Abstract
Clinical Practice Guidelines for Treating Tobacco Use and Dependence advocate for using counseling targeted at tobacco users’ motivation to quit during each office visit. We evaluate tobacco use screening and counseling interventions delivered during routine periodic health examinations by 44 adult primary care physicians practicing in 22 clinics of a large health system in southeast Michigan. 484 office visits were audio-recorded and transcribed. For this study, current tobacco users (N = 91) were identified using pre-visit surveys and audio-recordings. Transcripts were coded for the delivery of tobacco-related counseling interventions. The extent to which counseling interventions were used and/or targeted to the patients’ readiness to quit was the main outcome measure. The majority of tobacco users (n = 77) had their tobacco use status assessed, and most received some sort of tobacco-related counseling (n = 74). However, only 15% received the recommended counseling targeted to their readiness to quit. On the other hand, 19% received less counseling than recommended given their readiness to quit, 7% received only nonindicated counseling, and 59% received nonindicated counseling in addition to indicated counseling. Results illustrate physicians’ commitment to cessation counseling and also identify potential opportunities to improve the efficiency of tobacco-related counseling in primary care.
Introduction
Counseling recommendations in the Guidelines for Treating Tobacco Use and Dependence have remained largely unchanged since 1996 (Fiore, 2008; Fiore et al., 1996). According to these guidelines, physicians should counsel patients based on their readiness to quit (Fiore, 2008), as readiness to perform healthy behaviors, including smoking cessation, is strongly associated with success rates (Prochaska et al., 2005). According to the guidelines, current tobacco users should be counseled using an ask, advise, assess, assist, and arrange (5As) approach (Fiore, 2008; Lawson, Flocke, & Casucci, 2009). During the “Assess” step, if a patient does not express a readiness to quit within 30 days, the clinician should divert from the 5As and use the relevance, risks, rewards, and repetition (5Rs) approach which uses personalized motivational counseling to help sway ambivalent patients (Fiore, 2008).
Targeting tobacco counseling based on a patient’s readiness to quit has possible implications for counseling efficiency. Existing evidence indicates that motivational counseling (as done with the 5Rs) is effective in helping tobacco users to make a quit attempt (Steinberg, Ziedonis, Krejci, & Brandon, 2004). On the other hand, the use of motivational counseling with tobacco users willing to quit, though unlikely to have negative implications, is less likely to provide added value (Hettema & Hendricks, 2010). Given known time constraints on primary care physicians (PCPs) (Yarnall, Pollak, Østbye, Krause, & Michener, 2003), and repeated physician reports of lack of time as the main barrier to cessation counseling (Park et al., 2003; Vogt, Hall, & Marteau, 2005), it is imperative to consider whether there might be opportunities to address the efficiency of tobacco cessation counseling delivery in primary care.
Despite an extensive body of literature assessing the delivery of tobacco use counseling, including delivery of the 5As (Tessaro et al., 1997; Yarnall et al., 2003), few studies have examined how PCPs target tobacco use cessation counseling content in practice (Hollis et al., 2000; Lawson et al., 2009; Quinn et al., 2009). To our knowledge, only one previous study used office visit audio-recordings (Lawson et al., 2009), while the others relied on either patient questionnaires or medical records (Williams et al., 2014). While medical record abstraction and patient self-reports can capture whether or not office-based counseling took place, they are less ideal for capturing nuances regarding counseling content (Ha & Longnecker, 2010). Furthermore, no prior study has evaluated the potential efficiency of counseling in the practice setting. Using audio-recordings from primary care office visits, we describe the extent to which tobacco use cessation counseling is targeted based on patients’ readiness to quit and the resulting implications for the efficiency of office-based tobacco cessation counseling practices.
Method
Participant Eligibility
Eligible clinician and patient participants were those enrolled in an observational study of colorectal cancer screening discussions in 26 primary care clinics owned by a large health system in southeast Michigan (Shires et al., 2012). All family and general internal medicine physicians practicing with the health system’s affiliated medical group were invited to participate. Study-eligible patients were recruited sequentially from among those scheduling periodic health examinations between February 2007 and June 2009 with a participating physician (N = 64). Study patients (N = 484) were insured, 50 to 80 years old, and due for colorectal cancer screening. Physicians and patients were informed that the purpose of the study was to understand how physicians and patients discuss preventive health services. The study was approved by the institutional review boards at participating institutions.
