The main finding of this study is that local health promotion has primarily been developed on the basis of municipalities' own needs and areas of interest. Important local events and champions or enthusiasts were perceived as being more powerful driving forces in this development than external factors, such as national or international public-health policies.
The Swedish municipalities administer local welfare activities with regard to infrastructure and social care. They also have a number of statutory obligations of relevance to people's health, not only in terms of social care but also in relation to risks referred to in national legislation, concerning tobacco, alcohol, and the living and working environment. But neither these obligations nor their possible meanings in municipal public health were referred to in the local interviews or documents. Reasons for this might lie in the risk-factor orientation of the national laws and the absence of connections with unified public health. Also, the municipalities lack any local statements or definitions of local health promotion and what it stands for, and may not have ascertained relationships between the public-health laws and their own local health promotion. However, municipalities are local producers and actors with responsibilities for several areas of importance with regard to health determinants, and therefore have the opportunity to exert a positive influence on their residents' health and living conditions [5, 6, 20]. Community-based interaction between local movements increases concern about the importance of addressing determinants of health in health-promotion programmes . When actions are related to health determinants, they have been shown to be more efficient than those that are individually directed . Several of the relevant factors are addressed in this study.
The four municipalities' health-promotion activities have changed from being individual and lifestyle based to being founded in a more holistic view on health. And the prioritized areas have changed from disease and risk prevention towards a structural perspective with a focus on the determinants of health. These trends reflect a parallel development in health promotion in Sweden since the 1980s, which mirrors the international trends described by Catford . Municipalities are influenced by and dependent on external factors , but in this study such influence and dependency, with regard to local health promotion content, organization and development, were not explicitly expressed.
To implement sustainable and efficient local health promotion, the participation and active engagement of community leaders and organizations are essential [2, 24]. In all the municipalities, there have been one or more committed and knowledgeable core community members, both officials and politicians, who - implicitly or explicitly - have functioned as advocates and driving forces in the initiation of local health promotion.
In particular, in municipalities A and D, politicians and officials have played a key role in the development process. And, the higher the position they have occupied, the greater the impact they have had. They have had the function of operative champions in the field or as integrators - persons in leading positions who work with integration, cooperation and establishment . Since the 1980s, these core members have initiated health-promotion activities and driven them forwards, at the outset principally in the form of ad-hoc measures at departmental level in a decentralized and sectorally divided administration. This kind of engagement, combined with a position of power based on professional knowledge, has been shown to be of importance in health-development processes .
By virtue of the Swedish municipalities' increased degree of self-governance during the 1990s, with augmented areas of responsibility, their opportunities themselves to choose objectives and directions for their operations also increased. The turn of the century, in conjunction with reorganizations in three of our municipalities (A, B, C), saw the start of a trend towards more centralized goal-oriented local direction and control within the municipalitis.
The perceived lack of clarity among the municipalities under study on the issue of responsibility for local health promotion and the implementation of external goals (such as those in the PHP, as stipulated by the Swedish parliament in 2003) has only affected local developments to a limited extent. The municipalities, by taking their point of departure in their own needs and resources in the first instance, have themselves chosen the objectives and directions of their own health-promotion activities. Some safety work, which would conventionally be regarded as health promotion was, in several cases, not defined as local health promotion, but rather as an expression of a desire to offer municipal residents a safe community to grow up and live in.
All the municipalities (A-D) have established their health-promotion organization and development by both conscious and instinctive means and/or actions. A supportive factor in such development has been the presence of a strong and stable political majority, with a generally clear leadership, which was the case in municipalities A and C. In these municipalities, officials at departmental level perceived greater clarity in municipality-wide goals, which in turn facilitated intersectoral collaboration.
The municipalities' strained financial situation in the early 1990s and the shift back to more centralized governance in the municipalities seemed to be factors promoting the increased development of intersectoral collaboration and alliances within the municipal organization, and also with private and civic actors in the local arena. Different kinds of collaboration for health promotion have been developed, including networking and alliance formation. The view was expressed that this not only provided for efficiency and coordination gains, but also promoted public participation and influence. These enhancing factors, which facilitated action and implementation, and also increased participation and a sense of shared responsibility and ownership, have also been reported by Dressendorfer et al , Heward et al , Kegler et al , and Riley et al . In their conceptual model of community capacity development, Dressendorfer et al  stress three facilitating factors to consider in the successful implementation and maintenance of health-promoting initiatives: well-functioning and engaging leadership, policy-making, and an existing operational infrastructure. Further, the authors regard community-level capacity building as the foundation for sustainable and long-term development, where support for development is provided by a combination of dedication, resources and competences. These factors correspond to the influences found in the current study.
