This is an open-access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in the Journal of Medical Internet Research, is properly cited. The complete bibliographic information, a link to the original publication on http://www.jmir.org/, as well as this copyright and license information must be included.
Tobacco smoking leads to death or disability and a drain on national resources. The literature suggests that cigarette smoking continues to be a major modifiable risk factor for a variety of diseases and that smokers aged 18-30 years are relatively resistant to antismoking messages due to their widely held belief that they will not be lifelong smokers.
To conduct a randomized controlled trial (RCT) of a computer-generated photoaging intervention to promote smoking cessation among young adult smokers within a community pharmacy setting.
A trial was designed with 80% power based on the effect size observed in a published pilot study; 160 subjects were recruited (80 allocated to the control group and 80 to the intervention group) from 8 metropolitan community pharmacies located around Perth city center in Western Australia. All participants received standardized smoking cessation advice. The intervention group participants were also digitally photoaged by using the Internet-based APRIL Face Aging software so they could preview images of themselves as a lifelong smoker and as a nonsmoker. Due to the nature of the intervention, the participants and researcher could not be blinded to the study. The main outcome measure was quit attempts at 6-month follow-up, both self-reported and biochemically validated through testing for carbon monoxide (CO), and nicotine dependence assessed via the Fagerström scale.
At 6-month follow-up, 5 of 80 control group participants (6.3%) suggested they had quit smoking, but only 1 of 80 control group participants (1.3%) consented to, and was confirmed by, CO validation. In the intervention group, 22 of 80 participants (27.5%) reported quitting, with 11 of 80 participants (13.8%) confirmed by CO testing. This difference in biochemically confirmed quit attempts was statistically significant (χ2
1=9.0,
Demonstrating the detrimental effects on facial physical appearance by using a computer-generated simulation may be both effective and cost-effective at persuading young adult smokers to quit.
Australian New Zealand Clinical Trials Registry: ACTRN12609000885291; https://www.anzctr.org.au/Trial/Registration/TrialReview.aspx?ACTRN=12609000885291 (Archived by WebCite at http://www.webcitation.org/6F2kMt3kC)
Tobacco smoking leads to premature death or morbidity and places a drain on national resources. Consequently, health professionals and governments stress the importance of smoking cessation and a reduction in exposure to tobacco smoking [
The younger people are when they start smoking, the greater the risk of illness or death caused by smoking [
Currently in Australia, 19.7% of males and 16.3% of females aged 20 to 29 years smoke on a daily basis [
The objectives of this randomized controlled trial (RCT) were to test the efficacy and cost-effectiveness of an intervention based on personalized, vivid illustrations of “smoker’s face” among young smokers (18-30 years of age). Smokers face includes wrinkling of the face, gauntness of facial features, and a gray and plethoric complexion. Efficacy was assessed by comparing successful quitting, number of quit attempts, and change in smoking dependence (assessed by the Fagerström score) between the intervention and control groups. The study also aimed to explore the value (feasibility and cost) of an unfunded intervention within pharmaceutical practices.
This study was a RCT (Trial ID number: ACTRN12609000885291) that recruited 160 participants (80 participants assigned to both control and intervention groups) from 8 metropolitan community pharmacies located geographically around Perth city center, Western Australia, when presenting to collect prescribed medications or over-the-counter (OTC) medications.
Eligibility criteria included (1) age range of 18-30 years old (self-report); (2) smoker (defined as smoking 1 or more cigarettes per day via self-report); (3) able to give consent; (4) available for follow-up at 6 months; (5) no beards, mustaches, or non-removable facial accessories; (6) no body dysmorphia (participants screened using the Body Dysmorphic Disorder Questionnaire [BDDQ)]) [
The sample size of 80 participants per group was calculated to observe a medium effect size (
The APRIL Face Aging software is an Internet-based 3-dimensional age progression software package that creates a stream of aged images of faces from a standard digital photograph (the wrinkling/aging algorithms based upon normative data from people of a broad variety of ages, ethnicities, lifestyle habits, as well as published data regarding facial changes associated with aging). Additionally, the resulting aged images can be adjusted to compare how a person will age as a smoker versus as a nonsmoker (
Current age photo (25 years) and future digitally aged photo (65 years) of a female current smoker.
