Abstract
Objective
To assess the feasibility, acceptability, and treatment outcomes of a proactive motivational texting program with phone-based care coordination among people with chronic obstructive pulmonary disease who smoke (PCWS).
Design
Randomized, unblinded pilot among PCWS, regardless of intention to quit smoking.
Methods
Consenting participants were randomized 2:1 to the intervention vs a mailed offer of cessation services. Intervention consisted of a 12-week tapering-frequency motivational texting protocol, adapted from an existing clinically available protocol. Tailoring included COPD-specific messages and the option for phone-based care coordination, connecting participants to their choice of tobacco treatments. Primary outcomes were feasibility and treatment participation through 6 months. Results were analyzed using logistic regression, adjusted for imbalanced baseline characteristics deemed a priori likely to impact cessation behaviors.
Results
Of 697 invited, 125 enrolled and were randomized (83 intervention and 42 control) and 148 were ineligible, for an enrollment rate of 22.8% among eligible individuals. Primary reasons for ineligibility were: already quit or engaged in cessation efforts (n = 79), no text-capable phone (n = 50). Participants were older (mean age 68.4 years), predominantly male (82.3%) and white (78.2%). Most participants in the intervention arm completed the texting protocol (71/83, 85.5%) and responded to it at least once (63/83, 75.9%). Cessation outcomes were not statistically significant but favored the intervention. More participants in the intervention arm utilized any treatment (34.9% vs 21.9%, OR 2.37, 95%CI 0.94-6.40); 11 (13.2%) vs 1 (2.4%) (P = 0.053) completed a care coordination call. Acceptability of the texting protocol was high with 80.6% of participants finding the content relevant and 69.4% indicating they would recommend it. There were no identified harms. Study has completed all study procedures.
Conclusion
In this randomized pilot of a proactive, motivational texting program for PCWS, we found evidence of feasibility, acceptability, and possible benefits for cessation outcomes. Such a program is low burden and may be suitable for both future research and broad use.
Keywords
Introduction
Chronic obstructive pulmonary disease (COPD) remains a leading cause of morbidity and mortality, with most cases attributable to cigarette smoking. Historically, military personnel and Veterans have high rates of tobacco use, and Veterans suffer disproportionately from COPD. 1 Because most cases of COPD are attributable to smoking, patients with COPD smoke at rates double the national average,e 2 with 30%–40% still smoking.3,4 Quitting smoking is beneficial even in cases of advanced disease, 5 leading to improved symptoms2,6 and decreased risk of death from multiple smoking-related causes.5,7,8
Patients with COPD who smoke may be prompted to quit by the presence of respiratory symptoms, which may increase motivation to participate in treatment.9,10 They also often have significant barriers to cessation including high levels of addiction to nicotine, 11 low socioeconomic status12,13 and high burden of medical and psychiatric comorbidities. 14 Many have low motivation to quit, while others are motivated to quit, but are hampered in their success by a lack of self-efficacy 15 and quit confidence. 16 Despite these potential barriers, evidence suggests that patients with COPD can quit successfully at rates similar to patients in general if they receive the guideline-recommended combination of counseling and medication.17,18 Healthcare settings play a key role in connecting patients to cessation services. About 80% of patients with COPD seen in clinics receive brief advice, but few engage in supported quit attempts. 19 Therefore, motivating patients to quit and connecting them to treatment remains challenging in the context of sporadic visits when both clinicians and patients have competing clinical priorities. The motivation to quit or cut down can change rapidly, making every patient contact an important opportunity to address smoking behaviors.
Proactive patient engagement is an effective method for increasing participation in tobacco treatment and cessation. Proactive patient engagement involves systematically identifying and contacting an individual outside of a routine clinic visit and providing an intervention to boost treatment participation. It has been found to be effective for increasing smoking cessation in a number of patient populations (eg, hospitalized patients,20,21 Veterans in primary care 22 and mental health clinics 23 ). Effective methods include live phone calls, interactive voice response systems, mailed materials and patient portal messages. While effective, proactive outreach is rarely implemented, in part due to perceived high burden on the health system with more intensive outreach methods. 24
Information and communication technology (ICT) such as text messaging and patient portals are increasingly used in healthcare, 25 allowing for broad reach with low cost. Use of technology among older individuals has rapidly increased and ICT is now a key part of health promotion. Many studies confirm that text messaging protocols can boost successful cessation among individuals making a quit attempt,26-29 and these approaches are supported by a Cochrane systematic review. 30 Our goal was to develop and test a protocol for proactive tobacco outreach using text messaging, tailored to patients with COPD who smoke but are not already engaged in a quit attempt. Our primary goal was to assess the feasibility and acceptability of the program and to examine early outcomes of treatment engagement and cessation to generate estimated effect sizes for future large-scale studies.
