Abstract

In this systematic review, the authors aimed to assess the effectiveness of community programs for prevention of cardiovascular disease (CVD). They searched numerous electronic databases (CDSR, DARE, HTA, EED, and CENTRAL via the Cochrane Library, MEDLINE, MEDLINE In Process, EMBASE, CINAHL, PsycINFO, HMIC, and ASSIA) and relevant Web sites from January 1970 to mid-July 2008. Controlled studies of community programs for the primary prevention of CVD were included. Net changes in CVD risk factors were used to generate an overall index for net change in 10-year CVD risk. The authors identified 36 relevant community programs that took place between 1970 and 2008. These programs were multifaceted interventions employing combinations of media, screening, and counseling activities and environmental changes and were primarily evaluated using controlled before-after studies. In 7 studies, investigators reported changes in CVD/total mortality rates, and in 5 they reported net changes. In all cases, these net changes were positive but were largely nonsignificant. In 22 studies, investigators reported changes in physiologic CVD risk factors, and there was a positive trend in the calculated CVD risk score. The average net reduction in 10-year CVD risk was 0.65%. Community programs for CVD prevention appear to have generally achieved favorable changes in overall CVD risk and, with adaptation to current circumstances, deserve continued consideration as possible approaches to preventing CVD.

Cardiovascular disease (CVD) is the biggest cause of mortality in the Western world and contributes a sizeable burden of cost to health services (1). It has been demonstrated that risk factors such as smoking (2), obesity (3), and elevated cholesterol (4) and blood pressure (5) can increase the risk of developing CVD, but there is debate about the most effective approach for risk factor change (6).

Community approaches to CVD prevention are attractive, since they can target all groups in the community and, if effective, may achieve widespread behavioral change and risk reduction. A number of community CVD prevention programs have been implemented over the last 40 years. Despite their continued development and use, there is currently limited evidence to support or refute the effectiveness of these types of programs.

Few reviews have addressed the effectiveness of community CVD prevention. Of those that have, 1 included only studies targeting dietary factors (7), 1 evaluated only interventions in schoolchildren (8), and 2 included only randomized controlled trials and predominantly covered prevention in high-risk groups (9, 10). One review did cover community programs (11), but, as with all of these reviews, the investigators undertook literature searches prior to 1998 and did not include more recent programs, which are potentially of the most relevance and applicability to the present. To our knowledge, no reviewers to date have attempted to combine risk factor changes into a single measure of CVD risk.

The initial stages of this systematic review were conducted to inform the development of public health guidance on CVD prevention by the National Institute for Health and Clinical Excellence (NICE). NICE is an independent United Kingdom organization that is responsible for providing national guidance on promoting good health and preventing and treating ill health. We report results from this and subsequent work, in which a single index of overall CVD risk was generated to assess the effectiveness of community CVD programs.

MATERIALS AND METHODS

Search strategy

We searched MEDLINE, MEDLINE In Process, EMBASE, CINAHL, PsycINFO, HMIC, ASSIA, and the Cochrane Library databases (CDSR, DARE, HTA, EED, and CENTRAL) for systematic reviews and primary studies relating to population-level CVD programs that had been published between January 1970 and mid-July 2008. A highly sensitive search strategy was used covering terms relating to the 3 key concepts: CVD, health education, and risk factors for CVD. The MEDLINE search strategy is given in Appendix Table 1.

Eligibility criteria

Community programs were defined as those targeting whole populations living within a certain geographic area. The size of target areas was not specified, but programs were required to have aimed to bring about change across the whole community within that area. Programs were required to have targeted multiple CVD risk factors and to have included primary prevention strategies to tackle at least 2 of the following: smoking, poor diet, insufficient physical activity, high blood pressure, high blood cholesterol levels, obesity/overweight, diabetes, psychosocial stress, and high alcohol consumption. Acceptable study designs were interventional studies such as randomized controlled trials, controlled before-after studies, and controlled interrupted time series. Intervention programs had to have specifically aimed to address CVD with the goal of reducing morbidity/mortality from CVD or reducing levels of CVD risk factors and had to have evaluated 1 or more of the following outcomes: CVD mortality; CVD morbidity; biochemical precursors of CVD, including lipid levels, ratio of high density lipoprotein to low density lipoprotein, and triglyceride levels; physiologic precursors of CVD, including cholesterol, blood pressure, and body mass index; behaviors associated with the risk of CVD, including smoking, diet, physical activity, and alcohol consumption; knowledge, attitudes, and intentions regarding CVD; and adverse events.

Studies confined to populations clinically diagnosed as being at high risk of CVD or diagnosed with CVD were excluded, as well as any studies undertaken in countries that are not part of the Organisation for Economic Co-operation and Development. Studies were excluded if they did not include a control group or if the investigators did not give data on any relevant outcome measures. Books, book chapters, theses, and dissertations were excluded. Inclusion decisions were made by 1 reviewer, with reference to a second reviewer in the case of uncertainty.

Quality assessment and data extraction

Quality assessment of included studies was undertaken using the controlled before-after checklist in Methods for Development of NICE Public Health Guidance (12). Studies were scored with regard to 8 items: contemporaneous data collection, appropriate choice of control site, similarity of baseline measurements, similarity of study/control providers, blinded outcome assessment, protection against contamination, reliability of outcome measures, and follow-up of individuals (for cohort studies only), with higher scores indicating programs considered to be of higher quality. Study quality was assessed independently by 2 reviewers, with differences being resolved by consensus. Two reviewers independently extracted data on baseline and follow-up levels of CVD risk factors, and any differences were resolved by consensus.

Statistical analysis

Net changes in mortality rates were calculated as the change in mortality rate from baseline to the intervention/postintervention period in the intervention group minus the change in the control group. Where available, data for CVD mortality rates were recorded; for studies where this was not reported, total mortality rates were recorded.

Net changes in individual risk factors were calculated as risk factor change from baseline to follow-up in the intervention group minus change in mortality rate in the control group. Average values for risk factor changes were calculated from mean changes for each program—that is, average change in men and women and in cohort and cross-sectional surveys was first calculated for each program before calculation of the average net change across programs.

In order to estimate the number of cardiovascular events prevented, we determined the net change in CVD risk pre- to postintervention. CVD risk was calculated from age, sex, diabetes status, smoking status, systolic blood pressure, diastolic blood pressure, total cholesterol, and high density lipoprotein cholesterol using the Framingham equation (13). Because individual participants’ risk factor data were not available, mean ages and risk factor levels were used for the intervention population as a whole. Where a risk factor value was unavailable, we used a default value based on the age- and sex-stratified population average, derived from the Health Survey for England 1998 (14). This procedure followed a previously described method (15). Increases in age from the start to the end of the intervention period have a large impact on calculated CVD risk, but this does not reflect the effectiveness of the intervention. In order to avoid the spurious influence of increasing age in cross-sectional surveys, we also used baseline age to calculate CVD risk for the final survey, and for cohort data, we used baseline age plus duration of the intervention to calculate CVD risk at follow-up.

Graphical representations of net changes in 10-year CVD risk were generated on the basis of all programs with more than 2 relevant outcome measures. Formal meta-analysis could not be conducted because of the lack of statistical information in the majority of included studies. Mean net changes in 10-year CVD risk and 10-year CVD risk relative to baseline were calculated separately for programs that used cross-sectional surveys for evaluation and programs that used cohort surveys for evaluation. Mean net changes over all programs were calculated using both cross-sectional and cohort data; where evaluation had been done using both survey designs, cohort data were used, since this type of data was considered less at risk of bias compared with cross-sectional data (see Discussion). For studies where data were presented separately by sex, mean net changes were also calculated separately for men and women.

RESULTS

Trial flow and study characteristics

The search of the literature generated 3,153 reviews and 36,622 primary study citations (Figure 1). We identified 102 publications as relevant to the current review and identified 36 community CVD prevention programs (Table 1) (only the main index publications for each program are cited (16–50)).

Table 1.