Data Sources and Qualitative Analyses
Three data sources were used for the current analyses: a pre-visit patient telephone survey, health system records, and transcripts of office visit audio-recordings. To identify current tobacco users, we used office visit transcripts to identify any tobacco use discussion. This was done by the first author or a research assistant using Microsoft Word 2010 version to query for the following keys: “smok,” “cig,” “tobac,” or “nicot.” This resulted in the identification of n = 84 current smokers. In addition to identifying smokers via office visit discussions, seven tobacco users were identified via their pre-visit survey responses.
Once a tobacco-related discussion was identified, its content was coded by one of the two coders using a structured coding work sheet. Coded items focused on differentiating the presence or absence of a counseling technique that was either one of the 5As or any one of the first four components of the 5Rs (relevance, risks, rewards, and roadblocks). A similar approach has been used previously (Lawson et al., 2009).
The “Ask” step was assessed by determining whether there was any discussion of the patient’s smoking status, regardless of who initiated the discussion. The “Assess” step could be completed in one of the three ways: (1) the PCP assessed willingness to quit, (2) the patient volunteered an interest in quitting, or (3) the patient responded to a PCP inquiry with a declaration of interest (or lack of interest) in quitting. A current tobacco user was classified as having received cessation assistance if any cessation aid was recommended or classified as having received motivational counseling if any of the 5R techniques was delivered.
We labeled the counseling delivered as indicated if the patient reported willingness to quit and received cessation assistance, or if the patient was unwilling to quit and received motivational counseling. Counseling was nonindicated if the patient reported willingness to quit and received motivational counseling, or if the patient was unwilling to quit and received cessation assistance. Patients were classified as willing to quit if they (1) asked for help to quit, (2) responded affirmatively to a PCP inquiry regarding willingness to quit, or (3) accepted cessation assistance when offered.
Delivered counseling for each patient was classified into one of the four mutually exclusive categories: indicated only (only guideline-indicated interventions delivered), indicated plus (at least one nonindicated intervention in addition to indicated ones delivered), nonindicated only (only nonindicated interventions delivered), and none (no intervention delivered). The interrater agreement of the coded variables was assessed via a random sample of n = 49 transcripts coded independently by the two coders. Mean Cohen’s κ was .94 (range 0.68–1.00).
Statistical Analysis
We report the proportion of visits among current tobacco users for which tobacco use status was discussed (Ask) and among those, the proportion with their willingness to quit assessed (Assess). Tobacco use cessation counseling (Advise) is reported by the four mutually exclusive categories described earlier (indicated only, nonindicated only, indicated plus, and none). SAS 9.0 (SAS Institute Inc., Cary, NC) was used for data analysis.
Results
The tobacco users (N = 91, 18% of the overall sample) were seen by a total of 44 PCPs (range of recorded visits per PCP was 1–3, with a mean of two visits per PCP). The mean patient age was 59 years (±8.1), and the sample was 34% male and mostly White (66%) or African American (28%). Among current tobacco users, 92% (n = 84) were asked by their PCP about their tobacco use (Ask). Among those asked, 92% (n = 77) had their willingness to quit ascertained (Assess). Some sort of tobacco-related counseling was received by 81% (n = 74) of these patients. Among those with their willingness to quit assessed (n = 77), 66% (n = 51) expressed willingness and 34% (n = 26) reported being unwilling to quit (Figure 1).
Delivery of indicated and nonindicated counseling interventions among current tobacco users (N = 91).
Among those willing to quit, only 4% received no tobacco cessation counseling (none). That is, 96% (n = 49) received some type of cessation-related counseling. Upon further examination, 8 were offered only indicated counseling (indicated), 39 received both indicated and nonindicated counseling (indicated plus), and 2 received only nonindicated counseling (nonindicated). Among those unwilling to quit (n = 26), 88% (n = 23) received some type of cessation-related counseling. Upon further examination, 6 were offered only indicated counseling interventions (i.e., indicated motivational counseling), 15 received both indicated and nonindicated counseling (indicated plus), and 2 received only nonindicated counseling (i.e., cessation assistance). Willingness to quit was unknown when tobacco use was not discussed (n = 7) or willingness to quit was not assessed (n = 7). Although two of these tobacco users were offered motivational counseling, it was labeled as nonindicated because we were unable to determine the patient’s willingness to quit. The algorithm in Figure 1 shows that 15% of tobacco users received counseling specifically targeted for their expressed willingness to quit (indicated only), 59% received both indicated and nonindicated counseling (indicated plus), 7% received only nonindicated counseling (nonindicated only), and 19% received no counseling (none).
