|Home | About | Journals | Submit | Contact Us | Français|
This article examines the variables associated with the presence of smoking cessation interventions in drug abuse treatment units, as well as staff attitudes toward the integration of smoking cessation services as a component of care. Surveys were administered to 106 organizations, 348 treatment clinics, and 3,786 employees in agencies that participated in the National Drug Abuse Treatment Clinical Trials Network. Organizational factors, attributes of the treatment setting, and staff attitudes toward smoking cessation treatment were assessed. Use of smoking cessation interventions was associated with the number of additional services offered at clinics, residential detoxification services, and attitudes of the staff toward smoking cessation treatment. Staff attitudes toward integrating smoking cessation services in drug treatment were influenced by the number of pregnant women admitted, the number of ancillary services provided, the attitudes of staff toward evidence-based practices, and whether smoking cessation treatment was offered as a component of care.
The Department of Health and Human Services Clinical Guidelines for Treating Tobacco Use and Dependence recommends targeting drug and alcohol users for nicotine dependence treatment (Fiore, Bailey, & Cohen, 2000). Individuals with substance use disorders smoke at a rate higher than that of the general population (Batel, Pessione, Maitre, & Rueff, 1995; Bobo, Lando, Walker, & McIlvain, 1996; Burling, Ramsey, Seidner, & Kondo, 1997; Hurt, Eberman, Slade, & Karan, 1993; Kalman, 1998). Substance-abusing persons who smoke are more heavily addicted to nicotine (Hughes, 1996, 2002; Marks, Hill, Pomerleau, Mudd, & Blow, 1997; Sobell, 2002) and generally have ’more difficulty quitting smoking than nonsubstance-abusing smokers (Bobo, Gilchrist, Schilling, Noach, & Schinke, 1987; Joseph, Nichol, & Anderson, 1993; Kozlowski, Skinner, Kent, & Pope, 1989; Zimmerman, Warheit, Ulbrich, & Auth, 1990). Alcohol-dependent individuals who smoke have higher rates of cancer, and they die from smoking-related causes more frequently than from alcohol-related causes (Hurt et al., 1996).
Drug abuse programs are an optimal venue for delivering smoking cessation interventions. Many persons entering drug treatment express interest in quitting smoking when asked (Orleans & Hutchinson, 1993; Saxon, McGuffin, & Walker, 1997). When patients are referred to external smoking cessation clinics, they often do not follow through (Thompson et al., 1988). Integrating nicotine dependence treatment into drug abuse care reduces these attendance problems while also increasing patients’ sense of mastery and allowing programs to provide a consistent message that targets all addictive substances and focuses on positive lifestyle changes (Sussman, 2002).
Despite national guidelines that direct health care professionals to address nicotine dependence in drug abuse patients (Fiore et al., 2000), smoking is often overlooked in drug treatment. Barriers to using smoking cessation interventions include lack of available resources, little or no insurance coverage for tobacco dependence, and the cost of nicotine replacement therapy or other appropriate medications (e.g., buproprion). Other barriers include high rates of cigarette use among staff (approximately 40%), a culture amenable to smoking (e.g., “smoke breaks” structured into a treatment day), and the attitudes of treatment staff. Resistance to treating nicotine dependence among drug treatment staff has been documented (Bobo, Slade, & Hoffman, 1995; Capretto, 1993; Fishman & Earley, 1993; Goldsmith & Knapp, 1993; Hahn, Warnick, & Plemmons, 1999) and has been found to be rooted partly on traditional belief that those in treatment should avoid major life changes (including smoking cessation) during their first year of recovery and that stopping smoking may jeopardize recovery. Smoking may also be viewed as a low priority when compared to the more immediate harms of alcohol and illegal drug use (Bobo, 1992; Sees & Clark, 1993). Drug counselors may also believe that their patients are not interested in quitting smoking (Bobo, 1992; Sees & Clark, 1993). Counselors who smoke are more resistant to viewing client smoking as a treatment issue and are less likely to participate in discussions about clients’ nicotine dependence (Campbell, Krumenacker, & Stark, 1998).
A positive staff attitude toward the integration of smoking cessation interventions into drug abuse clinics is an important predictor of facilities that offer nicotine dependence services as a component of care. A majority of staff members with prior experience in implementing smoking cessation felt that such interventions had either a positive impact or no discernable impact on clients and staff. Only 10% felt that it had a negative impact (Williams, McGregor, Borrelli, Jordan, & Strecher, 2005).
