An EHR-automated and theory-based population health management intervention for smoking cessation in diverse low-income patients of safety-net health centers: a pilot randomized controlled trial

Abstract

This study tested the preliminary effectiveness of an electronic health record (EHR)-automated population health management (PHM) intervention for smoking cessation among adult patients of a federally qualified health center in Chicago. Participants (N = 190; 64.7% women, 82.1% African American/Black, 8.4% Hispanic/Latino) were self-identified as smokers, as documented in the EHR, who completed the baseline survey of a longitudinal “needs assessment of health behaviors to strengthen health programs and services.” Four weeks later, participants were randomly assigned to the PHM intervention (N = 97) or enhanced usual care (EUC; N = 93). PHM participants were mailed a single-page self-determination theory (SDT)-informed letter that encouraged smoking cessation or reduction as an initial step. The letter also addressed low health literacy and low income. PHM participants also received automated text messages on days 1, 5, 8, 11, and 20 after the mailed letter. Two weeks after mailing, participants were called by the Illinois Tobacco Quitline. EUC participants were e-referred following a usual practice. Participants reached by the quitline were offered behavioral counseling and nicotine replacement therapy. Outcome assessments were conducted at weeks 6, 14, and 28 after the mailed letter. Primary outcomes were treatment engagement, utilization, and self-reported smoking cessation. In the PHM arm, 25.8% of participants engaged in treatment, 21.6% used treatment, and 16.3% were abstinent at 28 weeks. This contrasts with no quitline engagement among EUC participants, and a 6.4% abstinence rate. A PHM approach that can reach all patients who smoke and address unique barriers for low-income individuals may be a critical supplement to clinic-based care.

Implications.

Practice: A population health management approach comprising clinical systems- and targeted individual-level intervention components may be an effective strategy for improving quitline access and engagement for socioeconomically disadvantaged individuals.

Policy: State telephone quitlines should devote more resources to establish low-cost and sustainable electronic and clinical linkages with safety-net health centers.

Research: Population-based intervention strategies automated via the EHR system and utilizing a treatment opt-out approach may be a critical supplement to clinic-based care in community health center settings.

Introduction

Persons from low-income households continue to be more likely to smoke and have limited access to effective smoking cessation treatment [1]. Persistent barriers to addressing smoking during primary care visits involve clinical systems, clinicians, and patient factors [2]. Electronic health record (EHR) systems and other system changes have been used to improve smoking status documentation and to prompt assistance during visits [3, 4] or outside of visits [5]. These changes have improved care, but more than one-third of patients seen in primary care are still not offered smoking cessation treatment [6, 7].

Individuals living in poverty face extreme barriers to equitable health care, including limited health literacy skills and social and economic determinants of health [8] and limited access to effective and affordable preventive services. These inequalities contribute to their disproportionate health burden [9]. While other marginalized populations encounter these same challenges, they occur much more frequently and with greater impact among poor racial/ethnic minorities [10]. In 2010, 38.6% of persons below the federal poverty line did not see a primary care physician, compared with 18.5% of higher-income persons [11]. Even when low-income persons receive health services, they receive fewer preventive services, including evidence-based smoking cessation treatment [12].

State quitlines are effective for reaching large populations and promoting smoking cessation [13, 14]. Because quitline services are free or low-cost, easily accessed, and have the potential to target treatment based on smoker characteristics, they are well positioned to partner with traditional health care settings to address inequities in treatment access and utilization. Many state quitlines, including the Illinois Tobacco Quitline (ITQL), provide reduction-focused treatment for individuals who choose smoking reduction as their initial goal or who do not achieve cessation by their target quit date and decline to set a new quit date [15]. Given the chronic nature of tobacco dependence, integrating state quitlines with community health care networks could advance the long-term management of smoking for underserved populations.

Electronic referral to state quitlines has been evaluated in prior studies of the Ask-Advise-Connect (AAC) and similar e-referral approaches [16–20], including two studies that assessed the effectiveness of the AAC approach in safety-net health centers [19, 21]. Although these studies significantly improved access to smoking cessation treatment, the AAC approach requires trained clinical staff to manage, and it only addresses smoking among patients who schedule and attend a primary care visit and who agree to accept the quitline referral.

Although EHR systems can be linked with quitline systems to enable the referral of patients, individual-level interventions are needed to promote treatment acceptance and engagement. While approximately 90% of smokers are not ready to quit, many are interested in cutting down [6, 22], increasing their likelihood of making a quit attempt and achieving smoking cessation [23–25]. According to self-determination theory (SDT) [26], smokers who believe that they are autonomous partners in their health care, and empowered to select treatment goals that match their level of readiness to quit, will be more interested and engaged in treatment. SDT-based intervention increases treatment utilization and cessation [27] and offering smoking reduction as one of the treatment options may increase treatment engagement [6, 28].

This study evaluated the effectiveness of a person-centered and EHR-automated population health management (PHM) intervention for smoking cessation that was embedded in a community health center primary care setting and designed to engage low-income smokers outside of the primary care visit. The intervention overcame system-, clinician-, and patient-level barriers to addressing smoking in low-income patients. Health care delivery and access barriers were addressed through an EHR system that automated the identification of smokers, the printing of targeted intervention material, referral for proactive quitline treatment, and the return of treatment outcomes to the EHR system to support follow-up care. Health literacy barriers were addressed via letter and text messaging (words, graphics, and style/format) targeted to low-literacy smokers. Psychological barriers (low motivation, belief that abrupt cessation is the only option) were addressed through a person-centered approach via the quitline counselors that encouraged smokers to choose their initial treatment goal. We hypothesized that the ability to choose between cessation and reduction, in the context of other SDT and educational elements, would promote engagement in quitline treatment.

Methods

Study design

The study used a two-group randomized controlled design to evaluate the effectiveness of a PHM intervention (called Choose to Change [CtC]), when compared with enhanced usual care (EUC), for promoting quitline engagement, utilization, and smoking cessation (see Figure 1). Randomization was stratified by racial group (African American/Black vs. other) and level of motivation to quit (low vs. high), as measured by a single item rated from 1 = not at all motivated to 10 = extremely motivated. Higher quit motivation was reflected in ratings ≥6. To measure smoking cessation among participants who did not engage in quitline treatment, all participants in the pilot randomized controlled trial had first enrolled in a 32-week “needs assessment of health behaviors to strengthen health programs and services.” The study was conducted between April 21, 2017 and September 5, 2019 (ClinicalTrials.gov, NCT03077737).

Fig 1.

Study design, intervention timeline, and assessment schedule.

Participants and setting

Participants were adult patients of Near North Health Service Corporation (NNHSC) in Chicago, Illinois. Patients were eligible for the longitudinal observational study, and therefore the pilot trial, if they were ≥18 years old, identified as a “current someday smoker” or “current everyday smoker” in the EHR and had ≥1 provider visits within the past 12 months. Patients documented as preferring a language other than English or Spanish for their health care, who did not have a telephone number or address listed in the EHR or did not have a cell phone with texting capability were excluded. Only one patient per household was eligible to participate.

Bidirectional e-referral system

Supported by Alliance-Chicago, a Health Resources and Services Administration-funded network that provides a health information technology and collaboration ­infrastructure for independent primary care community health centers, the GE Centricity EHR system was used by NNHSC for all clinical encounters. Standard procedure is to confirm patient telephone numbers and residential addresses at every visit. An automated e-referral order, developed and validated as a part of the current study, allowed all physicians to refer a patient to the ITQL electronically.

The order contained specific text information and variables related to the patient and a statement indicating patient authorization to be contacted by the quitline. A signed order fired within the EHR system triggered an outbound referral request via HL7 messaging through the interface engine (Qvera). The outbound e-referral order request was confirmed via an automated review of a valid email address through Qvera, confirming the bidirectional connection between the EHR system and the ITQL. Once this process was completed, the outbound message was sent via an automated secure file transfer procedure to the ITQL and was received in less than 24 hr.

