Key takeaways
- Retention needs a written definition of an active patient, usually a visit within a set window that fits your specialty.
- Calculate it as patients active at the start of a period who visit again during the period, divided by patients active at the start.
- Cohort tracking is more revealing: group new patients by the month of their first visit and track how many return within 6 and 12 months.
- Remove patients who moved, died or transferred care on purpose before calculating, and report them separately.
Why retention needs a careful definition
Patient retention is the share of a practice’s patients who keep coming back. It sounds simple, but a practice can produce very different numbers from the same data depending on how it defines an active patient, which period it measures and which patients it removes. A retention rate that changes because the definition changed is worse than none.
Retention matters because most visits come from patients the practice already knows. In 2019, established patients accounted for 83.2% of US physician office visits, and new patients 16.8%[1]. Every established patient who drifts away has to be replaced by a new patient, which costs more front office time and marketing than keeping the first one.
Patients have more alternatives than ever. In 2024, 27.6% of people in the US had at least one urgent care center visit in the past 12 months, and 19.0% had at least one retail health clinic visit[2]. Among adults aged 18 to 34, only 67.7% named a doctor’s office or health center as their usual source of care, compared with 88.8% of adults 65 and older[3].
Source: CDC National Center for Health Statistics, 2024[2]
Show the data
| Item | Value |
|---|---|
| Urgent care center | 27.6% |
| Retail health clinic | 19% |
Step one: define an active patient
An active patient is one who has had a completed visit within a set window. Choose a window that fits how often your patients should be seen. A primary care or pediatric practice might use 18 or 24 months, because most patients should have at least an annual visit. A specialty practice that sees patients for a course of treatment might use a shorter window for retention during treatment and a separate measure for return after discharge.
Write down three more rules. First, which visits count: completed visits only, or also telehealth and nurse visits. Second, which patients are excluded: deceased patients, patients who moved away, patients discharged from the practice and patients who formally transferred care. Third, how household or family accounts are treated in pediatrics. Keep the rules fixed for at least a year.
Step two: calculate the rate
Take a starting list
List all active patients on the first day of the period, using your definition. This is the denominator.
Remove known exits
Remove patients who died, moved out of the area, transferred care or were discharged during the period, and count them separately.
Count returns
Count how many of the remaining patients had at least one completed visit during the period. This is the numerator.
Divide and record
Retention rate equals returns divided by the starting list after exits. Record the definition and the exit count next to the rate.
Repeat on the same day each year
Compare like with like. A rate calculated in January will differ from one calculated in July if visits are seasonal.
Example: a fictional practice has 6,000 active patients on January 1. During the year, 240 are removed as known exits. Of the remaining 5,760, 4,550 have a completed visit by December 31. The annual retention rate is 4,550 divided by 5,760, about 79%.
Step three: track new-patient cohorts
The overall rate moves slowly and mixes long-standing patients with recent arrivals. Cohorts show more. Group new patients by the month of their first visit, then track what share have a second visit within 6 months and within 12 months. A practice that keeps long-standing patients well but loses many new patients after one visit has a first-impression problem, not a loyalty problem.
| Cohort (first visit) | New patients | Returned within 6 months | Returned within 12 months |
|---|---|---|---|
| January | 120 | 62% | 74% |
| February | 105 | 60% | 71% |
| March | 131 | 48% | 63% |
| April | 118 | 61% | 73% |
Example: in the table, the March cohort returned at a noticeably lower rate. The office manager checks what was different in March and finds that a provider left that month, and many of the patients they had seen as new were not offered a follow-up with another clinician. A cohort view surfaces that kind of problem in months; an annual rate might hide it for a year.
Where patients drift, and what the front desk can do
- No follow-up booked at check-out. Book the next visit before the patient leaves, or put them on a recall list with a date if the schedule is not open yet.
- Recall lists that are never worked. Assign an owner and a weekly time to work the recall list, and track how many recalled patients book.
- A first visit that went badly. Long waits, billing surprises or a rushed check-in leave new patients unlikely to return. Look at new-patient feedback and complaints by cohort.
