Author + information
- Received December 22, 2015
- Revision received March 30, 2016
- Accepted April 10, 2016
- Published online August 23, 2016.
- Jean-François Etter, PhD∗ ( and )
- Felicia Schmid, MA
- ↵∗Reprint requests and correspondence:
Dr. Jean-François Etter, Institute of Global Health, University of Geneva, CMU, 1 rue Michel-Servet, CH-1211 Geneva 4, Switzerland.
Background It is not known whether large financial incentives enhance long-term smoking cessation rates outside clinical or workplace settings.
Objectives The goal of this study was to test whether large financial incentives improved long-term smoking cessation rates in low-income smokers, in a general population setting, without face-to-face or telephone counseling.
Methods This was a 2-arm, parallel group, individually randomized controlled trial, with follow-up after 3, 6, and 18 months. Participants were 805 low-income smokers enrolled between 2011 and 2013 from the general population in Geneva, Switzerland. We randomly assigned participants to receive either booklets plus access to a smoking cessation website (control group, n = 404), or the same intervention plus financial incentives (intervention group, n = 401). Incremental financial rewards, to a maximum of U.S. $1,650, were offered for biochemically verified abstinence at 1, 2, and 3 weeks, and 1, 3, and 6 months. No in-person counseling, telephone counseling, or medications were provided. The primary outcome was continuous abstinence between 6 months (end of incentives) and 18 months (12 months after the incentives ended), verified by expired carbon monoxide and salivary cotinine. We also assessed biochemically verified 7-day abstinence at 3, 6, and 18 months.
Results Rates of continuous abstinence between months 6 and 18 were 9.5% in the incentive group and 3.7% in the control group (p = 0.001). Rates of 7-day abstinence were higher in the incentive group than in the control group at 3 (54.9% vs. 11.9%; p < 0.001), 6 (44.6% vs. 11.1%; p < 0.001), and 18 months (18.2% vs. 11.4%; p = 0.006).
Conclusions In low-income smokers who did not receive face-to-face or telephone smoking cessation counseling, large financial incentives increased long-term rates of smoking cessation. (Financial incentives for smoking cessation in low-income smokers; ISRCTN04019434).
The burden of smoking-related disease is borne disproportionately by the least affluent people because of the high prevalence of smoking in this group (1). To reduce smoking-related health disparities, it is important to design smoking cessation interventions that reach and are effective in low-income smokers. As this population may be harder to reach with traditional information and education interventions (2,3), other approaches need to be explored, particularly those that address the financial stress in this group.
Financial incentives are effective for smoking cessation while they are in place, but it is not clear whether they have a sustained effect, possibly because most previous studies used small incentives and did not assess long-term outcomes (4). To be effective, the value of financial incentives should be high enough to compensate for tobacco withdrawal symptoms and for the loss of a valued activity. In substance abusers, there is a dose-response association between the value of incentives and their effects on abstinence (5). Two studies of large financial incentives in smokers showed that in educated and relatively affluent employees of large U.S. companies, financial incentives of $750 and $800 increased smoking cessation rates, and that an effect on biochemically verified abstinence was maintained 6 months after the final distribution of incentives (6,7). Thus, large incentives appear to elicit long-term smoking abstinence, but it is not clear whether these findings apply outside workplace or clinical settings, or to other populations, or whether the effect extends beyond 6 months after the final distribution of incentives (4,6,7).
Therefore, the objective of this study was to test whether large financial incentives improved long-term smoking cessation rates in low-income smokers in a general population setting, without face-to-face or telephone counseling for smoking cessation. As there is a risk that the effects of incentives may disappear after the incentives end (4), our study addressed this important point by assessing outcome 12 months after receipt of the final incentives.
Participants were 805 smokers enrolled in Geneva, Switzerland, between August 2011 and May 2013. The financial incentives study was advertised via the press; on the Internet; in workplaces, hospitals, pharmacies, and medical and dental clinics; and by email. After answering the baseline questionnaire online, participants visited our research unit, where eligibility was assessed. Inclusion criteria included: ≥18 years of age; smokes every day; smokes ≥5 cigarettes per day; has smoked for ≥1 year; expired carbon monoxide ≥10 ppm; saliva cotinine ≥10 ng/ml (NicAlert reading ≥1) (8); sets a quit date within 1 month and commits to quit at that date by signing a quitter contract; commits to take part in all follow-up procedures; and taxable income ≤50,000 Swiss francs (CHF) ($55,000, single) or CHF ≤100,000 ($110,000, married), proven by the most recent tax assessment. These income limits correspond to the least affluent one-third of households in Geneva (9).