Methods
Study Design and Setting
We performed a pragmatic 31 randomized feasibility pilot of the proactive motivational texting program. The study was conducted at the Minneapolis VA Health Care System (VAHCS) including the main medical center and 14 affiliated community-based clinics. This encompasses a highly rural catchment area including Minnesota and Western Wisconsin. Participants were enrolled remotely from March to December 2023. This study was approved by the Minneapolis VA IRB (#1708415) and registered on https://www.clinicaltrials.gov (NCT04462289). Because the study was fully remote and minimal risk, patients were apprised of the elements of informed consent by phone prior to enrollment under a waiver of documentation of informed consent.
Subjects and Recruitment
Potentially eligible participants were identified using the electronic health record (EHR) if they met the following criteria: 2 or more ICD-10 codes for COPD in the past 2 years at any VA healthcare system, a visit to primary care or pulmonary medicine at the Minneapolis VAHCS with in the past 18 months, last tobacco assessment indicated current smoking, not enrolled in tobacco treatment, not on hospice or under treatment for malignancy or dementia. Our facility cares for predominantly white males; therefore we oversampled women (12.5%) and racial or ethnic minorities (25%). Potential participants were mailed an opt-out letter. They were then contacted by phone to confirm eligibility and offer enrollment. Participants were excluded during the call if they had already quit smoking, were in a cessation program, lacked a text-capable phone, or declined participation. Participants were enrolled regardless of intention to quit smoking. Consenting individuals were enrolled into the VA’s ‘Annie’ texting platform, the system through which the intervention was delivered, then randomized 2:1 to the intervention and control arms. Randomization utilized permuted blocks of size 3 and 6 and was stratified by female/male and minoritized race or ethnicity.
Intervention
We follow published recommendations for reporting proactive outreach intervention studies. 32 To enhance scalability, we adapted a clinically available motivational texting program for tobacco treatment engagement. The program was delivered using the ‘Annie’ platform. ‘Annie’ is the VA’s clinical texting system, designed to support patient self-management protocols.
Base Protocol
The texting program had been developed by a group of psychologists and patient educators and was available for use by any clinician in the ‘Annie’ database of protocols. The protocol was delivered over 12 weeks, with greater frequency at the beginning (daily message) and diminishing frequency during subsequent weeks (twice weekly message). The content of the messages included informational statements (eg, “Stopping smoking is the most important thing you can do for your lung health. Even cutting down can improve symptoms like cough and wheezing.”) and assessments of readiness to quit or importance of quitting. Participants were asked weekly for the first 8 weeks, then biweekly, if they were ready to set a quit date; those indicating ‘quit yes’ were encouraged to enroll in cessation support by calling the VA quitline or their primary care provider.
Tailored Protocol
Based on a mixed methods assessment, 32 we tailored the intervention to the needs of our patient population. We collaborated with the Minneapolis VA Learning and Technology Annie Lead to create the adapted version of the texting protocol. Changes included: COPD-specific information, tips/quotes from Veterans who successfully quit, the option to text or call back and participate in live phone-based care coordination, and offering several choices for cessation programs. We obtained quotes from patients with COPD who had successfully quit, obtained from participants in tobacco treatment groups and pulmonary clinics. We included the 11 quotes ranked most highly on a survey of Veterans with COPD who smoke. In addition to general information about how to access quit resources (eg, call the VA Quitline, contact primary care), participants were encouraged to call the coordinator for a motivational discussion and connection to treatment options. Alternatively, if they responded to a text that they wanted to quit (‘quit yes’), this information was displayed in the Annie clinician-facing platform. These participants were contacted within approximately 2 business days to offer cessation services. Care coordination was conducted by phone by the research coordinator using a script offering the most intensive options (group counseling class with pharmacist medication management) to the least (VA app or texting program). (See Appendix A for script). Appropriate referrals were placed in the EHR. Participants could terminate the protocol at any time. The texting protocol is available in the VA Annie Library in Midwest region.
