Persistence of Smoking-Cessation Decision Support Use in a Dental Practice Thomas E. Kottke, MD, MSPH, D. Brad Rindal, DDS, William A. Rush, PhD, Stephen E. Asche, MA, Chris J. Enstad, BS Introduction: A computer-assisted tobacco decision support tool increased dental practitioners’ (dentists and dental hygienists) advice to quit smoking and referral to a quitline during a group randomized trial. The purpose of this study is to document the extent to which use persisted after the trial. Methods: Electronic dental record (EDR) data from 2010 to 2013 were analyzed in 2014 for use of computer-assisted tobacco intervention tool advice scripts and referral to a quitline during four periods: during the trial and post-trial when only intervention clinic dental practitioners had access to the tool, and during full deployment, both before and after an EDR modification. Results: Intervention clinic dental practitioners (18.5 dentist full-time equivalents [FTEs] and 27.8 dental hygienist FTEs practicing in seven clinics) referred 19.0% of 1,368 smokers to a quitline during the trial and referred 15.4% of 4,011 smokers post-trial. After full tool deployment but pre-EDR change, these dental practitioners referred 15.6% of 2,214 intervention clinic smokers, whereas 18.3 dentist FTEs and 29.7 dental hygienist FTEs practicing in eight clinics referred 8.5% of 2,113 smokers. Post-EDR change, dental practitioners referred 12.2% of 2,214 intervention clinic smokers and 8.1% of 2,399 control clinic smokers to a quitline. In the last three quarters of observation, clinic script use ranged from 15.4% to 65.8% and referral to a quitline ranged from 2.0% to 18.7% of visits.
Conclusions: Although EDR design affected rates of referral, dental practitioners persisted in using a computer-assisted tobacco intervention tool to refer smokers to a quitline. (Am J Prev Med 2015;48(6):722–728) & 2015 American Journal of Preventive Medicine
Introduction
T
he 2008 update to the clinical practice guideline for treating tobacco use and dependence concluded that brief tobacco treatment is effective,1 and randomized trials have demonstrated that smokingcessation advice delivered in dental offices is efficacious.2–5 The guideline also concluded that tobacco quitlines are efficacious tools that help smokers quit. In addition to efficacy, however, the effectiveness of any intervention is influenced by at least four other factors: reach, adoption, implementation, and maintenance.6 The effectiveness of tobacco quitlines is limited by their From the HealthPartners Institute for Education and Research, Minneapolis, Minnesota Address correspondence to: Thomas E. Kottke, MD, MSPH, HealthPartners Institute for Education and Research, 8170 33rd Avenue South, Mail Stop 21110X, Minneapolis MN 55425. E-mail: thomas.e.kottke@ healthpartners.com. 0749-3797/$36.00 http://dx.doi.org/10.1016/j.amepre.2014.12.017
722 Am J Prev Med 2015;48(6):722–728
failure to reach smokers.7–9 This is also true for e-referrals to Internet-based smoking-cessation support programs.10,11 The context of the dental visit makes it a potential tool to increase quitline reach: Tobacco use has a strong negative effect on dental and oral health; there are nearly as many dental hygienists practicing in the U.S. (181,800) as primary care physicians (209,000)12,13; the number of dental hygienists is expected to increase to 250,000 by 202012; more than 60% of Americans visit a dental office each year14; and, in contrast to the 20minute primary care doctor’s visit,15 dental visits tend to be much longer and generally include education on improving oral health.16 In a 2011 Cochrane Review, Boyle et al.17 concluded that when physicians are expected to use the electronic health record to record and treat patient tobacco use at medical visits, their activity increases. HealthPartners Dental Group had developed a computer-assisted tobacco intervention (CATI) tool that is embedded in the electronic dental record (EDR) that, as originally
& 2015 American Journal of Preventive Medicine
Published by Elsevier Inc.
Kottke et al / Am J Prev Med 2015;48(6):722–728
deployed, required the dental practitioner (dentist or dental hygienist) to assess four health behavior variables for all patients who reported smoking cigarettes: two questions from the Heaviness of Smoking Index18 (number of cigarettes smoked daily and how soon on waking the first cigarette was smoked); interest in quitting; and number of previous quit attempts. The EDR automatically calculated the patient’s dependency level based on the Heaviness of Smoking Index, used rule-based algorithms to generate patient-centered evidence-based practitioner scripts, and displayed both the dependency level and scripts in the health history screen. The dental practitioner could use these scripts to guide their advice to the patient. During a 5-month randomized trial between November 15, 2010, and April 15, 2011, deploying the CATI tool nearly doubled the frequency at which smokers reported that their dental practitioner suggested specific strategies for quitting.19 The current report describes the extent to which intervention clinic practitioners maintained use of the CATI and referred smokers to a quitline after the trial ended. Adoption by control clinic practitioners after they were given access to the CATI tool is also reported. Finally, because an EDR update resulted in making use of the CATI optional, the change in dental practitioner activity after the EDR was modified is described.
723
Methods The HealthPartners Institute for Education and Research IRB approved this study. The analysis was conducted in 2014. The analytic denominator consisted of 13,993 encounters during the time period from November 15, 2010, to December 31, 2013, for routine, new, prophylaxis, and periodontal visits by 8,967 patients aged Z18 years who were listed in the EDR as current smokers at the time of the encounter. During the period of observation, the intervention group comprised 18.5 dentist full-time equivalents (FTEs) and 27.8 dental hygienist FTEs practicing in seven clinics, and the control group included 18.3 dentist FTEs and 29.7 dental hygienist FTEs practicing in eight clinics. When reporting activity by FTE, only the FTEs for dental hygienist are reported, so that dental practitioner activity is not overestimated by reporting both dentist and dental hygienist activity at the time of the same encounter. The proportion of smoker visits with script use and the proportion with a quitline referral are reported by month and divided into four analysis periods (Figures 1 and 2). The first period, the “trial” period, was from November 15, 2010, to April 15, 2011. During this period, only the intervention clinic practitioners had access to the CATI tool. Patients in both the intervention and control clinics were surveyed at this time to assess practitioner actions. In the second period, “post-trial,” only intervention clinic practitioners had access to the tool, and patients were not surveyed. The third time period, “pre-EDR change,” was a 7.5-month interval. During this time, the CATI was available to both intervention and control clinic practitioners, and the
Figure 1. The proportion of encounters with smoking cessation script use during four time periods for intervention group clinics and control group clinics. EDR, electronic dental record.
