Introduction

Despite decreasing cigarette smoking rates in the general US population, the rate of smoking among persons living with HIV (PLWH) remains disproportionately high. As of 2016, an estimated 15.5% of US adults 18 years or older smoked cigarettes [1], but recent estimated rates among PLWH range from 42.4% to 70% [2,3,4,5]. There are a variety of health risks related to smoking that may be enhanced by the presence of HIV infection, including increased susceptibility to pneumonia and chronic obstructive pulmonary disease and an increased risk of non-AIDS-defining cancers [4, 6,7,8,9]. The negative effects of smoking for PLWH can also limit the HIV treatment benefits of antiretroviral therapy (ART) [10, 11]. PLWH have an increased risk of cancer on average compared to the general population overall, and lung cancer is currently a leading cause of mortality among PLWH [11,12,13]. PLWH who are on ART with viral suppression but smoke cigarettes are predicted to lose more life years to smoking than to HIV infection [11, 13]. PLWH who smoke are also less likely to quit smoking than persons who do not have HIV (32.4% vs. 51.7%) [4].

Thus, smoking cigarettes is an important modifiable risk factor that should be a high priority among PLWH, even amid other important health care priorities. However, the need for smoking cessation and the potential health risks of smoking are often not thoroughly emphasized to PLWH by health care providers, in part due to multiple competing health care needs of this population. Therefore, interventions for PLWH and providers that care for PLWH are necessary to increase cessation among this population.

There are many evidence-based smoking cessation interventions with varying degrees of efficacy, efficiency, and cost. These include pharmacotherapy, counseling, information and communications technology (i.e., utilizing the Internet, email, cell phones, or text messaging), and clinic-level interventions. Pharmacotherapy used in smoking cessation interventions, such as nicotine replacement therapy (NRT), bupropion, and varenicline, can be applied among PLWH who smoke as there are limited to no drug interactions when combined with ART [14]. In addition to pharmacotherapy-based interventions, counseling-based interventions can include brief information sessions, individual or group therapy, and motivational interviewing. Many interventions combine pharmacotherapy with counseling to increase cessation-related outcomes [15]. Information and communications technology interventions, including the use of cell phones for delivery of counseling and Internet-based cessation interventions, have also recently been developed and used, including with PLWH [16].

Clinic-level interventions can also be implemented to address the challenges of engaging PLWH in smoking cessation by encouraging provider and clinic staff involvement in promoting cessation. Based on extensive evidence supporting the use of screening and brief interventions, the US Public Health Services Clinical Practice Guideline for Treating Tobacco Use and Dependence (PHS Guidelines) recommends clinic-level interventions, such as tobacco-user identification systems coupled with provider-led interventions that focus on robust patient care to help patients quit (e.g., the 5 As [Ask, Advise, Assess, Assist, Arrange] and pharmacotherapy) [17]. The importance of providers’ role in supporting smoking cessation through the 5As for PLWH has been well documented. Recommendations for smoking cessation by providers during regular care, even brief advice (e.g., 3 min), and further involvement in cessation interventions in clinical settings have been found to increase cessation attempts and abstinence rates in the general population [17, 18]. However, providers working with PLWH are less likely to emphasize smoking cessation recommendations than providers working with persons who do not have HIV [19]. For populations with multiple competing health care needs such as PLWH, the effective engagement of providers and clinic staff in smoking cessation efforts is an important goal.

There is also a need for particular attention to PLWH who may be especially susceptible to tobacco-induced disease. For example, persons living in the US South have among the highest rates of HIV infection in the country [20], live in a region with strong ties to tobacco growing and manufacturing, and have high rates of smoking [21]. Persons of color, particularly African American/Black and LatinxFootnote 1 communities, bear a disproportionate burden of HIV infection and also face smoking-related health disparities [22,23,24]. African Americans/Blacks, for example, have similar smoking rates to Whites but are more likely to die from tobacco-induced diseases and have fewer successful quit attempts, and some Latinx subgroups (e.g., Puerto Ricans) have higher smoking rates than the general population [22, 23]. Gay, bisexual, and other men who have sex with men (MSM) and transgender women are also disproportionately affected by HIV and have higher rates of smoking than their heterosexual and cisgender peers, respectively [25,26,27,28]. In addition, communities of color and lesbian, gay, bisexual, and transgender (LGBT) communities are more heavily targeted by tobacco industry marketing [22, 23, 27].