Characteristics of Programs Included in a Systematic Review of the Effectiveness of Community Programs for Prevention of Cardiovascular Disease, 1970–2008

Program (Reference No.) Country Start Date Length of Follow-up, years Population Size Outcome(s) Measured
 
Evaluation Method
 
QA Scorea 
Mean TC Level Mean HDL Cholesterol Level Mean SBP Mean DBP Smoking, % Cohort Cross-Sections 
Action Heart (16United Kingdom 1991 22,000      
Bootheel Heart Health Project (17United States 1989 ∼70,000      
CardioVision 2020 (18United States 1999 87,685      
Coalfields Healthy Heartbeat Program (19Australia 1990 10 46,520   
Coeur en Santé St.-Henri (20Canada 1992 3.5 25,000      
Di.S.Co. Project (21Italy 1982 26,000  
Danish Municipality Project (22Denmark 1989 8,000      
Dutch Heart Health Community Intervention (23Netherlands 1998 >6 180,000    
Eberbach-Wiesloch Project (24Germany 1980 16,000   
Epernon Cardiovascular Health Programme (25France 1993 5,500     
Health and Inequality in Finnmark Programme Norway 1988          
        Båtsfjord (26   2,500   
        North Cape (27   4,000   
Franklin Cardiovascular Health Program (28United States 1974 20 29,645       
German Cardiovascular Prevention Project (29Germany 1985 1,000,000   
Have a Heart Paisley (30United Kingdom 2000 2.5 74,170       
Heartbeat Wales (31United Kingdom 1985 >5 ∼3,000,000      
Kentucky Blood Pressure Control Program (32United States 1979 32,400     
Kilkenny Health Project (33Ireland 1985 73,000   
Martignacco Project (24Italy 1977 5,259   
Minnesota Heart Health Program (34United States 1980 232,000  
Norsjö Project (35, 36Sweden 1985 10 5,500  
North Coast Healthy Lifestyle Programme (37Australia 1978 22,083 Yb     
North Karelia Project (38Finland 1972 >20 180,000   
National Research Program (39Switzerland 1978 28,000   
Olöfstrom Community Program (40Sweden 1988 15,000    
Otsego-Schoharie Heart Health Program (41United States 1989 100,000 
Pawtucket Heart Health Program (42United States 1982 71,000  
Quebec Heart Health Demonstration Project (43Canada 1992          
        Rural    90,000       
        Suburban    35,216       
        Urban    41,625       
Project Schleiz (44Germany 1976 33,000   
South Carolina Cardiovascular Disease Prevention Project (45United States 1987 46,000      
Stanford Heart Disease Prevention Program (Three-Community Study) (46United States 1972 42,000     
Stanford Five-City Project (47, 48United States 1979 126,000  
Tessin Cardiovascular Health Programme (49Switzerland 1982 >300,000      
Zurich Cardiovascular Health Programme (50Switzerland 1974 ∼600   
Program (Reference No.) Country Start Date Length of Follow-up, years Population Size Outcome(s) Measured
 
Evaluation Method
 
QA Scorea 
Mean TC Level Mean HDL Cholesterol Level Mean SBP Mean DBP Smoking, % Cohort Cross-Sections 
Action Heart (16United Kingdom 1991 22,000      
Bootheel Heart Health Project (17United States 1989 ∼70,000      
CardioVision 2020 (18United States 1999 87,685      
Coalfields Healthy Heartbeat Program (19Australia 1990 10 46,520   
Coeur en Santé St.-Henri (20Canada 1992 3.5 25,000      
Di.S.Co. Project (21Italy 1982 26,000  
Danish Municipality Project (22Denmark 1989 8,000      
Dutch Heart Health Community Intervention (23Netherlands 1998 >6 180,000    
Eberbach-Wiesloch Project (24Germany 1980 16,000   
Epernon Cardiovascular Health Programme (25France 1993 5,500     
Health and Inequality in Finnmark Programme Norway 1988          
        Båtsfjord (26   2,500   
        North Cape (27   4,000   
Franklin Cardiovascular Health Program (28United States 1974 20 29,645       
German Cardiovascular Prevention Project (29Germany 1985 1,000,000   
Have a Heart Paisley (30United Kingdom 2000 2.5 74,170       
Heartbeat Wales (31United Kingdom 1985 >5 ∼3,000,000      
Kentucky Blood Pressure Control Program (32United States 1979 32,400     
Kilkenny Health Project (33Ireland 1985 73,000   
Martignacco Project (24Italy 1977 5,259   
Minnesota Heart Health Program (34United States 1980 232,000  
Norsjö Project (35, 36Sweden 1985 10 5,500  
North Coast Healthy Lifestyle Programme (37Australia 1978 22,083 Yb     
North Karelia Project (38Finland 1972 >20 180,000   
National Research Program (39Switzerland 1978 28,000   
Olöfstrom Community Program (40Sweden 1988 15,000    
Otsego-Schoharie Heart Health Program (41United States 1989 100,000 
Pawtucket Heart Health Program (42United States 1982 71,000  
Quebec Heart Health Demonstration Project (43Canada 1992          
        Rural    90,000       
        Suburban    35,216       
        Urban    41,625       
Project Schleiz (44Germany 1976 33,000   
South Carolina Cardiovascular Disease Prevention Project (45United States 1987 46,000      
Stanford Heart Disease Prevention Program (Three-Community Study) (46United States 1972 42,000     
Stanford Five-City Project (47, 48United States 1979 126,000  
Tessin Cardiovascular Health Programme (49Switzerland 1982 >300,000      
Zurich Cardiovascular Health Programme (50Switzerland 1974 ∼600   

Abbreviations: DBP, diastolic blood pressure; Di.S.Co., Sezze District Community Control; HDL, high density lipoprotein; QA, quality assessment; SBP, systolic blood pressure; TC, total cholesterol; Y, yes.

a

Quality assessment was conducted using the controlled before-after checklist in Methods for Development of NICE Public Health Guidance (12). Possible scores ranged from 0 to 8.

b

No control group data.

Figure 1.

Selection of studies included in a systematic review of the effectiveness of community programs for prevention of cardiovascular disease (CVD), 1970–2008.

Figure 1.

Selection of studies included in a systematic review of the effectiveness of community programs for prevention of cardiovascular disease (CVD), 1970–2008.

The programs were based in Europe (n = 20), the United States (n = 10), Canada (n = 4), and Australia (n = 2). All programs included a media-based approach, using combinations of radio, television, and printed material to communicate heart health messages. Many programs also included screening interventions (n = 19), individual (n = 21) and group (n = 17) counseling, and environmental changes (n = 14). Health departments (n = 22), local health committees (n = 13), voluntary organizations (n = 11), and community volunteers (n = 11) had roles in program delivery, and interventions were delivered in a variety of settings, including workplaces (n = 13) and schools (n = 21). Program lengths ranged from 1 year to more than 20 years during the period 1970–2008, and the sizes of the intervention populations ranged from approximately 600 people to over 1,000,000 people.

The majority of programs were evaluated using controlled before-after studies (n = 33), and 3 were evaluated using controlled interrupted time series. Risk factor levels in control and experimental groups were measured preintervention, postintervention, and, in some cases, during the intervention period. For data collection, researchers in the majority of programs (n = 28) used independent cross-sectional surveys, some followed up cohorts (n = 15), and some used both cross-sectional and cohort approaches (n = 7) (Table 1). Program evaluation varied in quality (Table 1). In the majority of studies, protection against contamination was not clear. In some studies, intervention and control groups were not similar at baseline and only self-reported outcome measures were used, resulting in overall poorer quality scores.

Programs assessed different outcome measures, including CVD and total mortality (n = 7) and physiologic and behavioral CVD risk factors, such as total cholesterol level (n = 21), blood pressure (n = 22), smoking (n = 27), diet (n = 18), and physical activity (n = 18). Of the 36 included programs, 22 provided sufficient information on physiologic CVD risk factors (cholesterol, blood pressure, and smoking) to calculate the net treatment effect on 10-year CVD risk.

Changes in mortality rates

In 7 studies, investigators reported changes in mortality rates (primarily CVD mortality), and results are given in Table 2. Data were reported as change and percent change in CVD mortality rate and change in total mortality rate. In 2 studies, researchers presented postintervention mortality rates but not baseline mortality rates; therefore, it was not possible to calculate net change for those studies. Where net changes could be calculated, all studies showed favorable changes in CVD/total mortality rates, but this was significant for only 1 study (32); in that case, only men showed a significant net reduction in the rate of CVD mortality.

