This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR Formative Research, is properly cited. The complete bibliographic information, a link to the original publication on http://formative.jmir.org, as well as this copyright and license information must be included.
Persistent smoking after a cancer diagnosis predicts worse treatment outcomes and mortality, but access to effective smoking cessation interventions is limited. Smartphone apps can address this problem by providing a highly accessible, low-cost smoking cessation intervention designed for patients with a recent cancer diagnosis.
This study aimed to summarize our development process and report the trial design, feasibility, participant acceptability, preliminary effectiveness, and impact on processes of change (eg, cancer stigma) of the first-known smoking cessation smartphone app targeted for cancer patients.
We used an agile, user-centered design framework to develop a fully automated smartphone app called Quit2Heal that provided skills training and stories from cancer survivors focusing on coping with internalized shame, cancer stigma, depression, and anxiety as core triggers of smoking. Quit2Heal was compared with the National Cancer Institute’s QuitGuide, a widely used stop smoking app for the general population, in a pilot double-blinded randomized trial with a 2-month follow-up period. Participants were 59 adult smokers diagnosed with cancer within the past 12 months and recruited through 2 cancer center care networks and social media over a 12-month period. The most common types of cancer diagnosed were lung (21/59, 36%) and breast (10/59, 17%) cancers. The 2-month follow-up survey retention rate was 92% (54/59) and did not differ by study arm (
Compared with QuitGuide participants, Quit2Heal participants were more satisfied with their assigned app (90% [19/21] for Quit2Heal vs 65% [17/26] for QuitGuide;
In a pilot randomized trial with a high short-term retention rate, Quit2Heal showed promising acceptability and effectiveness for helping cancer patients stop smoking. Testing in a full-scale randomized controlled trial with a longer follow-up period and a larger sample size is required to test the effectiveness, mediators, and moderators of this promising digital cessation intervention.
ClinicalTrials.gov NCT03600038; https://clinicaltrials.gov/ct2/show/NCT03600038
In the United States, 15% to 54% of cancer patients are cigarette smokers at the time of their diagnosis [
Unfortunately, up to 80% of smokers with cancer continue to smoke after their diagnosis [
In direct response to this need, the US NCI created a Moonshot-funded Cancer Center Cessation Initiative (C3I) to support the implementation of evidence-based cessation interventions at 42 NCI-designated cancer centers [
One method for all smokers with cancer to access effective and low-cost smoking cessation treatment is via smartphone-based smoking cessation software apps [
Smartphone apps for smoking cessation are showing solid promise among the general population of smokers [
To address this knowledge gap, we developed a smartphone app, called Quit2Heal
We used an iterative, user-centered design approach [
Our formative research led us to iteratively develop content on the (1) consequences of continued smoking versus quitting smoking for health domains such as daily functioning and cancer treatment outcomes, (2) skills for coping with depression and anxiety often associated with a cancer diagnosis, (3) self-compassion exercises for coping with cancer-related stigma and internalized shame, (4) advice on how to seek support for quitting smoking from cancer treatment providers (eg, oncologist), and (5) testimonials from cancer survivors describing how quitting smoking has allowed them to live more meaningful lives. The wireframes created by our user experience designer were iterated upon by our team. The content was user tested with 13 smokers currently receiving cancer treatment to get feedback on usability and content in 3 iterative rounds of testing. Our user testing also identified the cancer patients’ choice of the best name for the app, Quit2Heal. After our developer created an alpha version of the Quit2Heal app, the study team identified edits for the content and features as well as any technical bugs.
Our review yielded a beta version that was tested in a 7-day diary study with 5 adult smokers (3 women and 2 men) currently receiving cancer treatment who had varying levels of technical ability and confidence in quitting smoking. The diary study included a 30-min onboarding session, 7 nightly 10-min surveys about each participant’s experience of the app that day, a 10-min call on day 4 to discuss their impressions of the app so far, and a 45-min exit interview about their overall experience and the usability of the app. All participants rated the app as highly useful overall, were very satisfied overall, and would recommend the app to other cancer patients who smoke. They all liked the 5 content areas created specifically for cancer patients who smoke. The major problem area was that they were not clear where to start the app’s program. Our remedies included (1) adding an introduction with screenshots showing how to begin the program and (2) graying out the sections that come later in the program until they become available. After minor usability concerns were remedied, the final version of Quit2Heal was ready for testing in the pilot randomized controlled trial (RCT) described herein.