Control
The control arm received a mailed offer with information about cessation services, providing the same phone number to call for care coordination. All participants in both arms continued to have access to the tobacco treatment clinics and quitline services that were available to all patients via their clinical teams. Participants could be referred to these services by any clinicians or through the annual tobacco ‘clinical reminder’. This reminder prompts clinic staff to update smoking information and offer cessation services at one health-promotion visit each year.
Data Collection
Participants completed a brief baseline survey, either electronically via Qualtrics or during their initial phone intake. Surveys assessed demographics, health information, 33 smoking behaviors, prior tobacco treatments and cessation attempts, quit confidence, electronic health literacy 34 and views of tobacco treatments. Participants were surveyed again via Qualtrics 6 months after randomization. Participants who did not complete the survey after two electronic messages were contacted to complete a phone survey. Interactions with the texting program (eg, responses to queries, requests to ‘stop’) were extracted from the Annie platform. Participants were paid $20 following baseline survey and $20 following 6-month survey. Data team was blinded to treatment arm. Data team was provided a de-identified data set for analysis.
Outcomes
Six-month outcomes were assessed by self-report. Our primary outcomes were feasibility of enrollment, protocol completion and interaction, and acceptability of the texting protocol. Our primary outcome was any self-reported tobacco treatment participation since randomization. Additional secondary outcomes included 7-day cessation, cutting down, a composite of quitting or cutting down, quit attempts, use of individual tobacco treatments, participation in care coordination, and change in contemplation ladder. Cutting down was defined as a drop in frequency from daily to non-daily or a drop in at least one category level of cigarettes per day (defined in 10 cigarettes per day increments). Quitting or cutting down was a composite of 7-day abstinence or cutting down.
Statistical Analysis
Our primary analysis compared treatment participation rates between the intervention arms using a multivariable logistic regression. We identified several contextually important variables that were imbalanced between arms with the potential to explain variability in our primary outcomes; these included participant education level operationalized as “some college or less” vs the reference “college graduate or more”; participant’s time frame to quit smoking at baseline operationalized as “otherwise” vs the reference “no time frame to quit”; and the participant having had a quit attempt in the past year operationalized as “I have tried to quit” vs the reference “I haven’t tried to quit”. These were included, along with intervention arm, as predictors in the regression analysis. We report the model estimated odds ratio, corresponding 95% confidence interval, and P-value for the Wald test for each of the predictors included in the model. Analyses used RStudio version 4.4.1 35
Results
Recruitment and Enrollment
Figure 1 shows the results of recruitment. 697 individuals received opt-out letters. Of those, 563 (80.7%) were reached by phone, 148 of whom were ineligible. The primary reasons for ineligibility were already quit or engaged in cessation program (n = 79) or lack of a text capable phone to receive the messaging protocol (n = 50). Of those found to be eligible on phone outreach, 125 were randomized for an enrollment rate of 22.8% among eligible individuals. Results of Recruitment. *Administratively Withdrawn Due to Incomplete Baseline Data (<10% Questions Answered)
Participant Characteristics
Baseline Characteristics by Intervention Arm. Values Presented are n (%) or Mean (Standard Deviation) (n = 124)
aVariables identified with an * were imbalanced and considered for inclusion in adjusted models.
bEHEALS: 8-40, 8 being very low e-health literacy. Pack years calculated as years of smoking times average cigarette count per day. Respiratory symptoms questionnaire: scale from 0 to 16, higher being more symptoms.
Interaction with the Texting Program and Care Coordination
Successful completion of the 12-week protocol was high, with 71/83 (85.5%) in the intervention arm completing the 12-week protocol. Of those who did not complete it (n = 12), 4 requested to stop, 5 were dropped in error by the system, and 3 were dropped by the system due to repeated erroneous responses. Most participants responded to at least one query (n = 63, 75.9%), with a minority (n = 22, 26.5%) indicating an intention to set a quit date at least once. However, some responses (4/42 “quit yes” responses, 4/58 “quit no” responses) did not conform to the response format and were rejected by the system. These participants responded to the system with narrative text (eg, “not ready yet!”). Therefore, these responses were not displayed in the clinician-facing portal. Responsiveness to the prompts to set a quit date diminished over time. Following week 8, no participants responded to the quit date query (Figure 2). Responses to “Ready to Set a Quit Date?” by Week of Texting Protocol (N = 83)
Acceptability
Acceptability of the texting program was high. Among the 62 respondents, 50 (70.6%) agreed or strongly agreed that the content of the messages was understandable and relevant; 45 (72.6%) agreed or strongly agreed that the program was easy to use and navigate; and 43 (69.4%) would recommend the program to someone else. Approximately a third of participants (n = 21, 33.9%) felt that it motivated them to stop smoking.
