
Building a Doctor-First Preconception Care Assessment System with AWS and Ollama
A doctor-facing platform that structures preconception assessments, provides real-time analytics, preserves patient history, and uses local AI to transform structured clinical data into readable documentation.
Building a Doctor-First Preconception Care Assessment System with AWS and Ollama
Introduction
Preconception care involves collecting a wide range of information about a patient's medical and obstetric history before pregnancy. During a consultation, doctors need to ask many questions and organize the responses while simultaneously focusing on the patient.
We built a doctor-first Preconception Care Assessment and Clinical Documentation System to make this workflow more structured and easier to review.
The key principle behind our system is simple:
The doctor remains in control of the consultation. The system helps organize, analyze, store, and document the information.
How the System Works
Our application is exclusively designed for doctors.
The patient does not log in, answer questions through a chatbot, or interact with the application.
Instead, the doctor asks the patient the questions verbally and records the responses in the system using:
- Yes
- No
- Not Sure
The assessment is organized into clinical sections such as marital history, obstetric history, and other relevant sections.
Each assessment section has one optional Notes field. This allows the doctor to record additional information related to the entire section rather than attaching notes to individual questions.
From Assessment to Analytics
After completing the assessment, the doctor submits it.
The system stores:
- Patient details
- Assessment responses
- Section-level notes
- Assessment date and status
The system then generates structured JSON from the collected information.
The doctor is immediately taken to the Dashboard Analytics.
The dashboard is dynamically generated from the actual responses entered during the consultation.
It provides:
- Total questions answered
- Yes responses
- No responses
- Not Sure responses
- Section-wise analytics
- Doctor-entered notes
- A summary of the collected responses
This gives the doctor a quick overview of the completed assessment without having to manually go through every response again.
Persistent Patient History
Completed assessments are associated with the patient's record.
When the patient returns, the doctor can review previous assessments, responses, notes, analytics, and assessment dates.
Previous assessments are preserved rather than overwritten, allowing the doctor to maintain a history of consultations.
Structured JSON
One of the important parts of our architecture is converting the assessment into structured JSON.
A simplified representation looks like this:
{
"patient": {
"name": "...",
"phone": "...",
"address": "..."
},
"assessment": {
"date": "...",
"status": "completed"
},
"sections": [
{
"name": "Obstetric History",
"answers": [
{
"question": "Previous pregnancy?",
"answer": "YES"
}
],
"notes": "Previous pregnancy had complications."
}
]
}The JSON acts as a structured source of truth for the assessment.
Local AI with Ollama
We use Ollama as the local AI layer for documentation.
Importantly, Ollama does not conduct the assessment.
It does not:
- Ask the patient questions
- Generate assessment questions
- Change the doctor's answers
- Diagnose the patient
- Decide treatment
Instead, after the assessment has been completed, the structured JSON can be passed to the local model to generate a readable doctor-facing document.
This creates a clear separation between data collection, analytics, and AI-assisted documentation.
AWS in Our Architecture
AWS provides the cloud infrastructure layer for our application.
We use AWS services to support the application's backend, data handling, deployment, and scalable infrastructure.
This separation allows us to keep the application workflow structured while using cloud infrastructure for the parts of the system that need reliable and scalable deployment.
The architecture can be represented as:
Doctor
↓
Clinical Assessment
↓
Structured Responses
↓
Backend / Data Layer
↓
AWS Infrastructure
↓
JSON
↓
Analytics Dashboard
↓
Ollama
↓
Doctor-Readable Documentation
Why This Approach?
We deliberately designed the system as a doctor-first workflow rather than a patient-facing AI assistant.
The doctor remains responsible for asking the questions and interpreting the information.
Our system focuses on reducing the documentation burden and making the collected information easier to organize and review.
The combination of structured clinical data, analytics, persistent patient history, AWS infrastructure, and local AI documentation provides a foundation for a more efficient preconception-care workflow.
Conclusion
Our goal is not to replace clinical decision-making.
Our goal is to help doctors spend less effort organizing information and more time focusing on the patient.
By combining a structured assessment workflow, dynamic analytics, persistent patient records, JSON-based data, AWS infrastructure, and Ollama-powered documentation, we created a system that turns a consultation into an organized and reviewable digital record.
The doctor talks to the patient.
The doctor makes the decisions.
Our system handles the organization and documentation.
Enjoyed reading this content? Let the author know!
Your likes, comments, shares, and saves help creators reach more builders.
Loading recommendations
Loading article