Diagnosis Engine

The Health Navigator Diagnosis Engine generates a list of possible causes (pre-diagnoses) for a specific symptom or problem. This can be used in digital health assistants (health bots, diagnosis symptom checkers) and electronic health records.

The Health Navigator Diagnosis Engine is a proprietary knowledge-based inference engine. It accepts as inputs: age, gender, pregnancy status, symptoms, known medical problems, and other clinical factors. It uses Bayesian logic to output a list of causes sorted by likelihood. This expert system is a structured and evolving database of curated content. It is the result of nearly two decades of subject matter expert input, literature review, data mining and analytics, and testing against large clinical data sets.

Here is an example use case scenario. This example illustrates how the diagnosis engine might be used in a digital health assistant:

Maria Williams is 44 years old, female, and not pregnant. She is having constant right upper stomach pain with nausea that started 8 hours earlier. The pain is moderate and did not get better after taking Tylenol and an antacid. She has no chest pain, fever, or diarrhea. She has no medical problems and does not drink alcohol.

Maria is most interested in knowing what are the most likely causes of her pain. She enters her information into the symptom checker app on her computer. The app lists the most probable causes sorted by likelihood and also provides guidance on how serious these conditions are.

Maria learns that gallstones is the most likely cause. The app provides Maria an easy-to-read definition of gallstones, the common symptoms, and some possible complications. Further, the app gives some guidance on the seriousness of gallstones, the type of doctors who treat it, and a list of internet resources where she can learn more.

Diagnosis Engine Demo Application - Abdomen Pain Example

Diagnosis Engine Demo Application - Gallstones


How it works…

[su_spacer size=”10″]


The Health Navigator Diagnosis Engine is provided for use by our partners through an easy to implement application programming interface (API).

The graphic below shows a high-level workflow of a consumer diagnosis symptom checker or health bot.



  1. What’s Wrong? The first step is find out what is the user’s main symptom or problem. This is also called the chief complaint. This is best done using the Health Navigator Natural Language Processing (NLP) engine. For example, the user might enter “I have had tummy pain and nausea since early this morning.” A single call to the FindCCCbyPopulation method and the NLP technology converts this free-text into acuity-coded chief complaints (Abdomen Pain, Nausea). These clinical concepts and their associated concept IDs are used as data in the remainder of the process.
  2. Tell Us More. The second step is to ask the user to “tell us more” about their primary and secondary complaint(s). This is easily achieved by calling three Clinical Documentation Support (CDS) methods using the clinical concept IDs from the first step:
    • TellUsMorePrimaryCCC
    • TellUsMorePrimaryCCC_OPQRST
    • TellUsMoreSecondaryCCC_OPQRST_Severity
  3. List Possible Causes. The third and final step is to generate the list of possible causes. This is accomplished by a single call to the Diagnosis Engine method GetCauses. The input parameters for the GetCauses method include: primary chief complaint (concept ID), age, gender, pregnancy status, LanguageID, and all other clinical findings (concept ID; present or absent).

The Diagnosis Engine is available in eight languages. The Diagnosis Engine can be used along with the Triage Engine develop a combined triage and diagnosis symptom checker app or health bot.


How it improves care…

[su_spacer size=”10″]


The Health Navigator Diagnosis Engine provides a tool for our partners to use to educate and empower patients in the diagnostic process. The Institute of Medicine, in their report Improving Diagnosis in Healthcare, has stated that “Patients and their family members are essential members of the diagnostic team.” Furthermore, patients should not hesitate to ask their doctor, physician assistant, or nurse practitioner:

  • What else could it be?
  • Do all my symptoms match your diagnosis?
  • Could there be more than one thing going on?

Here are two examples of how the Health Navigator Diagnosis Engine could potentially improve patient care:

  • Artificial Intelligence, Digital Health Assistants, and Health Bots: Millions of people search the internet every day for answers to health-related questions. The Diagnosis Engine can help you deliver to your patients more accurate and reliable information about possible causes for their symptoms. Partners like Microsoft believe that the Health Navigator platform “… will help users easily receive accurate hands-on information to help them make healthcare decisions”.
  • Electronic Health Records: The Health Navigator Diagnosis Engine can provide decision support to clinicians and help broaden their differential diagnosis.


An integrated clinical vocabulary and database architecture…

[su_spacer size=”10″]


Underlying the Health Navigator platform is an integrated clinical vocabulary and database architecture. There are over 470 distinct Coded Chief Complaints (reasons for visit) and approximately 2,700 different diagnoses (conditions, diseases, injuries).

Diagnosis Engine - An Integrated Clinical Vocabulary and Database