Tobacco use among black South African university students : attitudes , risk awareness and health locus of control

Objectives: To provide data on African/black South Afri­ can university students’ tobacco use status, belief in the benefits to health of not smoking, risk awareness in terms of knowledge of the links between smoking and disease, health locus of control, value for health, subjective health status and well-being. Design: Cross sectional. Setting: University of the North Subjects: 793 Black University students from non-health courses chosen by random sampling, of these 370 (46.7%) were males and 423 (53.3%) were females in the age range of 18 to 25 years (M age 21.0 years, SD=3.48). Main Outcome Measures: A measure of smoking, the M ultidimensional Health Locus of Control Scale, the Health as a Value Scale, and a measure for subjective health and subjective well-being. Results: The average prevalence of current tobacco use was 15% in men and 1% in women. The proportion of tobacco users who were classified as light users (1-10 per day) averaged 10% in men and 1% in women. Age and being male were significantly positively associated with status and frequency of tobacco use. Awareness of the link between smoking and lung cancer was high (93%), but awareness of the role of smoking in heart disease was very low (16%). The importance to health of not smoking was associated with smoking status (non-smoking versus smok­ ing). Overall, 75% of the current smokers stated that they would like to reduce the amount they smoked. Poor sub­ jective health status and low subjective well-being was associated with smoking status. No significant differences were found among non-tobacco users and tobacco users in relation to the three subscales of the Health Locus of Con­ trol (Internal, Chance, and Powerful others) and Value for health. Conclusion: For about 9% of the male students investi­ gated, a high risk exists to become regular tobacco users for the next 30 years.


Introduction
Tobacco consumption is a major, worldwide health hazard.In South Africa tobacco use is a major public health concern as it has severe consequences for smokers, non-smokers and the economy (Rocha-Silva, De Miranda & Erasmus 1996: 5ff., Yach 1996: 29).Yach (1996: 3 If.)reports that tobacco use among Blacks is rising in South Africa.For the country as a whole, lung cancer already accounts for 24% of all death from cancer in men^and 10.6% of all such death in women.A study of cigarette smoking in the black township popula tion of Cape Town showed that the smoking prevalence among adults was 53% in men compared to 6% in women (Strebel, Kuhn & Yach 1989: 209f.).Steyn et al. (1994: 784f.)con ducted a similar study and found that about 52% of the men, but only 8% of the women used tobacco regularly.Men and women who smoked cigarettes, smoked 9.6 and 4.3 cigarettes on average per day, respectively.Peltzer and Phaswana (1999: 36) found in a pilot study that among South African university students past month prevalence rates of tobacco use were 13% in males and 0% in females.Among rural primary health care patients in the Northern Province of South Africa, Peltzer (1999b: 96f.) found past 6 months smoking of cigarettes of 42.6% among men and 0% among women and taking snuff 3.3% and 17.2% respectively.Ac cording to Russel (1990: 293) ninety percent of teenagers who smoke 3-4 cigarettes a day become regular smokers for 35-40 years.
Many factors affect smoking, including cultural norms, so cial influences and economic constraints.Psychological mod els of health behaviour emphasize the relevance of cognitive factors such as attitudes and risk awareness.Two major cog nitive variables relevant to tobacco consumption were included in this survey.The first was risk awareness, operationalized in terms of knowledge of the links between smoking and two diseases: lung cancer and coronary heart disease.Ignorance of the risk factors associated with tobacco use has been re ported in a number of surveys (Peltzer & Phaswana 1999: 38).Reddy et al. (1996Reddy et al. ( : 1391) ) found in a national sample in South Africa that only 58% were aware that cancer was asso ciated with smoking and 36% associated heart disease with smoking.It might therefore be predicted that fewer users than non-users will be aware of the role of tobacco use in major diseases.Bennett et al.(1997: 179) studied a representative sample including measures of health locus of control and value for health and found that smokers held stronger internal, chance, and powerful others beliefs, than never smokers.Moreover, the interaction between chance, health locus of control, and value for health, was a significant predictor of smoking status, suggesting that health value may moderate the relationship between health locus of control and smoking status.Therefore this study also investigated health locus of control and value for health in regard to smoking status and frequency.

M aterials and Methods
Sample and procedure: The sample included 793 Black University students from non-health courses chosen from ran domly selected classes from the University of the North, South Africa.The students were 370 (46.7%) males and 423 (53.3%) females in the age range of 18 to 25 years (M age 21.0 years, SD=3.48).Data were collected by a self-administered questionnaire in a class room situation after informed consent had been obtained.Participants were assured of complete anonymity.

