Validity and Reliability of the Fagerstrom Test for Cigarette Dependence in a Sample of Arabic Speaking UK-Resident Yemeni Khat Chewers

Cigarette smoking remains a leading cause of preventable disease (cancer, heart disease) and premature death in the United States and other countries (Surgeon General’s report, 2004, 2010). The prevalence of tobacco smoking in the developed world has decreased dramatically in the last decades whilst the trend for the increase in the developing countries is growing. It is estimated that by 2030, 80% of all tobacco-related deaths will occur in developing countries (Mathers & Loncar, 2006). In the Arabic speaking countries in the Middle East the consumption of cigarette smoking has increased by 24% between 1990-1997 and the prevalence is higher amongst males with the Yemen having the highest rate (75%) of male prevalence of smoking followed by Jordan, Tunisia, Syria, Lebanon and Palestine at 40% (Shafey, 2007). The use of khat chewing amongst the Yemeni in homeland and diaspora to facilitate social interaction is often associated with tobacco smoking. Khat is an established drug of dependence resembling amphetamine chemically and in mode of action (Kalix, 1992). High prevalence of regular cigarette smoking amongst khat chewers has been reported recently (Kassim & Croucher, 2011). According to the Surgeon General’s report (1988),


Introduction
Cigarette smoking remains a leading cause of preventable disease (cancer, heart disease) and premature death in the United States and other countries (Surgeon General's report, 2004, 2010).The prevalence of tobacco smoking in the developed world has decreased dramatically in the last decades whilst the trend for the increase in the developing countries is growing.It is estimated that by 2030, 80% of all tobacco-related deaths will occur in developing countries (Mathers & Loncar, 2006).In the Arabic speaking countries in the Middle East the consumption of cigarette smoking has increased by 24% between 1990-1997 and the prevalence is higher amongst males with the Yemen having the highest rate (75%) of male prevalence of smoking followed by Jordan, Tunisia, Syria, Lebanon and Palestine at 40% (Shafey, 2007).The use of khat chewing amongst the Yemeni in homeland and diaspora to facilitate social interaction is often associated with tobacco smoking.Khat is an established drug of dependence resembling amphetamine chemically and in mode of action (Kalix, 1992).High prevalence of regular cigarette smoking amongst khat chewers has been reported recently (Kassim & Croucher, 2011).According to the Surgeon General's report (1988) nicotine dependence treatment approaches should be based on assessment of their dependence.The FTCD scale, formerly FTND, has been adapted globally as a measure of physical dependence on nicotine in particular cigarette (Fagerstrom, 2012) due to its ease of understanding and rapidity of application (De Meneses-Gaya et al., 2009).Its validity and reliability has been assessed in different cultural settings (Becona & Vazquez, 1998;Uysal et al., 2004;Huang, 2006;De Meneses-Gaya et al., 2009;Jhanjee & Sethi, 2010) despite recognized variations in its psychometric properties.The psychometric properties of FTCD in Arabic speakers have not been tested, in particular amongst this group who smoke cigarette and chew khat concurrently.
The objective of this communication is to assess the cross cultural validity and reliability of the FTCD scale in Arabic speakers, specifically, a sample of UK-resident male adult Yemeni khat chewers who smoked cigarettes regularly.

Materials and Methods
Recruitment and sample selection have been previously reported (Kassim & Croucher, 2010).This study involves a subsample (91regular cigarette smokers) of randomly recruited residents Yemeni khat chewers.Socio-demographics, severity of dependency on khat chewing (SDS-Khat) and cigarette smoking data have been assessed.

Measures
Socio-demographic characteristics of the sample were collected using a pre-piloted questionnaire (Kassim & Croucher, 2011).Levels (high and low) of cigarette dependence were measured using FTCD (Heatherton et al., 1991) as well as different levels of cigarette dependence (very low, low and medium) (Fagerstrom & Kozlowski, 1990).We also used SDS-Khat (Kassim & Croucher, 2010) to assess participants' severity of dependency on khat chewing.Adaptation/translation of the scale (FTCD) was carried out before the pilot study following the process of adaptation proposed by Hunt & Bhopal (2004).Four bilingual Yemeni khat chewers who smoked forward translated the English version of the questionnaire.Monolingual (Arabic speakers) khat chewers who smoked were consulted during this process and field-testing was conducted.This was then followed with back translation from Arabic to English.The researcher along with one Yemeni medical professional (who was fully aware of tobacco and khat chewing among Yemeni community) reviewed the adaptation/translation (Table 1).In this Arabic version the word 'mosque' was used instead of 'church' due to its relevance to the religion of most Yemenis.

Ethical approval and confidentiality
This study was approved by the The East London and City Health Authority Local Research Ethics Committee (REC Ref. No. 05/Q06034/195).

Statistical analysis
The data were analysed using SPSS v 16.A descriptive statistical analysis was undertaken first to report sample characteristics.The following steps were then performed to evaluate the validity and reliability of the FTCD.Construct validity of the FTCD was assessed using factorial structural analysis.Adequacy of the data for factor analysis was assessed.Exploratory factor analysis was adopted using the principle component method without limit on factors extracted.The number of factors was assessed by the scree test plot adopting the eigenvalue criteria exceeding 1 and the variance explained was observed.A factor loading of.40 or above was a prior criterion for item inclusion in defining a factor (Steven, 2002).As for the reliability of the scale, the item-total correlations were calculated and the impact of removing an item on the internal reliability of the scale was assessed.To ease interpretation of factors both Promax and Varimax rotated solutions were conducted.The internal consistency of the scale and subscales extracted was tested using Cronbach's alpha coefficient.