Participants with the following eligibility criteria were included in the study: (1) aged 18 years or above, (2) diagnosed with cancer within the past 12 months or currently receiving cancer treatment or planning to receive cancer treatment in the next 3 months (consistent with prior trials of smoking cessation in cancer patients [
The study participant flow diagram is shown in
Our original recruitment goal was 200 participants (100 per arm) based on our experience with recruiting 200 participants in prior pilot randomized trials of mobile health (mHealth) and electronic health (eHealth) for smoking cessation for the
Participant flow diagram.
A total of 2 reminder emails were sent over a 14-day period to individuals who did not respond to the initial email invitation. Individuals who did not consent or complete the enrollment process within the 14-day period were sent an email indicating that they were not enrolled. Participants not enrolled (or ineligible) were referred to
The enrolled participants were randomized (1:1) to either the experimental intervention (Quit2Heal, n=29) or the control intervention (QuitGuide, n=30). We used computer-generated randomly permuted block randomization, stratified by Heaviness of Smoking Index (score>4 [
To ensure participants were blinded to their assigned intervention, each app was branded as Quit2Heal. Contamination between the interventions was avoided with a unique username and password provided only to the study participant and by having an eligibility criterion of not having family, friends, or other household members participating.
Quit2Heal [
The first 5 levels contain content and exercises designed to prepare the users for their chosen quit day. Level 1,
The last 4 levels contain content and exercises designed to help the user stay smoke-free after their quit date. These levels contain 25 exercises that focus on coping with cancer-related depression and anxiety, withdrawal symptoms, slips, and potential weight gain and building smoke-free life activities. All levels contain at least one
Through the main menu, participants access the education section, which has 3 components. The first component educates on the negative consequences of continuing to smoke after a cancer diagnosis: (1) impacts on radiation, chemotherapy, and postsurgical recovery; (2) risk of second primary cancers; and (3) mortality. The second component educates on the positive consequences of quitting smoking after a cancer diagnosis: (1) improved treatment outcomes, (2) lesser chance of a second primary cancer, and (3) lower chance of mortality. The third section contains education on how to talk to a cancer care provider about the user’s smoking and ask for the provider’s assistance and support in quitting—which can help reduce shame and stigma. The participants can edit their quit plan, review their progress (eg, smoke-free days), and view the badges they earned for making progress in the program.
The comparison was NCI’s QuitGuide app [
QuitGuide is a
Both interventions were available for log in at any time after randomization. Neither was modified during the study (the apps used in this study are available for download, with tester usernames and passwords available upon request [
The procedures for follow-up data collection were modeled after the procedures that have been successful in our previous trials at maximizing data retention [
At baseline, participants reported their demographics, cancer (eg, type and stage), alcohol use (Quick Drinking Screen [
Utilization was assessed via data logged automatically by the secured server on how many times the app was opened during the 60 days after randomization. Owing to a database error, these data were only available from the 32 participants who were randomized after June 29, 2018.
Treatment satisfaction outcomes were the extent to which a participant (1) was overall satisfied with the assigned app, (2) would recommend the assigned app to a friend, and (3) believed that the assigned app was made for someone like them. The response choices for all items ranged from
The brief process measures, assessed at baseline and the 2-month follow-up, were internalized shame (5-item internalized shame subscale of the Social Impact Scale [
For scientific rigor and comparability with other low-intensity behavioral intervention trials [
The demographic characteristics, smoking behavior, and process measures at baseline were compared between the study groups using 2-sample
We used logistic regression models to analyze the differences between the treatment arms on binary cessation and satisfaction outcomes. A negative binomial model was used to analyze the right-skewed app utilization data. Linear models were used to analyze the changes in process indicator measures, adjusting for baseline value of the measure. All models were adjusted for the 3 variables used in stratified randomization. The models were also adjusted for any baseline characteristic that was both imbalanced between the study arms at baseline (ie,
As shown in
Baseline characteristics of Quit2Heal study participants.
