This is an open-access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in the Journal of Medical Internet Research, is properly cited. The complete bibliographic information, a link to the original publication on http://www.jmir.org/, as well as this copyright and license information must be included.
Most smokers attempt to quit on their own even though cessation aids can substantially increase their chances of success. Millions of smokers seek cessation advice on the Internet, so using it to promote cessation products and services is one strategy for increasing demand for treatments. Little is known, however, about what cessation aids these smokers would find most appealing or what predicts their preferences (eg, age, level of dependence, or timing of quit date).
The objective of our study was to gain insight into how Internet seekers of cessation information make judgments about their preferences for treatments, and to identify sociodemographic and other predictors of preferences.
An online survey assessing interest in 9 evidence-based cessation products and services was voluntarily completed by 1196 smokers who visited the American Cancer Society’s Great American Smokeout (GASO) webpage. Cluster analysis was conducted on ratings of interest.
In total, 48% (572/1196) of respondents were “quite a bit” or “very much” interested in nicotine replacement therapy (NRT), 45% (534/1196) in a website that provides customized quitting advice, and 37% (447/1196) in prescription medications. Only 11.5% (138/1196) indicated similar interest in quitlines, and 17% (208/1196) in receiving customized text messages. Hierarchical agglomerative cluster analysis revealed that interest in treatments formed 3 clusters:
Smokers accessing the Internet for information on quitting appear to differentiate cessation treatments by how much interpersonal interaction or support the treatment entails. Quitting date, smoking level, and sociodemographic variables can identify smokers with varying levels of interest in the 3 classes of cessation methods identified. These results can potentially be used to more effectively target and increase demand for these treatments among smokers searching the Internet for cessation information.
Several effective tobacco-cessation products and services are available to help smokers who want to quit. These can double or triple the rate of cessation compared with quitting without help [
There has been little research on why smokers are not using evidenced-based treatments to quit. One study found that, although smokers planning to quit expressed several barriers around quitline use, none of the self-reported barriers predicted actual calls made to a quitline [
To increase utilization of cessation products and services, the National Tobacco Cessation Collaborative, an American and Canadian consortium of leading nonprofit and government agencies dedicated to reducing the burden of tobacco use, delineated several core strategies to increase demand for available treatments. One of these strategies includes the recommendation to understand what smokers need and want, instead of viewing them as “passive treatment beneficiaries rather than treatment consumers” (p. S308) [
Many smokers thinking of quitting access the Internet for general information on smoking cessation. The Internet may therefore be an effective medium for promoting evidenced-based cessation aids [
Identifying and targeting potential quitters by providing them information on or immediate access to treatments would be maximally effective if smokers’ preferences for treatments and predictors of those preferences were known. Only 1 study to our knowledge, however, asked smokers seeking cessation information on the Internet to rate their perceptions of various treatments [
Research on what methods are used most frequently by smokers when attempting to quit could provide some indication of what cessation products and services would be favored. Low use or nonuse of a particular product or service, however, may be due to a smoker simply not knowing it exists. Some products may also be used only because they were recommended or were available at the time. Nevertheless, data on actual use of evidence-based cessation treatments can be used to form hypotheses about what types of treatments might be preferred, or how preferences might vary as a function of sociodemographic characteristics. For example, in the 2005 National Health Interview Survey (NHIS) [
We also expected preferences to vary by gender. In the 2-year longitudinal National Youth Smoking Cessation Survey of smokers aged 16–24 years [