Table 2.

Changes in Rates of Mortality From Cardiovascular Disease After Implementation of Community Heart Health Programs, 1970–2008

Program (Reference No.) Outcome and Sex Control Group Intervention Group Net Change P Valuea 
Coalfields Healthy Heartbeart Program (19Change in CVD mortality (no. of deaths/100,000 persons/year)     
     Men −7.0 −10.9 −3.9 NS 
     Women −7.8 −14.2 −6.4 NS 
Franklin Cardiovascular Health Program (28Relative risk of death from CVD during program periodb (both sexes combined)  0.91 NR NR 
Kentucky Blood Pressure Control Program (32% change in CVD mortality     
     Men −8.6 −36.0 −27.4 <0.04 
     Women −11.6 −21.2 −9.6 NS 
Martignacco Project (24CVD mortality during projectb (mean annual incidence/1,000 persons)     
     Men 2.4 1.2 NR NR 
     Women 1.4 0.7 NR NR 
Project Schleiz (44% change in CVD mortality     
     Men 4.8 2.0 −2.8 NS 
     Women 5.4 −3.5 −8.9 NS 
Stanford Five-City Project (47, 48All-cause mortalityc (no. of deaths/1,000 persons/10 years)     
     Men 2.0 −1.1 −3.1 0.447 
     Women 0.9 0.4 −0.5 0.795 
Tessin Cardiovascular Health Programme (49% change in CVD mortality     
     Men −14.7 −26.9 −12.2 NR 
     Women −20.0 −25.0 −5.0 NR 
Program (Reference No.) Outcome and Sex Control Group Intervention Group Net Change P Valuea 
Coalfields Healthy Heartbeart Program (19Change in CVD mortality (no. of deaths/100,000 persons/year)     
     Men −7.0 −10.9 −3.9 NS 
     Women −7.8 −14.2 −6.4 NS 
Franklin Cardiovascular Health Program (28Relative risk of death from CVD during program periodb (both sexes combined)  0.91 NR NR 
Kentucky Blood Pressure Control Program (32% change in CVD mortality     
     Men −8.6 −36.0 −27.4 <0.04 
     Women −11.6 −21.2 −9.6 NS 
Martignacco Project (24CVD mortality during projectb (mean annual incidence/1,000 persons)     
     Men 2.4 1.2 NR NR 
     Women 1.4 0.7 NR NR 
Project Schleiz (44% change in CVD mortality     
     Men 4.8 2.0 −2.8 NS 
     Women 5.4 −3.5 −8.9 NS 
Stanford Five-City Project (47, 48All-cause mortalityc (no. of deaths/1,000 persons/10 years)     
     Men 2.0 −1.1 −3.1 0.447 
     Women 0.9 0.4 −0.5 0.795 
Tessin Cardiovascular Health Programme (49% change in CVD mortality     
     Men −14.7 −26.9 −12.2 NR 
     Women −20.0 −25.0 −5.0 NR 

Abbreviations: CVD, cardiovascular disease; NR, not reported; NS, not significant.

a

P value for significance of net change in mortality rate.

b

Changes from baseline were not reported.

c

Rates of CVD mortality were not reported.

Risk factor changes

Net changes in individual CVD risk factors were mixed but generally showed a trend towards a positive program effect (Table 3). There appeared to be reasonable consistency in outcome changes—that is, where positive effects were seen in 1 risk factor, there also tended to be positive changes in others. As a general indication of the size of effects, the mean net changes in systolic and diastolic blood pressure were −2.9 mm Hg and −1.1 mm Hg, respectively; mean net change in total cholesterol level was −0.01 mmol/L, and there was a net reduction in smoking prevalence of 1.7%. However, since there was no formal meta-analysis, these may not necessarily represent valid summary estimates.

Table 3.

Net Changes in Levels of Cardiovascular Disease Risk Factors and Overall 10-Year Risk of Cardiovascular Disease After Implementation of Community Heart Health Programs, 1970–2008

Program (Reference No.) and Survey Design Sex Net Change
 
SBP, mm Hg DBP, mm Hg TC, mmol/L Smoking, % 10-Year CVD Risk, % 
Coalfields Healthy Heartbeart Program (19      
    Cross-sectional Men −2.1 −2.2 −0.1 9.4 0.15 
    Cross-sectional Women 2.2 0.9 0.3 12.8 1.47 
Di.S.Co. Project (21      
    Cross-sectional Men −0.5 −0.3 0.3 −2.8 0.14 
    Cross-sectional Women −4.9 3.6 0.0 −5.6 −0.77 
    Cohort Men 0.5 0.3 0.1 0.3 0.38 
    Cohort Women −4.2 −1.8 0.0 −1.2 −0.55 
Dutch Heart Health Community Intervention (23      
    Cohort Men −8.2 −5.3 0.0  −2.31 
    Cohort Women −5.8 −4.5 0.1  −1.06 
Eberbach-Wiesloch Project (24      
    Cross-sectional Men −10 −6 0.3 12 −0.38 
    Cross-sectional Women −7 −2 0.1 −24 −1.30 
Epernon Cardiovascular Health Programme (25      
    Cohort Both −3.8   −1.9 −0.66 
Health and Inequality in Finnmark Programme       
    Båtsfjord (26      
        Cohort Men −2.2 −2.9 0.2 −4.8 −1.11 
        Cohort Women −2.5 −3.9 0.0 −0.6 −0.98 
    North Cape (27      
        Cohort Men 1.0 0.8 −0.1 1.9 0.12 
        Cohort Women 1.1 −0.2 −0.1 −7.2 −0.52 
German Cardiovascular Prevention Project (29      
    Cross-sectional Men −2.1 −1.3 −0.1 −4.5 −0.79 
    Cross-sectional Women −3.1 −1.9 −0.1 −0.4 −0.45 
Kentucky Blood Pressure Control Program (32      
    Cross-sectional Men −3.8 −7.6   −0.69 
    Cross-sectional Women 2.9 −2.3   0.31 
Kilkenny Health Project (33      
    Cross-sectional Men 0.0 5.9 0.4 1.9 0.03 
    Cross-sectional Women −1.1 1.9 0.0 −2.9 −0.11 
Martignacco Project (24      
    Cohort Men −14 −7 −0.4 −6.5 −5.16 
    Cohort Women −14 −3 −0.1 −2.36 
Minnesota Heart Health Program (34
    Cross-sectional Both 1.0 −0.2 0.0 −1.8 0.27 
    Cohort Both 1.4 0.6 0.1 2.6 0.10 
Norsjö Project (35, 36      
    Cross-sectional Men −5.6 −3.5 −0.6 −2 −1.84 
    Cross-sectional Women −4.4 −0.5 −0.3 −0.76 
    Cohort Men −3.6 −1.6 −0.7 −3 −2.30 
    Cohort Women −5.4 −1.7 −0.6 −4 −1.55 
North Karelia Project (38      
    Cross-sectional Men −4.4 −1.0 −0.2 −2.7 −1.35 
    Cross-sectional Women −7.7 −1.5 0.0 −0.1 −0.99 
National Research Program (39      
    Cohort Both 4.9 3.9 0.0 −3.6 0.40 
Olöfstrom Community Programme (40
    Cross-sectional Men  −5.4 −0.6  −0.90 
    Cross-sectional Women  −3.7 −0.5  −0.40 
    Cohort Men  −4.3 −0.4  −0.74 
    Cohort Women  −2.9 −0.3  −0.32 
Otsego-Schoharie Heart Health Program (41      
    Cross-sectional Both 0.4 2.0 0.0 −12.7 −0.90 
    Cohort Both −1.6 1.1 0.1 −7.5 −0.84 
Pawtucket Heart Health Program (42
    Cross-sectional Both −0.6 −1.2 0.0 0.3 0.00 
    Cohort Both 0.6 −0.7 0.0 −1.4 −0.14 
Project Schleiz (44      
    Cross-sectional Both −5.4 −5.0 −0.3 −3.2 −1.51 
Stanford Three-Community Study (46
    Cohort Both −10.1  −0.2  −1.45 
Stanford Five-City Project (47, 48      
    Cross-sectional Men −3.5 −2.5 −0.1 4.1 −0.50 
    Cross-sectional Women −4.8 −4.1 −0.1 6.3 −0.29 
    Cohort Both −5.2 −3.7 0.0 −3.9 −0.49 
Zurich Cardiovascular Health Programme (50      
    Cross-sectional Men 6.4 6.9 0.6 1.5 3.10 
    Cross-sectional Women 3.2 0.3 −0.1 0.91 
Program (Reference No.) and Survey Design Sex Net Change
 
