Bulletin of the World Health Organization

Do lifestyle interventions work in developing countries? Findings from the Isfahan Healthy Heart Program in the Islamic Republic of Iran

Nizal Sarrafzadegan a, Roya Kelishadi a, Ahmad Esmaillzadeh b, Noushin Mohammadifard a, Katayoun Rabiei a, Hamidreza Roohafza a, Leila Azadbakht b, Ahmad Bahonar a, Gholamhossein Sadri a, Ahmad Amani c, Saeid Heidari c & Hossein Malekafzali d

a. Cardiovascular Research Center, Isfahan University of Medical Sciences, PO Box 81465-1148, Isfahan, Islamic Republic of Iran.
b. School of Public Health, Isfahan University of Medical Sciences, Isfahan, Islamic Republic of Iran.
c. Arak University of Medical Sciences, Arak, Islamic Republic of Iran.
d. School of Public Health, Tehran University of Medical Sciences, Tehran, Islamic Republic of Iran.

Correspondence to Nizal Sarrafzadegan (e-mail: nizal.sarrafzadegan@gmail.com).

(Submitted: 23 November 2007 – Revised version received: 21 June 2008 – Accepted: 25 June 2008 – Published online: 04 November 2008.)

Bulletin of the World Health Organization 2009;87:39-50. doi: 10.2471/BLT.07.049841

Introduction

Non-communicable diseases currently represent 43% of the global burden of disease and are expected to account for 60% of the disease burden and 73% of all deaths in the world by 2020.1 Most of this increase will reflect non-communicable disease epidemics in developing countries resulting from the epidemiological transition, recent changes in diet and social environment, and the adoption of lifestyles resembling those of developed societies.24 In developing countries, lifestyle-related chronic diseases, particularly cardiovascular disease, heavily burden the health-care system.5,6 It has been estimated that an unhealthy diet and physical inactivity alone accounted for approximately 20% of the deaths among adults in the United States of America in 2000,7 and the figures could be even higher in developing countries.46 Cross-sectional and prospective studies have shown that the prevalence and incidence of many chronic conditions, including obesity, atherosclerosis, coronary heart disease and certain cancers, are increased by unhealthy lifestyles,812 particularly an unhealthy diet, physical inactivity, smoking and stress. Therefore, lifestyle modification, long considered the cornerstone of interventions, is extremely important in reducing the burden of chronic diseases.

Several intervention trials have reported the effects of lifestyle intervention programmes among high-risk populations.1318 Some have recently shown a 58% decrease in the incidence of diabetes in individuals with impaired glucose tolerance.19,20 Others have reported the beneficial effects of lifestyle modification on blood pressure control.21,22 Lifestyle interventions seem to be at least as effective as drugs.23

Despite the above, recent reviews have cast doubt on whether lifestyle interventions really help reduce multiple cardiac risk factors.24 For developing countries the evidence is less clear, and intervention studies in such countries have been scarce. Moreover, most intervention studies in the developing world have targeted specific population groups, rather than the whole community. In the Islamic Republic of Iran the Isfahan Healthy Heart Program (IHHP), which relies on comprehensive community-based lifestyle interventions to improve diet, physical activity, smoking behaviour and stress management,25 has provided an opportunity to assess whether such interventions really work in developing countries. The specific objective of this study is to evaluate the effects of this comprehensive, integrated community-based lifestyle intervention on diet, physical activity and smoking behaviours.