Cessation Support and Quit Behaviors
Nonresponse to the 6-month follow-up survey was low (8%) and not differential between arms. Results favored the intervention arm but were not statistically significant. In unadjusted analyses for our primary outcome, more participants in the intervention arm used cessation services (34.9% vs 21.9.0%, P = 0.14). More completed a care coordination call 13.2% vs 2.4%, P = 0.053), quit smoking over a 7--day timeframe (9.6% vs 7.3%, P = 0.928), cut down in heaviness or frequency of smoking (30.1% vs 22%, P = 0.609), and quit or cut down (38.6% vs 29.3%, P = 0.399). Interestingly, fewer participants in the intervention arm had tried to quit for at least 24 hours (42% vs 51.2%, P = 0.789) and there was no apparent impact on moving up the quit ladder (27.7% vs 31.7%, P = 0.906).
Treatment Participation and Cessation Behaviors at 6 months (n = 124)
Combination nicotine replacement: short-acting nicotine and patch.
NRT: Nicotine replacement therapy.
Adjusted Analyses
Adjusted Analyses Examining Cessation Behaviors and Participation in Quit Support (n = 109)*
*Models adjusted for educational attainment, timeframe to quit at baseline, past year quit attempt.
Discussion
In this randomized pilot of a proactive motivational texting program for patients with COPD who smoke, we found that the program was feasible and acceptable with high rates of completion of the 12-week protocol. While not statistically significant due to sample size limitations, the greater than 10% increase in use of cessation services and fairly high participation in care coordination calls suggests a positive impact. Recruitment was limited primarily by lack of interest in participation, though a significant number (8.8%) of potential participants lacked a text-capable phone. This finding is consistent with our prior research that identified an ongoing ‘technology gap’ in this population of older patients with COPD who smoke. 36 We designed the program to be fully remote and centralized, which allowed us to successfully recruit rural-residing individuals and demonstrate feasibility for this often hard to reach population. While the protocol was designed to be motivational it was notable that while more participants in the intervention arm engaged with cessation supports, there was no observed effect on quit attempts or change in the quit ladder. This suggests that the utility of the program may be primarily ease of connection to treatment for otherwise motivated individuals, rather than an effect on motivation per se. It is notable that despite many participants reporting no timeframe to quit at baseline, many of these individuals did in fact use assistance, attempt to quit smoking, or change their smoking behaviors over 6 months of follow-up. Unsurprisingly, more participants intending to quit at baseline successfully became abstinent from tobacco. Our results support that motivation to quit can change rapidly over time. Far from being refractory to cessation support, patients with COPD who smoke, even who state they are not planning to quit, made frequent supported quit attempts and were open to engaging in tobacco treatment. Longitudinal outreach programs such as this one can capitalize on this shifting motivation to engage patients in supported quit attempts.
In our preliminary work, we found that navigating technology to access care was a common source of frustration 36 and patients preferred live phone care to automated systems. Clinicians perceived the same preferences and strongly recommended that proactive programs incorporate live phone calls whenever possible. 24 However, the burden to the health system associated with exclusively live phone outreach is a barrier to implementation. Therefore, we chose a ‘hybrid’ approach, combining automated text outreach with phone-based care coordination for interested individuals, which was successful in connecting patients to care. Patients and clinicians also prefer to have choices when accessing tobacco treatments. In our prior work, clinicians perceive that patients’ lack of knowledge that there are choices is a barrier to cessation participation, particularly for patients with past experience with quit support that was unsuccessful. We therefore explicitly offered participants a choice of care, rather than an automatic referral. While these approaches increased the complexity beyond that of a fully technology-based intervention, the trade-off appears justified by the high acceptability and increased uptake of combined pharmacologic and behavioral treatments. Future work could compare the hybrid approach text-based outreach with phone coordination) with fully automatic options.
