- Email: [email protected]
Health-related behaviour in a small Scottish community
Public Health (1990), 104, 131-140
© The Society of Public Health, 1990
H e a l t h - R e l a t e d Behaviour in a Small Scottish Community A. Amos, C. Currie 2, S. M. Hunt 2 and C . J . Martin 2
1Department of Community Medicine, Edinburgh University, Medical School Teviot Place, Edinburgh, EH8 9AG eResearch Unit in Health and Behavioural Change, Edinburgh University, 17 Teviot Place, Edinburgh, EH1 2QZ, Scotland
A survey of health-related behaviours was carried out in a small Scottish town. A random sample stratified by age and sex was drawn from the records of the sole local health centre. A previously validated questionnaire was sent by post with a covering letter from the general practitioners and a reply paid envelope. An overall response rate of 71.3 % was achieved. The results indicated very' high rates of smoking for women with manual jobs, but differences between men in manual & non manual groups were not significant. Younger women reported higher alcohol consumption than older women, but there were no significant differences between occupational groups or employed and unemployed men. Only a minority of respondents reported regular consumption of high fibre products, low fat milk and polyunsaturated margarine and few claimed to undertake vigorous exercise. In general health-related behaviours showed little association with self-rated health. Whilst the data confirm the importance of socio-demographic factors in certain behaviours they also indicate that these are inadequate to predict the pattern of such behaviours within a community.
Introduction Scotland has one o f the lowest life expectancy rates in Europe, t'2 tops the league table for deaths f r o m heart disease and was the first country where deaths f r o m lung cancer a m o n g w o m e n o v e r t o o k those f r o m breast cancer. 3 There are, however, m a r k e d geographical differences and there is a need to have m o r e information a b o u t the prevalence o f risk factors for i m p o r t a n t causes o f p r e m a t u r e mortality and morbidity. F o r example, whilst Scotland has the highest rate o f smoking-related disease o u r knowledge a b o u t the pattern and prevalence o f smoking is limited. 4 There are no accurate Scottish statistics a b o u t smoking patterns and trends in the younger age groups or in the different social classes, and little information a b o u t regional or sub-regional differences. Scotland has higher smoking and alcohol consumption rates and a lower consumption of fresh fruit and vegetables than the rest o f the U K . 3~ However, it is unclear to what extent regional and socio-demographic factors affect differences in the pattern o f these behaviours within the population. Accurate information a b o u t the prevalence o f health-related behaviours and an understanding o f the factors which facilitate or inhibit behavioural change are i m p o r t a n t prerequisites for effective planning, implementation and evaluation o f health p r o m o t i o n strategies. As Green and others have pointed out, 7 health p r o m o t i o n and health education Correspondence: A. Amos
132
A. Amos et al.
programmes will only be successful if they take into account the needs of the target population and the process by which new ideas and behaviours diffuse through a population. This paper reports on the preliminary findings from a study designed to look at health-related behaviours and the process of behavioural change in a community by exploring, through postal surveys and in-depth interviews, people's perceptions of their own behaviour. Four health behaviours were selected for investigation: cigarette smoking, alcohol consumption, eating habits and physical activity. The first three because of their widely regarded importance as risk factors for major causes of mortality and morbidity, 8-~3 and physical activity because it is an important component of the healthy lifestyle approach to attaining positive health and well-being) 4 These behaviours were also of interest because over the past few years many initiatives have been developed with the aim of encouraging positive changes in these behaviours) 4-17 Method
The sample The sample was drawn from the records of a health centre located in an area to the north-east of Edinburgh. The location was chosen for its situation outwith urban Edinburgh, its social characteristics, for example having undergone economic changes as a result of the decline of the mining industry, its clear geographical boundaries and its socio-economic profile. The majority of residents belong to skilled or semi-skilled manual groups and the area is classed in the top 20% of deprived areas in Lothian Region. TM Quota sampling by five age groups and sex was used yielding a total of 2,310 residents aged 16-59 years at the time the sample was drawn. No more than one person of each sex from any address was chosen. Thus, because of the small population size, the sample covered more than half the households in the area. Procedure The overall study comprised two rounds of postal questionnaires followed by two rounds of interviews. The data reported here are based upon responses to the first postal questionnaire. The questionnaire was mainly in a closed format with respondents required to choose one of two or several alternative responses in relation to change during the past year in dietary habits, weight control, smoking, alcohol use and exercise. In addition, respondents were asked to give their height and weight, rate their health on a five-point scale and provide some socio-demographic detail. The general preactitioners at the health centre signed a covering letter and this together with the questionnaire and a pre-paid return envelope was sent out in July 1987. Two reminder letters were sent to individuals who had not responded after two and four weeks. Results
1,512 questionnaires were returned completed and 189 were returned to sender. The response rate was 71.3% and ranged from 64% for men aged 20-29 years to 81% for
Health-Related Behaviour in Scotland Table I
133
Social class composition by head of household Respondents %
I
II IIInm IIIm IV V other
1981 Census %
2.0
0
10.4 34.4 27.7 13.6 11.9 --
12 11 45 22 9 2
women aged 40-49 years. In relation to the 198l census for the area there were slightly fewer w o m e n aged under 30. Men were also underrepresented in younger age groups, but close to the census estimates for 40--49 and 50-59 year olds. Table I shows the social class distribution which has more Class I I I n m and fewer I I I m and IV than the 1981 census. However, in recent years there has been movement into the area by more affluent workers related to the sale of council houses. Twelve percent o f respondents described themselves as unemployed, two-thirds were married and 7% were divorced, widowed or separated. These figures were very similar to those for Lothian Region as a whole.