June 2015
724
Kottke et al / Am J Prev Med 2015;48(6):722–728
Figure 2. The proportion of encounters during which a smoker was referred to a quitline during four time periods for intervention group clinics and control group clinics. EDR, electronic dental record.
practitioners were required to update the smoking history by asking four questions: the number of cigarettes smoked daily, how soon on waking the first cigarette was smoked, interest in quitting, and the number of previous quit attempts. During the fourth period, “postEDR change” from March 7, 2013, to December 31, 2013, dental practitioners were able to update the dental record from the most recent prior visit with a single key stroke, and the CATI tool was not activated unless the dental practitioner manually entered responses to the tobacco questions and clicked on a button that opened the scripts routine. The proportion of encounters with script use and quitline referrals was plotted by month by study group for all time periods. When summarizing CATI tool use by study group and time period in the table, the exact time period date cut offs were used as described above. A generalized estimating equation (GEE) approach with empirical SEs to accommodate multiple visits per patient was used to analyze differences by study group and time period in proportion with script use and quitline referrals. The same GEE approach and a model that included terms for study group (or time period); month; and an interaction of these two terms were used to test differences in slopes by study group or time period. Variation in activity within and between clinics for the last three quarters of observation is described, but statistical testing was not applied to these data.
Results When the 13,993 encounters for 8,967 patients were stratified by study arm and month-long time period, the
cells had a median of 255 encounters and the number of encounters with smokers in the cells ranged from 194 to 323. Both the use of any tobacco intervention script and referral to a quitline were greatest with new patients seen in intervention clinics during the intervention period (Table 1). A control clinic dental practitioner used a script with about half of smoking patients during the initial 5 months of access to the CATI tool and referred 8.5% of smokers to a quitline. Although dental practitioners used scripts less frequently in the post-EDR change period, intervention group clinics still used scripts with 50% of new patients, and control group clinics used scripts with 31% of new patients. On average, clinicians in both groups referred more than 8% of their patients to a quitline in the postEDR change period. The number of quitline referrals per dental hygienist FTE per month was highest (1.88) during the trial period, and declined to 1.45 in the post-trial period, 1.39 in the pre-EDR change period, and 0.98 in the post-EDR change period in the intervention clinics. In control clinics, there were 0.82 quitline referrals per dental hygienist FTE per month pre-EDR change and 0.66 quitline referrals per dental hygienist FTE per month post-EDR change. www.ajpmonline.org
Kottke et al / Am J Prev Med 2015;48(6):722–728
725
a
Table 1. The Number of Visits by Smokers by Visit Type, Advice Rates, and Quitline Referral Rates Intervention clinics Time periodsb Duration (months)
Control clinics
During the trial
Posttrial
Pre-EDR change
Post-EDR change
Pre-EDR change
Post-EDR change
5
15
7.5
9.5
7.5
9.5
Number of visits with smokers by visit type during each time period Routine
844
2,481
1,096
1,401
1,248
1,469
New
415
1,223
594
634
577
669
Prophylaxis
66
191
81
97
198
173
Periodontal
44
116
56
82
90
88
1,369
4,011
1,827
2,214
2,113
2,399
All types
Visits during which a script was used to counsel the patient (%) Routine visits
56.4
61.1
63.4
51.3
43.9
30.2
New visits
62.4
71.0
69.2
50.2
50.4
31.5
Prophylaxis
25.8
30.4
46.9
25.8
29.8
17.9
Periodontal
25.0
31.0
42.9
28.1
41.1
23.9
All types
55.7
61.8
63.9
49.0
44.3
29.5
Visits during which the patient was referred to a quitline (%) Routine
18.8
13.7
13.3
11.4
7.8
7.6
New
21.7
20.9
20.9
14.8
11.1
10.6
Prophylaxis
15.2
7.3
13.6
13.4
5.6
5.2
Periodontal
2.3
8.6
7.1
3.7
8.9
2.3
19.0
15.4
15.6
12.2
8.5
8.1
All types a
The following CDT-5 visit codes defined the visit types: Routine (0120, 0120C); New (0150, 0150C); Prophylaxis (1110X); Periodontal (0180, 49103, 49104, 49105, 49106, 4910X). Time periods: during the trial, November 15, 2010, to April 15, 2011; post-trial, April 16, 2011, to July 25, 2012; pre-EDR change, July 26, 2012, to March 6, 2013; post-EDR change, March 7, 2013, to December 31, 2013. EDR, electronic dental record. b
The slope for script use was positive during all time periods except pre-EDR change in the intervention group (C), which had a slightly negative slope (OR¼0.989, p¼0.72) (Figure 1). Only the slopes for trial (A, OR¼1.255, po0.001) and post-trial (B, OR¼1.021, p¼0.003) were significantly different from horizontal. There was no evidence of a quadratic trend in any time period (all p40.09). The proportion of visits with script use post-trial (B) was higher than that during the trial (A, OR¼1.277, po0.001), and the positive slope during the trial period (A, OR¼1.255) was higher than that during the post-trial period (B, OR¼1.021, po0.001). The intervention group during the pre-EDR change period (C) had higher script use than during the post-EDR change time period (D), and script use higher than the control group pre-EDR change (E) (all po0.001). June 2015
Compared to pre-EDR change script use (E), the control group had lower script use post-EDR change (F, po0.001). Post-EDR change script use by the control group (F) was also lower than script use by the intervention group during the same period (D, po0.001). There were no pairwise differences in slopes for time periods C, D, E, or F. Although the slopes for quitline referrals during the post-trial time period (B) and pre-EDR change period tended toward negative for the intervention group (C), and all of the other slopes tended toward positive, none of the slopes were significantly different from horizontal (all p40.08) (Figure 2). Statistically significant quadratic trends were present for the intervention group in the post-trial period (B, p¼0.001) and for the control group in the pre-EDR change period (E, p¼0.03). Although the
726
Kottke et al / Am J Prev Med 2015;48(6):722–728
trial (A) and post-trial (B) slopes did not differ, quitline referrals were lower in the post-trial period (B) compared with the trial period (A, OR¼0.837, p¼0.04). In the intervention group, the pre-EDR change period (C) had higher quitline referral rates than the post-EDR change period (D), and the pre-EDR change period referral rate in the intervention group was higher than the pre-EDR change time period in the control group (E) (both po0.04). Although the quitline referral rate post-EDR change in the control group (F) was lower than the quitline referral rates by the intervention group during the same period (D, po0.001), the referral rate by the control group post-EDR change (F) was not different from the control group referral rate pre-EDR change (E, p¼0.92). There were no pairwise differences in slopes for time periods C, D, E, or F when treating all as linear (all p40.07). During the three quarters of the post-EDR change period, the percentage of visits with script use varied by clinic from 15.4% at the lowest, 41.1% at the median, and 65.8% at the highest. Variation in script use over time within each clinic was as small as 2 percentage points in one clinic and as large as 28 percentage points in another clinic. Referral to a quitline across clinics ranged from 2.0% to 18.7%, with a median of 8.9%.