Participants’ mean age was 38.4 years (SD 9.1) and they smoked an average of 15.9 cigarettes per day (SD 9.1). The majority of the sample was female (840, 74.3%). A college degree or higher was reported by 34.2% (388), with 39.6% (449) reporting some college, 21.1% (239) completing high school or its equivalent, and 5.1% (58) completing grade 11 or lower. A minority (43, 3.8%) were Latino/Hispanic, whereas the majority were not (1068, 94.8%), or indicated “Don’t know” (15, 1.4%). The majority of participants were white (984, 87%). Black or African American smokers constituted 5% (57) and Asian smokers, 1.6% (18) of the sample. The remaining racial groups were collapsed into 1 category that comprised 6.4% (75) of the sample (ie, Pacific Islander, American Indian, Alaska Native, Other, and Don’t know) (see
Descriptive statistics of the study population
Characteristic | n | % | |
|
|||
<10 | 230 | 22.1 | |
10–19 | 379 | 36.4 | |
20–29 | 329 | 31.6 | |
≥30 | 103 | 9.9 | |
|
|||
≤25 | 164 | 15.7 | |
26–40 | 438 | 41.9 | |
41–55 | 358 | 34.2 | |
≥56 | 86 | 8.2 | |
|
|||
Female | 840 | 74.3 | |
Male | 291 | 25.7 | |
|
|||
Grade ≤11 | 58 | 5.1 | |
High school graduate or GEDa | 239 | 21.1 | |
Some college | 449 | 39.6 | |
College graduate or higher | 388 | 34.2 | |
|
|||
White | 984 | 87.0 | |
Black/African American | 57 | 5.0 | |
Asian | 18 | 1.6 | |
Pacific Islander/American Indian/Alaskan Native/Other/Don’t know | 72 | 6.4 | |
|
|||
Yes | 730 | 63.2 | |
No | 425 | 36.8 | |
|
|||
In the next 24 hours | 266 | 22.2 | |
In next week or two | 362 | 30.3 | |
In next month | 209 | 17.5 | |
In next 6 months | 114 | 9.5 | |
In future/undecided | 245 | 20.5 | |
|
|||
Yes | 256 | 21.4 | |
No | 35 | 3.0 | |
Don’t know | 861 | 74.7 | |
|
|||
Yes | 112 | 9.7 | |
No | 884 | 76.6 | |
N/Ab—never tried to quit before | 158 | 13.7 |
a General equivalency diploma.
b Not applicable.
A 12-item online questionnaire was posted for 11 months on the American Cancer Society’s (ACS) Great American Smokeout (GASO) webpage. The survey was posted 1 week prior to the 2008 GASO event (November 20). The GASO webpage is the ACS’s online portal for information about quitting and received 92,946 unique views during the study period. An introductory paragraph on the GASO webpage explained that the ACS was interested in learning about how smokers quit, and that if they were interested to click on the link appearing below. The link led respondents to a page in which a consent section appeared above the survey. The survey was completed by 1594 current smokers over the entire study period. Approximately half of responses (845, 53%) were collected by 6 days after GASO, and the other half during the remaining months of the study period. We excluded 90 participants due to missing data on sociodemographic variables. Analyses reported below were restricted to participants who provided responses on sociodemographic variables and who did not choose the “don’t know” option on the items assessing interest in cessation products and services (N = 1196). The voluntary and anonymous survey, which was approved by the Emory University institutional review board, did not provide incentives, financial or otherwise, for completion.
The current study also subjected ratings of interest in cessation products and services to a cluster analysis. Cluster analysis is an assumption-free classification technique that is commonly used in market research to understand consumer behavior [
To examine the relationship between interest in cessation products and services, and sociodemographic and smoking behavior variables, a multivariate analysis of variance (MANOVA) was conducted. Cessation methods that formed clusters were averaged to obtain summary variables that were the set of dependent variables in the MANOVA. Independent variables (categorical) were gender, age, race, education, when smokers were planning to quit, prior use of a quitline, smoking rate, knowledge of quitlines’ free availability, and whether an attempt to quit was made in the past year. A second MANOVA was conducted that excluded independent variables with nonsignificant multivariate main effects (gender, quitline knowledge, and past-year attempt). Multivariate results from this final MANOVA are presented. Analyses that controlled for the number of days between when the survey was completed and the start date of the survey were also conducted.
Follow-up Tukey pairwise comparisons that controlled for the familywise error rate were conducted. These examined differences on dependent variables among levels of the significant independent variables (ie, significant according to the univariate results). For comparisons that the Tukey tests indicated were statistically significant (ie,
Participants were asked to indicate when they planned to quit (in the next 24 hours, in the next week or two, in the next month, in the next 6 months, sometime in the future but haven’t decided when, not applicable (N/A)—already quit, other), whether they knew if “free help from a counselor at a quitline” was available to all smokers in their state (yes, no, don’t know), whether they had ever called a quitline to help them quit smoking (yes, no, N/A—never tried to quit before), the number of cigarettes smoked per day, whether they tried to seriously quit in the past year (yes, no), their gender, and their age.