In the context of the National Drug Abuse Treatment Clinical Trials Network (CTN) (McCarty, in press), directors and staff were surveyed concerning organizational characteristics and the range of services provided. The CTN is an alliance of research centers and drug treatment programs sponsored by the National Institute on Drug Abuse (NIDA) to conduct randomized trials of drug abuse treatments. The NIDA CTN aims to improve drug abuse treatment through two goals: to determine the effectiveness of promising interventions in multisite clinical trials and to support the transfer of tested and effective interventions into clinical practice (Hansen, Leshner, & Tai, 2002). NIDA has sustained a commitment to these goals in the past 7 years at a cost of approximately US$40 million per year. The CTN has made substantial progress in testing promising interventions. The network includes 17 research centers and more than 100 community treatment programs. More than 7,000 participants have been enrolled into a series of 21 multisite research protocols in various stages of completion (CTN Bulletin, 2006).
Data were collected from drug abuse treatment programs participating in the CTN, with three surveys administered between February 2002 and August 2004. Organizational surveys were given to the president or the chief executive officer of the program, treatment unit surveys were given to directors, and workforce surveys were given to medical, management, and counseling staff.
The organizational survey characterized each program at a macro level, with a focus on funding sources, annual revenue, mission, and number of full-time employees. The treatment unit survey included additional items assessing patients served, services provided, and program philosophy. Finally, the workforce survey was distributed to staff within each treatment unit and requested information on years of experience, education, training, licensing, credentials, and job title.
Each node identified a node protocol coordinator who served as node liaison and the individual with whom the Oregon Node coordinated study implementation. Coordinators were responsible for facilitating data collection and communication with treatment programs. They confirmed contact individuals at each program and facilitated the distribution of paper surveys or passwords for optional web-based surveys. In addition, they worked with the Oregon Node to monitor response rates and promote participation. The Oregon Node trained coordinators and supported their efforts to promote survey participation. The research plan and materials were reviewed by the Institutional Review Board at the Oregon Health and Science University and by those in participating locations. Information sheets (or consent forms) were provided to all study participants.
For the organizational survey, 106 surveys were collected from 112 eligible treatment programs. The treatment unit survey collected data from 348 of 388 treatment units. The director of each treatment unit provided the names and addresses of staff members who are eligible to complete the workforce survey. For the workforce survey, 3,786 of 5,334 eligible respondents responded (71% of eligible individuals).
Specific items were derived from each of the surveys. The variables created for these analyses are detailed in Table 1.
First, logistic regression was used to model the probability of the treatment unit providing smoking cessation treatment. This outcome variable was defined as a dichotomous indicator of whether the treatment unit offered smoking cessation interventions as a part of their curriculum.
Second, a multiple linear regression model determined the relative contribution of several predictors of staff attitudes toward the integration of smoking cessation treatment as a component of care. This analysis was conducted at the workforce survey level. Every staff member who had a valid observation on an item assessing staff attitudes toward smoking cessation was included in this analysis.
A consistent problem with surveys of this type is the presence of missing data. A multiple imputation approach (Little & Rubin, 1987) replaced each missing value with a set of plausible values. This approach calculates accurate estimates of standard errors.
A set of regression parameters was generated for each of 20 datasets generated by PROC MI (SAS Institute, 1999). To condense the output so that 20 separate analyses could be presented together, PROC MIANALYZE in SAS averaged the values to one set of parameters. These values are a stable set of parameters that reduced sample-specific effects (i.e., significant findings that are found in one imputation but not found in others) and reflected the best estimates of a full dataset. The means and standard deviations presented in Table 2 are nonimputed values. The only imputed values are the regression parameters for the logistic regression presented in Table 3 and for the multiple regression presented in Table 4.
The logistic regression simultaneously entered 14 variables into a model to determine which were associated with the presence or the absence of smoking cessation treatment at each clinic. A multiple imputation approach, which estimated values for missing data in the model with the Expectation/Maximization (EM) algorithm for 20 imputations, was employed. A logistic regression procedure was conducted on each dataset and produced 20 sets of parameter estimates. All of the models were significant (p < .0001), with chi-square estimates ranging from 81.18 to 91.83, with 14 df.
A second regression model used data from the workforce survey to determine which of the 19 staff and organizational variables were significantly associated with staff attitudes toward nicotine dependence interventions. A multiple imputation procedure was used to replace missing values. Twenty imputations were computed, and each imputed model was significant (p < .0001). The F values with (19, 3104) df ranged from 9.68 to 10.66, with R2 ranging from .0520 to .0613. The low R2 was due to the outcome variable having only five potential values (e.g., 1–5). Regression analysis is relatively robust to ordinal-level outcome variables of this type.
The 342 treatment units included 106 (31%) units that offered some kind of smoking cessation intervention and 235 (69%) units that offered no treatment for nicotine dependence. Table 2 provides the means, standard deviations, and range of values for 14 organizational predictors, and the criterion (provision of smoking cessation treatment).
