Selected work · 05
Infant Guard
My BCA major project: a prototype web app for coordinating newborn care between parents, doctors and hospital staff, with health records, vaccination scheduling, alerts and doctor-approved nutrition plans.
Python · Flask · PostgreSQL · HTML / CSS / JavaScript · SMTP email
Context
Infant Guard was the major project for my BCA degree, built by a team of three. It was designed around the needs of a maternity hospital, but it's a prototype: it was never deployed or used with real patients, and nothing here is a clinical claim.
The problem it explores is real enough: in many smaller hospitals newborn records live on paper and spreadsheets, so vaccinations and follow-ups get missed and nutrition risk is judged informally.
Who can do what
| Capability | Admin / staff | Doctor | Parent |
|---|---|---|---|
| Register infants and parents | Yes | - | - |
| Record growth and medical history | - | Yes | View |
| Approve or edit nutrition plans | - | Yes | View approved |
| Appointments | - | Manage availability | Book |
| Reports and system records | Manage | - | - |
| Alerts and reminders | - | Receive | Receive |
Role permissions in the prototype.
What the prototype implements
- Health records: birth details, growth measurements, medications, illness history and vaccinations, in a normalised PostgreSQL schema.
- Role-based access: separate logins and dashboards for admin, doctors and parents.
- Vaccination scheduling: generates a schedule from birth to age five, with reminders before each due date.
- Alerts: email over SMTP, mirrored on each role's dashboard, for vaccinations, appointments, plans and risk flags.
- Risk flag: combines growth measurements, illness history and nutrition assessment into a score; crossing a configured threshold notifies the doctor and parent.
- Nutrition plan drafts: a suggested plan is generated from the infant's profile, and a doctor must approve or edit it before a parent sees it.
Limits
- The risk thresholds and plan suggestions were set for the project and haven't been clinically validated.
- It was tested with sample data only, and was assessed as coursework, not in a hospital.