SOWMYA ELURI

Health Informatics graduate student and Clinical Pharmacy professional passionate about leveraging the intersection of healthcare and technology to advance patient-centered care. Adept in managing and analyzing healthcare data to drive innovation and efficiency. Seeking a role that values technical expertise and offers hands-on experience in a dynamic healthcare environment.

Title: Pneumonia Detection Using Deep Learning on Chest X-ray Images 

 Description: This project focused on developing a deep learning model for the detection of pneumonia using chest X-ray images, leveraging ResNet-18, a pre-trained convolutional neural network (CNN) from the PyTorch library. The objective was to build a reliable diagnostic tool to assist healthcare professionals by automating pneumonia classification, thus improving the efficiency of patient care.  

Technologies: PyTorch, ResNet-18, Data Augmentation, Image Classification, Healthcare Analytics  

Key Achievements: Trained and evaluated a ResNet-18 model on chest X-ray datasets to classify images as NORMAL or  PNEUMONIA. Implemented data augmentation techniques, such as random cropping and horizontal flipping, to increase model robustness. Achieved a test accuracy of 87.82%, validating the model's effectiveness in real-world settings. Computed key performance metrics: Precision: 0.83, Recall: 0.99, Visualized confusion matrices and count plots to gain insights into classification performance and data distribution.  

Results: Demonstrated the model's reliability with high recall, ensuring that most pneumonia cases were detected correctly.  Highlighted a few false positives for pneumonia, identifying areas for further model optimization. Validated the model's performance on unseen test data, achieving robust predictions across multiple datasets. Created an interactive process to classify individual X-ray images in real-time, supporting clinical workflows. 

Title: Analysis of Florida Hospital Memorial Medical Center Readmissions Data.  

Description: Spearheaded a comprehensive analysis of readmission data at Florida Hospital Memorial Medical Center,  employing Tableau for in-depth data visualization and insights generation. The focus was on enhancing patient care quality,  comparing hospital performance within the state and nationwide. Technologies: Tableau, Data Visualization, Data Analysis, Healthcare Analytics. 

Key Achievements: Analyzed the excess readmission ratio for conditions like Acute Myocardial Infarction (AMI), Chronic  Obstructive Pulmonary Disease (COPD), Heart Failure (HF), Hip Knee, and Pneumonia (PN) across the nation and the local area., Identified areas for improvement in patient care, particularly emphasizing COPD readmissions., Utilized Tableau's visualizations, including bar charts, maps, and scatter plots, to represent key metrics like excess readmission ratios and timeliness of care.  

Results: Discovered that Florida Hospital Memorial Medical Center maintained readmission ratios below 1, indicating efficient  patient care and performance., Recognized a high-risk scenario for COPD readmissions, with an average Excess Readmission  Ratio (ERR) of 0.98, necessitating specific interventions for improvement.

• Licensed Clinical Pharmacist, India.