Sample characteristics
The mean/sd age of the sample was 41.36 (± 17.84) years (range 18-87).The mean (sd) age of tobacco smoking initiation was 19 years (± 6.21) (range 10-40), 68% were married, 65% with low level of education (up to high school education) and 35% only employed.The mean/sd of khat chewing dependency was 5.09 (± 3.57) and 45% were self reported more likely dependent on khat chewing.

The psychometric properties analysis of the FTCD
The psychometric properties analysis of the FTCD was as follows.First, the data set satisfied the criteria for factor analysis with a correlation matrix of many coefficients of .3 and above, a sampling adequacy (KMO) of 0.76 and a statistically significant value of Bartlett's Test of Sphericity 0.005.Second, a two factor solution accounting for 58.36% (41.64%, 16.74%) of the variance was obtained from exploratory factor analysis with principle components.Extraction of these factors was supported with the use of the Catell (1966) scree plot test which revealed a clear break after the second component.Third, the internal consistency of the FTCD was .68 overall and .60 and .62respectively for each subscale.Fourth, as for the reliability of the scale, item three of the scale performed least well on the construct reliability test (Table 2).A key item of nicotine addiction (item 2) the restoration of level of nicotine after abstinence during sleeping, showed a significant reduction in the reliability of the scale if omitted (Table 2) and similarly item 6 which indicates the maintenance of nicotine level during waking.Fourth, the analysis derived by using the orthogonal (Promax) and oblique (Varimax) rotations extracted two factors explaining 58.36% of the total variance (Table 3).The items loaded on Factor 1 (urgency of restoring the level of cigarette nicotine after abstinence during sleeping) were 'How soon after you wake up do you smoke your first cigarettes?','Which cigarette would you most hate to give up?' and 'How many cigarettes do you smoke per day?'.As for Factor 2 (maintaining the level of nicotine during waking) these were 'Do you smoke more frequently during the first hours after waking than during the rest of the day?', 'Do you smoke cigarettes even if you so ill that you are in bed much of the day?' and 'Do you find it difficult to refrain from smoking in places where it is forbidden, such as mosque, library or cinema?'.The loading items on two factors identified with the varimax rotation resembled that of the promax rotation apart from the items 'How soon after you wake up do you smoke your first cigarettes?' and 'Do you smoke more frequently during the first hours after waking than during the rest of the day?' which loaded on both factors with the promax rotation.

Discussion
In this study we have assessed the psychometric properties of FTCD in a sample of UK-resident Yemeni khat chewers.The key findings of the study were as follows: two factors were found to fit the data and the internal reliability of the scale was low alongside its underlying subscales.The study lends further support to the bi-dimensionality of the scale as reported in other cultural settings (Payne et al., 1994;Etter & Perneger, 1999;Haddock et al., 1999;Radzius et al., 2003).This study confirmed as well that Factor 1 reflects the degree of urgency to restore nicotine levels to a given threshold after waking whilst Factor 2 reflects the persistence with which nicotine concentration are maintained during waking hours (Radzius et al., 2003).The number of daily cigarettes has reported before to be loaded on the second factor (Radzius et al., 2003).According to Nunnally & Bernstein (1994) generalisability and validity of one psychometric characteristic to all contexts may not be possible as these features are influenced by the nature of the population being assessed.A recent systematic review has reported the bi-factorial structure of FTCD with small differences observed among the questions that compose the factors (De Meneses-Gaya et al., 2009).The two factors revealed from the promax rotation showed two distinguishable and correlated domains of nicotine dependence which lends further support to a construct of nicotine dependence that manifests itself by urgency of use and maintenance of nicotine levels (Jhanjee & Sethi, 2010).
The internal consistency of the Cronbach's alpha score was low (.68) and did not reach the standard Cronbach's alpha threshold of ≥ .70(Nunnally & Bernstein, 1994).This is not surprising and has been reported elsewhere (Pomerleau et al., 1994;Etter & Perneger, 1999).Item 3 'Which cigarette would you most hate to give up?' performed poorly on the reliability test as in other studies (Etter & Perneger, 1999;Uysal et al., 2004).
The findings of this study diverged from other studies (Etter & Perneger, 1999;Uysal et al., 2004) in terms of the high percentages (88% and 91% respectively) of smokers who reported their first daily cigarette within 30 minutes of waking and this cigarette as the least desirable to give up, supporting the pressing need to restore systemic nicotine levels.The SDS-khat established the psychological khat dependence, so there remains a knowledge gap with respect to the physical dependence on khat related stimulants such as cathinone measured by time of first intake after waking.Future research should  clarify the relationship between first intake of nicotine and khat related stimulants such as cathinone.This could be explored amongst khat chewing smokers and smokers only.
Limitations, this study reported cigarette dependence amongst khat chewers, and a possible overlap of nicotine and khat chewing dependence should be considered.Our study sampling framework was khat selling outlets which meant that khat chewer smokers who did not buy khat themselves from these places were excluded.Additionally, probability sampling was not employed; hence, extrapolation of findings from this study to the whole population would be inappropriate.The sample size for running factor analysis is contentious (Tabachnick & Fidell, 2007).The adequacy of this study sample was assessed and the number of cases per variable was five or more (Norman & Streiner, 2000).Further validation of this scale amongst larger samples of both male and female Arabic speakers is required.Test-retest has not been reported in this study.
In conclusions, the psychometric properties of the Arabic version of the FTCD scale in this sample of Yemeni khat chewers who smoked regularly confirmed what has been established in other cultural settings.The findings of this study have yet to be cross validated amongst other appropriately representative sample of Arabic speakers.