Methods

Population

The study design and rationale of the intervention methods employed in the IHHP have been described elsewhere.26 In this study, which was initiated in the year 2000, two intervention counties (Isfahan and Najaf-Abad) and a control area (Arak), all located in central Islamic Republic of Iran, were studied. According to the 2000 national census, the population was 1 895 856 in Isfahan and 275 084 in Najaf-Abad, a neighbouring county. Arak, a county with a population of 668 531 located 375 km north-west of Isfahan, was selected as a control area because it resembled the intervention areas in its socioeconomic, demographic and health profile and offered good cooperation.26 The intervention programme targeted the general population as well as specific groups in urban and rural areas within the intervention communities. Arak was monitored for evaluation purposes but did not receive any intervention. In each community, a random sample of adults was selected yearly by multi-stage cluster sampling. To achieve adequate sample size, those who declined to participate in the study were replaced by their neig hbours. Assessments of dietary intake, physical activity and smoking behaviour were made at baseline and annually for up to 4 years in the intervention areas and up to 3 years in the control area. For each annual evaluation, an independent random sample of adult community residents was randomly selected by a multi-stage cluster sampling method (Table 1). Informed written consent was provided by each participant.

Interventions

The IHHP conducts integrated activities in health promotion, disease prevention, and health-care treatment and rehabilitation. In all, the programme comprises 10 distinct projects, each targeting different groups. The projects’ target populations and main interventions are listed in Table 2. Key strategies linked to intervention activities include public education through the mass media, inter-sectoral cooperation and collaboration, community participation, education and involvement of health professionals, marketing and organizational development, legislation and policy development or enforcement, and research and evaluation. The IHHP promotes healthy nutrition and increased physical activity and conducts tobacco control and stress management activities. Interventions are targeted to individuals, populations and the environment depending on the results of baseline surveys on diet, physical activity, smoking and stress management and on an assessment of needs in these areas and of their coverage by existing health services. The directors of different interventions work intensively and closely with representatives of the mass media (television, newspapers, radio, etc.), health professionals (administrators, physicians, nurses, health workers and volunteers, social workers, etc.), business and market leaders (from the food industry, groceries, bakeries, fast food shops, etc.), key staff of nongovernmental organizations, and local political decision-makers (county, municipal and provincial leaders). The IHHP’s organization and interventions are described in greater detail elsewhere.27

Behaviour assessments

Assessments of diet, physical activity and smoking behaviour were performed at baseline (year 2000) and annually for 4 years in the intervention areas and for 3 years in the control area. An insufficient budget made it impossible to collect data in the control area during the fourth year. The usual dietary intake was assessed using a 49-item food frequency questionnaire (FFQ) listing foods commonly consumed by Iranians and administered by trained technicians. For each food item, participants were asked to report common portion sizes and consumption frequency during the previous year. The latter was recorded in terms of daily (e.g. bread), weekly (e.g. rice, meat) and monthly (e.g. fish) consumption, and the daily intake of each food was derived by dividing weekly consumption by 7 and monthly consumption by 30.

Data on physical activity, expressed as metabolic equivalent task (MET) minutes per week, were obtained through an oral questionnaire that included questions on four activity domains: job-related physical activity; transportation-related physical activity; housework and house maintenance activities; recreation, sport and leisure-time physical activity. We asked participants to think about all the vigorous and moderate activities they had performed in the last 7 days, considering the number of days a week and the time spent on these activities. Additional information regarding age, sex, smoking behaviour, place of residence and educational level was collected using a questionnaire. Several questions on smoking behaviours were asked, with the following key questions used to categorize individuals: “Are you currently smoking (cigarettes, pipe, and hookah)?” and “What is the frequency of smoking in a day, week or month”? We categorized individuals as current smokers if they smoked ≥ 1 times a day. In this study, smoking included the use of cigarettes, a pipe or a hookah.