SBP, mm Hg DBP, mm Hg TC, mmol/L Smoking, % 10-Year CVD Risk, % 
Coalfields Healthy Heartbeart Program (19      
    Cross-sectional Men −2.1 −2.2 −0.1 9.4 0.15 
    Cross-sectional Women 2.2 0.9 0.3 12.8 1.47 
Di.S.Co. Project (21      
    Cross-sectional Men −0.5 −0.3 0.3 −2.8 0.14 
    Cross-sectional Women −4.9 3.6 0.0 −5.6 −0.77 
    Cohort Men 0.5 0.3 0.1 0.3 0.38 
    Cohort Women −4.2 −1.8 0.0 −1.2 −0.55 
Dutch Heart Health Community Intervention (23      
    Cohort Men −8.2 −5.3 0.0  −2.31 
    Cohort Women −5.8 −4.5 0.1  −1.06 
Eberbach-Wiesloch Project (24      
    Cross-sectional Men −10 −6 0.3 12 −0.38 
    Cross-sectional Women −7 −2 0.1 −24 −1.30 
Epernon Cardiovascular Health Programme (25      
    Cohort Both −3.8   −1.9 −0.66 
Health and Inequality in Finnmark Programme       
    Båtsfjord (26      
        Cohort Men −2.2 −2.9 0.2 −4.8 −1.11 
        Cohort Women −2.5 −3.9 0.0 −0.6 −0.98 
    North Cape (27      
        Cohort Men 1.0 0.8 −0.1 1.9 0.12 
        Cohort Women 1.1 −0.2 −0.1 −7.2 −0.52 
German Cardiovascular Prevention Project (29      
    Cross-sectional Men −2.1 −1.3 −0.1 −4.5 −0.79 
    Cross-sectional Women −3.1 −1.9 −0.1 −0.4 −0.45 
Kentucky Blood Pressure Control Program (32      
    Cross-sectional Men −3.8 −7.6   −0.69 
    Cross-sectional Women 2.9 −2.3   0.31 
Kilkenny Health Project (33      
    Cross-sectional Men 0.0 5.9 0.4 1.9 0.03 
    Cross-sectional Women −1.1 1.9 0.0 −2.9 −0.11 
Martignacco Project (24      
    Cohort Men −14 −7 −0.4 −6.5 −5.16 
    Cohort Women −14 −3 −0.1 −2.36 
Minnesota Heart Health Program (34
    Cross-sectional Both 1.0 −0.2 0.0 −1.8 0.27 
    Cohort Both 1.4 0.6 0.1 2.6 0.10 
Norsjö Project (35, 36      
    Cross-sectional Men −5.6 −3.5 −0.6 −2 −1.84 
    Cross-sectional Women −4.4 −0.5 −0.3 −0.76 
    Cohort Men −3.6 −1.6 −0.7 −3 −2.30 
    Cohort Women −5.4 −1.7 −0.6 −4 −1.55 
North Karelia Project (38      
    Cross-sectional Men −4.4 −1.0 −0.2 −2.7 −1.35 
    Cross-sectional Women −7.7 −1.5 0.0 −0.1 −0.99 
National Research Program (39      
    Cohort Both 4.9 3.9 0.0 −3.6 0.40 
Olöfstrom Community Programme (40
    Cross-sectional Men  −5.4 −0.6  −0.90 
    Cross-sectional Women  −3.7 −0.5  −0.40 
    Cohort Men  −4.3 −0.4  −0.74 
    Cohort Women  −2.9 −0.3  −0.32 
Otsego-Schoharie Heart Health Program (41      
    Cross-sectional Both 0.4 2.0 0.0 −12.7 −0.90 
    Cohort Both −1.6 1.1 0.1 −7.5 −0.84 
Pawtucket Heart Health Program (42
    Cross-sectional Both −0.6 −1.2 0.0 0.3 0.00 
    Cohort Both 0.6 −0.7 0.0 −1.4 −0.14 
Project Schleiz (44      
    Cross-sectional Both −5.4 −5.0 −0.3 −3.2 −1.51 
Stanford Three-Community Study (46
    Cohort Both −10.1  −0.2  −1.45 
Stanford Five-City Project (47, 48      
    Cross-sectional Men −3.5 −2.5 −0.1 4.1 −0.50 
    Cross-sectional Women −4.8 −4.1 −0.1 6.3 −0.29 
    Cohort Both −5.2 −3.7 0.0 −3.9 −0.49 
Zurich Cardiovascular Health Programme (50      
    Cross-sectional Men 6.4 6.9 0.6 1.5 3.10 
    Cross-sectional Women 3.2 0.3 −0.1 0.91 

Abbreviations: CVD, cardiovascular disease; DBP, diastolic blood pressure; Di.S.Co., Sezze District Community Control; SBP, systolic blood pressure; TC, total cholesterol.

Calculation of overall net changes in 10-year CVD risk showed a more consistent trend (Table 3 and Figures 2 and 3), with an average net reduction in 10-year CVD risk of 0.32% for cross-sectional surveys (Figure 2) and 0.88% for cohort surveys (Figure 3) (an average 0.65% reduction over all programs). When expressed as a percentage of baseline risk, the average net reduction in 10-year CVD risk was 5.81% for cross-sectional surveys and 11.14% for cohort surveys (9.08% over all programs). For studies in which data were available for men and women separately (n = 15), the average reduction in 10-year CVD risk across all programs was 0.77% for men and 0.52% for women. When expressed as a percentage of baseline risk, this reduction was 6.35% for men and 9.71% for women.

Figure 2.

Mean net changes in 10-year cardiovascular disease risk in heart health programs evaluated by means of cross-sectional surveys for men (•), women (○), and all study participants (▪), 1970–2008. (For complete names of programs, see Table 1.)

Figure 2.

Mean net changes in 10-year cardiovascular disease risk in heart health programs evaluated by means of cross-sectional surveys for men (•), women (○), and all study participants (▪), 1970–2008. (For complete names of programs, see Table 1.)

Figure 3.

Mean net changes in 10-year cardiovascular disease risk in heart health programs evaluated by means of cohort follow-up for men (•), women (○), and all study participants (▪), 1970–2008. B, Båtsfjord; NC, North Cape; NRP, National Research Program. (For complete names of programs, see Table 1.)

Figure 3.

Mean net changes in 10-year cardiovascular disease risk in heart health programs evaluated by means of cohort follow-up for men (•), women (○), and all study participants (▪), 1970–2008. B, Båtsfjord; NC, North Cape; NRP, National Research Program. (For complete names of programs, see Table 1.)

DISCUSSION

There are a number of well-known community CVD prevention programs, but this systematic review identified an unexpectedly large number of additional programs, and many of these have not been featured in previous systematic reviews. Therefore, this systematic review brought together a large body of previously unreviewed evidence.

The favorable trend toward reduced rates of CVD mortality suggests that these programs may be beneficial for the prevention of CVD, but the limited number of studies presenting these data and the lack of statistically significant findings limit the interpretation of mortality results.

A larger amount of data was available on changes in physiologic CVD risk factors. Consistent trends in favor of CVD programs for individual risk factors were evident, but in isolation the sizes of these effects appeared clinically insignificant. On the basis of a potential cumulative effect of individual risk factor changes, it appeared likely that the overall effect might be more clinically significant, and this was the purpose of considering the impact on an overall risk score.

