PrepCheck

Inspiration

PrepCheck was inspired by a story close to our hearts. One of the team members reported experiencing a case where a family member had to go through surgery. Far from reassuring, key information was hurried over or left out when the patient felt lost in the sea of 'big words' used in pre-op consultations. The confusion and uncertainty this caused were unnecessary stresses added to an already difficult situation. It, therefore, was apparent that accessible information was a common problem within healthcare. We realized that patients had the right to understand what was happening to their body and how to prepare for surgery in terms they actually understood. The personal experience ignited our drive to create a tool that would bridge the communication chasm between caregivers and healthcare patients, thereby empowering them with the knowledge to feel informed and confident.

What it does

PrepCheck is an AI-powered pre-operative assistant built with the purpose of simplifying the pre-surgery process for both the patient and the healthcare provider. It clearly explains the surgical procedure, what one should expect, potential complications, and all the pre-surgery do's and don'ts instructions. The AI converts a complex medical explanation into understandable terms; hence, the patient will understand how to prepare and what to expect. Moreover, PrepCheck promotes accountability by normalizing information given in pre-surgical briefings and ensures that critical information does not get excluded.

How we built it

We built PrepCheck using a combination of Python and Flask to provide it with a simple yet effective backend infrastructure. Natural Language Processing was quite central in the development of clear, natural communication of the AI with patients. We trained it on sets of medical terminology and patient education materials so that the AI can provide responses which are not only accurate but also with appropriate detail. Our approach here was to break down the preparation process into smaller chunks and to address very specific questions and concerns that a patient may have. We also created a feedback loop for continuous learning - AI improves on real-world usage.

Challenges we ran into

One of the most major challenges we needed to face was to make sure the AI responses were not only factually correct but appropriately nuanced. Further, the most challenging tasks included simplifying difficult medical terms to make them understandable, yet without losing vital information in the process; hence, trying to find an optimal balance between clarity and completeness. Besides, AI training that would fit into various types of surgeries and individual needs of patients was harder than expected because of the wide scope of medical procedures. Similarly, there were technical questions regarding how to attach the NLP functionality seamlessly to the system without compromising the system's performance and responsiveness.

Things we are proud of

We are proud that, in this timeframe, we have managed to develop a working prototype that fills the existing gap in the field of communication for the care of patients before surgery. We consider it gratifying to realize that we will contribute to a solution aimed at minimizing anxiety in patients while improving safety within surgical procedures. We also consider ourselves proud because we have managed to overcome serious technical difficulties with the NLP part of the work-a thing that actually allows AI to craft explanations that could be easily understood by patients. But what really demonstrated the team collaboration and problem-solving was how we were continuing to work out PrepCheck and make it as useful and user-friendly as possible.

What we learned

In the development of PrepCheck, we came to learn about the various complications possible with healthcare communication, as well as about the diverse needs of patients. We realized how important it was to provide information in a manner that would be easy to digest, though complete. Equally, we improved our skills in natural language processing and AI development-skills of tuning algorithms for accuracy and user experience. Beyond the technical skills learned, which were empathy in technology design, at every turn, it had value: keeping the patient in mind. It truly matters in the success of our project.

What's next for PrepCheck

The future for PrepCheck is great, and we look forward to continuing to work on enhancing the capabilities of the AI to cover a wider range of surgical procedures and conditions. Further down the road, we will incorporate multi-language support and accessibility features for patients with special needs to ensure that PrepCheck can assist some of the most diverse populations. We also want to integrate with EHR systems so that patient information is streamlined, making the pre-surgery process even more automated. We envision future collaboration with hospitals and healthcare institutions on the eventual piloting of PrepCheck in real-world settings, continuous improvement in its accuracy and usefulness based on real patient feedback.