Within 24 hr after each quitline session, a HL7 file was automatically produced that provided a summary update of the services performed and the information documented during the call. Again, Qvera confirmed the email address and connection, and the HL7 message was sent via secure file transfer to Alliance-Chicago where it automatically populated structured EHR fields. A summary document within the EHR was sent simultaneously to the patient’s primary care provider to communicate the referral outcome. All information received from the ITQL was documented within the patient’s EHR, allowing for a longitudinal view of treatment utilization and smoking status. The study allowed 4 months after launching the e-referral system for referral rates to stabilize before launching the pilot clinical trial.

Interventions

PHM intervention

The PHM intervention comprised three components: (i) proactive e-referral to the quitline; (ii) a single-page letter developed through patient focus groups, guided by SDT, and printed via the EHR system; and (iii) five text messages that reinforced the central messaging of the letter. The letter and text messages were developed in English and Spanish and delivered according to the patient’s preferred language for health care as identified in the EHR. The letter was sent on behalf of the director of performance improvement and primary care provider (T.L.) at NNHSC and addressed low motivation by encouraging either smoking cessation or reduction as an initial step to cessation. The letter also provided information about treatment, including the safety and effectiveness of nicotine replacement therapy (NRT), and stated that a quitline counselor would be calling within 2 weeks to offer free treatment. The central message developed through focus groups was “Choose to change and make your own goals.” The letter also contained other elements of SDT designed to increase autonomous motivation (e.g., provided autonomy support through minimizing pressure and control) and standard supportive and motivational statements to encourage behavior change. The letter was designed for smokers with low health literacy. The written message was one page in length to minimize cognitive overload and burden and to maximize the message’s effectiveness [29, 30]. The estimated readability according to the Flesch–Kincaid Grade Level rating was 4.8 (fifth grade). Patients were asked to contact the NNHSC care coordinator if they did not want to be contacted by the quitline. Opt-outs and reasons for not participating were recorded. The letter was printed on NNHSC letterhead, enclosed in 6″ × 9″ NNHSC envelopes, and sent via first-class mail.

Five text messages were sent on days 1, 5, 8, 11, and 20 after letter mailing. The first message was: “From Near North Health Service: Look for a letter in the mail from your doctor about FREE services to help you cut down or quit smoking. You have choices!” The message sent on day 5 directed participants to view the mailed letter via a hyperlink. Text messages 2–5 included a simple opt-out of future messages with varying responses to opting out. For example, a “STOP” after the day 5 message triggered the following text response: “OK, no more texts. A Quitline counselor will be calling you and can help you quit or cut down. Treatment is FREE. You can make your own goals.” Opt-outs were recorded. The text messaging component of the PHM intervention was automated by TouchPointCare, a telehealth system service provider specializing in secure practice–patient communication.

Enhanced usual care

Participants assigned to the control condition were e-referred to the ITQL following standard procedures that were established by NNHSC in 2011. If the patient was seen in the health center for a routine evaluation or an acute problem, a medical assistant confirmed current smoking and advised the participant to quit using clear and strong language [2]. Current smoking and advice to quit were documented in the EHR system by the medical assistant. Next, the participant was seen by the physician who addressed the purpose of the visit. An e-referral was made if the physician also discussed the patient’s smoking, determined that the patient was willing to set a quit date in the next month, and the patient agreed to the referral.

Quitline treatment

The ITQL protocol delivers treatment determined to be effective in the 2008 U.S. Public Health Service Clinical Practice Guideline [2]. All participants who were e-referred and reached by the ITQL were offered either smoking cessation- or reduction-focused treatment in English or Spanish. A self-help packet (Quit-Kit) was mailed to provide detailed information on the topics covered during the initial counseling session. Participants who chose reduction as their initial goal were guided toward setting a quit date. A free 8-week course of NRT (patch, lozenge, or gum) was mailed in 2-week allotments to participants who were willing to set a quit date and were interested in NRT and able to use it safely. Because free NRT increases the use of quitlines [31, 32], NRT was offered at no cost to all participants to eliminate insurance type as a potential confounder of intervention effects. Treatment continued as either accepted or initiated by participants for 28 weeks.