Discussion Follow-up of the randomized trial phase of the CATI tool documented that the dental providers persisted in using scripts and referring patients to a quitline at a relatively high level, until the EDR was modified to allow the past dental history to be loaded into the current visit with a single key stroke and the use of the CATI became optional. This change was associated with a 20–percentage point decline in the use of scripts and a lesser decline in referrals to a quitline. Use of scripts by control group dental practitioners was about 20 percentage points lower than the use of scripts by intervention group dental practitioners both before and after the change in the EDR. The reason for this difference in performance may be due to the fact that the control group training was not as intensive as the training given to the intervention group. As did the intervention group dental practitioners, the control group dental practitioners reduced their activity after the change in the EDR. Although control group dental practitioners referred patients to a quitline less frequently, their referral rate did not change with the change in the EDR. In the last three quarters of observation, there was considerable variation in script use and referral to a quitline both within and between clinics. These observations suggest that making a CATI
tool a standard feature of EDRs would significantly increase the proportion of smokers who are counseled and referred to a quitline. Several meta-analyses of computer-assisted decision support systems have concluded that the systems can improve health professional behavior.20–22 However, the effectiveness of the systems varies. The source of at least some of this variation can be explained by examining the results of the current trial in the context of the behavioral model of B.J. Fogg of Stanford University.23 He posits that the presence of three components is both necessary and sufficient for a behavior to take place: ability, motivation, and a trigger. The natural experiment described in this paper is consistent with his hypothesis. The ability to use scripts and refer to a quitline did not change, so ability cannot possibly explain the change in behavior. It is possible that the dental practitioners coincidentally experienced a marked reduction in their motivation to counsel and refer patients at the time of the EDR change, but there is no evidence to support this hypothesis. The most compelling explanation is that the removal of the trigger—the requirement to ask and advise, and the opportunity to refer—is the explanation for the sudden decline in dental provider activity. The decline in activity after the CATI was no longer activated automatically suggests that restoring this trigger might permanently increase the use of scripts and referral to the quitline. The variation in activity between and within clinics suggests that activity might also be increased to a significant extent if other triggers can be identified and implemented in dental practices. Randomized trials have already demonstrated that e-mails providing educational material or performance feedback can trigger smoking-cessation advice.24,25 Another effective trigger is academic detailing.26 Perhaps the most significant limitation to the data analyzed here is that neither smoking-cessation rates nor quitline contact and engagement rates are available; only dental practitioner activity is. It is not possible to discern from the data set whether a dentist, a dental hygienist, or both used the scripts when a script was used. It is also not possible to precisely define who made a referral when one was made. The data are also limited by the fact that they are derived from a single dental group in the Midwestern U.S. followed over a limited time period. The experience reported here may not apply to all dental practitioners or all EDRs. The difference in dental practitioner behavior between intervention and control groups and the change in dental practitioner behavior associated with the change in the EDR suggests that dental practitioner behavior is very sensitive to both training and EDR design. It is interesting that whereas 37% of intervention clinic smokers who were interviewed during the trial reported www.ajpmonline.org
Kottke et al / Am J Prev Med 2015;48(6):722–728
that they were referred to a quitline, the rate calculated from medical record review was only 19%. There are at least two explanations for this discrepancy. A study in Australia that audio recorded entire medical encounters and compared the recordings to patient reports found that patients over-reported physician smoking-cessation counseling rates.27 The same discrepancy could be present in the HealthPartners Dental Group experience. It is also possible that the dental practitioners referred patients to a quitline even when not prompted by a script, or they failed to record their referral in the dental record. It has been documented that medical records do not accurately reflect practitioner smoking-cessation activity.27 Despite the fact that smoking has a powerful negative impact on oral health, dental practitioners have historically been less active and have felt less qualified than physicians to deal with smoking.28 The data presented here suggest that a CATI tool integrated into the EDR could significantly increase the proportion of patients who are advised to quit smoking, assisted in their efforts to quit smoking, and referred to a quitline. This is particularly true if use of the CATI with patients who smoke were obligatory.
7. 8.
9.
10.
11.
12. 13.
14.
15.
16.
CDC, Office on Smoking and Health, under contract number 200-2009-28537, provided the sole financial support for this analysis. No financial disclosures were reported by the authors of this paper.
17.
18.