Cigarette smoking 44% o f women and 46% of men reported that they smoked cigarettes (Table II) and these rates differed significantly with age in both sexes ( P < 0 . 0 0 1 ) . Overall the rate of smoking amongst women and men was higher than the average for Scotland. W o m e n aged 40-59 had particularly high smoking rates which were higher, though not significantly so, than the rates for men of that age. W o m e n classified into the manual groups, on the basis of their spouse's occupation, had a much higher smoking rate (52%) than women in non-manual
Table II
Cigarette smoking by age and sex amongst respondents compared to Scotland Respondents (1987) %
16-19 20-29 30-39 40-49 50-59 All (16-59)* * Weighted by age t see Ref. 3
Scotland1- (1986) %
Women
Men
Women
Men
28 43 43 53 49 44
38 44 52 49 45 46
16--24 25-34 35--49 50-59 All 16+
(36) 33 43 45 36
(41) 37 40 44 38
134
A. Amos et al.
groups (32%) (P<0.001). The difference in smoking rates was augmented when women were classified according to their own occupation. 61% of women with manual jobs s m o k e d compared to only 30°,/o of those with non-manual jobs. The high smoking rates amongst working class female respondents was due to a combination of a higher ever smoked rate in this group (P<0.001) and a much lower cessation rate. Only 17% of working class women who had ever-smoked had given up compared to 38% of middle class women. Differences between male manual and non-manual groups were not significant. Single people had the lowest smoking rates with divorced/separated and widowed men (P<0.05) and women (P<0.001) having the highest. Differences again were greatest amongst the women ranging from 34% amongst the single, to 47% amongst the married and 58% amongst those who were divorced, widowed or separated. Unemployed men (63%) had higher smoking rates (P
Alcohol consumption Men in all age groups consumed alcohol more frequently and in larger amounts than women of the same age. Average weekly consumption was estimated by multiplying the usual number of drinking occasions each week by the average intake at each session. 72% of men claimed to drink at least once a week compared to only 48 % of women. There were no significant differences in the alcohol intake in the different age groups in men but younger women reported significantly higher consumption levels than the older women (P<0.001). There were no significant differences in consumption levels between manual and non-manual groups or between employed and unemployed men. However women employed outside the home reported significantly higher consumption (P<0.05) than women who worked exclusively in the home or who were unemployed. Highest consumption rates in men were found amongst the divorced, widowed or separated (P < 0.001) whereas highest rates in women were amongst the single (P < 0.01).
Eating patterns A range of foods were included in the questionnaire to give some indication of current eating patterns. A summary of the frequency of eating certain types of food is set out in Table III. Women reported eating fruit and wholemeal bread more often but chips, crisps, sausage and meat pies less often than men. There seemed to be distinctly different eating patterns associated with certain foods which are taken as indicators of 'healthy' eating, in particular high fibre cereal, wholemeal bread, low fat milk and polyunsaturated margarine. Whilst only a minority of respondents claimed to eat these foods regularly, that is on most days, the frequency of consumption varied considerably between the different products. The pattern of polyunsaturated margarine consumption clearly differed from other foods such as a high fibre cereal or wholemeal bread where it appeared that many respondents alternated their consumption with other versions of these products such as white bread. Age was associated with the freqeuncy of eating certain foods ie red meat, sausages/meat pies, chips/crisps, sweets/biscuits, high fibre cereal and, in women only, fruit and wholemeal bread. Overall there was an indication that for women consumption of fruit, high fibre cereal and wholemeal bread increased with age whilst that of processed foods decreased.
Health-Related Behaviour in Scotland Table III
135
Frequency of eating certain foods by sex* Women %
Fruit High fibre cereal Wholemeal bread White bread Low fat milk Polyunsaturated margarine Red meat Sausages/meat pies Chips or crisps Sweets or biscuits
Most days
1 or 2 days a week
43 29 32 49 33 42 20 3 17 42
43 21 37 26 13 11 64 54 61 38
Men % Rarely/ Most days never 14 49 33 25 54 48 16 43 23 21
26 21 22 67 28 43 21 7 28 39
1 or 2 days a week
Rarely/ never
48 21 30 17 11 9 69 66 56 39
26 57 48 16 61 48 10 26 16 22
* Weighted by age
Obesity and weight control Overweight and obesity, as measured by calculating individuals' BMIs [weight (kg)/height 2 (m)] using self-reported weight and height, increased with age in both men and w o m e n ( P < 0 . 0 5 ) , a pattern similar to that found in Scotland as a whole. 6 A third o f men and women were classified as overweight (BMI 23.9-28.5) and 13% o f women and 8% o f men were classified as obese (BMI > 28.6 for w o m e n and > 30 for men). Single men and w o m e n ( P < 0.001) had the lowest BMIs. The majority o f w o m e n in all age groups, 69% overall, were attempting to control their weight but this did not vary significantly with age. Fewer men in all age groups (P < 0.001), 45% overall, were attempting to control their weight and this increased significantly with age from 35% (16-19 yrs) to 55% (50-59 yrs). M o s t o f the people trying to control their weight were classified as either acceptable weight (40%) or overweight (40 %). However the prevalence o f weight control increased significantly in b o t h sexes with increasing BMI. F o r example over 80% o f overweight and obese w o m e n claimed to be controlling their weight. There were no social class differences with respect to B M I or weight control. Housewives were more likely to be obese than other w o m e n (P < 0.05) but employment status had no effect on weight control in women. Employed men were more likely to be controlling their weight ( P < 0.01) than unemployed men. Whilst proportionally fewer men c o m p a r e d to w o m e n claimed to be attempting to control their weight, those that were doing so were m o r e likely to feel that they were being successful (P < 0.001).
