Physician smoking status may influence cessation counseling practices.
Meshefedjian, Garbis A. ; Gervais, Andre ; Tremblay, Michele 等
Primary care physicians have a unique opportunity to systematically
deliver effective smoking cessation treatment to their patients who
smoke, (1,2) and smoking cessation counseling by physicians is now
considered to be an evidence-based practice. In the 2005 Canadian
Tobacco Use Monitoring Survey (CTUMS), 73% of current smokers had
visited a physician in the past year, but only 51% said they had
received advice to quit or reduce smoking; and 57% were given
information on smoking cessation aids. (3) Several physician-related
characteristics are positively associated with favourable smoking
cessation interventions, including: older age; (4) female gender; (4-6)
working in urban areas4 and in private settings; (7) positive beliefs
and attitudes about the effectiveness of counseling; (8-12) favourable
perceptions about patient responsiveness to advice; (13,14) perceived
self-efficacy; (7) and having received training in smoking cessation.
(4,5) A number of reports (4,5,15-24) suggest that physician smoking
status relates to the quantity and quality of cessation counseling. In
this analysis, we hypothesized that physicians who smoke have less
favourable beliefs and attitudes and more perceived barriers to
counseling than non-smoking physicians and would therefore be less
likely to intervene with smokers.
METHODS
Data were collected in two cross-sectional surveys of general
practitioners (GPs) in Montreal. GPs were eligible to participate if: 1)
their name was registered in the Quebec College of Physicians database,
2) they had an active license, and 3) they had provided patient care in
Montreal in the year preceding the survey. English or French
questionnaires were mailed to randomly selected GPs in April 2000 and
May 2004. If questionnaires were not returned, reminder postcards were
mailed at three and five weeks after the initial mail-out.
Non-respondents were then telephoned by one of the study investigators
to encourage participation. Of 454 eligible participants in 2000, 316
returned a questionnaire; 302 of 463 eligible participants returned a
questionnaire in 2004, for a total of 618 participants across years.
Detailed information on the survey methods is available. (8)
Study variables
Data were collected on the socio-demographic characteristics of
GPs, their practice setting, smoking status, and psychosocial
characteristics related to cessation counseling, including: knowledge,
beliefs, and attitudes about counseling; self-efficacy to provide
effective counseling; perceived barriers to counseling related to both
the physician and patient; awareness of the "stages of change"
model; interest in learning more about cessation methods and/or in
updating skills; and past training in smoking cessation.
Cessation counseling practices in the past three months were
measured through four indicators: 1) ascertainment of patient smoking
status (8 items; Cronbach's [alpha]=0.82); 2) provision of advice
on how to quit (6 items; Cronbach's [alpha]=0.89); 3) provision of
adjunct support (4 items; Cronbach's [alpha]=0.74); and 4)
provision of complete cessation counseling coverage (3 items;
Cronbach's a=0.70). Scores for the first three indicators ranged
between 1 and 6 (corresponding to whether or not the GPs provided
counseling to all, almost all, more than half, less than half, a few, or
none of their patients who smoke); we designated a score [less than or
equal tol]3 (i.e., the GP provided counseling to all, almost all, or
more than half of patients who smoke) as "favourable". Scores
for the fourth indicator ranged between 1 and 5; again we designated a
score [less than or equal to]3 as "favourable" (i.e., the GP
provided counseling for >2 minutes on each occasion, in at least 1 of
3 visits to more than half of smokers). Appendix 1 describes the items
comprising each counseling practice.
<01_TB010>
Data analysis
The association between GP smoking status and psychosocial
characteristics was tested univariately. GP smoking status (never,
former, current) was tested as an independent correlate of counseling
practices in multivariable logistic regression analyses adjusting for
study year, age, sex, language and clinical setting. Data from the two
years (2000, 2004) were pooled together because the interaction between
"year of study" and "smoking status" variables was
not significant.
RESULTS
A total of 618 (67% of 917 eligible) GPs returned a questionnaire.
Data on smoking status were missing for 8 GPs, therefore the analytic
sample included 610 GPs. Among these, 7% were current smokers (2% daily
and 5% occasional), 32% were former smokers, and 61% were never-smokers.
The proportion of current smokers declined from 10% in 2000 to 5% in
2004, while the proportion of former smokers was 31% in 2000 and 34% in
2004. Sex, age, language and clinical setting were significantly
associated with GP smoking status (Table 1).
