Doctors prescribe medications based off of guidelines and standards of care based on disease states. However, every patient has a unique pharmacodynamic profile. By integrating lab values and patient reported outcomes, pharmacists have an opportunity to work with patients along with other health professionals to bring personalized care to our patients.

What it does

Blossom analyzes EHR data along with patient reported outcomes to predict vitamin supplement and self care products that will improve the patient's overall health and disease management specific to them.

How we built it

We built this product using an EHR dataset from VCU analyzing 130 US hospitals , ADA guidelines, and clinical judgement. Python and Excel were used to read, import, and analyze the data and provide the recommendation. Tensorflow will be used after a patient uses Blossom for machine learning to measure and analyze the patient reported outcomes of the recommendations. This will be used to better our future recommendations for patient populations.

Challenges we ran into

We ran into problems finding data sets surrounding the lab values and symptoms we wanted to analyze along with the right computer software to analyze and compute the recommendations we felt best.

Accomplishments that we're proud of

Our team consisted of 4 females: two Pitt students, a Penn State student, and a Penn student. At the beginning, it was difficult to make sure we were all on the same page with ideas and what we were going to accomplish.

What we learned

As two pharmacy students, we learned a lot about the different aspects of coding and the whole process. Before this experience, we were very unaware of the different aspects of computer science and computer science engineering. Working with computer science, it was very nice to explain the different aspects of diabetes, the disease state we focused on, and what different lab values/symptoms could mean.

What's next for Blossom

Our next stop for Blossom would be expanding to more disease states along with improving the code on both the front end and back end to improve the patient interface.

Built With

Share this project: