Project Overview +

This developmental project tests the effects of expert vs user tailoring and rich vs poor graphical treatment in a brief web-based smoking cessation and relapse prevention intervention. The site's content is based on a series of booklets called Forever Free, developed by Dr. Thomas Brandon at the H. Lee Moffitt Cancer Center & Research Institute. The project is an adjunct to Phase I of Project Quit and is offered to Project Quit participants at the end of that study's 6-month follow-up.

Aims +

Aim 1. Screen and identify factors influencing cessation among participants who completed a web based smoking cessation program.

Aim 2. Identify best sequence of treatments for the combined Project Quit/Forever Free Study.

Aim 3. Identify how graphical presentation (is it better for the visual graphics to be rich or poor?) and tailoring locus (is it more or less useful to have a librarian to help you decide - expert tailored - which booklets are best for you?) affect cessation results.

Participants +

Four hundred sixty seven people recruited from the Project Quit Phase I study. Participants agreed to stay involved in the study for an additional 6 months.

Intervention +

During the six-month follow-up phone CATI follow-up to Project Quit Phase I, participants are asked if they would like to participate in a follow-on program to help them stay quit or to help them continue the quitting process. Those willing to participate are sent an email message to log in to their Project Quit website for their new Forever Free program materials.

As participants log in to the website, they are randomly assigned to be in one of five study arms:

  • website rich in graphical presentation and expert tailored
  • website poor in graphical presentation and expert tailored
  • website rich in graphical presentation and user tailored
  • website poor in graphical presentation and user tailored
  • control website

Forever Free materials for arms 1-4 are adapted from Dr. Tom Brandon's 8 original Forever Free relapse prevention booklets. Versions of each are created to support participants as they either remain smoke free or as they get ready to quit smoking again. The booklets for arms 1 and 3 also receive an upgrade in graphic elements and the ability to complete type-in activities on the site.

Expert tailored participants receive a welcome page with recommendations and direct links for two or three of the 8 available web booklets. The recommendations are based on their 6-month follow-up survey responses. They can also navigate to all of the other web booklets if they so choose. User-tailored participants receive a welcome page that encourages them to read from any of the 8 web booklets they feel is most important to them in helping them in their cessation efforts. The Control website contains a brief message about the participant's current smoking status, but does not provide access to any of the 8 web booklets or expert tailored content.

Participants complete a follow-up survey at 3-months and receive updated information according to the study arm they are in, as described above. A final follow-up is completed at 6-months to determine smoking status and program satisfaction.

Findings +

The study was designed to investigate a data analysis method that can be used to construct adaptive health interventions from clinical trial data. An adaptive health intervention (also known as a dynamic treatment regime in the statistical science literature) is a sequence of individually-tailored decision rules that specify whether, how or when to alter the intensity, scope, type, or delivery of treatment at critical decision points based on the changing status of the patient/client. For example, an adaptive health intervention may specify which therapy or treatment (i) to provide first (first-line treatment), (ii) to provide for those who responded adequately to initial treatment (maintenance treatment), and (iii) to provide those who do not respond adequately to initial treatment (rescue treatment).

In particular, the statistical methods developed can be used by investigators to identify baseline patient characteristics (e.g., age, race, gender, co-morbidities) to choose among appropriate first-line treatment options, as well as to identify time-varying measures (e.g., patient response to first-line treatment, including adherence to first-line treatment) to choose among appropriate subsequent treatment options. After applying the method to clinical trial data, the result is an optimized, individually-tailored, adaptive health intervention that can be used to guide clinical practice. Both Project Quit and Forever Free clinical trial data sets were combined and used to illustrate these new statistical methodologies.

Conclusion +

This research led to new statistical methodologies that can be used by researchers to develop individually-tailored, adaptive health interventions. In addition to tailoring content, the methods can be used to learn how to tailor the delivery or modality or scope or intensity of the interventions over time depending on the changing status of the patient (i.e., as they respond or fail to respond to the intervention over time).

Forever Free

09/01/2003 - 08/31/2006


National Cancer Institute

Principal Investigator:

Susan A. Murphy, PhD