Definition of low-risk groups

We defined low-risk groups in terms of dietary intake, smoking habits and physical activity. For dietary intake, the 49 food items on the food frequency questionnaire were first classified into 12 food groups, as follows: (i) fruits, (ii) vegetables, (iii) dairy products, (iv) non-hydrogenated vegetable oils, (v) legumes, (vi) nuts, (vii) white meat, (viii) grains, (ix) hydrogenated vegetable oils, (x) red meat, (xi) processed meat, (xii) sweets and pizza. We then quantified participants’ intakes of foods from these groups and divided the participants into quintiles according to their intakes. Individuals in the two highest intake quintiles for fruits, vegetables, dairy products, non-hydrogenated vegetable oils, legumes, nuts and white meat were classified as having a healthy diet and were given a score of 1 for each food group, while those in the lowest, second and third intake quintiles of these food groups were given a score of 0. For unhealthy food groups like grains, hydrogenated vegetable oils, red meat and processed meat, sweets and pizza, the opposite was done: individuals in the lowest and second quintiles were given a score of 1 and those in the three highest quintiles were given a score of 0. All grains were classified as unhealthy because the ones ordinarily consumed in Iran are refined rather than whole. Pizza plus sweets were counted as a single unhealthy food group because both are commonly consumed in Iran and contain harmful fats, such as trans-fats. It was not possible to separate low- and high-fat dairy products because the distinction was not made in the consumption questionnaire, so all dairy products were classified as a single, healthy food group. The total dietary score was calculated as the sum of the scores given for all 12 food groups. Thus, the total dietary score for each individual could vary from 0–12. Individuals whose total dietary score was ≥ 8 were classified as being on a healthy diet and assigned a risk score of 1, whereas those with a dietary score of < 8 were classified as being on a poor diet and assigned a risk score of 0.

For smoking, the low-risk group was defined as being composed of individuals who had stopped smoking or had never smoked. In terms of physical activity, the low-risk group was composed of individuals who spent an average of half an hour per day doing vigorous or moderate exercise. Finally, we summed up the scores for diet (low-risk = 1, others = 0), smoking (low-risk = 1, others = 0) and physical activity (low-risk = 1, others = 0) to come up with a lifestyle score.

Statistical analysis

The Statistical Package for the Social Sciences version 10.0 (SPSS Inc., Chicago, IL, United States of America) was used for all statistical analyses. Linear trend analysis of variance was used to compare mean diet scores and MET minutes per week in the three different study sites across different years. The Mantel-Haenszel extension χ² test was used to assess the overall trend for categorical variables. To compare continuous and categorical variables between intervention and control areas, Student’s t test, analysis of covariance (controlling for age, sex and baseline values) and χ² test were used where appropriate. The means of lifestyle score were computed separately by residence area and linear trend analysis of variance was used to assess the overall trend shown by this score across annual evaluations. Cumulative proportions of individuals with different lifestyle scores (0–3) were calculated in different years separately for intervention and control areas.

Results

The characteristics of the study participants in terms of diet, smoking habits and physical activity at baseline and after the intervention in the different sites are presented in Table 3. The baseline mean dietary score in Isfahan and Najaf-Abad (intervention areas) and in Arak (control area) was 4.5, 4.8 and 5.7, respectively. Annual evaluations revealed a significant increasing trend in mean dietary scores in both intervention areas (P for trend < 0.05), but no significant change in the control area (P for trend = 0.41). A similar pattern was seen in the percentage of individuals who ate a healthy diet. In the fourth annual evaluation, almost 40% of all individuals in Isfahan (as compared to 18% at baseline, P for trend < 0.01) and 31% in Najaf-Abad (as compared to 14% at baseline, P for trend < 0.01) ate healthy diets, while no significant trend was observed in Arak (almost 11% in the third annual evaluation as compared to 13% at baseline, P for trend = 0.38). With regard to smoking, a significant decreasing trend was seen in Isfahan (P for trend < 0.001) but not in Najaf-Abad (P for trend = 0.21). Although, a significant decreasing trend (P = 0.004) was also evident in the control area, the percentage of daily smokers in the first and second yearly evaluations was not significantly different from baseline.

Data on energy expenditure for total daily physical activity showed a significant decreasing trend in both intervention areas (P for trend was < 0.001 in Isfahan, 0.002 in Najaf-Abad) and in the control area (P for trend < 0.001). However, the opposite trend was noted in energy expenditure for leisure time physical activity, both in the intervention areas and in the control area. A decreasing trend was also found in the percentage of individuals with ≥ 30 minutes of moderate or vigorous physical activity a day in all areas (P < 0.01 for all areas). The findings were almost the same when the data were analysed separately by gender.

