Introduction Concerns about the ageing population and formal responsibilities of local governments to promote social cohesion and to enhance participation of vulnerable groups in society placed loneliness prevention high on the local policy agenda of Dutch municipalities in the past decade. The study described in this thesis was part of the Healthy Ageing programme of the Academic Collaborative Centre AGORA and aimed to contribute to more effective, evidence-based and problem-oriented approaches to healthy ageing at the local level. Aim The general aim of this thesis was to evaluate the effectiveness of a local intervention project – called Healthy Ageing – targeting loneliness among non-institutionalised elderly people. Healthy Ageing consisted of five intervention components, namely, a mass media campaign, information meetings, psychosocial group courses, social activities organised by neighbours, Neighbours Connected, and training of intermediaries. Methods First, the influence of socio-demographic and health characteristics on changes in loneliness over time and municipal differences in the prevalence of loneliness were investigated. Data were gathered from 9,641 persons who participated in the Elderly Health Survey of the community health service, GGD Noord- en Oost- Gelderland (former GGD Gelre-IJssel), in 2005 or 2010. Second, the overall-effect of Healthy Ageing on the initial outcome loneliness literacy, intermediate outcome social support, and ultimate outcome loneliness was evaluated using a quasi-experimental pre-test post-test design, including an intervention and control community. Baseline and follow-up measurements, in 2008 and 2010 respectively, were available for 858 non-institutionalized elderly people. The Loneliness Literacy Scale was developed within the context of this thesis and was pre-tested in a separate study among 303 elderly persons who also participated in the quasi-experimental study. Finally, delivery, reach, and acceptance of the individual intervention components was studied in several satellite studies. Data were collected by different means, e.g. project records and surveys among participants. Furthermore, the acceptability of the mass media communication materials, information meetings, and psychosocial courses of Healthy Ageing was studied by in-depth interviews with 14 clients of the meal delivery service in the intervention community. Results Overall and across municipalities, average loneliness scores did not significantly differ between 2005 and 2010. However, among the subgroup with mobility disabilities, loneliness was significantly higher in 2010. Furthermore, mobility disabilities and marital status were the most important factors explaining differences between municipalities. With regard to the evaluation of Healthy Ageing, the satellite studies showed that the reach and intensity of the intervention components were modest. Furthermore, from the interviews it appeared that the mass media communication materials were not successful in attracting attention because interviewees did not expect health information from these communication channels, the perceived personal relevance of the message was low, and the presentation was not attractive. Moreover, the content of the intervention components was not well received because the objectives and intervention components did not connect well with the priority group’s daily life. In addition, it appeared from the quasi-experimental study that 39% of the study participants from the intervention community was familiar with Healthy Ageing at follow-up. Overall, the intervention group scored more favourably on the loneliness literacy subscales, motivation (4.4%), perceived social support (8.2%), and subjective norm (11.5%) than the control group. However, no overall effects were observed for the intermediate and ultimate outcomes, total social support and loneliness after two years. Conclusion Given the modest overall intervention exposure, the effect of Healthy Ageing on the loneliness literacy subscale, motivation, is plausible, whereas on the subscales, perceived social support and subjective norm, probable, and on the subscale, self-efficacy, unlikely. Furthermore, whether the initial effects will carry forward to the intermediate and ultimate outcomes needs to be confirmed. The modest effects of Healthy Ageing can partly be explained by the challenges on organisational level which delayed and suppressed project implementation. Furthermore, the project might have benefited from a more systematic approach in order to ensure better alignment between the intervention components and formulated objectives. Finally, target group differentiation is highly recommended. This evaluation of Healthy Ageing illustrates how researchers can cope with the evaluation challenges of complex interventions which cannot be fully controlled. In turn, this provides valuable lessons for the development of intervention programmes and evaluation designs in public health practice.