Tobacco Use Screening and Treatment by Outpatient Psychiatrists Before and After Release of the American Psychiatric Association Treatment Guidelines for Nicotine Dependence

Abstract

Objectives. We examined tobacco use screening and treatment by US psychiatrists before and after release of the 1996 American Psychiatric Association (APA) nicotine dependence treatment guidelines.

Methods. We used data from the National Ambulatory Medical Care Survey to identify rates of tobacco screening and treatment by psychiatrists before the release of the guidelines (1993–1996) and during 2 postguidelines periods: 2001–2005 and 2006–2010. Multiple logistic regression was used to compare preguidelines and postguidelines rates.

Results. Psychiatrists screened for tobacco use during 77% of visits from 1993 to 1996, 69% of visits from 2001 to 2005 (odds ratio [OR] = 0.69; 95% confidence interval [CI] = 0.64, 0.75), and 60% of visits from 2006 to 2010 (OR = 0.46; 95% CI = 0.43, 0.50). Psychiatrists provided cessation counseling to 12% of smokers from 1993 to 1996, 11% from 2001 to 2005 (OR = 0.97; 95% CI = 0.74, 1.26), and 23% from 2006 to 2010 (OR = 2.23; 95% CI = 1.74, 2.86). Psychiatrists prescribed nicotine replacement therapy to fewer than 1% of smokers during all 3 time periods.

Conclusions. Psychiatrists are screening for tobacco use at declining rates, and the proportion of smokers provided with treatment remains low.

Smoking is responsible for 435 000 deaths per year, representing the largest single cause of preventable mortality in the United States.¹ Individuals diagnosed with a mental health disorder have particularly high rates of tobacco use and consume more than 40% of cigarettes sold in the United States each year.² There are several smoking cessation treatments available that are effective among smokers with a mental illness, including behavioral counseling, nicotine replacement therapy (NRT), and medications such as bupropion and varenicline.³

Despite the availability of effective treatment, smokers diagnosed with a mental health disorder face many barriers to quitting, including high levels of nicotine dependence and stronger withdrawal symptoms than the general population,⁴ comorbid substance abuse or dependence, difficulties coping with mental health symptoms during a quit attempt, lack of support for quitting, and lack of regular access to preventive medical services.⁵ Furthermore, mental health providers often view tobacco cessation as a low priority for their patients, and historically psychiatric treatment settings have promoted the use of tobacco as a form of self-medication or behavioral control for patients with mental health disorders.⁶

In 1996, the American Psychiatric Association (APA) published clinical practice guidelines for the treatment of nicotine dependence.⁷ The guidelines recommended that all patients with a psychiatric diagnosis be screened for tobacco use and that those interested in quitting be provided with treatment. Prior to the release of the guidelines, the rate at which patients with a psychiatric diagnosis were screened for tobacco use in US outpatient settings was relatively high (76%), but rates of providing counseling and medications to smokers with a psychiatric diagnosis were low (23% and 2%, respectively), and psychiatrists screened and treated their patients for tobacco use at lower rates than other types of physicians.^8,9

It is currently unknown whether there have been any improvements since the release of the APA guidelines in the rates at which psychiatrists screen their patients for tobacco use and provide smokers with evidence-based cessation therapies. We used data from the National Ambulatory Medical Care Survey (NAMCS) to identify national rates of tobacco screening and treatment (provision of counseling, NRT, or both) by psychiatrists during outpatient visits before and after the introduction of the APA guidelines. We also examined associations of key patient and visit factors with tobacco screening and treatment.

METHODS

The NAMCS is an annual survey of US office-based physicians.¹⁰ Doctors of medicine (MDs) and doctors of osteopathy (DOs) participating in the survey are randomly assigned to a 1-week reporting period. A systematic random sample of each physician’s office visits during the reporting period is selected, and physicians or their office staff complete an encounter form for each selected visit. The form requests visit data including date, patient demographic characteristics, symptoms, diagnoses and source of payment, reason for visit, medications ordered, services provided, and planned future treatment. Physician data collected on the form include doctor type (MD or DO) and specialty (e.g., internal medicine, pediatrics, psychiatry).