Procedure

We used structured query language (SQL) to analyze EHR data monthly to identify eligible patients for the longitudinal needs assessment of health behaviors to strengthen health programs and services. The primary care physicians of identified candidates were sent a message alert via the EHR system requesting that they review the list of patients identified (provider-specific) and respond within 10 days to identify any patients not appropriate for treatment. Following the 10-day provider opt-out period, patients were contacted by telephone and invited to participate in a 32-week study being conducted at NNHSC. The study was described as a way for NNHSC and the NNHSC providers and staff to better understand the health habits, risk behaviors, and wellness service needs of their patients to strengthen health programs and services. Patients who enrolled in the study completed a baseline assessment of smoking and smoking cessation treatment history conducted by study staff. To disguise the primary purpose of the assessment, diet, alcohol use, physical activity, and sedentary behavior were also assessed during each interview. As a part of the baseline assessment, participant phone numbers and residential addresses were confirmed in the EHR system.

Four weeks after completing the baseline assessment of the longitudinal study, participants were randomized to either the PHM intervention or EUC, and participants assigned to PHM were mailed the “Choose to Change” letter. Two weeks after sending and following a 12-day opt-out period, participants’ contact details were transmitted to the ITQL via the EHR system. The automated quitline e-referral, with physician and patient opt-out procedures, was possible through a Business Associate Agreement involving NNHSC, Alliance-Chicago, and the American Lung Association of the Upper Midwest. The ITQL called participants within 2 days after being referred. Quitline engagement, utilization, and smoking cessation outcomes were transmitted directly from the ITQL server to the EHR system. Participants were reassessed by study staff at weeks 6, 14, and 28 after randomization and the start of the intervention, which was 10, 18, and 32 weeks after the baseline assessment. Assessments were conducted by telephone in English or Spanish. Participants were paid a total of $145 in electronic gift cards for completing all four assessments. Once the study was completed, EUC participants were sent a newsletter that described the pilot clinical trial, along with the reasons for withholding the true study purpose. EUC participants were then offered the PHM intervention and were told about forthcoming text messages, quitline referral, and the opt-out procedure.

Data on smoking cessation for PHM and EUC participants and smoking cessation treatment (non-quitline) utilization for EUC participants were obtained as a part of the longitudinal observational study. Data on quitline engagement and treatment utilization were obtained via the NNHSC EHR system. Only smoking-related data are reported herein.

Outcome measures

Primary study outcomes included quitline engagement at 6 weeks, utilization at 14 weeks, and smoking cessation at 28 weeks. Treatment engagement was defined as accepting the first proactive call from the quitline counselor, enrolling in treatment, and completing the first behavioral counseling session. Quitline treatment utilization was defined as completing one or more additional behavioral counseling sessions by week 14. Smoking cessation was defined as self-reported abstinence for at least 7 days before the week 28 assessment. Participants were classified as smoking if they reported smoking or could not be reached to provide data.

Statistical analysis

T-test and chi-square test compared intervention groups on baseline demographic and smoking-related continuous and categorical variables, respectively. Intervention effects on all binary outcomes of interest (engagement, utilization, smoking cessation) were estimated using odds ratios (ORs). Due to small cell counts, p-values and confidence intervals (CIs) were based on Fisher’s exact test as implemented using the base distribution of the statistical software package R (https://cran.r-project.org).

Results

Participant screening, accrual, and retention in the observational study

Figure 2 illustrates participant flow through the longitudinal observational study by the intervention arm. Of the 1606 patients who were identified via the EHR system and contacted, 387 (24.0%) completed the eligibility screen, and 190 (11.8%) enrolled, completed the baseline assessment, and were randomized in the pilot clinical trial to either the PHM intervention (N = 97) or EUC (N = 93). Of the participants screened, 96.9% (375/387) owned a cell phone with texting capability. Survey completion rates at weeks 6, 14, and 28 after randomization and letter mailing (10, 18, and 32 weeks after the baseline assessment) were comparably high between arms. Due to project timeline constraints, only the first 96 participants were followed for 32 weeks.

Fig 2.

Participant flow in the longitudinal observational study by intervention arm.

