Amrish Kumar
Associate Lead Software Engineer at Cerner
Currently working as an Associate Lead Software Engineer in Rev Cycle Dev Group. I previously worked at Priceline.com Inc. where I had given numerous talks on machine learning. I also gave a talk in DevCon 2018 on machine learning. Regarding my academic credentials, I have completed my MS from the State University of New York At Buffalo with an emphasis on Robotics and Artificial Intelligence. In addition, I have completed an MBA from Welch College of Business. I have also published numerous papers in the IEEE Conference and the American Journal of Urology and given various talks on machine learning.
Applying Machine Learning to HIM Patient Deficiency Analysis
Health Information Management (HIM) is the practice of acquiring, analyzing, and protecting digital and traditional medical information vital to providing quality patient care. HIM Patient Deficiency Analysis is primarily used to manage the identification and completion of patient chart deficiencies. We are in the process of designing a machine learning system to automate the addition and completion of patient deficiency requests to improve the quality and timeliness of the deficiency process. One example of a patient deficiency request is document deficiencies.
There are two methods we developed for predicting document deficiencies:
1. Via configurable rules that were created based on common historical sets of documents.
2. A machine learning model, which would be able to assist the HIM Analyst by predicting and eventually auto assigning a list of document deficiencies in the system.
In this talk, we will walk through this real-life data science and machine learning example and explain association rules and One vs Rest Classifier machine learning techniques that were applied to Document References data for predicting document deficiencies using HealtheDataLab. This will add value to our clients through improved HIM Analyst productivity and efficiency. This will modernize our new deficiency analysis product, which is a SaaS offering built on Orion and Alva technologies.
Target Audience:
Engineers, Data Analysts, Data Scientists, Data Strategists, Data Architects - This talk is beneficial for anyone interested in learning about machine learning and its practical applications. The goal of this talk is to help technical attendees to get familiar with various machine learning techniques and start thinking about different ways how machine learning can be applied in their fields.
Predictive Modeling of Patient Deficiency Requests for HIM Deficiency Management
HIM Patient Deficiency Management is primarily used to manage the identification and completion of patient chart deficiencies. Currently, we are in the process of designing a machine learning framework to automate the addition and completion of patient deficiency requests. This will enhance our new deficiency management suite, which is our SaaS offering built on cloud technologies.
The primary goals of this framework are:
1. To provide recommendations and capabilities with the current product line for improved deficiency assignment via configurable rules.
2. To train a machine learning model, which would be able to assist the user by predicting and eventually auto assigning deficiencies in the system, only available in the SaaS solution.
We will walk through real-life data science and machine learning algorithms being applied to Document References Data from HealtheDataLab to make predictions and find patterns that add value to our solutions and our clients. In addition, we will demonstrate visualization techniques that were instrumental in computing relationships within Document References Data. We will conclude with the impact of these relationships on the healthcare industry.
We will dissect the plethora of various machine learning strategies such as decision tree, clustering, principal component analysis, recommendation, association rules, random forest, resampling strategies, etc. The audience will get a thorough understanding of data cleaning, data retrieval, and predictive modeling strategies in HealtheDataLab.
Target Audience:
Data Analysts, Data Scientists, Data Strategists, Data Architects - This talk is beneficial for anyone interested in learning about machine learning and its practical applications. Technical attendees will learn about various machine learning algorithms, tools, and techniques to make those concepts a reality. Applied correctly, artificial intelligence and machine learning will transform the healthcare industry and help achieve the quadruple aim.
Amrish Kumar
Associate Lead Software Engineer at Cerner