Smoking status was not assessed in the 1997 to 2000 versions of the NAMCS, so we did not use data from those years. We selected records from the 1993 to 1996 and 2001 to 2010 NAMCS data files for all visits involving individuals 18 years or older who had been seen by an MD or DO with a specialty in psychiatry. We assessed the rate of screening for tobacco use by examining provider responses to the question “Does patient smoke cigarettes?” (on the 1993–1996 surveys) or “Does patient use tobacco?” (on the 2001–2010 surveys). Response options on the 1993 to 2004 forms were yes, no, and unknown; options on the 2005 to 2010 forms were current, not current, and unknown. If the provider checked yes or no on the 1993 to 2004 surveys or checked current or not current on the 2005 to 2010 surveys, we coded the patient as having been screened for tobacco use. If the provider checked unknown or did not respond to the question, we coded the patient as not having been screened for tobacco use.

We assessed whether patients who screened positive for tobacco use (i.e., yes on a 1993–2004 survey or current on a 2005–2010 survey) were provided with tobacco cessation counseling by examining provider responses to the section of the NAMCS forms asking which types of counseling or health education were provided during the visit. If a provider checked the box indicating he or she provided counseling or education on “smoking cessation” (on the 1993–1994 surveys) or “tobacco/use” (on the 1995–2010 surveys), we coded a smoker as having been counseled regarding tobacco cessation. If that box was not checked, we coded a smoker as not having been counseled on tobacco cessation. Nonsmokers and patients for whom no tobacco screening was completed were excluded from the analysis of this outcome.

We assessed whether patients who screened positive for tobacco use were provided with NRT by examining the list of medications prescribed during a visit. The NAMCS forms included fields for listing any medications prescribed during a visit. The 1993 and 1994 forms allowed up to 5 medications to be listed, the 1995 to 2006 forms allowed up to 6 medications to be listed, and the 2007 to 2010 forms allowed up to 8 medications to be listed. To avoid biasing this outcome in favor of the years after the APA guidelines by examining a greater number of medications in more recent years, we limited our analysis to the first 5 medications listed for all studied years. If any form of NRT appeared in the first 5 medications, a smoker was coded as having received an NRT prescription. If a form of NRT did not appear in the first 5 medications, a smoker was coded as not having received such a prescription. Nonsmokers and patients for whom no tobacco screening was completed were excluded from the analysis of this outcome.

We assessed whether each visit included noncessation health counseling by examining providers’ responses to the section of the NAMCS form asking which types of counseling or health education were provided during the visit. The NACMS forms in all studied years asked providers to check a box if they provided counseling or education on exercise and a separate box if they provided counseling or education on weight reduction. If the provider checked either of these boxes, we coded the visit as having included noncessation health counseling. If the provider did not check either of these boxes, we coded the visit as not having included such counseling.

We next coded the primary diagnosis of the patient seen during each visit. When completing the survey forms, physicians were asked to enter the full International Classification of Diseases, Ninth Revision (ICD-9)¹¹ codes for up to 3 diagnoses for the patient seen during the visit. The NAMCS data set includes a variable for the full ICD-9 code and a variable for only the first 3 digits of the code, which allows data users to examine broader diagnostic categories (e.g., a code of 296 indicates an affective disorder). We used the 3-digit broad category variable for our analyses and limited our analyses to the first code listed in each record to capture a patient’s primary diagnosis.

Finally, we coded the time period in which each visit occurred. We coded visits occurring from 1993 to 1996 as “preguidelines,” visits from 2001 to 2005 as “early postguidelines,” and visits from 2006 to 2010 as “late postguidelines.”

We used multiple logistic regression to compare rates of screening all patients for tobacco use, providing smokers with cessation counseling, and providing smokers with NRT between the preguidelines period and the 2 postguidelines periods. We controlled all analyses for patient race, age, and gender. We used the NAMCS race variable, which contains imputed values for missing data and categorizes all patients as White, Black, or “other.”

We then used multiple logistic regression to identify patient and visit factors independently associated with tobacco screening and treatment during the late postguidelines period (2006–2010). Factors included in the analyses were patient age, gender, and race; provision of noncessation health counseling during the visit; visit duration; diagnosis; and whether the provider had seen the patient before.

RESULTS

We analyzed data on 6267 outpatient visits from 1993 to 1996, 6803 visits from 2001 to 2005, and 5845 visits from 2006 to 2010. Table 1 shows descriptive statistics for key patient and visit variables during the preguidelines, early postguidelines, and late postguidelines periods. Approximately 60% of the visits during all 3 time periods were with female patients. Most visits were also with White patients and patients who had seen the psychiatrist before. Fewer than 5% of visits in all 3 time periods included counseling about weight or diet.

