Aim 1. Compare the effectiveness of low- versus high-depth tailored websites in achieving a 7-day abstinence measured at 6-month follow-up.

Aim 2. Compare the effectiveness of low- versus high-depth tailored websites on smoking cessation across baseline characteristics of self-efficacy, motivation, gender, age, income, and education.

1,848 participants from three HMOs: Group Health (Seattle, WA), HealthPartners (Minneapolis, MN) and the Health Alliance Plan (Detroit, MI).

Participants are randomized into one of 16 study arms. All participants receive a web-based guide to help them quit smoking, along with a free 10-week supply of nicotine replacement therapy patches. Email messages are sent to participants about various aspects of the program, including patch use and step down process, and the availability of new web guide materials.

The 6 components for each web guide include:

1. Source Letter: available immediately after baseline survey
2. Action Plan: available immediately OR 7 days before Quit date
3. Efficacy x 2: available immediately OR 1 day before and 7 days after Quit date
4. Outcomes: available immediately OR 14 days after Quit date
5. Testimonial: available immediately OR 21 days after Quit date

Additional factors:

• High vs. Low Depth: All components are tailored, using a high or low degree of depth.
• Single vs. Multiple Exposure: Participants receive web guide components either all at once (immediately after baseline survey), OR as listed above, over time.
• Gain vs. Loss Framed Messages: Outcomes messages are written either in terms of the benefits users stand to gain by quitting or the risks they stand to face by continuing to smoke.

To test the impact of tailoring depth, smokers were assigned to receive between 0 and 3 high-depth tailored cessation intervention components (success stories, outcome expectations, and efficacy expectations). Tailoring depth was significantly related to 6-month smoking cessation using per-protocol analysis, both across the entire range of cumulative high-depth components (OR1.91; CI1.18 –3.11) and for each high-depth component added (OR1.24; CI1.06 –1.45). Tailoring depth was marginally related (p0.08) to smoking cessation in the complete respondent and ITT analyses. Adjusted 6-month cessation rates among participants receiving all three high-depth tailored components were 38.6% in the per-protocol analysis, 37.9% in the complete respondent analysis, and 27.7% in the ITT analysis.

Other findings include:

• Abstinence was most influenced by high-tailored success stories (testimonials) and high-personalized message source (source letter).
• Cumulative assignment of the three tailoring depth factors (message source, testimonial, and efficacy) also resulted in increasing rates of six-month cessation, demonstrating an effect of tailoring depth.
• Greater engagement with a web-based smoking cessation program results in greater subsequent cessation.
• The depth of tailoring in smoking cessation messages results in greater perceptions of message relevance, which, in turn, results in greater engagement with a web-based smoking cessation program.
• A significant number of smokers can be recruited through Health Maintenance Organizations (HMOs) for web-based smoking cessation interventions.
• African Americans with a large number of people supporting their quit efforts were more than two times more likely to quit than those with few or no supporters, regardless of tailoring intensity.

The study identified relevant components of smoking cessation interventions that should be generalizable to other cessation interventions. The study also demonstrates the importance of higher-depth tailoring in smoking cessation programs. Finally, the fractional factorial design allowed efficient examination of the study aims. The rapidly changing interfaces, software, and capabilities of eHealth are likely to require such dynamic experimental approaches to intervention discovery.