|
Total (N=59)a | QuitGuide (N=30)a | Quit2Heal (N=29)a | |||
Age (years), mean (SD) | 45.2 (12.9) | 47.3 (13.5) | 42.9 (12.0) | .19 | ||
Male, % (n/N) | 25 (15/59) | 30 (9/30) | 21 (6/29) | .60 | ||
|
|
|||||
|
White | 78 (46/59) | 77 (23/30) | 79 (23/29) | ||
|
Black or African American | 12 (7/59) | 10 (3/30) | 14 (4/29) | ||
|
Native American | 2 (1/59) | 3 (1/30) | 0 (0/29) | ||
|
Asian | 2 (1/59) | 3 (1/30) | 0 (0/29) | ||
|
More than 1 race | 3 (2/59) | 7 (2/30) | 0 (0/29) | ||
|
Unknown race | 3 (2/59) | 0 (0/30) | 7 (2/29) | ||
|
Hispanic | 7 (4/59) | 3 (1/30) | 10 (3/29) | .58 | |
Married, % (n/N) | 44 (26/59) | 40 (12/30) | 48 (14/29) | .71 | ||
Working, % (n/N) | 46 (27/59) | 40 (12/30) | 52 (15/29) | .52 | ||
High school or less education, % (n/N) | 29 (17/59) | 17 (5/30) | 41 (12/29) | .07 | ||
Lesbian, Gay, or Bisexual % (n/N) | 22 (13/59) | 20 (6/30) | 7 (24) | .95 | ||
|
||||||
|
Positive depression screen, % (n/N) | 73 (43/59) | 70 (21/30) | 76 (22/29) | .83 | |
|
Positive anxiety screen, % (n/N) | 39 (23/59) | 30 (9/30) | 48 (14/29) | .24 | |
|
Internalized shamec, mean (SD) | 11.3 (4.2) | 11.5 (3.8) | 11.2 (4.7) | .77 | |
|
Cancer-related stigmad, mean (SD) | 30.1 (11.9) | 31.2 (10.4) | 28.9 (13.4) | .47 | |
|
||||||
|
|
|
||||
|
|
Lung | 36 (21/59) | 30 (9/30) | 41 (12/29) | |
|
|
Breast | 17 (10/59) | 20 (6/30) | 14 (4/29) | |
|
|
Skin | 7 (4/59) | 10 (3/30) | 3 (1/29) | |
|
|
Cervical | 5 (3/59) | 3 (1/30) | 7 (2/29) | |
|
|
Colorectal | 3 (2/59) | 3 (1/30) | 3 (1/29) | |
|
|
Leukemia | 3 (2/59) | 0 (0/30) | 7 (2/29) | |
|
|
Non-Hodgkin lymphoma | 3 (2/59) | 0 (0/30) | 7 (2/29) | |
|
|
Pancreatic | 3 (2/59) | 7 (2/30) | 0 (0/29) | |
|
|
Esophageal | 2 (1/59) | 3 (1/30) | 0 (0/29) | |
|
|
Liver | 2 (1/59) | 3 (1/30) | 0 (0/29) | |
|
|
Prostate | 2 (1/59) | 0 (0/30) | 3 (1/29) | |
|
|
Stomach | 2 (1/59) | 0 (0/30) | 3 (1/29) | |
|
|
Throat | 2 (1/59) | 3 (1/30) | 0 (0/29) | |
|
|
All others | 14 (8/59) | 17 (5/30) | 10 (3/29) | |
|
|
|
||||
|
|
0 | 11 (5/47) | 5 (1/21) | 15 (4/26) | |
|
|
I | 36 (17/47) | 48 (10/21) | 27 (7/26) | |
|
|
II | 30 (14/47) | 19 (4/21) | 38 (10/26) | |
|
|
III | 11 (5/47) | 5 (1/21) | 15 (4/26) | |
|
|
IV | 13 (6/47) | 24 (5/21) | 4 (1/26) | |
|
Months since initial diagnosis, mean (SD) | 4.7 (3.5) | 4.2 (3.7) | 5.3 (3.3) | .23 | |
|
|
|||||
|
|
Chemotherapy | 51 (21/41) | 48 (10/21) | 55 (11/20) | .87 |
|
|
Radiation | 41 (17/41) | 38 (8/21) | 45 (9/20) | .90 |
|
|
Surgery | 44 (18/41) | 38 (8/21) | 50 (10/20) | .65 |
|
|
Hormone therapy | 12 (5/41) | 14 (3/21) | 10 (2/20) | >.99 |
|
|
Stem cell transplant | 0 (0/41) | 0 (0/21) | 0 (0/20) | >.99 |
|
|
Immunotherapy | 0 (0/41) | 0 (0/21) | 0 (0/20) | >.99 |
|
||||||
|
Fagerstrom Test of Nicotine Dependence score, mean (SD) | 5.3 (2.2) | 5.4 (2.1) | 5.1 (2.4) | .69 | |
|
High nicotine dependence (FTND ≥6), % (n/N) | 51 (30/59) | 57 (17/30) | 45 (13/29) | .52 | |
|
Smokes more than half a pack of cigarettes per day, % (n/N) | 56 (33/59) | 53 (16/30) | 59 (17/29) | .88 | |
|
Smoked for 10 or more years, % (n/N) | 92 (54/59) | 93 (28/30) | 90 (26/29) | .97 | |
|
Used electronic cigarettes at least once in the past month, % (n/N) | 29 (17/59) | 37 (11/30) | 21 (6/29) | .29 | |
|
Made at least one attempt to quit smoking in the past 12 months, % (n/N) | 60 (35/58) | 69 (20/29) | 52 (15/29) | .28 | |
|
