The telemedicine sector of healthcare has become a vital and, oftentimes, inseparable component of successful care providers. During the COVID pandemic, many facilities encountered a key problem when trying to follow the guidelines and ensure a proper distancing of the patients coming to see their providers.
Reduction of the lines in the hospitals and private facilities was a crucial step to be taken and that’s where telemedicine comes in handy. Especially, when dealing with mental health, when the timing is crucial. With the help of our client’s solution, healthcare specialists get contacted via video sessions, therefore, there’s no risk of catching the virus.
Eventually, the growing popularity among the platform’s clients led to the need of updating the existing system since it handles large scopes of data, utilizes machine learning to help detect any deviations in the patients via video and notify the doctor, etc.
In the article, we are talking about the newest updates to the platform and how our client plans on expanding the range of adjacent healthcare spheres outside mental health solely. You can follow the link to read more on the given Clojure solution.
The Telemedicine Project Background
The mobile app is built utilizing Clojure, and the web version – ClojureScript. they also have extended functionality that contains but is not limited to the video/audio/language in-app analysis (for building emotional estimations), SMS notifications, push notifications, etc.
The team is also involved in the development of other essential components such as the AI/ML model from the early stages of the app. All the materials collected from the patients, including videos, texts, automatic scores, and estimations are available to the healthcare providers and allow them to get immediate access and estimate the patient’s condition.
Updated functionality for the improved user experience and simple technical support
As the number of registered healthcare providers, therefore, their patients was growing, the development team needed to introduce updates and changes to the platform, so it doesn’t fail while operating and processing large scopes of data. In addition, there were some minor changes to the UI aiming at making the further support process simpler and less costly for the client.
- An updated dashboard so that care providers could track the progress, i.e., the number of sessions for each patient, the number of patients and how it grows over a period of time, etc.
- The platform integrates with multiple EHR systems for storing health records. The integrations allow users to stay within the system without having to switch when they need a patient that’s registered in a different one. The records can be imported, and healthcare providers can monitor the patient’s progress after they check out of a healthcare facility. In addition, healthcare providers can schedule sessions and make records during the sessions.
- The next step for our client is to pass SOC 2 certification. This way we ensure proper data management and comply with all obligations regarding privacy, data security, and availability. The solution will allow our client to grow their customer base since many require SOC 2 certification completed before starting to collaborate with the platform.
- To simplify the log-in process, the development team is currently implementing Single-Sign-On solutions including Microsoft SSO.
- Implementation of analytics through the metabase system. This is the database solution written in Clojure for analysis of other databases.
- Introduction of Storybook which allows to organize and document UI components
- Since the platform handles large scales of data and files, it was crucial to set up an environment that will manage to process these data without the server failing and causing any disruption to the system. Our engineers implemented the solution which allows our server to control data upload to AWS. Therefore, the data streaming process is continuous and provides a great user experience to the platform users. When the video uploads to Amazon, Step Functions launches. In the previous versions, lambda functions were invoked that would run case analysis, change the formatting of videos, etc. The reason for switching to Step Function was that we needed to gain more control over data.
In the beginning, we had a function that would transcode video into different formats. Yet, this approach would lead to system failure and files didn’t run properly. Moving forward from the recommended Amazon approach, when an event is placed onto S3 Bucket, then lambda function triggers, and so on, we decided to incorporate an additional server for video upload, and wrote function stack, which triggers after the event. With Step Functions, we made a function reusable, if input parameters differed.
This solution allows gaining more control over the system functionality as well as error detection process takes less time. Moreover, it is easier to fix any malfunctioning as it would only limit to a specific segment of the platform, without affecting the functionality of the entire system. The client also got an improved interface of the system architecture and parameters.
To expand the number of users, the platform is also set to step outside the mental healthcare sector and grow into other adjacent ones such as fitness, pharmaceutics, nutrition and dietology, and more. There is a huge market gap that we are hopeful to fill with our product.
For the past two years, the world had to learn how to transition from the majority of things we were accustomed to for the sake of societal safety. The healthcare industry was the one to encounter probably the most difficulties. And that’s the reason why it was essential to create something that’ll help reduce face-to-face doctor visits unless they are an emergency. And here’s how telemedicine became the way out of the problem.
Our solution incorporates the best technologies along with a well-thought and tested platform where mental healthcare providers would be reached out by their patients whenever there is such a need without having to wait for the actual appointment. We use machine learning to teach the system how to differentiate multiple symptoms and features like tone of voice, eye motion, speech speed, and many more. To prove the accuracy of the automatic analysis, a doctor always checks the system’s report on a given patient. So that nothing important is missing.
The solution is a great way for establishing doctor-patient communication without wasting time commuting to the appointment or waiting for one for weeks.