AWS for Industries

HMS People Heart Study streamlines digital research with AWS

Evidence suggests that adverse consequences of heart disease, such as stroke and heart attack, can be prevented by appropriate lifestyle and medical management. Leveraging standards-compliant services, such as Amazon Web Services (AWS) HealthLake, the People Heart Study team has built a system that allows it to collect and analyze health data relevant to the study.

This data is then used to provide personalized recommendations through an iPhone app. It is aimed at helping study participants self-evaluate their risk of developing a heart attack, stroke, or other cardiovascular disease using data from their care providers.

Raising awareness of heart disease risk using technology

Heart disease remains the leading cause of death in the United States. The American College of Cardiology (ACC) and American Heart Association (AHA) have published guidelines to prevent the development of heart diseases. These guidelines include adopting healthier eating habits, physical exercise and, if necessary, medications to control risk factors such as cholesterol and blood pressure. These clinical practice guidelines (CPGs) provide guidance to doctors for assessing their patients’ heart health.

Yet it can be extremely challenging, if not impossible, for a physician to be knowledgeable about each aspect of each guideline and to recognize when one is applicable to a particular patient. The sheer volume of patients, the uniqueness of each patient, the large number of CPGs, the length of each CPG (often over 100-200 pages), and the frequency of CPG updates all contribute to that challenge.

Technology has an enormous potential to help both physicians and patients. Particular individuals may benefit from guideline-directed interventions that decrease their risk of significant adverse health events like strokes and heart attacks.

Recent advances in Healthcare IT, with the emergence of HL7 Fast Healthcare Interoperability Resources (FHIR) and the adoption of the SMART on FHIR API, have transformed electronic health records (EHRs) into platforms allowing access to third-party health apps. The Apple Health app (on most iPhones today) uses the SMART on FHIR API to retrieve patient records into its users’ personal devices.

These technologies can be leveraged to empower individuals to become aware of their own personalized risk of health issues, such as heart disease. The individual can then become their own advocate and partner with their physicians to decrease their risk for such events.

The People Heart Study

The People Heart Study, part of the People-Powered Medicine (PPM) group at Harvard Medical School, investigates how individuals can use their personal health data to understand the health of their heart and blood vessels (that is, their cardiovascular system). The People Heart Study aims to understand whether individuals can reliably self-assess using an app and receiving a personalized recommendation summary that uses their health data from their care providers, which is digitally accessed through the SMART on FHIR API.

Specifically, the People Heart Study app will empower study participants to evaluate and receive recommendations for the primary prevention of heart attack, stroke, and other cardiovascular disease, based on the authoritative guidelines from the ACC and AHA. To participate, the user downloads the app, enrolls by digitally consenting, and then completes four tasks. These tasks include short surveys on medical history. There is also a special task that guides the user on how to capture health record data stored in the Apple Health app, which is linked to their care provider’s EHR. The People Heart Study app then pulls lab test results, medications, and clinical conditions as required by the study’s protocol.

This study will assess the data quality and whether it is sufficiently complete for soundly executing a CPG’s decision tree. This is important in establishing the utility of this novel channel for patient-directed decision support. This study will also offer a qualitative assessment of how much clinical context is available through the SMART on FHIR API in relation to a respective CPG.

Leveraging standards and managed services to increase efficiency

The study’s architectural principle is to be efficient, lightweight, and simple to set up by leveraging current Health IT standards-based services. The entire architecture consists of a study-data repository in the cloud on AWS and an iOS research application for participants. As shown in Figure 1, the study data repository utilizes AWS HealthLake as its sole FHIR-based datastore.

Two public APIs are configured in Amazon API Gateway to govern “write-only” access to the AWS HealthLake datastore. The first API is a token API that verifies the requesting client (the research app) is indeed the People Heart Study app. It also returns a bearer token from Amazon Cognito after verifying the app’s credentials, which are stored in AWS Secrets Manager. The second API is a FHIR Resource API that relays FHIR resources to AWS HealthLake after verifying the access-token and the attached password key.

Using AWS HealthLake and other AWS services

High-level architecture used to build the People Heart Study app on AWS. Architecture shows an iOS App interfacing with Amazon API Gateway, which then interfaces with AWS Lambda. The AWS Lambda uses AWS Secrets Manager and AWS Cognito to verify credentials and then calls AWS HealthLake to verify the signed request and then return FHIR output. Figure 1: High-level architecture for the People Heart Study

The following is a further breakdown of the People Heart Study app architecture and how the two different APIs work:

1. API 1: The People Heart Study app first obtains an access-token for submitting FHIR resources to the repository.

a. Amazon API Gateway receives token requests from the app along with client secrets (“1” in Figure 1), which are relayed to an AWS Lambda function (“2” in Figure 1).

b. The function then verifies the study app and credentials stored in AWS Secrets Manager (“3” in Figure 1).

c. AWS Cognito issues an access-token to the app (“4” in Figure 1).

