AI Enabled Blood Test Report Analysis
Use Case Description
Blood Test Report Analysis
The global blood testing market size is over $10 billion with an annual growth rate of 8%. There are no true estimates on how many blood tests are performed each year, but based on various sources, we can estimate it to be over 1 billion tests a year.
There are different types of blood tests, but there are 5 types of blood tests that account for >70% of all tests done. These are: Glucose testing, A1C testing, Direct LDL testing, Lipid panel testing, and Prostate Antigen testing.
For a specific type of test, the contents of the test reports are similar, although they are reported in different formats depending on the lab. The generation of reports is largely automated now. However, the lab clinician needs to interpret the data and include it as their observation. The doctors often rely on the observations, unless certain parameters are off, then they would review in detail.
The proposed use case focuses on providing prescriptive observations for the lab clinicians and doctors, which they can use as an additional input in their analysis.
Part 1: Extraction
The solution will be available as both an unattended or attended automation. Large labs can run it as unattended automation, while smaller labs or doctors can run it on an ad-hoc basis as an attended automation. RPA robot will fetch the lab report and use UiPath Document Understanding to classify the lab report for the type of test and then extract the required fields from the report. The robot will save the data in a global database without any PII information.
Part 2: Analysis
The robot will send the report parameters to the ML models hosted in the UiPath AI Center. There will be 1 ML model per type of test. There are both open-sources and licensed ML models available for these cases today, which will be imported into AI Center. Over time, we can build our own models as well, once we have sufficient data available in our global database.
Part 3: Recommendation
In the case of the unattended automation, the robot will attach the recommendations as an additional page in the test report PDF and send it to the lab clinician. In the case of the attended automation, we will build a simple app using UiPath Apps for the lab clinicians and doctors.
Part 4: We will also be able to highlight to the lab clinician on similar cases that have been found in the database for the specific hospital or laboratory. The reason for not doing a global search is because we strip PII data and only have a unique ID, which is specific to the hospital or laboratory.
Part 5: We will also review past blood test reports of the same patient and provide additional recommendations on changes observed over time.
- Productivity Savings: Lab clinicians will be able to do the report assessments faster.
- Reduce Risk of Human Errors: A lab clinicians may have to clear 100 reports in a day and may overlook unique aspects in a test report. Through our solution, these will be provided to the lab clinicians as high alert for them to do a closer review.
- Improve Speed of Test Results Availability: In most countries, the healthcare professionals have a high load and it often takes longer to provide the reports. With this solution, we expect that lab clinicians will be able to process more reports in a day and provide the results faster.
AS-IS WORKFLOW, TO-BE WORKFLOW
Other information about the use case
Industry categories for this use case: Healthcare Pharma
Skill level required: Advanced
UiPath Products that were used: UiPath Studio, UiPath Action Center, UiPath AI Center, UiPath Data Services, UiPath Document Understanding
Other applications that were used: SQL, Excel, PDF, Health Care Applications
Other resources: -
What is the top ROI driver for this use case?: Accelerate growth and operational efficiency