This study was a single-center, unblinded, 2-arm, parallel group, individually randomized controlled superiority trial with follow-up after 3, 6, and 18 months. We compared an intervention group that received financial incentives plus Internet-based support to a control group that received Internet-based support, but no financial incentives. All participants signed a paper consent form during the enrollment visit. The study was approved by the ethics committee of Geneva University Hospitals, and registered in Current Controlled Trials (ISRCTN04019434). The study protocol was previously published (10) (Online Appendix).
Randomization was performed using sealed opaque envelopes drawn by participants. Neither the researchers nor the participants could know in advance the content of the envelopes. We did not use blocks for randomization. Participants could not be blinded to their assignment group. Researchers were not blinded, but online data collection at follow-up was computerized. Follow-up data was collected by postal mail from nonrespondents to the online surveys, and for nonrespondents to the postal questionnaires, a minimal set of questions on smoking behavior was asked over the phone.
In the incentive group, financial rewards of up to CHF 1,500 ($1,650 in 2013) were paid to those participants biochemically verified as abstinent. All biochemical tests were performed in person at our research unit. As delayed rewards are usually discounted, and as reinforcement works best when the target behavior is followed without delay by the reinforcer (5,11), self-reported quitters with negative tests of both expired carbon monoxide (0 to 3 ppm) and cotinine (<10 ng/ml; i.e., NicAlert = 0) were declared confirmed quitters, and received their reward immediately. Carbon monoxide tests were performed with a Micro Smokerlyzer (Bedfont, Maidstone, United Kingdom), and salivary cotinine was assessed with NicAlert (Nymox, Hasbrouck Heights, New Jersey) test strips, which can detect cotinine levels >10 ng/ml (8). Escalating rewards appear to produce better results than constant rewards (12). Because we wanted to reward sustained abstinence, rather than initial quit attempts, we used an escalating scheme and gave incentives 6 times during 6 months: CHF 100, 150, 200, 300, 350, and 400 at 1, 2, and 3 weeks, and at 1, 3, and 6 months, respectively ($110, $165, $220, $330, $385, and $440, respectively). If participants smoked or missed an assessment, the value of the next reward was reset to the value of the previous reward they had received. Incentives were not paid in cash, but in Migros gift cards, because Migros, a large supermarket chain, does not sell tobacco or alcohol.
Behavioral intervention for both study groups
Participants in both study groups were encouraged to use the Stop-tabac.ch smoking cessation website, which offers fact sheets, discussion forums, testimonials, and an interactive “coach” that provides automatically written, personalized feedback reports (2). This website has proven to be effective (2). Baseline and follow-up questionnaires were collected via Stop-tabac.ch; thus, all participants visited this website repeatedly. During the enrollment visit, all participants also signed a contract in which they committed to quit on a target quit date set no later than 1 month after enrollment; this contract was countersigned by the research assistant and by the optional social supporter. Participants who agreed designated a social supporter of their choice (family or friend), preferably a nonsmoker. The supporters committed to help and support participants during their quit attempts, and were invited to send back a consent form to us. Finally, all participants received self-help smoking cessation booklets (2,13). Research assistants who had not received any training in smoking cessation support delivered the behavioral intervention to both study groups only during the enrollment visit, before randomization. The follow-up visits included only biochemical tests (in both groups) and the distribution of rewards (in the incentives group). We did not provide in-person smoking cessation counseling, telephone counseling, or stop-smoking medications. In Switzerland, clinical services and medications for smoking cessation were not available free of charge at the time of our study. The intervention was implemented as described in the published protocol (10).
Control group procedures
After randomization, participants in the control group were informed about their group assignment, and they were contacted again only for the follow-up surveys, and for biochemical tests in quitters.
