Positive Psychotherapy for Smoking Cessation Enhanced with Text Messaging: A Randomized Controlled Trial

Abstract

Introduction:

Positive Psychology Interventions (PPIs) aim to enhance well-being and have shown promise in improving mental health and augmenting behavioral interventions. Although pilot studies have suggested PPIs may support smoking cessation, no full-scale randomized clinical trials (RCTs) have tested their efficacy in this context.

Method:

In this RCT, 271 adults who smoked ≥ 5 cigarettes per day (55.2% female, mean age = 49.2 (SD = 12.0); average cigarettes per day = 16.9 [SD = 7.0]) were randomized to six sessions of Positive Psychotherapy for Smoking Cessation with text messaging (PPT-S+) or a time-matched standard treatment with text messaging (ST+). All participants received eight weeks of nicotine replacement therapy (transdermal nicotine patch). The primary outcome was biochemically-confirmed 7-day point-prevalence smoking abstinence at 12, 26, and 52 post quit date.

Results:

Seven-day point-prevalence abstinence did not differ significantly by condition: 20.9%, 19.4%, and 19.4% at 12, 26, and 52 weeks in PPT-S+ compared to 19.1%, 20.6%, and 21.4% at 12, 26, and 52 weeks in ST+. Conditions did not differ in self-reported use of positive psychology-consistent strategies for quitting smoking at posttreatment, representing an implementation failure, and these strategies did not significantly predict subsequent smoking outcomes. Those in ST+ rated the text messaging program as significantly more helpful than those in PPT-S+.

Conclusions:

PPT-S+ did not improve smoking cessation outcomes compared to a time-matched control. The completion of positive psychology exercises through text messaging may have reduced the helpfulness of the text-messaging support program received in both conditions.

INTRODUCTION

Positive psychology interventions (PPIs), which focus on cultivating personal strengths, positive emotions, and positive cognitions,¹ have been found to enhance well-being.² Recently, the effects of PPIs have been tested on a range of health behaviors including physical activity³ and substance use⁴ with preliminary promise. Although mechanisms of PPIs’ effects have not been fully elucidated, theories suggest (a) that experiencing positive affect when engaging in health behaviors may strengthen nonconscious motives for behavior change maintenance⁵ and (b) that positive affect may facilitate building psychological resources that support behavior change and have a stress-buffering effect.^5,6 Thus, PPIs may influence both motivation and self-efficacy for health behavior change.

Positive Psychotherapy for Smoking Cessation (PPT-S) was the first smoking cessation program to integrate PPIs into smoking cessation counseling.⁷ PPT-S utilizes a gain-framed approach^8,9 to highlight that quitting smoking can be a rewarding challenge that enhances mental health.^10–12 It encourages smokers to consider how their signature strengths (e.g., humor or perseverance) may be employed in quitting efforts and includes exercises—completed outside of session and documented on written homework logs—that focus on enhancing nonsmoking means of experiencing positive affect. These exercises are intended to reduce the attraction to smoking that may persist even after a period of abstinence^13,14 and to support self-efficacy for remaining abstinent¹⁵ by building psychological resources to cope with challenges during a quit attempt. In a pilot trial (N = 66), participants randomized to PPT-S were more likely to remain abstinent during a 26-week follow-up than those receiving standard treatment (OR = 2.75, 95% CI = 1.02, 7.42),¹⁶ and this effect was stronger in participants with higher baseline positive affect. Changes in positive and negative affect did not differ by condition, but greater engagement in quit smoking strategies consistent with a positive psychology approach, such as focusing on good things that happened each day during a quit attempt, predicted sustained abstinence. Similarly, a different study, focusing on nondaily smoking and leveraging the smartphone app Smiling instead of Smoking (SiS) to deliver PPIs,¹⁷ showed that greater engagement in PPIs predicted smoking cessation.¹⁸ In a subsequent pilot trial (N = 226), those randomized to SiS had significantly higher self-efficacy to abstain from smoking, higher positive affect, and lower cigarette craving at posttreatment (6 weeks after target quit date) than those randomized to an active control app with effect sizes showing promise for SiS for smoking cessation at 6 weeks.¹⁹ Taken together, these studies highlight the promise of positive psychology-based cessation interventions and the importance of fostering ongoing engagement with PPIs.

