Introduction

Over the past few decades, considerable research effort and funding were invested to implement community-based policies or programs designed to reduce the risk of major chronic diseases in the population [1]. These programs often take place in an urban context due to the use of multi-component interventions, such as training of health professionals, media campaigns or policy action, and mobilize the community around the issue of public health importance [2]. Specifically, the programs offer three potential advantages: First, community-based interventions do not primarily depend on the professional health care system. Second, the interventions can reach persons of all risk groups. Third, these interventions can also influence contextual factors that shape lifestyle, but are not under control of the individual [3].

Results from the evaluation of the impact of these programs are mixed, raising some concerns about their effectiveness and about the ability to evaluate their impact against marked contextual influences of decreasing risk factor prevalence, changes in health-related behavior, and shifts in social and cultural tastes [4]. The situation is particularly complex in a time when numerous programs are being set up to tackle inequalities, in particular health inequalities. In many disadvantaged areas, more than one program operates side by side, often with similar and overlapping objectives [5]. Nevertheless, the public health impact makes community-based intervention compelling, especially in subgroups that continue to suffer a disproportionate burden, such as low income and migrant groups [6].

In addition to a well-tailored and effective concept and implementation, community-based health promotion also requires participation in the sense of active commitment of residents and professionals. This involvement in the planning, implementation and evaluation of health promotion programs is considered a key note of health promotion and community capacity [7]. Health promotion policies cannot be imposed ‘top down’, but are necessarily (co-) determined by the interests of local stakeholders, and should aim at maximizing the participation in the design and direction of actions. A high degree of participation and collaborative partnership is expected to contribute to the effectiveness and sustainability of community-based measures [8, 9].

Socio-ecologic Framework

Recent approaches of community-based health promotion suggest programs that apply multi-component or complex interventions addressing different aspects of health. This seems to be the most promising strategy for changing risk behaviors as well as the physical and social environment in which behaviors are shaped [10, 11]. In this socio-ecologic paradigm of health, we can differentiate between ‘user-addressed’ activities which are provided in the community, and ‘structural’ interventions in which people from the community are actively involved, either in planning or in implementing the activities [12].

User-addressed interventions aim at direct changes on at least three different levels: the individual level (e.g. health education, training), interpersonal level (e.g. develop new social ties/support, lay health advisors), and community level (e.g. media advocacy, environmental change) [13]. Therefore, not only unhealthy behavior, but also unhealthy environments need to be addressed to reduce general health problems and socioeconomic health inequalities. The impact of structural interventions is to build up community capacities (e.g. participation, local leadership, partnerships/networking, and resource mobilization), which indirectly promote the health outcome of the population [14, 15]. In this sense, structural interventions may enable user-addressed interventions, increase their effectiveness and ensure sustainability.

Socio-ecologic approaches assume that there are multiple determinants of health on different levels and that these have interactive and cumulative effects. Thus, interventions can be expected to be most effective and sustainable when all these factors/levels are targeted simultaneously. However, this may be impractical, and Stokols [16] recommends that interventions should focus at least on two levels. Furthermore, interventions addressing certain factors (e.g. healthy nutrition and physical activity) or singular settings (e.g. schools) have more successfully adopted a social ecological approach [11].

Current Evidence Gaps

The comprehensive review by Merzel and D’Afflitti [4] reflected the growing interest in addressing community-based health issues in the late 1980s and early 1990s. These included the Stamford Five-City Project [17], the Minnesota Heart Health Program [18], the Pawtucket Heart Health Program [19], and the North Karelia Project [20]. Apart from these large-scale programs in the USA and Finland, the review by Hills [5] largely focused on fairly small-scale UK interventions for health improvement. Both reviews found little evidence of many initiatives, projects and programs, and discussed crucial research challenges (e.g. need for intermediate outcomes, complexity and feasibility of interventions).

Overall, both reviews provide valuable information on activities in this field up to the early 2000s. However, they failed to take up a more international perspective, not being able to contribute to recent developments in the past 17 years. Community involvement has been boosted in many policy areas, supported by the current focus on health inequalities and the broader interest in community development [21].