Number of attempts to quit smoking in the past 12 months, mean (SD) | 2.1 (3.3) | 2.6 (4.0) | 1.6 (2.3) | .25 | |
|
Confidence of being smoke-free, mean (SD) | 71.7 (25.7) | 70.7 (26.5) | 72.8 (25.3) | .76 | |
|
||||||
|
Number of close friends who smoke, mean (SD) | 2.1 (1.8) | 2.3 (1.8) | 2.0 (1.9) | .53 | |
|
Number of adults at home who smoke, mean (SD) | 1.4 (1.1) | 1.5 (1.2) | 1.3 (1.1) | .52 | |
|
Living with partner who smokes, % (n/N) | 32 (19/59) | 37 (11/30) | 28 (8/29) | .64 | |
Heavy alcohol drinker, % (n/N) | 5 (3/57) | 11 (3/28) | 0 (0/29) | .22 |
aSample size, unless otherwise indicated in the cell.
b
cInternalized shame scores range from 5 to 20.
dCancer-related stigma scores range from 9 to 45.
e
fNumbers shown indicate that not all participants provided the stage of cancer.
Summarizing from
Primary and secondary study outcomes.
QuitGuide (N=30)a | Quit2Heal (N=29)a | |||
|
||||
|
Logged in at least once, % (n/N) | 97 (29/30) | 93 (27/29) | .43 |
|
Number of times the app was openedc,d, Mean (SD) | 6.1 (5.3)e | 10.0 (14.4)f | .33 |
|
Time spent per log in (in min) | 2.7 (2.1)e | 3.9 (3.2)e | .07 |
|
Number of days from the first use to the last usec, Mean (SD) | 19.8 (22.6)e | 25.1 (19.8)f | .32 |
|
||||
|
Satisfied overallg | 65 (17/26) | 90 (19/21) | .047 |
|
Will recommend to a friend | 57 (16/28) | 74 (17/23) | .21 |
|
App was made for someone like to meg | 62 (16/26) | 86 (18/21) | .04 |
|
||||
|
30-day quit rate, using all available outcome datad | 7 (2/29) | 20 (5/25) | .10 |
|
30-day quit rate, missing outcomes coded as smokingd | 7 (2/30) | 17 (5/29) | .17 |
|
||||
|
Change in internalized shame | 0.2 (3.5)i | −0.5 (4.7)j | .27 |
|
Change in cancer-related stigma | −1.3 (8.8)i | −3.0 (9.9)k | .48 |
|
Change in depression score | −0.9 (6.5)l | −3.5 (5.0)k | .38 |
|
Change in anxiety score | 0.0 (5.6)m | −2.8 (6.8)j | .56 |
aSample size, unless otherwise indicated in the cell.
bTwo-sided
cApp opening data are limited to a subset of participants for whom the objective utilization data were available. Owing to a technical error, automatic recording of the utilization data began 2 months after the beginning of the trial recruitment period.
dRegression model was adjusted for high school or less education because of its association with the outcome and slight imbalance between arms.
eN=15.
fN=17.
gResponses were dichotomized as “somewhat,” “mostly,” or “very much” versus “not at all” or “a little.”
hProcess indicators were calculated as follow-up score minus baseline score.
iN=29.
jN=25.
kN=24.
lN=27.
mN=28.
The self-reported 30-day point prevalence quit rate for those who completed the 2-month follow-up was 20% (5/25) for Quit2Heal versus 7% (2/29) for QuitGuide (odds ratio [OR] 5.16, 95% CI 0.71-37.29;
From baseline to the 2-month follow-up, Quit2Heal participants also reported greater improvement in internalized shame, cancer stigma, depression, and anxiety, although none of these changes were significant (all
The use of outside treatments to quit smoking during the 2-month study period did not differ by study arm: nicotine patch (20% for Quit2Heal vs 31% for QuitGuide;
This paper reports the trial recruitment, retention, participant acceptability, preliminary effectiveness, and impact on processes of change of the first-known smoking cessation smartphone app targeted for cancer patients. In general, the results supported all the aims of the pilot study.