Building on the PPT-S pilot trial, the present study utilized a randomized controlled trial design to test the efficacy of an enhanced version of PPT-S, called PPT-S+, which incorporated interactive text messaging to support greater daily engagement in PPIs, replacing the prior reliance in PPT-S on written homework logs and offering a mechanism of providing daily reminders to engage in the exercises. Text messaging has been used successfully to deliver PPIs in prior studies.^20,21 To enhance experimental rigor, we included daily text messaging in both the experimental and control conditions by providing to all participants the Text to Quit (T2Q) program. T2Q, which is not based on positive psychology, provides informational and supportive text messages based on a user’s motivational stage (e.g., preparing to quit) and has shown promise in enhancing smoking cessation counseling.²² We hypothesized that participants in PPT-S+, compared to those in a standard treatment control that also included T2Q (ST+), would have higher rates of abstinence across 12 months of follow-up and that this effect would be stronger among those with higher positive affect.¹⁶ Additionally, we hypothesized that the effect of PPT-S+ on smoking outcomes would be mediated by greater engagement in PPT-consistent quitting strategies, reduced attraction to smoking, and greater self-efficacy for smoking cessation.

METHODS

Study Setting

Study procedures were conducted at Brown University and Massachusetts General Hospital. After March 11, 2020, procedures transitioned to remote delivery due to the COVID-19 pandemic. Detailed methodology has been reported previously.²³

Participants

The sample included 271 adults (≥18 years) who smoked ≥5 cigarettes daily for over a year, were willing to use the nicotine patch, rated quitting smoking as at least moderately important (≥5 on a 10-point scale), and had a text-capable cell phone. Exclusion criteria included: current psychosis, affective or substance use disorder other than nicotine dependence; use of smoking cessation pharmacotherapy; pregnancy or nursing; contraindications for transdermal nicotine patch.

Procedures

Procedures were approved by the IRBs at Brown University and Massachusetts General Hospital. Participants were enrolled between February 2017 and September 2020, with a COVID-related pause from March-June 2020, and follow-ups concluded by October 2021. Recruitment occurred via advertisements in the Rhode Island and Boston areas. Individuals completed eligibility screening by phone followed by a baseline interview where they completed written informed consent, questionnaires, and interviews before randomization. They began counseling within one week of baseline, and quit dates were scheduled for the third counseling session. Follow-up assessments occurred 12, 26, and 52 weeks post-quit date. Compensation was $20 for baseline, $35 for the 12- and 26-week follow-up, and $50 for the 52-week follow-up. Participants who did not have unlimited texting received $40 to offset any additional costs on their cell phone bill. Research assistants conducting follow-ups were blinded to treatment conditions.

Randomization

Eligible participants were randomized to either PPT-S+ or ST+ by a staff member uninvolved in assessments or treatment. Urn randomization²⁴ was used to balance groups on gender (male/female), positive affect (PA) as measured by the Center for Epidemiologic Studies – Depression scale (CES-D PA),²⁵ and level of cigarette dependence, measured via the Fagerström Test for Cigarette Dependence (FTCD).^26,27

Treatment Conditions

All participants received six sessions of smoking cessation counseling over 6 weeks: an initial 50-minute first session followed by five 35-minute sessions. Target quit date was set for session 3. Participants received a text messaging support program and eight-week supply of transdermal nicotine patches (21mg or 14mg based on baseline smoking rate). Sessions were audio recorded and delivered by six counselors with master’s degrees or higher and experience in behavioral health. Counselors conducted both the PPT-S+ and ST+ sessions after 20 hours of training and received audio review of their first five cases and weekly supervision. Sessions were conducted in-person until March 11, 2020; thereafter, counseling occurred by telephone or videoconferencing. Content and timing of counseling sessions are described in Supplemental Table 1. Protocols are available upon request. One session from each case was rated for adherence by research assistants using checklists of session content across both conditions.

Standard Treatment+ (ST+)

ST+ was based on practice guidelines²⁸ with core components on recognizing and problem solving high-risk situations for smoking, providing support and reinforcement for success, and encouraging participants to seek support for smoking cessation outside of treatment.²⁹ To match contact time with PPT-S+, ~20 minutes of each ST+ session provided progressive muscle relaxation training, which has not been shown to improve smoking cessation outcomes.²⁸

PPT-S+

PPT-S+ shared core smoking cessation components with ST+ while integrating positive psychology concepts.¹⁶ Specifically, PPT-S+ included: (1) positive introductions including identifying signature strengths based on the Values in Actions Survey³⁰ and discussing how those can support quitting smoking; (2) employing signature strengths to help manage high-risk situations for smoking; (3) recording three good things that happened each day; (4) savoring positive experiences; (5) listening actively and constructively responding when people report positive events; and (6) savoring acts of kindness performed by oneself or others.

