Prevent CKD progression - use claims data to track behavior change and quantify the effectiveness of the awareness letter

In part 2, we share the evaluation methods and results of this project, future developments, and thoughts on collaboration.
（You can read Part 1 from here）

How is a "doctor visit" defined? Data is used to establish outcomes and the control group

―― Could you tell us how you evaluate behavior change?

Kento Takeyama：We use the health insurance claims data for the three months after the letter is sent to evaluate whether or not the patient has visited a doctor. However, it was very difficult to determine what constitutes a "doctor visit," since CKD is rarely diagnosed at the first visit.
Therefore, we extracted data on those with a confirmed diagnosis of CKD from our anonymous claims database of approximately 1.2 million patients, analyzed the patient journey until the diagnosis was made, and after discussions with specialists and AstraZeneca colleagues, defined outcomes as "biochemical test performed," "lifestyle disease-related diagnosis" and "kidney disease-related diagnosis."
Another important point is that we found people who were close to the target population from the data and created a control group.

Lei Liu：For example, even if we find out that "30% of the people who received the disease awareness letter visited a doctor," we don't know how we should evaluate that, do we? Because some people will see a doctor for some reason without any kind of campaign. We can only clarify the impact by comparing the results with those of the control group.

Takeyama ：Use data to define outcomes. Evaluate the impact in the control and intervention groups. I believe that these are all unique points that PREVENT has been able to achieve through our accumulated expertise.

Liu：Of the 740 subjects, 312 (42%) showed behavior change ,which was more than we had expected. This was one of the results that validated the hypothesis that "patients become aware of the risk, bring the awareness letter to the doctor, the doctor becomes aware of the kidney risk and performs tests, and the test results lead to a kidney disease-related diagnosis," and it also suggested a behavioral change in which doctors perform tests based on the letter brought by the patients. Since the difference between the two groups started to appear from the 0th month after the letter was sent, and the difference gradually increased, it would be very interesting to see what the results would have been if we had followed up longer.

●Outcome assessment

Medical practice among survey respondents who have seen a doctor or plan to see a doctor vs. control group

●Outcome results (Testing rate)

Comparative analysis of biochemical testing rates (urine, feces, and blood) based on claims data in the intervention and control groups

●Outcome results (name of lifestyle disease and kidney disease-related diagnoses)

Comparative analysis of rates of lifestyle disease-related diagnoses¹ and kidney disease-related diagnoses based on claims data in the intervention and control groups

• 

• 

―― Are there any challenges that you see in the process or results?

Takeyama ：Behavioral change varied by group. For example, we were not able to encourage those who "have not been to a hospital in the past year" to see a doctor. Additional measures for this segment of the population need to be further explored.

Liu：In the so-called "bedrock segment," who are at high-risk and need early behavioral change, we did not see behavioral change in this project either. We recognize that this is another issue that we need to continue to consider. In the meantime, it is difficult to clearly track the results of conventional disease awareness activities such as TV commercials and public lectures. For the first time, this project allowed us to quantify the results of our efforts. We can discuss ROI (Return on Investment) and say that we now have a KPI (Key Performance Indicator) for continued investment in disease awareness activities.

“Win x 4” model for both scale-up and rollout to other diseases

―― Is it possible to apply it to diseases other than CKD?

Takeyama ：If the disease is such that the right patients can be found from the health checkup or claims data, it is possible to roll out to other diseases.

Liu：Of the diseases my team is working on, it can be rolled out to respiratory and liver diseases. There are a number of diseases that have a long course, with low risk awareness in mild cases, and can be diagnosed early through disease awareness.

Takeyama ：The fact that we were able to raise awareness of diseases other than lifestyle diseases gave us confidence and reaffirmed the significance of disease awareness. In order to optimize medical costs and improve the quality of life of the insured, we would like to target a wider range of insurers and promote disease awareness for diseases other than lifestyle diseases.

―― What do you think is the motivation and significance of this project?

Takeyama ：PREVENT is a startup that implements a system to support "ichi-byo sokusai," which means people can live a healthier life while coping with a disease. This disease awareness campaign is the embodiment of that vision, and we are very proud of the results we have achieved!

Liu：I think we have created a prototype for AstraZeneca's "patient-centered business model.” By providing accurate information to patients, we can make them aware of their own risks and encourage them to take action. It is very significant that we were able to present this model with evidence.

Takeyama ： It has become a “win x 4” business, which makes patients, insurers, AstraZeneca and PREVENT happy.

Liu：For i2.JP, we want to introduce this as the first model that we can scale up.