Why Patients Abandon Intake Forms, and How to Fix It
Abandonment is almost never general fatigue. It is one or two specific questions, it is the same ones in most practices, and you can usually see exactly where it happens.
It is not length, it is a question
The instinct when people stop finishing a form is that it is too long, and the instinct is usually wrong. Patients will answer a great many questions about themselves if each one is easy. What they will not do is get stuck.
Getting stuck has a shape. A patient is moving through a form at a steady rate, hits a question they cannot answer from where they are sitting, and stops. They do not come back, because the thing that stopped them has not changed — they still do not have the card, still cannot remember the dose, still do not want to type a paragraph on a phone.
This is why drop-off clusters rather than spreading evenly. If the problem were length you would see a steady fade across the form. What practices actually see is a cliff at one question and a smaller one somewhere else.
Which means the fix is specific rather than general. You do not need to redesign the form. You need to find the two questions and change those.
Find the cliff before you change anything
If your forms tool can tell you where people stop, read that before touching anything. In Oraform it is the drop-off by question panel in the dashboard, and it ranks the questions patients quit on with the share of patients who stopped at each.
If your current tool cannot tell you, that is itself worth knowing, and there is a crude substitute: look at the partially-completed forms you do have and note the last answered question. Twenty of those will show you the cliff.
What you are looking for is not the question with the most blank answers — that is optional questions being skipped, which is fine. You are looking for the last question before people stop existing.
One caution about reading it: the first question is a special case. People who leave at question one did not get stuck, they changed their mind about starting, and that is a problem with the message carrying the link rather than with the form.
The questions people quit on
Four questions account for most of it, in roughly this order.
Anything requiring a document the patient does not have in their hand. Insurance policy and group numbers are the classic. A patient on the bus does not have the card, so they stop, intending to come back when they do. The fix is to accept a photograph instead of typed digits, and to make the question optional so it is not a wall. A card photographed later beats a policy number guessed now, which is a wrong answer rather than a missing one.
A long free-text box. “Please describe your medical history” on a phone is a request to write an essay with a thumb, and it reads as homework. Replace it with specific questions that have specific answers, and keep one optional box at the end for anything you missed.
A question that is unclear or feels intrusive without a reason. Income, immigration status, a question about substance use with no context. If you need it, say in one line why you are asking. “We ask so that we can check what your plan covers” converts a suspicious question into an ordinary one.
A grid. The conditions checklist, forty items in a table. On a desktop it is dense; on a phone it is a horizontal-scrolling trap with tap targets too small to hit. Break it into groups and present it as a list of large, tappable items.
Short, easy screens, and an honest estimate
The single biggest structural change is to stop showing the whole form at once. A long page of fields is an evaluation: the patient reads it, decides how long it will take, and frequently decides not now. A short screen cannot be evaluated the same way — it can only be answered.
It works for a reason beyond psychology: a screen holding one question, or a few that belong together, has room for the label to be a sentence, for the help text to exist, and for the tap targets to be large enough to hit.

Pair it with a progress indicator and a time estimate, and make the estimate honest. “About three minutes” that turns out to be eight does more damage than no estimate at all, because it teaches the patient that the form lies.
One real caveat. One-question-at-a-time can make a short form feel longer than it is — six questions as six screens can feel like more work than six fields on one page. Group the trivial ones: first name and last name belong on the same screen, and so do city, state and ZIP.
Phone-first means more than responsive
Most patients will open your form on a phone, including the ones who have a computer. Responsive layout is the minimum; the things that actually cost completions are smaller than layout.
- The right keyboard. A phone number field that brings up a QWERTY keyboard is a small insult repeated on every form. Numeric fields should ask for the numeric keypad, email fields for the email one.
- Autofill that works. Name, address, email and phone can all be filled by the phone if the fields are marked properly. A patient who fills four fields with one tap is a patient who is still there at question twelve.
- Dates as a picker. Asking a patient to type a date of birth in a particular format produces errors and friction. Let them pick.
- Tap targets that are large. Radio buttons sized for a mouse are a precision task on a phone.
- No pinch-zoom. If the patient has to zoom to read a label, the form is not phone-first, whatever the layout does.
None of these are interesting individually. Together they are most of the difference between a form that gets finished on a bus and one that gets postponed.
Let people stop without starting over
Some patients will stop regardless — the appointment comes up, the bus arrives, the child needs something. The question is whether stopping costs them everything.
Save as they go, so no answer is lost. Then give them a route back: a private link that returns them to the question they left on, sent to them rather than requiring an account. The account is worth emphasising, because the most common way to destroy this feature is to protect it with a password the patient now has to invent and remember in order to tell you their address.
A patient who stops at question twelve and finishes that evening is a completed form. The same patient, without this, is a blank form and a phone call from your desk.
When the link arrives decides most of it
Everything above is about the form. The largest single factor is often not the form at all — it is the moment the link lands.
A link in an appointment reminder the evening before is answered, because the patient is thinking about the appointment and has a free minute. The same link in a welcome email three weeks earlier is not, because there is nothing to act on yet and the message is gone by the time there is.
If you can send it twice, do. Once at booking for the organised, once the day before for everyone else. And if a patient arrives without having done it, a QR code at the desk should open the same form — not a paper version, the same form, so the answers land in the same place.
It is also worth asking what the message says. “Please complete your new patient paperwork” sounds like twenty minutes. “Four minutes, on your phone, before Thursday” sounds like something to do now.
Change one thing and look again
Resist the redesign. If drop-off is concentrated on two questions, change those two and leave everything else alone, so that when the number moves you know what moved it.
Then look again in a fortnight. The usual pattern is that the first cliff disappears and a smaller one somewhere else becomes visible, which is progress rather than a setback — it was always there, hidden behind the bigger one.
And accept a floor. Some patients will never fill in a form before an appointment, and chasing the last few percent costs more than keeping a tablet at the desk. The goal is that enough patients arrive finished that your desk has time for the ones who did not.
Questions
What percentage of patients abandon online intake forms?
We are not going to give you a number, because any number we invented would be worse than none — it varies enormously by practice, specialty, form length and how the link is sent. The useful number is your own, and the way to get it is a tool that shows drop-off by question. If yours cannot, that is worth fixing before anything else.
Does making a form shorter actually help?
Cutting questions nobody reads helps. Cutting length for its own sake helps less than people expect, because abandonment clusters at specific hard questions rather than spreading with length. Find the cliff first; you may find the fix is a photograph instead of a policy number rather than ten fewer questions.
Should we require every field?
No. A required field the patient cannot answer from where they are sitting is a wall, and a wall is where they stop. Require what you genuinely cannot proceed without, make the rest optional, and accept a photograph where typing is the problem.
How do we find out where patients stop?
In Oraform the dashboard ranks drop-off by question, with the share of patients who stopped at each. Without a tool that reports it, look at twenty partially-completed submissions and note the last answered question on each — the cliff shows up quickly.
Templates and guides
Get a free sample of your form.
Send us the forms you hand patients today and we'll turn them into working Oraform samples — your logo on them, finished on a phone or a computer. We rebuild all of your forms free, and free setup comes with it. No cost, no commitment, nothing to cancel.
We'll send your sample within one business day, with your logo on it — a link you can open on a phone or a computer and hand straight to a patient.
Nothing to do in the meantime. If you'd rather talk it through, email forms@oraform.health or call (312) 635-5523.