A comparison of lifestyle variables between intervention and control areas showed that at baseline the mean dietary score was significantly lower in the intervention areas than in the control area (4.6 versus 5.7; P < 0.01). In contrast, the percentage of individuals on a healthy diet was significantly higher in the intervention areas than in the control area (16.9 versus 12.8%, respectively; P < 0.05). Other variables did not differ significantly between intervention and control areas at baseline (Table 4), even when the data were analysed separately by gender.

Mean lifestyle scores at baseline and at each annual evaluation are shown in Fig. 1. Residents of Isfahan had a slightly better mean lifestyle score at baseline than did residents of Najaf-Abad and Arak (1.32 versus 1.22 and 1.23, respectively, P < 0.05). Following lifestyle interventions, the mean lifestyle score increased progressively in Isfahan (from baseline to the fourth annual evaluation: 1.32, 1.44, 1.57, 1.69, 1.72, respectively; P for trend < 0.05) and in Najaf-Abad (1.22, 1.31, 1.49, 1.56, 1.69, respectively; P for trend < 0.05), but not in Arak (from baseline to the third annual evaluation: 1.23, 1.17, 1.24, 1.25, respectively; P for trend = 0.15). A significantly greater change from baseline in mean lifestyle score was noted in the intervention areas when compared to the control area, even after controlling for age, sex and baseline values (data not shown).

Fig. 1. Mean lifestyle scorea in Isfahan and Najaf-Abad (intervention areas) and Arak (control area) before and after lifestyle interventions in a study of the effects of the Isfahan Healthy Heart Program, the Islamic Republic of Iran
a The lifestyle score is the mean of the sum of the scores for diet (low-risk=1, others=0), smoking (low-risk=1, others=0) and physical activity (low-risk=1, others=0). Residents of Isfahan had a slightly healthier lifestyle at baseline than did residents of Najaf-Abad and Arak.

After the lifestyle intervention, changes from baseline in mean dietary score were significantly different between intervention and control areas (+2.1 points versus –1.2 points, respectively; P < 0.01). Similarly, the change from baseline in the percentage of individuals who ate a healthy diet differed significantly between the intervention and control areas (+14.9% versus –2.0%, respectively; P < 0.001). Covariance analysis after controlling for sex, age, and baseline values yielded the same findings.

Daily smoking decreased in both intervention and control areas after the lifestyle intervention, but the changes from baseline were not significantly different between areas. When genders were analysed separately, a significant decreasing trend was found among men, but not among women (Fig. 2). Although changes from baseline in energy expenditure for both total daily physical activities and leisure time physical activities were in a downward and upward direction, respectively, when intervention and control areas were compared, mean changes from baseline in total daily physical activities were significantly lower in the intervention areas than in the control area (–68 versus –114 MET minutes per week; P < 0.05). Such findings were obtained even after adjusting for age, sex and baseline values. No significant difference was seen in the percentage of individuals with ≥ 30 minutes a day of moderate or vigorous physical activity between intervention and control areas after 4 years of the lifestyle intervention. However, in the first and second annual evaluations, participants in the intervention areas were more likely to engage in such activity.

Fig. 2. Prevalence of daily smoking, by gender, in Isfahan and Najaf-Abad (intervention areas) and in Arak (control area) before and after lifestyle interventions in a study of the effects of the Isfahan Healthy Heart Program, the Islamic Republic of Iran

The percentage of individuals with different lifestyle scores in intervention and control areas is presented in Fig. 3 and Fig. 4. At baseline, 8.7% of the individuals in the intervention areas had a lifestyle score of 0 (unhealthy lifestyle), whereas only 3.6% had a score of 3 (healthy lifestyle). In these areas, the percentage of individuals with an unhealthy lifestyle was lower after the lifestyle intervention, while the percentage of individuals with a healthy lifestyle was higher. Thus, by the fourth yearly evaluation almost two-thirds of the population had at least two (out of three) healthy lifestyle components. Such a trend was not so clear in the control area; however, the percentage of individuals with a score of 3 had increased significantly by the third evaluation as compared to baseline (P < 0.05).