The calculated risk score findings suggest that there is a net favorable effect associated with community CVD prevention programs. Although it was not obtained via meta-analysis, the average net reduction of 0.65% in 10-year CVD risk (a 9.08% reduction relative to baseline) provides support for the view that changes in individual risk factors, when considered together, produce a clinically important improvement. This risk reduction equates to a number needed to treat of 154 in order to avoid 1 case of CVD over 10 years. In population terms, this may have important benefits, where large-scale, whole-population approaches are being implemented.

Most interventions appear to have achieved net reductions in 10-year CVD risk of approximately 1%, but some showed smaller or greater risk reductions. An important question in public health practice is, What type/mode of intervention is most effective for achieving improvements in population heart health? Investigators have sought to untangle possible determinants of program success (51–53), and in 1 review, Sellers et al. (11) used regression analysis to attempt to determine correlates of program effectiveness. They showed that the biggest predictor of risk factor changes was the method of program evaluation (length of follow-up, matching of intervention and control communities, and number of intervention/reference communities), with little apparent influence of intervention or population type (11).

Although variation in both the nature of programs and the apparent effectiveness of programs was observed, the limited information on the nature of each program, the large number of other factors which also varied between programs, and the lack of statistical information made further investigation of the effect of variation in intervention futile. This key question thus remains unanswered by this systematic review. However, it is interesting to note that the North Karelia program, highlighted for its particular effectiveness, showed a reduction in risk (average of −1.17% for men and women) similar to that of some other programs. This intervention appears not to have been unique, and there may be more uniformity in the ability of programs to bring about reduction in risk than previously thought.

Net increases in CVD risk were observed in some cases, and 4 programs showed net negative effects (Coalfields (19), Minnesota (34), National Research Program (39), and Zurich (50)). On the whole, the size of these increases was relatively small, but for the Zurich study, there appears to have been a large increase in predicted CVD risk associated with the intervention. The study investigators discussed this finding and noted that there may have been some systematic error in the measurement of cholesterol levels (50). They also commented on the limited intensity of the intervention program and suggested that this may have been responsible for the unfavorable results (50). These factors may have contributed to the apparent ineffectiveness of this particular program, but the potential for harm cannot be ruled out.

It is possible that a program of this kind, if received badly or implemented in a way that increases social inequalities, could result in adverse effects. After 5 years of intervention in Norsjö, participants of a lower social class were less likely to recall having seen newspaper or television news on the CVD project or to have heard radio coverage of the project (54). Of those that did recall media messages, the less educated and manual workers, particularly men, were less likely to report having been influenced by those messages (54). However, when health outcomes were examined according to social stratum, net changes in overall estimated risk were found to be similar, if not more beneficial, in persons with a low educational level compared with persons with a high educational level, and investigators concluded that this project had, if anything, reduced health inequalities (55). Despite these findings, it remains uncertain whether programs of this type have the potential to increase health inequalities, and the potential for harm may be an important consideration.

The observed results are affected by uncertainty arising from limitations in the primary research and review method. All of the included studies used a controlled before-after or controlled interrupted time series design, and a potentially important methodological issue is the mode of outcome data collection. The scale of the intervention populations dictated that only samples of the whole population could be evaluated, and differences in approach are potentially important. In some studies, researchers employed a cohort approach, using the same subpopulation to measure outcomes throughout the study, while others used a cross-sectional approach, where different randomly selected samples of the whole population were chosen at different time points. These study designs are open to different sources of bias (contamination in cross-sectional surveys and differential attrition in cohort studies). Cohort studies may provide a better measure of an intervention's effect on a stable population, whereas cross-sectional studies may provide a better measure of the total population effect (including the effect of contamination).

Where treatment and control areas are similar with respect to social and geographic factors that govern rates of dropout, differential attrition may be unlikely. In the majority of studies, the investigators stated that communities with similar age ranges and socioeconomic profiles were selected as controls. Although differential attrition cannot be discounted, it appears unlikely that this factor had a major influence on results for cohort evaluations. Results for cross-sectional surveys appear to have been less favorable than those for cohort surveys, and a potential cause may be cross-community contamination. In order to select controls with similar socioeconomic profiles and secular trends in CVD, investigators often chose control communities that were located reasonably close to intervention communities. The spread of media coverage and individual migration between the intervention and control communities may, to some extent, have reduced the apparent intervention effect in programs evaluated by means of cross-sectional surveys.

Another potential source of bias associated with these types of studies is the choice of control group. Differences in intervention and control groups at baseline may not in themselves be a source of bias, since baseline measurements are taken into account in the calculation of net change. However, differences present at baseline may reflect differences in the types of people living in each community. Where control and intervention groups are well-matched at baseline, secular trends are more likely to be equivalent, and there may be less risk of bias. However, where there are substantial differences at baseline, these may reflect differences in population groups with different underlying secular trends. It is unclear in which direction this source of bias may act, but the quality of control group selection is an important consideration in assessing the validity of these types of studies.

The applicability of programs included in this review to the present is also uncertain. The majority of programs were conducted in the 1970s (n = 9), 1980s (n = 17), and 1990s (n = 9), with only 1 being implemented since 2000. Changes in the prevalence of risk factors and changes in attitudes, lifestyles, and community settings may have an impact on the effectiveness of these types of programs and may limit the generalizability of findings.

Areas for intervention were often selected on the basis of elevated CVD risk, and this may have implications for their applicability to future program implementation. It is unclear from the current review whether these types of programs have a lesser or greater effect in high-risk communities, but the possibility of differential program effectiveness should be considered.

Another issue relating to the applicability of programs is the nature of subjects taking part in evaluation surveys. Persons who lead less healthy lifestyles, younger persons, ethnic minorities, and persons of lower socioeconomic status may be less likely to participate in surveys, and this can lead to response bias. Weinehall et al. (56) compared survey participants with nonparticipants and found that younger persons, the unemployed, and those with the lowest incomes were less likely to participate. However, only minor differences regarding socioeconomic status (assessed by employment type) and educational level were found. Survey respondents had a more favorable total cholesterol level but higher average blood pressure, and there were no differences in body mass index or rates of smoking (56). However, these findings may not be true in all cases, and the extent to which response bias affects the applicability of findings to whole populations is unclear.

In addition to limitations arising from the nature of the primary research, there are some limitations arising from the review method. In this review, because of a lack of statistical information in the majority of included studies, it was not possible to calculate confidence intervals or to carry out meta-analysis. For approximately half of the included studies, no information relating to within-study error for risk factor changes was given. This hinders the interpretation of the overall findings, since overall statistical significance could not be assessed.

Additionally, risk factor data that could be used to generate a CVD risk score were not reported for all included programs. Results for the 14 programs that were not included in the calculation of CVD risk were mixed (Appendix Table 2). Although some showed favorable results, overall these programs appeared less promising than those for which CVD risk was calculated, and their exclusion may have produced an overly optimistic result. However, for programs that were included in the calculation, missing values for certain risk factors (diabetes status and high density lipoprotein cholesterol level were missing for most programs, smoking status was missing for 3 programs, diastolic blood pressure and total cholesterol level were missing for 2 programs, and systolic blood pressure was missing for 1 program) would tend to counteract this. Missing values are likely to reduce apparent effect size, since, where values are missing, risk factors are presumed to remain constant, resulting in smaller apparent treatment-versus-control differences. Therefore, the calculated average net CVD risk reduction may be a realistic estimate of the effectiveness of community CVD programs.

Finally, an ever-present threat in systematic reviews is the possibility of publication bias. We were able to guard against this to a considerable degree because of the comprehensive search used, which included possible sources of gray literature. It is therefore improbable that there were large numbers of negative studies missed from the review, and publication bias appears unlikely to overturn the current findings.

Community interventions for the prevention of CVD appear to have generally achieved favorable changes in overall CVD risk. Considerable uncertainty remains, but this review provided no evidence that community prevention can be rejected as a useful approach to preventing CVD. However, programs implemented in the past need to be adapted to current circumstances, and revised approaches should be reevaluated before widespread implementation.

Abbreviations

    Abbreviations
  • CVD

    cardiovascular disease

  • NICE

    National Institute for Health and Clinical Excellence

Author affiliations: Unit of Public Health, Epidemiology and Biostatistics, College of Medical and Dental Sciences, University of Birmingham, Birmingham, United Kingdom (Mary Pennant, Clare Davenport, Sue Bayliss, Wendy Greenheld, Tom Marshall); and Peninsula Technology Assessment Group, Peninsula College of Medicine and Dentistry, University of Exeter, Exeter, United Kingdom (Chris Hyde).

