School-based behavioral intervention to reduce the habit of smokeless tobacco and betel quid use in high-risk youth in Karachi: A randomized controlled trial

There have been recent surges in the use of smokeless tobacco (SLT) and betel quid (BQ) chew among adolescents in South East Asian countries, with an increase, on average, of 7% to 15% between 2004 and 2013, necessitating interventional investigations to modify this behavior. The current intervention was aimed towards changing adolescents’ perceptions regarding the harmful effects of SLT and BQ use and encouraging them to quit. This randomized control trial involved 2140 adolescents from 26 private and public-sector schools in Karachi, Pakistan. After randomization, 1185 individuals were placed in the intervention group and administered a behavior changing intervention (BCI), while 955 individuals constituted the control group. A generalized estimating equation was employed to measure differences in repeated measures for both groups. The beta coefficients were reported after adjusting the covariates with the 95% confidence interval, and the p-value was considered significant at <0.050. Cohen’s d was employed to report the effect size of the intervention. The BCI resulted in a 0.176-unit (95% CI 0.078–0.274, p-value <0.001) increase in knowledge scores regarding the health hazards of SLT and BQ, a 0.141-unit (95% CI 0.090–0.192, p-value <0.001) increase in use perception scores, and a 0.067-unit (95% CI 0.006–0.129, p-value 0.031) increase in quit perception scores in the intervention group compared with those in the control group. A knowledge related module (p-value 0.024) and quit preparation module (p-value 0.005) were found to be helpful by adolescents in either changing their perceptions regarding SLT and/or BQ chew use or in quitting. The role of BCI is promising in improving adolescents’ knowledge and changing their perceptions in a positive manner regarding their harmful SLT and BQ use. Convincing results may be achieved if interventions are tailored, with an emphasis on the identification of the products that are used by adolescents in addition to highlighting their ill effects and how students may manage to quit them. If included in the schools’ curricula, this BCI method may help in developing schools that are free of SLT and BQ use. Trial registration: ClinicalTrials.gov NCT03488095.


Background
Smokeless tobacco (SLT) and Betel Quid (BQ) are known risk factors for Oral, Pharyngeal and Esophageal carcinoma. Three-quarters global SLT consumption is among the South Asian Population. The habit of SLT and BQ chewing commences at a very young age which has an underpinning of socio-cultural dimension in South Asian population which perhaps due to its more addictive potential, has more dependency. Youth also perceives SLT and BQ as a part of confectionery, which is socially served in South Asian ceremonies. There exists very little or no evidence regarding efficacy of SLT and BQ cessation interventions in this population. Most of the existing interventions are based in western world with little or no cultural sensitivities pertinent to South Asia. We developed a culturally rooted behavior changing intervention (BCI) to alter SLT use prevalence, perceptions pertinent to the deleterious effects of SLT and BQ use in the etiology of oral cancers and help youth quit SLT and BQ in Karachi, Pakistan.

Methods
Baseline demography and SLT & BQ use prevalence will be ascertained among 11-16-year-old school going children from both government and private schools. Cluster sampling will randomly recruit 26 clusters (schools) from within 6 districts of Karachi, which will then be divided into intervention and control groups (clusters) using block randomization based on proportionate number of each school type present (Government and Private).
Both groups' participants will complete all questionnaires pre and post intervention as described elsewhere and will also undergo screening for oral cancer and oral potentially malignant lesions (OPMLs). Students in intervention cluster will be given BCI and printed pamphlets along with a gift pack (reminder for SLT and BQ quit, a branded tooth paste and a tooth brush) while students in control cluster will only differ in that they will not be exposed to BCI. BCI was designed after reviewing literature and consulting specialist group for all untoward effects of SLT and BQ use in Oral Cavity that have a potential of transforming into oral carcinoma. Identification of all determinants of SLT and BQ use was fundamental for the identification of behavior changing techniques (BCTs) which can possibly modify the behavior of SLT & BQ users and help youth in quitting. The BCTs were also sensitively incorporated into BCI by keeping South Asian culture as a backdrop. The accuracy of BCI will be assessed by fidelity index modified to score adherence and quality of the components of BCI.
A follow up after 12 weeks will be conducted to re-assess their SLT and BQ use prevalence, perceptions regarding hazardous effects of SLT & BQ use in oral cavity, dependency on SLT & BQ and success in quit (among users in both groups), and perception regarding warning labels on SLT & BQ product packet.