Smokers were asked to indicate how interested they would be in using the following cessation products and services on a scale from 1 (not at all) to 5 (very much). Specific items, as written, were (1) “Using a telephone quitline (a quitline has trained counselors help you over the phone with your quit attempt,” (2) “Using a website that gives professional advice about quitting smoking that is customized for you,” (3) “Using the Internet to chat with other smokers who are trying to quit,” (4) “Receiving emails timed around your quit date that contain professional advice about quitting that is customized for you,” (5) “Attending a program led by a professional and that involves a few meetings with other smokers trying to quit,” (6) “Using nicotine replacement therapy (eg, the patch) which is available without a prescription,” (7) “Using booklets or other printed materials that give professional advice on how to quit,” (8) “Receiving text messages on your cell phone timed around your quit date that contain professional advice about quitting that is customized for you,” and (9) “Using prescription medications for quitting such as Zyban (bupropion) or Chantix (varenicline).”
The majority of respondents had tried to quit in the past year (730, 63.2%) (see
A large majority of smokers (861, 74.7%) did not know whether free help from a quitline counselor was available to all smokers in their state; 3.0% (35) indicated that such help was
The cessation method that received the greatest proportion of respondents who indicated being “quite a bit” or “very much” interested was NRT (572/1196, 47.8%), followed by a website that provides customized quitting advice (534/1196, 44.6%) and prescription medications (447/1196, 37.4%). Only 11.5% (138/1196) of respondents indicated being “quite a bit” or “very much” interested in using quitlines, and only 17.4% (208/1196) reported similar interest in receiving customized text messages. The proportion of respondents in each sociodemographic category who were “quite a bit” or “very much” interested in a particular cessation method are presented in
Number and percentage of respondents in each sociodemographic category who were “quite a bit” or “very much” interested in each cessation method
Quitline | Website | Web peer |
Emails | Group |
NRTa | Prescription |
Cessation |
Text |
n | ||
≤25 | 14 |
64 |
33 |
47 |
30 |
65 |
51 |
34 |
45 |
164 | |
26–40 | 44 |
94 |
94 |
162 |
85 |
217 |
187 |
146 |
79 |
438 | |
41–55 | 54 |
87 |
87 |
143 |
94 |
185 |
138 |
121 |
52 |
358 | |
≥56 | 9 |
21 |
21 |
35 |
24 |
47 |
30 |
28 |
10 |
86 | |
Male | 24 |
120 |
50 |
90 |
56 |
142 |
93 |
71 |
44 |
291 | |
Female | 106 |
391 |
200 |
324 |
193 |
411 |
338 |
284 |
154 |
840 | |
Grade ≤11 | 10 |
23 |
15 |
21 |
9 |
17 |
17 |
17 |
9 |
58 | |
High school graduate/GEDb | 24 |
103 |
53 |
82 |
47 |
122 |
87 |
65 |
30 |
239 | |
Some college | 46 |
196 |
99 |
157 |
99 |
226 |
184 |
154 |
79 |
449 | |
College graduate or higher | 51 |
190 |
84 |
155 |
95 |
189 |
144 |
120 |
81 |
388 | |
White | 105 |
452 |
210 |
359 |
206 |
490 |
376 |
303 |
164 |
984 | |
Black/African American | 15 |
28 |
18 |
28 |
22 |
26 |
20 |
24 |
18 |
57 | |
Asian | 2 |
8 |
5 |
5 |
4 |
8 |
8 |
8 |
4 |
18 | |
Other | 9 |
23 |
16 |
23 |
17 |
28 |
27 |
21 |
13 |
72 |
a Nicotine replacement therapy.
b General equivalency diploma.