This work was supported by the Centre for Public Health Excellence of the United Kingdom National Institute for Health and Clinical Excellence as part of its public health guidance development program.

Conflict of interest: none declared.

Appendix Table 1.

MEDLINE Search Strategy Used in a Systematic Review of the Effectiveness of Community Programs for Prevention of Cardiovascular Disease, 1970–2008

cardiovascular disease$.mp. or exp cardiovascular diseases/ 
CVD.mp. 
coronary disease$.mp. 
heart disease$.mp. 
atherosclerosis.mp. 
arteriosclerosis.mp. 
hypertension.mp. 
blood pressure.mp. 
exp hyperlipidemias/or hyperlipidaemia$.mp. 
10 hyperlipidemia$.mp. 
11 exp cholesterol/or cholesterol.mp. 
12 exp stroke/or stroke$.mp. 
13 peripheral vascular disease$.mp. 
14 peripheral arterial disease$.mp. 
15 hypercholesterol$.mp. 
16 hyperlipid$.mp. 
17 or/1–16 
18 health education.mp. or exp health education/ 
19 health promotion.mp. or exp health promotion/ 
20 primary prevention.mp. or exp primary prevention/ 
21 campaign$.mp. 
22 media.mp. or exp mass media/ 
23 exp counseling/or advice$.mp. 
24 counsel$.mp. 
25 program$.mp. 
26 (policy or policies).mp. 
27 or/18–26 
28 exp smoking/or smoking.mp. 
29 exp tobacco/or tobacco.mp. 
30 exp diet/or diet.mp. 
31 exercise.mp. or exp exercise/ 
32 obesity.mp. or exp obesity/ 
33 diabetes.mp. or exp diabetes mellitus/ 
34 stress.mp. or exp stress/ 
35 exp cholesterol/or cholesterol.mp. 
36 exp hypertension/or hypertension.mp. 
37 blood pressure.mp. or exp blood pressure/ 
38 alcohol$.mp. 
39 drinking.mp. or exp alcohol drinking/ 
40 (cardiovascular adj3 risk$).mp. 
41 multiple risk$.mp. 
42 or/28–41 
43 17 and 27 and 42 
44 limit 43 to (humans and year = “1970–2008”) 
45 limit 44 to “therapy (sensitivity)” 
46 epidemiologic studies/ 
47 longitudinal studies/ 
48 (control$ before and after).mp. 
49 cohort.mp. 
50 case control.mp. 
51 interrupted time series.mp. 
52 or/46–51 
53 44 and 52 
54 45 or 53 
cardiovascular disease$.mp. or exp cardiovascular diseases/ 
CVD.mp. 
coronary disease$.mp. 
heart disease$.mp. 
atherosclerosis.mp. 
arteriosclerosis.mp. 
hypertension.mp. 
blood pressure.mp. 
exp hyperlipidemias/or hyperlipidaemia$.mp. 
10 hyperlipidemia$.mp. 
11 exp cholesterol/or cholesterol.mp. 
12 exp stroke/or stroke$.mp. 
13 peripheral vascular disease$.mp. 
14 peripheral arterial disease$.mp. 
15 hypercholesterol$.mp. 
16 hyperlipid$.mp. 
17 or/1–16 
18 health education.mp. or exp health education/ 
19 health promotion.mp. or exp health promotion/ 
20 primary prevention.mp. or exp primary prevention/ 
21 campaign$.mp. 
22 media.mp. or exp mass media/ 
23 exp counseling/or advice$.mp. 
24 counsel$.mp. 
25 program$.mp. 
26 (policy or policies).mp. 
27 or/18–26 
28 exp smoking/or smoking.mp. 
29 exp tobacco/or tobacco.mp. 
30 exp diet/or diet.mp. 
31 exercise.mp. or exp exercise/ 
32 obesity.mp. or exp obesity/ 
33 diabetes.mp. or exp diabetes mellitus/ 
34 stress.mp. or exp stress/ 
35 exp cholesterol/or cholesterol.mp. 
36 exp hypertension/or hypertension.mp. 
37 blood pressure.mp. or exp blood pressure/ 
38 alcohol$.mp. 
39 drinking.mp. or exp alcohol drinking/ 
40 (cardiovascular adj3 risk$).mp. 
41 multiple risk$.mp. 
42 or/28–41 
43 17 and 27 and 42 
44 limit 43 to (humans and year = “1970–2008”) 
45 limit 44 to “therapy (sensitivity)” 
46 epidemiologic studies/ 
47 longitudinal studies/ 
48 (control$ before and after).mp. 
49 cohort.mp. 
50 case control.mp. 
51 interrupted time series.mp. 
52 or/46–51 
53 44 and 52 
54 45 or 53 
Appendix Table 2.

Programs Not Included in the Calculation of 10-Year Risk of Cardiovascular Disease in a Systematic Review of Heart Health Programs, 1970–2008

Program (Reference No.) Net Change(s) Relative to Control Group 
Action Heart (16Significant reduction in smoking rates and an increase in low-fat milk consumption. No other significant changes. 
Bootheel Heart Health Project (17Significant increase in leisure-time physical activity. No significant effects on smoking, fruit and vegetable consumption, or prevalence of overweight. Significantly more participants with cholesterol checked in the past 2 years. Black population showed improvements in physical activity, smoking, fruit and vegetable consumption, weight status, and cholesterol checks. 
CardioVision 2020 (18No significant difference in trends for smoking or body mass index. Significantly worse trends for physical activity. Significantly more people eating 5 fruits and vegetables per day. 
Coeur en Santé St.-Henri (20Trend toward net reduction in smoking prevalence and frequency (not significant). Trend toward increased leisure-time physical activity (not significant). No difference in body mass index or blood pressure. Significantly more participants with blood pressure and cholesterol checked in the past year. No difference in dietary outcomes. 
Danish Municipality Project (22No differences in smoking, diet, or physical activity. 
Franklin Cardiovascular Health Program (28Lower rates of death from stroke and significantly lower rates of death from coronary heart disease. 
Have a Heart Paisley (30No significant effects on any self-reported outcomes. 
Heartbeat Wales (31No significant effects on any self-reported outcomes. 
North Coast Healthy Lifestyle Programme (37Significant decrease in smoking prevalence. 
Quebec Heart Health Demonstration Project (43 
    Rural Significant increase in consumption of low-fat processed meats. No significant difference in dietary index measures. 
    Suburban Significant increase in consumption of low-fat milk in women and of low-fat meats in men and women. No significant difference in dietary index measures. 
    Urban No significant difference in dietary index measures. 
South Carolina Cardiovascular Disease Prevention Project (45Significantly smaller increase in the prevalence of high cholesterol and significantly higher rates of cholesterol screening. No significant difference in smoking or physical activity. Significant net negative effect on blood pressure. Significantly smaller increase in the prevalence of overweight. Significant net reduction in consumption of animal fats. 
Tessin Cardiovascular Health Programme (49Significant reductions from baseline in 12 out of 15 risk-factor measures (related to smoking, blood pressure, body mass index, physical activity, and diet) in intervention area as compared with 3 out of 15 for the control area. No direct comparison of individual risk-factor changes in intervention area versus control area. 
Program (Reference No.) Net Change(s) Relative to Control Group 
Action Heart (16Significant reduction in smoking rates and an increase in low-fat milk consumption. No other significant changes. 
Bootheel Heart Health Project (17Significant increase in leisure-time physical activity. No significant effects on smoking, fruit and vegetable consumption, or prevalence of overweight. Significantly more participants with cholesterol checked in the past 2 years. Black population showed improvements in physical activity, smoking, fruit and vegetable consumption, weight status, and cholesterol checks. 
CardioVision 2020 (18No significant difference in trends for smoking or body mass index. Significantly worse trends for physical activity. Significantly more people eating 5 fruits and vegetables per day. 
Coeur en Santé St.-Henri (20Trend toward net reduction in smoking prevalence and frequency (not significant). Trend toward increased leisure-time physical activity (not significant). No difference in body mass index or blood pressure. Significantly more participants with blood pressure and cholesterol checked in the past year. No difference in dietary outcomes. 
Danish Municipality Project (22No differences in smoking, diet, or physical activity. 
Franklin Cardiovascular Health Program (28Lower rates of death from stroke and significantly lower rates of death from coronary heart disease. 
Have a Heart Paisley (30No significant effects on any self-reported outcomes. 
Heartbeat Wales (31No significant effects on any self-reported outcomes. 
North Coast Healthy Lifestyle Programme (37Significant decrease in smoking prevalence. 
Quebec Heart Health Demonstration Project (43 
    Rural Significant increase in consumption of low-fat processed meats. No significant difference in dietary index measures. 
    Suburban Significant increase in consumption of low-fat milk in women and of low-fat meats in men and women. No significant difference in dietary index measures. 
    Urban No significant difference in dietary index measures. 
South Carolina Cardiovascular Disease Prevention Project (45Significantly smaller increase in the prevalence of high cholesterol and significantly higher rates of cholesterol screening. No significant difference in smoking or physical activity. Significant net negative effect on blood pressure. Significantly smaller increase in the prevalence of overweight. Significant net reduction in consumption of animal fats. 
Tessin Cardiovascular Health Programme (49Significant reductions from baseline in 12 out of 15 risk-factor measures (related to smoking, blood pressure, body mass index, physical activity, and diet) in intervention area as compared with 3 out of 15 for the control area. No direct comparison of individual risk-factor changes in intervention area versus control area. 