Key words
Tobacco, smokeless, chewing, betel quid, South Asian, behavioral support, behavior change, behavior changing technique, fidelity index, cessation, Oral cancer, adolescents, screening 2 Introduction 2.1 Significance of the problem Smokeless tobacco(SLT) is a widely used term that includes varying types of tobacco products that are used orally or put in nose (1). It includes betel quid (contains areca nut with or without tobacco), ghutka, naswar, zarda, mawa, snuff (both moist and dry), chewing tobacco and many other dissolvable tobacco products (2). SLT is consumed worldwide in many forms which may include unprocessed that is dried in sunlight or processed products, that are either chewed, snuffed, sucked or placed in the oral cavity between lips and buccal mucosa (3).
In approximately 115 countries, more than 300 million adults consume SLT in various forms(4); the majority of the consumers (89%) are concentrated in South Asian countries (5). In South Asian Countries, SLT use prevalence is more in males than females; less than one third of female population consume SLT in Mauritania (28.3%), and India (18.4%) whereas, more than one third of male population of Myanmar (51.4%), and India (32.9%) consume SLT. Within Europe, SLT in the form of Snus is more commonly consumed with similar trend of higher consumption in males (24% and 20% in Sweden and Norway respectively) as compared to females which is 7% and 6% in Sweden and Norway respectively (4). In Pakistan, the regular consumption of SLT is 16.3% and 2.4% amongst males and females respectively (6). In school going adolescents; according to a study based on GYTS (global youth tobacco survey)(7), 18.5% boys and 8.4% girls chew SLT that is similar to the more male and less female consumption trend as in adults. According to another study in India, within 10 % SLT use prevalence 80.9% users were males (8). In Karachi, school going adolescents of 13-15 years; according to GYTS-2008, 16.7% boys and 9.6% girls consume SLT (9).
SLT consumption is accepted in South Asian countries as part of their cultural and social norms (10). At a very young age, individuals start consuming it as they are readily available and considered as sweets (11). A variety of factors and list is not exhaustive to family values, easy access, cheap in cost, as dental pain relieving medicine, no enforcement of legislation on its use, are responsible for SLT wide spread use (12,13). The women and young individuals in India, Pakistan and Bangladesh refrain from smoking due to cultural and traditional beliefs but there is no such hesitation in SLT use; rather parents encourage their children to use it (14).
Considerably variable modes of marketing like in Indian cinema, celebrities use these ghutka products as something very "cool" and then it is sold in very attractive packaging. The youth considers it as very popular, acceptable product and subsequently becomes addictive to it (15). School going children also get fascinated by the undesirable marketing of SLT products and spend the substantial amount of their pocket money on the purchase of these dangerous substances (8,16). Majority of the users get engaged in this habit even before 15 years of age due to influence by media (40.2%) and family & friends pressure (30.8%)(17) while in another study, peer pressure (76%) was reported (8). In a study by Khan A. (18); in India, Pakistan, Nepal and Bangladesh, there are either no policies to control sell and purchase of SLT or they are very inadequate. Once adolescents start consuming SLT, it becomes an essential component of their lives since SLT is the fourth commonly abused substances used worldwide and develop dependency after alcohol, nicotine & caffeine with both stimulating and relaxing effect (19)(20)(21).

Statement of the problem
SLT contains nicotine and carcinogenic nitrosamines which cause more than a quarter million deaths mainly due to oral and pharyngeal cancers (4). Betel Quid(pan) with or without tobacco, areca nut (supari, chalia), ghutka, niswar, tobacco smoking in all forms are proven oral carcinogens by IARC (International Agency for Research on Cancer) (22). Areca nut alone, in formulation as betel quid and other SLT chewable are proven oral cavity carcinogens (19,22) and are chewed by 600 million people worldwide specially in South Asia (23). SLT is also a risk factor for oral potentially malignant lesions like oral sub mucus fibrosis, leukoplakia, dysplastic mucosal changes, and oral ulcers(3).
Oral cancer incidences are increasing with decreasing survival (24) rates to approximately below 50% despite of extensive treatments given at tertiary care hospitals. The World Health Organization (WHO) reports that the trend of such cancers would increase further in South Asian countries where incidences of oral cancers are more as compared to rest of the world (25). In developed world, alcohol and tobacco are established risk factors of oral cavity tumors (26). According to a systematic review regarding SLT use and oral cancer in South Asia (27), there is a very profound association between SLT consumption and oral cancer where betel quid with tobacco chewers have seven times more risk of developing oral cancer than non-chewers. Also, the risk of oral cancer in SLT chewers other than betel quid is five times more than in non-chewers (27). In Pakistan, according to Shaukat Khanum cancer registry, in both genders above 18 years of age; lip and oral cavity carcinoma is third most carcinoma (28). Areca nut, betel quid, and tobacco increase the risk of oral cancer by 8.5-10 times as per reported in a study in Pakistan(10).