Digitally aged photos of female participant at 65 years as a nonsmoker (left) and as a smoker (right).
At recruitment, all participants were asked to complete a baseline questionnaire consisting of demographic data, the Fagerström Smoking Dependence Scale (score from 0-10) [
Participants in the intervention group were also screened for body dysmorphia using the BDDQ. In addition, they were photographed and their images were digitally aged as both a smoker and a nonsmoker (using the Internet-based APRIL Face Aging software), and invited to view the age-processed images (
If participants stated at the 6-month follow-up that they had quit smoking, they were required within 48 hours to undertake a carbon monoxide (CO) breath test to validate their nonsmoking status. The CO monitor utilized was a portable, battery-operated Smokerlyzer (Bedfont Scientific Ltd, Kent, England, UK) that provided a CO level reading in parts per million (ppm).
Intervention being delivered by pharmacist.
The primary outcomes measured were (1) the effect of the intervention by using successful quitting, quit attempts, and progression along the transtheoretical stages of change model, and (2) nicotine dependence using the Fagerström scale. These were measured at baseline and at 1-, 3-, and 6-month follow-ups.
The demographic and baseline smoking habit profiles of the recruited participants were compared between groups using Fisher’s exact test and Pearson’s chi-square test for categorical variables, and the Student’s
The secondary outcomes measured were (1) the cost-effectiveness of the intervention from a health sector perspective in terms of the incremental cost per additional quitter and per additional lifetime quitter, and (2) the business viability of delivering the intervention in a community pharmacy. These were calculated at the conclusion of the study.
Two perspectives were adopted: a health sector perspective and the perspective of a community pharmacy on the assumption that the intervention was not government funded. The direct costs of providing the digital aging service over and above providing standard cessation advice were calculated based on the time taken to provide the service and the cost to a pharmacy of purchasing tokens to use the online software to photoage participants. The cost of a pharmacist’s time was valued based on a published recommended rate of pay in Western Australia [
To assess the robustness of the study results, a scenario sensitivity analysis with the best-case and worst-case scenarios were performed [
In the best-case scenario, the pharmacist’s time was adjusted down by 25%, the exchange rate for converting American dollars to Australian dollars was varied to the lowest level in the past 5 years, and a discount rate of 0% was used [
Parameter values for base case and sensitivity analyses.
Itema | Base case | Scenario sensitivity analysis | |
|
|
Best case | Worst case |
Pharmacist time per participant to deliver service (mins) | 4.8 | 3.6 | 6.0 |
Award wage rate per week for a pharmacist (AU$) | 907.40 | - | - |
Cost of a token (AU$) | 3.87 | 3.63 | 6.53 |
Exchange rate (AU$) | 0.9687 | 0.9067 | 1.6321 |
Discount rate (%) | 3 | 0 | 5 |
a Compared to US $1.
Customers were screened for eligibility to the RCT from 8 community pharmacies (
In total, 1259 customers were screened for eligibility; 213 customers were eligible and 160 were recruited, the others declined the invitation to participate for a range of very different reasons. Eighty participants were recruited to the control group and 80 to the intervention group.
The smoker’s face simulations were created using a digital photograph (6.0 megapixels) taken of the intervention participants and uploaded to APRIL Face Aging software version 2.5 on a laptop computer.
The RCT was conducted between January 2010 and December 2010 and all follow-up surveys were completed by June 2011. The final 6-month follow-up showed a response rate of 77.5% for the control group and 72.5% for the intervention group. The demographic and baseline smoking behaviors of recruited participants are shown and compared between groups (intervention versus control) in
There were more females and lighter smokers (smoking up to 5 cigarettes per day) in the intervention group; however, there were no statistically significant differences between the control and intervention groups on demographic or smoking dependence variables at baseline. No participants were revealed to have body dysmorphia. A number of questions on the survey were designed to gather the respondents’ opinions of self-perceptions and their attitudes toward their smoking behavior. These questions were taken from an earlier survey [
Profile of the randomized controlled trial (using CONSORT guidelines).
Demographic and baseline smoking profile of study participants (N=160).