Participant characteristics

The sample reflected the diversity of the NNHSC community: 64.7% were women, 82.1% were African American/Black, and 8.4% were Hispanic/Latino (Table 1). Of the 190 participants, 70.5% reported daily smoking, averaging about 9 cigarettes/day. The remaining participants reported non-daily smoking, averaging about 4 cigarettes per day on days smoked. A quarter of participants reported a past year quit attempt and three-quarters reported reducing their smoking during the past year. The level of motivation to quit was moderately high. The PHM intervention and quitline treatment were delivered in Spanish for the seven participants who preferred Spanish for their health care.

Table 1.

Participant characteristics by intervention arm

Characteristic	Enhanced usual care (N = 93)	PHM choose to change (N = 97)	Overall sample (N = 190)
Age, mean (SD)	48.8 (11.9)	49.2 (10.7)	49.0 (11.3)
Sex (% female)	69.9	59.8	64.7
Race (% African American/Black)	80.6	83.5	82.1
Ethnicity (% Hispanic)	8.6	8.2	8.4
Preferred language for health care (% English)	95.7	96.9	96.3
Cigarettes/day daily smokers, mean (SD)	9.5 (5.8) N = 65	8.2 (5.1) N = 69	8.9 (5.5) N = 134
Cigarettes/day non-daily smokers, mean (SD)	5.4 (3.3) N = 28	3.3 (2.2) N = 28	3.9 (2.8) N = 56
Motivation to quit smoking	7.1 (2.9)	7.0 (2.7)	7.1 (2.7)
Past year quit attempt (% yes)	22.6	28.1	25.4
Past year reduced smoking (% yes)	78.5	71.9	75.1

Motivation to quit was rated 1 (not at all motivated) to 10 (extremely motivated).

Effects of intervention on treatment engagement, utilization, and smoking cessation

None of the PHM intervention participants opted out of the proactive quitline call and 8.2% (8/97) opted out of text ­messaging. By week 6, no EUC participants (0/92) and 25.8% (25/97) of PHM participants had engaged in quitline treatment (OR = 0, 95% CI: 0 to 0.13, p < .0001). No additional PHM participants engaged in treatment after week 6. Of the 25 PHM participants who engaged in quitline treatment, 44.0% were women, 80.0% were African American/Black, and 12.0% were Hispanic/Latino. By week 14, no EUC participants (0/92) and 21.6% (21/97) of PHM participants were treatment utilizers (OR = 0, 95% CI: 0 to 0.17, p < .0001).

At week 6, 4.3% (4/93) of EUC participants and 2.1% (2/97) of PHM participants were abstinent (OR = 2.13, 95% CI: 0.30 to 24.05, p = .437). At week 14, 3.2% (3/93) of EUC participants and 9.3% (9/97) of PHM participants were abstinent (OR = 0.33, 95% CI: 0.06 to 1.37, p = .134).

At week 28, the primary endpoint, 6.4% (3/47) of EUC participants, and 16.3% (8/49) of PHM participants were abstinent (OR = 0.35, 95% CI: 0.06 to 1.60, p = .199).

Quitline utilization among PHM intervention participants

Among the 25 PHM participants who engaged in quitline treatment, all received cessation-focused behavioral counseling and 92% (23/25) received NRT. Of the participants who received NRT, 14 used the patch, 5 used gum, 3 used lozenge, and 1 used lozenge and gum. On average, PHM participants completed 4.0 (SD = 2.9) behavioral counseling sessions across 23.0 (SD = 17.0) days. The average duration of NRT use was 3.4 (SD = 2.3) weeks.

Non-quitline smoking cessation treatment reported by EUC participants

During the 32-week observation period, 4 EUC participants reported receiving smoking cessation counseling and 20 participants reported using NRT. Additional smoking cessation treatment reported by EUC participants included: varenicline or bupropion (n = 7); mobile app (n = 1); text messaging program (n = 1); web-based program (n = 5); books, pamphlets, or other written material (n = 10); or alternative therapy, such as acupuncture, meditation, herbal supplements, vitamins, yoga, or massage therapy (n = 5).