References 1. Fiore MC, Jaén CR, Baker TB, et al. Treating Tobacco Use and Dependence: 2008 Update. Clinical Practice Guideline. Rockville, MD: USDHHS. Public Health Service, 2008. 2. Nohlert E, Tegelberg A, Tillgren P, Johansson P, Rosenblad A, Helgason AR. Comparison of a high and a low intensity smoking cessation intervention in a dentistry setting in Sweden—a randomized trial. BMC Public Health. 2009;9:121. http://dx.doi.org/10.1186/ 1471-2458-9-121. 3. Nohlert E, Ohrvik J, Tegelberg A, Tillgren P, Helgason AR. Long-term follow-up of a high- and a low-intensity smoking cessation intervention in a dental setting—a randomized trial. BMC Public Health. 2013;13(1):592. http://dx.doi.org/10.1186/1471-2458-13-592. 4. Gordon JS, Andrews JA, Crews KM, Payne TJ, Severson HH. The 5Aʼs vs 3Aʼs plus proactive quitline referral in private practice dental offices: preliminary results. Tob Control. 2007;16(4):285–288. http://dx.doi.org/ 10.1136/tc.2007.020271. 5. Gordon JS, Andrews JA, Albert DA, Crews KM, Payne TJ, Severson HH. Tobacco cessation via public dental clinics: results of a randomized trial. Am J Public Health. 2010;100(7):1307–1312. http://dx.doi.org/ 10.2105/AJPH.2009.181214. 6. Glasgow RE, Vogt TM, Boles SM. Evaluating the public health impact of health promotion interventions: the RE-AIM framework.
June 2015
19.
20.
21.
22.
23. 24.
727
Am J Public Health. 1999;89(9):1322–1327. http://dx.doi.org/10.2105/ AJPH.89.9.1322. Keller PA, Feltracco A, Bailey LA, et al. Changes in tobacco quitlines in the United States, 2005-2006. Prev Chronic Dis. 2010;7(2):A36. Cummins SE, Bailey L, Campbell S, Koon-Kirby C, Zhu SH. Tobacco cessation quitlines in North America: a descriptive study. Tob Control. 2007;16(suppl 1):i9–i15. http://dx.doi.org/10.1136/tc.2007.020370. Kaufman A, Augustson E, Davis K, Finney Rutten LJ. Awareness and use of tobacco quitlines: evidence from the Health Information National Trends Survey. J Health Commun. 2010;15(suppl 3): 264–278. http://dx.doi.org/10.1080/10810730.2010.526172. Ray MN, Funkhouser E, Williams JH, et al. Smoking-cessation e-referrals: a national dental practice-based research network randomized controlled trial. Am J Prev Med. 2014;46(2):158–165. http://dx.doi.org/10.1016/ j.amepre.2013.10.018. Sadasivam RS, Hogan TP, Volkman JE, et al. Implementing point of care “e-referrals” in 137 clinics to increase access to a quit smoking internet system: the Quit-Primo and National Dental PBRN HI-QUIT Studies. Transl Behav Med. 2013;3(4):370–378. http://dx.doi.org/10.1007/ s13142-013-0230-3. Occupational Outlook Handbook. 2012; www.bls.gov/ooh/healthcare/ dental-hygienists.htm. The Number of Practicing Primary Care Physicians in the United States. www.ahrq.gov/research/findings/factsheets/primary/pcwork1/ index.html. Health, United States, 2012. With Special Feature on Emergency Care. Table 90. Hyattsville, MD: National Center for Health Statistics, 2013. Abbo ED, Zhang Q, Zelder M, Huang ES. The increasing number of clinical items addressed during the time of adult primary care visits. J Gen Intern Med. 2008;23(12):2058–2065. http://dx.doi.org/ 10.1007/s11606-008-0805-8. Wotman S, Demko CA, Victoroff K, Sudano JJ, Lalumandier JA. A multimethod investigation including direct observation of 3751 patient visits to 120 dental offices. Clin Cosmet Investig Dent. 2010;2: 27–39. http://dx.doi.org/10.2147/CCIDEN.S6592. Boyle R, Solberg L, Fiore M. Use of electronic health records to support smoking cessation. Cochrane Database Syst Rev. 2011;12: CD008743. http://dx.doi.org/10.1002/14651858.CD008743.pub2. Heatherton TF, Kozlowski LT, Frecker RC, Rickert W, Robinson J. Measuring the heaviness of smoking: using self-reported time to the first cigarette of the day and number of cigarettes smoked per day. Br J Addict. 1989;84(7):791–799. http://dx.doi.org/10.1111/ j.1360-0443.1989.tb03059.x. Rindal DB, Rush WA, Schleyer TK, et al. Computer-assisted guidance for dental office tobacco-cessation counseling: a randomized controlled trial. Am J Prev Med. 2013;44(3):260–264. http://dx.doi.org/10.1016/ j.amepre.2012.10.023. Hunt DL, Haynes RB, Hanna SE, Smith K. Effects of computer-based clinical decision support systems on physician performance and patient outcomes: a systematic review. JAMA. 1998;280(15): 1339–1346. http://dx.doi.org/10.1001/jama.280.15.1339. Garg AX, Adhikari NK, McDonald H, et al. Effects of computerized clinical decision support systems on practitioner performance and patient outcomes: a systematic review. JAMA. 2005;293(10): 1223–1238. http://dx.doi.org/10.1001/jama.293.10.1223. Boyle JP, Honeycutt AA, Narayan KMV, et al. Projection of diabetes burden through 2050: impact of changing demography and disease prevalence in the U.S. Diabetes Care. 2001;24(11):1936–1940. http://dx.doi.org/ 10.2337/diacare.24.11.1936. Fogg BJ. BJ Foggʼs Behavior Model. Stanford University. www. behaviormodel.org/. Houston TK, Richman JS, Ray MN, et al. Internet delivered support for tobacco control in dental practice: randomized controlled trial. J Med Internet Res. 2008;10(5):e38. http://dx.doi.org/10.2196/ jmir.1095.
728
Kottke et al / Am J Prev Med 2015;48(6):722–728
25. Houston TK, Coley HL, Sadasivam RS, et al. Impact of content-specific email reminders on provider participation in an online intervention: a dental PBRN study. Stud Health Technol Inform. 2010;160(Pt 2):801–805. 26. Sheffer MA, Baker TB, Fraser DL, Adsit RT, McAfee TA, Fiore MC. Fax referrals, academic detailing, and tobacco quitline use: a randomized trial. Am J Prev Med. 2011;42(1):21–28. http://dx.doi.org/ 10.1016/j.amepre.2011.08.028.
