Challenges in Collecting Data for WebBased Smoking Cessation Research

1
Challenges in Collecting Datafor Web-Based
Smoking Cessation Research

• Peter Selby MD
• Peter Farvolden PhD
• Trevor van Mierlo BA
• Centre for Addiction and Mental Health
• Toronto, Ontario
• Web Assisted Tobacco Interventions
• January 20, 2004

2
Evidence for Web Based Self Help

• To date, there are not enough studies of
  Web-Based self-help for smoking cessation to draw
  any conclusions about their effectiveness
• However, there is emerging evidence that
  web-based self-help smoking cessation programs
  are effective, high-reach, relatively low cost
  interventions

3
Efficacy vs. Reach

• The Impact of tobacco cessation interventions is
  a function of the interventions potential reach
  (i.e. the proportion of smokers who receive the
  treatment) and effectiveness (i.e., long term
  abstinence rates associated with the treatment)
• Intensive clinical interventions have high
  cessation rates (25-40) but reach few smokers
• Public health interventions (e.g., information
  booklets) reach more smokers but are associated
  with much lower cessation rates

4
Open Trials

• In the absence of data from controlled trials we
  must rely on other evidence
• Is it best viewed as an intervention or is it a
  medium for knowledge transfer and should be
  evaluated as such?

5
Design and Pilot Evaluation of an Internet
Smoking Cessation Program (Lenert et al., 2003)

• Web-based open trial
• N 49 (recruited via email)
• Ss prepared to quit in the next 30 days
• Outcome data collected at least 30 days later
• Outcome data for 26/49 Ss
• Overall quit rate at 30 days 18
• 16 of Ss reported a reduction in cigarettes
  smoked

6
Evaluation of an Internet-Based Smoking Cessation
Program Lesson Learned from A Pilot Study (Feil
et al., 2003)

• 370 Ss recruited via email
• Ss at least Contemplators
• Follow up data at 3 months
• Cessation rate at 3 months was 18 with
  non-responders (n161) considered smokers
• Social support group feature used most frequently

7
Web Assisted Smoking Cessation High Reach,
Sustainable and Effective (Selby Farvolden,
2003)

• Cumulative tobacco dependence and program usage
  data from 4,644 treatment program registrants and
  153,083 Support Group visits were analyzed
  according to gender and place of residence.
• Open trial of all users regardless of stage of
  change

8
Web Assisted Smoking Cessation High Reach,
Sustainable and Effective (Selby Farvolden,
2003)

• The majority of participants were American
  (67.45) and 65.7 were women.
• On average, users were 37.1 years of age, smoked
  21.8 cigarettes per day and had smoked for 18.9
  years.
• The mean FTND scores were 5.77. Users had a
  median of 3 previous quit attempts and 43 had at
  least one co-resident smoker.

9
Quit Rates

• Number who set a quit date 1244 (100)
• One diary entry 1085 (87.21)
• Two or more entries 784 (63.02)
• 1085 users used the diary to help maintain their
  quit attempt for an average of 45.05 (SD53.30)
  days.

10
Quit Rates

• Last entry 802 (73.92) were maintaining their
  quit,
• 189 (17.42) reported that they had slipped and
• 94 (8.66) reported that they were back to
  smoking.
• Can it be this good?

11
Quit Rates

• N 1085 who made at least one journal entry
• 4 weeks 302/1085 27.83
• 8 weeks 228/1085 21.01
• 12 weeks 141/1085 13.00
• 16 weeks 93/1085 8.57
• 20 weeks 65/1085 5.99
• 24 weeks 29/1085 2.67

12
Quit Rates

• Now N 784 (2 or more entries)
• 4 weeks 302/784 38.52
• 8 weeks 228/784 29.08
• 12 weeks 141/784 17.98
• 16 weeks 93/784 11.86
• 20 weeks 65/784 8.29
• 24 weeks 29/784 3.70

13
In the Absence of Compelling Data from RCTs

• We need more data from RCTs as well as more data
  from Open and Wide Open Trials. Data from
  either alone is not as compelling. We need data.
• Can we assume efficacy?
• What are legitimate data?