Discussion

Improving access to effective, free, or low-cost smoking cessation treatment is critical for low-income smokers who are underserved and who carry a disproportionate burden of tobacco-related disease. This pilot study tested the preliminary efficacy of a population-based intervention for promoting quitline engagement, treatment utilization, and smoking cessation for patients of a large community health center in Chicago. The intervention, called Choose to Change, comprised system-level and individual-level components. Automated using an EHR system, access to free behavioral and pharmacological treatment was enabled through the electronic referral of patients to the ITQL using an opt-out approach [33]. To promote acceptance of proactive quitline treatment, patients were mailed a letter and sent text messages before quitline outreach. The content of the letters and text messages was guided by SDT, developed through a series of focus groups involving patients of the health centers, and designed to address unique barriers for low-income individuals.

An important advantage of the PHM intervention evaluated in this study is that it does not require trained clinical staff to manage due to the integration of a robust EHR system and automated functionality. A second significant advantage is that the intervention does not rely on patients being seen in the health center and consenting to a quitline referral. Study results, which are preliminary and require evaluation in a fully powered randomized trial, indicate that the PHM intervention can achieve population-level reductions in smoking among diverse low-income individuals who receive their health care in community health centers.

The clinic-based approach represented by the EUC condition is comparable to the AAC approach developed by Vidrine et al. [16], which has been shown to engage 7.8%–11.8% of smokers in safety-net health centers [19, 21] and achieve an abstinence rate of 16.6% at 6 months [19]. The main difference was the use of trained, licensed vocational nurses who were responsible for carrying out the AAC steps during patient visits. Before the launch of the e-referral system at NNHSC as a part of this study, 4 months before starting the pilot trial, the study team conducted a health center-wide provider training in the use of the e-referral procedure. Training also covered the simple process to access the quitline treatment progress and outcomes updates in the EHR system. Refresher training was conducted periodically during the study, but more frequent training may have been needed to sustain provider and staff awareness of the e-referral system. The absence of referrals among the EUC participants highlights the potentially critical importance of the use of trained clinic staff, as a part of the AAC approach, who are dedicated to asking, advising, obtaining patient authorization, and making the quitline e-referral.

The PHM intervention was designed to engage individuals who are not ready to quit smoking by promoting quitline treatment engagement and smoking reduction as essential steps to smoking cessation. A potential limitation of this preliminary evaluation was the moderately high baseline level of motivation to quit among participants who enrolled in the 32-week observation study, which established the patient pool for the pilot clinical trial. The intervention arms were well balanced on baseline motivation to quit, but 29% of participants overall were at the ceiling of the single-item assessment of quit motivation. Only 30.6% of participants rated their motivation ≤5 on the 10-point Likert scale. Among the 30 PHM participants with lower baseline motivation, as reflected in ratings ≤5, 13.3% engaged in quitline treatment. This engagement rate is comparable to the engagement rate achieved with the AAC approach implemented during patient visits [19]. Given that e-referral is often conditional upon patient acceptance, it is possible that patients who engaged in those studies had higher levels of motivation to quit. Among the 66 participants in the PHM arm who had a higher level of quit motivation >5 (1 participant was missing data), 31.8% engaged in quitline treatment.

Another potential limitation is that smoking abstinence was based on self-report. However, three features give us confidence in the validity of self-report in this study. First, participants were asked about a range of health behaviors. The stated purpose of the observation study assessments was to obtain opinions about health promotion programs and services needed at the health center. Participants were told that their answers would not be entered into their medical records and only be reported in aggregate form. Thus, the social desirability to report non-smoking was reduced. Second, the assessment of smoking status was interviewer-administered, which yields higher levels of sensitivity and specificity than self-report questionnaires [34]. Third, the waiver of informed consent for the pilot trial resulted in a research context comparable to an observational study. Compared with treatment studies, observational studies yield higher levels of sensitivity and specificity [34].

