General Information
Project description
In partnership with Godot Inc. the Yomitan Village Office in Okinawa (Japan) conducted an AI-driven sludge audit of the medical check-up process. Since the 2022 financial year, the village office and Godot Inc. have been working together to apply behavioural science to medical check-up notifications delivered via SMS and postcards.
In 2021, Yomitan Village had the lowest general medical check-up rate among all 41 municipalities in Okinawa Prefecture. However, the application of behavioural science to local policy and communication has reversed this trend. The screening rate increased by 7.8 percentage points in FY2022 and by 5.3 percentage points in FY2023 compared to the previous year, whereas the average increases across Okinawa’s municipalities were only 1.7 percentage points in FY2022 and 1.3 percentage points in FY2023.
Amid these improvements, in order to further effectively apply behavioural science and provide better public services, an AI-driven sludge audit was conducted in fiscal year 2024 based on methods provided by the OECD and the government of New South Wales, Australia.
The sludge audit was successfully completed based on the methodologies advocated by the OECD and the government of New South Wales, with some context-specific adjustments made to reflect language, cultural, and communication differences in Japan. Some of the suggested improvements included (1) Changing the message from an administrative perspective to one that highlights the benefits for village residents, (2) reducing the number of clicks required to reach the reservation screen within the app and (3) sending an explanation of what happens after the reservation via the app to help users prepare for the next steps and reduce post-reservation drop-off.
Detailed information
Final report: Is there a final report presenting the results and conclusions of this project?
Pre-analysis plan: Is there a pre-analysis plan associated with this registration?
Hypothesis
The sludge audit method advocated by the OECD and the government of New South Wales helps government officials in Japan improve their service processes, provided that context-specific adjustments are made, thereby increasing the screening take-up rate by minimising drop-outs during the booking and screening attendance processes.
How hypothesis was tested
After conducting the sludge audit, we presented the UX score and points for service improvement to the public servants. The first step in hypothesis testing was for the public servants to review these and determine whether they can be used to actually improve the service flow. They consider insights gained through a sludge audit were valuable and they can improve the service based on the suggestions.
For the next step, we will verify the hypothesis by checking data on the actual service improvements, such as the screening or the proportion of villagers who opened the SMS and completed the screening reservation.
Dependent variables
The number of sludges reduced through service improvements, along with variables such as search cost, decision cost, cognitive cost, emotional cost, the appointment completion rate of those who opened the SMS screening recommendation notification, and the screening participation rate, can be considered. If the appointment completion rate of those who opened the SMS screening recommendation notification is obtained, it is believed to be the primary outcome of the sludge audit.
Analyses
By comparing the reduction rates of sludge and behavioural burdens, as well as the appointment completion rate of those who opened the SMS screening recommendation notification, with the previous service flow, we believe the effectiveness can be analysed.
Additional information
Who is behind the project?
Project status:
Completed
Methods
What is the project about?
Date published:
23 June 2025