Fig. 3. Percentage of individuals with different lifestyle scoresa in Isfahan and Najaf-Abad (intervention areas) before and after lifestyle interventions in a study of the effects of the Isfahan Healthy Heart Program, the Islamic Republic of Iran
a The lifestyle score is the sum of the scores for diet (low-risk=1, others=0), smoking (low-risk=1, others=0) and physical activity (low-risk=1, others=0).
Fig. 4. Percentage of individuals with different lifestyle scoresa in Arak (control area) before and after lifestyle interventions in a study of the effects of the Isfahan Healthy Heart Program, the Islamic Republic of Iran
a The lifestyle score is the sum of the scores for diet (low-risk=1, others=0), smoking (low-risk=1, others=0) and physical activity (low-risk=1, others=0).

Discussion

The findings of the present study, which was performed in a representative sample of the Iranian population, suggest that lifestyle habits can be improved by a community-based lifestyle intervention programme even in a developing country setting. After the comprehensive, integrated community-based lifestyle intervention programme, beneficial changes were noted in diet and physical activity but no substantial changes were seen in the smoking behaviour of the population, particularly among women. Total lifestyle scores and the percentage of individuals with a healthy lifestyle increased significantly in the intervention areas. The approaches followed in the IHHP closely resemble those of the North Karelia project in Finland, since both projects were controlled comprehensive, community-based interventions targeting the general population and the environment. Although Finland is not a developing country, the North Karelia project was originally carried out in fairly low-resource, semi-rural settings.

Although assessing the effects of comprehensive lifestyle intervention programmes on diet, physical activity and smoking behaviours is not new, few reports are available from developing countries.2830 Most integrated community-based programmes have been carried out in developed countries; however, a remarkably rapid increase in the burden of non-communicable diseases in developing countries has made researchers in these places undertake similar activities. Several lifestyle interventions have been carried out in these countries.3,2833 The major limitations of previous reports from developing countries have been the lack of a control area,28 the lack of annual and process evaluations, and a focus on risk factor prevalence rather than on actual changes in diet, physical activity and smoking habits.29,30 Therefore, it is not clear to what extent the changes documented in this study represent the effect of the intervention as opposed to underlying secular trends. Through comparing changes from baseline in diet, physical activity and smoking behaviour in intervention areas and in a control area, the present study shows that lifestyle interventions can work in developing countries.

Community-based lifestyle interventions have been conducted with varying results. Some have documented beneficial changes in respondents’ dietary habits;22,34 others have failed to document such changes.35 In this study, the most pronounced change from baseline after the IHHP lifestyle intervention programme was noted in the dietary habits of intervention area residents. The mean dietary score and percentage of individuals following a healthy diet increased significantly in areas where the lifestyle intervention programme was in place. Previous reports from the IHHP have also suggested that the lifestyle interventions encouraged people to choose healthy foods and prompted government authorities to make them available. Thus, policy changes in the framework of the IHHP intervention programme were such that between 2000 (baseline) and 2002, the distribution of hydrogenated and non-hydrogenated vegetable oils in intervention areas changed from 82% to 68% and from 18% to 32%, respectively, while in the control area it changed from 97% to 95% and from 3% to 5%, respectively.25,26 A decline in salt intake has also been reported in the intervention areas.25 All these data suggest that dietary habits have been affected by the IHHP intervention programme. However, in this study the consumption of specific nutrients or foods was not assessed because it was felt that the multivariate, whole diet approach would provide more information and eliminate concerns about confounding factors and co-linearity in food and nutrient intakes. Dietary changes across the population have been noted for as long as 20 years after the cessation of lifestyle modification programmes.36 Although the direct and independent effects of such beneficial dietary changes on the risk of non-communicable diseases in the target population are relatively unknown, it has been suggested that a relatively small shift towards a healthier diet in the entire population may lead to a reduction in the incidence of non-communicable diseases.37,38 The fact that in this study the most prominent changes from baseline were found in dietary habits is in line with findings reported from North Karelia,39,40 where half of the decline in mortality from coronary disease since 1972 can be explained by dietary changes across the population.39

