Development of strategies for evaluating a community intervention programme for cancer prevention through dietary change

Abstract

One of the main strategies of the Stockholm Cancer Prevention Programme (SCPP) is to lower cancer risk in the population of Stockholm County through changes in dietary habits. It is planned to achieve these changes through a ‘community intervention’ approach: that is, an intervention programme aiming at geographically defined populations, driven by a small group of SCPP staff but working with and through ‘participant organizations’ already active in the communities concerned. Supply will be influenced through organizations such as food producers, wholesalers and caterers; demand will be influenced through primary and occupational health care, schools, voluntary organizations and the mass media. This diet-cancer programme poses particular problems for evaluation. Some of these problems follow specifically from the community intervention approach, and are associated with the diffusion of control over targets and levels of intervention, the many intervention ‘channels’, and estimation of the time and effort devoted to the programme by members of participant organizations. Other problems are associated with measuring dietary habits and attitudes, and with the aetiology and natural history of cancer. Furthermore, four distinct issues are involved, each requiring a different kind of evaluative approach. First, there is the question of how successful the SCPP staff have been in stimulating appropriate activity in participant organizations. This will be investigated through budgetary data, diaries, registers of contacts, and activity reviews. Second, in a programme expected to last 10 years or more, evaluation has a role to play in programme development; questions arise such as whether specific programme activities are being effective, who they are reaching, and how they might be improved. These will be investigated through a range of relatively short-term studies which will generally take as their sampling frames the ‘membership’ of selected participant organizations. Third, there is the question of the extent to which the programme as a whole has achieved its objective of changing the diet of people in Stockholm County. This will be investigated by a cohort design based on random samples from populations of different municipalities (administrative areas within the County). The design proposed exploits the way in which the intervention will be phased over its 10-year period, with some municipalities being intensively involved from the beginning, and some being less involved until relatively late. The cohort study will also include control areas outside Stockholm County. In support of this there will be small cross-sectional studies to investigate issues such as learning effects, and to compensate to some extent for the expected cohort attrition. Fourth, there is the aetiological question: how effective is dietary change in preventing cancer? Data on cancer morbidity and mortality will be available from routine sources, but changes in cancer incidence resulting from the programme are not expected for a decade at least. In 10 to 15 years it will be possible to relate aggregate data for municipalities on dietary change to aggregate data on change in cancer incidence, but detection of any link that may exist between diet and cancer cannot be ensured within a moderate time period and budget. © 1988 National Association for Music Therapy, Inc.