Why OPD waiting time is rarely a doctor problem

Ask a hospital administrator why the OPD is backed up and the answer usually involves consultation time. Ask the patients queueing at 9:40am and you get a different picture: most of them are not waiting for a doctor at all. They are waiting for registration, for a file to be found, for a payment to be recorded, or for someone to tell them which floor to go to.

That distinction matters, because the two problems have completely different fixes, and pushing doctors to consult faster is both the most common response and the least effective one.

Average waiting time hides the problem

A hospital reporting a 22-minute average OPD wait can still have patients routinely waiting ninety minutes. Averages flatten exactly the peak you need to see. Two numbers are more useful:

• Wait at the 90th percentile, which tells you what your worst-served patients actually experience. This is the number that generates complaints.
• Wait by hour of day, which almost always reveals that the problem is a two-hour window, not a whole-day condition.

Most Indian OPDs have a sharp morning peak. If your intervention does not change what happens between roughly 9am and 11am, it will not change anything a patient notices.

Where the minutes actually go

Before optimising, measure the segments separately. A patient’s total time splits into at least four distinct waits, and they are rarely all broken at once:

1. Arrival to registration complete
2. Registration to being called into the consultation room
3. Consultation itself
4. Consultation to pharmacy or billing complete

In our experience with Indian hospitals, the first and fourth segments are where the avoidable time sits, and they are the two nobody measures. A patient who saw the doctor within twenty minutes but then spent forty more at the pharmacy counter remembers a slow hospital, and your consultation-time metric shows everything is fine.

Four fixes that hold up at peak

1. Move registration off the critical path

Returning patients should not queue behind new registrations. If a returning patient can be identified by phone number and pulled up in seconds, you have removed a large fraction of your morning queue from the counter entirely. This is worth doing before anything else, because it costs nothing structurally, it is purely about how records are indexed.

2. Stagger appointments against real consultation times

Booking every patient at 9am and processing them in arrival order is common and it guarantees a queue. Staggering only works if the slot length reflects what a consultation actually takes for that department, which you can only know by measuring it rather than assuming fifteen minutes across the board. A follow-up review and a first orthopaedic consultation are not the same event.

3. Tell patients where they are in the queue

A known ninety-minute wait is tolerated far better than an unknown thirty-minute one. Perceived waiting is a real variable and it is much cheaper to fix than actual waiting. A queue position sent to the patient’s phone lets them sit somewhere comfortable instead of standing near the door watching for their name, which also thins the crowd at the counter and speeds up the staff working there.

4. Stop bills being assembled at the end

If pharmacy and lab charges are collected manually at checkout, every patient pays a compounding tax on their exit. Charges that flow into the bill as they are incurred turn checkout into a payment step rather than a construction step. This is usually the single largest recoverable block of time in the whole visit.

What not to do

Two popular interventions tend to disappoint. Adding a counter at peak helps only if registration is genuinely the bottleneck, otherwise you have simply moved the queue downstream to a place with less floor space. And asking clinicians to shorten consultations trades a visible metric for an invisible cost, in repeat visits, in missed detail, and in the quality of the thing patients actually came for.

Measuring whether it worked

Fix one segment at a time and watch the 90th percentile for that segment specifically. A change that improves your overall average while leaving the morning peak untouched has not improved anything a patient experiences. Give each change at least two weeks before judging it, since OPD volume varies enough day to day that a single good Tuesday tells you nothing.

If you are evaluating software as part of this, our buyer’s guide to choosing an HMS covers what to test during a demo, and the busy-morning test is the one that matters most here.