OPD queue management: tokens, display screens or live wait times?

Walk into almost any OPD at ten in the morning and you’ll see the same thing: a full hall, a crowd at the doctor’s door and someone at the counter asking “मेरा नंबर कब आएगा?” (when will my turn come?). Hospitals are offered paper tokens, display screens that call numbers, SMS updates, apps and appointment systems, and many are sold as if they shorten the wait. Some do. Others only make the same wait look tidier.

This guide sets out which is which, and how to give patients an estimate of their wait that you can stand behind.

Why does the OPD queue build up?

A queue grows when patients arrive faster than the doctor can see them. Nothing on a screen changes that sum.

The clearest Indian evidence comes from AIIMS New Delhi, where as many as 10,000 patients reach the OPD every day. Its computer facility found that 79% of patients reported before 10 am, but only 19% of them had been seen by 10 am. Patients were arriving out of step with the times they could be seen, and AIIMS concluded that this clogged the OPD and pushed up the average wait. It didn’t answer that with a bigger display. It gave each patient a time slot.

A 2024 study of four OPDs at an Indian medical college found the same pattern: in every OPD, patients arrived faster than the doctors could see them, and in orthopaedics, waiting to see the doctor took 85.71% of patients’ total time.

Here is what published studies found. All five are teaching hospitals, most of them government-run, where crowds are often bigger than in private hospitals. Still, this is the experience patients bring with them to every hospital.

The last row is a reminder that most hospitals have two queues: the registration counter and the doctor’s door. The government says its ABDM Scan and Share service, where patients scan a QR code to get a digital queue token, cut patients’ wait at registration from about an hour to 2 to 5 minutes, citing a study by the Indian Institute of Health Management Research. That fixes the first queue. The second one only gets shorter when arrivals are spread out.

Tokens, screens, messages, live waits and time windows compared

Do token display screens reduce waiting time?

Not on their own. They make the wait orderly, and that is worth having: a token ends arguments at the door, protects the patient who came first and, with a voice caller in Hindi and English, helps patients who can’t read the screen. In a big hall, that matters.

What a token can’t tell anyone is when. “Token 47” says nothing about whether a patient will sit for ten minutes or two hours, so everyone stays in the hall, and everyone still comes at nine. If you install a screen:

• It is only as true as the last button press. If “next” isn’t pressed for every patient, the screen goes wrong and patients stop believing it.
• Queue-jumping becomes visible. Unfair waits feel longer than fair ones, as David Maister argued in his classic essay on the psychology of waiting lines. Patients in a study at a government medical college in Nanded named VIP patients jumping the queue among the causes of their long waits. Decide who goes first (emergencies, very elderly patients, those only showing a report), put it on a notice, and stick to it.

SMS and WhatsApp updates: only with a yes

Messages let patients step out to the canteen or the pharmacy and come back in time, and they reach patients who haven’t left home yet. Before you send any:

1. Ask first, and record the answer. Consent must be clear and informed, and a patient can withdraw it at any time, as the government’s summary of the DPDP Rules, 2025 puts it. Ask at booking, keep the yes, and stop when a patient says stop.
2. Register SMS templates in advance. Business SMS in India go out only under a registered sender header and a registered template with fixed and variable parts, under TRAI’s rules. “Dr {name} is running {minutes} minutes late” has to be approved before the morning you need it.
3. Send a time, not a position. “You are 5th” means ten minutes with one doctor and an hour with another.

Reminders matter too. In a 2026 study of patients who missed appointments at an Indian tertiary hospital, forgetting and not getting a reminder were among the leading reasons.

How to work out a patient’s waiting time fairly

A fair estimate rests on two things: the doctor’s real pace today, and the number of patients ahead who are actually there.

1. Use the doctor’s pace, not the slot length

A doctor booked in 10-minute slots may see a patient every 6 minutes, or every 14. In the studies above, average consultations ran from 6.5 to 12.8 minutes. Take the minutes per patient over the doctor’s last few patients today, and keep updating it.