Beyond these explicit external and internal influencing factors, the initiation of targeted local health-promotion actions showed a latent implicit connection with external health targets. One external influencing factor was the parallel national preparatory work for the national PHP, which started in 1995. Issues of public health were reported in the national and local media, and also revealed in local health statistics. Yanovitzky  has declared that, even if there is no clear evidence of a connection between media reporting and healthier behaviour, the policy-making of official authorities is affected, which has a positive influence on health promotion in the community.
The results of the three phases of development described in this study (initiation, action, achievements) also correspond to the three streams (problem, politics, policy) in the Kingdon model . In all four municipalities, youths' alcohol habits were a known and defined problem. There was also a united political will to solve the problem. In Municipality D, for example, a renewal of alcohol policy was seen as one solution. And there were both officials and politicians who could be identified as window-opening policy entrepreneurs, and who made efforts to reach a broad and intersectoral foundation for the work. In Municipality D, the policy process led to a new alcohol policy, which was adopted and implemented. Another example is the policy process involved in the child impact analyses pursued in Municipality B.
The processes and factors related to health-promotion development in four Swedish municipalities can also be discussed from the perspective of theories of social change. Thompson and Kinne  describe a holistic view, where individuals are seen as actors in the context of a community system. There are four levels from a top-down perspective, starting with the external environment, and then descending in turn to the community, organization and individual. In this context, there are several connections with the development process of initiation, action and achievement that we have described. Even though the municipalities did not have any clearly defined or organized local health-promotion objectives, they did manage to develop measures and programs aimed at a healthier community.
This study's results correspond to those of Baum et al , in which successful health promotion is seen in terms of its sustainability. Several of the essential factors referred to by Baum and colleagues also emerged in this study. An important one is the ability to transform projects into well-integrated and long-term programmes. Others are a clearly defined leadership, adjustment to local preconditions, wide-ranging local support, intersectoral working, and an ability to tackle competing interests.
The weakest connection is the absence of a clearly expressed vision of health in the four municipalities. The municipalities' own conditions and needs have had an obvious influence on the development of long-term and intersectoral local health promotion. This is regarded as a key issue in health development by Dahlgren and Whitehead . Further, that individuals and societies have the will to do something, if feasible, is a more important factor than what can be done technically.
The components presented and the increasing awareness of community health as a resource and important factor in the development of a well-functioning municipality, and also concern for a sustainable environment and economy, have led to increasing attention being paid to public-health issues and health promotion.
Studying local health promotion and its growth inductively and from a communal bottom-up perspective faces several difficulties. Each case is complex, contextualized, and not possible to control.
The municipalities in this study were not selected on the basis of how their local health-promotion activities were organized or developed. The purposeful sample was based on general factors, such as demography or geographical location; only thereafter was an investigation conducted into the municipalities' local health-promotion content, organization and development.
The aim of the study was exploratively to identify the municipalities' local health promotion from their own perspectives, and thereafter the factors that have influenced and characterized their development. In this context, the case-study approach is a suitable research strategy, since cases offer the opportunity for the study of processes and connections, in each particular case, and also a systematic way of handling and analyzing data [22, 37]. We chose the inductive approach in order to capture the municipalities' views in an unprejudiced manner. We rejected a theory-driven deductive study design and analysis for the same reason; it would not have given us the municipalities' own objective views on their local health promotion. But, as Yin  puts it, an explorative study must have some stated purpose even though it may not present specific propositions.
Rootman's  seven principles for health promotion were taken as our starting point. Also, the design of this study, which encompassed a case-study protocol and several data-collection methods and data sources, enabled triangulation; validity is thereby strengthened through the interpretations of reasonableness that can be made from the different parts of the results.
The processes involved in the development of local health promotion could therefore be followed horizontally over time, and vertically, by monitoring, inter alia, decision-making processes, and also by the strategic sampling of interviewees at different levels in the municipal organizations and the examination of policy documents.
A common criticism of case studies is that, by virtue of their focus on a strictly delimited area, they hinder or prevent transformability. According to Yin , this criticism can be counteracted through the use of multiple cases. More cases make for a more robust study, in that analytic conclusions can be drawn from independent cases, which thereby strengthen their credibility. The number of inhabitants in Sweden's 290 municipalities ranges between 2 500 and 810 000. In relation to population size, the four municipalities in this study belong to the most common group of Swedish municipalities. There are no Swedish studies showing that bigger municipalities give greater priority to health promotion than smaller ones. But differences related to size have been found with regard to how municipalities have organized their local health promotion . In order to promote population health and reach national and global goals for public health, intersectoral and multi-level cooperation and action are important. The municipality's role in this work as a local-level public sector actor has been stressed in, for example, the Ottawa Charter  and the Commission on Social Determinants of Health .