605 (100) 398 (100) 540 (100) 988 (100) Sex (<0.001) Male 98 (15.8) 87 (21.4) 111 (20.1) 304 (29.5) Female 521 (84.2) 320 (78.6) 440 (79.9) 727 (70.5) Total 619 (100) 407 (100) 551 (100) 1031 (100) Education (<0.001) Less than high school 17 (2.7) 9 (2.2) 10 (1.8) 16 (1.5) High school/trades certificate or diploma/college 365 (59.0) 164 (40.1) 218 (39.6) 524 (50.2) University certificate/ degree 237 (38.3) 236 (57.7) 323 (58.6) 503 (48.2) Total 619 (100) 409 (100) 551 (100) 1043 (100) BMI (<0.01) <20 11 (4.0) 7 (4.0) 14 (5.8) 32 (3.9) (32.2) 39 (22.4) 71 (29.5) 205 (24.7) (15.4) 40 (23.0) 47 (19.5) 112 (13.5) > (48.4) 88 (50.6) 109 (45.2) 482 (58.0) Total 273 (100) 174 (100) 241 (100) 831 (100) χ 2 (p) due to possible constraints on the use of s and websites in delivering health behaviour change programs. 12 Suggested strategies to increase engagement of participants in future website initiatives include activities such as: integrating web-based systems with other social/environmental support systems (e.g., walking groups, workplace physical activity policies, group lists/sign-ups); and providing more specific and tailored goal-setting, self-monitoring and feedback. 6 Such approaches may enable researchers to recruit and retain larger, more representative samples in future interventions. Indeed, once recruited, participant commitment can be achieved as reflected by the finding in our study among multi-time users: approximately 5% of multi-time users visited the website more than twice. Most of our sample learned about the website through media sources. These results are comparable with Canada s Physical Activity Monitor describing the sources that provide Canadians with most of their information about physical activity. 26 Canadians report television (42%) as the source where they have seen or heard information about physical activity most often, with 12% reporting newspaper, 5% from the Internet, 3% from the radio, and 19% from somewhere else. 26 It is also worthy of note that the COTM media campaigns reached mostly Canadians from Ontario (42%). This reflects Ontario s larger population (39% of the total population in Canada), or possibly the greater intensity of the media campaign/cereal box pedometer distribution in Ontario. National health surveillance studies are currently conducted every few years. The Internet offers the opportunity to continuously track physical activity levels and other health indicators in those who participate, which could facilitate more rapid shifts in population health policy in response to Canadians physical activity/health habits. Moreover, our findings suggest that many Canadians would be willing to provide more objective data such as step counts. Indeed, over this study s 10-month reporting period, over 4 million steps were donated to the project. The Internet may also allow us to more precisely and immediately track the effect of different population health campaigns and to implement changes. The Internet may also be less costly than pen and pencil or phone-administered surveys. Although it may not be possible to reach as representative a sample as with national health surveys, accessing important population subgroups (such as women and those overweight) can provide important information relevant to national health policy efforts. 29 Overall, this study suggests that website assessment and intervention systems have the potential to reach large albeit selfselected samples of Canadians. To expand the reach of these website systems, and to recruit a larger and more representative sample, it is likely that additional recruitment channels will be needed. Such strategies might be created by linking the website with other organizations promoting physical activity in collaborative partnerships. For example, the news media could give a daily walking levels across Canada report, in addition to the weather report. Physical activity and recreational groups could provide prompts to access the website system and integrate the system with their physical activity/recreational programs. Worksite health promotion programs could also have participants log their steps on the website and introduce corporate step-count challenges. Strengths of this study include the broad potential reach and accessibility of the website to participants in both urban and CANADIAN JOURNAL OF PUBLIC HEALTH S33