Inspection of the dendrogram from the cluster analysis and membership at each cluster stage suggested that a 3-cluster solution was appropriate (
Dendrogram indicating clusters from cluster analysis of interest in cessation products and services (Msgs = messages, NRT = nicotine replacement therapy, Prgrm = program, PrscrpMed = prescription medication)
A second cluster consisted of using a customized website, receiving timed emails, and using printed materials that give professional advice on how to quit. We labeled this group of treatments
Composite variables to represent interest in each of the 3 types of cessation methods were computed as the average of the variables in each cluster. The means and standard deviations for each cluster were as follows: nonsocial–informational methods (mean 2.9, SD 1.2); interpersonal–supportive methods (mean 2.1, SD 1.0); and pharmacotherapy (mean 3.0, SD 1.3). Paired samples
Results from multivariate analyses are reported in
Results of multivariate analysis of variance indicating main effects of independent variables on smokers’ interest in cessation treatments
Multivariate | Univariate | |||||
Independent variable | Pillai trace |
|
Dependent variable |
|
df |
|
Quit date | 4.21 | <.001 | Pharmacotherapy | 4.11 | 4, 938 | .003 |
Interpersonal–supportive | 4.04 | 4, 938 | .003 | |||
Nonsocial–informational | 8.10 | 4, 938 | <.001 | |||
Used a quitline before | 3.79 | <.001 | Pharmacotherapy | 4.18 | 2, 938 | .016 |
Interpersonal–supportive | 8.69 | 2, 938 | <.001 | |||
Nonsocial–informational | 1.26 | 2, 938 | .28 | |||
Smoking rate | 1.95 | .04 | Pharmacotherapy | 3.02 | 3, 938 | .03 |
Interpersonal–supportive | 2.74 | 3, 938 | .04 | |||
Nonsocial–informational | 1.48 | 3, 938 | .22 | |||
Age group | 2.48 | .008 | Pharmacotherapy | 3.30 | 3, 938 | .02 |
Interpersonal–supportive | 0.31 | 3, 938 | .82 | |||
Nonsocial–informational | 2.50 | 3, 938 | .06 | |||
Race | 2.59 | .006 | Pharmacotherapy | 0.31 | 3, 938 | .82 |
Interpersonal–supportive | 6.71 | 3, 938 | <.001 | |||
Nonsocial–informational | 1.64 | 3, 938 | .18 | |||
Education level | 2.16 | .02 | Pharmacotherapy | 2.15 | 3, 938 | .09 |
Interpersonal–supportive | 3.37 | 3, 938 | .02 | |||
Nonsocial–informational | 0.84 | 3, 938 | .47 |
a Indicates significance of multivariate relationship between the independent variable and the set of 3 dependent variables.
b Indicates significance of univariate relationship between the independent variable and each dependent variable.
The later smokers planned to quit, the greater their interest in pharmacotherapy. While 48% (69/143) of smokers planning to quit within a day were “quite a bit” or “very much” interested in pharmacotherapy, a significantly greater proportion (≥75%); planning to quit later were similarly interested, either “in the next week or two” (132/175, 75%; n = 318, χ2
1 = 25.0,
Interest in nonsocial–informational methods was greater among smokers planning to quit in the next week or two (121/144, 84%) compared with those quitting in the next 24 hours (72/107, 67%; n = 251, χ2
1 = 9.7,
A greater proportion of smokers who had used a quitline before were “quite a bit” or “very much” interested in pharmacotherapy (49/63, 78%), than the proportion who had never before used a quitline (284/439, 65%; n = 502, χ2
1 = 4.2,
Interest in pharmacotherapy was significantly associated with smoking level. Among lighter smokers (<10 cigarettes per day), 57% (63/111) were “quite a bit” or “very much” interested in pharmacotherap
Although the MANOVA and univariate tests indicated a significant relationship between cigarettes per day and interest in interpersonal–supportive methods, none of the pairwise Tukey test comparisons were significant (all
A significantly smaller proportion of younger smokers (≤25 years) were “quite a bit” or “very much” interested in pharmacotherapy (51/88, 58%) than the proportion (163/220, 74%) of smokers in the 26 to 40 age group (n = 308, χ2
1 = 7.2,
The univariate
White and black/African American smokers differed in their interest in interpersonal–supportive methods. While only 18% (44/242) of white smokers were “quite a bit” or “very much” interested in interpersonal–supportive methods, a much greater percentage (7/15, 47%) of black/African American smokers were similarly interested (n = 257, χ2
1 = 7.2,
Although the univariate
Increasing the demand for cessation products and services will lead to more quit attempts, higher cessation rates, and greater reductions in smoking prevalence [
