The client is the single largest financial and administrative healthcare network in the United States, reaching approximately 750,000 physicians, 105,000 dentists, 60,000 pharmacies, 5,000 hospitals, 600 vendors, 450 laboratories and 1,200 government and commercial payers. Client has developed this network of payers and providers over 30 years in the industry, connecting virtually all private and government payers, claim submitting providers and pharmacies in a hybrid cloud based, user-centric and secure infrastructure environment. In 2013, client was included as part of the Information Week 500 list of Top Technology Innovators in the United States.
The business operations team uses Chart Splitters to Split the Member Chart to multiple Encounter files. The Encounter files split are used by the Natural Language Processing (NLP) system to extract the International Classification of Disease (ICD) codes having Hierarchical Condition Category (HCC). Chart splitting is not required in cases where the HCC is not present in the chart. The client needed an ability to avoid the Charts being sent for Splitting where there was no or low HCC.
Partnered with the internal AI Data Science team to identify HCCs in a Chart. Integration solution with AI Data Science where the system can accept the Chart and identify if the chart has high probability of HCCs or not. Charts with High HCCs are assigned for internal Splitting and extraction is done using the NLP system. Charts with low or No HCCs are manually entered in system thus avoiding the need for splitting or sending it to NLP with less successful results.