Overall, 91% of GPs ascertained the smoking status of their
patients; 76% provided advice on how to quit, but only 26% provided
adjunct support (i.e., offered written educational materials, follow-up
visits, and referred to community resources). Eighty-two percent
provided complete cessation counseling coverage within their practice.
<01_TB011>
Two thirds (69%) of current smokers had favourable beliefs and
attitudes about cessation counseling compared to approximately 83% of
former and never-smokers (p=0.08). Smoking status was not associated
with self-efficacy, physician or patient barriers, awareness of the
"stages of change" model, or training in cessation counseling
(data not shown). Current smokers were however more interested in
learning about methods to support cessation attempts than GPs who had
never smoked or who had quit (64%, 56%, and 45% respectively; p=0.018).
GPs who smoked were markedly less likely than never-smokers to
ascertain the smoking status of their patients (OR 0.6, 95% CI 0.21.6),
to provide advice on how to quit (OR 0.6, 0.3-1.3), and to provide
complete cessation counseling coverage (OR 0.5, 0.2-1.1) within their
practices (Table 2). Former smokers were more likely than never-smokers
to provide adjunct support (OR 1.5, 1.0-2.4).
DISCUSSION
The proportion of current smokers among GPs in Montreal declined
from 10% in 2000 to 5% in 2004. Steady declines in smoking among
physicians have been reported in most developed countries, with the
prevalence as low as 2% in the USA in 2000. (25) The prevalence reported
herein is similar to the 3% reported in the 2008 Canadian Physician
Health Survey, (26) but lower than the 22% reported for Canadian
physicians in the international "Smoking: The Opinions of
Physicians" (STOP) survey in 2006. (24)
In our analysis, smoking status was associated with several
cessation counseling practices. Compared to non-smokers, GPs who smoke
were less likely to ascertain the smoking status of their patients, to
provide advice on how to quit, and to provide complete counseling
coverage. While the confidence intervals on the estimates include unity
(likely related to the small number of smokers), the ORs indicate very
strong negative associations between smoking status and these three
components of the counseling intervention.
While fewer current than never-smokers in this sample had
favourable beliefs and attitudes about counseling, there was little
difference between groups in the other psychosocial characteristics
investigated. The lower level of intervention among GPs who smoke may
reflect that GPs who smoke are reluctant to advise patients on how to
quit when they themselves smoke. If GPs are unable to quit themselves,
they may feel that they are ill-equipped to help others quit.
Interestingly, GPs who had quit smoking were significantly more
likely than current or never-smokers to provide adjunct support for
cessation, possibly because they themselves had experienced difficulty
quitting and were more aware of the need for concrete support to help
smokers to quit.
Our findings are consistent with results from the recent STOP
survey, wherein 80% of physicians who smoke compared to 85% of
non-smoking physicians asked their patients how much they smoke; 85%
compared to 90% advised patients to stop smoking; and 40% compared to
48% assisted patients in developing a plan to quit. (24) Similarly, both
Ohida et al. (2001) and Underner et al. (2006) reported that non-smoking
physicians were more active in their smoking cessation practices than
physicians who smoke. (18,19) Squier et al. (2006) reported that
non-smoking physicians were more likely than physicians who smoked to
record patients' tobacco use, but failed to show statistically
significant differences in the provision of advice to quit. (4) A study
among GPs in Finland also concluded that there was no difference in
anti-smoking advice given to patients between GPs who smoke and those
who do not, with the exception that, compared to their counterparts who
smoke, non-smoking male GPs gave more smoking cessation advice to
patients with tobacco-related diseases. (5)
Differences between reports may relate to the measure of smoking
cessation counseling practices. While most studies (15-17,19,24) use
single-item indicators, we used a composite indicator which incorporated
several items to measure each component of cessation counseling.
Single-item measures may not have the same threshold as composite
measures in terms of capturing the underlying concept, which may result
in discordant findings. (27,28) In addition to differences in the
measurement of counseling practices, our study included former smokers
as a separate category of exposure.
Limitations
Study limitations include that the cross-sectional design does not
permit causal inference. Self-reports of both smoking status and
counseling practices may result in misclassification bias which could
have attenuated the findings towards the null. Selection bias related to
non-response may have limited external generalizability of the results.
CONCLUSION
While very few physicians continue to smoke, our findings suggest
that their smoking status is associated with the content of the
counseling they provide for their patients who smoke. Taking physician
smoking status into consideration in the design of cessation training
programs may improve cessation counseling interventions.
Appendix 1.