TABLE 1—

Patient and Visit Characteristics in the Preguidelines (1993–1996), Early Postguidelines (2001–2005), and Late Postguidelines (2006–2010) Periods: National Ambulatory Medical Care Survey

		2001–2005 (n = 6803)		2006–2010 (n = 5845)
Variable	1993–1996 (n = 6267), %	%	ORa (95% CI)	%	ORa (95% CI)
Age, y
18–25	7	10	1.47 (1.30, 1.67)	11	1.68 (1.48, 1.91)
26–35	18	13	0.68 (0.62, 0.75)	14	0.71 (0.64, 0.78)
36–45	31	24	0.71 (0.65, 0.76)	21	0.60 (0.55, 0.65)
46–55	23	28	1.25 (1.16, 1.35)	26	1.15 (1.06, 1.25)
56–65	9	14	1.66 (1.48, 1.85)	18	2.21 (1.98, 2.46)
> 65	10	10	0.98 (0.87, 1.09)	9	0.85 (0.75, 0.96)
Male gender	40	39	0.96 (0.90, 1.03)	40	0.98 (0.92, 1.06)
Race
White	92	92	0.93 (0.82, 1.05)	89	0.68 (0.60, 0.77)
Black	6	6	1.04 (0.90, 1.21)	7	1.33 (1.15, 1.55)
Other	2	2	1.17 (0.92, 1.48)	3	1.74 (1.39, 2.18)
Visit duration, min
0–15	7	18	2.58 (2.30, 2.89)	17	2.59 (2.31, 2.91)
16–30	29	36	1.34 (1.24, 1.44)	34	1.27 (1.17, 1.37)
31–45	21	18	0.81 (0.74, 0.88)	18	0.81 (0.74, 0.89)
46–60	36	24	0.55 (0.51, 0.60)	24	0.58 (0.53, 0.62)
> 60	2	2	0.78 (0.61, 0.99)	2	0.91 (0.71, 1.16)
Noncessation counseling delivered	3	3	0.91 (0.75, 1.12)	3	1.02 (0.84, 1.26)
Established patient	89	93	1.49 (1.32, 1.68)	95	2.14 (1.86, 2.46)
Primary diagnosis
Schizophrenia/other psychosis	8	8	0.91 (0.80, 1.03)	8	0.91 (0.79, 1.03)
Affective disorder	41	49	1.38 (1.29, 1.48)	49	1.38 (1.28, 1.48)
Anxiety disorder	27	21	0.75 (0.69, 0.81)	18	0.61 (0.56, 0.67)
Alcohol- or drug-related disorder	4	1	0.27 (0.21, 0.35)	3	0.73 (0.61, 0.89)
Personality disorder	4	3	0.62 (0.51, 0.75)	3	0.60 (0.49, 0.74)
Other	14	14	1.05 (0.95, 1.16)	15	1.14 (1.03, 1.26)
Rate of smoking among screened patients, overall and by diagnosis
All patients	25	22	0.83 (0.75, 0.91)	23	0.87 (0.79, 0.97)
Schizophrenia/psychosis	44	41	0.94 (0.71, 1.24)	30	0.55 (0.39, 0.76)
Affective disorder	22	22	1.00 (0.87, 1.16)	21	1.00 (0.86, 1.17)
Anxiety disorder	18	15	0.82 (0.66, 1.01)	16	0.93 (0.73, 1.18)
Alcohol- or drug-related disorder	76	44	0.30 (0.17, 0.55)	63	0.54 (0.33, 0.87)
Personality disorder	17	15	0.89 (0.51, 1.53)	18	1.06 (0.58, 1.95)
Other	16	10	0.71 (0.53, 0.96)	11	0.87 (0.65, 1.18)

Note. CI = confidence interval; OR = odds ratio.

^aIn comparison with 1993–1996 data.

Most visits in the preguidelines time period were more than 30 minutes long, whereas most visits in the postguidelines periods were less than 30 minutes long. Approximately 41% of visits in the preguidelines period were with patients who had an affective disorder diagnosis. This number grew to about 49% during the postguidelines periods. About 8% of visits during all 3 time periods were with patients who had a diagnosis of schizophrenia or other psychosis. Current smoking was identified most often during visits with patients with an alcohol- or drug-related diagnosis and those with a diagnosis of schizophrenia or psychosis.