2. API 2: FHIR Resource API

a. The app sends a FHIR-compliant request with the access-token to the FHIR Resource API setup on Amazon API Gateway (“1” in Figure 1) and relayed to the AWS Lambda function that handles FHIR calls (“2” in Figure 1).

b. The AWS Lambda function calls the corresponding AWS HealthLake endpoint using AWS Signature Version 4 (Sigv4) authentication (“5” in Figure 1).

c. AWS HealthLake is configured to verify the signed request and, upon success, return a FHIR output before returning back to the study app.

FHIR is now the standard of health information exchange across health systems in the United States. FHIR data is becoming more pervasive and in the hands of individuals (through the efforts of Apple Health and CommonHealth on Android). Researchers can utilize ready cloud-managed FHIR stores (like AWS HealthLake) as the main study data repository through an institutional controlled, HIPAA-ready service account with standardized access controls (OAuth2 or Amazon Cognito).

This brings significant advantages for research teams:

1. More time to focus on the participant-facing app experience

2. Use of FHIR-enabled frameworks (for example, SMART Markers) to further streamline and reduce developmental burdens

3. Predictable standard data formats to analyze using notebooks

Conclusion

The People Heart Study is an example of how digital research can be streamlined by leveraging interoperability standards like HL7 FHIR by using AWS HealthLake. Individuals can be empowered to better understand their own cardiovascular disease risk through a patient-facing app aligned with published clinical practice guidelines.

AWS cloud computing technologies are providing researchers the ability to quickly deploy research studies and enable the use of personal health data to help inform patients about their individualized disease risk. This creates the ability to collaborate with their clinicians to improve their outcomes.

Where to learn more

You can learn more about this study at https://peoplepoweredmedicine.org/heart and enroll in the study by downloading the app from the Apple App Store.

Contact an AWS Representative to know how we can help accelerate your business.

To learn more about AWS for Healthcare & Life Sciences—curated AWS services and AWS Partner Network solutions used by thousands of healthcare and life sciences customers globally—visit the AWS for Healthcare & Life Sciences and AWS Healthcare Solutions webpages. For more information about AWS HealthLake and other AWS health data services, check out the AWS Health Data Portfolio site. You can also read more blogs about AWS healthcare stories.

Raheel Sayeed, MBBS

Raheel Sayeed, MBBS

Dr. Raheel Sayeed is a physician and biomedical informatics researcher at Harvard Medical School, working at the intersection of clinical AI, digital health infrastructure, and health data standards. He leads the People Heart Study, a national initiative exploring patient-centered cardiovascular risk assessment using real-time EHR data. His ongoing research also investigates the role of large language models in clinical decision-making and value alignment.

Christine Tsien Silvers, MD, PhD

Christine Tsien Silvers, MD, PhD

Christine Tsien Silvers, MD, PhD, serves as Healthcare Executive Advisor at AWS. Her research at MIT and Harvard Medical School since the 1990s focused on AI/ML and their use to improve patient care. Trained at Massachusetts General Hospital and Brigham and Women’s Hospital, she is Board certified in Emergency Medicine as well as Clinical Informatics. In the 20+ years prior to joining AWS, Chris worked clinically and then served as Chief Medical Officer at two healthcare technology startups. She is passionate about leveraging technology to improve health.

Isaac “Zak” Kohane, MD, PhD

Isaac “Zak” Kohane, MD, PhD

Isaac “Zak” Kohane, MD, PhD, is the inaugural chair of Harvard Medical School’s Department of Biomedical Informatics, whose mission is to develop the methods, tools, and infrastructure required for a new generation of scientists and care providers to move biomedicine rapidly forward by taking advantage of the insight and precision offered by big data. Kohane develops and applies computational techniques to address disease at multiple scales, from whole health care systems to the functional genomics of neurodevelopment. He also has worked on AI applications in medicine since the 1990’s, including automated ventilator control, pediatric growth monitoring, detection of domestic abuse, diagnosing autism from multimodal data and most recently assisting clinicians using whole genome sequence and clinical histories to diagnose rare or unknown disease patients. He is the inaugural Editor-in-Chief of NEJM AI and co-author of a recent book “The AI Revolution in Medicine." He is a member of the National Academy of Medicine, the American Society for Clinical Investigation and the American College of Medical Informatics.

Paul R. Saxman

Paul R. Saxman

Paul Saxman is the Software Lead for Advanced Spatial Modeling at the Gates Foundation, where he develops innovative software and AI tools to support infectious disease research and improve predictions of disease dynamics. Previously, he was a Principal Solutions Architect and the Global Tech Lead for Education and Academic Research at AWS, where he led technical initiatives that helped institutions adopt advanced cloud technologies.