Skip to content

Web App Operation Manual

1. Introduction

1.1 Purpose of this manual

This manual introduces how to use the demo of the policy making support technology "Policy Twin".

1.2 Prerequisites

  • User registration for the Fujitsu Research Portal (Fujitsu employees can use without registration)

1.3 Operating environment

  • Recommended browser: Google Chrome

2. Procedure - Experience Policy Twin in a sample project

2.1 Log in to the demo page

From the Fujitsu Research Portal, click 'Learn more' button for Policy Twin in the Technology List.

Click the 'Demo' button in the page to use the demo. The landing page of Policy Twin demo is as follows.

2.2 Select the sample project

This demo allows you to preemptively assess and compare the effectiveness of policies. On this page, sample projects are displayed, so click the 'Edit' button for the project you want to experience.

This demo runs in three steps: Registration, Policies, and Evaluation. On this page, you can view and check pre-registerd multiple policies in flowchart format. Click the '<' or '>' button at the top of the flowchart to view other policies. Click the 'Policies' tab to go to the next step.

2.3 Comparison of policies

In this page you can view and compare multiple policies. Select the checkboxes from the 'Policy candidates' list on the left to compare the policies. Check the differences in flowchart structures and branching conditions. Click the 'Evaluation' tab to see the effects of multiple policies.

2.4 Evaluation results of policies by simulation

In this page you can compare the effects of multiple policies. The 'Evaluation results' table shows each policy's name and target indicators, with numerical values representing simulated results. To view the details of the policies you want to compare, select the checkboxes in the 'MULTIPLE VIEWS' column of the policies you want to compare, and click the 'Comparison' tab.

The flowcharts and radar charts for multiple policies will be displayed.

Radar charts and scatter plots display the simulation results for each target indicator when policies are implemented, facilitating comparison to find the most effective policy.

To review details of each policy, click the 'PATHWAY' button below each flowchart. The diagram below shows the estimated number of people who would follow each route in the flowchart when the policy is executed. In this diagram, starting from the left, 5000 people are initially targeted by the policy. In the first diamond-shaped branching point, these people are divided into two groups: 869 and 85 people, who proceed to the right, while the remaining 4046 people do not receive the services after this branching point. Please scroll down the screen further.

The Sankey diagram below uses line thickness to indicate the flow of people through each service when the Policy is executed. Starting from the far left with the total number of people targeted by the policy, the diagram visually represents how this group divides at each branching point, ultimately showing the number of people who reach the final service on the far right(in this case, HOME-VISIT GUIDANCE).

3. Procedure - Perform the simulation for each use case

The following will explain the specific use cases.

3.1 Use case: Multiple Duplicated Medications

3.1.1 Overview

This use case addresses the issue of "multiple duplicated medications" in the medical field. For details on what constitutes multiple duplicated medications issue, please refer here. Also, please get sample data from here. Specifically, this use case is designed with the following usage scenarios:

You are a local government official of a fictional local government (Municipality A) and need to review this year's medication guidance project. While there is already a basic polocy, you have received a request to “consider whether the policy could be improved for greater effectiveness”. Nearby fictional local governments (Municipality B, Municipality C), separate reviews of their medication guidance policies are also underway. Using "Policy Twin", you aim to examine and incorporate insights from other municipalitys’ policies to help develop a more effective policy.

3.1.2 Creating a policy and checking the flowchart

Let's first use Policy Twin to check the flowchart of the digitized medication guidance policy for Municipality A. Following the instructions in "2. Procedure", log in to the demo screen and click the 'New policy' button on the landing page. First, enter "Digital Rehearsal for Municipality A" as the name.
Up to 10 projects, including sample projects, can be created in 'New policy' on the landing page. If you want to create more projects, wait until the created projects are deleted. (All created projects are automatically deleted at 00:00 (JST) 24 hours after creation.)
Please note that you cannot create more than one project with the same name.

Upload and register the "Municipality_A.bpmn" file from the sample data. In the window that appears, enter "Municipality_A" or similar as the name of the uploaded file.

This process displays the pre-improvement medication guidance policy for Municipality A as a flowchart, enabling confirmation. The flowchart illustrates the following steps from left to right:

  1. To identify residents who meet the criteria for duplicate medical treatment, check if a resident meets the criteria of "they have prescriptions from two or more medical institutions for the same medicine" and "these prescriptions are for 60 days or longer"

  2. If the criteria are met, extract them as the target of this policy.

  3. If they have a certain disease (e.g., cancer or an incurable disease), exclude them from the target group as an exception.

  4. The final remaining targets are notified by postcard or through pharmacy to confirm their agreement to participate in medication guidance. (Those who agree will then receive instruction from the responsible staff)

3.1.3 Running a simulation with the required data settings

For this example, assume that "medication guidance policies in multiple municipalitys have already been digitized and accumulated." Sample data includes flowchart files for Municipalitys B and C. These files allow you to run a simulation to evaluate the implementation of policies from Municipalitys B and C in municipality A.