Table 3 shows the solution for logistic regression parameter estimates aggregated in the 20 datasets. This solution suggests that the presence of smoking cessation at each treatment unit was positively related to three variables: (a) the mean attitude of the staff about smoking cessation treatment; (b) the number of additional mental health and medical services offered at the clinic; and (c) the presence of a residential detoxification program.
Staff attitudes toward integrating smoking cessation interventions (measured on a 5-point Likert scale) were more positive in agencies that offered some kind of treatment for nicotine dependence (M = 3.7) compared with units that did not (M = 3.5). Treatment units providing smoking cessation interventions also provided more ancillary services (M = 8.2) compared with those clinics that did not (M = 5.8). Finally, residential detoxification facilities were more common among treatment units with smoking cessation services (36% with vs. 19% without).
Table 2 also presents the means and standard deviations for the variables in the multiple regression model. Table 4 provides the imputed regression parameter estimates, standard errors, t values, and p values for each variable in the model. Positive attitudes toward smoking cessation were associated with: (a) the use of smoking cessation interventions as part of treatment; (b) a higher number of women admitted; (c) a higher number of pregnant women admitted; (d) being a program in a Veterans Administration (VA) Medical Center; (e) positive attitudes toward evidence-based practices; and (f) familiarity with American Society of Addiction Medicine (ASAM) placement criteria. Attitudes were less supportive if the unit also contained a residential detoxification service. The zero-order correlation of this latter variable was r = −.107, indicating that this negative weight is not due to model-specific effects. This finding and its apparent contradiction with the conclusions of the logistic model are discussed below.
A national sample of drug abuse treatment units provided information on variables associated with the adoption of smoking cessation treatment and the factors associated with positive staff attitudes about integrating nicotine dependence interventions as a component of care. This study examines the provision of smoking cessation services in a wide range of programs, including drug-free residential, methadone maintenance, outpatient, inpatient, and detoxification facilities. Factors at both treatment unit and staff levels were associated with the presence or the absence of smoking cessation services within drug abuse clinics.
Smoking cessation treatment was more likely to be available in units that offered other ancillary services, including detoxification. Treatment units providing multiple medical and mental health services appear to be more likely to offer smoking cessation interventions. Stand-alone drug abuse treatment programs were less likely to offer smoking cessation interventions. Treatment programs that provided a more comprehensive level of service were more likely to have the resources to provide nicotine dependence treatment. It is unclear whether the smoking cessation treatment was provided through ancillary services such as primary medical care. This might imply that smoking cessation was generally more acceptable in other health care settings but not in independent drug abuse treatment settings.
The second model examined staff attitudes. Employees were more likely to have a positive view of smoking cessation treatment if the clinic operated a nicotine dependence program, admitted certain populations (veterans, women, and pregnant women), and did not offer residential detoxification services. Counselor attitudes about evidence-based practices and ASAM placement criteria contributed to a positive attitude toward smoking cessation treatment.
These results are a mix of intuitive and paradoxical findings. Clinics providing smoking cessation care were more likely to have staff members with a supportive attitude toward such services. This is consistent with previous findings (Hahn et al., 1999; Hurt, Croghan, Offord, Eberman, & Morse, 1995; Williams et al., 2005). Staff members with a positive view toward smoking cessation may be more likely to refer patients to the program. This raises the question as to whether the presence of nicotine dependence treatment improves staff attitudes or whether having a supportive staff increases the likelihood that a clinic would offer smoking cessation interventions. The association between staff attitude and the provision of smoking cessation treatment is likely bidirectional.
The results demonstrate that staff members who worked in clinics with a high number of pregnant women were more likely to support the integration of smoking cessation into drug abuse treatment. This finding may reflect that individuals in clinics serving pregnant and perinatal women are more aware of the negative impacts of smoking on fetal development and are more ready to integrate smoking cessation services into their clinics. On the other hand, the proportion of youth admissions was a predictor neither for staff attitudes nor for the provision of smoking cessation services.
A curious finding is the negative relationship between staff attitudes toward smoking cessation treatment and residential detoxification services. Additional analyses confirmed that this effect was due to most detoxification facilities operating in large hospital settings or any other level-of-care effect. Although staff attitudes toward the integration of smoking cessation services were less positive in stand-alone detoxification facilities than in multiservice agencies (3.09 vs. 3.43), this difference was not significant. Staff attitudes toward smoking cessation interventions in these settings may be negative because the staff members are focused on patients on withdrawal and may believe that removing smoking during this period will only make the patient more uncomfortable. This may contribute to the ideation that patients are likely to leave detoxification facilities prematurely because of cigarette cravings. Concerns such as these may lead staff members to have more negative attitudes toward the use of smoking cessation treatment. With staff education and administrative commitment, these attitudes usually change (Williams et al., 2005). There is no evidence that more patients actually leave treatment because of smoking restrictions.
