Building Tech for Rural Realities: What Happens When a Clinician Codes with AI?

Duncan Miller

BSc Physiotherapy (UCT) | MPH (Imperial College London)

1 minute AI Demo

Gemini CLI

duncan@hadadahealth.com:~$ |

Prompt: Please create the front-end for a simple equipment and assistive device tracking app. This application will allow clinicians to issue wheelchairs and crutches and walking frames and any other assistive devices, as well as return them and include putting people on a waiting list if there is no wheelchair. This app will allow clinicians to keep track of all issues, returns, and orders.

There needs to be a form of login, but for now, for this demo purpose, that is not necessary. I will eventually use a Superbase backend, but for now, please create an HTML, CSS, and JavaScript front-end for me to try out.

Coding for Non-Coders

No-code tools

Glide Logo

AppSheet Logo

AI coding tools

ChatGPT Logo

Claude Code Logo

Gemini Logo

24-hour Demo

QR Code
Referral App

What I used:

Gemini CLI to program
TypeScript, React, and Next.js
Google authentication
Google Sheets

Back to the 1-minute Demo

Let's jump back to the 1-minute demo and see what the AI generated.

Practical AI-Assisted Coding Tips for Clinicians

design_services

Planning

Start with clear requirements and user stories

build

Incremental Development

Build and test features one at a time

layers

Choosing the right stack

Pick familiar technologies that AI can help with

cloud_upload

Deployment Simplicity

Use platforms that minimize infrastructure complexity

web

Frontend Frameworks

Choose well-documented frameworks with good AI support

Barriers and Solutions

group

Buy-in

Users need to see the benefit

Ordering wheelchairs worked because clinicians got what patients needed; logging 'issued' had no direct benefit.

signal_wifi_off

Connectivity

Progressive Web Apps (PWA)

We download clinic patient notes to the phone; clinicians can work offline and sync later.

data_usage

Data Usage

Quantify usage and communicate

A wheelchair order is a few KB; the PDF you printed to attend this meeting was MBs.

smartphone

Computer literacy or confidence

Cell-phone-based solutions

Design for phones first — staff use them daily even if computer skills are limited.

alt_route

Workflow

Be as flexible as needed

Allow SOAP or a big free-text box; speech-to-text works better for some clinicians.

Data Protection Responsibility

POPIA

Protection of Personal Information Act

Data stewardship is a clinical responsibility.

Conclusion

favorite

Good systems support good care

Technology should enhance patient outcomes

local_hospital

Clinicians know what is needed

Healthcare professionals understand the real problems

psychology

AI is an amazing tool we should utilise

Artificial intelligence makes development accessible to clinicians