Building mHealth Applications (Part 2)

by Christopher Crosbie MPH, MS

As I mentioned in the first post, health market trends are driving major demand for mobile health (mHealth) applications. This is not to say there aren’t industry hurdles such as data flooding, security and privacy, clinical validation, and ease of adoption. In this post, I’ll show how these hurdles present a major opportunity for startups to leverage the functionality of AWS to directly address the mobile health segment of the healthcare industry.

Addressing mHealth Challenges with an Assist from AWS

The healthcare industry today is at an inflection point where new regulations, incentive programs, and patient demand are pushing healthcare providers to look for new and innovative solutions built on the same consumer (i.e., smartphone) technologies that we have all incorporated into nearly every other aspect of our daily lives. It is also the startup community that is expected to deliver this mobilization of our healthcare.

Traditional healthcare IT vendors are not necessarily in a position to simply absorb these applications into their products. This has often been the case as technology improvements were made to healthcare. For one, there are too many niche use cases and applications that need to be developed. Startups are much more equipped to build these numerous, targeted applications. Additionally, startups have already achieved a considerable amount of success in the mobile space, in particular those who have chosen to build on top of AWS. Leading mobile apps such as Netflix, Yelp, Airbnb, and Shazam have developed mobile applications on top of AWS that are pushing the limits of what can be done on a mobile device. There will be many more interesting technical challenges that will need to be addressed by the upcoming fleet of mHealth applications. The startup community’s expertise in building mobile applications will be necessary to develop mHealth applications that meet the bar set for adoption by tech savvy patients and physicians.

In the following sections, I discuss some of the broad requirements that healthcare has challenged startups to address. I also provide guidance about how to leverage AWS to remove undifferentiated tasks with each challenge and put an AWS mHealth startup at an advantage over other applications competing for this new segment of the health tech industry.

Data Flooding

All of the monitors and sensors associated with mHealth applications generate a lot of data that can be very difficult to interpret. There is no good way today for physicians to confirm the veracity of this highly diverse data for all of their patients. Physicians need summarizations and alerts in order to understand the information generated by the mass amount of data coming in. Such a solution is an analytics problem — algorithms are needed that can interpret and distinguish what is clinically relevant from the noise. There will be high value in creating algorithms targeted at commonplace use cases that are expected to be prevalent across many mHealth applications (e.g., continuous glucose monitoring, heart monitoring, and more).

The data required to do this analysis is very large, but Amazon Elastic MapReduce (Amazon EMR) makes it easy to quickly and cost-effectively process vast amounts of data. Amazon EMR is a HIPAA-eligible web service that lets you launch a Hadoop cluster in minutes. You don’t need to worry about node provisioning, cluster setup, Hadoop configuration, or cluster tuning. Amazon EMR optimizes the setup for the cloud environment, but still gives you root access to every instance in the cluster so you can customize to your needs. Amazon EMR can also make integration between other AWS services seamless. For example, you can keep all of your data stored in the low cost Amazon Simple Storage Service (Amazon S3) and use that instead of HDFS file storage. The benefit to this approach of breaking apart the compute cluster from the storage medium is that you pay only for the data once and not for HDFS replicated copies. You also eliminate the need to extract, transform, or load (ETL) the data into a separate system. Additional integration with Amazon EC2 Spot Instances allows you to build a steeply discounted Hadoop cluster by bidding on spare Amazon EC2 computing capacity. This low cost integration feature makes it possible to try out lots of new ideas and scientific tests for these new mobile data summarization algorithms without the risk of spiking to unaffordable compute costs.

Security and Privacy

Depending on the features of a given mHealth application and its user base, an application may fall between one and three regulating bodies: ONC (HIPAA), FDA (medical device), or the FTC (PHR). Each of these has separate reporting metrics, processes, and compliance controls.

AWS Cloud Compliance enables customers to understand the robust controls in place at AWS to maintain security and data protection in the cloud. Building an mHealth application on top of these controls eliminates a substantial set of compliance controls, which traditionally would have been needed at the physical layer (server room access, security guards, cameras, etc.).

AWS has developed a HIPAA compliance program and will even offer startups best practices and recommendations on ways to achieve a HIPAA architecture that will meet your regulatory needs. With the addition of a GxP program, which is set of standards required by the FDA, AWS now offers an official stance on how to achieve a successful regulatory framework in the cloud, regardless of the US healthcare reporting body. In fact, AWS also offers many services that can help you meet or exceed healthcare compliance and security standards found throughout the industry today.

Clinical Validation

To truly be accepted in clinical care and to move an mHealth application into a reimbursable validated part of care, an application will need to undergo randomized clinical trials. This is a long and expensive process, but establishes the “gold standard” for what the application can be expected to do for a patient. AWS understands that mHealth startups may need some additional resources to get started and offset the initial cost. The AWS Activate program was built to address this very need, and some of the fastest-growing startups in the world began in this program.

Easily Adoptable Solutions

Healthcare is an industry that responds well to fully wrapped, end-to-end solutions. There is a lot of value that can be derived from tying together healthcare data in the cloud, but this software will need to be developed by a mobile ecosystem that can build the software for the medical industry. Mobile health startups will not have the luxury of waiting for consulting firms or overworked hospital IT staff to learn and adopt new mobile technologies. Startups will need to build production-ready code deployments of infrastructure to move their mHealth applications into the legacy healthcare environment.

AWS offers a variety of ways for startups to build automated deployments with a few clicks. AWS CloudFormation gives developers and systems administrators an easy way to manage a JSON representation of their AWS resources that can be provided as templates. AWS Service Catalog allows startups or organizations to create and manage catalogs of IT services that are approved for use on AWS.

The AWS Marketplace is also a sales channel that makes it easy for AWS Partners to offer their mHealth solutions pre-configured for an end customer.

Collecting the Data

Startups that want to build their own monitors and sensors may also find the AWS IoT service an easy way to wrap security around their patient’s sensor data. AWS IoT is a managed cloud platform that lets connected devices easily and securely interact with cloud applications and other devices. Many device protocols used in IoT such as MQTT are inherently very insecure. With AWS IoT, you can use an end-to-end encrypted MQTT with dual authentication to help ensure that data is never exchanged between devices and AWS IoT without proven identity. If there ever is an issue with a device, certificates can easily be revoked. There are many other features of AWS IoT, such as device shadows and rules, that make it easy to control complicated interactions of thousands of devices with little to no programming.

Building the App

Of course, at the heart of mHealth is the mobile application itself. I will conclude by saying that any startup that is considering building a mobile health application on top of AWS should review the AWS Mobile Hub console. AWS Mobile Hub simplifies the process of building, testing, and monitoring mobile applications that make use of one or more AWS services. It helps you skip the heavy lifting of integrating and configuring services by letting you add and configure features to your apps, including user authentication, data storage, backend logic, push notifications, content delivery, and analytics — all from a single, integrated console. The output is working code comprising AWS best practices that developers can take and immediately incorporate into clinically valuable code that differentiates their mHealth applications.

Bringing It all Together

In this post, I’ve introduced you to some of the AWS services that are useful to startups that want to quickly build and scale an mHealth application. With the diversity of services available and the ease of integration, startups that choose to build on top of the AWS platform should be able to continually expand their business without having to find and integrate additional infrastructure components. This puts them at a significant advantage in the mHealth space, and allows them to focus on the industry issues instead of the technical plumbing underlying sophisticated mobile technology.