Assessment of outcome variables
Existing data indicate that the relapse rate is usually relatively small after 6 to 12 months of abstinence (14,15). We therefore ended the intervention after 6 months and conducted the final follow-up 12 months after the final distribution of incentives. Participants automatically received an email message inviting them to take part in the online follow-up surveys at 3, 6, and 18 months after their target quit date. After 6 email reminders, nonrespondents received the follow-up questionnaires by post; after 3 reminders by post, nonrespondents were contacted by phone (16). Tests of cotinine in saliva and expired carbon monoxide were conducted at baseline in all participants to verify that they were smokers, and at each follow-up (3, 6, and 18 months) in participants who declared that they had not smoked even a puff of tobacco in the previous 7 days. When it was not clear whether the NicAlert result was 0 or 1, salivary cotinine was assessed using gas chromatography (performed by ABS Laboratories, Welwyn Garden City, United Kingdom). In self-reported nonsmokers who reported use of nicotine medications at follow-up and had a negative carbon monoxide test (0 to 3 ppm) but a NicAlert test ≥1 (that is, cotinine >10 ng/ml) (17,18), salivary thiocyanate was used to verify smoking abstinence (cutoff: 100 mg/ml) (19,20). Thiocyanate tests were performed at the University of Greifswald (Germany). Participants in both study groups received a Migros gift card of CHF 25 ($27) for providing saliva and breath samples at follow-up, but not at baseline, and not when another financial incentive was given on the same visit.
Assessment of efficacy
The primary outcome was continuous smoking abstinence between 6 and 18 months, defined as self-report of no puff of tobacco in the previous 7 days at 6 months plus self-report of no smoking between the 6- and 18-month surveys, verified by carbon monoxide and either cotinine or thiocyanate measurements at both time points (6 and 18 months), as described in the preceding text. Secondary outcomes were self-reports of not smoking even a puff of tobacco in the previous 7 days and 4 weeks after 3, 6, and 18 months (point prevalence abstinence), verified by biochemical tests. Carbon monoxide can verify abstinence in the previous 24 h and cotinine in the previous week (21). Thus, our biochemical tests could only verify abstinence during the 7 days before each assessment (3, 6, and 18 months).
All randomized participants were included in the denominator and participants with missing data at follow-up were counted as smokers. We used chi-square tests and odds ratios (ORs) with 95% confidence intervals (CIs) to compare proportions of abstinent smokers. The p values were 2-sided, and were not adjusted for multiple comparisons. Despite randomization, there was imbalance between groups for the distributions of sex and previous attempts to quit smoking (Table 1). Thus, we used multivariate logistic regression models to adjust for sex and previous quit attempts. We used logistic regression models with interaction terms to assess interactions between the incentives and the personal characteristics of participants. We also conducted a complete case analysis including only participants for whom we had complete data. On the basis of previous research (4–6), we determined a priori that a sample of 800 participants would enable us to detect a difference between quit rates of 10% in the control group and 17% in the incentive group (OR: 1.84; power = 80%; p = 0.05) (10).
Participants were not involved in the development, design, or conduct of the study.
We enrolled 805 smokers, 401 of whom were randomly assigned to the incentive group and 404 to the control group (Figure 1). The mean age of participants was 32 years, and many were students. They smoked a mean of 16 cigarettes per day, and were moderately dependent (mean score of 4.0 on the Fagerström Test for Nicotine Dependence) (22) (Table 1).
Follow-up and biochemical confirmation
Of the 805 participants, 743 (92%) were followed up at 3 months, 742 (92%) at 6 months, and 724 (90%) at 18 months. More participants were lost to follow-up in the control group than in the incentive group at 3, 6, and 18 months (Figure 1). Among participants who said they had not smoked in the past 7 days, rates of biochemical confirmation of abstinence were similar in the intervention and control groups at 18 months (63% in both groups; p = 0.99), but confirmation rates were higher in the incentive group at 3 months (82% vs. 67%; p = 0.004) and at 6 months (83% vs. 68%; p = 0.010) (Online Table 1).
Use of incentives and support
In the incentive group, participants received financial rewards of CHF 1,028 on average ($1,130 ± 436); 87% of participants received at least 1 reward and 48% received all 6 rewards. At 6 months, more participants in the incentive group than in the control group reported having obtained help to quit smoking on the Internet over the past 6 months (20% vs. 14%; p = 0.03), and having read a smoking cessation booklet (64% vs. 55%; p = 0.01). Supporters for 8% of participants in the incentive group and 3% of the control group (p = 0.005) returned consent forms.
Efficacy of the incentives
The primary outcome (no smoking in the previous 12 months at 18 months plus no smoking in the previous 7 days at 6 months, biochemically verified at both time points) was 9.48% in the incentive group and 3.71% in the control group (OR: 2.72; difference 5.76%; 95% CI: 2.35% to 9.17%; p = 0.001). Rates of 7-day and 4-week smoking abstinence were higher in the incentive group than in the control group at 3, 6, and 18 months (Table 2). These effects remained statistically significant and of similar size after adjustment for sex and past quit attempts (Online Table 2). In subgroup analyses, none of the interaction effects was statistically significant (Table 3). At 6 months, 82% of participants in the incentive group reported having made a serious attempt to quit smoking after entering the study, compared with 63% in the control group (p < 0.001).