Text Messaging

Starting at session 1, all participants received T2Q,²² an 8-week program of daily texts tailored to the user’s phase of quitting that includes behavioral and motivational messages and information on dealing with nicotine dependence and withdrawal. Participants received one message per day leading up to quit date, and two messages per day in the six weeks after quit date for—a total of 96 texts if the participant quit on their target quit date. Participants also could receive on-demand messages in response to a texted key word (e.g., “crave”).²² The content of the text messaging program is provided in Supplemental File 1. Participants were informed that all text messaging was automated.

Text Messaging in PPT-S+

In addition to T2Q, participants in PPT-S+ received daily text reminders to complete PPT-S+ exercises discussed within counseling. Morning texts were static messages relevant to a given exercise, whereas evening texts asked for open-ended descriptions of how participants completed that day’s exercise, with a reminder sent if no response was received. Text messages were specific to the exercise discussed in the most recent counseling session and continued until that exercise was replaced by a new exercise or a maximum of two weeks. At session 6, participants chose which exercise(s) to continue for two more weeks. Excluding reminders to respond, participants could receive up to 119 text messages specific to PPT-S+.

Measures

Demographics and Covariates

Participants completed demographic questions, the 7-item FTCD,²⁷ and the CES-D²⁵, a well-validated 20-item scale (α in the current sample = .73). The CESD-PA subscale comprised four items with higher scores indicating greater PA (α = .73).

Primary Outcome

The primary outcome was biochemically verified 7-day point-prevalence abstinence at 12, 26, and 52 weeks. Participants provided breath samples for expired carbon monoxide (CO) analysis at baseline and follow-ups. At 12-, 26-, and 52-week follow ups, self-reported smoking abstinence with no other use of nicotine-containing products—including nicotine replacement therapy and electronic cigarettes—was verified by both CO <5ppm and saliva cotinine radioimmune assay analysis ≤ 15 ng/ml.³¹ For those reporting smoking abstinence but with past 7-day use of other nicotine-containing products, abstinence was verified only by CO. After March 11, 2020, abstinence was biochemically verified only by saliva cotinine test strip. Continuous smoking abstinence was defined as self-reporting no cigarettes smoked since target quit date, with abstinence biochemically confirmed at each follow-up. Participants who withdrew from the study or did not complete a follow-up were coded as smoking (see Figure 1); smoking status for the one participant who died during follow-up was considered missing.

Figure 1:

CONSORT diagram showing the flow of participants through each stage of the trial.

Mediators

Quitting Strategies.

Treatment Strategies Questionnaire at session 6 (i.e., 4 weeks after quit date) assessed use of six standard smoking cessation strategies (α = .77), such as avoiding high-risk situations and using nicotine patch as directed, and nine PPT-consistent strategies (α = .88). PPT-consistent strategies were worded generally so participants in each condition could rate whether they used that strategy and included intentionally focusing on mental health benefits of quitting, using signature strengths, savoring positive experiences, and engaging in positive social interactions.¹⁶ Single items asked participants whether they used relaxation techniques and whether “the text messaging program helped me quit smoking.” Items were rated on a 1 = strongly disagree to 5 = strongly agree scale.

Self-Efficacy.

Smoking cessation self-efficacy was measured at each counseling session using a well-validated 9-item scale.³² Participants rated their confidence that they would not smoke in various situations, such as “with friends at a party,” on a scale from 1 = not at all to 5 = extremely. A total score was calculated by averaging across responses, with higher scores indicating greater self-efficacy; α at session 6 = .95.

Residual Attraction to Smoking.

Residual attraction to smoking was assessed at session 6 with a validated 3-item scale.¹⁴ Items assessed (1) “Do you still miss cigarettes?” with response options of 0 = no or 1 = yes; (2) “Does smoking have any attraction for you now?” with response options ranging from 0 = no, not at all to 3 = yes, a lot; (3) “Would you return to smoking if it was not harmful?” with response options ranging from 0 = definitely not to 3 = definitely. A total score was calculated by adding items together; α = .79.

