Abstract
Objective The consultation rate in Japan for mammography screening for breast cancer continues to drop. In order to examine this issue based on demand for mammography, the present study was conducted using Discrete Choice Experiments (DCEs), a type of Stated Preference (SP) method. The objectives of this study were as follows: 1) To consider what attributes of mammography screening are being potentially evaluated by general recipients in the target age group; 2) To verify the validity of the SP method by separating the sample group into sub-groups of previous mammography recipients and non-recipients, and to compare their findings; 3) To predict selection behavior by setting scenarios for screening options possibly in demand in the future.
Methods 800 subjects aged between 40 to 59 years and with no history of breast cancer were randomly selected from the general population of Tokyo. A DCE was conducted using postal self-administered survey forms. A total of 301 survey responses were obtained. Subjects were presented with a pair of hypothetical screenings, including 5 attributes regarding mammography screening, and asked which screening they would prefer to receive. For the entire sample and sub-groups, estimations for parameters were made using the conditional logit model setting the screening attributes as independent variables and the selection of whether or not to receive each screening as a dependent variable. Based on these results, short-time/high-cost and long-time/low-cost screening options were set and selection behavior was predicted.
Results The five attributes regarding mammography screening for all samples -total amount of time taken for the screening; degree of breast pain; possibility of breast cancer being missed during the screening; the effectiveness of reducing deaths caused by breast cancer; and the total cost required for the screening -were each estimated to be significant at the 5% level with coefficient signs consistent with expectations for the entire sample group. Next, comparing the estimated results of the sub-groups, a positive correlation was confirmed between behavior and preference. As for forecast of selection behavior, the percentage of respondents choosing the short-time option was predicted to be the same or higher than for those choosing long-time/low-cost option when the short-time option was offered for ¥7,500 or less.
Conclusion This study shows that subjects place significant value on attributes regarding the screening process as well as actual health outcomes. It also suggests the validity of using the SP methods in examining screening preferences. We found that short-time screening was able to compete against long-time/low-cost screening when it was offered for ¥7,500 or less.
These results suggest that offering mammography screening with favorable settings could increase the demand.