Purpose

In order to reduce smoking rates among PLWH and promote cessation to improve their health, interventions are needed to address their unique needs. The purpose of this qualitative systematic review was to assess the literature on smoking cessation interventions for PLWH and to identify gaps and opportunities for future intervention-oriented research.

Methods

A qualitative systematic review [29] of the published literature was conducted using three online databases: PubMed, PsychInfo, and EBSCO Academic Search Premier. The Boolean terms and search terms used were: (HIV OR HIV/AIDS OR AIDS) AND (tobacco OR smoking) AND (cessation OR quit OR abstinence) AND (intervention OR program OR treatment). Databases were searched from their inception through December 31, 2017. Citations from previous systematic reviews of smoking cessation interventions with PLWH were also searched for relevant papers.

After initial search results were retrieved from the databases, articles written in English were further screened. Articles from peer-reviewed journals describing interventions for smoking cessation in adults (≥ 18 years old) living with HIV were included. Duplicates were removed, and search results were screened by title and abstract to determine whether the articles met criteria for inclusion in the review. Finally, relevant articles were selected and interventions described by those articles were identified.

Data Collection

An abstraction form was created to abstract data from the identified articles. This form collected information on intervention study location, participant description (i.e., sample size and demographic characteristics), theories/models, intervention description, study design, primary cessation-related behavioral outcome measures, and study results. Articles that did not describe an intervention or did not report sufficient information to be abstractable were excluded. Data from multiple articles that described the same intervention study were abstracted under the same intervention.

Results

The initial search retrieved 1031 English articles. From these articles, 65 potentially relevant articles were identified by screening their titles and abstracts. Screening the full text of the potentially relevant articles initially identified led to the exclusion of 15 articles, and an additional 18 were excluded during data abstraction based on the criteria described above. From the remaining 32 articles, 28 unique intervention studies were identified, and data were abstracted from these articles (See Fig. 1).

Fig. 1
figure 1

PRISMA diagram of steps of the qualitative systematic review of the literature

Full size image

Intervention Locations and Participant Characteristics

Of the 28 intervention studies outlined in Table 1, seven were conducted in the US Northeast [31, 33, 44, 45, 47, 51, 55, 56, 62], five in the South [34,35,36, 46, 58,59,60], five in the Midwest [32, 40, 49, 52, 61], and four in the West [42, 48, 54, 57]; six were conducted in international settings [30, 37,38,39, 41, 53] and the location of one study was not specified [43].

Table 1 Review of smoking cessation interventions for persons living with HIV
Full size table

Intervention study sample sizes ranged from 15 [61] to 1689 [41], with half of the studies having sample sizes under 100 [30, 34, 37,38,39, 43, 46, 48, 49, 52, 54, 57, 58, 61]. Ten of the 28 intervention studies had samples that were majority African American/Black [33, 35, 36, 43,44,45,46, 49, 51, 55, 58,59,60], but only one intervention was tailored specifically for African American/Black communities [49]. Eleven had Latinx representation that ranged from 8 to 100% [31, 42, 44, 45, 47, 48, 51, 54,55,56,57,58,59,60], of which five presented any Latinx-specific findings [31, 47, 51, 54, 56, 57]. Four intervention studies explicitly noted American Indian/Native American representation [31, 42, 47, 51, 55], but these groups only constituted 2 to 6% of the sample population, and thus no significant results were identified specifically for this group. Only one intervention study included an Asian or Asian American participant [58]. Seven intervention studies included men who identified as gay, bisexual, or other MSM, who constituted between 25 to 84% of study samples [30, 38, 41, 42, 49, 58,59,60], and six documented participation of transgender persons but only in small numbers constituting 1 to 3% of study samples [30, 42,43,44, 51, 55]. Only one study described an intervention that was specifically designed to address the needs of MSM in a culturally congruent manner [49], and none were focused specifically on transgender persons.

Theoretical Underpinnings

Twenty-four of the 28 interventions described a theoretical model or framework. The most commonly used theories included cognitive behavioral therapy (CBT), which was used in 10 of the interventions [30, 32, 39, 42, 43, 46, 49, 53, 58,59,60]; the transtheoretical model of change (TTM), which was used in four interventions [31, 39, 41, 47, 53]; social cognitive theory (SCT), which was used in two interventions [44, 45, 51]; and the Screening, Brief Intervention, and Referral for Treatment (SBIRT) model, which was also used in two interventions [34,35,36]. Two intervention studies did not report a theoretical underpinning or framework beyond the PHS Guidelines [37, 40], which were used in eight of the intervention studies overall [32, 37, 40, 46, 48, 49, 58,59,60]. Four intervention studies did not report any theoretical underpinning [33, 52, 54, 57].

