Key takeaways
- Ask five to seven questions, each tied to one step of the visit and one owner.
- Send the survey within a day by text or email, and keep a paper or phone option.
- Report the share of top-box answers per question and by location, not one average.
- Tell patients and staff what changed because of the survey, every quarter.
What a satisfaction survey is for
A patient satisfaction survey earns its place only if someone changes something because of it. Most practices already sense the general mood. What they lack is a precise read on which step of the visit is failing, for which patients, and at what time of day. That calls for short surveys, sent soon after the visit, with questions that each point to a step someone owns.
Access questions matter most to the front desk. When MGMA asked practice leaders to name their top patient access focus for 2026, answers split between no-shows (27%), online scheduling (24%), phone access (22%) and wait times (21%)[1]. Each of those can be measured from the patient's side, and each has an owner in the office.
Patients also talk in public. In one survey of US patients, 57% said they read online reviews of their physician[2]. A private survey gives an unhappy patient a direct channel and gives you the first chance to fix the problem.
Standard instrument or your own survey
If a payer program, accountable care organization or health system requires a standardized instrument, use it. The CAHPS Clinician & Group survey, for example, includes measures on getting timely appointments, care and information, on provider communication, and on helpful, courteous and respectful office staff[3]. Standard surveys allow comparison with other practices, but they come with fixed wording, sampling rules and approved vendors. Our guide to CG-CAHPS access questions explains how those items behave.
A home-grown pulse survey is faster to change and cheaper to run. Many practices use both: the required instrument for benchmarking and a short internal survey for weekly operating decisions. Use the criteria below to decide.
| Question to ask | If yes | If no |
|---|---|---|
| Does a payer, ACO or health system require a specific survey? | Use the required instrument and vendor; do not add items that break its rules | You are free to design your own |
| Do you need to compare scores with other practices? | Standardized items through a qualified vendor | An internal pulse survey is enough |
| Will someone review results every month? | Send a short survey after every visit | Survey a quarterly sample instead; unread results waste patient time |
| Do many patients prefer a language other than English? | Translate each version and have a fluent reader test it | One version in plain English |
A question bank built around the visit
Write each question so a low score points to one step. Avoid double questions such as “Was the staff friendly and efficient?” because the answer cannot tell you which half failed. Keep one answer scale throughout. A frequency scale (never, sometimes, usually, always) or a five-point agreement scale both work if you never mix them.
| Visit step | Sample question | Owner |
|---|---|---|
| Reaching us | How easy was it to reach us by phone or online when you needed an appointment? | Phone lead |
| Scheduling | Did you get an appointment as soon as you needed one? | Scheduling lead |
| Check-in | Was check-in quick and clear? | Front desk lead |
| Waiting | If you waited more than 15 minutes, did someone tell you why? | Front desk and clinical lead |
| Courtesy | Did the front desk staff treat you with courtesy and respect? | Practice manager |
| Billing | Did you understand what you might owe before you left? | Billing lead |
| Open comment | What one thing should we change? | Practice manager |
Pick five to seven. Then rotate one module question each quarter, such as portal sign-up, parking or after-hours calls, so you can explore a new issue without making the survey longer.
Keep exactly one open-text question. It produces the comments staff remember, and it often names the exact moment that went wrong.
When and how to send it
Send within a day
Memory of the front desk fades quickly. A text or email the same evening or the next morning gets more accurate answers about check-in and waiting than one sent a week later.
Use the channel patients already carry
About 91% of US adults own a smartphone[4], so a text link reaches most patients. Keep an email option, plus a short paper card or a call-back for patients who do not text.
Exclude with care
Skip surveys after visits where a message could reveal something private to whoever sees the phone, respect every opt-out, and limit each patient to one survey per month.
Offer the right languages
About 9.0% of US residents aged 5 and over speak English less than very well[5]. Where that share is high, translate the survey and have a fluent speaker read it before launch.
Design for a phone screen
One question per screen, a progress indicator and no login. Ask for a name and number only when the patient wants a call back.
Keep the invitation neutral and short. A workable text reads: “Thanks for visiting Example Family Medicine today. Would you answer 5 quick questions about your visit? [link] Reply STOP to opt out.” Do not hint at the answer you hope for, and do not send a survey and a public review request in the same message.