Objectives
In lieu of the increasing disease burden of oral cancer in South Asian countries due to increasing popularly fashionable SLT consumption; our objective is to develop (i) a BCI pertinent to socio-cultural aspect for South Asian SLT users, followed by (ii) assessing its efficacy in changing perception and dependency regarding SLT use in high risk group of 11-16 years old school going children of Karachi of both private and public sector schools.
We also aim to deconstruct our BCI and establish which components of our BCT were more useful in making an intervention most efficient by using sub-indices of fidelity index to assess the adherence and quality of each component of intervention. Each component will be scored using fidelity index by a neutral trained listener each time intervention is delivered by a trained personnel.
Quantitatively, our objectives of the study are to: i. measure SLT and BQ use prevalence amongst adolescents of 11-16 years ii. assess the reasons behind their use of SLT and BQ in users (both intervention and control clusters) by using Reasons for Betel Quid Chewing Scale (RBCS).
iii. evaluate the dependency on SLT and BQ in users (both intervention and control clusters) using

Fagerstörm Tobacco and Nicotine Dependency scale for Smokeless Tobacco (FTND-ST) and Betel
Quid Dependency by using Betel Quid Dependency Scale (BQDS).
iv. assess perceptions of adolescents (both intervention and control clusters) regarding ill effects of SLT and BQ use in oral cavity causing oral cancer and OPMLs.
v. assess the perception of adolescents regarding oral cancer picture of SLT and BQ product pack in cessation of its use with the help of Behavior Changing Intervention in intervention cluster and a questionnaire item in control cluster.
vi. evaluate each component of Behavior Changing Intervention in its efficacy by using fidelity index and self-perceived efficacy by participants of intervention cluster.

Literature Review
Individuals at a very young age start consuming SLT with a mean age of initiation being 13 years (SD 7.07) which earlier was 15 years, suggestive of an early initiation of SLT in this era (8). Key factors that played role in initiation also play role in quitting decision along with dentists' advice (13). But there is also no adequate support available for people who wish to quit (18). Thus, no support for quit and loneliness lead to relapse (30). The effectiveness of behavioral interventions in helping SLT users in quitting was greatly supported by a Cochrane review but its effect size was limited since there was ambiguity in which part of the intervention helped the most to have its impact (31). It was also concluded in the review that pharmacological interventions like varenicline and nicotinic lozenges may also support quitters in abrupt cessation (31). SLT cessation was effective in one intervention where success was attributed to follow up session along with oral examination (32), this being coherent with a Also, the rationale of incorporating more social and cultural components in BCI in South Asian countries is more justified. Most of the trials designed for SLT cessation are adapted by already conducted trials based on smoking cessation that are insensitive of socio-cultural impact of SLT use (37). A systematic review and meta-analysis strongly suggests that since SLT use has a very strong cultural backdrop attached in South Asian countires and along with scanty awareness regarding deletirious effects on oral health and illiteracy; they are bound to play a signifcant role in BCI to control SLT use amongst high risk group of 13-15 years old (38). Also, since increasing oral cancer trend in south asian countries is because of SLT which is completely avoidable thus it can serve as a lucrative initiating point to control for oral cancer and plan primary preventive interventions (38).  With a population of approximately 16 million, Karachi is the largest city of Pakistan (42). Karachi City government has divided the city into six (6) districts in 2013 with each district being further divided into administrative towns of total 18 in number (43).
Cluster sampling will be done. Out of six districts, twenty six clusters (secondary schools) will be randomly selected proportionate to the number of each school type (ensuring equal participation of government and private schools). The students of 11-16 years of age, studying in grade VI -Grade X will be randomly recruited.

Complete lists of all private and government schools of Karachi have been arranged by requesting respective
Directors (Appendix 6). If any randomly selected school declines to participate than another school will be randomly selected from the same district; also, a government school will be selected if government school will decline and vice versa. Fifty to 100 students per school will be randomly included in the study. Depending upon the size of each class, students will be randomly selected from each of the secondary classes present on the day of our visit till a total number of 50-100 is reached.
This will therefore summit a total sample size of 2200 that may be considered Students of younger or older age group than mentioned already will be excluded and students who already are undergoing treatments for Oral Cancer will be excluded.

Research Design
This study is a "Cluster-Randomized Trial".