Screening

Table 2 shows our comparisons of tobacco screening and treatment rates during the 3 studied time periods. Psychiatrists were significantly less likely to screen their patients for tobacco use during the early and late postguidelines periods than during the preguidelines period. Psychiatrists screened their patients for tobacco use in 77% of visits from 1993 to 1996, 69% of visits from 2001 to 2005, and 60% of visits from 2006 to 2010.

TABLE 2—

Tobacco Use Screening and Treatment by Outpatient Psychiatrists Before and After the Release of the 1996 American Psychiatric Association Guidelines: National Ambulatory Medical Care Survey

Variable	Screening or Treatment, %	ORa (95% CI)
Screening
1993–1996 (Ref)	77	1.00
2001–2005	69	0.69 (0.64, 0.75)
2006–2010	60	0.46 (0.43, 0.50)
Counseling
1993–1996 (Ref)	12	1.00
2001–2005	11	0.97 (0.74, 1.26)
2006–2010	23	2.23 (1.74, 2.86)
Nicotine replacement therapy
1993–1996	0.3	…b
2001–2005	0.2	…b
2006–2010	0.6	…b

Note. CI = confidence interval; OR = odds ratio.

^aAnalyses adjusted for patient age, race, and gender.

^bRates of prescribing nicotine replacement therapy were too small to generate stable estimates.

Table 3 shows the associations of patient and visit factors with tobacco screening and treatment during the late postguidelines period. Psychiatrists were less likely to screen Black patients than White patients for tobacco use; they were more likely to screen a patient for tobacco if the visit also included noncessation health counseling or if the visit was with a patient who had an alcohol or drug use disorder. In addition, as the time a psychiatrist spent with a patient increased, so did the odds that the psychiatrist screened the patient for tobacco use. Patient gender, age, and race other than White or Black were not found to be significantly associated with an increase or decrease in the odds of a psychiatrist screening a patient for tobacco use, and whether the provider had seen the patient before did not affect the odds of completing a tobacco screening.

TABLE 3—

Associations of Patient and Visit Factors With Rates of Tobacco Screening and Treatment by US Psychiatrists: National Ambulatory Medical Care Survey, 2006–2010

Variable	Screening		Counseling		Nicotine Replacement Therapy,a %
	%	OR (95% CI)	%	OR (95% CI)
Race
White (Ref)	61	1.00	22	1.00	0.7
Black	52	0.74 (0.60, 0.91)	30	1.67 (0.92, 3.05)	0.0
Other	61	0.97 (0.71, 1.33)	25	1.45 (0.53, 3.98)	0.0
Gender
Women (Ref)	60	1.00	20	1.00	0.5
Men	61	1.02 (0.91, 1.46)	27	1.48 (1.02, 2.14)	0.8
Age
18–25 (Ref)	60	1.00	17	1.00	0.0
26–35	60	1.03 (0.83, 1.29)	24	1.50 (0.76, 2.97)	0.8
36–45	60	1.06 (0.87, 1.30)	22	1.10 (0.58, 2.10)	0.5
46–55	60	1.03 (0.85, 1.26)	24	1.38 (0.75, 2.56)	1.3
56–65	60	1.04 (0.85, 1.28)	28	2.37 (1.16, 4.81)	0.0
> 65	63	1.12 (0.87, 1.44)	12	0.65 (0.17, 2.51)	0.0
Visit duration, min
0–15 (Ref)	41	1.00	26	1.00	0.0
16–30	56	1.79 (1.53, 2.09)	28	1.10 (0.66, 1.84)	0.9
31–45	69	3.04 (2.52, 3.67)	23	0.83 (0.44, 1.58)	0.8
46–60	74	3.91 (3.28, 4.68)	16	0.61 (0.33, 1.14)	0.6
> 60	68	2.64 (1.75, 3.98)	19	0.54 (0.13, 2.21)	0.0
Seen before
No (Ref)	66	1.00	13	1.00	0.0
Yes	60	0.95 (0.74, 1.23)	24	1.51 (0.67, 3.42)	0.7
Noncessation counseling
No (Ref)	60	1.00	21	1.00	0.7
Yes	81	2.97 (2.00, 4.39)	54	5.50 (2.80, 10.75)	0.0
Primary diagnosis
Affective disorder (Ref)	62	1.00	21	1.00	0.3
Schizophrenia/psychosis	52	0.85 (0.69, 1.06)	35	1.59 (0.87, 2.92)	1.4
Anxiety disorder	66	1.11 (0.95, 1.30)	16	0.83 (0.46, 1.48)	2.5
Alcohol- or drug-related disorder	63	1.53 (1.07, 2.19)	44	2.48 (1.39, 4.46)	0.0
Personality disorder	65	0.90 (0.63, 1.30)	28	2.02 (0.65, 6.30)	0.0
Other	59	0.95 (0.81, 1.12)	19	0.98 (0.53, 1.82)	0.0

Note. CI = confidence interval; OR = odds ratio. Data were based on 5845 visits.

