Our goal was to develop a program that was feasible and acceptable to patients, clinicians, and hospital leadership. We therefore convened an advisory stakeholder panel that included a recently hospitalized patient with tobacco dependence, clinicians, and Boston Medical Center (BMC) hospital leaders (VP of Ambulatory Operations and Professional Services and Sr. Manager of ACO Operations). Clinicians included an outpatient primary care physician-investigator with expertise in developing and implementing a financial incentive program at our safety-net hospital primary care practice (KL); a pulmonologist who directs the tobacco treatment center (HK); and tobacco-trained specialists comprised of nurse practitioners (AGF, CW), a respiratory therapist (CO), and a community health worker (JH). In the 6 months prior to implementing the financial incentive intervention, we met with stakeholders to inform intervention development.

After reviewing data with the advisory panel on the high smoking rates of 26% at BMC, demographics (largely Medicaid-insured), low engagement with postdischarge tobacco treatment [19,20], and the effectiveness of financial incentives in promoting smoking cessation [21], we concluded that adults hospitalized with tobacco dependence, particularly if low-income, could potentially benefit from such a program. BMC hospital leaders on the advisory panel recommended focusing the intervention on the 1-month postdischarge period and relayed that the total financial incentives that health care plans would likely find feasible to implement is US $50 per patient hospitalization. Coupled with limited funding for this pilot study, our research team made a pragmatic choice to develop and implement the financial incentive intervention in the 1-month postdischarge period.

At BMC, the standard of care is to offer treatment to all individuals who smoke, regardless of readiness to quit [22]. Since individuals participating in the financial incentives intervention would be in various stages of readiness to quit, we agreed that it was essential to include financial incentives tailored to an individual’s preferences for contingent (rewards based on achieving abstinence), noncontingent (rewards for participation, regardless of achieving abstinence), or combination (rewards for participation and enhanced for achieving abstinence) rewards.

We collaborated with Vincere Health Inc. to tailor their mobile application that uses facial recognition features, a portable breath test CO monitor, and smartphone technology to instantly display and deliver financial incentives to a participant’s digital wallet after completing each CO test. At the end of the 30-day period, participants could redeem their incentives from their digital wallet through the app by choosing from 4 options (Venmo, CashApp, digital check, or digital gift card).

Individuals from Vincere Health Inc. (HJ, JK, SdS, and TC) programmed the app to encourage individuals to conduct 2 CO breath tests daily for 30 days by sending reminder SMS text messages before each scheduled test. Twice daily testing made it less plausible that participants were smoking between assessments since the half-life of CO is about 4.5 hours [23].

Incentives distributed to the digital wallet were dependent on the track chosen and the number of tests performed (maximum of 60 tests in a 30-day period). Total potential incentives earned centered around US $50, the amount deemed feasible by BMC leadership. Slightly higher incentives could be earned for demonstrating CO levels <10 parts per million (ppm); according to the Society on Nicotine and Tobacco Subcommittee on Biochemical Verification (2002), CO levels ≥8 to 10 ppm suggest recent cigarette smoking [24]. Of note, an update was published in late 2020 suggesting that an appropriate cutpoint may be 4 to 10 ppm for research and clinical purposes depending on local smoke-free legislation, smoking prevalence, and air pollution levels [25]. For this study, we chose <10 ppm because secondhand smoke exposure is higher among people with low socioeconomic status and because air pollution levels are higher in communities served by BMC.

The tracks were as follows:

Individuals were compensated up to US $50 for participation in surveys and interviews: US $10 for completing the baseline survey or interview, US $15 for completing a 2-week follow-up survey or interview, and US $25 for completing a 4-week follow-up survey or interview. Participants could additionally earn up to US $60 in financial incentive payments. Vincere Health, Inc. donated disposable CO monitors, loaner phones, and funds to cover the financial incentive payments. Research funds covered the costs of gift cards for participation in surveys and interviews. All data were linked to patients by a unique study identification number unrelated to any Health Insurance Portability and Accountability Act identifiers. Data were stored on a secure server to which only designated individuals had access, thus providing a secure environment for all project data. The Boston University Medical Center Institutional Review Board approved this study (H-39365). All procedures performed in studies involving human participants were in accordance with ethical standards of the institutional review board. Informed consent was obtained from all individual participants included in the study.

We pilot-tested the program from September to November 2020 with a convenience sample of 33 individuals who smoked cigarettes and were hospitalized at BMC, the largest safety-net hospital in New England. Although we initially planned to recruit participants starting in July 2020, based on BMC guidance on conducting research studies during the COVID-19 pandemic, we limited recruitment from September to November 2020. Participants were identified from a list of hospitalized individuals who triggered consultation to the Tobacco Treatment Consult (TTC) service based on current smoking status in the electronic health record (EHR) (20). Eligible participants were (1) ≥18 years old, (2) hospitalized at BMC, (3) able to speak and read English, (4) currently smoking cigarettes (defined as smoking at least 1 cigarette up until the day of hospital admission), and (5) able to provide informed consent. Participants were excluded if cognitively impaired or diagnosed with COVID-19. At the time of enrollment, the mobile app was only compatible with an Android operating system (OS); loaner phones were provided to individuals who did not have a mobile phone or did not have an Android OS at the time of hospitalization. Of note, by the time of publication of this study, Vincere Health, Inc. created this app for use in Apple OS (Apple iOS).