Sample Size Calculations

We determined sample size needs using abstinence rates estimated from our developmental trial⁷ and pilot randomized controlled trial¹⁶; for PPT-S+: 45%, 30%, and 25% abstinence at 12, 26, and 52 weeks, with 20% continuous abstinence; for ST+: 30%, 17.5%, and 12.5% at 12, 26, and 52 weeks, with 8% continuous abstinence. Power analyses indicated a sample size of 272 participants would achieve power of .80 in repeated measures analyses, whereas a sample size of 298 participants would be needed for the continuous abstinence outcome. To allow for up to 10% of participants being lost to follow-up, the original targeted sample size was 340. Recruitment goals lagged projections early in the project, which was exacerbated by the recruitment pause necessitated by the COVID-19 pandemic. Given budgetary constraints as the end of the project neared, we set a new recruitment goal of 272 participants given that primary analyses used a “missing=smoking” assumption.

Modifications Due to COVID-19

After March 11, 2020, counseling, baseline (n = 54) and follow-up (n = 290) assessments were done remotely via videoconference or phone and with Qualtrics surveys. To confirm eligibility, newly enrolling participants were mailed a saliva cotinine test strip (NicoTest^®; lower level of detection = 30ng/ml) after screening eligible and then completed a baseline videoconference call with a research assistant, in which they showed identification and completed the cotinine test. For follow-ups, those reporting smoking and other nicotine abstinence were mailed a test strip and completed the test while on videoconference. There were 19 instances where test strips were unavailable due to product shortages during the pandemic, and in these cases, self-report alone was accepted.

Analysis Plan

We conducted chi-square tests and logistic regressions to examine group differences in biochemically-validated, 7-day point prevalence smoking abstinence at each follow-up as well as continuous abstinence across follow-ups. The primary outcome analysis utilized Generalized Estimating Equations^33,34 to test the effect of treatment condition on abstinence at 12, 26, and 52 weeks. The primary, between-groups, independent variable was treatment group assignment with ST+ as the reference. The model contained a linear effect of time as well as variables included in the urn randomization—gender, cigarette dependence, and CES-D PA—following clinical trial analysis recommendations.³⁵ Next, we tested the interaction between treatment condition and time to determine whether effects of PPT-S+ changed over time. We also tested whether baseline PA moderated the effect of PPT-S+ on smoking outcome.

To examine mediators, we first tested whether intervention condition predicted quitting strategy use, reduced attraction to smoking, or greater self-efficacy as assessed at 6 weeks. We then ran a GEE model with these variables added to the main effects GEE model described above. Because participants reported on strategy use during their quit attempt, whereas attraction and self-efficacy reflected how participants felt at the time of assessment, two separate GEE models were conducted: one model with strategy use only and one with both attraction and self-efficacy.

All GEE models controlled for COVID-19 timing due to its association with smoking outcome; however, conclusions did not differ when analyses did not include this covariate. We also repeated primary outcome analyses including only participants who were recruited prior to the pandemic; results were equivalent to those including all participants. Data and analysis code are available on request to the corresponding author.

RESULTS

Figure 1 shows the CONSORT diagram of participant flow through the study in which 271 participants were randomized to PPT-S+ (n = 140) or ST+ (n = 131). Table 1 shows the sample characteristics. Only baseline cigarette dependence appeared to differ by treatment condition in a way that could be clinically meaningful, with those randomized to PPT-S+ reporting an average score of 5.07 (SD = 2.01) out of 10 and those randomized to ST+ reporting an average score of 4.55 (SD = 2.08), Cohen’s d = −0.26. The a priori analysis plan included FTCD score as a covariate.

Table 1.