Ascertaining smoking status was measured by eight items (Cronbach
[alpha]=0.82): (i) Do you use a system (such as a medical problem list,
stamp or label,...) to identify patients who smoke? Responses were:
always, usually, sometimes, rarely, never. In the past 3 months, for how
many patients in each of the following patient groups did you ascertain
the smoking status? (ii) new patients on the first visit, (iii) patients
who were smokers at their last visit, (iv) recent ex-smokers, (v)
adolescents (age 13-19), (vi) patients presenting with smoking-related
symptoms or diseases, (vii) patients not presenting with smoking-related
symptoms or diseases, and (viii) in the past 3 months, for how many of
your new patients who smoke did you ascertain number of cigarettes
smoked per day? Responses were: all, almost all, more than half, about
half, less than half, and few/none. Scores 1 to 6 were assigned to each
response choice. Their prorated sum was dichotomized into favourable
outcome (i.e., score of 3 or less, which generally corresponded to
responses "all; almost all; and, more than half") and
unfavourable outcome (i.e., score of 4 or more, which generally
corresponded to "about half; less than half; and, few/none").
Provision of advice on how to quit was measured by six items
(Cronbach a=0.89): During the past 3 months, for how many of your
patients who did not want to quit did you: (i) express concern about
their continued smoking, (ii) recommend that they think about quitting;
and during the past 3 months, for which patients who were preparing to
quit smoking did you: (iii) discuss withdrawal symptoms, weight gain or
other concerns, (iv) discuss strategies to quit smoking, (v) discuss
setting a quit date, and (vi) recommend nicotine replacement therapy.
Responses were: all, almost all, more than half, about half, less than
half, and few/none. Scores 1 to 6 were assigned to each response choice.
Their prorated sum was dichotomized into favourable outcome (i.e., score
of 3 or less, which generally corresponded to responses "all;
almost all; and, more than half") and unfavourable outcome (i.e.,
score of 4 or more, which generally corresponded to "about half;
less than half; and, few/none").
Provision of adjunct support was measured by 4 items (Cronbach
=0.74): During the past 3 months, for how many of your patients who did
not want to quit, did you: (i) offer written educational material on
smoking or smoking cessation, and during the past 3 months, for which
patients who were preparing to quit smoking did you: (ii) offer written
educational material on smoking or smoking cessation, (iii) refer them
to community resources, and (iv) offer a follow-up visit 1 to 2 weeks
after the expected date of cessation. Responses were: all, almost all,
more than half, about half, less than half, and few/none. Scores 1 to 6
were assigned to each response choice. Their prorated sum was
dichotomized into favourable outcome (i.e., score of 3 or less, which
generally corresponded to responses "all; almost all; and, more
than half") and unfavourable outcome (i.e., score of 4 or more,
which generally corresponded to "about half; less than half; and,
few/none").
Counseling completeness was measured in three items (Cronbach
=0.70): (i) In the past 3 months, to how many of your patients who smoke
did you offer advice or assistance for smoking cessation? Responses
were: all, almost all, more than half, about half, less than half; (ii)
How often do you offer smoking cessation counseling to patients who
smoke? Responses were: every visit, 1 in 2 visits, 1 in 3 visits, 1 in 4
visits, and less than 1 in 4 visits; (iii) When you offer smoking
cessation counseling during a patient visit, how many minutes on average
do you devote to the counseling? Responses were: more than 10 minutes, 6
to 10 minutes, 2 to 5 minutes, less than 2 minutes, never offer smoking
cessation counseling. Scores 1 to 5 were respectively assigned to each
response choice. Their prorated sum was then dichotomized into
favourable outcome (i.e., score of 3 or less, generally corresponding to
responses "all, almost all, or more than half" in item (i);
"1 in 3 or more visits" in item (ii), and, "2 minutes or
more" in item (iii)) and unfavourable outcome (i.e., score of 4 or
more).
Received: July 3, 2009
Accepted: April 16, 2010
REFERENCES
(1.) Fiore MC, Jaen CR, Baker TB, Bailey WC, Benowitz NL, Curry SJ,
et al. Treating Tobacco Use and Dependence: 2008 Update, Clinical
Practice Guideline. Rockville, MD: U.S. Department of Health and Human
Services. Public Health Service. May 2008. Available at:
www.ahrq.gov/path/tobacco.htm#Clinic (Accessed January 6, 2010).
(2.) World Health Organization: WHO Tobacco Free Initiative: The
Role of Health Professionals in Tobacco Control. Geneva, Switzerland:
WHO, 2005.
