Increasing Attendance to Cervical Cancer-screening Among Immigrants

Goals and objectives of the project The present study is designed to evaluate the effectiveness of two different community-based strategies to increase the attendance to the cervical cancer screening prevention program. Interventions have been developed together with stakeholders and conducted in two different geographical areas in Norway. Each intervention group will be compared to a control population in the same county in the first place. The main outcome to compare will be the percentage of women in each group attending to the screening program. However, the investigators will not only focus on this quantitative outcome, but also on the underlying processes determining how communities and individuals change their behaviour.

Secondly, the effect of the interventions measured quantitatively will be compared to each other in order to be able to prioritise one of the two strategies if needed, although strategies will probably be complementary. A particular interest of the interventions is to examine whether and how immigrant women respond to different intervention strategies.

Methods This study is designed as a community-based health intervention trial. A key feature of such trials is the allocation of intact communities or clusters of individuals rather than individuals themselves to different intervention groups.

The study will rely on both quantitative and qualitative evaluation of the interventions. For the quantitative part,each of the two interventions will be compared to its control. In addition, the effects of the interventions will be compared. Additionally, to know how the interventions increased (or not) attendance to the screening program post-intervention focus groups among intervention subjects will be conducted.

Interventions, participants and participating areas Two different intervention strategies have developed, one of them meant to increase awareness and attendance among users, and the other one targeting health professionals.

1. User intervention Area and design The developed User intervention will be implemented in the counties of Akershus and Buskerud. Women from Somalia and Pakistan were chosen as targets for this intervention after previous analyses for register data. Sample size was calculated for an increase in cancer screening participation from 0.45 to 0.55 with 80% power and 5% significance level. The investigators tried several intra class correlation (ICC) levels, and ended dividing the area in 16 clusters that were matched according to the calculated number of female immigrants aged 20 to 66 from Somalia and Pakistan. After assuring that pairs were not naturally linked by a mosque or other natural known gather centre for these populations, clusters were randomly assigned to intervention or control groups. Participants All women from Somalia and Pakistan living in the intervention clusters compose the intervention groups; immigrant women from the same countries of origin living in control clusters will be control participants. It is estimated that 625 and 915 women in screening age from Somalia and Pakistan respectively live in the study area (Akershus and Buskerud). Approximately half part of the women from each country will be in the intervention clusters. Intervention Bilingual key women from the two chosen countries were recruited and trained as peer educators for their respective communities to increase awareness regarding the importance of cervical cancer prevention and to help peers overcome detected barriers and increase attendance to screening. The strategy to achieve these goals relied upon meetings with the women organised according to their practical needs, using videos and other interactive materials.
2. Heath professional intervention Area The developed Heath professional intervention will be implemented in Bergen. Sample size was calculated for an increase in cancer screening participation for non-Western immigrants (all groups together) from 0.45 to 0.55 with 80% power and 5% significance level. The investigators tried several intra class correlation (ICC) levels, and finally divided the area in 20 clusters that were matched according to the cluster size and the calculated number of female non-Western immigrants aged 20 to 66. The clusters were thereafter randomly assigned to intervention or control groups. Participants All General practitioners working in the intervention cluster areas were assigned to the intervention group; similarly, health professionals working in the other clusters will be chosen as controls. Women will for the analyses be linked to their general practitioner through the General Practice Database. Intervention Based on the existing evidence and on interviews conducted with health professionals, an appropriate intervention was developed to increase awareness of the health professionals involved in screening working in the chosen areas. Health professionals in the intervention areas will be contacted by letter and a personal appointment was made to explain the intervention. No change in "screening as usual" will be implemented for health professionals working in the control areas.
3. Post-intervention qualitative studies Among participants included in the intervention areas, focus groups/in depth-interviews will be conducted right after the interventions are implemented among both users and health professionals to understand whether and how each of the interventions worked.

Measurements The study relies on both quantitative and qualitative evaluation of the interventions.

1. Outcome information on screening attendance for intervention and control groups will be obtained from the Norwegian Cytology registry, where all pap smears taken in Norway are registered. The leader of the Norwegian Cytology register is part of the Reference group for this study, and this register can provide data for the variables Morphology (if the test is taken, if it is normal or pathological) and date for test for all women in Norway after the necessary permissions are given. This register will be then linked at the individual level to the Norwegian National Registry in order to obtain information regarding immigrant background, country of origin as explained below, reason for migration and length of stay in Norway for all 25 to 69 years old registered in Norway, in addition to other socioeconomic variables (age, education, income level, work status, cluster area for those in intervention and control areas, geographical centrality, civil status, husbands' immigrant background and GP's immigrant background) that will be used to adjust our results, and to two variables from the GP database (GP's age and gender) to identify the GPs in the intervention and control groups in the Health professional strategy. Data will be linked by Statistics Norway before being accessible to the researchers. However, since ethnicity is considered sensitive data, and in order to make the dataset completely anonymous for the research group, we will only apply for the following groups with regard to country of origin (see page 1 for approximate numbers): Norway, Western Europe, Eastern Europe, Middle-East, South-East, South Asia, Africa, North America, Central and South America, Poland, Sweden, Pakistan and Somalia. As explained below, data will be obtained for several years (2012-2018) in order to be able to distinguish secular trends in attendance to cervical screening versus the effect due to our intervention. Data will be stored in a secure server at the University of Bergen (BIOS, Core Facility for Biostatistics and Data Analyses) in an anonymised form.
2. In addition, the process of how communities and individuals change their behaviour will be studied using a qualitative approach to the intervention groups after the interventions. These data will do not include any information that can lead to personal identification of the women.