Sample Size Calculation
Sample size calculation was based on efficacy of BCI for SLT and BQ use cessation. In openEpi; using two sided significance level at (1-alpha) 95%, an intra-cluster correlation coefficient of 0.05 (to cater for within school influences on cessation) with 90% power to detect a change of minimum 7% over a control cluster for BCI efficacy in intervention arm; 1606 sample size was calculated (9). Adding approx. 25% attrition rate, it was upsized to 2200 participants. Due to inaccessibility to software for cluster randomized trial to us, cluster size and number of clusters were calculated manually using equations which turned out to be 13 clusters in each arm with 50-100 participants in each cluster (45).

Informed consent and Recruitment
For recruitment purposes, we will contact principals of selected schools (both government and private) to provide them with details regarding the intervention, its merits and efficacy, and will request for their inclination to participate in the trial. If any selected school refuses to participate then another school of same profile will be sent an invitation.
Schools' heads will then be requested to send consent forms (will be provide to schools) to the parents along with a covering letter in which they will be provided with all relevant details regarding intervention and will be encouraged to contact principal investigator via a text on given contact number who will then call parents to respond to all queries since we do not have toll-free numbers.
Parents will be requested to sign an acceptance or refusal on the form and sent it back to school by a specific decided date.

Questionnaire (appendices 1 & 2) -Baseline measures
All students finally selected will complete a questionnaire (adapted from GYTS(44)) at the baseline session (appendix 1). The questionnaire includes demographic information (e.g. age, gender, grade in which they are studying), past and present SLT and BQ use history, past smoking history in any form, family history, perceptions regarding adverse health effects of SLT and BQ use, motivation to quit, perception regarding warning labels if pasted on SLT and BQ packets in Pakistan). SLT and BQ users will then be identified to complete questionnaires on reasons to chew betel quid (46) to assess their rationale behind this habit, to assess nicotine dependency Since this BCI is first of its kind at such a mega level in youth of Karachi, Pakistan; a literature review was conducted to assess all untoward effects of using SLT and BQ on oral health specifically, to assess habit forming reasons of using SLT & BQ and to identify potentially malignant lesions (OPMLs) (50).
With an aim to influence youth's perception regarding discontinuing the use of these products; most recent taxonomy of BCTs was adapted to translate these key contributing factors associated with SLT & BQ use, into BCI by focusing more or oral cancer prevention /or early detection (39). This BCI was then designed into an intervention which was culturally sound in South Asian population keeping all cultural sensitivities as a backdrop by a panel of experts who are well aware of designing such activities in South Asian communities. It was essential to be mindful of the cultural sensitivities as for south Asian individuals, it is part of their cultural norms and their parents encourage SLT use. Thus, throughout BCI, harmful effects of BCI are highlighted without stigmatizing its socio-cultural use.
A guide for behavioral support for smoking cessation(51) was kept as a guideline for designing this intervention into pre-quit, quit and post quit management of withdrawal symptoms.

The components of BCI include:
 Identifying the SLT and BQ products, their harmful effects, and why is it important to quit along with setting quit date  Preparation to quit along with managing urges for relapse, and  Recognizing and managing withdrawal symptoms

Intervention and Control Groups (Randomization and concealment)
Once schools are selected; they will be recruited into intervention and control groups by block randomization based on private or public schools so as to control for any differences in socio-economic status as the fee structure in each significantly differs (52) and wealth is directly and positively related to SLT consumption intensity for already users (53). Also, schools (clusters) will first agree to participate and then will be recruited in randomly in each intervention and control groups (33) to avoid any selection bias. However, all intervention material will be delivered to control clusters at the end of the study.
We will recruit clusters (schools) in each intervention and control group to avoid any diminution of the intervention outcome through discussion regarding intervention amongst the participants of the same school.
Gansky and colleagues (33) assigned the failure of their intervention to the "spill over" effect due to the contact between athletes in intervention and control groups.
Both groups' participants will complete all questionnaires pre and post intervention as described elsewhere.

Students in the intervention group will be shown a well-structured visual presentation which will assess their
perceptions regarding SLT and BQ mentioned elsewhere while controls will not be exposed to BCI. The intervention will be delivered by trained personnel to avoid any researcher bias.

All participants will be given a gift hamper comprising of a branded tooth paste (identity hidden to avoid conflict of interest), a tooth brush (identity hidden to avoid conflict of interest) and a pencil with imprints "SAY
NO TO CHALIA AND GHUTKA". They will also be given a quit calendar to keep record of their quit attempt.