A total of 81 individuals met screening criteria by EHR review (individuals listed as “current” for smoking status). Of the indviduals, 41% (n=33) agreed to participate; provided informed consent; chose their preference for tracks 1, 2, or 3; and enrolled. Figure 1 shows the CONSORT (Consolidated Standards of Reporting Trials) flow diagram. Per standard of care at BMC, all patients were provided bedside tobacco treatment counseling and medication recommendations, postdischarge medication recommendations to the primary inpatient team, and referral to an internal tobacco treatment clinic and/or state quitline at discharge [20].

During hospitalization, the study team gave participants the CO monitor and taught them how to use the device and the mobile app. We provided loaner smartphones to 12 participants. The onboarding process consisted of having the patient download the app, select their track choice (track 1, 2, or 3), and enter their unique 3-digit Smokerlyzer CO monitor PIN number. Prior to hospital discharge, patients were instructed to conduct their first breath test by following on-screen instructions to ensure that everything was working. Videos such as “How to do a Breath Test” were available to patients and the study team to help in the process, now updated to highlight newer features including the ability to connect wirelessly to smartphones [11]. Study staff administered baseline questionnaires in-person at the time of hospitalization and conducted 2- and 4-week questionnaires and interviews by telephone.

At the time of enrollment, we collected baseline demographics and smoking characteristics. Engagement was assessed by how frequently participants conducted CO tests. Program satisfaction was measured at 2- and 4-week questionnaires, thus limiting recall bias. Overall program satisfaction was rated on a scale of 1 to 7 (range: 1=not satisfied at all to 7=very much satisfied). The 2- and 4-week questionnaires additionally assessed the perceived impact of the program on motivation to stop smoking.

We qualitatively measured feasibility and acceptability at both 2 and 4 weeks through semistructured interviews. The interviews assessed participants’ (1) perceived impact of the intervention in motivating smoking cessation, (2) experiences with the program, and (3) suggestions for improvement. Interviews were audio-recorded and transcribed verbatim.

Descriptive statistics were calculated using the SPSS v18 (IBM Corp) and R (R Foundation for Statistical Computing) statistical programming software. For qualitative interviews, we used inductive content analysis to analyze transcripts and performed unstructured coding of transcripts to identify themes. Three members (SS, RVC, and HK) developed a codebook and independently reviewed all transcripts and added codes until the team reached consensus. We finalized conceptual categories, grouped themes in each category, and identified quotes best highlighting themes.

The median age of participants was 46 (range 30-74) years. Participants were 42% (14/33) female and 55% (18/33) non-Hispanic Black American (Table 1). Individuals smoked 6.8 (SD 4.7) cigarettes daily.

Of the 33 participants, 15% (n=5) enrolled in track 1, 45% (n=15) in track 2, and 27% (n=13) in track 3 (Figure 1). Overall, 61% (n=20) of participants engaged with the program, as measured by conducting at least 1 CO test each week. The program was discontinued by 30% (n=10) of participants within 24 hours of hospital discharge. On the first day after hospital discharge, the mean CO levels were 11.9 (range 2.6-36.1) ppm. The average incentive earned per participant was US $14.19. Seven patients had consecutive CO levels <10 ppm during the last 7 days of the program; 4 of these patients were enrolled in track 3. Table 2 shows data by enrolled track.

Of the 33 participants, 61% (n=20) completed either the 2- or 4-week questionnaires. Responses were similar for 2- and 4-week data; for individuals who completed both 2- and 4-week questionnaires, we only analyzed 4-week responses (Figure 1). We assessed end-of-program smoking behavior at 4 weeks by self-report (n=15). Of these participants, 87% (n=13) reported stopping smoking or smoking less. The 3 participants who reported stopping smoking were in track 3.

Of 20 participants, 90% (n=18) of participants were (1) very likely or likely to recommend the program to others and (2) strongly agreed or agreed with the statement “the program is making me think about quitting smoking,” and 95% (n=19) rated the program very highly or highly. All participants wanted to extend the program beyond 30 days. Table 3 details responses on satisfaction and likability of the program, stratified by enrolled track.

Of the 33 participants, 61% (n=20) participated in the 2- and/or 4-week interviews; we analyzed both 2- and 4-week interviews (Figure 1). Supporting quotes are identified by patient number, enrolled track, and interview week.

Financial incentives paired to measurements of exhaled CO concentration levels is a novel smartphone-based tobacco cessation approach that is feasible and acceptable. Future studies should examine the efficacy of the intervention after it is refined to add a counseling and/or SMS text messaging component.