We conducted a cluster-randomized trial (2005–2008) to evaluate http://www.dtc.cme.uab.edu, a web-assisted tobacco control quality improvement program designed to increase the quality and quantity of tobacco product cessation advice delivered at the point of care by dental providers. A total of 143 participating practices were randomized to either a provider-facingweb-delivered multimodal educational intervention or usual-care wait list control, and completed follow-up data collection. Dental patients who used tobacco products were recruited from the http://www.dtc.cme.uab.edu quality improvement program intervention and control practices. These dental patients consented to participate in six-month follow-up calls to report their tobacco use behavior after receiving dental care. This evaluation was funded by the National Institutes of Health (National Institute of Drug Abuse and the National Institute on Dental and Craniofacial Research), and was approved by the University of Alabama at Birmingham Institutional Review Board (IRB).
Clinical setting and provider sample
Community-based dental practices were recruited from Alabama, Georgia, Florida and North Carolina, identified using dental licensure lists and mailing lists from the National Dental PBRN, a dental practice-based research network . These were community-based dental practices with varying numbers of providers and varying setting, rural and urban. Accounting for clustering of patients within practices, a sample size of 130 practices (65 per arm) was needed to detect a 10% increase in tobacco use cessation advice, comparing intervention and control. Practices were randomized to the intervention or control groups using a permuted block randomization sequence generated by our biostatistician. Of the randomized practices who provided follow-up data (N=143; 70 intervention practices, 73 control practices), 79.2% were solo dental practices, 94.5% were general practice dentistry (others included a periodontal practice) and 78.1% had 4 or less dental hygienist or assistants.
The OralCancerPrevention.org intervention
In the tobacco control quality improvement program intervention practice providers were sent information and instructions to log on to OralCancerPrevention.org. This interactive web-assisted system included educational cases, patient education and practice tools, a forum for chatting, opportunities to ask questions, and a presentation of headlines, with pushed email educational reminders to cue participation , see Additional file 1 for screenshots of the website. Developed by a team of dentists, hygienists, tobacco and health informatics experts, the http://www.dtc.cme.uab.edu website provided the dental practices with strategies for advising patients on tobacco control . Tobacco users in these practices thus received usual dental care and had the opportunity to receive the guideline-concordant tobacco cessation advice provided through oralcancerprevention.org. The educational cases provided information related to the 5A’s tailored approach to smoking cessation (Ask about smoking, Advise to quit, Assess readiness to quit, Assist quitting (providing individualized skills advice and supportive treatments (nicotine replacement therapy), and Arrange follow-up discussions) and the 5R’s (Relevance, Risks, Rewards, Roadblocks, and Repetition) . Tobacco users in control practices received usual dental care and services. Practices were blinded to the target of the intervention to minimize bias related to the effect of observation on dental practice behavior.
Pre-intervention dental practice usual care for tobacco Use cessation
Usual care for tobacco use cessation in the dental practices was reported in baseline, pre-intervention, data collection from patients. Each dental was given 100 patient exit cards, and provided instructions on handing out these exit cards to one hundred consecutive adult patients after their visit. These brief postcard-sized surveys were designed to be completed in one to two minutes while the patient was awaiting follow-up instructions and completing payment, and assessed patient tobacco use, age and gender. These exit cards were used to collect baseline data on dental patients and included questions regarding what the dentist said about smoking (ask and advice). Tobacco users who completed the exit cards indicated whether they had been asked about tobacco use, and whether they had been advised to quit if a tobacco user.
Patient participants and six-month follow-up data
After the http://www.dtc.cme.uab.edu intervention was implemented in dental practices each intervention (and comparison usual care control) practice was again given 100 patient exit cards with the same questions on the postcard size survey as at baseline. These cards also asked tobacco users to indicate on the card if they were willing to be contacted for a follow-up call and, if so, provided their name and telephone number. Once all 100 cards were distributed, the dental practice returned the sealed collection box to the study’s coordinating center. Out of the cards returned, 1,361 were from smokers from intervention practices; 1,210 were from smokers from control practices. The dental practices did not have access to the patient responses after being entered into the sealed collection box.
To assess tobacco use cessation, a six-month follow-up telephone survey was conducted among those patients who completed the exit card and indicated willingness to participate by providing a phone number. Participation is summarized in Additional file 2. Patients were called and asked if they would be willing to complete a 10-minute follow-up survey. After verbal consent was obtained, the survey was conducted and each participant was mailed a $10.00 gift card for completing the survey. The six-month follow-up survey confirmed general demographic characteristics, and six-month tobacco use behavior.
Thus, the main dependent (outcome) variable in this report is six-month point prevalent tobacco use abstinence as reported in the follow-up phone survey. The specific question was “Do you smoke cigarettes, cigars, or use smokeless tobacco (dip, chew or snuff) now? (Modification of Behavioral Risk Factor Surveillance System, BRFSS Historical Questions, 2003) We calculated a range of effect sizes for different sample sizes agreeing to participate in follow-up. With the sample size we achieved (over 280 per group), we would have had 80% power to detect a difference of 9% six month tobacco product use quitting, assuming a base rate of quitting of 11%.
The http://www.dtc.cme.uab.edu study was a cluster-randomized trial of a practice-level quality improvement intervention. Thus, the level of randomization was at the practice level and dental patients who used tobacco, the subjects of this analysis, were clustered within these practices. Because cluster-randomized trials are at risk from imbalance of characteristics at levels below the level of randomization (e.g.: patient demographics), we first compared the demographic characteristics of dental patients who used tobacco at baseline recruited from intervention and control practices.
To confirm the impact of our http://www.dtc.cme.uab.edu intervention on dental practices, we then compared patient’s baseline report of their provider performance of asking about tobacco use cessation and advising the tobacco user to quit. We also compared tobacco users’ six-month follow-up reports related to attitudes about tobacco and use of over-the-counter nicotine replacement therapy , comparing tobacco users for intervention and control practices.
To test our main hypothesis, that tobacco users seen at http://www.dtc.cme.uab.edu intervention dental practices would have higher rates of six-month tobacco use cessation (as compared with those tobacco users seen at control practices) we began by comparing the proportion of tobacco users reporting point prevalent tobacco use abstinence at six months using a chi-square test. For our primary analysis, we used an intent-to-treat approach  including patients regardless of whether they actually reported that their dental provider advised them to quit tobacco. In accordance with current guidelines for analysis of tobacco use cessation trials from the Society for Research on Nicotine and Tobacco Research, our intent-to-treat analysis assigned tobacco users who did not complete six-month follow-up as continued using tobacco (not abstinent) . Because patients were clustered within practices, we accounted for variance inflation due to the clustering by conducting a generalized latent linear and mixed model (GLLMM), with a logit link. The GLLMM regression model was tested with multiple correlation matrices, and was confirmed with a generalized estimating equation. As GLLMM is more robust to variation in cluster size (as in this study), the results from GLLMM are used as the main outcomes. We further adjusted the GLLMM regression model for patient variables (demographic characteristics) that were not balanced in the intervention and control practice patient groups. Again, the main dependent variable was patient-reported six-month point-prevalent tobacco use cessation abstinence and the independent variable was intervention or control practice status. All analyses were conducted using the STATA statistical program, version 11.