References

1.
Leal
J
Luengo-Fernández
R
Gray
A
, et al.  . 
Economic burden of cardiovascular diseases in the enlarged European Union
Eur Heart J
 , 
2006
, vol. 
27
 
13
(pg. 
1610
-
1619
)
2.
Buechley
RW
Drake
RM
Breslow
L
Relationship of amount of cigarette smoking to coronary heart disease mortality rates in men
Circulation
 , 
1958
, vol. 
18
 
6
(pg. 
1085
-
1090
)
3.
Mayor
S
Obesity in middle age is an independent risk factor for cardiovascular disease
BMJ
 , 
2006
, vol. 
332
 
7533
pg. 
71
 
4.
Manninen
V
Tenkanen
L
Koskinen
P
, et al.  . 
Joint effects of serum triglyceride and LDL cholesterol and HDL cholesterol concentrations on coronary heart disease risk in the Helsinki Heart Study. Implications for treatment
Circulation
 , 
1992
, vol. 
85
 
1
(pg. 
37
-
45
)
5.
Law
MR
Morris
JK
Wald
NJ
Use of blood pressure lowering drugs in the prevention of cardiovascular disease: meta-analysis of 147 randomised trials in the context of expectations from prospective epidemiological studies
BMJ
 , 
2009
, vol. 
338
  
b1665
6.
Blackburn
H
Population strategies of cardiovascular disease prevention: scientific base, rationale and public health implications
Ann Med
 , 
1989
, vol. 
21
 
3
(pg. 
157
-
162
)
7.
Brunner
E
White
I
Thorogood
M
, et al.  . 
Can dietary interventions change diet and cardiovascular risk factors? A meta-analysis of randomized controlled trials
Am J Public Health
 , 
1997
, vol. 
87
 
9
(pg. 
1415
-
1422
)
8.
Nicholson
SO
The effect of cardiovascular health promotion on health behaviors in elementary school children: an integrative review
J Pediatr Nurs
 , 
2000
, vol. 
15
 
6
(pg. 
343
-
355
)
9.
Ebrahim
S
Smith
GD
Systematic review of randomised controlled trials of multiple risk factor interventions for preventing coronary heart disease
BMJ
 , 
1997
, vol. 
314
 
7095
(pg. 
1666
-
1674
)
10.
Ketola
E
Sipilä
R
Mäkelä
M
Effectiveness of individual lifestyle interventions in reducing cardiovascular disease and risk factors
Ann Med
 , 
2000
, vol. 
32
 
4
(pg. 
239
-
251
)
11.
Sellers
DE
Crawford
SL
Bullock
K
, et al.  . 
Understanding the variability in the effectiveness of community heart health programs: a meta-analysis
Soc Sci Med
 , 
1997
, vol. 
44
 
9
(pg. 
1325
-
1339
)
12.
National Institute for Health and Clinical Excellence
Methods for Development of NICE Public Health Guidance
 , 
2006
London, United Kingdom
National Institute for Health and Clinical Excellence
13.
Anderson
KM
Odell
PM
Wilson
PW
, et al.  . 
Cardiovascular disease risk profiles
Am Heart J
 , 
1991
, vol. 
121
 
1
(pg. 
293
-
298
)
14.
United Kingdom Department of Health
Health Survey for England 1998
 , 
1999
London, United Kingdom
Department of Health
 
15.
Marshall
T
Identification of patients for clinical risk assessment by prediction of cardiovascular risk using default risk factor values
BMC Public Health
 , 
2008
, vol. 
8
 pg. 
25
  
(doi: 10.1186/1471-2458-8-25)
16.
Baxter
T
Milner
P
Wilson
K
, et al.  . 
A cost effective, community based heart health promotion project in England: prospective comparative study
BMJ
 , 
1997
, vol. 
315
 
7108
(pg. 
582
-
585
)
17.
Brownson
RC
Smith
CA
Pratt
M
, et al.  . 
Preventing cardiovascular disease through community-based risk reduction: the Bootheel Heart Health Project
Am J Public Health
 , 
1996
, vol. 
86
 
2
(pg. 
206
-
213
)
18.
Kottke
TE
Thomas
RJ
Lopez-Jimenez
F
, et al.  . 
CardioVision 2020: program acceptance and progress after 4 years
Am J Prev Med
 , 
2006
, vol. 
30
 
2
(pg. 
137
-
143
)
19.
Higginbotham
N
Heading
G
McElduff
P
, et al.  . 
Reducing coronary heart disease in the Australian Coalfields: evaluation of a 10-year community intervention
Soc Sci Med
 , 
1999
, vol. 
48
 
5
(pg. 
683
-
692
)
20.
O'Loughlin
JL
Paradis
G
Gray-Donald
K
, et al.  . 
The impact of a community-based heart disease prevention program in a low-income, inner-city neighborhood
Am J Public Health
 , 
1999
, vol. 
89
 
12
(pg. 
1819
-
1826
)
21.
Giampaoli
S
Poce
A
Sciarra
F
, et al.  . 
Change in cardiovascular risk factors during a 10-year community intervention program
Acta Cardiol
 , 
1997
, vol. 
52
 
5
(pg. 
411
-
422
)
22.
Osler
M
Jespersen
NB
The effect of a community-based cardiovascular disease prevention project in a Danish municipality
Dan Med Bull.
 , 
1993
, vol. 
40
 
4
(pg. 
485
-
489
)
23.
Schuit
AJ
Wendel-Vos
GC
Verschuren
WM
, et al.  . 
Effect of 5-year community intervention Hartslag Limburg on cardiovascular risk factors
Am J Prev Med
 , 
2006
, vol. 
30
 
3
(pg. 
237
-
242
)
24.
Regional Office for Europe, World Health Organization
Comprehensive Cardiovascular Control Programmes. (EURO Reports and Studies 106)
 , 
1988
Copenhagen, Denmark
WHO Regional Office for Europe
25.
Bruckert
E
Thomas
D
Emmerich
J
, et al.  . 
Impact of an information campaign on cardiovascular risk factors. 5-year results at the study town Epernon [in French]
Presse Med
 , 
1999
, vol. 
28
 
10
(pg. 
517
-
522
)
26.
Lupton
BS
Fønnebø
V
Søgaard
AJ
The Finnmark Intervention Study: is it possible to change CVD risk factors by community-based intervention in an Arctic village in crisis?
Scand J Public Health
 , 
2003
, vol. 
31
 
3
(pg. 
178
-
186
)
27.
Lupton
BS
Fønnebø
V
Søgaard
AJ
, et al.  . 
The Finnmark Intervention Study. Better health for the fishery population in an Arctic village in North Norway
Scand J Prim Health Care
 , 
2002
, vol. 
20
 