Response rates and results you can act on
A low response rate is not automatically a problem, but a lopsided one is. If most answers come from one location, one provider or one age group, the results describe them rather than your practice. Track who answers as well as how many.
- Count invitations sent, surveys started and surveys completed each month.
- Break results down by location, provider and day of week before acting on them.
- Treat any group with only a handful of responses in a month as a hint, and wait for the next month before changing staffing.
- Read phone-access answers next to call logs. MGMA notes the phone often remains the real front door for many patients[6], so patient answers and abandoned-call data should tell the same story.
Report the share of top-box answers, such as “always” or “strongly agree”, for each question rather than an average score. Averages hide a group of very unhappy patients behind a crowd of satisfied ones, and top-box shares move visibly when you fix something.
Mistakes that spoil survey data
- Letting staff hand surveys only to patients who seem happy. Send to everyone, or to an automatic random sample.
- Changing the wording every month. Trends mean something only when the question stays the same, so revise wording once a year at most.
- Surveying the same loyal patients again and again. Cap frequency per patient so frequent visitors do not drown out everyone else.
- Collecting comments nobody reads. If no one has time to read open text each week, shorten the survey until someone does.
- Using scores to punish individuals. Staff who fear the survey will steer patients away from it. Use results to fix processes and to thank people by name.
- Ignoring the patients who never answer. Look at who is missing, such as older patients or those who prefer another language, and offer them a phone or paper version.
Example: one month of results
Example. A fictional three-provider practice sends 1,200 invitations in March and receives 216 completed surveys. Courtesy is strong at 88% top-box. Reaching us is weak at 54%, and open comments mention busy signals before 9 a.m. The practice manager checks phone logs and finds most abandoned calls between 8:00 and 9:30.
In the example, two changes follow: a second person on the phones from 8:00 to 9:30, and a call-back option when the queue is long. In April, reaching us rises to 63% top-box, and comments shift to parking. The manager keeps the phone change, rotates a parking module question into May, and posts a short note in the waiting room explaining what changed.
What made this work was not the survey tool. It was that each question had an owner, the manager checked a second data source before acting, and the change was small enough to test within one month. If the phone score had not moved, the next step would have been to listen to recorded calls from that window rather than add another question.
Close the loop with patients and staff
Patients keep answering when they can see that answers lead somewhere. Post a short “You said, we did” note in the waiting room and on the portal each quarter. Share results with staff monthly, naming individuals only for praise, and treat low scores as a process question first: scripts, staffing at peak hours, handoffs.
Any comment that describes a safety concern, a privacy concern or a billing error should get a personal reply within one business day and go into your complaint log, following your process for handling patient complaints. For a monthly routine that turns comments into visible changes, see turning patient feedback into front desk changes.
Frequently asked questions
How many questions should a patient satisfaction survey have?
Five to seven closed questions plus one open comment is enough for operational decisions. Rotate one extra question each quarter instead of adding more.
When is the best time to send a post-visit survey?
Within a day of the visit, while the patient still remembers check-in and waiting. Later surveys tend to reflect the overall relationship rather than the visit.
Should we offer an incentive for completing the survey?
Usually it is not needed. If you offer one, give it to everyone who responds regardless of their answers, and never tie it to posting a public review.
Can we quote survey comments on our website?
Only with the patient's written permission, and only comments that reveal nothing about their health. Most practices keep survey comments internal.
References
Medical Group Management Association (MGMA) (2025). Patient access priorities for 2026: Tackling wait times, phones, no-shows and more (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) ↑
Agency for Healthcare Research and Quality (AHRQ) Academy (current). Consumer Assessment of Healthcare Providers and Systems - Clinician & Group Measures (opens in a new tab) ↑
Pew Research Center (2026). Americans' Use of Mobile Devices in 2026 (Mobile Fact Sheet) (opens in a new tab) ↑
U.S. Census Bureau (2024). GCT1603: Percent of People 5 Years and Over Who Speak English Less Than 'Very Well' (2024 ACS 1-year geographic comparison table) (opens in a new tab) ↑
Medical Group Management Association (MGMA) (2026). Phones are still a bottleneck costing medical practices time they can't afford (opens in a new tab) ↑