Exercise and physical activity Only a minority of men and women claimed to undertake some form o f vigorous exercise or physical activity on most days of the week and this declined significantly with age (Table IV) (P < 0.05). Considerably m o r e men than women, particularly those under 40, claimed to be taking regular exercise. Most people felt that on the whole they led quite active lives
A. A m o s et al.
136 Table IV
Participation in vigorous exercise or sport and self-rated overall level of physical activity by age and sex
Women 16-39 40-59 All (16-59)*
Daily/most days
Weekly
Les than weekly/ never
25 20 23
19 11 15
56 68 61
41 33 37 Very active 16 22
18 14 16 Quite active 64 57
42 52 47 Not very active 20 20
Men
16-39 40-59 All (16-59)* All women (16-59)* All men (16-59)* * Weighted by age
(Table IV). Age had little effect in w o m e n but men in older age groups felt they were less active ( P < 0.05). Generally men rated their lives as being m o r e active than women. Teenage boys had the highest activity ratings but teenage girls had the lowest. In all age groups the percentage o f respondents who regarded their daily life as being very active was lower than those claiming to exercise vigorously on m o s t days. E m p l o y m e n t status was not associated with exercise participation but single men (P < 0.001) and women (P < 0.05) had significantly higher overall rates. W o m e n in manual groups had higher rates ( P < 0.001) than n o n - m a n u a l w o m e n but no difference was found amongst the men. Participation in exercise and physical activity did not appear to be associated to any great extent with BMI. Health status
More men (28%) than w o m e n (22°/O), particularly in the younger age groups, felt their health was very good, and m o r e women (32%) than men (29%) rated their health as being fair to very poor. Older men rated their health m o r e poorly than younger men ( P < 0 . 0 5 ) but there were no significant differences a m o n g s t the women. M o r e men and women in non-manual occupations rated their health as being very good c o m p a r e d to those in manual occupations but this was only significant a m o n g s t the w o m e n ( P < 0 . 0 5 ) . Differences between the employed and unemployed were not significant. The highest self-reported rates o f fair to very p o o r health were a m o n g s t divorced, separated and widowed men, 47% c o m p a r e d to 32% a m o n g s t other m e n ( P < 0 . 0 1 ) . Marital status was not significantly associated with health in women. Overall health-related behaviours showed very little association with self-rated health, and in all cases associations were limited to the 20--39 age groups. Significant associations were found with smoking in 30-39 year old men ( P < 0 . 0 0 1 ) and w o m e n ( P < 0 . 0 5 ) , with smokers feeling less healthy than non-smokers. P o o r self-rated health was also associated with higher levels o f alcohol intake (P < 0.05) and attempts to control weight in 20-29 year
Health-Related Behaviour in Scotland
137
old men ( P < 0.05). Exercise frequency was not significantly associated with health in any age group. The only significant associations with different B M I categories was found amongst 20-29 ( P < 0 . 0 5 ) and 30-39 year old ( P < 0 . 0 0 1 ) w o m e n and 20-29 year old men (P < 0.001) with those in the higher weight category rating their health m o r e poorly. Inter-relationships between behaviours and health status The inter-relationship between people's smoking behaviour and their level o f alcohol consumption with other behaviours and health status was analysed using the Z2 test. Male smokers were likely to drink more, exercise less and be less likely to be attempting to control their weight than men who had given up or had never smoked. Smokers were also less likely to be eating healthy food such as fruit and wholemeal bread but more likely to be eating less healthy foods such as chips and sweets. Whilst ex-smokers had the highest levels o f obesity, smokers rated their health more poorly than ex or never smokers. In contrast, whilst smoking amongst the w o m e n was associated with the frequency o f eating certain foods there was no association with their levels o f drinking, exercise or attempts to control weight. However, women who smoked felt less healthy than those who had stopped or never smoked. B M I was not associated with smoking behaviour. Drinking levels in both men and w o m e n showed very little association with other behaviours. Where such associations did exist they were all, as with smoking, in the negative direction. T h a t is the m o r e people reported drinking the less likely they were to undertake 'positive' health-related behaviours. C o n s u m p t i o n o f polyunsaturated margarine was again distinguished f r o m other eating habits in that it was not associated with either smoking behaviour or alcohol consumption in either men or women. These findings suggest that different factors m a y be involved in determining the pattern and prevalence o f different health-related behaviours and that these factors m a y be different or operated differently amongst men and women. Discussion F r o m what was k n o w n from the census a b o u t the social class composition of this community, and its geographical location, certain predictions could have been m a d e a b o u t the pattern and prevalence o f health-related behaviours. In particular it might have been expected that smoking rates and alcohol consumption would be higher than the average for the U K and that consumption o f 'healthy' foods, attempts to control weight and participation in exercise would be lower. However, the picture that emerges f r o m the survey is somewhat m o r e complex. Smoking rates in both men and w o m e n were considerably higher than in the U K 4 and Scotland as a whole? However, the very high rates amongst older and working class w o m e n could not have been predicted. The rate was even higher amongst w o m e n who had manual jobs in contrast to the U K as a whole where smoking rates differed little between w o m e n classified according to their own occupational grouping, a° With respect to alcohol consumption although slightly m o r e men and w o m e n c o m p a r e d to Scotland as a whole claimed to be regular drinkers, the p r o p o r t i o n classified as moderate or heavy drinkers was either similar to (in men) or less (in women) than that obtained in other studies. 6 It also appears that underage drinking was quite c o m m o n in this c o m m u n i t y as over half the 16-19 year olds claimed to be regular drinkers. Only limited data are available a b o u t weight control and exercise habits in Scotland and the U K and much o f the data is not directly c o m p a r a b l e with this study. However, there are