27. Ward J, Sanson-Fisher R. Accuracy of patient recall of opportunistic smoking cessation advice in general practice. Tob Control. 1996;5(2): 110–113. http://dx.doi.org/10.1136/tc.5.2.110. 28. Secker-Walker RH, Solomon LJ, Flynn BS, Dana GS. Comparisons of the smoking cessation counseling activities of six types of health professionals. Prev Med. 1994;23(6):800–808. http://dx.doi.org/ 10.1006/pmed.1994.1137.
www.ajpmonline.org
Conclusions: Although EDR design affected rates of referral, dental practitioners persisted in using a computer-assisted tobacco intervention tool to refer smokers to a quitline. (Am J Prev Med 2015;48(6):722–728) & 2015 American Journal of Preventive Medicine
Introduction
T
he 2008 update to the clinical practice guideline for treating tobacco use and dependence concluded that brief tobacco treatment is effective,1 and randomized trials have demonstrated that smokingcessation advice delivered in dental offices is efficacious.2–5 The guideline also concluded that tobacco quitlines are efficacious tools that help smokers quit. In addition to efficacy, however, the effectiveness of any intervention is influenced by at least four other factors: reach, adoption, implementation, and maintenance.6 The effectiveness of tobacco quitlines is limited by their From the HealthPartners Institute for Education and Research, Minneapolis, Minnesota Address correspondence to: Thomas E. Kottke, MD, MSPH, HealthPartners Institute for Education and Research, 8170 33rd Avenue South, Mail Stop 21110X, Minneapolis MN 55425. E-mail: thomas.e.kottke@ healthpartners.com. 0749-3797/$36.00 http://dx.doi.org/10.1016/j.amepre.2014.12.017
722 Am J Prev Med 2015;48(6):722–728
failure to reach smokers.7–9 This is also true for e-referrals to Internet-based smoking-cessation support programs.10,11 The context of the dental visit makes it a potential tool to increase quitline reach: Tobacco use has a strong negative effect on dental and oral health; there are nearly as many dental hygienists practicing in the U.S. (181,800) as primary care physicians (209,000)12,13; the number of dental hygienists is expected to increase to 250,000 by 202012; more than 60% of Americans visit a dental office each year14; and, in contrast to the 20minute primary care doctor’s visit,15 dental visits tend to be much longer and generally include education on improving oral health.16 In a 2011 Cochrane Review, Boyle et al.17 concluded that when physicians are expected to use the electronic health record to record and treat patient tobacco use at medical visits, their activity increases. HealthPartners Dental Group had developed a computer-assisted tobacco intervention (CATI) tool that is embedded in the electronic dental record (EDR) that, as originally
& 2015 American Journal of Preventive Medicine
Published by Elsevier Inc.
Kottke et al / Am J Prev Med 2015;48(6):722–728
deployed, required the dental practitioner (dentist or dental hygienist) to assess four health behavior variables for all patients who reported smoking cigarettes: two questions from the Heaviness of Smoking Index18 (number of cigarettes smoked daily and how soon on waking the first cigarette was smoked); interest in quitting; and number of previous quit attempts. The EDR automatically calculated the patient’s dependency level based on the Heaviness of Smoking Index, used rule-based algorithms to generate patient-centered evidence-based practitioner scripts, and displayed both the dependency level and scripts in the health history screen. The dental practitioner could use these scripts to guide their advice to the patient. During a 5-month randomized trial between November 15, 2010, and April 15, 2011, deploying the CATI tool nearly doubled the frequency at which smokers reported that their dental practitioner suggested specific strategies for quitting.19 The current report describes the extent to which intervention clinic practitioners maintained use of the CATI and referred smokers to a quitline after the trial ended. Adoption by control clinic practitioners after they were given access to the CATI tool is also reported. Finally, because an EDR update resulted in making use of the CATI optional, the change in dental practitioner activity after the EDR was modified is described.
723
Methods The HealthPartners Institute for Education and Research IRB approved this study. The analysis was conducted in 2014. The analytic denominator consisted of 13,993 encounters during the time period from November 15, 2010, to December 31, 2013, for routine, new, prophylaxis, and periodontal visits by 8,967 patients aged Z18 years who were listed in the EDR as current smokers at the time of the encounter. During the period of observation, the intervention group comprised 18.5 dentist full-time equivalents (FTEs) and 27.8 dental hygienist FTEs practicing in seven clinics, and the control group included 18.3 dentist FTEs and 29.7 dental hygienist FTEs practicing in eight clinics. When reporting activity by FTE, only the FTEs for dental hygienist are reported, so that dental practitioner activity is not overestimated by reporting both dentist and dental hygienist activity at the time of the same encounter. The proportion of smoker visits with script use and the proportion with a quitline referral are reported by month and divided into four analysis periods (Figures 1 and 2). The first period, the “trial” period, was from November 15, 2010, to April 15, 2011. During this period, only the intervention clinic practitioners had access to the CATI tool. Patients in both the intervention and control clinics were surveyed at this time to assess practitioner actions. In the second period, “post-trial,” only intervention clinic practitioners had access to the tool, and patients were not surveyed. The third time period, “pre-EDR change,” was a 7.5-month interval. During this time, the CATI was available to both intervention and control clinic practitioners, and the
Figure 1. The proportion of encounters with smoking cessation script use during four time periods for intervention group clinics and control group clinics. EDR, electronic dental record.