Variable | Control group |
Treatment group |
|
|
|
|
|
|
|
|
Male | 35 (43.8) | 25 (31.3) | .10 |
|
Female | 45 (56.2) | 55 (68.7) |
|
Age, mean (SD) | 25.1 (4.1) | 24.2 (4.1) | .16b | |
|
|
|
.71 | |
|
Year 10 high school | 15 (19.0) | 17 (21.3) |
|
|
Year 12 high school | 31 (39.2) | 29 (36.3) |
|
|
Technical and further education (TAFE) qualifications | 17 (21.5) | 22 (27.5) |
|
|
Degree (university/college) | 16 (20.3) | 12 (15.0) |
|
|
|
|
.35 | |
|
1 | 11 (13.8) | 19 (23.8) |
|
|
2-5 | 9 (11.3) | 10 (12.5) |
|
|
6-10 | 21 (26.3) | 14 (17.5) |
|
|
11-20 | 27 (33.8) | 29 (36.3) |
|
|
>21 | 12 (15.0) | 8 (10.0) |
|
Fagerström score, mean (SD) | 2.96 (2.52) | 2.87 (2.48) | .82b | |
|
|
|
.92 | |
|
0-2 | 39 (48.8) | 39 (49.4) |
|
|
3-4 | 19 (23.8) | 18 (22.8) |
|
|
5 | 8 (10.0) | 10 (12.7) |
|
|
6-7 | 10 (12.5) | 10 (12.7) |
|
|
8-10 | 4 (5.0) | 2 (2.5) |
|
a From chi-square test (unless otherwise marked) comparing the treatment groups.
b From
A similar model using confirmed quitting as the dependent variable showed an adjusted
Pattern of survey completion and change in smoking behavior at 6-month follow-up (N=160).
Variable | Control group |
Treatment group |
|
|
|
|
|
|
|
|
All surveys completed | 56 (70.0) | 48 (60.0) | .38 |
|
|
|
|
|
|
6 month | 6 (7.5) | 10 (12.5) |
|
|
3 month | 8 (10.0) | 14 (17.5) |
|
|
1 month | 3 (3.8) | 4 (5.0) |
|
|
No follow-up | 7 (8.8) | 4 (5.0) |
|
|
|
|
|
|
|
Self-report (questionnaire) | 5 (6.3) | 22 (27.5) | <.001 |
|
Confirmed (CO-validated) | 1 (1.3) | 11 (13.8) | .003 |
|
|
|
<.001b | |
|
Reduced dependence | 11 (13.8) | 41 (51.3) |
|
|
No change | 68 (85.0) | 39 (48.8) |
|
|
Increased dependence | 1 (1.3) | 0 |
|
|
|
|
<.001c | |
|
At 1-month follow-up | –0.14 | –0.83 |
|
|
At 3-month follow-up | –0.38 | –1.34 |
|
|
At 6-month follow-up | –0.26 | –1.88 |
|
a From chi-square tests unless otherwise specified.
b From Fisher’s Exact test.
c Obtained from a repeated measures analysis including all available surveys.
A random effects regression model was used to model the mean change in Fagerström score from baseline by using data from all follow-up surveys The control group did not experience a significant drop in Fagerström score over the study (
Although there were no differences between participants at baseline, the regression models were extended to adjust for the gender and age of the participant, and the number of cigarettes smoked at baseline. The models were fitted to the control and intervention group separately because it was clear that changes in score appeared only in the intervention group. For the control group, there were no associations between change in score and age (
Total costs of implementing the intervention from a health sector perspective were AU $463, or the equivalent of AU $5.79 per participant (
Economic analysis of photoaging service.
Item | Base case | Scenario sensitivity analysis | ||
|
|
Best case | Worst case | |
Mean cost per participant of service (AU$) | 5.79 | 5.07 | 8.93 | |
Total cost of servicea (AU$) | 463 | 406 | 714 | |
|
|
|
|
|
|
Cost per additional quitter (AU$) | 46 | 41 | 71 |
|
Cost per additional lifetime quitter (AU$) | 74 | 64 | 113 |
Cost offset from reduction in health care costs (AU$) | 2144 | 2660 | 1867 | |
Net total cost savings (AU$) | 1778 | 2346 | 1316 | |
|
|
|
|
|
|
Mean (SD) | 20.25 (15.32) | - | - |
|
Median (IQR) | 20.00 |
- | - |
a For all 80 participants.