Study Designs

There were 15 randomized controlled trials (RCTs) among the 28 intervention studies, all of which involved randomization at the participant level [31, 34,35,36, 42,43,44,45, 47, 48, 51, 52, 55,56,57,58,59,60,61]. Of these RCTs, cessation rates at follow-up ranged from 4 to 63%, and seven produced statistically significant improvements in cessation-related outcomes [35, 36, 44, 45, 48, 56, 58,59,60,61].

The 13 intervention studies that were not RCTs were primarily single-group longitudinal studies with pre- and post-test and pilot studies with non-randomized control groups [30, 32, 33, 37,38,39,40,41, 46, 49, 53, 54, 62]. Among these non-randomized studies, cessation rates at follow-up ranged from 6 to 75%, and six demonstrated significant improvements in cessation-related outcomes [30, 37, 39,40,41, 49].

Among the 21 intervention studies that included a comparison group, comparison group participants received usual care [31, 34,35,36, 39, 44, 45, 47, 48, 51, 52, 55, 57,58,59,60,61] or a comparison intervention such as self-guided reading or brief advice sessions [40, 42, 43, 56].

Outcome Measures

Across the intervention studies, the measures used to determine cessation-related outcomes varied. Reported outcome measures included self-reported 24-h [58,59,60], 7-day [31,32,33, 37, 40, 42, 44, 45, 47,48,49, 51, 55,56,57,58,59,60,61], and 2-week [46] point prevalence abstinence (PPA); self-reported 4-week [37], 30-day [59, 60], 90-day [57], 6-month [41, 52], 12-month [39], or continuous abstinence [30, 43, 53, 59,60,61] or cessation [38]; self-reported number of cigarettes smoked per day [34,35,36,37, 43, 48, 49, 56] or daily nicotine intake [38]; self-reported quit attempts [32, 35, 36, 46]; and adherence to NRT [54]. The most frequently used measures were 7-day PPA, usually at 3 or 6 months after baseline, and self-reported number of cigarettes smoked per day. Self-reported measures were often biochemically verified using expired carbon monoxide measures (e.g., three [43], five [57], seven [58], eight [32, 44, 45, 49, 61], or 10 [30, 31, 40, 42, 47, 51, 53, 55, 56] parts per million) or cotinine or nicotine levels in blood [37], saliva [32, 40, 52], or urine [48].

Follow-up periods for collecting data on outcome measures varied from immediate post-test [46] to 1 month [35, 36, 48] to 1 year [32, 42, 56, 59, 60].

Intervention Results by Strategy

Most interventions used multiple strategies, including combinations of counseling, pharmacotherapy, and information and communications technology.

Counseling was a widely used intervention strategy, and several different kinds of counseling approaches were applied. Twenty-four of the 28 interventions used one or more forms of counseling [30,31,32,33,34,35,36,37,38,39,40, 42,43,44,45,46,47,48,49, 51, 53, 54, 56,57,58,59,60,61]. Thirteen interventions used face-to-face individual counseling delivery methods [30, 33,34,35,36,37,38,39,40, 42, 54, 56, 57, 61], 9 used phone-based counseling [31, 32, 40, 44,45,46,47, 56, 58,59,60,61], and 3 used group counseling [49, 51, 53]. Four of the counseling interventions used motivational interviewing techniques [31, 43,44,45, 47, 48]. Eight interventions included a single counseling session [33,34,35,36, 38, 43, 48, 54, 57] and 16 were multi-session, ranging from 2 to 15 sessions [30,31,32, 37, 39, 40, 42, 44,45,46,47, 49, 51, 53, 56, 58,59,60,61]. Fourteen of the intervention studies using counseling were RCTs [31, 34,35,36, 42,43,44,45, 47, 48, 51, 55,56,57,58,59,60,61].

All of the 13 intervention studies that used face-to-face individual counseling also included pharmacotherapy, and 7 showed statistically significant improvements in cessation-related outcomes [30, 35,36,37, 39, 40, 56, 61]. Three of these successful interventions also included a cell phone-based counseling component [40, 56, 61].