7 TABLE III Demographic Characteristics and Comparisons between One-time and Multi-time Users, Source of Hearing About the Website Demographic Source* Characteristics Back of a Cereal Box Surfed/Linked from Another Word of Mouth N=559 (21%) Media: Television, Magazine, N=627 (24%) Website N=410 (16%) Newspaper, Radio N=1052 (40%) One-time Multi-time χ 2 (p) One-time Multi-time χ 2 (p) One-time Multi-time χ 2 (p) One-time Multi-time χ2 (p) Users Users Users Users Users Users Users Users n (%) n (%) n (%) n (%) n (%) n (%) n (%) n (%) Age 2.06 (0.73) 3.30 (0.51) 9.48 (0.05) (2.5) 2 (2.1) 4 (1.2) 1 (1.4) 10 (2.3) 1 (1.0) 11 (1.3) 2 (1.2) (7.6) 6 (6.3) 43 (13.2) 4 (5.6) 37 (8.5) 7 (6.8) 68 (8.3) 5 (2.9) (51.8) 46 (48.4) 170 (52.1) 32 (44.4) 221 (50.6) 49 (47.6) 336 (41.2) 63 (36.6) (33.1) 38 (40.0) 101 (31.0) 32 (44.4) 158 (36.2) 45 (43.7) 348 (42.6) 87 (50.6) (4.9) 3 (3.2) 8 (2.5) 3 (4.2) 11 (2.5) 1 (1.0) 53 (6.5) 15 (8.7) Total 510 (100) 95 (100) 326 (100) 72 (100) 437 (100) 103 (100) 816 (100) 172 (100) Sex 0.57 (0.45) 0.00 (0.97) 0.03 (0.86) 0.64 (0.43) Male 80 (15.3) 18 (18.9) 71 (21.2) 16 (22.2) 91 (20.4) 20 (19.0) 257 (30.1) 47 (26.7) Female 444 (84.7) 77 (81.1) 264 (78.8) 56 (77.8) 355 (79.6) 85 (81.0) 598 (69.9) 129 (73.3) Total 524 (100) 95 (100) 335 (100) 72 (100) 446 (100) 105 (100) 855 (100) 176 (100) Education 0.24 (0.87) 1.16 (0.56) 0.62 (0.73) 0.08 (0.96) Less than high school 15 (2.9) 2 (2.1) 8 (2.4) 1 (1.4) 8 (1.8) 2 (1.9) 13 (1.5) 3 (1.7) High school/ Trades certificate or diploma/ College 306 (58.7) 58 (60.4) 138 (41.1) 26 (35.6) 180 (40.4) 38 (36.2) 435 (50.3) 87 (49.4) University certificate/ diploma/ degree 200 (38.4) 36 (37.5) 190 (56.5) 46 (63.0) 258 (57.8) 65 (61.9) 416 (48.1) 86 (48.9) Total 521 (100) 96 (100) 336 (100) 73 (100) 446 (100) 105 (100) 864 (100) 176 (100) BMI 0.63 (0.89) 15.76(<0.01) 8.67 (0.03) 1.40 (0.71) <20 (underweight) 10 (4.3) 1 (2.6) 4 (2.7) 3 (10.7) 8 (4.1) 6 (13.0) 29 (4.2) 3 (2.2) (acceptable weight) 74 (31.6) 14 (35.9) 29 (19.9) 10 (35.7) 63 (32.3) 8 (17.4) 172 (24.8) 33 (24.1) (some excess weight) 37 (15.8) 5 (12.8) 30 (20.5) 10 (35.7) 39 (20.0) 8 (17.4) 94 (13.5) 18 (13.1) >27 (overweight) 113 (48.3) 19 (48.7) 83 (56.8) 5 (17.9) 85 (43.6) 24 (52.2) 399 (57.7) 83 (60.6) Total 234 (100) 39 (100) 146 (100) 28 (100) 195 (100) 46 (100) 694 (100) 137 (100) * People who heard about the website through another grocery product were not included (n=7). Characteristics of COTM website participants. Could not be computed due to a low cell frequency. S34 CANADIAN JOURNAL OF PUBLIC HEALTH VOLUME 97, SUPPLEMENT 1

8 rural areas of Canada. In addition, mediators of physical activity are rarely assessed in national health surveys 29 as was done in the current study. This study also provides support for a potentially useful research platform for trialing subsequent interventions. Lack of consistency in the delivery of the website intervention makes it unclear whether it was the sample s demographic characteristics alone, characteristics of the website, or a combination of the two, that influenced the number of repeat visits. In addition, the goal-setting/ feedback aspects of the intervention might have benefitted from greater fidelity to theory. 6 Finally, although the sample was larger than in other website programs, it still represents a self-selected sample. In summary, the Canada on the Move website presents a viable approach to reaching diverse demographic groups. To become a more effective method of reaching large numbers of the Canadian population, further work needs to be undertaken to: 1) develop engaging websites; 2) detail the monitoring of the web-base access information; 3) integrate the website with other organizations promoting physical activity; and 4) expand the number of recruitment sources. REFERENCES 1. Craig CL, Cameron C. Increasing physical activity: Assessing trends from Ottawa, ON: Canadian Fitness and Lifestyle Research Institute, Plotnikoff RC, McCargar LJ, Wilson PM, Loucaides CA. Efficacy of an intervention for the promotion of physical activity and nutrition behaviour in the workplace context. Am J Health Promot 2005;19(6): Statistics Canada. Household Internet Use Survey [Internet article]. Available from: d040708a.htm (Accessed July 8, 2004). 