138
A. Amos et al.
indications that, in some respects, residents in this community may have a somewhat higher than expected prevalence of positive health-related behaviours. With respect to eating habits the frequency of consumption of many foods was indeed similar or slightly worse than that reported for Scotland and the rest of the UK. 6 However, wholemeal bread was eaten twice as often as that reported for Britain 7 and three times as often as reported individuals in another community in Lothian. ~9Similarly, even taking into account the fact that low fat milk consumption is higher in Scotland compared to the rest of Britain, consumption in this community where 40% of respondents claimed to drink it at least once a week, was considerably higher than that in Britain where only 26°/0 of households buy it every week. 7 Reported consumption of polyunsaturated margarine was three times as high as that in Wales. 17 Before conclusions can be drawn from these findings it is necessary to consider the extent to which data from different studies are comparable, and the validity of such data. Comparability and validity are likely to vary depending on the health behaviour being considered. For example, whilst data collected in this way about cigarette smoking is generally regarded as being reasonably accurate and comparable between studies, it is known that surveys of this type tend to considerably underestimate alcohol consumption. 2~ However, the figures which are obtained about alcohol consumption have been found to be reasonably consistent between differentstudies. In contrast very little is known about the validity of questions about eating habits and even less is known about weight control, physical activity and exercise. In part this is due to methodological difficulties. For example, truly accurate measures of nutritional intake can only be obtained using the weighted intake diary method, a method which is highly demanding of both the participant and the researcher. Therefore most surveys, with some modifications,7 have been limited to collecting subjective data about eating patterns. Despite these methodological limitations there still are some behavioural patterns which could not have been predicted from, and cannot be totally explained by, the known socio-economic and demographic factors. Amongst the most striking is the very high cigarette smoking rates amongst men and women. Whilst this was clearly associated with social class, employment and marital status, these factors do not fully explain why the rate should be so high in this community or why for example in the younger age groups more men smoke than women, but this is reversed in the older age groups. These findings highlight some of the problems inherent in using often limited national data to deduce the prevalence and patterns of behaviour at the community level. Whilst the data from this survey to a large extent confirm the importance of socio-demographic factors such as age, sex, social class, employment status and marital status, in relation to health-related behaviours, they indicate that this information may be inadequate when attempting to predict these behaviours within a population or community. This clearly has implications for the planning and implementation of health promotion programmes. The findings tend to support those of other studies which have looked at preventive behaviours, both health care behaviours such as breast self-examination and health behaviour such as smoking, which have concluded that preventive behaviour is not unidimensional.22 24 It has been suggested that there may be several dimensions of preventive behaviours which may be independent of each other and the dynamics of which may differ. At the individual level this means that health behaviours may not be strongly associated with each other. The evidence from this study indicates that while some health-related behaviours might be inter-related, these associations are limited and differ between the sexes. In order to fully understand the patterning of health-related behaviours it is necessary to
Health-Related Behaviour in Scotland
139
have further information both structural and personal. F o r example, the low consumption o f fresh fruit m a y be related to its high price and the difficulty o f obtaining good quality fruit in Scotland. High levels o f smoking have been associated with certain types o f stress 25 and the degree o f concern with weight m a y be m o r e a consequence o f anxieties a b o u t appearance than a b o u t healthfl 6 It is hoped that further data to be reported from the second state o f this study will help elucidate these and other factors.
Acknowledgements This research was funded by a grant from the H S R C , S H H D . We wish to thank Jane H o p t o n for her help in the data analysis and F i o n a Foley and J e m i m a Porter for their typing.
References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23.
World Health Organisation. (1986). World Health Statistics Annual 1985. Geneva: WHO. World Health Organisation. (1987). World Health Statistics Annual 1986. Geneva: WHO. Information Services Division. (1988). Scottish Health Statistics 1986. Edinburgh: CSA. OPCS. (1988). Cigarette smoking 1972 to 1986. OPCS Monitor, ss 88/1. London: HMSO. Health Promotion Research Trust. (1987). The Health and Lifestyle Survey. London: Health Promotion Research Trust. Ministry of Agriculture, Fisheries and Food. (1988). Household Food Consumption and Expenditure Survey, 1986. London: HMSO. Green, L. (1985). Health education strategies and approaches to education in community and mass media. In: Smoking Control. Crofton, J. & Wood, M. (eds). 35-45. London: Health Education Council. National Advisory Council on Nutrition Education. (1983). Proposals for Nutrition Guidelinesfor Health Education. London: Health Education Council. Committee on Medical Aspects of Food Policy. (1984). Diet and Cardiovascular Disease. London: DHSS. Royal College of Physicians (1983). Health or Smoking. London: Pitman. Royal College of Psychiatrists. (1986). Alcoholour Favourite Drug. London: Tavistock. Scottish Health Education Co-ordinating Committee. (1983). Health Education in the Prevention of Smoking-Related Diseases. Edinburgh: SHHD. Scottish Health Education Co-ordinating Committee. (1985). Health Education in the Prevention of Alcohol-Related Problems. Edinburgh: SHHD, Scottish Health Education Group. (1988). Annual Report 1986-1987. Edinburgh: SHEG. National Forum for Coronary Heart Disease Prevention. (1988). Coronary Heart Disease Prevention, Action in the UK 1984-1987. London: Health Education Authority. Health Education Authority. (1988). Working for Health. London: Health Education Authority. Welsh Heart Programme Directorate. (1986). Pulse of Wales: Preliminary Report of the Welsh Heart Health Survey 1985. Cardiff: Heartbeat Wales. Lothian Regional Council Health Committee. (1987). An Analysis of Mortality Differentials in the 0-64 year old Population of Lothian, 1974-85. Edinburgh: Lothian Regional Council. Hunt, S. M., Currie, C. & Martin, C. J. (1987). Food and Fibre in a Scottish Community. Health Education Research, 2, 223-229. ASH. (1986). Women andsmoking--a Handbook for Action. London: Health Education Council. Dunbar, G.C. & Morgan, D. D. V. (1987). The changing pattern of alcohol consumption in England and Wales 1978-85. British Medical Journal, 295, 807-810. Calnan, M. (1985). Patterns in preventive behaviour: a study of women in middle age. Social Science & Medicine, 20, 263-268. Health and Welfare Canada. (1988). Canada's Health Promotion Survey: Technical Report. Ottawa: Minister of Supply and Services Canada.
140
A. Amos et al.
24. Pill, R. & Stott, N. C. H. (1985). Preventive procedures and practices among working class women: new data and fresh insights. Social Science & Medicine, 21,975-983. 25. Graham, H. (1987). Women's smoking and family health. Social Science and Medicine, 25,47-56. 26. Hayes, D. T. & Ross, C. E. (1987). Concern with appearance, health beliefs and eating habits. Journal of Health and Social Behaviour, 28, 120-130.