June 2015
724
Kottke et al / Am J Prev Med 2015;48(6):722–728
Figure 2. The proportion of encounters during which a smoker was referred to a quitline during four time periods for intervention group clinics and control group clinics. EDR, electronic dental record.
practitioners were required to update the smoking history by asking four questions: the number of cigarettes smoked daily, how soon on waking the first cigarette was smoked, interest in quitting, and the number of previous quit attempts. During the fourth period, “postEDR change” from March 7, 2013, to December 31, 2013, dental practitioners were able to update the dental record from the most recent prior visit with a single key stroke, and the CATI tool was not activated unless the dental practitioner manually entered responses to the tobacco questions and clicked on a button that opened the scripts routine. The proportion of encounters with script use and quitline referrals was plotted by month by study group for all time periods. When summarizing CATI tool use by study group and time period in the table, the exact time period date cut offs were used as described above. A generalized estimating equation (GEE) approach with empirical SEs to accommodate multiple visits per patient was used to analyze differences by study group and time period in proportion with script use and quitline referrals. The same GEE approach and a model that included terms for study group (or time period); month; and an interaction of these two terms were used to test differences in slopes by study group or time period. Variation in activity within and between clinics for the last three quarters of observation is described, but statistical testing was not applied to these data.
Results When the 13,993 encounters for 8,967 patients were stratified by study arm and month-long time period, the
cells had a median of 255 encounters and the number of encounters with smokers in the cells ranged from 194 to 323. Both the use of any tobacco intervention script and referral to a quitline were greatest with new patients seen in intervention clinics during the intervention period (Table 1). A control clinic dental practitioner used a script with about half of smoking patients during the initial 5 months of access to the CATI tool and referred 8.5% of smokers to a quitline. Although dental practitioners used scripts less frequently in the post-EDR change period, intervention group clinics still used scripts with 50% of new patients, and control group clinics used scripts with 31% of new patients. On average, clinicians in both groups referred more than 8% of their patients to a quitline in the postEDR change period. The number of quitline referrals per dental hygienist FTE per month was highest (1.88) during the trial period, and declined to 1.45 in the post-trial period, 1.39 in the pre-EDR change period, and 0.98 in the post-EDR change period in the intervention clinics. In control clinics, there were 0.82 quitline referrals per dental hygienist FTE per month pre-EDR change and 0.66 quitline referrals per dental hygienist FTE per month post-EDR change. www.ajpmonline.org
Kottke et al / Am J Prev Med 2015;48(6):722–728
725
a
Table 1. The Number of Visits by Smokers by Visit Type, Advice Rates, and Quitline Referral Rates Intervention clinics Time periodsb Duration (months)
Control clinics
During the trial
Posttrial
Pre-EDR change
Post-EDR change
Pre-EDR change
Post-EDR change
5
15
7.5
9.5
7.5
9.5
Number of visits with smokers by visit type during each time period Routine
844
2,481
1,096
1,401
1,248
1,469
New
415
1,223
594
634
577
669
Prophylaxis
66
191
81
97
198
173
Periodontal
44
116
56
82
90
88
1,369
4,011
1,827
2,214
2,113
2,399
All types
Visits during which a script was used to counsel the patient (%) Routine visits
56.4
61.1
63.4
51.3
43.9
30.2
New visits
62.4
71.0
69.2
50.2
50.4
31.5
Prophylaxis
25.8
30.4
46.9
25.8
29.8
17.9
Periodontal
25.0
31.0
42.9
28.1
41.1
23.9
All types
55.7
61.8
63.9
49.0
44.3
29.5
Visits during which the patient was referred to a quitline (%) Routine
18.8
13.7
13.3
11.4
7.8
7.6
New
21.7
20.9
20.9
14.8
11.1
10.6
Prophylaxis
15.2
7.3
13.6
13.4
5.6
5.2
Periodontal
2.3
8.6
7.1
3.7
8.9
2.3
19.0
15.4
15.6
12.2
8.5
8.1
All types a
The following CDT-5 visit codes defined the visit types: Routine (0120, 0120C); New (0150, 0150C); Prophylaxis (1110X); Periodontal (0180, 49103, 49104, 49105, 49106, 4910X). Time periods: during the trial, November 15, 2010, to April 15, 2011; post-trial, April 16, 2011, to July 25, 2012; pre-EDR change, July 26, 2012, to March 6, 2013; post-EDR change, March 7, 2013, to December 31, 2013. EDR, electronic dental record. b
The slope for script use was positive during all time periods except pre-EDR change in the intervention group (C), which had a slightly negative slope (OR¼0.989, p¼0.72) (Figure 1). Only the slopes for trial (A, OR¼1.255, po0.001) and post-trial (B, OR¼1.021, p¼0.003) were significantly different from horizontal. There was no evidence of a quadratic trend in any time period (all p40.09). The proportion of visits with script use post-trial (B) was higher than that during the trial (A, OR¼1.277, po0.001), and the positive slope during the trial period (A, OR¼1.255) was higher than that during the post-trial period (B, OR¼1.021, po0.001). The intervention group during the pre-EDR change period (C) had higher script use than during the post-EDR change time period (D), and script use higher than the control group pre-EDR change (E) (all po0.001). June 2015
Compared to pre-EDR change script use (E), the control group had lower script use post-EDR change (F, po0.001). Post-EDR change script use by the control group (F) was also lower than script use by the intervention group during the same period (D, po0.001). There were no pairwise differences in slopes for time periods C, D, E, or F. Although the slopes for quitline referrals during the post-trial time period (B) and pre-EDR change period tended toward negative for the intervention group (C), and all of the other slopes tended toward positive, none of the slopes were significantly different from horizontal (all p40.08) (Figure 2). Statistically significant quadratic trends were present for the intervention group in the post-trial period (B, p¼0.001) and for the control group in the pre-EDR change period (E, p¼0.03). Although the
726
Kottke et al / Am J Prev Med 2015;48(6):722–728
trial (A) and post-trial (B) slopes did not differ, quitline referrals were lower in the post-trial period (B) compared with the trial period (A, OR¼0.837, p¼0.04). In the intervention group, the pre-EDR change period (C) had higher quitline referral rates than the post-EDR change period (D), and the pre-EDR change period referral rate in the intervention group was higher than the pre-EDR change time period in the control group (E) (both po0.04). Although the quitline referral rate post-EDR change in the control group (F) was lower than the quitline referral rates by the intervention group during the same period (D, po0.001), the referral rate by the control group post-EDR change (F) was not different from the control group referral rate pre-EDR change (E, p¼0.92). There were no pairwise differences in slopes for time periods C, D, E, or F when treating all as linear (all p40.07). During the three quarters of the post-EDR change period, the percentage of visits with script use varied by clinic from 15.4% at the lowest, 41.1% at the median, and 65.8% at the highest. Variation in script use over time within each clinic was as small as 2 percentage points in one clinic and as large as 28 percentage points in another clinic. Referral to a quitline across clinics ranged from 2.0% to 18.7%, with a median of 8.9%.