Employees working in VA Medical Centers tended to have more positive attitudes toward smoking cessation than the rest of the workforce and reflected governmental regulations requiring a smoke-free environment in VA hospitals (as well as most other health care facilities). Although a small number of drug treatment clinics were a part of VA Medical Centers (n = 15), there was still a significant effect on the regression equation (confirmed by a significant univariate correlation).
Respondents who valued evidence-based practices and those who perceived themselves to be knowledgeable of the ASAM placement criteria were also more likely to favor the integration of smoking cessation into drug treatment units. These staff members may be more aware of both the importance and the current techniques of treating nicotine dependence.
As with any large-scale survey, missing data were a limitation. The use of data imputation in constructing these models is a technological tool that allows the best use of incomplete data. In this study, most of the incomplete data were single omissions of items rather than large spans of missing data. In this case, imputation was the best way to use these data to determine underlying relationships.
Because the focus of the survey instruments was not on the provision of smoking cessation treatment, no details regarding specific types of services were obtained. For some clinics, smoking cessation treatment could mean providing nicotine replacement while the patient is on detoxification. For others, a program might involve nicotine replacement therapy, social support groups, and psychoeducational counseling or medication. This variation was not assessed and, thus, it is unclear what the respondent meant when indicating the presence of smoking cessation treatment at the clinic.
Staff surveys indicate that lack of demonstrated efficacy and lack of client interest are big barriers for the implementation of smoking reduction interventions while on treatment (Walsh, Bowman, Tzelepis, & Lecathelinais, 2005). A large proportion of the workforce who smoke cigarettes was less likely to suggest smoking cessation treatment to their clients (Bobo & Gilchrist, 1983). Some staff members believe that it is therapeutic to occasionally smoke with their clients (Walsh et al., 2005).
Research has indicated that nicotine dependence treatment does not jeopardize drug treatment and may actually help recovery (Burling, Marshall, & Seidner, 1991; Hurt et al., 1994; Martin et al., 1997; Stuyt, 1997; Toneatto, Sobell, Sobell, & Kozlowski, 1995). Some research studies demonstrate that smoking cessation interventions improve long-term abstinence from alcohol or drugs, but not tobacco use (Bobo, McIlvain, Lando, Walker, & Leed-Kelly, 1998; Prochaska, Delucchi, & Hall, 2004). Although smoking interventions started early in residential treatment have been shown to affect abstinence rates, these effects are largely short term (Joseph, Willenbring, Ngent, & Nelson, 2004). Studies examining the effectiveness of smoking cessation treatment in drug treatment show short-term (6-month) reductions in cigarette use, but do not show long-term (18-month) effects (Prochaska et al., 2004). It is unclear to what extent treatment staff members are aware of these findings and how much this lack of evidence influences staff attitudes. In addition, inconclusive is how generalizable the current findings are when compared to treatment agencies not affiliated with the CTN. Although this is likely a good sample of treatment units, the CTN may have more multifaceted clinics than a random sample of agencies would contain.
Future research in this area should focus on the types of smoking cessation interventions that are provided at each agency to better characterize the types of services available. Additionally, a qualitative study of staff attitudes toward smoking cessation would greatly improve the understanding of the complex attitudes of staff members, particularly in detoxification centers. Research into programs designed to educate staff about the importance of integrating smoking cessation into drug abuse clinics is also merited. An examination is also needed to determine the how attitudes of patients and providers are influenced by the large smoking population in the workforce.
This study presents some challenges to the treatment field to focus on evidence-based services regarding smoking cessation treatment and raises some ethical issues as well. Pregnant clients who do not receive nicotine dependence treatment have limited ability to eliminate tobacco use, leading to more fetal complications. Treatment clinics for youths that ignore smoking cessation education do a disservice to a vulnerable population that may face a lifetime of tobacco addiction. The incorporation of evidence-based practices can be enhanced by the adoption of concurrent tobacco cessation services during rehabilitation—clearly an asset to good health and client recovery.
Cooperative agreements from the NIDA supported the design, distribution, collection, and analysis of the organizational, treatment unit, and workforce surveys within the CTN: Oregon Node (U10 DA13036), California–Arizona Node (U10 DA15815), Delaware Valley Node (U10 DA13043), Florida Node (U10 DA13720), Great Lakes Node (U10 DA13710), Long Island Node (U10 DA13035), Mid-Atlantic Node (U10 DA13034), New England Node (U10 DA13038), New York Node (U10 DA13046), North Carolina Node (U10 DA13711), Northern New England Node (U10 DA15831), Ohio Valley Node (U10 DA13732), Pacific Node (U10 DA13045), Rocky Mountain Node (U10 DA13716), South Carolina Node (U10 DA13727), Southwest Node (U10 DA15833), and Washington Node (U10 DA13714). We appreciate the support and participation of executive directors, treatment unit directors, and the work-force in the participating clinics.