© The Society of Public Health, 1990
H e a l t h - R e l a t e d Behaviour in a Small Scottish Community A. Amos, C. Currie 2, S. M. Hunt 2 and C . J . Martin 2
1Department of Community Medicine, Edinburgh University, Medical School Teviot Place, Edinburgh, EH8 9AG eResearch Unit in Health and Behavioural Change, Edinburgh University, 17 Teviot Place, Edinburgh, EH1 2QZ, Scotland
A survey of health-related behaviours was carried out in a small Scottish town. A random sample stratified by age and sex was drawn from the records of the sole local health centre. A previously validated questionnaire was sent by post with a covering letter from the general practitioners and a reply paid envelope. An overall response rate of 71.3 % was achieved. The results indicated very' high rates of smoking for women with manual jobs, but differences between men in manual & non manual groups were not significant. Younger women reported higher alcohol consumption than older women, but there were no significant differences between occupational groups or employed and unemployed men. Only a minority of respondents reported regular consumption of high fibre products, low fat milk and polyunsaturated margarine and few claimed to undertake vigorous exercise. In general health-related behaviours showed little association with self-rated health. Whilst the data confirm the importance of socio-demographic factors in certain behaviours they also indicate that these are inadequate to predict the pattern of such behaviours within a community.
Introduction Scotland has one o f the lowest life expectancy rates in Europe, t'2 tops the league table for deaths f r o m heart disease and was the first country where deaths f r o m lung cancer a m o n g w o m e n o v e r t o o k those f r o m breast cancer. 3 There are, however, m a r k e d geographical differences and there is a need to have m o r e information a b o u t the prevalence o f risk factors for i m p o r t a n t causes o f p r e m a t u r e mortality and morbidity. F o r example, whilst Scotland has the highest rate o f smoking-related disease o u r knowledge a b o u t the pattern and prevalence o f smoking is limited. 4 There are no accurate Scottish statistics a b o u t smoking patterns and trends in the younger age groups or in the different social classes, and little information a b o u t regional or sub-regional differences. Scotland has higher smoking and alcohol consumption rates and a lower consumption of fresh fruit and vegetables than the rest o f the U K . 3~ However, it is unclear to what extent regional and socio-demographic factors affect differences in the pattern o f these behaviours within the population. Accurate information a b o u t the prevalence o f health-related behaviours and an understanding o f the factors which facilitate or inhibit behavioural change are i m p o r t a n t prerequisites for effective planning, implementation and evaluation o f health p r o m o t i o n strategies. As Green and others have pointed out, 7 health p r o m o t i o n and health education Correspondence: A. Amos
132
A. Amos et al.
programmes will only be successful if they take into account the needs of the target population and the process by which new ideas and behaviours diffuse through a population. This paper reports on the preliminary findings from a study designed to look at health-related behaviours and the process of behavioural change in a community by exploring, through postal surveys and in-depth interviews, people's perceptions of their own behaviour. Four health behaviours were selected for investigation: cigarette smoking, alcohol consumption, eating habits and physical activity. The first three because of their widely regarded importance as risk factors for major causes of mortality and morbidity, 8-~3 and physical activity because it is an important component of the healthy lifestyle approach to attaining positive health and well-being) 4 These behaviours were also of interest because over the past few years many initiatives have been developed with the aim of encouraging positive changes in these behaviours) 4-17 Method
The sample The sample was drawn from the records of a health centre located in an area to the north-east of Edinburgh. The location was chosen for its situation outwith urban Edinburgh, its social characteristics, for example having undergone economic changes as a result of the decline of the mining industry, its clear geographical boundaries and its socio-economic profile. The majority of residents belong to skilled or semi-skilled manual groups and the area is classed in the top 20% of deprived areas in Lothian Region. TM Quota sampling by five age groups and sex was used yielding a total of 2,310 residents aged 16-59 years at the time the sample was drawn. No more than one person of each sex from any address was chosen. Thus, because of the small population size, the sample covered more than half the households in the area. Procedure The overall study comprised two rounds of postal questionnaires followed by two rounds of interviews. The data reported here are based upon responses to the first postal questionnaire. The questionnaire was mainly in a closed format with respondents required to choose one of two or several alternative responses in relation to change during the past year in dietary habits, weight control, smoking, alcohol use and exercise. In addition, respondents were asked to give their height and weight, rate their health on a five-point scale and provide some socio-demographic detail. The general preactitioners at the health centre signed a covering letter and this together with the questionnaire and a pre-paid return envelope was sent out in July 1987. Two reminder letters were sent to individuals who had not responded after two and four weeks. Results
1,512 questionnaires were returned completed and 189 were returned to sender. The response rate was 71.3% and ranged from 64% for men aged 20-29 years to 81% for
Health-Related Behaviour in Scotland Table I
133
Social class composition by head of household Respondents %
I
II IIInm IIIm IV V other
1981 Census %
2.0
0
10.4 34.4 27.7 13.6 11.9 --
12 11 45 22 9 2
women aged 40-49 years. In relation to the 198l census for the area there were slightly fewer w o m e n aged under 30. Men were also underrepresented in younger age groups, but close to the census estimates for 40--49 and 50-59 year olds. Table I shows the social class distribution which has more Class I I I n m and fewer I I I m and IV than the 1981 census. However, in recent years there has been movement into the area by more affluent workers related to the sale of council houses. Twelve percent o f respondents described themselves as unemployed, two-thirds were married and 7% were divorced, widowed or separated. These figures were very similar to those for Lothian Region as a whole.