Discussion
Our findings demonstrate the clear commitment of PCPs in our sample to address tobacco use. Despite competing demands for clinician time (Yarnall et al., 2003), the overwhelming majority of tobacco users in our sample had their smoking status assessed and received some sort of tobacco-related counseling. Such findings are consistent with PCPs’ self-reported counseling behavior (Tong, Strouse, Hall, Kovac, & Schroeder, 2010) and previously documented use of cessation counseling in primary care (Williams et al., 2014). However, counseling rates reported here are substantially higher than those reported by Jamal, Dube, Malarcher, Shaw, and Engstrom (2012) among a national sample of physicians practicing in diverse primary care practice settings. Despite this clear commitment, over a quarter of tobacco users in our sample received either no tobacco-related counseling or received no counseling appropriate to their willingness to quit, representing a missed opportunity to deliver known effective counseling. As such, our findings illustrate a potential opportunity to improve the efficiency of counseling delivery.
Targeted counseling is critical to tobacco cessation efforts and is deemed the standard of care for treatment of tobacco dependence (Fiore, 2008). Physicians’ discussions with current tobacco users are the most significant prompts for assisted quit attempts (Ussher, Brown, Rajamanoharan, & West, 2013). Thus, capitalizing on opportunities to deliver known effective cessation assistance to a patient who is willing to quit is central to the ability of office-based counseling to impact tobacco use. On the other hand, nonindicated counseling may be ignored by patients (Fiore, 2008) while competing for scarce visit time. A recent review found that the main negative belief regarding tobacco cessation counseling among PCPs was that it was too time consuming (Vogt et al., 2005). Thus, helping PCPs reduce the time they allocate to the delivery of nonindicated tobacco use counseling, while ensuring the delivery of appropriately targeted counseling, could not only impact the efficiency of tobacco-related counseling delivery but also allow precious office visit time to be allocated to other things.
Limitations
Our study had limitations. The sample of physicians and adult patients was limited to those practicing with and receiving care from one integrated health system, where physicians are salaried and the electronic health record includes smoking status within vital signs. Furthermore, the system offers smokers an in-house smoking cessation program. As such, care should be taken when generalizing findings to other settings and populations, as tobacco use screening and counseling practices are known to vary by settings (Jamal et al., 2012; Tong et al., 2010). Such variation may reflect differences in clinician knowledge of tobacco counseling guidelines but may also reflect differing logistical barriers and facilitators. It is also possible that the presence of the observer and the audio-recorder may have led to different office visit interactions. However, in similar studies, the Hawthorne effect has been negligible (Pringle & Stewart-Evans, 1990; Stange et al., 1998). In addition, we observed only one visit per patient and therefore do not have information on previous conversations between the patient and their PCP. Furthermore, the pre-visit survey assessed cigarette smoking and not tobacco use in general, so there is a possibility that users of other forms of tobacco who did not get their tobacco use status discussed were inappropriately excluded from our sample. Additionally, we were unable to test the impact of counseling, regardless of whether with or without pharmacotherapy, on patients’ subsequent tobacco use and cessation attempts. Further studies should utilize a longitudinal design to enable testing the differential impact of indicated and nonindicated counseling on quit rates to inform the appropriateness of applying the guideline counseling recommendations in their current form in primary care practice.
Conclusion
Among a sample of PCPs who routinely address tobacco use during periodic health examinations, we found opportunities to improve the targeting of their cessation counseling to a patient’s motivation to quit. Improving such targeting may have implications for the efficiency with which tobacco cessation counseling is delivered and thus save precious office visit time.
Footnotes
Acknowledgments
We are grateful to all the PCPs, study participants, and research assistants who took part in this study. We also want to especially acknowledge Ms. Amber Cox who assisted with coding and Mr. Scott Ratliff who provided advice on the coding and statistical analysis.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by NIH, National Cancer Institute Grant R01 CA112379 (principal investigator [PI]: Elston Lafata).