Custom Curriculum Development – Case Study

When you need a customized curriculum, you want a partnership to make the process easy and one that won’t cost and arm and a leg.  You know that this course development will be key to helping your employees reach their intended learning goals.  You also know that customizing the material will help you save money and will not waste anyone’s valuable time. That’s where Lumious can help.  Our talented SMEs (Subject Matter Experts) and Instructional Designers are at the ready to complete your most unique requests.

In recent work with a client, Lumious quickly and innovatively solved our client’s unique business challenge. Below outlines the specific project that Lumious recently completed for our client, a large health care company.

Business Challenge: A large, health care organization needed to train approximately 2,000 personnel to become data analysts, developers or data scientists.  Having developers and analysts on-staff would save the client significant time and money.

Solution: Develop 50 hours of ILT (Instructor Led Training) with Big Data content that modelized, enabling the client to deliver what they need to and when they need to.  Additionally, they can pull only the topics for a specific audience and skill set.

In order to develop the most customized course material for the client,  below are the considerations that were taken into account while developing the course:

    • Volume: Healthcare generates a very large volume of data, which creates problems in its storage and management.  Having scalable and abundant amount of storage is key.
    • Velocity: Not only is the data generated in large quantities, but it is also generated quickly. More over, it often needs to be acted upon in a timely fashion.
    • Veracity: Much of the data in healthcare is redundant, incomplete, or misleading; which can make it difficult to know what to trust.
    • Variability: There is a large degree of variation between different sources of healthcare data. Some sources, such as in-patient clinical notes, may be comprehensive but have a high degree of duplication or redundancy. Others, such as discharge summaries, are more concise but lack important context. The quality of data, even for the same patient, may be extremely variable.
    • Complexity: Given its volume and variability, working with healthcare data is complex. Analyzing it to gain insights often involves complex algorithms, logic, and specialized tooling or techniques.

A significant portion of the course development involved specific use cases:

Deep Learning. Chest x-rays are used to identify diseases and abnormalities in the thorax.  In recent years, machine learning technologies have emerged which allow for disease detection to be done automatically and then verified by a trained radiologist. Using machines to read X-rays allows for the whole process to be streamlined and for patients with severe problems to be seen first. Sometimes, the abnormalities are so small in size that a human eye becomes unable to detect it. With doctors not being able to detect such an abnormality, it could cost a patient’s life. AI helps in detecting such abnormalities which the human eye cannot detect.

Training exercise- Use Kafka and deep learning techniques to train a model on a subset of the data and then assess the model’s accuracy against a larger group of 100,000 chest x-rays.

Data Acquisition and Enrichment: Healthcare data management requires systems which are able to ingest different types of information from multiple sources; aggregate and enrich it in meaningful ways; resolve duplicates, copy with missing data; and prevent false alarms which might arise from misleading information.

Training exercise- Using large clinical and claims datasets, students will learn how to aggregate data across tables, re-code variables, work with missing values, and prepare summaries of the data as we prepare for its use in a machine learning.

Healthcare System Analysis:  Using AI and Data Science tools, a healthcare organization can analyze treatments from a database of patients and identify inefficiencies in treatment and unnecessary hospitalization.

Cyber Security: Hospitals and clinics store confidential information. This information needs protection from hackers and cyber attacks. Even the smallest data leak can cause significant harm to both patients, as well as, healthcare companies/professionals. It’s not possible for the hospitals’ cybersecurity teams to figure out every potential threat to their systems. AI can help cybersecurity teams to figure out every potential issue or risk. The algorithm can also rank them based on their priority and present it to the team.

Impact: The health care organization is assigning about 20 use cases per year to their newly trained data analysts and will save about $1-2 million dollars per use case.