Building a multi-channel, data driven patient engagement platform with AWS
In today’s digitally transformed world, it’s more important than ever to have a deeper understanding of your patient’s attitude and behavior, and to engage them with personalized content through the channels that they prefer in order to provide efficient and personalized patient care. In addition, with today’s value-based healthcare, the patient is the center of the healthcare ecosystem, so it’s important to understand them and deliver healthcare solutions to maximize their care while minimizing cost.
In this post, we discuss the current challenges in the healthcare industry and how AWS services can be leveraged to tackle this problem. The solution functional areas discussed later in this post are the foundation for building and operating a multi-channel, data-driven patient engagement platform.
Patient engagement is one of the pillars of population health management and value-based care, but traditional, print-heavy attempts at engagement have failed. That’s because an estimated 90 million adults have low health literacy, and printed brochures and flyers are often lengthy and have a lot of medical jargon. That makes it difficult for many patients to remain adherent to physicians’ treatment plans, especially the millions of Americans who have chronic conditions like COPD, diabetes, hypertension, and depression.
A recent study conducted by the Council for Affordable Health Coverage found that two-thirds of patients in the United States don’t adhere to their medication treatment plans. Non-adherent patients cost the U.S. healthcare system an estimated $100 billion or more each year. These patients tend to have poor outcomes, which costs physicians money because of Medicare’s Value-Based Payment Modifier and value-based contracts with private insurers.
The need of the hour is to build a data-driven recommendation engine to drive patient engagement, creating a more patient-centric, multi-channel, data analytics platform. The global patient engagement solutions market size is expected to reach US$ 56.92 billion by 2026, according to a new report by Grand View Research Inc. The supporting government initiatives and rising usage and awareness regarding mobile healthcare services are expected to drive the growth of patient engagement solutions across the globe.
Key technology considerations
- Building a channel-aware platform to maximize the patient engagement, enabling omni-channel switching capabilities and delivering a user experience expected from modern applications
- Personalizing the contents, notifications for patients, payers and HCPs that are timely and within context
- Ensuring the solution is compliant, secure and protects patient information and at all times
- Ensuring delivery teams can execute at speed, are agile in responding to changing market conditions while keeping costs low
- Can handle the data tsunami from a multi-channel solution; be it real world evidence, EHR, sales data, real-time patient behavior on social media channels etc.
- Applying appropriate user segmentation to enable more informed business process decisions
- Ability to leverage new innovations in data analytics, AI and ML and future communication channels without having to re-architect the entire platform
Given these technology considerations we’ll now walk you through AWS solution reference architecture that leverages AWS services to deliver a patient engagement platform in a reliable, scalable, and consistent way.
A reference architecture for building the multi-channel, data driven patient engagement is shown below:
This solution architecture focuses on catering for patient and payer interactions across multiple channels, handling requests in a consist manner in AWS and communicating back to patients and payers in a timely and meaningful way on a preferred channel for them.
This solution is broken down into six functional areas and as we walk through the architecture we will dive deep into each one looking at the AWS services applicable to the functionality we are looking for.
Amazon Lex is used as a conversational chatbot service to automate simple requests and provide answers to FAQs, or routing to call center agents.
For ease of interaction, patients and payers can use a number of developed channels and integrations with this solution, with the specific channels of mobile apps, digital therapeutics devices or voice-enabled devices called out specifically for key interactions such as patient registrations, appointments, payer status alerts, prescriptions, medicine distribution wellness advice and more.
Security and compliance
Amazon Cognito is used as the authentication mechanism in this architecture, which lets you add user sign-up, sign-in, and access control to web and mobile apps quickly and easily. Amazon Cognito scales to millions of users and supports sign-in with social identity providers, such as Facebook, Google, and Amazon, and enterprise identity providers via SAML 2.0.
Amazon Cognito also supports multi-factor authentication and encryption of data at rest and in-transit and is HIPAA eligible, PCI DSS, SOC, ISO/IEC 27001, ISO/IEC 27017, ISO/IEC 27018, and ISO 9001 compliant.
Additional AWS services such as CloudTrail allows for audit trail of activity throughout the solution and Amazon Macie is one of our security services that is applicable here that uses machine learning to automatically discover, classify, and protect sensitive data in AWS.
Event orchestration is managed through Amazon Kinesis, receiving real-time data from multiple application channels, social media, wearables, website, and mobile app clickstreams and IoT telemetry data from wearables. It also enables you to process and analyze data as it arrives and respond instantly instead of having to wait until all your data is collected before the processing can begin.
This architecture uses the NoSQL datastore, DynamoDB to track patients and activity, and is not only a single-digit millisecond performance database, but includes features such as DynamoDB streams that allows the architecture to use a publisher/subscriber pattern on the submission of data to interested parties such as payers, primary care systems and patients.
Proliferation of data from medical devices and social media data makes a scalable event orchestration layer essential.
Once data requests are stored in DynamoDB, we can classify structured and unstructured data using one or more of Amazon’s feature-rich AI services. In this illustration, we are using Amazon Comprehend Medical to classify and understand medical terminology that could be present in free-text that describes a condition. The classifications identified within the content are then added as metadata to the data record that enables accurate reporting and improved aggregation.
Data Storage, cataloging and enrichment
Segregating of storage and compute is a core concept on AWS, which is often a major contributor in reducing overall cost of solutions. In this solution, we achieve this through pushing structured and unstructured data generated from different channels into a data lake solution.
Data residing in the data lake is cataloged, and combined with other correlative data such as environmental data, medical journals and research, time series data, social trends etc. and then using the managed machine learning framework, Amazon SageMaker, we can then develop recommendation engine endpoints for next best actions that can be integrated back into internal or external applications.
An important factor for the health industry as well as many other regulated industries is that access control and operations on data in S3 can be managed right down to an object level with bucket policies and IAM roles, and data can be encrypted using either client or server-side encryption, with management of keys handled by AWS KMS.
Lastly, within the data storage layer of the solution, AWS has a feature called Intelligent Tiering, which can automatically detect access patterns in data and transition data to less expensive tiers to save customer’s money.
With new data being ingested and processed from various channels continuously, we can then trigger Amazon Personalize in real-time to help segment patients and payers based upon historical interactions and events, and then use Amazon Pinpoint to target audience segments in a very meaningful and timely way across channels such as email, SMS, voice, or push notification. This personalized and context aware communication moves healthcare organizations away from mass-marketing campaign with limited success to a more patient-centric approach that is designed to improve patient and payer experience, increase retention rates and ultimately improve the well-being of patients.
In this blog post, we have talked about the challenges in the traditional patient engagement approach, and how through technology we can create an experience that delights the patient and provides an automated, data-driven platform backbone as the future of patient engagement.
Healthcare industry is in the midst of a transformation driven by overall healthcare reform, outcome-based patient care, and advancement in technology such as artificial intelligence, machine learning, cloud, Internet of Things (IoT), and so on. Historically, these problems have been difficult to address. However, with the advent of cloud technology, healthcare customers are addressing this critical component of the value chain in new ways by making data securely available at the right place and time, and accessible through right channels.
The innovative AWS services used in this architecture describes a matured state for a lot of our healthcare customers that is potentially years away and nestles deep within a digital transformation program of work. With solutions like Amazon Pinpoint and the newly released Amazon Personalize, taking that first step to create an exciting patient experience is easier than you’d think.