Accuracy of intervention delivery (appendix 3)
The accuracy and fidelity to the intervention will be assessed by applying two sub-indices of pre-defined fidelity index. The fidelity index is composed of (i) Adherence Scale, and (ii) quality scale. There are 20 items for adherence scale in our intervention to be assessed by fidelity index and 5 items for quality scale. Each scale has a scoring system composed of three point Likert scale; 0=not implemented, 1=partially implemented, and 2=fully implemented (54). This is adapted from a fidelity index used in a study (55).
This index will be filled up by a trained neutral person each time BCI is delivered in small groups.
The significance of this index is to correlate intervention delivery with the outcome in the form of behavior change regarding SLT and BQ quit and identify which component worked or did not work in attaining the desired modification in the behavior. This method has an advantage of being more objective assessment of fidelity of the intervention than self-reporting (56). Participants' adherence to the intervention will be objectively assessed by the number of quit calendars that we will get back after 12 weeks. They will also help us to assess attrition and thus attrition bias if any.

Quit measures
Self-reported SLT and BQ quit will be measured by quit calendars from users of both control and intervention groups.

Oral examinations
Oral examinations will be performed using conventional screening methods(57) for oral cancer detection and for evaluation of the prevalence of OPMLs in users. This will be performed by trained dentist/dental hygienist to advise SLT and BQ users to quit all hazardous products so as to live a healthy life. Also, it will be performed in the non-users so as to encourage them to remain SLT and BQ free (33). Individuals detected with OPMLs and oral cancers will be referred for further diagnosis and treatment.

Follow up after two weeks
After 2 weeks, BCI related visual presentation will be shown to the students of intervention schools.

Follow up after twelve weeks
After 12 weeks, a follow up visit will ensue to: 1. Re-assess their perceptions regarding SLT and BQ use in users and non-users of both intervention and control clusters.
2. Re-assess their dependency on SLT and BQ in users of both intervention and control clusters.
3. Re-evaluate their perceptions (of users and non-users of intervention cluster) regarding warning labels on SLT and BQ product packet 4. Collect their quit calendars to assess quit from users of both intervention and control clusters, and 5. Collect perceived efficacy of each component of BCI that helped them quit, attempt to quit or change perceptions from all users and non-users of intervention cluster to assess efficacy of each component of BCI.
Any differences in outcome based on BCI will be assessed and documented between two arms of clusterrandomized trial.
The change in perceptions if at all, after 12 weeks, regarding imprinting of pictorial warning labels on SLT and BQ packaging will help us in recommending its implementation at National level to curb SLT & BQ use.

Reasons for Betel quid chewing scale(46) -users of both intervention and control arms
This validated scale is to measure the reasons for betel quid chewing. It comprises of 10-items that will be assessed on five-point scale (0=not important to 4=extremely important).

Betel Quid Dependence Scale (BQDS)(48) -users of both intervention and control arms
The validated BQDS is a 16-item scale which is intended to measure users' dependency on BQ. Each item on the scale marks a binary outcome (No=0 and Yes =1).

Fagerstörm Tobacco and Nicotine Dependency Scale (FTND-ST)(47) -users of both intervention and control arms
This modified and validated scale is a 06-item scale. The total 10 points can be scored on this scale to assess dependency on nicotine in SLT.

Perceptions regarding ill effects of SLT/or BQ use among adolescents (intervention arm)
There are five items that are related to behaviors amongst adolescents regarding harmful effects of SLT and BQ use in the oral cavity. All the harmful consequences are scored (No = 0 and Yes = 1) depending on participants' awareness about them. One item is based on their perception regarding warning label on SLT and BQ product pack (No = 0 and Yes = 1). Pre and post intervention differences in individual scores will be noted to assess efficacy of the BCI.

Quit measures
The participant's attempt in trying to quit, if it is important for them to quit, if they are confident regarding quit and if they are ready to quit, are the 4-items that will be scored (No = 0 and Yes = 1). Pre and post intervention differences in individual scores will be noted to assess efficacy of the quit component of the BCI. Along with this the quit will be assessed subjectively by quit calendars that will be submitted at the 12 weeks follow-up.

Oral examinations
Prevalence of OPMLs and oral cancer will be evaluated in users of SLT and BQ by conventional method of screening and test positives will be referred for further diagnosis and treatment.
The data will be a nested data thus multilevel modelling will be done to have accuracy in the analysis of the hierarchical data. Analysis of covariates will be performed to control for the potential confounders including age.

Ethical considerations
 Formal consents will be sought from schools and parents for the study.
 All relevant details will be provided to the all concerned.
 Interim analysis(58) will be performed to cater for significant differences between the two arms of the study.
 All intervention material will be provided to the control arm at the end of the study.