Cigarette smoking 44% o f women and 46% of men reported that they smoked cigarettes (Table II) and these rates differed significantly with age in both sexes ( P < 0 . 0 0 1 ) . Overall the rate of smoking amongst women and men was higher than the average for Scotland. W o m e n aged 40-59 had particularly high smoking rates which were higher, though not significantly so, than the rates for men of that age. W o m e n classified into the manual groups, on the basis of their spouse's occupation, had a much higher smoking rate (52%) than women in non-manual
Table II
Cigarette smoking by age and sex amongst respondents compared to Scotland Respondents (1987) %
16-19 20-29 30-39 40-49 50-59 All (16-59)* * Weighted by age t see Ref. 3
Scotland1- (1986) %
Women
Men
Women
Men
28 43 43 53 49 44
38 44 52 49 45 46
16--24 25-34 35--49 50-59 All 16+
(36) 33 43 45 36
(41) 37 40 44 38
134
A. Amos et al.
groups (32%) (P<0.001). The difference in smoking rates was augmented when women were classified according to their own occupation. 61% of women with manual jobs s m o k e d compared to only 30°,/o of those with non-manual jobs. The high smoking rates amongst working class female respondents was due to a combination of a higher ever smoked rate in this group (P<0.001) and a much lower cessation rate. Only 17% of working class women who had ever-smoked had given up compared to 38% of middle class women. Differences between male manual and non-manual groups were not significant. Single people had the lowest smoking rates with divorced/separated and widowed men (P<0.05) and women (P<0.001) having the highest. Differences again were greatest amongst the women ranging from 34% amongst the single, to 47% amongst the married and 58% amongst those who were divorced, widowed or separated. Unemployed men (63%) had higher smoking rates (P
Alcohol consumption Men in all age groups consumed alcohol more frequently and in larger amounts than women of the same age. Average weekly consumption was estimated by multiplying the usual number of drinking occasions each week by the average intake at each session. 72% of men claimed to drink at least once a week compared to only 48 % of women. There were no significant differences in the alcohol intake in the different age groups in men but younger women reported significantly higher consumption levels than the older women (P<0.001). There were no significant differences in consumption levels between manual and non-manual groups or between employed and unemployed men. However women employed outside the home reported significantly higher consumption (P<0.05) than women who worked exclusively in the home or who were unemployed. Highest consumption rates in men were found amongst the divorced, widowed or separated (P < 0.001) whereas highest rates in women were amongst the single (P < 0.01).
Eating patterns A range of foods were included in the questionnaire to give some indication of current eating patterns. A summary of the frequency of eating certain types of food is set out in Table III. Women reported eating fruit and wholemeal bread more often but chips, crisps, sausage and meat pies less often than men. There seemed to be distinctly different eating patterns associated with certain foods which are taken as indicators of 'healthy' eating, in particular high fibre cereal, wholemeal bread, low fat milk and polyunsaturated margarine. Whilst only a minority of respondents claimed to eat these foods regularly, that is on most days, the frequency of consumption varied considerably between the different products. The pattern of polyunsaturated margarine consumption clearly differed from other foods such as a high fibre cereal or wholemeal bread where it appeared that many respondents alternated their consumption with other versions of these products such as white bread. Age was associated with the freqeuncy of eating certain foods ie red meat, sausages/meat pies, chips/crisps, sweets/biscuits, high fibre cereal and, in women only, fruit and wholemeal bread. Overall there was an indication that for women consumption of fruit, high fibre cereal and wholemeal bread increased with age whilst that of processed foods decreased.
Health-Related Behaviour in Scotland Table III
135
Frequency of eating certain foods by sex* Women %
Fruit High fibre cereal Wholemeal bread White bread Low fat milk Polyunsaturated margarine Red meat Sausages/meat pies Chips or crisps Sweets or biscuits
Most days
1 or 2 days a week
43 29 32 49 33 42 20 3 17 42
43 21 37 26 13 11 64 54 61 38
Men % Rarely/ Most days never 14 49 33 25 54 48 16 43 23 21
26 21 22 67 28 43 21 7 28 39
1 or 2 days a week
Rarely/ never
48 21 30 17 11 9 69 66 56 39
26 57 48 16 61 48 10 26 16 22
* Weighted by age
Obesity and weight control Overweight and obesity, as measured by calculating individuals' BMIs [weight (kg)/height 2 (m)] using self-reported weight and height, increased with age in both men and w o m e n ( P < 0 . 0 5 ) , a pattern similar to that found in Scotland as a whole. 6 A third o f men and women were classified as overweight (BMI 23.9-28.5) and 13% o f women and 8% o f men were classified as obese (BMI > 28.6 for w o m e n and > 30 for men). Single men and w o m e n ( P < 0.001) had the lowest BMIs. The majority o f w o m e n in all age groups, 69% overall, were attempting to control their weight but this did not vary significantly with age. Fewer men in all age groups (P < 0.001), 45% overall, were attempting to control their weight and this increased significantly with age from 35% (16-19 yrs) to 55% (50-59 yrs). M o s t o f the people trying to control their weight were classified as either acceptable weight (40%) or overweight (40 %). However the prevalence o f weight control increased significantly in b o t h sexes with increasing BMI. F o r example over 80% o f overweight and obese w o m e n claimed to be controlling their weight. There were no social class differences with respect to B M I or weight control. Housewives were more likely to be obese than other w o m e n (P < 0.05) but employment status had no effect on weight control in women. Employed men were more likely to be controlling their weight ( P < 0.01) than unemployed men. Whilst proportionally fewer men c o m p a r e d to w o m e n claimed to be attempting to control their weight, those that were doing so were m o r e likely to feel that they were being successful (P < 0.001).