Measures:
The measure of smoking was adapted from the questions used for sreening adolescent smoking behaviour in the UK (Steptoe et al. 1995: 574f.)Tobacco consumption was assessed by asking participants to endorse one of eight re sponse options: "1 have never smoked a cigarette or taken snuff, not even a puff'; "1 have only ever tried one or two cigarettes or snuff1; "1 used to smoke sometimes, but I don't now"; "1 don't smoke cigarettes, but smoke a pipe or cigars'1; "1 smoke cigarettes but not as many as one per day"; "1 usu ally smoke between one or ten cigarettes per day"; "1 usually smoke between ten and twenty cigarettes per day"; "1 usually smoke more than twenty cigarettes per day".Participants in the first two categories were classified as never smokers, while respondents in the third category were classified as former smokers.For analysis of smoking prevalence the five smok ing categories were combined, and compared with the three non-smoking categories.For analysis of the prevalence of "ever smoking", former smokers were included with the present smokers.For analysis of the strength of smoking hab its, the five smoking categories were reduced to three: very light smokers (<l/day), light smokers (1-10/day) and moder ate/heavy smokers (> 10/day).Only one respondent stated he smoked cigars or a pipe.As this was a very small number, this category was grouped with the very light cigarette smok ers in the analysis of levels of smoking.Both the Cronbach alpha and the split-half reliability coefficient for the tobacco use measure were .6 for this sample.Smokers were asked whether they would like to reduce the amount they smoke, with responses in a Yes/No format.Be liefs were assessed by asking participants to rate their beleifs in the importance of not smoking for health on a 10-point scale where 1 =low importance and 10=very great importance.Risk awareness for smoking-disease links was asked in a Yes/ No format if they believed that lung cancer and heart disease were influenced by smoking.Data were also collected by us ing additional instruments that assess factors es tablished as im portant to health behaviour in re search in health p s y c h o lo g y .These included the Multidimen sional Health Locus of Control Scale (MHLOC) (Wallston, Wallston & DeVellis 1978: 160), and the Health as a Value Scale (Lau & Ware 1981: II47).Both the Cronbach alpha and the split-half reliability coefficient for the MHLOC and the Health as a Value Scale measures were .7 for this sample.In addition, one question was included on the subjective health status (rated from l=excellent to 5=poor), one question on subjective well-being (rated from l=very satisfied to 5=very dissatisfied, and 6 items on sociobiographic data: age, sex, religion, income, car ownership in family, education of mother and father.

Prevalence of tobacco consumption
The prevalence of current snuff users was very small (1% in men and 1 % in women) and lifetime users of snuff (once or twice) was 5% in men and 4% in women.Therefore, in the following snuff users are grouped together with tobacco smok ers and called tobacco users.

Table 1 indicates tobacco use status and frequency
The average prevalence of current smoking was 15% in men and 1 % in women.The proportion of smokers who were clas sified as light smokers (1-10 per day) averaged 10% in men and 1% in women.Age and being male were significantly positively associated with smoking status [for age: r=.132; p<.001 and for sex: r=.420; p<.001] and frequency of smok ing [for age: r=.141; p<001 and for sex: r=.262; p<001].In each case, there was a higher prevalence among men.The socio-economic background was neither related to smoking status [r=.012; ns] nor to freqency of smoking [r=.90; nsj.Religion was divided into Christian (Protestant and Roman Catholic) (50.6%) and healing churches (Zion Christian Church, Apostolic, and African) (49.4%).Smoking status and major religious orientation was weakly associated, e.g.never smokers were positively related to being a member of a Zion Christian, Apostolic church, or African/traditional religion (r=3.919,p<.044).

Beliefs about smoking, risk awareness and smoking
The mean scores on the 10-point rating of the importance of health of not smoking was among women 7.69 and men 7.52, which indicated no significant sex difference (F=2.187,ns).The highest mean of the importance of health of not smoking was rated for former smokers (8.29) and the lowest mean rat ing was for moderate/heavy smokers (5.30) (see Table 2).
As far as risk awareness was concerned 93% of the partici pants were aware of the association between smoking and lung cancer but only 16% knew the relationship between smoking and heart disease.There was no relationship be tween the level of risk awareness (smoking and heart dis ease) and smoking status.
Wish to reduce smoking levels Overall, 75% of the current smokers stated that they would like to reduce the amount they smoked.In the very light smok ing group 71.4% wanted to reduce smoking, compared with 88.9% in the light smoking group and only 44.4% in the mod erate/heavy smoking group.The wish to reduce smoking was positively associated with tobacco use status [r=.378; p<001] and amount of smoking [r=.278;pc.OOl].Seventy six per cent of the men and 24.1% of the female smokers stated that they would like to reduce the amount they used.Analysis of variance further indicated that the wish to reduce smoking was associated with current smokers, light smokers, moderate/heavy smokers, very light smokers, and the belief in the importance for health of not smoking.
Poor subjective health status was associ ated with moderate/heavy, current and light smokers and low subjective well-be ing was related to moderate/heavy smok ing (see Table 3).Regarding tobacco use status and fre quency no significant differences were found among non-tobacco users and to bacco users in relation to the three subscales of the Health Loucus of Control (Internal, Chance, and Powerful others) and Value for health except for a relationship between moderate/heavy tobacco use and internal locus of control.

