PoDiaCar is a digital healthcare platform designed to support pediatricians in monitoring cardiometabolic risk in children and adolescents. The goal was not to build a simple clinical management tool, but a system capable of integrating longitudinal data, automatically calculating key indicators, and providing a structured view of the patient’s evolution over time.

In healthcare, this means reducing manual error, improving follow-up quality, and ensuring that every technological decision aligns with strict security standards.

Fragmented clinical data and manual calculations

Pediatric cardiometabolic monitoring involves multiple parameters, including blood pressure, BMI, lipid profile, glucose levels, and related indicators. These values are often analyzed individually at each visit, with calculations performed manually or using separate tools.

This approach makes it harder to evaluate long-term trends and increases the risk of inconsistencies in data recording.

The project originated from the need to centralize clinical information, automate indicator calculations, and provide physicians with a tool capable of transforming isolated measurements into a coherent and continuously updated clinical overview.

An integrated platform with a predictive model

We developed a web-based platform that combines data collection, automated indicator calculation, and decision support.

At the core of the system is a machine learning model that analyzes patient parameters and estimates eight cardiometabolic indicators. The model combines regression and classification algorithms, selected through comparative testing and cross-validation to balance predictive performance and clinical interpretability.

Each new visit updates the patient’s profile and recalculates risk panels in real time. This shifts the perspective from static data points to a longitudinal analysis, which is essential for preventive care.

A dynamic digital representation of the patient

One of the key components of the platform is the pediatric Digital Twin: a structured digital representation of the patient that evolves as new clinical data is added.

The Digital Twin is not merely a visual element, but a data-driven model designed to support clinical interpretation. Dedicated dashboards allow physicians to identify meaningful changes over time and assess whether further evaluation or intervention may be required.

This approach simplifies the interpretation of longitudinal data and supports more informed decision-making.

Cloud-native design for sensitive healthcare data

Given the sensitivity of healthcare data, the system architecture was designed with strong emphasis on security, environment separation, and role-based access control.

The infrastructure was built on AWS, with distinct environments for development, staging, and production. Infrastructure as Code was implemented using Terraform to ensure reproducibility and control.

The platform defines differentiated roles (physician, parent, administrator), each with specific permissions and traceability of relevant operations. Data handling processes were structured to ensure confidentiality, integrity, and availability.

In healthcare software, these aspects are not secondary features but fundamental design requirements.

Supporting clinical practice with structured data

PoDiaCar reduces the time spent on manual calculations, improves accuracy in visit documentation, and provides a structured view of cardiometabolic risk over time.

The platform does not replace clinical judgment. Instead, it strengthens it by providing consistent, automated analysis that supports long-term monitoring and early detection of risk factors.