4. Napolitano MA, Fotheringham M, Tate D, Sciamanna C, Leslie E, Owen N, et al. Evaluation of an internet-based physical activity intervention: A preliminary investigation. Ann Behav Med 2003;25(2): Sciamanna CN, Lewis B, Tate D, Napolitano MA, Fotheringham M, Marcus BH. User attitudes toward a physical activity promotion website. Prev Med 2002;35: Rovniak LS, Hovell MF, Wojcik JR, Winett RA, Martinez-Donate AP. Enhancing theoretical fidelity: An -based walking program demonstration. Am J Health Promot 2005;20(2): Tate DF, Wing RR, Winett RA. Using internet technology to deliver a behavioral weight loss program. JAMA 2001;285: Tate DF, Jackvony EH, Wing RR. Effects of internet behavioral counseling on weight loss in adults at risk for type 2 diabetes. JAMA 2003;289(14): Veverka DV, Anderson J, Auld GW, Coulter GR, Kennedy C, Chapman PL. Use of the stages of change model in improving nutrition and exercise habits in enlisted air force men. Milit Med 2003;168(5): McKay GH, King D, Eakin EG, Seeley JR, Glasglow RE. The diabetes network Internetbased physical activity intervention: A randomized pilot study. Diabetes Care 2001;24: Marshall AL, Leslie ER, Bauman AE, Marcus BH, Owen N. Print versus website physical activity programs: A randomized trial. Am J Prev Med 2003;25: Leslie E, Marshall AL, Owen N, Bauman A. Engagement and retention of participants in a physical activity website. Prev Med 2005;40: Harvey-Berino J, Pintauro SJ, Casey Gold E. The feasibility of using internet support for the maintenance of weight loss. Behav Modif 2002;26(1): Silver RC, Holman EA, McIntosh DN, Poulin M, Gil-Rivas V. Nationwide longitudinal study of psychological responses to September 11. JAMA 2002;288(10): Ritter P, Lorig K, Laurent D, Matthews K. Internet versus mailed questionnaires: A randomized comparison. J Medical Internet Res 2004;6(3):e29. Available from: (Accessed November 9, 2005). 16. Statistics Canada Census. Available from: index.cfm (Accessed November 9, 2005). 17. Statistics Canada Census. Available from: ensus96.cfm (Accessed November 9, 2005). 18. Canadian Community Health Survey (CCHS) Cycle 1.1: Extending the wealth of health data in Canada. Available from: english/concepts/health/cchsinfo.htm (Accessed November 9, 2005). 19. Rodgers WM, Sullivan MJL. Task, coping, and scheduling self-efficacy in relation to frequency of physical activity. J Appl Soc Psychol 2001;31(4): Ajzen I. Nature and operation of attitudes. Ann Rev Psychol 2001;52: Courneya KS, Plotnikoff RC, Hotz SB, Birkett NJ. Predicting exercise stage transitions over two consecutive 6-month periods: A test of the theory of planned behaviour in a population-based sample. Br J Health Psychol 2001;6: Sallis JF, Hovell MF, Hofstetter CR, Barrinton E. Explanation of vigorous physical activity during two years using social learning variables. Soc Sci Med 1992;34: Craig CL, Marshall AL, Sjostrom M, Bauman AE, Booth ML, Ainsworth BE, et al. International physical activity questionnaire: 12-country reliability and validity. Med Sci Sports Exerc 2003; International Physical Activity Prevalence Survey: Environmental Module [2002 Nov.]. Available from: IPAQIPS.pdf (Accessed November 9, 2005). 25. Spence JC, Plotnikoff RC, Rovniak LS, Martin Ginis KA, Rodgers W, Lear SA. Perceived neighbourhood correlates of walking among participants visiting the Canada on the Move website. Can J Public Health 2006;97(Suppl 1):S36-S Physical Activity Monitor, Creating Effective Communication for Physical Activity: Trend Available from: pam (Accessed November 9, 2005). 27. Craig CL, Russell SJ, Cameron C, Bauman A. Twenty-year trends in physical activity among Canadian adults. Can J Public Health 2004;95(1): Katzmarzyk PT, Ardern CI. Overweight and obesity mortality trends in Canada. Can J Public Health 2004;95(1): Fridinger F, Macera C, Cordell K. The use of surveillance data and market research to promote physical activity. Am J Prev Med 2002;23(2S): MARCH APRIL 2006 CANADIAN JOURNAL OF PUBLIC HEALTH S35