During the 10-day study period, 2265 eligible US residents were intercepted. Of those, 35.8% (N = 811) clicked on the “survey” link, 48% (N = 1088) clicked on the link to take them directly to the site (“declined”), and 16.2% (N = 366) did neither (“abandoned”). Of the 811 individuals who clicked through to the survey, 87.2% (N = 655) completed the full survey, yielding an overall response rate of 29% (). Of the survey completers, 29 individuals reported having never smoked, leaving a final sample of 626 respondents.
Eligibility and Recruitment Results
To assess generalizability, we compared all survey participants (N = 655) with nonrespondents who proceeded to register with QuitNet (N = 243). Overall, nonrespondents (N = 1454, abandoned and declined) were significantly less likely than survey respondents to register on QuitNet (16.7 vs 51.4%, Χ2
2 = 303.7, P < .001). Compared to survey respondents, nonrespondents spent less time on QuitNet (4.5 vs 12.0 minutes, t = 13.4, P < .001) and viewed fewer pages (5.9 vs 15.3 pages, t = 16.0, P < .001) on the website. Nonrespondents were more likely to be female (59.4 vs 51.9%, Χ2
2 = 4.2, P = .02) but did not differ by age, smoking status, time of survey invitation, or specific search engine used.
As shown in , the relative volume of participants referred from each search engine was consistent with national usage patterns (Χ2
= 1.06, P
= .59). In this study, 57% of participants were referred from Google, 29% from Yahoo!, and 14% from MSN. At the time of this study, 60% of all Internet search queries were estimated to be conducted using Google, 23% with Yahoo!, and 17% with MSN [14
Comparison of search engine usage to Nielsen/NetRatings statistics
The use of key search terms (“quit smoking,” “quitting smoking,” “stop smoking,” or “stopping smoking”) by survey respondents was also consistent with search patterns captured by Overture and Wordtracker. As shown in , the most commonly used search term was “quit smoking,” which constituted 52.9% of study queries, 59.1% of Overture queries, and 47.8% of Wordtracker queries. “Stop smoking” was the second most frequently used search term, which constituted 24.9% of study queries, 31.1% of Overture queries, and 36.5% of Wordtracker queries.
Frequency of smoking-related search terms in search engine queries
As shown in , the majority of study participants were female (61.2%, n = 383) and between the ages of 26 and 44 years (62.7%, n = 393); 18.7% (n = 117) were aged 18-25 years, 17.1% (n = 107) were aged 45-64, and less than 1% were 65 or older (n = 4) or under age 18 (n = 5). Adjusted to local time of the participant, more than half (53.4%) of search engine queries for cessation information occurred during work hours (8 am-5 pm), 26.6% occurred between 5-9 pm, and 20% occurred at night (9 pm-6 am).
Demographic and smoking characteristics of study participants (N = 626)
Participants were asked the reason they were searching for smoking cessation information. The majority of survey respondents (90.1%, n = 590) indicated that they were looking for help or support for themselves; 5.6% (n = 37) were looking for general information; 3.4% (n = 22) were looking for help for someone else; and 1% (n = 6) were health professionals or researchers looking for information. Further analyses were limited to individuals looking for cessation help or support for themselves or for general cessation information (N = 626). Among these individuals, 75.4% (n = 472) were current smokers, 17.4% (n = 109) had quit within 7 days (“recent quitters”), and 7.2% (n = 45) had quit more than 7 days ago (“longer-term quitters”).
The majority of current smokers (52.8%, n = 249) planned to quit in the next 30 days, 47.0% (n = 222) planned to quit in the next 6 months, and one person (0.2%) was not thinking about quitting. Smokers had made an average of 5.1 quit attempts (SD = 14.7; median = 1) during the past year.
As shown in , information preferences varied by smoking status. Current smokers were more likely than recent quitters and longer-term quitters to be interested in information about how to quit smoking (88.1%, 54.1%, and 40.0%, respectively; Χ2
2 = 104.7, P < .001) and medication usage (30.7%, 5.5%, and 4.4%, respectively; Χ2
2 = 41.0, P < .001). Not surprisingly, both recent quitters and longer-term quitters were more interested than current smokers in information about withdrawal (77.1%, 66.7%, and 59.7%, respectively; Χ2
2 = 11.7, P = .003).
Information sought by smoking status (N = 626)
Perceived Helpfulness of Cessation Services
Participants were also asked to rate the perceived helpfulness of various smoking cessation treatment interventions on a scale from 1 to 5, with 1 representing “very helpful” and 5 representing “not helpful at all.” As shown in , the three features that were rated most highly by all participants were (1) individually tailored information (mean = 1.90, SD = 1.18); (2) information on withdrawal (mean = 1.84, SD = 1.15); and (3) a meter that keeps track of personal data (mean = 2.14, SD = 1.37). The three features rated the lowest by all participants were (1) support from a telephone counselor (mean = 3.21, SD = 1.35); (2) email support (mean = 2.95, SD = 1.40); and (3) support from others (mean = 2.90, SD = 1.38). Ratings of perceived helpfulness varied according to smoking status. Current smokers rated information about medications, assistance in setting a quit date, and assistance in choosing a medication as more helpful than did recent quitters and ex-smokers. Support from others and information about withdrawal received higher ratings of perceived helpfulness from recent quitters and ex-smokers than from current smokers. As detailed in , information of withdrawal, individually tailored information, and tracking meters were rated as “helpful” or “very helpful” by over half of the participants, while telephone counseling was thought to be helpful by less than 30% of participants.
Perceived helpfulness of Internet features by smoking status
Proportion of participants (N = 626) rating Internet cessation services as helpful or very helpful
Estimating Incidence of Cessation Queries
Over the course of 9 months, 541685 searches were extracted from MetaSpy, of which a total of 38 were smoking cessation related. Assuming a total search engine volume of 52 billion searches per year [14
], this ratio yields an estimate of 3.6 million (99% CI = 2.5-4.8 million) cessation-related searches per year in the United States alone.