Relapse and cessation between surveys
For the analysis of relapse and cessation between surveys, we used the criterion of 4-week biochemically verified abstinence and counted nonrespondents as smokers. Among those abstinent at 3 months, rates of relapse to smoking at 6 months were similar in the incentive (61 of 204, 30%) and control (16 of 39, 41%; p = 0.17) groups. Among those abstinent at 6 months, rates of relapse to smoking at 18 months were also similar (incentive: 108 of 165, 66%; control: 22 of 39, 56%; p = 0.29). Among those smoking at 3 months, rates of smoking cessation at 6 months were higher in the incentive group (22 of 197, 11%) than in the control group (16 of 365, 4%; p = 0.002). Among those smoking at 6 months, rates of smoking cessation at 18 months were similar in the incentive (9 of 236, 4%) and control groups (21 of 365, 6%; p = 0.29).
In the complete case analysis, the primary outcome was 10.4% in the incentive group and 4.4% in the control group (OR: 2.50; p = 0.003) (Online Table 3).
A recent Cochrane review concluded that financial incentives boost smoking cessation only while they remain in place (4). This may be because many previous trials gave small financial incentives and did not assess the long-term effects. The only 2 previous trials that achieved biochemically verified, sustained abstinence several months after the final distribution of incentives were conducted at U.S. work sites in educated and relatively affluent workers (6,7), in a context where smoking cessation counseling was otherwise available free or reimbursed, and used large incentives ($750 and $800). The Cochrane review concluded that this “approach may only be feasible where independently funded smoking cessation programs are already available, and within a relatively affluent and educated population” (4). Our study supports the idea that large financial incentives produce sustained results, but it differs from previous reports in that it showed that this effect was also observed outside workplace or clinical settings, in a context where personal counseling was not provided to participants and was not available for free elsewhere, and that the effect was sustained 12 (rather than 6) months after the final distribution of incentives, in relatively low-income smokers (Central Illustration).
The odds ratios and quit rates at final follow-up were similar in our study (OR: 2.72; between-group difference in quit rates: 5.76%) and in the 2 previous studies of financial incentives with biochemically verified long-term outcomes (OR at final follow-up: 2.48 and 3.28; between-group differences in quit rates: 4.0% and 5.8%) (4,6,7). The congruence between these 3 studies reinforces the idea that large incentives produce sustained effects, even though the studies differed. Compared with these 2 previous trials, our trial was conducted in the general population, rather than in workplace settings; we enrolled low-income smokers only, behavioral counseling was not provided and was not available free of charge elsewhere; we assessed outcomes at 12 (rather than 6) months after the final distribution of incentives; the value of incentives was twice as high in our trial; smoking status was biochemically verified at entry in the study in all participants; we gave incentives 6 instead of 3 times; we incentivized only smoking abstinence, and not participation in a smoking cessation program (6); we gave incremental rather than fixed incentives (7); and our dropout rates were lower.
The effect observed in our study (a between-group difference of 5.76 percentage points after 18 months; number needed to treat: 17.4) is similar to the 12-month effects of nicotine gum (6% above placebo), bupropion (5% above placebo), or intensive smoking cessation interventions by physicians (5% above usual care) (4,23–25). Studies are needed that directly compare the effectiveness and cost-effectiveness of financial incentives, medications, and medical interventions. In our study, among those smoking at 3 months, rates of smoking cessation at 6 months were higher in the incentive group than in the control group, which suggests that the 6-month reward was useful. Relapse rates between 6 and 18 months were high, but similar in both study groups (56% and 66%), which suggests that relapse at 18 months was not due to the discontinuation of incentives. In the 2 previous long-term studies of financial incentives, relapse rates were also high (36% to 63% after discontinuation of incentives), and these rates were also similar in the incentive and control groups (6,7). With stop-smoking medications, such as nicotine replacement therapy, bupropion, or varenicline, about half of the smokers abstinent at treatment end also relapse after a few months (26–28). Further studies should test how to increase the long-term effects of incentives, and could reward not only abstinence, but also participation in face-to-face counseling and use of nicotine medications, among others (29). Further studies should also assess whether this intervention is effective in other groups and places, whether it has an impact on health outcomes, and its cost effectiveness.