Cluster analysis of smokers’ ratings suggested that smokers’ interest in behavioral treatments centered on the degree of social or interpersonal involvement or social support the treatment would provide. Specifically, one cluster we obtained comprised products and services that involve high levels of interpersonal interaction and/or support. These included group-support cessation programs, telephone counseling, and using the Internet to chat with other smokers trying to quit. A contrasting cluster consisted of cessation methods that would provide tailored or individualized information on cessation, but that would
A separate cluster was also obtained that, as hypothesized, consisted of pharmacotherapies for cessation. This result suggests that motivation to use pharmacotherapy extends to both nicotinic and non-nicotinic medications. We had also hypothesized that interest in pharmacotherapy would be greater than interest in behavioral methods. We found, however, that smokers’ interest in medications as a whole was comparable with their interest in nonsocial–informational methods. Interest in interpersonal–supportive methods received lower ratings than pharmacotherapy or nonsocial–informational methods. These results are consistent with Cobb and Graham’s [
As noted earlier, compared with traditional media such as radio or television advertising, online advertising has been shown to be a more cost-efficient mode of recruiting smokers to Internet and telephone-based cessation treatments [
Our results indicated that more smokers who were quitting the next day were not at all interested in pharmacotherapy compared with smokers quitting later on. Smokers planning to quit the next day most likely have made up their minds about quitting right away. They therefore may not want to spend the time to learn about and choose a medication, go to the drugstore, or ask a doctor to write a prescription. However, given the efficacy of pharmacotherapy, efforts might be aimed at these smokers to set a quit date with enough time to consider pharmacotherapy.
As hypothesized, results indicated that a greater proportion of older smokers (26–40 years) than younger smokers (≤25 years) were strongly interested in pharmacotherapy. This suggests that older smokers would be more receptive than younger smokers to efforts that encourage pharmacotherapy use. It is not immediately clear why a smaller proportion of younger smokers were interested, but one possibility is that more of them may believe they cannot afford medications. Alternatively, younger smokers may hold myths about medications that older smokers know are not true. Another possibility is that older smokers are more accustomed to taking medications for various health conditions. These possibilities can be examined in future research addressing preferences for cessation methods. The affordability of medications, or myths about them, may also explain why a much smaller proportion of less educated smokers were interested in pharmacotherapy than more educated smokers. Understanding the main reasons for less educated smokers’ reluctance to use pharmacotherapy through further research would be important in devising strategies to increase demand for its use in this population.
Strategies to increase demand for pharmacotherapy should also consider smoking level, as the number of cigarettes smoked per day was associated with interest in pharmacotherapy. Fewer of the lightest smokers (defined as smoking <10 cigarettes per day) and fewer of the heaviest smokers (≥30 cigarettes per day) were interested in pharmacotherapy than were those who smoked between 10 and 29 cigarettes per day. A possible explanation is that lighter smokers may be more likely to believe that they are not so addicted that they need pharmacotherapy to help them quit. In contrast, the heaviest smokers may be more likely to believe that they are so addicted that even medications cannot help them quit. If evidence is obtained supporting these reasons, they can be addressed in online messages to increase demand for pharmacotherapy such as NRT.
Based on the results obtained, behavioral marketing could potentially be used to increase demand for NRT by targeting smokers 26–40 years old who have at least a high school education (as these individuals appear to be most interested in pharmacotherapy). Messages for these individuals could emphasize the ease of purchasing NRT, compare its cost relative to continued smoking, clarify concerns about using nicotine for treatment, and note its effectiveness relative to no medications. Focus-group research would be helpful in determining the precise content of messages that would resonate most for this demographic segment.