Discussion Follow-up of the randomized trial phase of the CATI tool documented that the dental providers persisted in using scripts and referring patients to a quitline at a relatively high level, until the EDR was modified to allow the past dental history to be loaded into the current visit with a single key stroke and the use of the CATI became optional. This change was associated with a 20–percentage point decline in the use of scripts and a lesser decline in referrals to a quitline. Use of scripts by control group dental practitioners was about 20 percentage points lower than the use of scripts by intervention group dental practitioners both before and after the change in the EDR. The reason for this difference in performance may be due to the fact that the control group training was not as intensive as the training given to the intervention group. As did the intervention group dental practitioners, the control group dental practitioners reduced their activity after the change in the EDR. Although control group dental practitioners referred patients to a quitline less frequently, their referral rate did not change with the change in the EDR. In the last three quarters of observation, there was considerable variation in script use and referral to a quitline both within and between clinics. These observations suggest that making a CATI
tool a standard feature of EDRs would significantly increase the proportion of smokers who are counseled and referred to a quitline. Several meta-analyses of computer-assisted decision support systems have concluded that the systems can improve health professional behavior.20–22 However, the effectiveness of the systems varies. The source of at least some of this variation can be explained by examining the results of the current trial in the context of the behavioral model of B.J. Fogg of Stanford University.23 He posits that the presence of three components is both necessary and sufficient for a behavior to take place: ability, motivation, and a trigger. The natural experiment described in this paper is consistent with his hypothesis. The ability to use scripts and refer to a quitline did not change, so ability cannot possibly explain the change in behavior. It is possible that the dental practitioners coincidentally experienced a marked reduction in their motivation to counsel and refer patients at the time of the EDR change, but there is no evidence to support this hypothesis. The most compelling explanation is that the removal of the trigger—the requirement to ask and advise, and the opportunity to refer—is the explanation for the sudden decline in dental provider activity. The decline in activity after the CATI was no longer activated automatically suggests that restoring this trigger might permanently increase the use of scripts and referral to the quitline. The variation in activity between and within clinics suggests that activity might also be increased to a significant extent if other triggers can be identified and implemented in dental practices. Randomized trials have already demonstrated that e-mails providing educational material or performance feedback can trigger smoking-cessation advice.24,25 Another effective trigger is academic detailing.26 Perhaps the most significant limitation to the data analyzed here is that neither smoking-cessation rates nor quitline contact and engagement rates are available; only dental practitioner activity is. It is not possible to discern from the data set whether a dentist, a dental hygienist, or both used the scripts when a script was used. It is also not possible to precisely define who made a referral when one was made. The data are also limited by the fact that they are derived from a single dental group in the Midwestern U.S. followed over a limited time period. The experience reported here may not apply to all dental practitioners or all EDRs. The difference in dental practitioner behavior between intervention and control groups and the change in dental practitioner behavior associated with the change in the EDR suggests that dental practitioner behavior is very sensitive to both training and EDR design. It is interesting that whereas 37% of intervention clinic smokers who were interviewed during the trial reported www.ajpmonline.org
Kottke et al / Am J Prev Med 2015;48(6):722–728
that they were referred to a quitline, the rate calculated from medical record review was only 19%. There are at least two explanations for this discrepancy. A study in Australia that audio recorded entire medical encounters and compared the recordings to patient reports found that patients over-reported physician smoking-cessation counseling rates.27 The same discrepancy could be present in the HealthPartners Dental Group experience. It is also possible that the dental practitioners referred patients to a quitline even when not prompted by a script, or they failed to record their referral in the dental record. It has been documented that medical records do not accurately reflect practitioner smoking-cessation activity.27 Despite the fact that smoking has a powerful negative impact on oral health, dental practitioners have historically been less active and have felt less qualified than physicians to deal with smoking.28 The data presented here suggest that a CATI tool integrated into the EDR could significantly increase the proportion of patients who are advised to quit smoking, assisted in their efforts to quit smoking, and referred to a quitline. This is particularly true if use of the CATI with patients who smoke were obligatory.
7. 8.
9.
10.
11.
12. 13.
14.
15.
16.
CDC, Office on Smoking and Health, under contract number 200-2009-28537, provided the sole financial support for this analysis. No financial disclosures were reported by the authors of this paper.
17.
18.
References 1. Fiore MC, Jaén CR, Baker TB, et al. Treating Tobacco Use and Dependence: 2008 Update. Clinical Practice Guideline. Rockville, MD: USDHHS. Public Health Service, 2008. 2. Nohlert E, Tegelberg A, Tillgren P, Johansson P, Rosenblad A, Helgason AR. Comparison of a high and a low intensity smoking cessation intervention in a dentistry setting in Sweden—a randomized trial. BMC Public Health. 2009;9:121. http://dx.doi.org/10.1186/ 1471-2458-9-121. 3. Nohlert E, Ohrvik J, Tegelberg A, Tillgren P, Helgason AR. Long-term follow-up of a high- and a low-intensity smoking cessation intervention in a dental setting—a randomized trial. BMC Public Health. 2013;13(1):592. http://dx.doi.org/10.1186/1471-2458-13-592. 4. Gordon JS, Andrews JA, Crews KM, Payne TJ, Severson HH. The 5Aʼs vs 3Aʼs plus proactive quitline referral in private practice dental offices: preliminary results. Tob Control. 2007;16(4):285–288. http://dx.doi.org/ 10.1136/tc.2007.020271. 5. Gordon JS, Andrews JA, Albert DA, Crews KM, Payne TJ, Severson HH. Tobacco cessation via public dental clinics: results of a randomized trial. Am J Public Health. 2010;100(7):1307–1312. http://dx.doi.org/ 10.2105/AJPH.2009.181214. 6. Glasgow RE, Vogt TM, Boles SM. Evaluating the public health impact of health promotion interventions: the RE-AIM framework.