Baseline participant characteristics by treatment group

Demographic and clinical characteristics	Overall sample n (%) or M (SD)	PPT-S+ (n = 140) n (%) or M (SD)	ST+ (n = 131) n (%) or M (SD)
Female	149 (55.2%)	78 (55.7%)	71 (54.6%)
Race
Black	29 (10.8%)	12 (8.6%)	17 (13.1%)
White	230 (85.5%)	119 (85.6%)	111 (85.4%)
Other racial identities1	10 (3.7%)	8 (5.8%)	2 (1.5%)
Hispanic/Latino	14 (5.2%)	8 (5.7%)	6 (4.7%)
Years of age	49.20 (12.02)	49.19 (12.05)	49.20 (12.03)
Unemployed	84 (31.1%)	40 (28.6%)	44 (33.8%)
Years of education	14.13 (2.90)	14.24 (2.94)	14.02 (2.87)
Number of cigarettes per day	16.89 (7.00)	17.35 (7.07)	16.40 (6.93)
FTCD	4.82 (2.06)	5.07 (2.01)	4.55 (2.08)
CES-D positive affect	2.26 (0.71)	2.24 (0.73)	2.28 (0.69)
CES-D total score	10.63 (8.57)	10.50 (7.86)	10.78 (9.29)

Note. FTCD = Fagerström Test for Cigarette Dependence; CES-D = Center for Epidemiological Studies-Depression Scale; ST+ = standard treatment; PPT-S+ = positive psychotherapy for smoking cessation.

¹Other racial identities included four American Indian/Alaskan Native, two Asian American, and four multi-racial participants.

Of those randomized, 100% completed session 1. Participants attended an average of 4.66 (SD = 1.67) sessions in ST+ (48.1% completed all six sessions) and 4.62 (SD = 1.76) sessions in PPT-S+ (51.4% completed all sessions). In ST+, 90.1% (SD = 11.4) of intended topics were covered in the 112 counseling sessions that were rated, compared to 91.9% (SD = 13.04) in the 117 PPT-S+ sessions that were rated. In ST+, only 3.8% of intended PPT-S+ topics were rated as present across sessions, indicating minimal contamination. Participants in PPT-S+ received interactive texting requests an average of 53.0 (SD = 11.9) days and responded on an average of 27.8 days (SD = 18.6), or 52.3% (SD = 32.5) of the days they were prompted.

Follow-up rates at 12, 26, and 52 weeks were 80.8%, 78.2%, and 76.8%, respectively. Participants who completed all follow-ups, compared to those who missed one or more, did not differ significantly in condition assignment or characteristics shown in Table 1.

Smoking Outcomes

As shown in Table 2, ~20% of participants achieved biochemically confirmed 7-day point-prevalence abstinence at each follow-up, and ~9% were continuously abstinent. These rates did not differ significantly by condition. As seen in GEE models in Table 3, participants were 1.80 times more likely to be abstinent at follow-ups that occurred after COVID-19 began compared to those that occurred before. No other variables, including PPT-S+, significantly predicted abstinence. The interactions between PPT-S+ and time [OR = 0.76, 95% CI = 0.41, 1.43, p = .40], baseline CES-D PA [OR = 0.91, 95% CI = .42, 1.99, p = .81], and whether participants were recruited after COVID-19 restrictions [OR = 1.12, 95% CI = 0.38, 3.32, p = .83], were nonsignificant.

Table 2.

Intervention group differences in biochemically confirmed 7-day point-prevalence abstinence by timepoint and continuous smoking abstinence

Timepoint	PPT-S+	ST+	OR	95% CI	p
12 weeks	20.9%	19.1%	1.12	0.62, 2.03	.72
26 weeks	19.4%	20.6%	0.93	0.51, 1.69	.81
52 weeks	19.4%	21.4%	0.89	0.49, 1.60	.69
Continuous abstinence	9.4%	9.2%	1.02	0.45, 2.33	.96

Note. OR indicates odds ratio between PPT-S+ and ST+ groups (ST+ as the reference group). Continuous abstinence was defined as self-reporting no cigarettes smoked since their target quit date, with abstinence biochemically confirmed at each follow-up.

Table 3.

Adjusted effects of PPT-S+ and quitting strategies on smoking abstinence at 12, 26, and 52 weeks

	Main effects model			Quitting strategy use model
Variable	OR	95% CI [LL, UL]	p	OR	95% CI [LL, UL]	p
Covariates
Female	1.13	0.68, 1.86	.64	1.11	0.63, 1.96	.72
Baseline FTCD	0.93	0.84, 1.04	.20	0.97	0.85, 1.11	.65
Baseline CES-D PA	1.33	0.89, 1.98	.16	1.082	0.69, 1.71	.73
Post-COVID-19 follow-up	1.80	1.24, 2.60	.00	1.77	1.16, 2.70	.01
Time effects
26-week follow-up	0.89	0.65, 1.22	.48	0.74	0.51, 1.07	.11
52-week follow-up	0.87	0.62, 1.22	.42	0.81	0.55, 1.21	.31
Treatment effect
PPT-S+	1.02	0.62, 1.69	.94	0.85	0.47, 1.52	.58
Quitting strategy use 1
Standard strategies	-	-	-	1.61	1.04, 2.50	.04
PPT-strategies	-	-	-	0.82	0.43, 1.55	.53

