Health Analytics - Clustrex Data

Case Study

1. Healthcare Analytics for Practices

Full AWS architecture and setup provisioning EC2, EBS, ECS/Fargate, Route 53, S3+Glacier, Athena, IAM, MFA, Secrets Manager, VPC, ALB, Workspace, Lambda, API Gateway, SFTP, SES, Systems Manager, Cloudwatch, SecurityHub, OpenVPN, KMS, Certificate Manager.
Large scale ETL ( Extract, Transform, Load ) of data across several US based healthcare practices using Talend / Postgres, across different data formats : CCDA/HL7, CSV. Full-scale migration of Data and Application from another cloud provider.
Processing daily patient data for over 25+ US based healthcare practices Flow/Stream based processing using Apache Nifi, Drill and Talend ETL for cross-format data, MirthConnect, AWS Lambda, SQS, AWS RDS, Cloudwatch. Automation scripts in Unix Shell scripts, Python. Data Analysis and visualization using Tableau.
Mirth Connect / NextGen Connect Healthcare data processing and Building Applications on top of the data layer.
Value delivered:
Technology Expertise, Cost Savings, Timezone support, Teamwork, HIPAA certified.
Challenges:
Variety of Data Interoperability with many systems Strict Regulatory Compliance Domain Knowledge depth.

2. Patient Mobile App for Clinics

Clustrex built a mobile app that enables patients to book, reschedule, cancel appointments with their clinic, view their medical records for each visit, update their profile information, view the payment balances and request their medication refills.
Specialities:
Ophthalmology, Pain Management, and more.
On the backend, the clinic staff can handle patient requests through a well defined workflow. All the patient records are made available integrating with the practice EMR and Practice management systems.
Mobile App is built with React Native targeted for Apple and Android platforms. Web backend is based on Python, AWS Lambda, RDS Postgresql. Data lakes is built on a HIPAA compliant data infrastructure.
Value delivered:
Improved patient experience, self-services workflows, and streamlining practice operations.

3. Clinical Trial Patient Database

Clustrex built a web application for an US based client, creating a database of patient prospects with their demographics information. Based on any specified filter criteria, select a list of prospects for a particular campaign/study. Bulk import of prospect information is enabled. Agencies can login and provide new patient prospects for Clinical Trial study.
Technologies used:
React JS, Python, AWS Lambda, and Postgresql.
Value Delivered:
Ability to quickly zero-in on a prospect list for Clinical Trials.

4. Pharma Analytics: Drug Switch Benchmarking

Introduction:
In the dynamic field of pharmaceutical analytics, understanding the patterns and impacts of drug switches is crucial for optimizing treatment strategies and improving patient outcomes. This case study focuses on developing and applying the Drug Switch Benchmark Dashboard, a data-driven tool designed to monitor and analyze drug switch trends over time.
Objective:
The primary goal of the dashboard is to track drug switches within a rolling 12-month window, ensuring that the analysis reflects the most recent trends in medication usage. Additionally, the dashboard incorporates data from the past 24 months to accurately identify new drug starts, providing a comprehensive view of both new prescriptions and switch behaviors.
Methodology:
- 1. Data Collection:
  - The dashboard aggregates data from electronic health records (EHRs) from multiple vendors.
  - A 24-month data window is utilized to capture new drug initiations and historical medication patterns.
- 2. Data Processing:
  - Patients are identified based on their prescription history, focusing on transitions from a previous drug to a current drug.
  - The analysis filters out incomplete data.
  - Patients with no treatment in the past 12 months are considered new starts.
- 3. Dashboard Design:
  - Interactive visualizations display switch rates, trends, and patient demographics.
  - Metrics include switch frequency, time to switch, and switch outcomes (e.g., efficacy, adherence).
Results:
The dashboard revealed key insights into drug switch behavior:
- Switch Rates: A noticeable trend of increasing switches in certain therapeutic areas, suggesting shifts in clinical guidelines or patient preferences.
- New Starts: The 24-month data window successfully identified new drug initiations, providing a baseline for comparing switch patterns.
- Impact Analysis: Correlation analyses showed that drug switches were often influenced by factors such as adverse events, cost considerations, and changes in clinical practice.
Conclusion:
The Drug Switch Benchmark Dashboard proved to be an effective tool for monitoring and analyzing medication switches over time. By leveraging a rolling 12-month window and a comprehensive 24-month data collection strategy, the dashboard offers valuable insights that support informed decision-making in pharmaceutical management and healthcare policy.

5. SOC 2 Compliance Implementation for Healthcare Data Security

Project Highlights:
To meet stringent healthcare security and enterprise compliance standards, the company implemented SOC 2 compliance to protect sensitive patient data, ensure HIPAA-aligned security measures, and establish audit-ready policies for regulatory adherence. This initiative enhanced data protection, access control, and continuous monitoring to safeguard healthcare information systems.
Technologies Used:
- AWS IAM & Security Groups for controlled access to electronic health records (EHRs).
- CloudTrail & CloudWatch for logging and real-time monitoring of healthcare applications.
- SIEM Solutions for proactive security event management and threat detection.
- Encryption (AES-256, TLS 1.2+) to secure patient data in transit and at rest.
- Automated Compliance Audits to ensure ongoing alignment with SOC 2 and HIPAA.
Challenges Faced:
- Defining and documenting security policies that align with SOC 2 and HIPAA requirements.
- Managing role-based access controls for sensitive healthcare data across multiple cloud environments.
- Implementing continuous monitoring to detect and mitigate security threats in real time.
- Ensuring compliance across third-party integrations, including healthcare SaaS providers.
Value Delivered:
- Successfully achieved SOC 2 Type II certification, reinforcing compliance with healthcare security standards.
- Strengthened cloud security posture with 24/7 monitoring and automated threat detection.
- Reduced compliance risk by automating security controls and access management.
- Enhanced incident response readiness with a structured framework for healthcare data breaches.
- Ensured data integrity and confidentiality for electronic health records (EHRs), improving patient trust.
This SOC 2 compliance implementation ensures that healthcare providers, insurers, and technology partners can confidently manage sensitive patient data while meeting regulatory standards and industry best practices.

