Decision aid on breast cancer screening reduces attendance rate: results of a large-scale, randomized, controlled study by the DECIDEO group

Controversies regarding the benefits of breast cancer screening programs have led to the promotion of new strategies taking into account individual preferences, such as decision aid. The aim of this study was to assess the impact of a decision aid leaflet on the participation of women invited to participate in a national breast cancer screening program. This Randomized, multicentre, controlled trial. Women aged 50 to 74 years, were randomly assigned to receive either a decision aid or the usual invitation letter. Primary outcome was the participation rate 12 months after the invitation. 16 000 women were randomized and 15 844 included in the modified intention-to-treat analysis. The participation rate in the intervention group was 40.25% (3174/7885 women) compared with 42.13% (3353/7959) in the control group (p = 0.02). Previous attendance for screening (RR = 6.24; [95%IC: 5.75-6.77]; p < 0.0001) and medium household income (RR = 1.05; [95%IC: 1.01-1.09]; p = 0.0074) were independently associated with attendance for screening. This large-scale study demonstrates that the decision aid reduced the participation rate. The decision aid activate the decision making process of women toward non-attendance to screening. These results show the importance of promoting informed patient choices, especially when those choices cannot be anticipated.

management structures in the participating departments took part in the study. Women already diagnosed with cancer were excluded.

Randomisation
Women aged between 50 and 74 and registered with the French Health Insurance System in the 11 departments were randomly selected through a list-based sample to participate in the study. Women were randomly assigned in a 1:1 ratio via a computer-generated, centralized randomization sequence, which was done with a block randomization of four, to the DECIDEO or usual invitation group. The randomization was balanced through stratification according to the following hierarchy: the department, the age according to 2 classes (above or below 65), and the number of invitations already received by the women (leading or not, to participation in national screening) The study was approved by our institutional review board (Ethical Committee of Saint Etienne University Hospital, December 4 th 2008), which waived the need for signed and informed consent according to French law, since the study involved healthy women and did not involve treatment.

Decision aid group
Women allocated to the decision aid group received an invitation to participate in the national breast cancer screening program and the specially-designed decision aid (a leaflet), by mail.
The paper-based leaflet DECIDEO is a 12-page pocket leaflet providing scientific information about the advantages and disadvantages of participating in the national breast screening program, understandable by all, constructed to conform with the International Patient Decision Aid Standards 33,34 . The aim of DECIDEO was to inform women and to help them reveal their true preferences for breast cancer screening based on scientific evidence.
(Supplementary appendix online only). The DECIDEO leaflet was designed by a multidisciplinary team including physicians, methodologists, sociologists and economists.
The leaflet was validated for its psychometric qualities (understandability, reliability, validity) and tested in a sample of targeted women in a monocentre, pilot study (unpublished results).

Control group
Women in the control group also received an invitation and the usual standard information, by mail. This invitation is an administrative letter sent by all the screening management structures to women scheduled to be invited to participate in the national screening program every two years from the age of 50 onwards. The usual standard information document that was sent with the administrative letter was specific to each screening management structure.

Data collection
The French Health Insurance System transmits contact details to the cancer screening management structure in each department. The screening management structures retain individual variables for each woman aged 50 to 74 included in their database. For women included in the DECIDEO study, all screening management structures supplied us with these variables: age, department of living, number of invitations sent, last screening date and postal address. We used those variables to compare baseline characteristics by study groups.
The postal address of each woman was geocoded on global positioning system (GPS) coordinates using Google Map® (Google Inc, California, USA); these coordinates allowed us to attribute 2 sets of variables to women:  Coordinates allow us to identify for each woman if she was living into a urban or rural area  Coordinates were linked to the smallest French geographic scale available, the IRIS zone. Each woman were attributed an IRIS zone. The mean household income assigned to the IRIS zone where the woman lived in (source : National Institut of Statistics and Economics Studies (INSEE)) was then attributed to each women.
All those variables were compared by study groups into table 1

Outcome measures
Twelve months after the invitations were sent, each screening management structure collected the participation status of the randomized women. Each structure has the contact details of all women aged between 50 and 74, eligible to receive an invitation to the national screening program for the entire department in a database. This database also records if and when a woman attends screening through automatic transmission of information from the radiologists.

Primary outcome
The primary outcome was the women's attendance rate for the breast cancer screening program during the 12 months following the invitation.

Secondary outcome
The secondary outcome was the delay between the invitation and the date of attendance for breast cancer screening. In addition, demographic details and other characteristics of the women were collected. The level of the household income was estimated based on the mean household income level from the town where the women lived.