Note. FTCD = Fagerström Test for Cigarette Dependence; CES-D = Center for Epidemiological Studies Depression Scale; PPT-S+ = positive psychotherapy for smoking condition. For the time effects, the 12-week follow-up served as the reference group. For the treatment effect, ST+ is the reference group. Bolded associations are significant at p < .05.

¹Strategies were assessed at Session 6 (end of counseling, 4 weeks after target quit date).

Quitting Strategies, Self-Efficacy and Residual Attraction to Smoking

Participants in PPT-S+ did not report significantly greater use of PPT-consistent strategies while quitting smoking compared to those in the ST+ (see Table 4), nor did they significantly differ in their use of standard smoking cessation strategies. Those in ST+ more strongly agreed that they used relaxation techniques while quitting compared to those in PPT-S+. Unexpectedly, ST+ participants, compared to PPT-S+ participants, more strongly agreed that the text messaging program helped them quit smoking. There were no significant condition differences in post-quitting residual attraction to smoking or cessation self-efficacy.

Table 4.

Quitting strategies and putative mediators at Session 6 (end of counseling) by group

	PPT-S+	ST+
	m (SD)		t	p
Quitting Strategies 1
PPT-consistent strategies	4.09 (0.57)	4.05 (0.51)	0.48	.63
Standard smoking cessation strategies	3.88 (0.82)	4.07 (0.64)	1.62	.10
Used relaxation techniques (1 item)	3.78 (1.05)	4.08 (0.86)	2.03	.04
Help from text-messaging (1 item)	3.55 (1.26)	3.92 (0.96)	2.04	.04
Putative Mediators
Residual attraction to smoking2	1.67 (1.21)	1.69 (1.23)	0.10	.92
Average smoking cessation self-efficacy3	3.80 (1.03)	3.75 (1.12)	0.23	.82

Note. Bold typeface indicates a significant difference between groups at p < .05.

¹Items were rated on a 1 = strongly disagree to 5 = strongly agree scale with scores indicating the item average.

²Three items were rated on different scales with a total possible score range of 0–7, with higher averages indicating a stronger attraction to smoking.

³Nine items were rated on a scale from 1 = not at all to 5 = extremely and averaged.

Greater use of standard smoking cessation strategies was associated with significantly greater odds of smoking abstinence across follow-ups (Table 3). Use of PPT-consistent strategies was not significantly associated with outcome. In subsequent models, neither self-efficacy (OR = 1.13, 95% CI = 0.84, 1.52, p = .41) nor residual attraction to smoking (OR = 0.81, 95% CI = 0.62, 1.05, p = .11) at session 6 were significantly associated with abstinence.

DISCUSSION

Results of this randomized controlled trial did not indicate that a positive psychology-based smoking cessation treatment produced better outcomes than a rigorous standard treatment control that included relaxation training to ensure equal contact time in counseling, with both conditions providing nicotine patch and text messaging support. The reasons for the differences in results between the present trial and the pilot trial that showed promise for PPT-S¹⁶ are unclear, but results suggest the current intervention failed to activate the PPT-consistent strategies for quitting posited as a primary mechanism of action. In the pilot trial of PPT-S, the mean on the PPT-consistent strategies scale was 4.3 (SD = 0.5) out of 5 for those in the PPT-S condition vs 4.1 (SD = 0.6) in the current trial for those in PPT-S+, a small effect size indicating less engagement in PPT-consistent strategies.

One primary difference between this trial and the pilot, which may explain the lack of effect on the posited mechanism of action, was that PPT-S+ exercises were completed by texting rather than handwritten responses; this change was made because text messaging offers the capacity to prompt participants daily to complete PPIs and a ready means of tracking exercise completion for research purposes. Because of text messaging delivery, participants were not asked to write and deliver a letter of gratitude, as they were in the pilot trial. Texting to deliver daily PPIs has been used successfully in prior trials.^20,21 In the present trial, participants in PPT-S+ responded to just over 50% of interactive texting requests, which is similar to exercise completion rates seen in an app-based positive psychology smoking cessation intervention that showed promising effects.¹⁸ It is possible that the rapid, brief form of communication used in texting may not have engaged the deep self-reflection that written PPIs are intended to engender.³⁶ Whether steps could be taken to enhance depth of processing of PPI exercises via texting—for example, by chatbot-facilitated interaction—is worthy of study.