^aRates of prescribing nicotine replacement therapy were too small to generate stable estimates.

Counseling

Rates of providing cessation counseling to smokers were low during all 3 time periods, but there was a significant increase in the rate of providing counseling in the late postguidelines period relative to the preguidelines period. Psychiatrists provided smoking cessation counseling in 12% of visits with smokers from 1993 to 1996, 11% of visits with smokers from 2001 to 2005, and 23% of visits with smokers from 2006 to 2010.

Psychiatrists were more likely to counsel men, patients aged 56 to 65 years (vs young adults aged 18–25 years), and patients with an alcohol- or drug-related disorder. Psychiatrists were also more likely to provide cessation counseling during a visit in which noncessation health counseling was delivered. Visit duration, patient race other than White or Black, and whether the psychiatrist had seen the patient were not associated with the odds of cessation counseling being delivered.

Nicotine Replacement Therapy

Rates of providing NRT to smokers were less than 1% in all 3 study periods. Because these rates were so small, we were unable to generate stable estimates of the relationships between our 3 outcomes, the study periods, and key patient and visit factors.

DISCUSSION

Smokers with psychiatric diagnoses are a vulnerable and high-risk population. Those with serious mental illnesses, such as schizophrenia or bipolar disorder, can spend up to one third of their monthly income on cigarettes and die up to 30 years earlier than the general population.¹² The increased mortality seen in this population is largely attributable to diseases for which smoking is a major risk factor, including cancer and cardiovascular disease.^13,14

APA released clinical practice guidelines for the treatment of nicotine dependence in 1996, recommending that heath care providers screen all patients with a psychiatric disorder for tobacco use and provide those interested in quitting with evidence-based treatments such as behavioral counseling and medications. There are several reasons why psychiatrists should follow these guidelines.

First, many mental health care patients do not have a regular source of nonpsychiatric medical care, and public mental health clinics are their first or only source of contact with the health care system.⁵ Therefore, these patients’ mental health providers are key sources of tobacco cessation information, recommendations, and treatment.

Second, smoking affects the metabolism of some antipsychotic and antidepressive medications, and smokers may require dosages of these medications up to 50% higher than those of nonsmokers.¹⁵ Having knowledge of a patient’s smoking status and assisting the patient with a quit attempt helps optimize a psychiatrist’s ability to maintain safe and effective psychiatric medication dosing.

Third, smoking cessation can negatively affect the mood of psychiatric (and nonpsychiatric) patients, and psychiatric symptoms can be strong barriers to successful cessation. As a patient’s primary mental health provider, with knowledge of the patient’s psychiatric symptoms and treatment history, psychiatrists are uniquely positioned to treat mood changes and mental health symptoms during quit attempts.

Despite the benefits of treating tobacco use in a mental health care setting, our analyses showed that US psychiatrists have been screening their patients for tobacco use at declining rates since the release of the APA guidelines. In the 4 years prior to the 1996 guidelines, psychiatrists were screening patients for tobacco use during 77% of visits, and this rate declined to an average of 60% from 2006 to 2010. Our findings suggest that lack of time may have played a role in this decline. Lack of time has been cited by primary care physicians as a common barrier to completing tobacco screening with patients.¹⁶ We found that short psychiatric visits were less likely to include screening for tobacco use.

Furthermore, the time psychiatrists spent with their patients decreased significantly after the release of the APA guidelines. Table 1 shows that from 1993 to 1996 the majority of visits were longer than 30 minutes and that fewer than 10% of visits were 15 minutes or shorter in duration. By the 2006 to 2010 time period, the majority of visits were shorter than 30 minutes, and almost 20% of visits were 15 minutes or shorter in duration. Table 1 data also show that the percentage of visits occurring with established patients increased over time (from about 89% during 1993–1996 to about 95% during 2006–2010). Thus, psychiatrists may be spending more of their time for brief visits with regular patients, and it may be more difficult to find time during such visits to initiate a discussion about tobacco.