June 2015
19.
20.
21.
22.
23. 24.
727
Am J Public Health. 1999;89(9):1322–1327. http://dx.doi.org/10.2105/ AJPH.89.9.1322. Keller PA, Feltracco A, Bailey LA, et al. Changes in tobacco quitlines in the United States, 2005-2006. Prev Chronic Dis. 2010;7(2):A36. Cummins SE, Bailey L, Campbell S, Koon-Kirby C, Zhu SH. Tobacco cessation quitlines in North America: a descriptive study. Tob Control. 2007;16(suppl 1):i9–i15. http://dx.doi.org/10.1136/tc.2007.020370. Kaufman A, Augustson E, Davis K, Finney Rutten LJ. Awareness and use of tobacco quitlines: evidence from the Health Information National Trends Survey. J Health Commun. 2010;15(suppl 3): 264–278. http://dx.doi.org/10.1080/10810730.2010.526172. Ray MN, Funkhouser E, Williams JH, et al. Smoking-cessation e-referrals: a national dental practice-based research network randomized controlled trial. Am J Prev Med. 2014;46(2):158–165. http://dx.doi.org/10.1016/ j.amepre.2013.10.018. Sadasivam RS, Hogan TP, Volkman JE, et al. Implementing point of care “e-referrals” in 137 clinics to increase access to a quit smoking internet system: the Quit-Primo and National Dental PBRN HI-QUIT Studies. Transl Behav Med. 2013;3(4):370–378. http://dx.doi.org/10.1007/ s13142-013-0230-3. Occupational Outlook Handbook. 2012; www.bls.gov/ooh/healthcare/ dental-hygienists.htm. The Number of Practicing Primary Care Physicians in the United States. www.ahrq.gov/research/findings/factsheets/primary/pcwork1/ index.html. Health, United States, 2012. With Special Feature on Emergency Care. Table 90. Hyattsville, MD: National Center for Health Statistics, 2013. Abbo ED, Zhang Q, Zelder M, Huang ES. The increasing number of clinical items addressed during the time of adult primary care visits. J Gen Intern Med. 2008;23(12):2058–2065. http://dx.doi.org/ 10.1007/s11606-008-0805-8. Wotman S, Demko CA, Victoroff K, Sudano JJ, Lalumandier JA. A multimethod investigation including direct observation of 3751 patient visits to 120 dental offices. Clin Cosmet Investig Dent. 2010;2: 27–39. http://dx.doi.org/10.2147/CCIDEN.S6592. Boyle R, Solberg L, Fiore M. Use of electronic health records to support smoking cessation. Cochrane Database Syst Rev. 2011;12: CD008743. http://dx.doi.org/10.1002/14651858.CD008743.pub2. Heatherton TF, Kozlowski LT, Frecker RC, Rickert W, Robinson J. Measuring the heaviness of smoking: using self-reported time to the first cigarette of the day and number of cigarettes smoked per day. Br J Addict. 1989;84(7):791–799. http://dx.doi.org/10.1111/ j.1360-0443.1989.tb03059.x. Rindal DB, Rush WA, Schleyer TK, et al. Computer-assisted guidance for dental office tobacco-cessation counseling: a randomized controlled trial. Am J Prev Med. 2013;44(3):260–264. http://dx.doi.org/10.1016/ j.amepre.2012.10.023. Hunt DL, Haynes RB, Hanna SE, Smith K. Effects of computer-based clinical decision support systems on physician performance and patient outcomes: a systematic review. JAMA. 1998;280(15): 1339–1346. http://dx.doi.org/10.1001/jama.280.15.1339. Garg AX, Adhikari NK, McDonald H, et al. Effects of computerized clinical decision support systems on practitioner performance and patient outcomes: a systematic review. JAMA. 2005;293(10): 1223–1238. http://dx.doi.org/10.1001/jama.293.10.1223. Boyle JP, Honeycutt AA, Narayan KMV, et al. Projection of diabetes burden through 2050: impact of changing demography and disease prevalence in the U.S. Diabetes Care. 2001;24(11):1936–1940. http://dx.doi.org/ 10.2337/diacare.24.11.1936. Fogg BJ. BJ Foggʼs Behavior Model. Stanford University. www. behaviormodel.org/. Houston TK, Richman JS, Ray MN, et al. Internet delivered support for tobacco control in dental practice: randomized controlled trial. J Med Internet Res. 2008;10(5):e38. http://dx.doi.org/10.2196/ jmir.1095.
728
Kottke et al / Am J Prev Med 2015;48(6):722–728
25. Houston TK, Coley HL, Sadasivam RS, et al. Impact of content-specific email reminders on provider participation in an online intervention: a dental PBRN study. Stud Health Technol Inform. 2010;160(Pt 2):801–805. 26. Sheffer MA, Baker TB, Fraser DL, Adsit RT, McAfee TA, Fiore MC. Fax referrals, academic detailing, and tobacco quitline use: a randomized trial. Am J Prev Med. 2011;42(1):21–28. http://dx.doi.org/ 10.1016/j.amepre.2011.08.028.
27. Ward J, Sanson-Fisher R. Accuracy of patient recall of opportunistic smoking cessation advice in general practice. Tob Control. 1996;5(2): 110–113. http://dx.doi.org/10.1136/tc.5.2.110. 28. Secker-Walker RH, Solomon LJ, Flynn BS, Dana GS. Comparisons of the smoking cessation counseling activities of six types of health professionals. Prev Med. 1994;23(6):800–808. http://dx.doi.org/ 10.1006/pmed.1994.1137.
www.ajpmonline.org
