The team clinician uses the software on a portable computing device to record what the patient relates to them during their clinical assessment. They use it the way they might historically have used a clip board to write things down. This can be about a new problem, an existing one, or a general discussion on prevention or wellness for example. The images on the left show a few of the ways that the tool allows the clinician to rapidly collect the results of the interview while it's underway using a highly graphical interface. This information is stored as a medical record and serves as part of the clinical documentation of the patient encounter. This interface can be used to collect demographic data, reconcile medications and review the patient's past medical record and other tasks typically associated with an office visit.


Dynamic medical analytics:

This is the part that is truly unique. While the clinician is recording the patient's responses to their questioning, behind the scenes the core of the software, the Adaptive Knowledge Engine (AKE), is independently analyzing the data as it comes in. This data is analyzed against an ever-expanding library of medical knowledge which uniquely considers the differential diagnosis associated with the presentation, the qualitative pattern match of conditions that might fit the circumstances and the dynamic prevalence of these conditions in the specific population being interviewed including the presence of co-morbid conditions. And not just the new data. The AKE also considers every piece of data that has ever been recorded about the patient and stored in its record. That is a groundbreaking advance. For the first time, the complete medical record is used in assessing every situation relevant to the patient's well-being. The clinical implications to individually tailored best practices are profound.


Interaction with the clinician:

As it analyses the data stream, the AKE returns a report to its user on the screen next to where the interview is being recorded. This report contains a list of the differential diagnoses that fits the scenario, a measure of the strength of the pattern match of each condition with respect to what has been recorded during the interview or inferred from the medical record, the prevalence of the condition in the population being assessed, and a transparent accounting as to why it is reporting the condition to the user at this time. It also uniquely dynamically generates a list of questions, observations, or tests that would help to sort out the various possibilities. These questions are key to how the system improves safety. By reminding and prompting the user to ask the relevant questions (instead of a canned set of general questions), a much more complete and tailored assessment is performed and recorded into the record.


Stateless design:

A huge problem plaguing human decision-making is confirmation bias. Once we head down a path and begin to formulate a theory, evidence supporting that theory is sought out and data which contradicts it is often ignored. The AKE eliminates that bias by being 'stateless'- each time the program runs, it behaves as if this is the first time it has ever seen the data submitted to it. Previous considerations are therefore not relevant to its new assessment. Each instance is handled independently allowing the AKE to immediately shift gears as needed in response to new information. As the clinician records the patient's answers to the new questions, the AKE reassesses the data and revises both its observations as well as its prompts. This provides a very dynamic interaction between, patient, clinician, and computer. This process continues until a reasonable end point is achieved. This end point can be a decision to treat, a follow-up visit, or termination of the interview and referral to another level of clinician. These decisions are predicated on the skill level of the using clinician and their scope of practice. The AKE alerts a non-MD to stop the interview and refer on to another practice level though the use of alerts and prompts. Future versions will automatically alert the higher level clinician to expect the referral. These two functions of analysis and interaction are the key methods used by the system to safely elevate the clinical capabilities of non-MD's to provide expanded services.


Documentation:

While all this is going on, the data is being stored in a compliant medical record format that is designed to support fluid data mining, process improvement and medical research. Key business practices such as automated billing and coding can also be included. This record facilitates communication amongst care providers and is intended to more readily support hand offs from one care level to another without the need for redundancy in assessments (an element to be studied during the phase III trial). A goal is to reduce the need for documentation other than at the bedside as a means to improve throughput and clinician compliance.


Summary:

The AKE or diagnostic engine considers the patient's condition, symptoms and past medical history as well as the level of the healthcare worker in both the questions that it poses and its recommendations. Conservative escalation recommendations, which insure that every patient receives the proper level of care, take into consideration the severity and acuity of all possible diagnoses as well as the skill level of the provider in making the clinical determinations needed to refine the diagnosis. Patient safety and efficacy are prime considerations. The software tools escalate a patient to a higher level practitioner through the presentation of triage alerts to its users.