Click 'SELECT DATA' to set up the data required for the simulation.

  • Settings
    • Reconfiguration mode: Select 'Off'.
    • Training data: Upload 'traindata.csv' from the sample data and click to select it.
    • Testing data: Upload 'testdata.csv' from the sample data and click to select it.
    • Flowcharts: Since you want to reference not only Municipality A but also other regions, upload 'Municipality_B.bpmn' followed by 'Municipality_C.bpmn' Ensure that three flowcharts are registered.
    • Click the gear icon in the upper right to move to the configuration screen. Click the gear document icon in the upper right and upload 'changeDisplayWord.json' from the sample data. Click the 'Save' button to return to the overview page.

3.1.4 Display the results of the simulation for each municipality's policy

Click the Policy Twin logo in the upper left of the screen to return to the landing page. Once the green 'SUCCESS' message appears, confirming that the simulation was completed successfully, click 'Edit' and then 'Evaluation' tab at the top of the page to view the results of the simulation.

As a result of the simulation, the effects of the policies in Municipalitys A, B, and C were calculated from five perspectives.

  • Medical cost savings: The amount of medical cost savings.
  • Number of notification: The number of targets who needed to be notified via postcard, pharmacy, etc.
  • Number of Health Guidance: The number of targets who agreed to participate in home visit guidance among those who were notified
  • Health Improvement Effects: The number of participants in the health guidance who are expected to show improvement.
  • Resources: The correlation between the predetermined maximum resource capacity for health guidance and the number of health guidance mentioned above.

The following insights can be draw from the results.

  • Looking at the 'Evaluation results' on this page, when the policy is implemented in Municipality A, the number of people who can receive health guidance is very small. In contrast, implementing the policy in Municipalitys C or B would allow for more health guidance.

  • In the 'Indicator scatter plot' on this page, if the Y axis is set as 'cost' and the X axis is set as 'medical cost savings', a positive correlation appears between the two. Municipality A is located at the bottom left, Municipality C in the middle, and Municipality B at the upper right, indicating that the medical cost savings increases as the number of health guidanc rises.

From these analysis, the policies in Municipalitys B and C may appear more favorable than those in Municipality A, However, in reality, the decisions must be made taking into account the circumstances of each municipality, such as the staff resources required for health guidance. For example, suppose Municipality A has only limited staff resources (just a few people). Implementing the same policy in Municipality B and under these conditions could lead to a dramatic increasing the number of people needing guidance, which may place a strain on resources.

To prepare for such cases, let's use the policy reconfiguration function to generate new policy candidates that consider the policies in Municipalitys A, B, and C.

3.1.5 Automatic generation of new policy candidates

Let's automatically generate new policy candidates. From the landing page, click the 'New Policy' button and enter a name different from the previous one, such as "Auto-generation of Municipality A".

The rest of the process will be the same as before, however please make sure to select 'On' for the Reconfiguration Mode in the settings.

  • Settings
    • Reconfiguration mode: Select 'On'.
    • For the remaining of the process, select the same file.

3.1.6 Displaying the results of the simulation for the automatically generated policies

Return to the landing page as you did previously. When the green 'SUCCESS' appears, confirming that the simulation has completed successfully, click 'Edit' and then click 'Evaluation' tab at the top of the page to check the results of the simulation.

As a result of the simulation, numerous candidates were automatically generated based on the policies in Municipalitys A, B, and C, and the effects were calculated from five perspectives as previous done. The generated candidates are automatically given names such as "New_policy_candidate0","New_policy_candidate1", and so on.

The results can also be used to make the following observations.

  • In the 'Evaluation results' section, click 'Number of Health Guidance' to sort in ascending order. When the policy in Municipality A is implemented, the number of people who can receive health guidance is very small. In contrast, the automatically generated candidates offer a wide range of options, and some candidates allow for increased health guidance. Here, considering realistic resource constraints, attention is given to candidates where the number of people receiving guidance falls between the values for Municipalitys A and C.
  • Next, use 'COMPARISON' to select Municipality A and multiple candidates of interest and compare the results. For example, if you look closely at one of the candidates, you can see that it is basically similar to Municipality A, but the conditions in the first half of the flowchart have been changed to those for municipality C.
  • In the 'Indicator scatter plot' section, if you set the Y axis to 'Number of Health Guidance' and the X axis to 'medical cost savings', as previous done, there is a roughly positive correlation between the amount of medical cost savings and the number of health guidance. Municipality A is in the lower left, Municipality C in the middle, and Municipality B in the upper right, and there are many policies that have been automatically generated to fill them in.

These considerations show that candidates reflecting the characteristics of each municipality are automatically generated, and that it is possible to consider better policies for Municipality A while taking into account real-world resources.