- Hard to reach the practice. If the phones are hard to get through, patients book the next visit somewhere easier.
- Coverage changes. Patients who change plans may believe the practice no longer accepts them. Ask at every contact whether insurance has changed, and tell them if you are still in network.
- Provider departures. When a clinician leaves, contact their patients with a named alternative clinician and an easy way to book.
Reputation plays a part too. In a Kansas survey of 284 patients, 57% reported reading online reviews of their physician[4]. Patients who leave unhappy often say so publicly, which affects new-patient demand as well as retention.
A sanity check from national visit rates
National survey data gives a rough check on whether a retention rate is plausible. In 2025, 84.7% of US adults and 94.7% of children had a visit with a doctor or other health care professional in the past year[5]. Those figures cover all providers, not one practice, so they are not a target. But if a primary care practice finds that well under half of its active adult patients had any visit in a year, the active-patient definition may be too loose, or the practice may be losing patients to other sources of care.
Pediatric practices can make the same check with the children’s figure. A pediatric practice with a much lower annual return rate than the national visit rate should look first at well-child recall: whether reminders go out, whether they reach the right parent, and whether the schedule has room for them when they are booked.
A recall call that works
Recall script: “Hi, this is [practice name] calling for Sam. Dr. Rivera asked us to check in because Sam is due for a follow-up visit. I have Tuesday at 10:20 or Thursday at 3:40 next week. Would either of those work? If now isn’t a good time, I can text you a link to our scheduling line.”
Offer specific times, name the clinician who asked for the visit and give an easy alternative. Record the outcome of every recall attempt, booked, declined, no answer or moved away, so the next cohort report can tell a recall problem from a demand problem.
Common calculation mistakes
- Counting every patient ever registered as the denominator. Old records of people who left years ago make retention look terrible and hide recent changes.
- Not removing known exits. Patients who moved away are not a retention failure; report them separately.
- Changing the active-patient window between years. Keep it fixed, or recalculate past years with the new window.
- Mixing visit types. Decide once whether nurse visits, telehealth and procedure-only visits count.
- Reporting a single practice-wide rate. Split by provider, age band and payer to see where the drift is.
Make retention part of the routine
Calculate the annual rate once a year on a fixed date, and the cohort table monthly. Review the cohort table at the monthly management meeting and the recall list results weekly at the front desk huddle. When a cohort returns at a lower rate, ask what changed that month before changing anything else.
Retention links to schedule utilization, revenue per visit and how a practice compares itself using MGMA benchmarks.
Frequently asked questions
How do you calculate patient retention rate?
Take patients active at the start of a period, remove known exits such as moves and deaths, then divide the number who had a completed visit during the period by the adjusted starting number.
What counts as an active patient?
A patient with a completed visit within a window you define, such as 18 or 24 months for primary care. Write the definition down and keep it fixed.
What is cohort retention?
Grouping new patients by the month of their first visit and tracking what share return within set periods, such as 6 and 12 months.
Can the front desk improve retention?
Yes: by booking follow-ups at check-out, working recall lists, keeping the phones reachable and contacting patients when their clinician leaves.
References
CDC National Center for Health Statistics (2019). National Ambulatory Medical Care Survey: 2019 National Summary Tables (opens in a new tab) ↑
CDC National Center for Health Statistics (2024). Urgent Care Center and Retail Health Clinic Use: United States, 2024 (NCHS Data Brief No. 562) (opens in a new tab) ↑
CDC National Center for Health Statistics (2024). Source of Usual Health Care for Adults Age 18 and Older: United States, 2024 (NCHS Data Brief No. 558) (opens in a new tab) ↑
Kansas Journal of Medicine (via PubMed Central) (2022). Millennials Seeking Healthcare: Examining the Degree to Which Patients Utilize Online Resources (Kansas Journal of Medicine, 2022) (opens in a new tab) ↑
CDC National Center for Health Statistics (2025). FastStats - Physician office visits (opens in a new tab) ↑