Medibridge
Healthcare management system
2025
PROBLEM STATEMENT
Medical records are still commonly stored as paper files, scanned documents, PDFs, or images. This makes it difficult for hospitals and clinics to quickly access, organize, and transfer accurate patient information, especially during inter-facility transfers. Healthcare staff often spend time manually reviewing documents, extracting information, and handling fragmented records across different systems. The lack of structured and interoperable medical data slows down workflows and affects continuity of care.
OBJECTIVE
Design a platform that helps hospitals and clinics digitize medical records, extract structured patient information from unstructured files, and streamline patient record transfers between healthcare facilities. The goal is to reduce administrative workload, improve accessibility of patient records, and create a faster and more reliable transfer process with review and patient approval workflows.
SOLUTION OVERVIEW
Medibridge is an AI-powered healthcare platform designed to convert physical and unstructured medical records into structured digital patient data. The platform allows healthcare staff to upload scanned documents, PDFs, and images, automatically extract patient information using AI, review the extracted data, and manage patient records transfers between hospitals through a centralized workflow. The system also includes transfer approval flows, record sharing, and structured patient profiles to improve record management and continuity of care.
ROLE
Product Engineer
DURATION
Ongoing
Picture this: a patient is rushed from a clinic in Surulere to a specialist hospital in Victoria Island. She’s unconscious. Her family does not know her medication history. The only record that travels with her is a folded piece of paper, handwritten and partially smudged, tucked inside a Ghana-must-go bag. The receiving doctor has maybe ten minutes to make decisions that matter. She’s starting with almost nothing. This is not hypothetical. This happens every day.
I started paying closer attention to how patient records move between hospitals. Files stacked in cardboard boxes. Discharge summaries photographed on someone’s Tecno phone and sent over WhatsApp. Lab results saved as scanned images that doctors have to zoom into just to read clearly.
I looked at platforms like Epic Systems and Oracle Health. They are powerful systems, but they were built for hospitals with strong IT infrastructure, large budgets, and fully digital workflows already in place. That is not the reality for many hospitals in Nigeria. Even then, a lot of medical data still has to be typed in manually. The challenge is not just storing records. It is the amount of human effort needed to organize them in the first place.
That is where Medibridge begins. With a simple question: what if we worked with the records hospitals already have? The PDFs. The scanned handwritten notes. The photographed lab results. The discharge summaries written in different formats across different clinics. What if AI could read those records, understand them, and turn them into something doctors can quickly search, review, and transfer? Medibridge is designed to fit into the way hospitals already work instead of asking them to start over. The goal is simple: help the next doctor see the full picture before making critical decisions. Because somewhere right now, a doctor is treating a patient without the information they need. That gap can be closed.
Patient Onboarding
One of the first problems I tried to solve was reducing the amount of manual work involved in adding patients into the system. Many hospitals already have patient information stored as printed documents, scanned files, PDFs, or images. Re-entering this information manually can be repetitive and time-consuming, especially when handling large numbers of records. I wanted the onboarding process to feel faster and more structured, so I built an AI-assisted flow that allows healthcare staff to upload existing medical documents and extract patient’s personal information automatically. The flow supports different file formats and guides staff through reviewing extracted information before saving it into the patient profile.
Patient onboarding is often the first interaction hospital staff have with a new record, so the experience needed to feel simple and reliable. I limited the initial extraction to essential information such as the patient’s name, age, gender, contact details, and identifiers. Keeping the scope focused makes the results easier to review and reduces the risk of incorrect information entering the patient’s history.
AI-extracted information is never saved automatically. Every field remains editable, allowing staff to correct errors and complete missing details before confirming the patient. The upload flow also supports the document formats hospitals already use, including PDFs, scanned files, images, and Word documents, so it can fit into existing workflows without requiring hospitals to change how their records are prepared.
Patient Management
After onboarding patients into the system, I designed the day-to-day experience of managing patient records inside the hospital workflow. Medical information changes frequently during treatment, so the interface needed to support quick updates while keeping records easy to review during consultations, admissions, transfers, and follow-up care.
Healthcare staff can manage patient information directly from a centralized patient profile with dedicated sections for medications, allergies, diagnoses, procedures, immunizations, laboratory results, and medical history. Grouping information into sections helps staff quickly locate relevant details, review patient history, and update records without moving through multiple pages or relying heavily on uploaded documents.
I also designed the system to support ongoing updates as patient care continues. New medications can be added, diagnoses reviewed, allergies updated, and laboratory results recorded directly within the patient profile. Uploaded files and attachments remain available when needed, while extracted and manually entered information becomes easier to maintain as part of the patient record over time.
The goal was to create a record management experience that feels practical during real hospital workflows where information needs to be reviewed quickly, updated regularly, and carried across different stages of patient care.
Once the information was divided into clinical sections, the next challenge was keeping a growing patient record easy to navigate. Each area uses list and detail views, allowing staff to scan existing records before opening a specific entry for more information. Several record types also include history views, making it easier to understand how individual details have changed over time.
Source documents and attachments remain available when additional context is needed, but staff do not have to search through every file during routine care. Confirmed patient information is presented in a form that is easier to review, while the original documents remain connected to the record. This creates a balance between quick access and the supporting information behind each patient’s history.
Patient Records Transfer
Another area I focused on was improving how patient records are transferred between hospitals. Transfers often rely on email, printed documents, messaging apps, or other disconnected channels, making records difficult to track, review, and verify. Medibridge brings this process into a structured workflow that begins from an existing patient profile.
Healthcare staff can prepare a transfer using verified records and provide any necessary context before sending it for patient approval. Once approved, the selected records are made available to the receiving hospital through a secure, time-limited, read-only link instead of sending raw files through disconnected channels.
The most important part of the transfer experience was keeping the patient involved. A hospital cannot release the selected records immediately after creating a request. The patient first receives an email containing a secure approval link, where they can review the receiving hospital, the requested records, and the transfer details before approving or declining.
Approval does not give the receiving hospital permanent access to the patient’s full history. The recipient must verify a short-lived code before opening an expiring, read-only view limited to the approved records. Transfer statuses also help the initiating hospital see whether a request is pending, approved, declined, or completed.
Medibridge is in active development. Follow along in the GitHub repository.
Next
Nomba