The average dose of behavioral treatment delivered to participants who engaged in quitline treatment was four sessions. This outcome compares with one session (M = 1.2, SD = 1.4) completed on average for participants in the Pineiro et al.’s [19] study connected to the Texas Quitline. As quitline treatment dose predicts smoking cessation among patients of safety-net health centers [35], an important focus of future research is to identify strategies to increase treatment utilization. Integrating cultural competency training for counselors is one potential strategy to increase continuing quitline engagement and utilization by diverse low-income individuals [36].

The next step in this intervention development work is to evaluate the effectiveness of the PHM intervention using a fully powered pragmatic randomized trial. Given that women comprised 59.8% of the PHM arm, but only 44.0% of participants who engaged in treatment, the interventions letter and text messaging components may need to be refined to highlight counseling content that is especially relevant to women [37]. Another aim should be to examine the effect of quitline outcomes feedback to providers via the EHR system. This “closing the loop” with physicians could enhance clinic-based efforts during patient visits to support quit attempts, smoking cessation, and maintenance of short- and long-term abstinence. Future research should develop and evaluate innovative low-cost strategies to increase the use of closed-loop e-referral systems like the one developed for this pilot clinical trial. As suggested by the absence of referrals in the EUC arm, a validated e-referral system and physician training alone in the safety-net health center setting are likely insufficient to promote quitline referrals. Trained and dedicated patient navigators [19] may not be feasible in many safety-net health centers.

In conclusion, an automated, low-cost PHM approach that can reach outside of safety-net health center visits to patients who smoke and address unique barriers to smoking cessation for low-income individuals could supplement clinic-based care. If found to be effective in a large-scale trial, the PHM intervention could be readily scaled to community health centers and quitlines throughout the USA.

Funding

This study was supported, in part, by the National Institutes of Health’s National Cancer Institute (grants U54CA202995, U54CA202997, and U54CA203000), the National Institutes of Health’s National Center for Advancing Translational Sciences (grant UL1TR001422), and the Agency for Healthcare Research and Quality (grant AHRQ P01HS021141). Additional funding was provided by the Robert H. Lurie Comprehensive Cancer Center of Northwestern University, the University of Illinois Cancer Center, and the North American Quitline Consortium through a grant from Pfizer IGLC (Independent Grants for Learning and Change).

Conflict of Interest: All authors declare that they have no conflicts of interest.

Authors’ Contributions: Dr. Hitsman had full access to the data and takes responsibility for the integrity of the data and the accuracy of the data analysis. Concept and design: Hitsman, Matthews, Papandonatos, Cameron, Mohanty, Long, Ackermann, Ciecierski, Simon. Acquisition of data: Rittner, Mohanty, Ramirez, Carr, Cordova, Bridges, Flowers-Carson, Giachello, Hamilton. Analysis and interpretation of data: Hitsman, Matthews, Papandonatos, Cameron. Drafting of the manuscript: Hitsman, Papandonatos, Ramirez, Cordova. Critical revision of the manuscript for important intellectual content: Hitsman, Matthews, Cameron, Rittner, Mohanty, Long, Carr, Ackermann, Bridges, Ciecierski, Simon. Statistical analysis: Papandonatos. Obtained funding: Hitsman, Matthews, Ciecierski. Administrative, technical, or material support: Hitsman, Matthews, Cameron, Rittner, Ackermann, Ramirez, Carr, Cordova, Bridges, Flowers-Carson, Giachello, Hamilton, Ciecierski, Simon. Supervision: Hitsman, Rittner, Mohanty, Long, Flowers-Carson.

Ethical Approval: All procedures were conducted in accordance with the ethical standards of the Institutional and National Research Committee and with the 1964 Helsinki Declaration and its later amendments. The study was approved by the IRBs of Northwestern University (IRB of Record), the University of Illinois at Chicago, and Northeastern Illinois University.

Informed Consent: Verbal informed consent was obtained from all participants in the longitudinal observational study. A waiver of informed consent was granted for the pilot clinical trial.

Transparency Statement: The study was registered at ClinicalTrials.gov (NCT03077737). The analytic plan was not registered. Deidentified data, statistical analysis code, and materials used to conduct the study are not publicly available. Deidentified data can be requested by contacting the corresponding author; however, data sharing will be contingent on institutional approval.