Discussion
The sample of black university students studied had relatively low rates of current tobacco use, especially for females (15% in males and 1% in females), as compared to other samples in western countries (Webb et al. 1997: 145. Steptoe et al. (1995: 578f.)found among 21 European countries that the average prevalence of current smoking across country sam ples was 33.1% in male and 29.0% in female university stu dents.The same applies for the rate of former smokers (11% male and 3% female).Steptoe et al.( 1995: 576) showed that as many of the university students population (30.9% men and 34.2% women) stated that they had sometimes smoked in the past.Never smokers were positively related to being a member of a Zion Christian, Apostolic church, or African/traditional reli gion as opposed to Christian (Protestant and Roman Catho lic) denominations.Especially the Zion Christian Church and Apostolic church have prohibitions against tobacco use, which may explain the higher abstinence towards tobacco use (Peltzer 1999a: 400).According to Russell (1990: 293) 90% of teenagers who smoke 3-4 cigarettes a day become regular smok ers for 35-40 years.Thus, for about 9% of the male students in vestigated a high risk ex ists to become regular to bacco users at later age.Some of the risks and protective factors identified in this study can provide the basis for effective prevention and cessation programmes (Editorial 1999: 353, Mzileni et al. 1999: 400).

Curationis May 2001
The assessment of risk awareness showed that knowledge of the links between smoking and lung cancer in this sample (93%) was similarly higher than that found among European university students (97.4%) (Steptoe et al. 1995: 578).How ever, the knowledge of the association between smoking and heart disease (16%) was much lower than that among Euro pean university students (63.7%) (ibid.)and a national sam ple of South Africans (36%) (Reddy et al. 1996(Reddy et al. : 1389).In accordance with other studies (Steptoe et al. 1995: 579) the wish to reduce smoking was positively related with to bacco use status.Overall, 75% of the current smokers stated that they would like to reduce the amount they smoked; this was highest among the light smoking group (88.9%) and low est among the moderate/heavy smoking group (44.4%).This means that the level of wanting to cut down with smoking did not increase with higher smoking levels, as it was found among European university students (ibid.).Similarly as among European male students (69%), in this sample, 76% of the males wanted to reduce the amount they smoked.How ever, only 24.1 % of the female smokers wanted to reduce the amount they used.Poor subjective health status and low subjective well-being seems to be associated with smoking status, which needs fur ther investigation.Regarding tobacco use status and frequency, no significant differences were found among non-tobacco us ers and tobacco users in relation to the three subscales of the Health Locus of Control (Internal, Chance, and Powerful oth ers) and Value for health except for a relationship between moderate/heavy tobacco use and internal locus of control.This finding seems contrary to other research findings (Bennett et al. 1997: 181).The model by Wallston et al. (1978: 161 f.) suggests that those who score highly on the internal dimen sion regard their health largely within their own control and are likely to engage in health-maintaining behaviours, and those who score highly on the chance dimension view their health as relatively independent of their behaviour and ac cordingly, are more likely to engage in health-damaging be haviours than those with lower scores.Value for health was also not found to be related to smoking status as was found in a representative sample studied by Bennet et al. (1997: 180).

Conclusion
The proportion of tobacco users who were classified as light users (1-10 per day) averaged 10% in men and 1% in women.

Table 4 indicates
ANOVA between tobacco use, Health Locus of Control and Value for Health.

Table 2 :
M ean beliefs in the im portance for health of not smoking

Table 3 :
A N O V A between tobacco use status and wish to reduce sm oking, subjective health and w ell being *** pc.001; ** pc.01; * pc.05 (1) Rated from l=excellent to 5 poor, and (2) Rated from l=very satisfied to 5=very dissatisfied

Table 4 :
A N O V A between tobacco use, Health Locus of Control and Value for Health