Key takeaways
- A single practice-wide no-show rate hides the slots that drive most of the waste; break it down before you act.
- Five cuts cover most of what you need: day of week, time block, provider, visit type and lead time.
- Ignore differences built on small numbers; look for patterns that hold for at least two or three months.
- Change the template where a pattern is strong, then check the same cut a month later to confirm the change worked.
Why one number is not enough
Most practices know their overall no-show rate. Fewer know that it might be twice as high on Monday mornings as on Thursday afternoons, or that one visit type accounts for a third of all misses. The overall figure tells you there is a problem. The breakdown tells you where to fix it.
Rates vary widely even inside one organization. In a study of ten main clinics in one health system, the mean no-show rate was 18.8%, and the highest rates were in subspecialty clinics[1]. The same spread usually exists between providers, sessions and appointment types within a single practice.
This matters more as no-shows rise. In MGMA's 2025 poll, 27% of practices reported higher no-show rates than the year before[2]. A practice that knows exactly which slots are failing can respond with a targeted template change instead of a blanket policy.
Step 1: pull the right report
You need one row per scheduled appointment for at least the last six months, ideally twelve. Most practice management systems can export this as a standard appointment report. Ask for these fields:
- Appointment date and start time.
- Provider and location.
- Appointment type, and whether the patient is new or established.
- Date the appointment was booked, so you can calculate lead time.
- Final status: kept, cancelled with notice, late cancellation, no-show, rescheduled.
- Reminder status, if your reminder system writes it back: sent, delivered, confirmed.
- A patient identifier, so you can count repeat misses without names leaving the system.
Keep the export inside your secure environment. If you analyze it in a spreadsheet, store it on a practice-managed drive, remove names and contact details first, and delete working copies when you are done.
Step 2: clean the data before counting
- Remove appointments cancelled by the practice, such as provider leave or weather closures. They are not patient no-shows.
- Decide how to treat late cancellations. Count them separately, because they need different fixes from true no-shows.
- Collapse visit types into a short list, such as new patient, annual or preventive, follow-up, procedure, acute and telehealth.
- Calculate lead time in days and group it into bands: same or next day, two to seven days, eight to thirty days and more than thirty days.
- Group start times into blocks: first hour, rest of morning, midday, early afternoon and last hour.
Define the rate the same way every time: no-shows divided by kept visits plus no-shows, excluding practice cancellations. Write the definition at the top of the report so nobody compares numbers built differently.
Step 3: build the five cuts
Day of week
No-show rate for each weekday. Look for one day that stands out, often the start or end of the week, or the day after a holiday.
Time block
Rate for each block. Early and late slots often behave differently from mid-morning, and the direction depends on your patients' work patterns.
Provider
Rate for each provider, alongside their mix of visit types and lead times. A high rate may reflect the panel or the booking pattern, not the clinician.
Visit type
Rate for each collapsed type. New patients and long-lead follow-ups are common problem areas; acute visits booked the same day rarely are.
Lead time
Rate for each lead-time band. In a study of a New York community health center network, lead time was among the strongest predictors of no-shows[3], and the same pattern shows up in many practices.
Then cross the two strongest cuts. If lead time and visit type both stand out, a small table of visit type by lead-time band usually shows exactly which bookings to change.
Step 4: read the patterns honestly
Small numbers create false alarms. A slot type with 20 appointments a month can swing widely from chance alone. Use a few simple rules before calling something a pattern:
- Only compare groups with enough volume, at least 100 appointments over the period, before drawing conclusions.
- Look for differences that hold for at least two or three consecutive months.
- Compare each group with the practice average, not with the single best group.
- Check whether a provider or day pattern is really a lead-time or visit-type pattern in disguise.
| Session | Visits | No-shows | Rate per 100 visits |
|---|---|---|---|
| Monday first hour | 160 | 29 | 18 |
| Monday rest of morning | 410 | 45 | 11 |
| Wednesday early afternoon | 380 | 30 | 8 |
| Friday last hour | 150 | 24 | 16 |
| All sessions | 4,200 | 420 | 10 |
In this fictional Example, the first hour on Mondays and the last hour on Fridays both stand well above the practice average on reasonable volumes. The next question is what kinds of visits are booked into those slots, and how far ahead.