The trial provided useful information on recruitment sourcing and budgeting. Social media (primarily Facebook Ads) yielded 66% (39/59) of the study sample, which showed that it was an effective method for recruiting cancer patients who smoke for an mHealth intervention trial. Recruitment of this population via Facebook was worth the investment because it was highly useful for recruitment, and an intervention, if proven effective, could be broadly disseminated through Facebook given its high reach to this population.
The outcome data collection protocol yielded a strong overall retention rate of 92%, which is consistent with our past experience with this protocol [
The recruitment methods yielded demographics broadly representative of adult cancer patients. There was variability of cancer diagnoses, including cancer diagnoses not typically thought of as caused by smoking but whose treatment would greatly benefit from quitting smoking (eg, breast cancer). In a fully powered trial, it would be worth exploring whether patients with cancers known to be attributable to smoking (eg, lung/head and neck cancers vs patients with all other cancers) are more likely to respond to an app targeted to cancer patients who smoke. The rates of inclusion of participants with mental health problems (eg, depression) and participants who belonged to racial or ethnic minority, were male, identified themselves as LGB, and had high school or less education were encouraging. These sociodemographic groups are typically underrepresented in eHealth and mHealth smoking cessation research [
Although not statistically significant, among those with available log-in data, Quit2Heal participants opened their app more often than QuitGuide participants. Quit2Heal participants were highly satisfied with their app on multiple indicators—substantially more than the QuitGuide participants. Particularly encouraging was the finding that, as an app targeted for cancer patients, 86% of the assigned study participants rated Quit2Heal as being made for someone like them. Taken together, these results suggest that the Quit2Heal content was engaging, acceptable, and seen as relevant by cancer patients.
Tests of the encouraging quit rates were underpowered as this was a pilot trial. Indeed, the 95% CIs for the ORs for the comparison of quit rates were wide, which is expected in pilot trials with low sample sizes. However, if similar quit rates are found in a fully powered RCT, the overall effect size could have high public health significance.
The results from Quit2Heal on improvements in internalized shame, cancer stigma, depression, and anxiety are important. They suggest that Quit2Heal may have impacted the processes hypothesized to impede smoking cessation among cancer patients. A future larger trial can determine the extent to which these processes mediate the effects of Quit2Heal on smoking cessation.
The study has several important limitations. As a pilot randomized trial, the sample size was not powered to detect statistically significant differences in quit rates or to conduct formal moderation or mediation analysis of the hypothesized treatment effects. Moreover, substantial smoking relapse naturally occurs after a 2-month follow-up [
The study results suggest 3 main lines of future research: (1) provide a definitive test of the effectiveness of smoking cessation of smartphone-delivered Quit2Heal compared with QuitGuide—an app that follows US Clinical Practice Guidelines, (2) demonstrate that the smoking cessation outcomes of Quit2Heal are mediated by processes that impede cancer patients’ cessation (eg, internalized shame and cancer stigma), and (3) explore the baseline moderators of treatment effectiveness.
In a pilot trial with a high short-term follow-up rate, Quit2Heal showed promising acceptability and effectiveness for helping cancer patients stop smoking. Testing in a full-scale RCT is required to definitively determine the effectiveness of Quit2Heal for smoking cessation.
CONSORT-EHEALTH checklist (V 1.6.1).
Cancer Center Cessation Initiative
Completely Automated Public Turing test to tell Computers and Humans Apart
Center for Epidemiologic Studies
electronic health
Food and Drug Administration
Fagerstrom Test of Nicotine Dependence
Generalized Anxiety Disorder-7
lesbian, gay, or bisexual
mobile health
National Comprehensive Cancer Network
National Cancer Institute
odds ratio
randomized controlled trial
Seattle Cancer Care Alliance
television
The authors thank the SCCA team members, especially Katie Brown, for assisting with the Quit2Heal content development. The authors thank Daniella Kim and Anna Serra for their work on the Quit2Heal app design and user testing. The authors also acknowledge Seattle-based Moby Inc for their development work, Eric Meier and his team at Datatope for their online recruitment and database development, and the cancer patients who volunteered for the study. This study was funded by Consumer Value Store Health Foundation, the NCI (R01CA166646 and R01CA192849), and the National Institute on Drug Abuse (R01DA038411). The funders had no role in the study.
In the past, JB has served on the Scientific Advisory Board of Chrono Therapeutics. JH has received research support from Pfizer. Other authors have no declarations.