Felix and Kirsten both worked at coop placements in the health industry. While working in very different settings, they both experienced one common issue: recording and transferring patient information and health records. While health records and electronic health records have become commonplace in many clinics, physicians seem less-than-inclined to use them due to the copious amount of writing or typing required. This makes sense, as a physician's expertise lies in diagnosing ailments, treating them, and working with patients, and not inputting values into tables or signing paperwork. This doesn't seem like a very large problem until one factors in the consequences of a physician rushing through this information-gathering process to focus on patients. 50 out of 175 million patient accidents were caused by typing errors on medical forms. This is outrageous and is only exacerbated by something like a global pandemic increasing the traffic in hospitals and, consequently, the number of forms physicians must fill out. To fix this major issue, our team decided to create a website that automates the EHR recording process. Drawing from our own experiences, we realized that physicians spend most of their time during check-ups talking to the patient. Instead of talking to the patient and looking down at a tablet to type information, why not just use the information rushing out of their own mouths? Hence, our solution is to use voice-to-speech and an efficient information processing algorithm to store patient information without the need to type a single word!

What it does

Our website has several features. Firstly, users sign in with a username, password, and occupation. Once signed in, users will enter the main page, where they will be directed to 3 links: the patient information form, the prescription form, and a link for contacting us in case they need help using our site.

The patient information form is the main feature of this website, as this is what automates filling out EHRs by hand or by typing information manually. The physician simply has to click "Start Recognition" and allow the site to access their microphone. Then, anything they say will be converted to text by the Web Speech API in our project. Once the physician stops speaking for over 15 seconds or the physician clicks "Pause Recognition," the voice-to-text API will stop receiving input. During pauses of speech, transcripts will appear inside of the main text box. Physicians can correct any mistakes made by the voice-to-text API inside this text box. Once the information is confirmed to be accurate, the physician simply needs to click "Save Information." Our website instantly cuts away the unnecessary words, leaving behind all the important details from the transcript pertaining to the patient by identifying keywords and only saving parts of the transcript surrounding these keywords. This is stored on the webpage with a date and time stamp of when information was saved. Physicians can delete this with a simple button click as well. This process streamlines EHRs tremendously, since physicians can speak with a patient and fill out the patient's EHR at the same time without looking at a screen or typing. Physicians can have all their patient's crucial information saved in as few as 3 button presses.

How we built it

We built this website using HTML, CSS, Javascript, and Web Speech API. The Web Speech API received microphone input and provided us with easy-to-use methods for transcribing vocal input to text. HTML was used for the skeleton of each webpage, CSS was used to style the webpages, and Javascript was used to allow button clicks to trigger actions, to receive vocal input, and to output relevant patient information.

Challenges we ran into

Some challenges we ran into were properly receiving the voice input. We had to read documentation on Web Speech API and access tutorials on the internet about integrating the API into Javascript code. We also ran into issues with styling the website and storing the voice input properly. However, we were able to solve all our problems by researching and finding alternative methods to achieve a similar end result.

Accomplishments that we're proud of

We're quite proud of making an aesthetic website in under 24 hours. We are even prouder of getting the Web Speech API to properly receive vocal input and enter it into the text box. Lavan was quite jovial when he managed to properly filter vocal input into a concise format when saving patient information and stamp it with the date and time. Overall, we are proud of the amount that we were able to achieve considering that most of our group had never worked with web development tools prior to this hackathon. In fact, most of our team had never attended a hackathon prior to this one, so this has been a wonderful experience and a great learning opportunity!

What we learned

We learned how to create a webpage using HTML, how to style elements using CSS, how to create functions triggered by button clicks in Javascript, and how to use an API to convert vocal input into text. Finally, we were able to turn our project into a live demo site using Github Pages! (UPDATE: no longer online)

What's next for Effortless Health Records

With more time, our group plans to complete the prescription form and improve on the storage of patient information using databases, so that patient records can be amended after an initial recording has been saved. We also want to work on security, ensuring that login information is stored safely and not displayed in the web browser link.

Built With

Share this project: