Understanding preferences for low-dose aspirin use in primary prevention of colorectal cancer among clinicians and the general public in Hong Kong: A discrete choice experiment

Professor Chan, Esther Wai Yin   (Principal Investigator (PI))

Dr Wong Carlos King Ho   (Co-Investigator)

Professor Wong Martin Chi-sang   (Co-Investigator)

Professor Tsoi Kam Fai, Kevin   (Co-Investigator)

Professor Li Xue   (Co-Investigator)

Dr Yu Yee Tak Esther   (Co-Investigator)

Professor Cheung Ka Shing Michael   (Co-Investigator)

24

2021-08-01

2025-03-17

797676

Understanding preferences for low-dose aspirin use in primary prevention of colorectal cancer among clinicians and the general public in Hong Kong: A discrete choice experiment

Colorectal cancer, Discrete choice experiment, Low-dose aspirin, Pharmacoeconomics, Preferences

Others - Medicine, Dentistry and Health

18191851

Health and Medical Research Fund - Full Grant

2020

Completed

In Hong Kong, colorectal cancer is the most common cancer with 5635 new cases in 2017. Low-dose aspirin has been recently reported to lower the risk of colorectal cancer. There has been no studies evaluating patients and clinicians preferences regarding the treatment. Objectives To elicit clinicians and the general public’s preferences for low-dose aspirin therapy in the primary prevention of colorectal cancer. Design and subjects A discrete choice experiment survey will be conducted among clinicians and the general public. A telephone survey agency will be used to randomly select a representative sample of individuals, which will then be sent the questionnaire by mail or email. Individuals aged 40 years or above will be eligible to participate. All clinicians will be eligible and invited to participate. Main outcome measures The participants’ preferences of low-dose aspirin therapy for the prevention of colorectal cancer will be measured and analysed according to the survey data of stated-preference choices as decided by the participants. Preferences for efficiency, safety, dosage, duration and out-of-pocket cost will be elucidated. Data analysis The random effect logit model will be used for measuring the relative importance of each attribute. Log-likelihood, pseudo R-squares and Akaike and Bayesian information criteria will be reported to inform on the goodness-of-fit of our regression models. P-values <0.05 (two-tail) will be considered as statistically significant. Expected results The most and least important attributes in determining the treatment choice will be identified. The relative importance between any paired attribute would be assessed.