Step 5: change the template where it matters
Match the change to the cause. Typical patterns and responses:
- One session is consistently high and full of long-lead follow-ups: move those follow-ups to a recall list and fill the session with shorter-lead bookings.
- Early slots are high: reserve them for established patients who have chosen them, and stop assigning them by default to new patients.
- A provider is high because of a panel with many barriers: add live confirmation calls and transport checks for that panel instead of overbooking.
- New-patient slots are high at long lead times: hold some new-patient slots for short-notice booking and keep a cancellation list.
- A slot is reliably high after all other fixes: consider a limited, clinician-approved overbook for that slot only.
Change one thing per pattern and note the date. A month later, look at the same cut. If the rate has not moved, undo the change and try the next idea.
Do not forget the calendar
Months behave differently. Flu season in the United States usually runs through the fall and winter, and most of the time activity peaks between December and February[4]. Across 40 seasons, the peak most often came in February[4]. Illness drives same-day cancellations and no-shows for routine visits, while acute demand rises.
Source: Centers for Disease Control and Prevention, 1982-2024[4]
Show the data
| Item | Value |
|---|---|
| December | 7 seasons |
| January | 6 seasons |
| February | 17 seasons |
| March | 6 seasons |
Compare each month with the same month last year, not with the month before. Plan winter templates with more same-day acute slots and fewer far-ahead routine bookings, and expect higher cancellation counts in the peak weeks.
Give late cancellations their own view
A late cancellation, made within a day of the visit, leaves almost as large a hole as a no-show, but the patient has told you something. Run the same five cuts for late cancellations separately. They often point to different fixes: a spike in late cancellations on Monday mornings suggests weekend illness or work rosters published late, while a spike in no-shows in the same slots suggests reminders are not landing. Pair each late cancellation with its reason code where you have one, and track how often the freed slot was refilled. A slot that was cancelled late but refilled from the waitlist is a process working well.
Make the report run itself
- Save the export as a scheduled report that runs on the first business day of each month, covering the previous twelve months.
- Keep a template spreadsheet or dashboard with the five cuts already built, so the person running it only refreshes the data.
- Add a short notes column for changes made each month, such as "moved Monday 8:00 slots to established patients", so later readers can link changes to results.
- Share a one-page summary with the scheduling lead, the practice manager and the clinical lead before the monthly operations meeting.
Talking to providers about their numbers
Provider-level numbers can feel like a performance score. Share them as information about the schedule, not as judgment. Show each provider their rate alongside their mix of visit types and lead times, and ask what they see in it. Clinicians often know why a session fails, such as a clinic that runs late every Friday, and they are the ones who can agree to change follow-up intervals or overbooking rules.
Use what you find to target outreach for high-risk no-show patients, shape no-show letters and plan family-friendly sessions for pediatric no-shows.
Frequently asked questions
How do I calculate a no-show rate by provider?
Divide each provider's no-shows by their kept visits plus no-shows over the same period, excluding practice cancellations. Show it alongside their visit-type and lead-time mix so the number is read in context.
How much data do I need?
At least six months, ideally twelve, and at least about 100 appointments in any group you compare. Smaller groups swing too much to trust.
Which day has the most no-shows?
It varies by practice and patient population. Pull your own report: the pattern is often tied to lead time and visit type more than to the day itself.
Should we overbook the worst slots?
Only after trying other fixes, only in slots with a reliable pattern and only with clinician agreement and a cap. Overbooking without those limits creates long waits when everyone shows up.
References
BMC Health Services Research (via PubMed Central) (2016). Prevalence, predictors and economic consequences of no-shows (opens in a new tab) โ
Medical Group Management Association (MGMA) (2025). Patient no-shows in 2025: What's changing and what to do about it (opens in a new tab) โ
BMC Health Services Research (via PubMed Central) (2020). Nudging New York: adaptive models and the limits of behavioral interventions to reduce no-shows and health inequalities (opens in a new tab) โ
Centers for Disease Control and Prevention (1982-2024). Flu Season (opens in a new tab) โ