Another way the present study differed from the pilot was that both conditions received a text messaging support program (T2Q), which was absent in the pilot study. We believed it was essential to include T2Q in ST+ so that any effect of PPT-S+ could be attributed to the content of the intervention rather than the absence versus presence of text messaging. T2Q may have provided positive effects in both conditions by supporting active coping and self-efficacy, thereby washing out benefits of PPT-S+ vs. ST+ counseling; overall, results indicate a lack of a unique effect of PPT-S+ content. Surprisingly, participants in PPT-S+ rated the overall texting program as less helpful than those in ST+. Using texting to support both smoking cessation and PPIs may have diluted the value of either or both. Although higher intensity smoking cessation text messaging interventions have not yielded superior outcomes than lesser intensity interventions,³⁷ we are unaware of research examining whether multi-target text messaging produces weaker effects than more focused programs. Because there was not a condition that received only PPT-S+ specific texts without T2Q, we cannot disentangle whether the combination of the two programs had iatrogenic effects. We do not have ratings from participants regarding whether the text messaging was overly burdensome in PPT-S+, but those in PPT-S+ did receive substantially more texts than those in ST+ (an intended maximum of 215 vs. 96 texts, respectively).

Given the lack of intervention effect on self-reported use of PPT-consistent strategies, it was not surprising that PPT-S+ and ST+ did not differ significantly in residual attraction to smoking or self-efficacy for quitting at the end of counseling; however, it was unexpected that neither of these putative mediators were associated with subsequent smoking outcomes. Use of standard smoking cessation strategies—like using nicotine patch and avoiding high risk situations—did predict longer-term smoking outcomes. As anticipated, conditions did not differ on this variable, which may explain their very similar smoking outcomes.

Limitations

The study was interrupted by the COVID-19 pandemic, and procedures were altered after March 11, 2020. Biochemically confirmed abstinence was higher after that date. This could reflect real differences in outcomes due to increased motivation to quit³⁸ and reduced exposure to smoking cues while social distancing protocols were in place. It also may reflect that the sensitivity of the cotinine test used at home was lower than tests conducted at the study sites, that CO was unavailable for remote verification of abstinence, and that cotinine test strips were unavailable for remote verification in nineteen cases. However, sensitivity analyses did not indicate that the treatment effect differed for participants recruited prior to versus after the pandemic started.

Slower-than-anticipated recruitment and pandemic-related delays in the study meant that we were unable to reach our desired recruitment goal, resulting in reduced statistical power. The overall abstinence rates in the trial were modest, perhaps reflecting the fact that pharmacotherapy involved nicotine patch only rather than combined NRT or varenicline. The exclusion of those with current affective and substance use disorders limits generalizability of results. Finally, the study design is unable to elucidate whether the lack of effects may have been due to using text messaging to deliver both the PPT-S+ exercise tracking and smoking cessation support.

Conclusions

The current trial provided no evidence for the efficacy of PPT-S+ for smoking cessation and a failure of the program to engage participants in PPIs sufficiently. Future analyses can address patterns of engagement with the PPIs via analysis of text responses to determine whether deeper engagement in the PPIs resulted in more positive outcomes. Finally, PPIs can contain a variety of components and be delivered in a variety of formats. Given promising results of prior positive psychology-based smoking cessation¹⁹ and health behavior interventions,³ greater research is needed to understand which specific PPIs or combinations of PPIs best enhance health behavior interventions and which delivery formats best facilitate PPI engagement.

IMPLICATIONS.

Results indicated no added benefit of incorporating a positive psychology intervention into smoking cessation counseling that also included text messaging support for cessation and nicotine patch. One potential explanation for this negative result is that there may be interference effects when additional text messaging-based interventions are added to smoking cessation support; participants who did not have this additional content rated the text messaging program as more helpful overall to their smoking cessation efforts than those who did, which suggests caution in delivering text messaging programs that may have too much or too diverse a content.

Data Availability

The data underlying this article will be shared on request to the corresponding author.