Of the three group counseling interventions, all used combined strategies that also included pharmacotherapy, but only one demonstrated a significant improvement in cessation-related outcomes (significant reduction in cigarettes smoked per day) [49]. Another resulted in a trend toward higher cessation rates in the intervention group and produced significant results when participants who were lost to follow-up or did not attend intervention sessions were excluded [51].

All four interventions using motivational interviewing also used pharmacotherapy, and two showed significant improvements in cessation-related outcomes [44, 45, 48], one of which also incorporated cell phone voice and text communications [44, 45].

Eleven interventions incorporated the use of information and communications technology [31, 32, 40, 42, 44,45,46,47, 55, 56, 58,59,60,61]. These included all nine phone-based counseling interventions, of which all but one [46] incorporated pharmacotherapy and six had statistically significant improvements in cessation-related outcomes [40, 44, 45, 56, 58,59,60,61]. In addition to the three phone-based counseling interventions with significant results that also included face-to-face individual counseling [40, 56, 61], another also included text messaging and a motivational interviewing component [44, 45]. Two interventions were web-based [42, 55], both in combination with NRT and one with a face-to-face individual counseling component [42]; however, neither produced significant improvements in cessation-related outcomes.

Nearly all (n = 24) of the reviewed interventions used some form of pharmacotherapy [30,31,32,33,34,35,36,37,38,39,40, 42,43,44,45, 47,48,49, 51, 53,54,55,56,57,58,59,60,61], which included NRT, bupropion, and varenicline; three out of the four interventions that did not use pharmacotherapy were clinic-level interventions [41, 52, 62]. Of the interventions that used pharmacotherapy, 17 offered NRT alone [30, 31, 34, 38, 39, 42, 43, 47,48,49, 51, 54,55,56,57,58,59,60,61], 2 offered varenicline alone [37, 44, 45], 2 offered both NRT and varenicline [32, 40], 1 offered both NRT and bupropion [53], and 2 offered all three forms of pharmacotherapy [33, 35, 36]. Fourteen of the intervention studies that used pharmacotherapy were RCTs [31, 34,35,36, 42,43,44,45, 47, 48, 51, 55,56,57,58,59,60,61], 7 of which showed statistically significant improvements in cessation-related outcomes [35, 36, 44, 45, 48, 56, 58,59,60,61]; 5 of the 10 intervention studies that involved pharmacotherapy and did not randomize participants also produced statistically significant results [30, 37, 39, 40, 49]. All of the studies using pharmacotherapy paired this with another intervention strategy.

There were only five clinic-level smoking cessation interventions [34,35,36, 41, 52, 62]. Clinic-level interventions involved changes to the clinic environment in order to promote smoking cessation for PLWH. Intervention components included trainings for staff [41, 62], checklists or electronic processes for screening and documenting patients’ smoking status [34, 41, 62], electronic algorithms to create personalized treatment plans [35, 36], and processes for making referrals for smoking cessation services [52, 62]. These strategies were often implemented in combination with individual-level strategies such as counseling and pharmacotherapy [34,35,36]. All of the clinic-level smoking cessation intervention studies reported increases in smoking cessation, though not all were statistically significant. Of the five intervention studies that involved clinic-level strategies, three were RCTs (with randomization at the patient level) [34,35,36, 52] and one had significant results [35, 36]; this intervention also involved the use of NRT and face-to-face individual counseling. One of the clinic-level intervention studies that did not involve randomization or any other intervention strategies also produced significant improvements in cessation-related outcomes [41].

Discussion

The importance of developing interventions tailored to specific at-risk populations has been noted by other reviews of smoking cessation interventions among PLWH [63], yet there are still many gaps within the literature in terms of interventions for PLWH with identities that are part of marginalized groups. Although a few studies evaluated interventions that were tailored specifically to the racial/ethnic, sexual, and/or gender identities of their participants, more studies aimed at developing, implementing, and evaluating culturally congruent interventions are needed. In particular, there is a need for more studies that focus on PLWH who are from underrepresented racial/ethnic groups such as Asian Americans or American Indian/Native Americans and PLWH who are sexual and gender minorities, particularly MSM and transgender women of color. It is important that these interventions consider the specific cultural, social, and structural factors that influence smoking within these groups, as behaviors, barriers to cessation, and resources may differ across different groups of PLWH. Such culturally congruent interventions may be more likely to achieve successful results among these particularly vulnerable groups than interventions not tailored to these specific contexts.

