◷ Reading Time: 5 minutes
Introduction
The sample project will show how to build a project that detects incompatible medical diagnosis codes in medical claims.
Scenario
In this example, we used a dataset containing around 70,000 diagnosis code pairs that cannot claim together. This decision service gets a set of diagnosis codes as inputs and then determines the incompatible diagnosis code combinations. Once it detects, as an output it generates error messages as “Diagnosis {code1} cannot reported together with {code2}“.
Running the Sample
- Open the file ClaimProcessing.xml
- Click on the Logic Run template.
- Click on any given template.
- Debug the sample by clicking on the debug button in the Logic Run Template window.
- You can click the ‘Next Step’ button to go step by step to follow the flow.
- Now you can see the output values in the ‘Parameters’ window.
Process Steps
The process contains the following steps to determine “Incompatibility Codes”
- Read the dataset, build a lookup table using the dataset and save the lookup table to cache memory when the project runs for first time.
- Search through the lookup table and create error messages for incompatible diagnosis code.
Project Description
The project contains the following documents.
Generic Flow
- ClaimProcessing.xml – Shows the main flow
- BuildLookupTable.xml – Reads the data, build the lookup table and set it to the cache memory.
Decision Table
- IncompatibilityCheck.xml – Determines the incompatible codes and generate an error messages
Boxed Expression
- CombinationChecker.xml – Search the code pair in the lookup table
The Flow Design
1. Create look up table and set it into the cache memory for the first time. This flow execute only if the cache memory doesn’t have the lookup table.
- Incompatible Code List – This is the data file(Excel). It reads the dataset(Excel) and save it to IncompatibleCodeList variable.
- Create Lookup Table – It creates a lookup table by indexing the dataset to increase the search efficiency.
- Set Cache – In this expression, the lookup table set into cache reference.
2. Create a Decision Graph to detect incompatible code pairs
As you can see in the below model, all the previous processes are in a subgraph called Build Lookup Table to make the process far easier to understand.
In this Build Lookup Table node, set a skip condition in the properties window to skip if the lookup table is exist in the cache memory.
3. IncompatibilityCheck – In this decision table, inputs diagnosis codes are checked against the dataset for incompatibility using the defined boxed expression.
4 CombinationChecker.xml (BoxedExression) – In this boxed expression, it reads the lookup table in the cache memory, search incompatible code pairs, and returns a Boolean value based on the search. This boxed expression is linked to the Decision Table( IncompatibilityCheck .xml).
Result
Once you run/debug this model, results will appears in the Notifications window under Notifications tab as an Error messages.
Download the project
Use the attachment at the end of the page to download the sample project.