SOLACE
Bridging the gap in Palliative and Hospice Care
This intelligent app goes beyond simple monitoring, providing healthcare providers and caregivers with real-time predictions of patient vital signs and symptom progression. Enable timely interventions, leading to improved patient outcomes and reduced complications.
Introduction
Palliative and hospice care in the Philippines face significant challenges, with only 10% of hospitals providing these essential services and a critical shortage of trained professionals.
As a pioneer in home-based palliative and hospice care, the Ruth Foundation contends with workforce limitations and an urgent need for advanced tools to deliver consistent, personalized care.
To address these pressing issues, SOLACE was developed in collaboration with The Ruth Foundation, introducing real-time symptom monitoring powered by AI integration to revolutionize patient monitoring and elevate the standard of care.
Advocacy
By bridging the gap in palliative and hospice care, SOLACE, together with The Ruth Foundation, seeks to improve the patient monitoring to provide the utmost care the end-of-life patients deserve.
Authors & Partners
This project is a collaborative effort by the following individuals and organizations, who have dedicated their expertise to enhance palliative and hospice care through innovative solutions.
Renzo Viñas
Project Lead, Frontend Developer, UI/UX
Earl Tacda
Backend Developer, AI Engineer
Cassandra Roxas
Data Analyst, Database Manager
Dr. Jerian Peren
Thesis Adviser
In partnership with
Methodology
Research Methodology
This study used Mixed Methods Approach to combine both quantitative and qualitative data collection and analysis. The quantitative data was collected through structured survey forms, while qualitative data was gathered through interviews and focus group discussions with healthcare professionals, caregivers, and patients.
Data Collection
eICU dataset for training the AI Model & Structured Survey forms for Problem Identification and Evaluation. The survey forms were designed to gather insights from healthcare professionals, caregivers, and patients, ensuring a comprehensive understanding of the challenges and needs in palliative and hospice care.
AI Model
Extreme Gradient Boosting (XGBoost) to predict future symptom and vital sign flare-ups. XGBoost learns from every input. It compares the actual and predicted values to get the residual error to help adjust the XGBoost model to create more accurate predictions in the future.
Tools and Instruments
Survey Questionnaires are based on ISO/IEC 25010:2023 Standards along with Technology Acceptance Model (TAM). XGBoost Model Prediction Accuracy is evaluated using Python.
System Development
Agile-Scrum Methodology was used to organize the software development lifecycle. Each sprint was planned to deliver specific features and improvements, ensuring continuous feedback and adaptation.
Data Analysis
Evaluate XGBoost Model Using MAE (Mean Absolute Error), MSE (Mean Squared Error), RMSE (Root Mean Squared Error), and R2 (R-squared). Evaluate User Satisfaction, Perceived Benefits, and Acceptance through Descriptive Analysis (Quantitative) and Thematic Analysis (Qualitative).
Development Process
SOLACE was developed using Flutter for frontend and Python, Google Cloud, and FastAPI for backend. Firebase was used for authentication and database management. Figma and Photoshop were used for UI/UX design.
System Architecture
Users interact with the system by role. Each activity is bonded to the database, thus providing real-time data processing. The XGBoost model accesses the database to record and provide timely predictive interventions.
Main Features
Patient Tracking
Collects vital signs and symptom inputs from the patient. Vital signs are obtained through manual input, while symptom assessments are collected using sliders. A summary of the tracking input is displayed to ensure data validity and accuracy.
Real-time Dashboard
Provides a comprehensive view of predicted critical vitals, analysis of symptom tracking, and tracking history.
Real-time Intervention
Generates non-pharmacological interventions and steps from the detected symptoms of the patient based off the patient tracking module.
Real-time Alerts and Notifications
Provides real-time communication between healthcare providers about the patient's status and activities.
Other Features
SOLACE includes additional features such as Note Taking, Patient Scheduling, Task Assignment, and Medicine Prescription to enhance the overall healthcare management experience.
Results
Metrics (MAE, MSE, RMSE) typically increase due to greater uncertainty in long-term predictions. R2 shows the inability to capture the underlying patterns, but still provides assurance when making predictions.
For most vital signs, particularly Temperature, SaO2, and Heart Rate, all metrics remain near zero, indicating high predictive accuracy across all horizons.
User Satisfaction
Healthcare providers were generally satisfied with the system, particularly about its reliability and usability.
Perceived Impact
Confirms that SOLACE has a 'High Impact' on enhancing patient monitoring and simplifying symptom tracking, and better patient-caregiver communication.
Healthcare Provider Acceptance
Healthcare providers find the system to be acceptable in performance efficiency, usability, and ease of use.
System Developer Acceptance
System developers find the system to be highly acceptable due to the user interface, functionality, and impact on palliative and hospice care settings.
Conclusion
SOLACE effectively addressed the Ruth Foundation of the Philippines’ challenges in manual symptom monitoring and timely response by enhancing home-based caregiving, empowering decision-making with helpful predictions, and delivering real-time non-pharmacological interventions.
The development and study of SOLACE have significantly contributed to the United Nations’ Sustainable Development Goals, particularly in the areas of Good Health and Well-being, Industry, Innovation, and Infrastructure, as well as Partnerships for the Goals.
Recommendations
Although the results are promising, there is still room for improvement. Enhancing the AI model with more comprehensive datasets and hybrid modeling approaches could significantly boost its performance. Additionally, integrating wearable devices for real-time data input and deploying SOLACE in home-based palliative and hospice care institutions affiliated with the Ruth Foundation are crucial steps to unlock its full potential.