Our findings may have limited generalizability because the participants lived in a relatively small urban area and many were students (but the outcome was similar in students and nonstudents). Compared with a representative sample of smokers in Geneva, smokers in this study were younger, more educated, more motivated to quit smoking (per inclusion criteria), and smoked more cigarettes per day (30). Moreover, although we enrolled participants with the lowest third of earnings in Geneva, this was still a relatively affluent population in international comparison. Also, it may not be feasible to use large financial incentives everywhere. It is unlikely that many participants started smoking to enroll in the study, but we cannot exclude the possibility that some smokers cheated (29,31,32). However, participants needed to abstain from smoking for a week before the biochemical tests in order to get negative cotinine values and obtain the rewards, and the co-occurrence of incentive distribution and outcome assessment did not occur at 18 months, when no incentives were given. It is, however, possible that participants in the incentives group learned how to abstain from smoking before assessments during the incentive period, and transferred this to the latter assessment point. Also, control participants were aware that they were not receiving financial rewards, which may have introduced a disincentive in this group. Finally, the biochemical tests themselves, even without counseling, may have had some impact, and control participants were subjected to fewer biochemical tests. Incidentally, the quit rate observed in our control group was within the range of quit rates reported in reviews of studies of unaided smokers (33,34).
This study also has several strengths, including the longer duration of follow-up after receipt of the final incentives (12 months) and the larger value of incentives than in any previous study (4), the large number of participants, the very high follow-up rates, the biochemical testing at 18 months, the assurance that all participants were truly smokers at baseline, and the general population setting (rather than clinical or workplace settings). Our criteria for biochemical validation were more stringent than in most previous studies because we used carbon monoxide and thiocyanate in addition to cotinine for the outcome assessment, and cotinine in addition to carbon monoxide for the incentive distribution (4). Finally, this intervention can be easily implemented elsewhere, without medical staff.
This study showed that large financial incentives increased rates of continuous smoking abstinence in relatively low-income smokers, and that an effect was maintained 12 months after receipt of the final incentive, in a context where personal counseling for smoking cessation was not provided to participants and was not available for free elsewhere. The results were maintained in a complete case analysis.
COMPETENCY IN PATIENT CARE: Substantial financial incentives have a sustained impact on smoking cessation among relatively low-income smokers, both inside and outside of workplace or clinical settings, even when personal counseling is not provided. The effect is sustained 12 months after the final distribution of incentives.
TRANSLATIONAL OUTLOOK: Further studies are needed to assess the net health care costs and savings that accrue from programs that offer large financial incentives for smoking cessation across a wide array of socioeconomic groups and practice settings.
The Swiss Tobacco Prevention Fund, which funded the study, suggested that the follow-up should be extended to 12 instead of 6 months after the final incentives were received, but had no other role in the conduct of the study.
Vincent Baujard, from the Health on the Net Foundation, developed the software for data collection. Dr. Harald Below, at the University of Greifswald (Germany), performed the thiocyanate analyses.
For an expanded Methods section and supplemental tables, please see the online version of this article.
The study was funded by the Swiss Tobacco Prevention Fund (Swiss Federal Office of Public Health), grant 11.001733. Dr. Etter’s salary was paid by the University of Geneva. The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- Swiss francs
- confidence interval
- odds ratio
- Received December 22, 2015.
- Revision received March 30, 2016.
- Accepted April 10, 2016.
- American College of Cardiology Foundation
- Townsend J.,
- Roderick P.,
- Cooper J.
- Cahill K.,
- Hartmann-Boyce J.,
- Perera R.
- Cooke F.,
- Bullen C.,
- Whittaker R.,
- et al.
- Office Cantonal de la Statistique (OCSTAT)
- Etter J.F.,
- Stapleton J.A.
- Dillman D.A.
- Etter J.F.,
- Vu Duc T.,
- Perneger T.V.
- Hughes J.R.,
- Stead L.F.,
- Hartmann-Boyce J.,
- et al.
- Stead L.F.,
- Buitrago D.,
- Preciado N.,
- et al.
- Jorenby D.E.,
- Hays J.T.,
- Rigotti N.A.,
- et al.,
- Varenicline Phase 3 Study Group
- Tappin D.,
- Bauld L.,
- Purves D.,
- et al.,
- Cessation in Pregnancy Incentives Trial Team