Although the number of interventions developed to promote smoking cessation among PLWH in the South is similar to the number of interventions developed in other regions, it is important to note that there have only been two research teams in two states within the South working on these types of intervention studies, which explore the efficacy of cell phone and electronic algorithm interventions. There remains little diversity in these studies, both in terms of intervention strategies and population groups for which the interventions are designed. Furthermore, contextual factors in the South may mean that some intervention strategies may be more appropriate and feasible than others; for example, a high proportion of persons in the South live in rural settings [64] and may face challenges traveling long distances to clinics to participate in multi-session face-to-face interventions. Given the disproportionate burden of HIV and smoking and the unique context of the South, more varied research is needed in this region.

There have been a number of pilot studies and quasi-experimental studies that show promising results; however, more robust studies with more statistical power are needed to obtain more generalizable results. Due to the wide variety of measures used to assess smoking cessation, it is difficult to compare across studies and to gauge the relative success of interventions. It is important that future studies utilize reliable and standardized measures for cessation to ensure internal and external validity; it is also recommended that a standard be established in order to more effectively compare the results of different types of interventions among different populations of PLWH.

Most of the studies used a theoretical framework to design their interventions. The TTM is based on the understanding that there are discrete stages in the process of behavioral change that relate to the likelihood of successful behavior modification and assumes that individuals must be ready to change their behavior before being encouraged to act. The SBIRT model, on the other hand, does not require expressed interest or readiness for change in order for an individual to be offered an opportunity to participate in a behavior change intervention. Interventions informed by both frameworks have been found to be feasible and promising, but there are fewer interventions that explore the idea of offering cessation opportunities to all PLWH who smoke, regardless of stated openness to or readiness for behavioral change. Studies that further explore the theoretical differences in these two frameworks would be helpful in determining the best theoretical foundation for smoking cessation among PLWH. In particular, more studies using SBIRT and other models that do not rely on patient interest and readiness to quit smoking are warranted.

Despite having been found to be effective for promoting smoking cessation in the general population [65,66,67], intervention studies using motivational interviewing have had mixed results. However, no findings indicated that the unique characteristics of PLWH would make this strategy inappropriate or not well-suited for this population.

The use of information and communications technology in interventions produced promising results. Cell phone-based counseling tended to yield positive results, and intervention studies that provided prepaid cell phones to study participants may have been particularly effective in part because some PLWH do not have regular access to cell phone service. Web-based interventions had less success in producing statistically significant outcomes. Web-based intervention studies may be able to increase their efficacy and cessation rates by ensuring that participants have regular and easy access to computers and a reliable Internet connection, and further studies with improved experimental designs are necessary to better gauge the efficacy and feasibility of these types of interventions.

Clinic-level interventions are especially promising for PLWH who smoke because this population tends to come into contact with providers on a regular basis due to HIV-related medical care. Studies have found that care provider recommendations for smoking cessation increase attempts; however, cessation has not been found to be a priority among HIV care providers. Providers may need additional training to understand smoking cessation for PLWH and the unique barriers to cessation that this population faces. For example, studies have documented that PLWH indicate that they would welcome education or smoking cessation information from their provider, but that the topic is rarely addressed by providers [68]. Clinic-level interventions have the potential to help address these issues; some of those reviewed showed high feasibility and efficacy in their results, but there were few examples of these types of interventions. Thus, more studies are needed to identify effective models for clinic-level interventions to increase smoking cessation in an HIV care setting.

Conclusions

Cigarette smoking is a modifiable factor that poses great health risk to PLWH. This review identified a need for more robust studies focused on specific vulnerable groups to explore various strategies and theoretical frameworks for smoking cessation among PLWH. In particular, more smoking cessation interventions for PLWH are necessary in the South and for historically marginalized and underrepresented groups such as PLWH of color and PLWH who are sexual and gender minorities. Pilot or quasi-experimental studies have been conducted for PLWH, but more rigorous studies are necessary to better assess intervention effectiveness and increase generalizability of findings. Additionally, standardized measures for cessation must be identified to better compare and gauge the relative efficacy of various interventions across different populations. Intervention types that merit further development and research are those that utilize the SBIRT model, operate at the clinic level, and utilize web-based strategies.