Mission and History


The mission of Healthcare Inference Systems is to improve the quality and delivery of healthcare by providing advanced natural language technology to implement fundamental change in the provider reimbursement process, to reduce costs, and to enable its principal customers to capture and search clinical information as natural language.


Healthcare Inference Systems (HIS) is a West Virginia limited liability company that was formed in 2002 to assist in the funding, development and marketing of a breakthrough natural language inference technology to address deficiencies in the health care coding and billing process and to facilitate more widespread analysis of medical data. Through the efforts and support of HIS, a private software developer has completed the development of an advanced inference engine and artificial intelligence internet-based software solution to capture and access patient history and treatment detail by applying natural language processing to medical terminology. Currently known under the trade name InferRx, this Internet-based application is the only product in existence that can capture and analyze clinical data from natural language in conformity with the World Health Organization's International Classification of Disease (ICD-9) medical data coding standard that is used in the United States and 37 other countries. It is also the only such application that can be augmented rapidly and efficiently for use with ICD-10 and ICD-11.

Since 2003, using independent coders, HIS has used the InferRx application to analyze thousands of patient records from hospitals ranging from major referral centers to small critical access hospitals. The results have demonstrated a consistent hospital coding error rate of 10 to 15 percent, even in hospitals that pre-screened their coding records for accuracy before submitting the patient records to HIS. Inaccuracies in coding in the records reviewed caused a projected average decrease in annual hospital revenue of approximately three percent. The automated review of records by HIS also quickly identified systematic up-coding in hospitals that were attempting to generate revenue using ICD-9 codes that were not supported by the patient records.

In 2007 the VA provided the company with the electronic records of 150 patients from one of its "Centers of Excellence" hospitals that the VA perceived to have the most accurate and efficient coding. Using the HIS natural language system, a single independent coder re-coded the 150 VA records in approximately 16 hours, compared to the VA average of fourteen charts per coder per day, and found an error rate of sixteen percent and lost revenue of approximately ten percent. As was the case with all hospitals, the majority of inaccuracies in the coding of VA hospital records were the result of the failure of coders to recognize and include complications and co-morbidities. This oversight was also found to exist in disproportionate numbers in the private sector in tertiary care centers where the most seriously ill and complex patients are treated.

The company recently completed testing of InferRx to incorporate automated ICD-10 recognition and verification. The addition of ICD-10 to InferRx will enable hospitals and health care providers to reduce dramatically the costs of conversion to ICD-10 based billing both in the short and the long term. The upgraded version of the InferRx system will identify simultaneously all ICD-9CM and ICD-10 codes that are supported by the text of a patient’s electronic record. Manual coding in ICD-10 will be time-consuming and likely will require additional personnel and increase coding costs significantly. The use of natural language processing such as the InferRx system will eliminate the need for additional coders, will improve accuracy, and will minimize ICD-10 conversion and training costs.”