The energy expended for total daily physical activity declined in both intervention and control areas but was less pronounced in the former than in the latter. However, the energy expended for leisure time physical activity increased in both areas, and much more so in the intervention areas than in the control area. Therefore, if the physical activity levels in the control area are assumed to be representative of the secular trend, one can conclude that the IHHP lifestyle intervention programme influenced physical activity levels in the target populations, albeit not substantially. Other community-based programmes have not been accompanied by substantial changes in physical activity, although the prevalence of physical inactivity might have decreased.41,42 However, some intensive programmes have been accompanied by improved maximum oxygen uptake,18 increased time spent on physical activities20,34 and reduced physical inactivity.34 Promoting physical activity calls not only for educating the target population, but often also for expensive facilities that can seldom be afforded by a research programme, particularly in a developing country,38 because many people mistakenly consider them necessary. Cultural factors that may hinder physical activity must also be taken into account.38

Community-based lifestyle intervention programmes have been associated in varying degrees with improvements in smoking behaviour in target populations.14,43 In the current study, daily smoking decreased in both the intervention areas and the control area after 3 years of lifestyle intervention, although changes from baseline were not significantly different between these two areas. Previous short-term reports from the IHHP have shown a significant effect of lifestyle interventions among men but not women,25 which points to the need to modify interventional activities in this group. In this study, the percentages of daily smokers in the control area in the first and second yearly evaluations were not significantly different from baseline, and there was a significant difference in changes from baseline between the intervention and control areas after 2 years of the intervention. This is in line with other studies and highlights the fact that long-term control of smoking is very difficult to attain. While various programmes have been accompanied by increased attempts at quitting and have moved participants towards being ready to quit, very few programmes have influenced long-term cessation rates.43,44 Any change that affects the prevalence of smoking in a community is also likely to affect community norms, which will in turn lead to even greater change in the community. Bringing about a shift of this kind, however, is not a trivial goal and usually requires more time than research projects allow.4547 These results serve to underscore the addictive properties of nicotine and suggest that long-term behavioural interventions and ongoing counselling may be required to influence cessation. As little evidence exists to support the effectiveness of any specific intervention on long-term cessation rates, further studies are needed to advance this field.

Several points need to be considered when interpreting the findings of this study. First, the effect of lifestyle interventions on risk factor prevalence has not been examined in the current study; data from the annual evaluations was used to document changes in the population’s behaviour. To what extent the IHHP lifestyle interventions have affected biochemical measures and biological risk factors will be assessed in the post-intervention phase of the study. Second, methodological differences between community-based intervention programmes might explain to some extent the observed differences in lifestyle changes from baseline. Third, although lifestyle interventions for the primary prevention of non-communicable diseases are cost-effective according to previous reports,48 this would differ among countries because of country-specific intervention costs.49

The high response rates in our study can be explained by the fact that the samples for different years were independent and that the authors followed the whole community, not just the same individuals who comprised the sample in the first year of the study. Because of this, the response rate in this study is higher than is commonly found in cohort studies. Another reason is that free services were provided to study participants. In the invitation letters to residents, it was highlighted that all medical services would be free for participants, even if these involved additional medical consultations and expensive treatments. This point is important in developing countries like the Islamic Republic of Iran, where many are barely able to afford medical services.

In conclusion, comprehensive community-based lifestyle interventions, even in developing countries, can change the environment of the entire community to support healthier lifestyles. This would reduce health-risk behaviours, which would in turn decrease chronic disease morbidity and mortality. ■


Funding: The Isfahan Cardiovascular Research Center funded the project.

Competing interests: None declared.

References

Share