With an additional 10 quitters confirmed in the intervention group compared to the control group (11 and 1, respectively), the incremental cost-effectiveness ratio (ICER) was AU $46 per additional quitter, or the equivalent of AU $74 per additional lifetime quitter. Cost offsets of AU $2144 from a reduction in the health care costs of quitters resulted in the intervention potentially generating net total cost savings of AU $1778. In the best-case scenario, the ICER was AU $41 per additional quitter and net total cost savings were AU $2346. Corresponding figures for the worst-case scenario were AU $71 per additional quitter and AU $1316, respectively.
The mean cost that the participants indicated that they were willing to pay for the digital aging service was AU $20.25, which exceeded the mean cost per participant for delivering the service (AU $5.79). The median cost they were willing to pay was AU $20, similar to the mean value. More than 80% of participants said they would be more likely to use the pharmacy to purchase future smoking cessation therapies and to use it more for other purchases. More than 80% of participants also thought their friends would be willing to pay for the service, and all but 2 participants said they would recommend the photoaging intervention to 1 or more friends who were smokers.
The impact of the photoaging innovation on confirmed quit attempts by the young people recruited to this study was statistically significant (
The pharmacies selected to take part in the study were chosen to cover a range of socioeconomic areas and the equal number of study participants selected for each treatment group at each pharmacy aimed to diminish any potential biases. However, because of the nature of the intervention, the participants and researcher could not be blinded to the study group. Allocation to groups was not performed as eligible participants were recruited, but according to the treatment being used at the pharmacy during that week. In this setting, there was a substantial risk of contamination between treatment and control groups if participants had been randomized at the point of recruitment rather than by week of attendance at the pharmacy.
The baseline comparisons showed that the 2 groups were very similar on smoking dependence scores, and the 6-month follow-up response rate was high (over 70% for both groups). Follow-up to 12 months may have been preferable, but impractical, in this case. However, follow-up at 6 months was augmented by biochemical verification of tobacco use and cessation [
Although there were more females and light smokers in the intervention group, this was not statistically significant and appeared not to diminish the significant statistical association between treatment group and quitting smoking.
Although many individualized smoking cessation interventions have been implemented in the past few decades, few have had as marked an impact as reported here. With the advent of digital technology, quit messages can now be delivered by mobile telephone, email, text messaging, and online social networks [
To date, there have been few studies reporting on a personalized photoaging intervention [
The economic analysis demonstrates that this personalized smoking cessation intervention is cheap and cost-effective, and it could be readily adopted in community pharmacies. It targets young smokers who are at significant risk of adverse effects of smoking if they continue lifelong smoking.
With an ICER of AU $74 per additional lifetime quitter, the intervention is cost-effective compared with other individualized smoking cessation programs. For example, a systematic review of economic evaluations of a range of smoking cessation interventions found ICERs of between US $260 to US $3263 per lifetime quitter (2002) for counseling or self-help programs versus usual care [
A review commissioned by the Australian Commonwealth Department of Health and Ageing concluded that interventions delivered by health care providers significantly increased the number and success of quit attempts made in Australia each year [
Health care providers, such as pharmacists, are accessible and highly trained [
A significant development for the use of mass media in delivering antismoking messages is the advent of digital technology. Technologies such as the Internet, social networking sites, and smartphones have the potential to reach large populations of younger people [
Further experimental research deploying photoaging technology is needed to answer these questions.
CONSORT-EHEALTH checklist V1.6.2 [
Body Dysmorphic Disorder Questionnaire
carbon monoxide
incremental cost-effectiveness ratio
nicotine replacement therapy
over-the-counter
randomized controlled trial
willingness to pay
Acknowledgement for use of progression software: MJ, OB, and OC thank APRIL Face Aging software for supplying the progression software v2.5 license per gratis for the duration of the study. MJ, OC, and OB conceptualized and designed the study; OB acquired all data; OB, RP, DH, OC, and MJ drafted the manuscript; RP performed all statistical data analyses; DH performed all economic analyses; MJ, OC, RP, DH, and OB critically reviewed the manuscript for intellectual context.
None declared.