4
(pg. 
213
-
218
)
28.
Record
NB
Harris
DE
Record
SS
, et al.  . 
Mortality impact of an integrated community cardiovascular health program
Am J Prev Med
 , 
2000
, vol. 
19
 
1
(pg. 
30
-
38
)
29.
Hoffmeister
H
Mensink
GB
Stolzenberg
H
, et al.  . 
Reduction of coronary heart disease risk factors in the German Cardiovascular Prevention Study
Prev Med
 , 
1996
, vol. 
25
 
2
(pg. 
135
-
145
)
30.
Blamey
A
Ayana
M
Lawson
L
, et al.  . 
Final Report of the Independent Evaluation of Have a Heart Paisley
 , 
2004
Glasgow, United Kingdom
University of Glasgow
31.
Tudor-Smith
C
Nutbeam
D
Moore
L
, et al.  . 
Effects of the Heartbeat Wales Programme over five years on behavioural risks for cardiovascular disease: quasi-experimental comparison of results from Wales and a matched reference area
BMJ
 , 
1998
, vol. 
316
 
7134
(pg. 
818
-
822
)
32.
Kotchen
JM
McKean
HE
Jackson-Thayer
S
, et al.  . 
Impact of a rural high blood pressure control program on hypertension control and cardiovascular disease mortality
JAMA
 , 
1986
, vol. 
255
 
16
(pg. 
2177
-
2182
)
33.
Shelley
E
Daly
L
Collins
C
, et al.  . 
Cardiovascular risk factor changes in the Kilkenny Health Project. A community health promotion programme
Eur Heart J
 , 
1995
, vol. 
16
 
6
(pg. 
752
-
760
)
34.
Luepker
RV
Murray
DM
Jacobs
DR
Jr
, et al.  . 
Community education for cardiovascular disease prevention: risk factor changes in the Minnesota Heart Health Program
Am J Public Health
 , 
1994
, vol. 
84
 
9
(pg. 
1383
-
1393
)
35.
Weinehall
L
Westman
G
Hellsten
G
, et al.  . 
Shifting the distribution of risk: results of a community intervention in a Swedish programme for the prevention of cardiovascular disease
J Epidemiol Community Health
 , 
1999
, vol. 
53
 
4
(pg. 
243
-
250
)
36.
Weinehall
L
Hellsten
G
Boman
K
, et al.  . 
Can a sustainable community intervention reduce the health gap?—10-year evaluation of a Swedish community intervention program for the prevention of cardiovascular disease
Scand J Public Health
 , 
2001
, vol. 
29
 
suppl 56
(pg. 
59
-
68
)
37.
Egger
G
Fitzgerald
W
Frape
G
, et al.  . 
Results of large scale media antismoking campaign in Australia: north coast “Quit for Life” programme
Br Med J (Clin Res Ed)
 , 
1983
, vol. 
287
 
6399
(pg. 
1125
-
1128
)
38.
Puska
P
Salonen
JT
Nissinen
A
Change in risk factors for coronary heart disease during 10 years of a community intervention programme (North Karelia Project)
Br Med J (Clin Res Ed)
 , 
1983
, vol. 
287
 
6804
(pg. 
1840
-
1844
)
39.
Gutzwiller
F
Nater
B
Martin
J
Community-based primary prevention of cardiovascular disease in Switzerland: methods and results of the National Research Program (NRP 1A)
Prev Med
 , 
1985
, vol. 
14
 
4
(pg. 
482
-
491
)
40.
Isacsson
A
Lindholm
LH
Schersten
B
, et al.  . 
Community intervention against non-insulin dependent diabetes mellitus (NIDDM) and cardiovascular disease: a study based on Swedish health care
Cardiovasc Risk Fact
 , 
1996
, vol. 
6
 
3
(pg. 
164
-
171
)
41.
Nafziger
AN
Erb
TA
Jenkins
PL
, et al.  . 
The Otsego-Schoharie Healthy Heart Program: prevention of cardiovascular disease in the rural US
Scand J Public Health Suppl
 , 
2001
, vol. 
29
 
suppl 56
(pg. 
21
-
32
)
42.
Carleton
RA
Lasater
TM
Assaf
AR
, et al.  . 
The Pawtucket Heart Health Program: community changes in cardiovascular risk factors and projected disease risk
Am J Public Health
 , 
1995
, vol. 
85
 
6
(pg. 
777
-
785
)
43.
Huot
I
Paradis
G
Ledoux
M
Effects of the Quebec Heart Health Demonstration Project on adult dietary behaviours. Quebec Heart Health Demonstration Project Research Group
Prev Med
 , 
2004
, vol. 
38
 
2
(pg. 
137
-
148
)
44.
Heinemann
L
Heine
H
Eckstein
M
, et al.  . 
Project Schleiz—national demonstration project for the population-wide prevention of heart, circulatory and other non-communicable chronic diseases
Z Klin Med
 , 
1986
, vol. 
41
 
7
(pg. 
536
-
539
)
45.
Goodman
RM
Wheeler
FC
Lee
PR
Evaluation of the Heart To Heart Project: lessons from a community-based chronic disease prevention project
Am J Health Promot
 , 
1995
, vol. 
9
 
6
(pg. 
443
-
455
)
46.
Farquhar
JW
Maccoby
N
Wood
PD
, et al.  . 
Community education for cardiovascular health
Lancet
 , 
1977
, vol. 
1
 
8023
(pg. 
1192
-
1195
)
47.
Winkleby
MA
Taylor
CB
Jatulis
D
, et al.  . 
The long-term effects of a cardiovascular disease prevention trial: the Stanford Five-City Project
Am J Public Health
 , 
1996
, vol. 
86
 
12
(pg. 
1773
-
1779
)
48.
Farquhar
JW
Fortmann
SP
Flora
JA
, et al.  . 
Effects of communitywide education on cardiovascular disease risk factors. The Stanford Five-City Project
JAMA
 , 
1990
, vol. 
264
 
3
(pg. 
359
-
365
)
49.
Domenighetti
G
Casabianca
A
Wietlisbach
V
, et al.  . 
Prevention of cardiovascular diseases is effective. Initial program evaluation of the canton Tessin (1984–1989) [in French]
Hygie
 , 
1992
, vol. 
11
 
1
(pg. 
24
-
31
)
50.
Stransky
M
Schwarzenbach
FH
Epstein
FH
, et al.  . 
Prevention of coronary heart disease. A study of 2 rural communities of Zurich [in German]
Soz Praventivmed
 , 
1982
, vol. 
27
 
4
(pg. 
178
-
186
)
51.
Rimer
BK
Glanz
K
Lerman
C
Contributions of public health to patient compliance
J Community Health
 , 
1991
, vol. 
16
 
4
(pg. 
225
-
240
)
52.
Pearson
TA
Wall
S
Lewis
C
, et al.  . 
Dissecting the “black box” of community intervention: lessons from community-wide cardiovascular disease prevention programs in the US and Sweden
Scand J Public Health
 , 
2001
, vol. 
29
 
suppl 56
(pg. 
69
-
78
)
53.
Rudholm
N
A comparison of population versus individual based cardiovascular disease prevention programs in Västerbotten, Sweden
Health Policy
 , 
2006
, vol. 
78
 
1
(pg. 
70
-
76
)
54.
Brännström
I
Persson
LA
Wall
S
Gender and social patterning of health: the Norsjö cardiovascular preventive programme in northern Sweden 1985–1990
Scand J Prim Health Care
 , 
1994
, vol. 
12
 
3
(pg. 
155
-
161
)
55.
Weinehall
L
Lewis
C
Nafziger
AN
, et al.  . 
Different outcomes for different interventions with different focus!—a cross-country comparison of community interventions in rural Swedish and US populations
Scand J Public Health
 , 
2001
, vol. 
29
 
suppl 56
(pg. 
46
-
58
)
56.
Weinehall
L
Hallgren
CG
Westman
G
, et al.  . 
Reduction of selection bias in primary prevention of cardiovascular disease through involvement of primary health care
Scand J Prim Health Care
 , 
1998
, vol. 
16
 
3
(pg. 
171
-
176
)