Sample size
The study sample size was calculated with an assumption of 50% of attendance rate (which was the mean participation rate observed in the 11 participating departments, in 2007). We estimated that there would be a 3% modification in the attendance rate (a 6% relative modification). With an alpha risk of 5% and a beta risk of 95%, we calculated that we needed to include 7 209 women in each group for a bilateral test (since a positive or a deleterious effect of the intervention could equally be possible). Taking into account the bias of lost to follow-up as well as the risk of contamination , we increased the group size by 10%, giving a sample size of 8 000 women in each group.

Statistical analysis
In compliance with the statistical analysis plan written before study completion, the results were analyzed using a modified intention-to-treat population, defined as all women who were randomly assigned, except those who had had a mammography before randomization. The statistical analyses were performed with a Pearson's Chi Square test (Fisher exact test if statistical conditions were not satisfied) for ordinal variables or a Student's T test ( Wilcoxon test if statistical conditions were not satisfied) for continuous variables . Variables that were significantly associated with attendance in univariate analyses (p. value <0.05, conservative as regards to the statistical power of the study) were introduced in a stepwise manner in a multivariate logistic regression model to identify independent predictive factors (exit p.value<0.05). Variables tested positively for correlation were not included in the multivariate analysis. We also compared the primary outcome in post-hoc defined subgroups ( Age, Departments, the Household Income and Having or not recieved a previous invitation).
Heterogeneity was tested among subgroups. All p-values are two sided, with the threshold of significance set at p<0.05. All analyses were carried out using SAS version 9.3.
A steering committee was established to guarantee the accuracy and completeness of the data and the analyses and the respect of the study protocol. The committee validated the decision to submit this manuscript for publication.

Breast cancer….What about you ?
What choices exist?
Page 2

Being aware of breast cancer risks
The risk of breast cancer is related to the woman's age. Among 100 women followed from the age of 50 to the age of 80, 10 will have breast cancer.

The seriousness of breast cancer is related to the relapse of the disease:
The relapse in the localized region of the breast or the evolution of the illness at a distance (also referred to as metastases) The risk of cancer death is related to this evolution at a distance.
The risks of relapse and death caused by breast cancer in the 10 years following its diagnosis, depend on the size of the tumor at the moment of diagnosis.  less than 1 cm  Between 1 and 2 cm  More than 2 cm The smaller the breast cancer, the easier it is to cure. Treatment is thus less traumatic.

Knowing about breast cancer screening
What is breast cancer screening?
It is the repeated realization of a mammography (a breast X-ray). It allows the detection of breast cancer, usually at an early stage, before one can feel it.
 The mammography can be prescribed by your physician : it is opportunistic screening  The national insurance system invite all women between 50 and 74, every 2 years by mail, for a mammography : population-based national screening

Figures shown in this document come from the data-base of the national population-based screening since no data are available for the opportunistic screening
Page 4

Breast cancer screening results
A mammography allows the search for anomalies, some of which are cancerous.
For 1000 women screened, 897 have a regular mammography (true negative). For 1000 women going through screening, 7 are found to have breast cancer (true positive). Other anomalies found are benign. 94 women among the 1000 are concerned (false positive).
Among those 94, 37 will need to be followed more frequently with an extra ultrasound and/or a mammography To confirm the absence of cancer, it is sometimes necessary to make a needlesample (puncture, biopsy) or a surgical sample.
Two women among 1000 will undergo surgery for a benign anomaly.
Some cancers are unseen or cannot be seen through screening. Among 1000 women screened, 1 to 2 will be concerned (false negatives). These women will develop cancer between two mammographies. During screening, 18% of the cancer tumors are too small to be diagnosed. Breast density, which varies in women, can be responsible for the non-detection of cancer.
For example a 1 cm cancer tumor could be detected in a "clear" breast and be undetected seen in a "dense" breast. (cf pictures) The population-based national screening aims at lowering this risk of false negatives

Screening advantages
In the case where breast cancer is discovered, the smaller the tumor is, the lower the chances are for relapse within the 10 following years. Women who have regular mammographies have more chance of being diagnosed with a small tumor.

Screening allows a decrease in death caused by breast cancer.
 Among 100 women who have never undergone screening and who have been diagnosed with and treated for breast cancer, 40 will die from this cancer  Among 100 women screened every two years between 50 and 74 years old, diagnosed with and treated for breast cancer, 26 will die of this cancer