Although the rate at which psychiatrists screened their patients for tobacco use declined after the release of the APA guidelines, the rate at which they provided cessation counseling to smokers nearly doubled from 12% during 1993 to 1996 to 23% during 2006 to 2010. This improvement may have been due to increased coverage of smoking cessation counseling in public and private insurance programs during this time. In July 2005, the Centers for Medicare and Medicaid Services implemented coverage for up to 8 intermediate or intensive outpatient counseling sessions per year for smokers, and by 2009 almost 65% of state Medicaid programs reported covering individual cessation counseling for eligible enrollees.¹⁷

This trend has also been seen in some private insurance sectors. In 2008, McMenamin et al. reported data from employer-sponsored health programs in California showing that the percentage of preferred provider organization plans offering coverage for cessation counseling increased from 29% in 2000 to 43% in 2005 and that the percentage of point of service plans offering such coverage increased from 16% to 38% during the same period.¹⁸ Such expansions in insurance coverage may have prompted increases in cessation counseling by psychiatrists.

Despite our finding of a significant increase in the percentage of smokers receiving cessation counseling, the absolute percentage of smokers counseled by their psychiatrists remained low. Even lower were rates of providing smokers with NRT. Psychiatrists prescribed NRT in fewer than 1% of visits with patients identified as current smokers during all 3 study periods. These results reflect a trend among US physicians in general of not providing guideline-concordant cessation treatment to their smoking patients. Thorndike et al. found that US physicians counseled only 20% of all smokers from 2001 to 2003 and provided cessation medications in fewer than 2% of visits with smokers.¹⁹ When breaking their results down by physician type, they found that even primary care providers counseled only 26% of smokers, whereas specialists provided cessation counseling during 14% of visits with smokers. In a separate 2011 survey of US outpatient oncologists and nonphysician oncology providers, fewer than 30% of respondents reported that they frequently or always provided cessation interventions to smoking patients.²⁰

Previous research has shown that many health care providers lack confidence, skills, or knowledge in effectively addressing tobacco use.^21,22 Our analyses showed that visits that included noncessation health counseling (e.g., on exercise or weight reduction) were more likely to include screening for and treatment of tobacco use. In addition, we found that psychiatrists were more likely to screen and counsel smokers with an alcohol- or drug-related disorder. Thus, psychiatrists who include healthful behavior change approaches in their routine care or who are already addressing problem behaviors such as alcohol or drug use may not only prioritize healthful behavior change but may also feel more comfortable and efficacious in helping their patients change unhealthy behaviors.

Limitations

There are several limitations to this study. First, the NAMCS relies on self-reported behavior by respondents, and there is a risk of misreporting of screening and treatment behavior by psychiatrists on the NAMCS forms. Second, we were unable to analyze tobacco screening and treatment rates in visits from 1997 to 2000 because the NAMCS did not include questions about tobacco screening during those years.

Third, we did not examine non-NRT smoking cessation prescriptions such as bupropion or varenicline. Bupropion is also prescribed for depression, and it would not have been possible for us to distinguish prescriptions for smoking and prescriptions for depression. Varenicline was not approved by the Food and Drug Administration until 2006, and thus it would have been available only during the late postguidelines period. It is possible that psychiatrists are more comfortable with, or knowledgeable of, non-NRT cessation medications and prescribed these medications to smoking patients instead of NRT. It is also possible that most smokers did not express an interest in quitting smoking, in which case the APA guidelines do not recommend prescribing cessation medications. Finally, we limited our analysis to the first 5 medications listed on the NAMCS. It is possible that some smokers were prescribed NRT but it was not listed among the first 5 medications. However, given the stable and very low rate of NRT prescriptions, these alternative hypotheses regarding the low rate of NRT prescriptions seem unlikely.

Conclusions

Despite these limitations, this study provides an important look at the limited adherence to the APA nicotine dependence treatment guidelines by US psychiatrists. Low rates of provider adherence to other types of clinical practice guidelines, such as those for HIV screening and oncology care, have been reported in the literature,^23,24 and studies have demonstrated that systems-level interventions are often needed to enable changes in clinical practice.^25–27 Our findings provide support for the systems-level interventions recommended in the Public Health Service guidelines for the treatment of tobacco.³ Systems interventions such as implementing a tobacco-user identification system; providing education, resources, and feedback to promote provider intervention; providing insurance coverage for evidence-based tobacco treatments; and reimbursing clinicians for delivery of evidence-based treatments may be needed to incorporate tobacco screening and treatment into routine psychiatric care.