Mid-march 2020 we delivered AI-powered Covid-19 pre-diagnostic tool
4 weeks later we had 1000s API calls / day from all around the world.
Technology behind EndlessMedical allowed us to create synthetic, individual-patient level data for the COVID-19 cohort, which was used to model AI/ML. API can be used now to triage patients with flu-like and respiratory symptoms, fever, "COVID toes" etc. Upon request we will make it free to use and available to you until 2021.
EndlessMedical API supports over 500 features
Unlike the competition, EndlessMedical API supports symptoms, signs, results of blood work, imaging, and physical examination findings - just to mention few categories of features.
You can use the API to build true medical applications, beyond just "symptom checkers". Not all already defined features (see API dictionaries) are yet active.
EndlessMedical API doesn’t allow to input any patients identifying data, like date of birth, or name
There are significantly fewer concerns about patients’ privacy.
We only process de-identified numbers and labels’ identifying categories for categorical features.
EndlessMedical API, at any point, can suggest the next "best step” on diagnostic pathway
It may be next physical examination maneuver, next question to the patient, next blood test or imaging study to be done, which will most efficiently narrow down the list of likely diagnoses, recommended tests, and additional recommendations.
EndlessMedical API will produce explanation on why this "next step” is recommended
Use this feature to optimize the time of care by guiding providers through interviews, physical examination sessions, workup, and treatment processes in the most efficient way.
Avoid unwanted adverse events from unnecessary testing and save resources spent on unnecessary testing.
Diagnoses are flagged by, including “life-threatening emergency”, “high risk”, or other flags
Use this feature to triage patients and suggest prepare billing and coding suggestions consistent with E/M guidelines.
Avoid being misled by big-data
How can we trust AI/ML modeling built on big-data when the research and experience repeatedly show that electronic medical record data is not trustworthy?
Predicted list of differential diagnoses/diseases are further subcategorized into medical areas and specialties
Use this to refer patients to a specialist specific to patient's signs and symptoms.
EndlessMedical API results is regularly verified
Founder, triple boarded practicing physician Lukasz Kiljanek MD, spent days transposing his clinical experience, and reviewing the literature on adding the diseases, tests, recommendations.
Documentation generation happens in the background real-time and on-the-go, as your users use EndlessMedical API
Generate documentation of the entire patient encounters.
This soon will cover encounter documentation (provider's notes), requisitions for recommended tests, and patient instructions.
Documentation generated by EndlessMedical API is compatible in structure to the current United States documentation standards
Soon, notes generated by API will have a chief complaint, history of present illness, review of systems, past medical, social and family histories, physical exmaination, imaging and laboratory studies, and impression with differential diagnosis and plans.
EndlessMedical API generate synthetic, individual patient-level data, and use for AI/ML modeling.
Endless Medical API uses patent-pending technologies (USPTO # 20200118691, PCT/US2019/055747) to generate a knowledge database containing synthetic, individual patient- level data based on research, clinical experience and literature.
Given recent concerns about transparency and data integrity (with 2 recent publications in Lancet and NEJM) with big-data, we encourage everybody to review the patents information to make sure they understand the advantages and disadvantages of EndlessMedical API technologies.
EndlessMedical API uses synthetic, individual patient-level data
EndlessMedical API does not use "big-data". Big data is inaccurate, has errors, typos, and biases, brings security, privacy and data ownership concerns.
We are using synthetic patient-level data for AI/ML modeling to overcome some of the big-data issues.
Big-data under-represents rare diseases and tends to over-represent billable diagnoses and pertinent positive findings in physical examination and history, ignoring other normal and negative findings.
Big-data acquisition for research or commercial use requires costly and time-consuming IRB approvals.
Synthethic patient-level data is not perfect either, but maybe an alternative to big-data for some usages. For example, Endless Medical API uses these data for AI/ML modeling.
Endless Medical API technologies, allow creating synthetic databases echoing the true prevalence of diseases and clinical findings in a given environment.
For example, the incidence of strep throat in general practice office is different than the incidence of strep throat in nursing homes. Historically, AI and ML models trained on data from one environment can not be effectively used in another. EndlessMedical technologies would allow us to transpose the data from one environment to another.
The EndlessMedical API's knowledge database, currently is created only by Lukasz Kiljanek MD and his review of literature when deemed necessary. It can, on-demand, be transposed, to a different clinical setting, like an intensive care unit or an emergency department.
In the future, data sourced from multiple experts or data sources (i.e., research papers) will be merged to increase the robustness of the medical knowledge database and thereof sensitivity and specificity of AI/ML models.
PRE-DIAGNOSE PATIENTS TRIAGE TO SPECIALTY CARE OPTIMIZE WORKUP TRIAGE BY ACUITY