(3.) Smoking-Cessation Advice from Health-Care Providers--Canada,
2005. MMWR 2007;56(28):708-12. Available at:
http://www.cdc.gov/mmwr/preview/ mmwrhtml/mm5628a3.htm (Accessed January
6, 2010).
(4.) Squier C, Hesli V, Lowe J, Ponamorenko V, Medvedovskaya N.
Tobacco use, cessation advice to patients and attitudes to tobacco
control among physicians in Ukraine. Eur J Cancer Prev
2006;15(5):458-63.
(5.) Barengo NC, Sandstrom HP, Jormanainen VJ, Myllykangas MT.
Attitudes and behaviors in smoking cessation among general practitioners
in Finland 2001. Soz Praventivmedn 2005;50(6):355-60.
(6.) Tremblay M, Gervais A, Lacroix C, O'Loughlin J, Makni H,
Paradis G. Physicians taking action against smoking: An intervention
program to optimize smoking cessation counseling by Montreal general
practitioners. CMAJ 2001;165(5):601-7.
(7.) Goldstein MG, DePue JD, Monroe AD, Lessne CW, Rakowski W,
Prokhorov A, et al. A population-based survey of physician smoking
cessation counseling practices. Prev Med 1998;27(5 Pt 1):720-29.
(8.) O'Loughlin J, Makni H, Tremblay M, Lacroix C, Gervais A,
Dery V, et al. Smoking cessation counseling practices of general
practitioners in Montreal. Prev Med2001;33(6):627-38.
(9.) Vogt F, Hall S, Marteau TM. General practitioners' and
family physicians' negative beliefs and attitudes towards
discussing smoking cessation with patients: A systematic review.
Addiction 2005;100(10):1423-31.
(10.) Vogt F, Hall S, Marteau TM. General practitioners'
beliefs about effectiveness and intentions to recommend smoking
cessation services: Qualitative and quantitative studies. BMC Family
Practice 2007;8:39-50.
(11.) Meredith LS, Yano EM, Hickey SC, Sherman SE. Primary care
provider attitudes are associated with smoking cessation counseling and
referral. Medical Care 2005;43:929-34.
(12.) Josseran L, King G, Guilbert P, Davis J, Brucker G. Smoking
by French general practitioners: Behaviour, attitudes and practice. Eur
J Public Health 2005;15(1):33-38.
(13.) Coleman T, Murphy E, Cheater F. Factors influencing
discussion of smoking between general practitioners and patients who
smoke: A qualitative study. Br J Gen Pract 2000;50(452):207-10.
(14.) Pollak KI, Arredondo EM, Yarnall KS, Lipkus I, Myers E,
McNeilly M, Costanzo P. How do residents prioritize smoking cessation
for young "high-risk" women? Factors associated with
addressing smoking cessation. Prev Med 2001;33(4):292-99.
(15.) Cornuz J, Humair JP, Seematter L, Stoianov R, van Melle G,
Stalder H, Pecoud A. Efficacy of resident training in smoking cessation:
A randomized, controlled trial of a program based on application of
behavioral theory and practice with standardized patients. Ann Intern
Med 2002;136(6):429-37.
(16.) Malarcher A, Easton A, Husten C, Frank E. Smoking cessation
counselling: Training and practice among women pediatricians. Clinical
Pediatrics 2002;41(5):341-49.
(17.) Kawakami M, Nakamura S, Fumimoto H, Takizawa J, Baba M.
Relationship between smoking status of physicians and their enthusiasm
to offer smoking cessation advice. Intern Med 1997;36(3):162-65.
(18.) Ohida T, Sakurai H, Mochizuki Y, Kamal AM, Takemura S, Minowa
M, Kawahara K. Smoking prevalence and attitudes towards smoking among
Japanese physicians. JAMA 2001;285(20):2643-48.
(19.) Underner M, Ingrand R, Allouch A, Laforgue AV, Migeot V,
Defossez G, Meurice JCl. Influence of smoking among family physicians on
their practice of giving minimal smoking cessation advice. Rev Mal
Respir 2006;23(5 Pt 1):426-29.
(20.) Pizzo AM, Chellini E, Grazzini G, Cardone A, Badellino F.
Italian general practitioners and smoking cessation strategies. Tumori
2003;89(3):250-54.
(21.) Nardini S, Bertoletti R, Rastelli V, Donner CF. The influence
of personal tobacco smoking on the clinical practice of Italian chest
physicians. Eur Respir J 1998;12(6):1450-53.
(22.) Kawahara K, Ohida T, Osaki Y, Mochizuki Y, Minowa M,
Yamaguchi N, Kusaka Y. Study of the smoking behavior of medical doctors
in Fukui, Japan and their antismoking measures. J Epidemiol
2000;10(3):157-62.
