Collaborative Projects Between Merck and Moravian University

OpenMRS

OpenMRS is an open source EMR used in more than 23 countries throughout the world, including 400+ installations in Nigeria. Providing efficient access to this data would allow researchers to more effectively answer important questions about population health. The goal of this project was to create an add-on module for OpenMRS to provide aggregate, anonymized data through a REST interface.

“The opportunity to meet industry professionals and work together on a project challenged me to apply my studies to the problem solving I would be faced with daily as a full-time employee.” – Alek Szilagyi

D3 Map Visualization

The A1c blood test provides a way to see average blood sugar levels of a patient over two to three months, a test doctors recommend be administered to diabetes patients four times per year. The goal of this project was to determine whether compliance to this recommendation could be seen in Medicare billing data. The project focused on visualizing this dataset simultaneously with other datasets on an interactive web-based map.

“I was able to apply my education to practical use cases that provided me with new appreciation for my coursework.” – Myles Barros

FHIR

The goal of interoperability is for patients to be able to seamlessly share their medical records – even if their doctors are using different EMR systems. Fast Healthcare Interoperability Resources (FHIR) is a standard created by Health Level 7 that describes data formats to achieve this goal. In this project, students created a program to translate Medicare billing data into the FHIR format.

“This opportunity provided me invaluable experience regarding the difference between software development in an academic setting and development of software that goes into production.” – Jason Boccuti

Hydra / Linea

Joining multiple datasets together can result in signi cant insights, but for this to be possible, researchers must know what datasets are available. Hydra provides Merck employees with an easy-to-use tool to catalog and search datasets, and a version of the tool (called Linea) was given to DocGraph prior to completion. The goal of this project was to incorporate features of the nal version of Hydra into Linea and to add features required by the DocGraph community.

“Prior to working on these projects at Merck, I had never really considered how much opportunity existed for computer scientists in the healthcare eld.” – Anna Lamoureux

k-Means on Medicare Billing Data

The datasets available from the Centers for Medicare and Medicaid Services (CMS) offer researchers a wealth of information. The goal of this project was to explore the accuracy of doctor classi cation in the Medicare billing datasets. Using the k-means algorithm to cluster by billing patterns, doctors who fell outside the most prevalent groups were agged as potentially mis-classi ed.

“This project showed me how computer science and data mining can be used to improve datasets and assist in a eld that is not always thought to be linked to computing technology.” – Bill Collins

Referral Patterns for ACOs

The Affordable Care Act (ACA) introduced the concept of an Accountable Care Organization (ACO), and it is believed that doctors in ACOs will have lower referral rates. The goal of this project was to determine whether changes to referral rates could be seen in the six years of Medicare billing data available from CMS.

“One of the benefits of my experience at Merck was learning how to communicate with other professionals— figuring out what level of information a person needs and then how to deliver that information clearly” – Alec Gerhart

Home Health Kit

Medical professionals make decisions based on available data, and
an increased frequency of measurement enables physicians to provide more informed diagnoses. The goal of this project was to develop an application for patients to collect blood pressure and blood oxygen data using low-cost sensors connected to a Raspberry Pi. The data was automatically uploaded to a server via a REST interface, and doctors could access the data to monitor their patients.

“Classroom projects follow an organized, predictable path, but in the workplace things change, and you have to adjust, go with the flow.” – Nick Zambelli

AWS Automation

There is a demand to manage and provide solutions for Merck’s internal framework, research, and outfacing applications in the cloud. Using multiple virtual private cloud (VPC) local networks, developers can embrace tools such as Amazon Web Services (AWS) and other cloud alternatives to meet their computing needs. The goal of this co-op project was to develop applications that automated or con gured AWS resources while ensuring compliance to Merck policies.

“The experience forced me to work with people with different backgrounds and to think about projects at a larger scale than anything I’d done at school.” – Josh Russett

Health IT Platform

Data scientists are often hindered by low-level data manipulation tasks. The Health IT Platform provides cutting-edge tools to automate ingestion and cleansing of data into a data lake, as well as consistent REST APIs to access the data. In this co-op, students created an SDK to provide data ltering, aggregation, and ingestion.

“My Merck experience showed me that there is a demand for learning in the eld—for picking up new technologies on the y to work on new and exciting projects.” – Charlie McDonald

Blockchain

Trust for digital documents is dif cult to establish because documents can be modi ed without detection. The Blockchain technology allows users to add a permanent record of a transaction, which can be used
to store the hash of a signed document. The goal of this co-op project was to create an API that allows users to upload a document hash to the blockchain and to see the path of a document through the company.

“I’ve become much better at researching ideas and concepts, and I can now switch between programming languages and get up to speed with them quickly.” – Michael Turnbach

Cloud Foundry

The infrastructure for a scalable app can require signi cant setup and con guration, and Cloud Foundry aims to simplify these details. The goal of this co-op project was to con gure the Cloud Foundry at Merck to provide monitoring and logging tools as well as to support LDAP authentication.

“We had goals that we had to achieve, but we also had opportunities to explore—to learn a new technology and try it out on a specific problem.” – John Vonelli

IoT for Alzheimer’s

The widespread usage of smartphones, tness trackers, and home automation provides an opportunity to improve patient care. The goal of this project was to explore how the Internet of Things (IoT) could
be used to improve Alzheimer’s care by collecting data on the patient, providing reminders, and allowing caregivers to remotely determine the wellness of the patient.

“At Merck, I was exposed to technologies that I don’t work with in the classroom. I developed leadership skills, and I learned how to interact with professionals in the workplace.” – Megan Biernat