Exercise and physical activity Only a minority of men and women claimed to undertake some form o f vigorous exercise or physical activity on most days of the week and this declined significantly with age (Table IV) (P < 0.05). Considerably m o r e men than women, particularly those under 40, claimed to be taking regular exercise. Most people felt that on the whole they led quite active lives
A. A m o s et al.
136 Table IV
Participation in vigorous exercise or sport and self-rated overall level of physical activity by age and sex
Women 16-39 40-59 All (16-59)*
Daily/most days
Weekly
Les than weekly/ never
25 20 23
19 11 15
56 68 61
41 33 37 Very active 16 22
18 14 16 Quite active 64 57
42 52 47 Not very active 20 20
Men
16-39 40-59 All (16-59)* All women (16-59)* All men (16-59)* * Weighted by age
(Table IV). Age had little effect in w o m e n but men in older age groups felt they were less active ( P < 0.05). Generally men rated their lives as being m o r e active than women. Teenage boys had the highest activity ratings but teenage girls had the lowest. In all age groups the percentage o f respondents who regarded their daily life as being very active was lower than those claiming to exercise vigorously on m o s t days. E m p l o y m e n t status was not associated with exercise participation but single men (P < 0.001) and women (P < 0.05) had significantly higher overall rates. W o m e n in manual groups had higher rates ( P < 0.001) than n o n - m a n u a l w o m e n but no difference was found amongst the men. Participation in exercise and physical activity did not appear to be associated to any great extent with BMI. Health status
More men (28%) than w o m e n (22°/O), particularly in the younger age groups, felt their health was very good, and m o r e women (32%) than men (29%) rated their health as being fair to very poor. Older men rated their health m o r e poorly than younger men ( P < 0 . 0 5 ) but there were no significant differences a m o n g s t the women. M o r e men and women in non-manual occupations rated their health as being very good c o m p a r e d to those in manual occupations but this was only significant a m o n g s t the w o m e n ( P < 0 . 0 5 ) . Differences between the employed and unemployed were not significant. The highest self-reported rates o f fair to very p o o r health were a m o n g s t divorced, separated and widowed men, 47% c o m p a r e d to 32% a m o n g s t other m e n ( P < 0 . 0 1 ) . Marital status was not significantly associated with health in women. Overall health-related behaviours showed very little association with self-rated health, and in all cases associations were limited to the 20--39 age groups. Significant associations were found with smoking in 30-39 year old men ( P < 0 . 0 0 1 ) and w o m e n ( P < 0 . 0 5 ) , with smokers feeling less healthy than non-smokers. P o o r self-rated health was also associated with higher levels o f alcohol intake (P < 0.05) and attempts to control weight in 20-29 year
Health-Related Behaviour in Scotland
137
old men ( P < 0.05). Exercise frequency was not significantly associated with health in any age group. The only significant associations with different B M I categories was found amongst 20-29 ( P < 0 . 0 5 ) and 30-39 year old ( P < 0 . 0 0 1 ) w o m e n and 20-29 year old men (P < 0.001) with those in the higher weight category rating their health m o r e poorly. Inter-relationships between behaviours and health status The inter-relationship between people's smoking behaviour and their level o f alcohol consumption with other behaviours and health status was analysed using the Z2 test. Male smokers were likely to drink more, exercise less and be less likely to be attempting to control their weight than men who had given up or had never smoked. Smokers were also less likely to be eating healthy food such as fruit and wholemeal bread but more likely to be eating less healthy foods such as chips and sweets. Whilst ex-smokers had the highest levels o f obesity, smokers rated their health more poorly than ex or never smokers. In contrast, whilst smoking amongst the w o m e n was associated with the frequency o f eating certain foods there was no association with their levels o f drinking, exercise or attempts to control weight. However, women who smoked felt less healthy than those who had stopped or never smoked. B M I was not associated with smoking behaviour. Drinking levels in both men and w o m e n showed very little association with other behaviours. Where such associations did exist they were all, as with smoking, in the negative direction. T h a t is the m o r e people reported drinking the less likely they were to undertake 'positive' health-related behaviours. C o n s u m p t i o n o f polyunsaturated margarine was again distinguished f r o m other eating habits in that it was not associated with either smoking behaviour or alcohol consumption in either men or women. These findings suggest that different factors m a y be involved in determining the pattern and prevalence o f different health-related behaviours and that these factors m a y be different or operated differently amongst men and women. Discussion F r o m what was k n o w n from the census a b o u t the social class composition of this community, and its geographical location, certain predictions could have been m a d e a b o u t the pattern and prevalence o f health-related behaviours. In particular it might have been expected that smoking rates and alcohol consumption would be higher than the average for the U K and that consumption o f 'healthy' foods, attempts to control weight and participation in exercise would be lower. However, the picture that emerges f r o m the survey is somewhat m o r e complex. Smoking rates in both men and w o m e n were considerably higher than in the U K 4 and Scotland as a whole? However, the very high rates amongst older and working class w o m e n could not have been predicted. The rate was even higher amongst w o m e n who had manual jobs in contrast to the U K as a whole where smoking rates differed little between w o m e n classified according to their own occupational grouping, a° With respect to alcohol consumption although slightly m o r e men and w o m e n c o m p a r e d to Scotland as a whole claimed to be regular drinkers, the p r o p o r t i o n classified as moderate or heavy drinkers was either similar to (in men) or less (in women) than that obtained in other studies. 6 It also appears that underage drinking was quite c o m m o n in this c o m m u n i t y as over half the 16-19 year olds claimed to be regular drinkers. Only limited data are available a b o u t weight control and exercise habits in Scotland and the U K and much o f the data is not directly c o m p a r a b l e with this study. However, there are