(23.) Parna K, Rahu K, Barengo NC, Rahu M, Sandstrom PH,
Jormanainen VJ, Myllykangas MT. Comparison of knowledge, attitudes and
behaviour regarding smoking among Estonian and Finnish physicians. Soz
Praventivmed 2005;50(6):378-88.
(24.) Pipe A, Sorensen A, Reid R. Physician smoking status,
attitudes towards smoking, and cessation advice to patients: An
international survey. Patient Educ Couns 2009;74(1):118-23.
(25.) Smith DR, Leggat PA. An international review of tobacco
smoking in the medical profession: 1974-2004. BMC Public Health
2007;7:115.
(26.) Kapur A. Smoking among Physicians in Canada. Presentation at
the 6th National Conference on Tobacco or Health--Montreal, Quebec,
November 3, 2009.
(27.) Sloan JF, Aaronson N, Cappelleri JC, Fairclough DL,
Varricchio C, and the Clinical Significance Consensus Meeting Group.
Assessing the clinical significance of single items relative to summated
scores. Mayo Clinic Proc 2002;77(5):479-87.
(28.) Gardner DG, Cummings LL. Single-item versus multiple-item
measurement scales: An empirical comparison. Educ Psycholog Measurement
1998;58(6):898 915.
Garbis A. Meshefedjian, mph, PhD, [1] Andre Gervais, MD, [1,2]
Michele Tremblay, MD, [3] Diane Villeneuve, MD, [1] Jennifer
O'Loughlin, PhD [4,5]
Author Affiliations
[1.] Agence de la sante et des services sociaux de Montreal,
Direction de sante publique, Montreal, QC
[2.] Department of Medicine, Universite de Montreal, Montreal, QC
[3.] Institut national de sante publique du Quebec, Montreal, QC
[4.] Centre de recherche CHUM, Montreal, QC
[5.] Department of Social and Preventive Medicine, Universite de
Montreal, Montreal, QC
Correspondence and reprint requests: G.A. Meshefedjian, Agence de
la sante et des services sociaux de Montreal, Direction de sante
publique, 1301, rue Sherbrooke Est, Montreal (Quebec) H2L 1M3, E-mail :
gmeshefe@santepub-mtl.qc.ca
Conflict of Interest: None to declare.
Table 1. Socio-demographic Characteristics and Practice
Profile of General Practitioners According to
Smoking Status, Montreal, 2000/4
Total Never- Former Current
smokers Smokers Smokers
(n=610) (n=370) (n=195) (n=45)
% % % p-value
Male 55 48 71 51 <0.001
Age, years
<40 25 31 14 27 <0.001
40-54 52 50 54 60
[greater than or
equal to] 55 23 19 32 13
Language
French 76 73 80 89 0.021
English 24 27 20 11
Year
2000 51 51 48 67 0.072
2004 49 49 52 33
Clinical setting *
Solo 26 23 32 24 0.067
Group 48 48 46 56 0.500
Hospital 33 31 36 31 0.400
CLSC ([dagger]) 20 25 11 18 <0.001
Other 29 31 24 29 0.201
* Categories are not mutually exclusive.
([dagger]) Centre Local de Services Communautaires (community clinic).
Table 2. Odds Ratio (95% Confidence Interval) for GP
Smoking Status for Selected Indicators of Smoking
Cessation Counseling (n=610)
GP Smoking Status [OR.sub.crude] [OR.sub.adj] *
(95% CI) (95% CI) *
Ascertains smoking status of
patients
Never 1.0 1.0
Former 0.8 (0.5-1.7) 1.0 (0.5-2.0)
Current 0.6 (0.2-1.4) 0.6 (0.2-1.6)
Provides advice on how to quit
Never 1.0 1.0
Former 0.9 (0.6-1.4) 1.1 (0.7-1.7)
Current 0.6 (0.3-1.3) 0.6 (0.3-1.3)
Provides adjunct support
Never 1.0 1.0
Former 1.1 (0.7-1.6) 1.5 (1.0-2.4)
Current 1.1 (0.6-2.3) 1.2 (0.6-2.5)
Provides complete cessation
counseling coverage
Never 1.0 1.0
Former 1.1 (0.7-1.7) 1.0 (0.6-1.7)
Current 0.6 (0.3-1.2) 0.5 (0.2-1.1)
* Adjusted for year (2000, 2004), sex, age, language and type of
clinical setting.