6. Telehealth Extension

Objective:
The extension allows the providers to attend appointments over video call. It also helps reduce the provider's time by taking SOAP notes from the recorded call and uploading the SOAP document against the appointment’s encounter.
Scope:
Deliver a rapid platform that offers easy ways to provide and attend appointments over the call, along with taking SOAP notes and uploading them against the appointment’s encounter by converting the notes to PDF format. Thus reducing the provider’s time for the next appointment session.
Technologies Used:
- 1. Frontend:
  - HTML, CSS – for clean, responsive UI design.
  - JavaScript – for dynamic content and user interactions.
  - Chrome extension - Delivers the application as a small program to be used in the browser itself.
- 2. Backend:
  - AWS Lambda – for handling business logic and API interactions.
  - Amazon API Gateway – for managing API requests and routing.
  - Vonage Video APIs - for making video calls and recording.
  - Amazon Transcribe - for speech to text conversion.
  - OpenAI - for SOAP note generation.
- 3. Cloud Infrastructure :
  - Amazon S3 — for storing the recorded video, audio and transcribed content.
  - Amazon Lambda and ECR - for serving backend on demand.
Value Delivered:
- Enabling video-based consultations: The extension enhances virtual care by enabling providers to conduct appointments via video calls.
- Automatic SOAP note generation: Converts provider-patient conversations to text format and extracts the SOAP notes.
- Minimizing documentation workload: Converts the generated SOAP notes to a properly formatted SOAP document and uploads against the appointment’s encounter.
- Improving provider efficiency and patient focus: Providers can concentrate on clinical decisions rather than administrative tasks. Allows providers to handle more appointments per day without compromising quality.
Challenges Faced:
- Vonage Video API Integration: Integrated Vonage video APIs into the platform and implemented polling mechanisms to reliably fetch recorded video sessions.
- Speech-to-Text Optimization: Evaluated and adopted a more accurate speech-to-text model, selecting AWS Transcribe for better transcription quality and language handling.
- Cost-Effective Backend Deployment: Deployed the backend application as an AWS Lambda using ECR images to reduce infrastructure costs, replacing the earlier ECS-based setup.
- PDF Generation in Serverless Lambda Environment: Encountered limitations in handling file I/O and PDF generation within the Lambda environment, making storage and uploading to encounters challenging.

7. Scribe Extension

Objective:
To automate clinical note-taking by capturing live audio during doctor-patient appointments and extracting answers to predefined medical questions. This tool acts as a virtual scribe, reducing the need for manual note-taking and enhancing clinical workflow efficiency.
Scope:
Build a Chrome Extension that records audio from live medical consultations, uploads it to Amazon S3, transcribes it using Amazon Transcribe, and processes the transcription with a large language model (LLM). The LLM identifies and extracts answers to predefined clinical questions, presenting the structured output within the same interface to assist providers in real-time or post-visit documentation.
Technologies Used:
- 1. Frontend:
  - HTML, CSS – For designing a clean, intuitive interface within the Chrome extension.
  - JavaScript – Manages audio recording and backend interactions.
  - Chrome Extension – Delivers seamless, browser-based functionality for immediate use during appointments.
- 2. Backend:
  - API Gateway – Secures and routes API requests to appropriate Lambda functions.
  - Lambda Authorizer – Ensures authenticated and authorized access to backend services.
  - Scribe Extension Backend (Lambda) – Handles audio uploads, transcription triggering, and coordination with the LLM for extracting structured answers.
- 3. Cloud Infrastructure:
  - Amazon S3 – Stores appointment audio files and their corresponding transcriptions.
  - Amazon Transcribe – Accurately converts audio into text using speech-to-text processing.
  - OpenAI / LLM – Extracts answers to predefined clinical questions from the transcribed conversation, replacing the need for a human scribe.
Value Delivered:
- Hands-free clinical documentation: Providers can capture and review appointment notes directly from their browser.
- High-quality transcription: Ensures accurate text generation from spoken conversations using Amazon Transcribe.
- Answer extraction from LLM: Uses an LLM to identify and summarize key clinical points in response to predefined questions.
- Reduced administrative burden: Automates the documentation process, freeing providers to focus on patient care.
Challenges Faced:
- Latency in transcription workflow: Managed smooth audio upload and processing to avoid blocking user interaction.
- Secure data handling: Implemented secure, compliant audio storage and transmission using pre-signed S3 URLs and IAM policies.
- Lambda optimization for LLM calls: Tuned Lambda functions for performance and cost-efficiency while invoking the LLM.
- Maintaining UI responsiveness: Maintained a responsive user experience while handling multi-step processing involving audio, transcription, and summarization.

Clustrex Data Private Limited

Say Hello!

Email

Phone

Address

Madipakkam Office 1

Madipakkam Office 2

Request A Demo