Main findings
In our study, an information leaflet on the risk-benefit balance of breast screening did not appear to influence overall participation in the program. However, in one hospital attending a population with lower socioeconomic status the leaflet reduced participation (− 1.4% in the intervention group in Hospital A). Furthermore, the per-protocol analysis showed an overall lower participation in the intervention group (− 2.6%) in both hospitals (− 3.9% in Hospital A and − 2.4% in Hospital B), although this difference was not statistically significant in Hospital B.
Comparison with previous studies
We found that providing explicit information about the risk-benefit balance of screening for the first time to women may not dramatically affect participation in the program. However, this intervention reduced participation in areas with a low socioeconomic status. Similar to our overall results, a randomized controlled study in Italy found no differences in participation among women who received extensive information material [
15]. That study evaluated the impact of a comprehensive leaflet with additional information, which explained the balance in greater depth than our leaflet. Our results are also consistent with those of a German study reporting that providing more information about screening did not seem to affect women’s intention to participate in breast cancer screening [
16]. Conversely, several other studies have suggested that providing information on the benefits and adverse effects of breast cancer screening decreases women’s intention to be screened [
11,
12,
14,
17]. In agreement with these studies, the participation rate decreased in one of our hospitals.
Of note, how the information is presented may influence its effect. One of the studies was criticized because the decision aid placed too much weight on informing women of the risks of screening [
18]. We believe that our leaflet is well-balanced and it clearly shows that the Catalonian Health Department invites women to be screened. However, the presentation of the leaflet could partly explain the differences in participation found in one of the centres. Overall, we did not find that providing information increased attendance in any group, or in the published literature.
A systematic review found that the use of decision aids made younger women (those around the age at which women begin to receive screening invitations) reluctant to participate in breast cancer screening programs [
13]. In our trial, we found no differences in age between the intervention and control groups. Moreover, another study found an inverse relationship between age and reading information on breast cancer screening, a finding that contradicts the possible explanation of a lower participation of young women in Hospital A due to the leaflet [
19]. On the other hand, another study argued that it was unlikely that women who have never participated in screening read screening information [
19]. In Hospital B, most women had never participated in our screening program and so there is a possibility that a lower willingness to read screening information in this group could explain the absence of an effect of the leaflet in that hospital. However, we were unable to assess this possibility as we did not have a measure of reading the leaflet.
Women who have never participated in screening have been found to assign more weight to adverse effects than women who have participated at some point in the screening program [
9,
20]. In our study, there was a similar proportion of these women in the intervention and control groups, and therefore this tendency seems unlikely to have influenced our results. However, another factor that has been related to a higher perceived importance of screening harms and lower perceived benefits is awareness of recent breast screening recommendations [
10]. This hypothesis may be reflected in our results: women in Hospital B could be less willing to read the leaflet than those in Hospital A; in the intervention groups, attendance was unaffected in Hospital B but was reduced in Hospital A.
In a study focusing solely on non-participant women in Scotland, 55.5% of them had a history of non-attendance at screening [
21]. In our study, around 85% of non-participant women had never attended our screening program. A possible explanation for our results could be that the areas included in our study have a high proportion of women who are disengaged from our screening program, and thus are not influenced by the information that we provide.
Another possibility could be that people do not usually read the information provided in the invitation letter, and decide to participate in the screening program based on information known beforehand through general practitioners, family or other sources of information.
The participation rate in this study was lower than that for the whole screening program in the last decade of around 55%. The low participation rate in Hospital B is probably explained by the high uptake of private breast screening in the area covered by this hospital. In addition, the COVID-19 pandemic started at the beginning of March 2020. Although this was the end of the study period, it could have negatively affected the expected participation to some extent.
Strengths and limitations
The main strength of our study is the randomized controlled design, with information about women’s actual participation in the breast cancer screening program. This allowed us to assess whether providing information on the benefits and adverse effects of screening along with the invitation to the program affects participation.
Our trial also has several limitations. One of them is that we randomized letter processing days rather than individual women. However, processing days are similar to natural homogeneous grouping of the population, and within each day the population is heterogeneous, so we do not consider that this influenced the internal validity of the study. Another limitation is that the catchment areas included may not be representative of the general population, and thus the results may not be applicable to other populations, as they include areas with high socioeconomic status and others with low socioeconomic status. For that reason, we conducted an analysis stratified by hospital. Consequently, the stratified analysis may not have sufficient statistical power to detect statistically significant differences in Hospital B.
The follow-up was initially established at 90 days after the scheduled date of the mammogram, but this period overlapped with the SARS-CoV-2 pandemic, during which time mammogram appointments were cancelled. Thus, once the trial had started, we were forced to limit the follow-up to 30 days. However, around 90% of the women who were actually screened did so before that period, so we believe that shortening the follow-up probably did not alter the results of the study. In addition, we retrospectively registered the trial. However, we explained throughout the manuscript all deviations from the protocol study.