Research Questions

The main goal of this paper is to assess what has been learned since the beginning of 2000s regarding the impact of community-based interventions on public health. The focus of our review is limited to urban areas due to complexity of the research field; it can also be assumed that the context for health promotion strategies in cities or urban neighborhoods is quite different than in rural areas [22]. Hence, we want to answer the central research question: What are the effects of multi-component health promotion activities on health behavior and health in urban areas? A sub-ordinated question concerns the issue to what extend the (in)effectiveness is associated with community participation as a key domain of capacity building [9, 14].

Methods

Protocol

We conducted a systematic review according to the preferred reporting items for systematic reviews and meta-analyses (PRISMA) guidelines [23]. An additional checklist shows this in more detail (see Additional File 1).

Search Strategy

The search was limited to articles published in English and German during the period January 1, 2002, to December 31, 2018. The reason for choosing the years since 2002 is the above-mentioned review by Merzel and D’Afflitti [4]. They had already screened studies on the same topic 20 years before. Our review used the string (effectiveness OR efficacy OR efficiency OR benefit* OR outcome*) AND (neighbo* OR district* OR communit* OR urban OR residential area*) AND (health promot* OR disease prevention) in Pubmed and PsycINFO. Pubmed research resulted in 3171 publications, while PsycINFO came to 1131 entries; after duplicates were removed 3646 records remained (advanced search: title/abstract) (see Fig. 1). However, the vast majority of publications did not fulfill our further criteria of inclusion: (i) urban areas, (ii) intervention studies, (iii) focus on impact or outcome evaluation, and (iv) multi-component interventions, settings and outcomes.

Fig. 1
figure 1

Flow diagram

Full size image

Data Extraction and Synthesis

Screening followed a two-step process with articles filtered by title/abstract, and full text. First author of this review (SN) initially screened the potentially relevant studies. The second author (OK) independently reviewed articles retained for inclusion in the preliminary phase. In the case of a persisting difference of opinion, a third colleague reviewed the paper in order to reach a consensus between the two review authors. Full texts of the remaining 101 articles were read and checked for eligibility, of which 72 were excluded. Main reasons for exclusion were (please also see Fig. 1): no interventional study on health outcomes, no urban area, and the focus on a single intervention or sub-setting (e.g. school). After checking the bibliographical references of the 29 eligible primary studies, we took three additional studies into the review. Finally, we extracted data on characteristics of 32 included studies, including author and publication date, study research design, country, baseline sample size, quality of study, type of intervention, impacts/outcomes, and key results.

We did not conduct a meta-analysis because the included studies were too heterogeneous regarding methodological aspects (i.e. target populations, outcomes, measurement instruments, and statistical methods).

Patient and Public Involvement

There was no direct patient or public involvement in this review.

Findings

Characteristics of the Studies

Of the 32 studies included in this review, 11 were conducted in Europe [24,25,26,27, 29,30,31,32,33,34,35], 9 in Asia [37,38,39,40,41,42,43,44,45], 7 in North America [46,47,48,49,50, 52, 53], 2 in Australia [54, 55], and there was one each in Central [56] and South America [57], and Africa [58] (see Table 1). The studies used various designs, including RCTs (n = 5) [30, 42, 45, 48, 49], cluster randomized trials (n = 7) [31, 34, 35, 38, 44, 52, 58], quasi-experimental designs (n = 12) [24, 26, 27, 29, 33, 35, 41, 43, 50, 54,55,56], pre-post-tests without control group (n = 7) [32, 39, 40, 46, 49, 50, 57], and secondary analysis (n = 1) [25]. The sample sizes at baseline varied from 69 to 16,228, with study populations having a variety of ethnic, cultural and socioeconomic backgrounds.

Table 1 Studies analyzing community-based health promotion interventions in urban areas
Full size table

Twenty-nine studies focused on the improvement of health behaviors (e.g. physical activity, nutrition, smoking), nine studies examined self-reported health outcomes, eight included health literacy, and 15 studies additionally examined clinical/anthropometric outcomes. Sixteen articles reported on interventions at three levels of community-based interventions (individual-level, group-level, community-level), 15 at two levels of interventions, and one at one level. Interventions for building community capacities were mentioned in 23 studies, including civic participation, leadership, partnership/networking, organizational development, resource mobilization and evaluation.

