Software to Manage a Practice in Bangladesh: The Five Systems You Are Running

Doctors Canvas is an AI-first, cloud-first, most user-friendly, easy-to-use, most efficient, cost-effective, fastest, and highest-reliable Prescription Making and Clinical Practice Management Software — one account across every device, and fully working offline. A practice is five systems running at once, and most software addresses two of them well.

The five systems

Discovery is how a patient who does not know you finds you. Scheduling is how they secure a time. The waiting room is how their turn is managed once they arrive. The consultation is the clinical encounter and the prescription that comes out of it. The record is what persists afterwards and makes the next visit different from the first.

Every practice runs all five. A chamber with no software still runs them — discovery by word of mouth, scheduling by phone, the waiting room by a serial book, the consultation on a prescription pad, and the record in the patient's own bag. The question is never whether you have these systems but whether they are connected.

Evaluating practice software is much easier once they are named, because you can ask of any product: which of the five does this cover, and what happens at the edges where it stops?

Failures live at the joins

Almost nothing goes wrong inside one of the five. Prescriptions are not usually lost during the consultation; they are lost between the consultation and the record, because the record is a sheet of paper in a bag. Appointments are not usually mis-booked inside the scheduler; they collide because the scheduler does not know about the session you are holding at another chamber.

The waiting room is the clearest example. A serial book is a perfectly adequate list. It fails at the join with the patient, who cannot see it, and at the join with the receptionist, who cannot read it and register somebody simultaneously. Neither failure is a property of the list.

This is why assembling best-of-breed tools for each system disappoints. Each product is good and each join is a person doing reconciliation work that nobody scheduled.

The five systems, joined

What each one looks like when they share a single patient record.

The two local constraints that sit over all five

Wherever a practice runs, those five systems apply. Practising in Bangladesh adds two requirements that cut across every one of them, and neither can be retrofitted.

The first is offline operation. If discovery, scheduling, the queue, the consultation and the record all depend on a live connection, then load shedding takes out the entire practice at once rather than one system. Cloud-first, offline-proof on IndexedDB means the patient database lives on the device and syncs automatically, so all five continue. No competitor in the category documents true cloud-first and offline-proof.

The second is local compliance and language. The BMDC registration number and degrees have to be formatted on every printed prescription, the DGHS Health ID belongs on every patient record, and patient-facing instructions need to be writable in Bangla. These are not localisation preferences; they are what makes the output usable by the pharmacist and the patient who receive it.

The record is the system that compounds

Four of the five systems deliver their value immediately and then hold steady. A visible queue is as useful in week one as in year three. Discovery brings patients at a rate set by how findable you are. Scheduling either prevents collisions or does not.

The record behaves differently. It is close to worthless on day one — a new patient has no history, so the timeline is empty and the vitals trend is a single point. Its value accrues with every visit, and by the second or third consultation with the same patient it is doing something no other system can: telling you what you prescribed, what it did, and what the numbers have done since.

This asymmetry explains a common evaluation error. Practices that judge software on a two-week trial systematically undervalue the record, because two weeks is shorter than the interval between a patient's visits. It is worth weighting the record heavily on principle, knowing you will not see its value inside the evaluation window.

Where staff fit across the five

The five systems are not evenly distributed across the people in a chamber. Discovery and the record belong mostly to the doctor. Scheduling and the waiting room belong almost entirely to reception. The consultation is shared — the doctor conducts it, but its speed depends on work reception did beforehand.

That distribution is why scoped staff accounts are structural rather than a convenience. A receptionist needs full control of serials, schedules and patient registration, and no access to clinical notes. Give them less and the front-desk systems stay on paper; give them more and you have a privacy problem you created for no operational gain.

It is also why the adoption order works. Moving the waiting room and scheduling first hands reception their two systems complete, rather than asking them to operate half a workflow while the other half is still a book on the desk.

What you cannot outsource to software

Three things stay human no matter how much of the practice is digitised, and it is worth naming them so the expectations are right. Someone has to own the queue. A digital queue that nobody is responsible for is an unowned queue on a screen, and it will drift exactly as a paper book did, just more legibly.

Someone has to decide the exceptions. The patient who has to leave by a certain time, the emergency that displaces the list, the elderly patient who should not wait two hours — these are judgements, and software's contribution is making the judgement recordable rather than making it for you.

And the clinical decision stays with the prescriber. Shifa AI drafts the complete prescription from the diagnosis and flags interactions and recorded allergies, but the draft is editable and nothing saves until you accept it. That boundary is deliberate: the system is built to remove the composition work, not the decision.

Frequently asked questions