In general, smokers were less interested in interpersonal–supportive cessation methods than in pharmacotherapy or nonsocial–informational methods. Moreover, interest in interpersonal–supportive methods could not be explained by whether smokers knew that free quitlines were available (as results were no different when this variable was controlled). In spite of the generally lower interest in interpersonal–supportive methods, however, results indicated that there were differences in interest as a function of quit date and whether smokers had previously used a quitline. Specifically, more smokers planning to quit “in the next week or two” were interested in interpersonal–supportive methods than were smokers who planned to quit at some undecided time in the future. This could be interpreted as a function of the greater seriousness about quitting among smokers planning to quit in the next week or two than among smokers who simply report planning to quit at some undecided time in the future.
Smokers who had previously used a quitline were also much more likely to be interested in interpersonal–supportive methods than smokers who had never before called a quitline. Smokers who had previously used a quitline may have had a positive enough experience to consider using an interpersonal–supportive method again. Alternatively, these methods may appeal to these smokers because of preexisting personal characteristics. For example, smokers higher on the personality trait of extraversion may find social interactional methods more appealing.
We had hypothesized that male smokers would be less interested than female smokers in interpersonal–supportive methods. Gender, however, was unrelated to interest in these treatments or, for that matter, in the two other types of treatments. Previous research has found gender differences among young adult smokers in the seeking of help from a professional [
Overall, these results indicate that it may take a greater amount of effort to persuade smokers using the Internet for cessation advice to use interpersonal–supportive methods such as quitlines, group counseling, or Internet forums for peer support. However, given the demonstrated efficacy of quitlines, efforts should be made to encourage these options, at least among those smokers who would be more receptive (ie, smokers planning to quit earlier and who may have used a quitline before).
Nonsocial–informational methods include Internet-based treatments for tobacco use, which reviews and meta-analyses have concluded are effective compared with minimal or no treatments [
Only age and when smokers were planning to quit were significantly related to interest in nonsocial–informational methods. While among all age groups most were “quite a bit” or “very much” interested in these methods, a greater majority of
The sample of individuals in this study may not be representative of the population of smokers who use the Internet for cessation advice. Nevertheless, some of our results for interest in cessation methods parallel actual use found by studies using different recruitment methods. For example, adult smokers in the 2005 NHIS reported a low rate of having used behavioral treatments (which included in-person or telephone counseling or group cessation programs) compared with pharmacotherapy. This is consistent with the current study’s finding of greater interest in pharmacotherapy than in behavioral methods. The NHIS also found that young adult smokers were less likely than older smokers to use pharmacotherapy [
It is also not certain whether interest in a type of cessation method would translate into actual use of that method if it were made available. The aim of the study, however, was not to predict actual use but rather to first understand the basis of smokers’ preferences. This is information that can then be used to increase adoption of evidenced-based treatments. Self-report is a reasonable approach to understanding these preferences and is typically a first step in market research. Future research may then focus on how to increase demand based on knowledge of these preferences by developing appropriate messaging for particular segments of consumers. Making preferred treatments appealing and available at the time smokers are seeking cessation advice on the Internet would help translate preferences to actual behavior, especially given the convenience of the Internet for purchasing products and services. There is also no reason to believe that ratings of preferences for particular cessation methods would be subject to social desirability biases. In the case of tobacco use behavior, self-reports have been shown to be reliable and valid [
Due to concerns for brevity, the study was unable to test specific underlying reasons for associations between predictors and interest in treatments. For example, why a greater proportion of older smokers were strongly interested in nonsocial–informational cessation methods than were younger smokers could not be answered by the present study. Understanding smokers’ reasons behind their interest (or noninterest) in cessation methods would be a fruitful topic for further research. Results of that research would help in refining strategies to promote use of cessation products and services for all sociodemographic groups. Evaluating smokers’ interest in cessation aids with no evidence of effectiveness (eg, hypnosis, acupuncture) would also be informative.
Smokers who want to quit have available to them several effective cessation products and services. Demand for these is relatively low, however. Fortunately, many smokers access the Internet to help them quit, so targeting these smokers to promote cessation aids is a potentially effective way of increasing demand. Results of the current study indicated that they have relatively greater interest in pharmacotherapy and in cessation methods that provide tailored information or advice, but that do
None declared
American Cancer Society
Great American Smokeout
multivariate analysis of variance
National Health Interview Survey
nicotine replacement therapy