Included studies were assessed for risk of bias using the ‘quality assessment tool for quantitative studies’ developed by the effective public health practice project (EPHPP) [59]. Studies were scored against six criteria (selection bias, study design, confounders, blinding, data collection method, withdrawals and drop-outs), and the number of weak ratings was summed up to give a global quality score. Of the 32 studies reviewed, 14 (43.8%) were found to be of strong quality [26, 30, 34, 35, 37, 42,43,44, 48, 51, 54, 56,57,58], 13 (40.6%) of moderate quality [24, 27, 29, 32, 33, 38, 39, 41, 45, 47, 50, 53, 55], and five (15.6%) were weak in quality [25, 31, 40, 46, 49]. Amongst all of the studies, the most (37.5%) showed poor ratings in withdrawals and drop-outs. An additional file shows this in more detail (see Additional File 2).

Effectiveness of Interventions

Out of the 32 studies that met our inclusion criteria, 24 (75.0%) reported small to medium improvements in at least one outcome parameter, including all programs in low and middle income countries [24, 25, 30, 32, 35, 37,38,39,40,41,42,43,44,45,46,47, 49, 50, 53,54,55,56,57,58]. Among these relatively successful interventions were education and counselling services (n = 17), peer education projects (14), exercise provision (11), health promotion policies (10), media campaigns (6) and ‘walkability’ initiatives to enhance the physical environment (5). Regarding health outcomes that were most affected by community-based interventions, we found physical activity (11), clinical/anthropometric parameters (9), and nutritional behavior (6). Furthermore, all studies involving only one specific risk group within a community (e.g., the elderly [41, 48, 56], people with development disabilities [50] or metabolic syndrome [45, 46]) showed substantial improvements through selective or indicated interventions.

Large-scale community programs, however, often resulted in limited or no population-wide changes in primary health outcomes [27, 29, 33, 34, 50, 54]. Here, statistically significant effects particularly referred to social conditions as secondary outcomes (e.g. promote a ´sense of community´ or increased perception that people in the neighborhood keep together). In contrast to a low level of evidence at the area level, an additional process evaluation revealed that the sub-group of participants in any activities reported that the projects improved their perceived health and mental well-being [60]. Likewise, an intervention success is the finding that most community capacities (e.g. participation and community structures) were growing, but this process required a significant “up front” time investment before health outcomes data demonstrate change [51].

Discussion

This is the first review since 2002 that gives a systematic overview of the effectiveness of community-based interventions in urban areas worldwide. Despite the high number of studies that found positive effects, the impact of the interventions seemed to be rather limited. This holds especially true for large-scale community programs. In the first part, we will discuss the reasons for this in the light of our main findings. The second section will address participation as a key component of community capacities and health promotion. The third section points out the limitations of our systematic review.

Explanations for the Limited Evidence

One of the major explanations might be that the interventions were not sufficient enough to result in measurable effects. Most activities were provided on a modest scale (low intervention “dose”), not all interventions were feasible or acceptable, and some reached only small groups [24, 29, 33, 34, 38, 51]. Activities and actions also dealt with many diverse topics, which made a cumulative effect unlikely [24]. Furthermore, the implementation phase may have been too short as some of the planned activities had not been completed at the time of the follow-up study. It is known that such a process takes time and needs to be intense at different levels [24, 27, 29, 33, 34, 51].

The fact that small to moderate effects were found might also be due to limitations in the study designs and other methodological issues. The high non-response and dropout rates—caused by decreasing interest of target groups or population movement—may have threatened the population validity, and this could result in either an under- or overestimation of true effects [24, 27, 34]. Furthermore, a large number of program outcomes resulted in the use of self-reported, not always validated measures with possibly reduced sensitivity [24, 34, 51].