138
A. Amos et al.
indications that, in some respects, residents in this community may have a somewhat higher than expected prevalence of positive health-related behaviours. With respect to eating habits the frequency of consumption of many foods was indeed similar or slightly worse than that reported for Scotland and the rest of the UK. 6 However, wholemeal bread was eaten twice as often as that reported for Britain 7 and three times as often as reported individuals in another community in Lothian. ~9Similarly, even taking into account the fact that low fat milk consumption is higher in Scotland compared to the rest of Britain, consumption in this community where 40% of respondents claimed to drink it at least once a week, was considerably higher than that in Britain where only 26°/0 of households buy it every week. 7 Reported consumption of polyunsaturated margarine was three times as high as that in Wales. 17 Before conclusions can be drawn from these findings it is necessary to consider the extent to which data from different studies are comparable, and the validity of such data. Comparability and validity are likely to vary depending on the health behaviour being considered. For example, whilst data collected in this way about cigarette smoking is generally regarded as being reasonably accurate and comparable between studies, it is known that surveys of this type tend to considerably underestimate alcohol consumption. 2~ However, the figures which are obtained about alcohol consumption have been found to be reasonably consistent between differentstudies. In contrast very little is known about the validity of questions about eating habits and even less is known about weight control, physical activity and exercise. In part this is due to methodological difficulties. For example, truly accurate measures of nutritional intake can only be obtained using the weighted intake diary method, a method which is highly demanding of both the participant and the researcher. Therefore most surveys, with some modifications,7 have been limited to collecting subjective data about eating patterns. Despite these methodological limitations there still are some behavioural patterns which could not have been predicted from, and cannot be totally explained by, the known socio-economic and demographic factors. Amongst the most striking is the very high cigarette smoking rates amongst men and women. Whilst this was clearly associated with social class, employment and marital status, these factors do not fully explain why the rate should be so high in this community or why for example in the younger age groups more men smoke than women, but this is reversed in the older age groups. These findings highlight some of the problems inherent in using often limited national data to deduce the prevalence and patterns of behaviour at the community level. Whilst the data from this survey to a large extent confirm the importance of socio-demographic factors such as age, sex, social class, employment status and marital status, in relation to health-related behaviours, they indicate that this information may be inadequate when attempting to predict these behaviours within a population or community. This clearly has implications for the planning and implementation of health promotion programmes. The findings tend to support those of other studies which have looked at preventive behaviours, both health care behaviours such as breast self-examination and health behaviour such as smoking, which have concluded that preventive behaviour is not unidimensional.22 24 It has been suggested that there may be several dimensions of preventive behaviours which may be independent of each other and the dynamics of which may differ. At the individual level this means that health behaviours may not be strongly associated with each other. The evidence from this study indicates that while some health-related behaviours might be inter-related, these associations are limited and differ between the sexes. In order to fully understand the patterning of health-related behaviours it is necessary to
Health-Related Behaviour in Scotland
139
have further information both structural and personal. F o r example, the low consumption o f fresh fruit m a y be related to its high price and the difficulty o f obtaining good quality fruit in Scotland. High levels o f smoking have been associated with certain types o f stress 25 and the degree o f concern with weight m a y be m o r e a consequence o f anxieties a b o u t appearance than a b o u t healthfl 6 It is hoped that further data to be reported from the second state o f this study will help elucidate these and other factors.
Acknowledgements This research was funded by a grant from the H S R C , S H H D . We wish to thank Jane H o p t o n for her help in the data analysis and F i o n a Foley and J e m i m a Porter for their typing.
References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23.
World Health Organisation. (1986). World Health Statistics Annual 1985. Geneva: WHO. World Health Organisation. (1987). World Health Statistics Annual 1986. Geneva: WHO. Information Services Division. (1988). Scottish Health Statistics 1986. Edinburgh: CSA. OPCS. (1988). Cigarette smoking 1972 to 1986. OPCS Monitor, ss 88/1. London: HMSO. Health Promotion Research Trust. (1987). The Health and Lifestyle Survey. London: Health Promotion Research Trust. Ministry of Agriculture, Fisheries and Food. (1988). Household Food Consumption and Expenditure Survey, 1986. London: HMSO. Green, L. (1985). Health education strategies and approaches to education in community and mass media. In: Smoking Control. Crofton, J. & Wood, M. (eds). 35-45. London: Health Education Council. National Advisory Council on Nutrition Education. (1983). Proposals for Nutrition Guidelinesfor Health Education. London: Health Education Council. Committee on Medical Aspects of Food Policy. (1984). Diet and Cardiovascular Disease. London: DHSS. Royal College of Physicians (1983). Health or Smoking. London: Pitman. Royal College of Psychiatrists. (1986). Alcoholour Favourite Drug. London: Tavistock. Scottish Health Education Co-ordinating Committee. (1983). Health Education in the Prevention of Smoking-Related Diseases. Edinburgh: SHHD. Scottish Health Education Co-ordinating Committee. (1985). Health Education in the Prevention of Alcohol-Related Problems. Edinburgh: SHHD, Scottish Health Education Group. (1988). Annual Report 1986-1987. Edinburgh: SHEG. National Forum for Coronary Heart Disease Prevention. (1988). Coronary Heart Disease Prevention, Action in the UK 1984-1987. London: Health Education Authority. Health Education Authority. (1988). Working for Health. London: Health Education Authority. Welsh Heart Programme Directorate. (1986). Pulse of Wales: Preliminary Report of the Welsh Heart Health Survey 1985. Cardiff: Heartbeat Wales. Lothian Regional Council Health Committee. (1987). An Analysis of Mortality Differentials in the 0-64 year old Population of Lothian, 1974-85. Edinburgh: Lothian Regional Council. Hunt, S. M., Currie, C. & Martin, C. J. (1987). Food and Fibre in a Scottish Community. Health Education Research, 2, 223-229. ASH. (1986). Women andsmoking--a Handbook for Action. London: Health Education Council. Dunbar, G.C. & Morgan, D. D. V. (1987). The changing pattern of alcohol consumption in England and Wales 1978-85. British Medical Journal, 295, 807-810. Calnan, M. (1985). Patterns in preventive behaviour: a study of women in middle age. Social Science & Medicine, 20, 263-268. Health and Welfare Canada. (1988). Canada's Health Promotion Survey: Technical Report. Ottawa: Minister of Supply and Services Canada.
140
A. Amos et al.
24. Pill, R. & Stott, N. C. H. (1985). Preventive procedures and practices among working class women: new data and fresh insights. Social Science & Medicine, 21,975-983. 25. Graham, H. (1987). Women's smoking and family health. Social Science and Medicine, 25,47-56. 26. Hayes, D. T. & Ross, C. E. (1987). Concern with appearance, health beliefs and eating habits. Journal of Health and Social Behaviour, 28, 120-130.