Ergo for personalized health care recommendations

A particular challenge for decision automation in treatment guidance is personalization.  While protocols are typically made for diseases or treatments in isolation, patients often present with multiple conditions and existing medications, each of which may present complications that contraindicate the standard protocol.  Rules or recommendations based only on the standard protocols will always fall short unless they allow for these contraindications and variations.

Ergo’s advanced knowledge representation and reasoning is capable of supporting decisions based on the full complexity of a patient’s particular characteristics and history. Ergo can answer such queries as ‘What treatment should be given – or not given – to a particular patient for a particular condition’.  It can automatically alert the care giver if they have chosen a treatment that is contraindicated.   Moreover, Ergo automatically provides an interactive explanation – the Why – for each answer, based on the logical chain of reasoning that was used to generate the answer.  Studies have shown that patients forget 80% of what a doctor advises them soon after leaving the doctor’s office.  Imagine having a tool that could provide both the guidance and the reasons that were used to generate the guidance.

Decision support for clinical guidance and insurance

The medical profession is challenged by the need to keep costs down while serving ever large numbers of patients in an increasingly short amount of time.  Federal and state mandates increasingly require the use of automation to aid various aspects of diagnosis and treatment, including to provide alerts and guidance for proper care.  Medical institutions and insurance providers have developed detailed protocols for treatment and management of diseases that must be followed.  In addition, proof of compliance with protocols is often required. In addition, detailed understanding of the treatment process is needed by the patients for them to comply with treatment guidelines and make informed decisions for care.

Example demo of personalized treatment guidance

In the scenario shown in the video to the right, a busy intern encounters an elderly woman in a rehabilitation facility complaining of knee pain.  The question is asked: What treatment should be given – or not given – to a particular patient for a particular condition?  The electronic health record (EHR) for the patient contains prior diagnoses and medication prescriptions, existing medical drug and disease diagnosis ontologies, and other structured data sources. These facts are used to answer basic queries, asked as logical sentences, and then answered (automatically) – accompanied by a logical explanation tree (generated automatically) to explain why the answer was given.

“Advanced Decision Analytics via Decision Analysis on Diverse Data: For Health Care and More” (30-minute presentation) by Benjamin Grosof and Janine Bloomfield, presented at DecisionCAMP 2016, July 8, 2016, Stony Brook, NY.

The demo of Ergo for health care treatment guidance begins at the 10:40 minute mark.