Thirdly, possible effects may have been masked by contextual or ‘spill-over’ effects, both in the intervention and the control areas. Examples are a city-wide renovation project in adjacent areas [24, 33] or other community-oriented health promotion activities launched by local, regional or national authorities or by other stakeholders that may have an impact on other communities, e.g. media campaigns and sickness funds’ initiatives [34]. More fundamentally, people in deprived areas, characterized by low socio-economic status, high unemployment and a large percentage of migrants, have restricted options and scope for action. This can overlay the effectiveness of community-based health promotion interventions [24].

It seems likely that the impact of interventions is to some extent proportional to the magnitude of the problem that is being addressed [27]. If the prevalence of health problems is below the population threshold, it may well be that the set of activities implemented in a program fails to have the intended effect. Thus, effects of community-wide interventions that comprise more ‘low risk’ than ‘high risk’ target-groups will be much smaller than typically expected from selective or targeted health interventions [33]. In other words, the more selective, not community-wide the interventions were, the larger the changes in the outcomes [40, 49].

Participation in Socially Deprived Areas

In the following, we will examine the components of community participation that have been studied and found to lead to effective intervention delivery. As mentioned above, 23 out of 32 projects used such capacity building strategies. Most of the reviewed articles did not include detailed information regarding the specific nature and outcomes of community participation. However, it is seen as a key note that interventions, especially in socially deprived areas, are the more successful the more they involve the environments of the target group, and the more the target group is involved in the planning, implementation and evaluation [15]. Community participation not only follows democratic ideals, but enhances the “utility” of health promotion by fostering commitment and a synergy of action and outcome [61].

In a systematic review by Milton [62], however, no evidence was found for impacts of ‘community engagement’ on health and service quality, but on non-health-related outcomes, such as housing, social capital/cohesion and empowerment. A qualitative study of the ‘Well London’ project participants (defined as residents who received activities) described a similarly positive impact of the project activities. In terms of the implementation process, this study clearly showed two key findings: First, changes at neighborhood-level did not lead to benefits among those who did not directly participate in project activities. Second, the social and physical environment of the neighborhoods was crucial for understanding people’s participation in the project activities and the extent of the intervention effects [60].

A recent review by Cyril et al. [63] examined community engagement levels according to a “ladder of participation” from informing residents to organizing themselves. They found a relation between low levels of community engagement (e.g. provide information to the public, consultation/hearings) and poor health outcomes in three studies. In contrast, studies ensuring high levels of community engagement (e.g. partnerships, codetermination, and decision-making power) resulted in positive health outcomes.

Similarly, for practice and participative research, it is true that while there is much known about motivations for citizen participation, this is also largely compatible with the goal of promoting good health [64]. Unfortunately, this does not guarantee that more democracy or less social inequality will be achieved [65]. Too little is generally known about the silent majority, their needs and interests. For example, a methodically demanding study of a random sample of 1160 Swedish citizens found that those who had previously used other forms of participation more frequently participated in neighborhood development processes, which may lead to an over-representation of certain particular interests [65, 66]. This may imply that the program objectives were not supported by all community members [24].

Limitations of the Review

Initial searches of databases identified several thousand references, but the small number of eligible studies suggests that few outcome evaluations have been published in peer-reviewed journals in the last two decades. Thus, the review described here is possibly not exhaustive in spite of searching the reference lists of the included studies. Our search strategy may not have revealed a complete list of all studies describing community-based interventions in urban areas in the relevant time period because of limitations of the Pubmed and PsycINFO search systems. It is likely that some health promotion projects are documented in ‘grey literature’ (e.g. unpublished reports and papers pending publication, conference abstracts). Thus, publication bias may have led to an overestimation rather than an underestimation of positive results.

Conclusion

Though the review is possibly not exhaustive, it captures major recent community-based health promotion interventions in urban areas, allowing for a systematic assessment of their impact on health outcomes. Our results confirm that community-based interventions are promising for health promotion and disease prevention but so far, their potential is not fully realized. For the future, it is recommended to plan, implement and evaluate interventions in the long term and in a participative manner with various sociodemographic groups and stakeholders of the neighborhood. These interventions should aim at proximal outcomes (e.g. risk behavior, sense of community) rather than distal outcomes (e.g. health status, life satisfaction) as well as the initial time investment in community capacity building.