Previous studies have provided proactive outreach using a variety of modalities and intensities and our findings align with these results. In a randomized study of mailed vs phone-based outreach with free nicotine replacement offered to low-income individuals who smoke, at one year more individuals in the outreach arm had quit (16.5% vs 12.5%) and used combined behavioral and pharmacologic treatment (17.4% vs 3.6%), similar to what we observed. 37 Another four-arm study compared the impact of brief phone advice vs mailed NRT alone vs mailed NRT and text messaging vs usual care on short-term cessation behaviors at 12 weeks, finding that more participants in the texting, NRT, and combined texting and NRT arms attempted to quit, with positive effects on 7-day abstinence. However, they found no differences between the three intervention arms. 38 A large study (5,123) of Veterans in primary care compared usual care to a mailed + phone offer of cessation services (counseling by phone or in person) finding increased prolonged cessation at one year (13.5% vs 10.9%) and increased treatment utilization 39 in the intervention arm. Another study examining 3 monthly mailed offers of cessation services (10% acceptance of treatment, 5.3 vs 1.1% 7-day abstinence at 3 months 21 ) found fairly similar effects as our program. When formally assessed, all proactive outreach interventions have been found to be cost-effective well below usual thresholds. 40 While we did not assess the cost in this pilot, the overall cost of the intervention is expected to be low and would be expected to be similar to prior programs.
To minimize costs and enhance scalability, we elected to use a clinically available texting platform and tailor it to our target patient population. We found a moderate rate of system errors (5/83, 6%) resulting in the protocol being terminated early. A greater concern was that some participants struggled to conform to the response format required by texting system, resulting either in discharge from the protocol or failure of an attempted response to be recorded. Our participants included a significant proportion with low electronic health literacy. In our preliminary work, 36 very few patients with COPD who smoke had ever used any sort of text messaging program or health-related app. This lack of experience likely explains their challenges with the system. In future iterations of the texting protocol, we plan to simplify the required responses (eg, a single number or letter response). We plan to include more ‘training’ messages up front about how to interact with the system. We hope to incorporate more personalized options in the program, for example allowing participants who want to continue to opt for ongoing messaging, and to streamline the process of contacting the team for care coordination. The requirement to contact patients by phone or in person prior to enrolling them in the program is a limitation to broad uptake. We are currently working on improving the program in preparation for a larger-scale multisite study, considering alternative strategies for recruitment such as clinic-based outreach and enrollment.
Our study has limitations. This was a small, single center pilot including mostly white, male Veterans. Results may therefore not be generalizable to other settings and should be replicated on a larger scale. However, this is a group expected to be particularly refractory to treatment suggesting that a similar protocol would work as well or better in other healthcare settings. We chose to make relatively moderate adaptations to an existing protocol and there may have been other areas that would have benefitted from further tailoring. We identified several areas that would benefit from refinement, including ‘dropping’ of participants, and plan to make these changes for further research and clinical use. Outcomes were relatively short and measured by self-report. As this was a pilot study, sample size was determined by precision of estimates and we were underpowered for statistical significance. There was baseline imbalance in no intention to quit, which was adjusted for by statistical means. If anything, this strengthens the results given that this is a population particularly refractory to cessation.
Conclusion
In conclusion, we found that a proactive motivational texting program with phone-based care coordination was feasible and acceptable for patients with COPD who smoke and may result in increased use of cessation resources. Such a protocol is suitable for implementation in many clinical settings including pulmonary clinics where patients could be enrolled at or around a planned visit. Future research will test the protocol on a broader scale and integrate further improvements to increase the usability of the texting interface.
Footnotes
Acknowledgements
The authors would like to acknowledge Lara Carson for her invaluable help tailoring the texting protocol.
Author Contributions
AM obtained funding, wrote protocol, completed primary manuscript drafting and is the overall guarantor of study findings. MC assisted in protocol development, completed data collection, and assisted in manuscript drafting. AK and DN designed and completed statistical analysis, assisted in data interpretation and manuscript drafting. SF and HH contributed to project and protocol development, interpreted data, and participated in manuscript drafting.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by a Department of Veterans Affairs (VA) Health Systems Research Career Development Award to Dr Melzer (VA HSR CDA HX003067-01A1). Funder had no role in design or conduct of the study.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Guarantor
Dr Melzer takes responsibility for the content of this manuscript, including data and analysis.
IRB Approval
Data Availability Statement
De-identified data can be shared on reasonable request to the principal investigator.
Disclaimer
The views expressed in this article are those of the authors and do not necessarily reflect the position or policy of the Department of Veterans Affairs or the United States government.
Prior Abstract Publication/Presentation
A portion of these results were presented in abstract form at the American Thoracic Society meeting May 2025.