Peter S enter left
14
Problems in Data Collection

• Recruitment
• Who are you recruiting?
• Open Trial vs. Wide Open Trial
• Assessing motivation
• committed quitters vs. tire kickers
• free vs. pay

15
Problems in Data Collection

• Use
• Honesty current research suggests that people
  are at least as honest on-line as they are in
  person
• Self-report vs. objective data (cotinine, CO)
• Use difficult to control how people use the
  web-based program
• Augmentation difficult to know what other
  concurrent interventions Ss are using, for
  example, three other websites plus a smoking
  cessation clinic

16
Problems in Data Collection

• Missing Data
• What does it mean when people stop returning to
  the program/support group/diary function?
• Have they quit or have they slipped or relapsed?

17
Problems in Data Collection

• Confidentiality
• Pros and cons of anonymity
• Electronic entities such as emails are not stable

18
What is Legitimate Data?

• Data that qualifies for clinical research
• Adheres to International Guidelines of Privacy
• Is obtained through express consent
• Is obtained through an unbiased process (no gifts
  or incentives, user fees, subjects must have free
  and consistent access)
• Has been washed to ensure legitimacy

19
Privacy Guideline Summary

• Canada The Personal Information Protection and
  Electronic Documents Act (PIPEDA) January 1,
  2004
• Imposes limitations on the purposes for which an
  organization may collect, use or disclose
  personal information
• All organizations collecting data must implement
  safeguards to protect the privacy of data
• Organizations collecting data are required to
  inform on the purposes for which information is
  being collected and must obtain express consent
  from clients
• United States Health Insurance Portability and
  Accountability Act of 1996 (HIPPA)
• Applies primarily to Health Plans, Health Care
  Clearinghouses and HMOs who transmit health
  information in electronic format
• Privacy Rule protects all individually
  identifiable health information
• No restrictions on the use or disclosure of
  de-identified health information
• European Union Directive 2002/58/EC (Directive
  on Privacy and Electronic Communications)
• Consent may be given by an appropriate method
  enabling a freely given specific and informed
  indication of the users wishes, including by
  ticking a box when visiting an Internet website
• Electronic communications services over the
  Internet should inform users and subscribers of
  measures they the user can take to protect the
  security of their communications
• Service providers should always keep subscribers
  informed of the type of data they are processing
  and the purposes and duration for which this is
  done

20
Privacy Policy Basics

• Subject must be expressly advised as to how data
  is used
• Subject must be permitted to remove data from
  analyzed database
• Subject must be advised on how to obtain
  information on Privacy Practices
• See attached sample (SCC Program Sign-Up
  Procedure, Privacy Policy and Legal Information)
• What if someone uses a program in the US and
  wishes to sue for damages?

21
Checks and Balances

• What survey questions can be asked to help ensure
  legitimacy?
• 1. Number of cigarettes smoked per day
• 2. Number of years smoked
• 3. Dollar amount spent per week on cigarettes
• 4. Age
• 5. Country of residence

• Bad Data
• 3
• 25
• 112.00
• 29
• Canada
• This 29 year old smokes either
• one pack OR 14 packs a week,
• and has smoked since they were
• 4 years old.

• Good Data
• 24
• 18
• 56.00
• 34
• Canada
• This 34 year old smokes a pack a
• day, and has smoked since they
• were 16 years old.

22
Other Limiting Factors

• Time How long did it take for the user to
  complete the program? Is the subject an actual
  quitter or a researcher/physician reviewing the
  program?
• Hours of Access Does the subject have daily
  access to the program? If not, did limited hours
  usage affect success?
• Usability Is the site easy or cumbersome to
  navigate? Does it adhere to common usability,
  navigation and access principles?
• Other Cessation Methods What other programs,
  products or services is the user employing? Can
  the success or failure rates be attributed to
  another smoking cessation intervention?
• Socioeconomic Do all demographics have access to
  the Internet?
• Stage of Change Status Are subjects highly
  motivated quitters who have paid an enrolment fee
  or purchased an NRT product?
• Incentives Did the subject receive payment or
  NRT products as an incentive to participate?
• Geography What is the cultural or societal
  background of the subject?
• Other Illness or Co-morbid Condition What
  percentage of users have a comorbid condition
  that may affect progress through the Stages of
  Change? Was the user forced to quit smoking due
  to another condition or was the quit
  self-imposed?