2. Count the patients ahead who are here

Count those who have arrived and are waiting, then add booked patients ahead who are still expected, but not as certainties: some won’t come. Across 105 studies worldwide, about 23% of booked patients didn’t turn up; your own figure is the one to use. AIIMS, working with the Indian Statistical Institute, used its past turn-up data to predict how many patients would actually come to each department, to within 10%.

3. Multiply, then give a range

An illustration with made-up names: Dr Rao started at 10:00 and has seen 9 patients by 10:54, so her pace today is 6 minutes a patient. Sunita has 7 patients ahead of her in the hall: 7 × 6 = 42 minutes. But this morning Dr Rao’s consultations have taken anywhere from 4 to 9 minutes, so tell Sunita “about 30 to 60 minutes” (7 × 4 = 28 at the quick end, 7 × 9 = 63 at the slow end, rounded).

4. Update it, and say when

Recalculate each time a patient goes in, and at once if the doctor is called away. Show “updated 10:54” beside the estimate so patients can trust it. Before the day of the visit you can’t know how the session will run, so give booked patients a window based on how full it is, and switch to the live estimate on the day.

Why a range, and not “42 minutes”? Consultations vary and emergencies happen, so a single figure is usually wrong, and every missed figure is a broken promise. A range that you keep updating gives patients a wait that is known and finite, and Maister’s point was that uncertain waits feel longer than known, finite ones. In a US emergency-department study of 1,631 patients, satisfaction followed whether the wait was shorter than they expected and how well delays were explained, not the minutes they actually waited.

Appointment windows: the fix for the 9 am crowd

Of the five options, time windows do the most to change when patients arrive, and that decides how long they sit in the hall. AIIMS staggered its morning OPD into three reporting slots (8 to 9, 9 to 10 and 10 to 10:30), sized to each department’s usual numbers, and printed the reporting time on the appointment slip. The need is the same outside Delhi: in the Uttarakhand radiation oncology OPD above, two in three patients (66.7%) arrived between 8:30 and 10:30 am.

A simple way to set windows (an illustration, not a rule):

1. Measure how many patients each doctor actually sees in an hour; say 8.
2. Keep places for walk-ins and urgent cases, based on how many you usually get; say 2 an hour.
3. Book the rest in half-hour windows: 3 patients for 10:00 to 10:30, 3 for 10:30 to 11:00, and so on.
4. Tell every patient their window when they book, on the phone and on the slip. Give walk-ins a token and an expected time.
5. Start the first window when the doctor really starts. If ward rounds end at 10:15, the first window is 10:15.

Some patients will still come early at first. Keeping the times you give is what teaches them to trust the window.

What to count for a week before you buy anything

A week of tallies at the counter will tell you which of these tools you need:

• arrivals per half hour at each doctor’s OPD;
• when each doctor actually starts and finishes;
• minutes per patient, for each doctor;
• the share of booked patients who don’t come;
• walk-ins against booked patients;
• where the longest wait is: registration, the doctor, billing or the pharmacy. In the North Indian study, the pharmacy stage was the second-biggest bottleneck, at 119.4 minutes.

If the crowd is at registration, fix registration. If it’s at the doctor’s door at 10 am, a screen will tidy it, but only windows will shrink it.

Where DrMate fits

DrMate’s front desk puts every doctor’s day on one screen and gives each patient a private page, with no app, showing their doctor, time or window, queue number and expected wait. Before the day the wait is a forecast shown as a range; on the day it comes from the doctor’s actual pace and the patient’s place among those who have arrived. When a doctor runs late, reception types the minutes, every later patient’s time moves, and patients who said yes to WhatsApp are told. On the phone, the AI receptionist books each caller into a real free time and tells them when to come.

For the words to use when the doctor is delayed, see what to tell patients when the doctor is running late, and for the phone desk, our hospital call centre script.