Ask two dentists in the same practice what a complete patient record contains and you will usually get two different answers, both reasonable. One will describe what is clinically necessary, the other what the practice has always kept. Neither is wrong, and the gap between them is where inconsistent records come from.
This is about settling that question deliberately: what a record should hold, how long different parts of it matter, and how to find out whether yours are as complete as you assume.
The four things a record has to do
Rather than starting from a list of fields, it helps to start from what the record is for. Everything worth keeping serves at least one of four purposes.
Continuity of care. The next clinician, who may not be you, needs to understand what was found, what was done and what was planned. This is the primary purpose and the one that should settle arguments about what to include.
Safety. Allergies, medical conditions, current medication and anything else that changes what is safe to do. This is a small part of the record and the part with the least tolerance for being out of date.
Something recorded once at registration and never revisited satisfies the letter of this and not the purpose.
Evidence of what was agreed. What was explained, what was consented to, what was quoted. This matters most in the rare case where it is disputed, which is exactly why it cannot be reconstructed afterwards.
Running the practice. What was charged, what was paid, what is owed, who referred them, when they are next due. Administratively essential and clinically irrelevant, which is why it is often kept somewhere else and then needed at the same moment as the rest.
What that means in fields
| Part of the record | Primary purpose | Common weakness |
|---|---|---|
| Identity and contact | Finding the patient, reaching them | Never updated after registration |
| Medical history and alerts | Safety | Captured once, treated as permanent |
| Examination findings | Continuity | Written for the author, not a successor |
| Treatment plan and status | Continuity and follow up | Lives in a note rather than as items |
| Treatment performed | Continuity and evidence | Recorded later from memory |
| Consent and what was explained | Evidence | Form signed, discussion unrecorded |
| Charges, payments, balance | Running the practice | Kept in a separate system |
| Referral source | Running the practice | Optional field, frequently blank |
The third row is the one most worth attention. Clinical notes are usually written as a memory aid by the person who was there, which makes them terse and full of private shorthand. The test that matters is whether a different dentist could pick up this patient's care in two years without you. Notes that fail that test are notes, not records.
The sixth row is the most common real weakness. A signed consent form proves a form was signed. What was actually discussed, including alternatives and risks raised, is the part that carries weight, and it is rarely written down because the conversation felt complete at the time.
Writing notes a successor can use
Three habits separate records that work from records that merely exist, and none of them takes longer.
Write the reasoning, not only the action. "Extracted 26" says what happened. "Extracted 26, unrestorable, patient declined implant, discussed bridge" tells the next clinician why, which is what they actually need to know.
Record the negatives that mattered. If you examined for something and it was absent, saying so is more useful than silence, because silence is ambiguous between not present and not checked.
Avoid private abbreviation. Every practice develops shorthand, and it is efficient until the person who coined it leaves. A short glossary kept with the clinic's conventions is worth having, but the safer habit is writing the thing out when it is not universal.
Separate observation from interpretation where both appear. What was seen and measured is one kind of statement; what you concluded from it is another. Keeping them distinguishable helps a successor who may reach a different conclusion from the same findings, and it is also the part that ages best, since observations remain true while interpretations can be superseded.
Write at the time, or as close to it as the day allows. Notes written at six o'clock for a nine o'clock patient are reconstructions, and they are reliably shorter and less accurate than the writer believes.
Keeping the record current
Completeness at registration is easy. Completeness three years later is the real problem, and it needs a moment in the workflow rather than good intentions.
The workable approach is to attach verification to events that already happen. Contact details confirmed whenever a patient pays, because they are standing there anyway. Medical history reviewed at each recall examination rather than annually in the abstract, because the recall is a moment someone is already asking questions.
Flag what has not been checked rather than assuming currency. A medical history last confirmed four years ago should look different on screen from one confirmed last month. Where the record shows when something was last verified, staleness becomes visible instead of being invisible by default.
Where a patient's history, plans and payment record sit on one profile, updating during a routine visit is a single task rather than a decision about which system to open, which is the difference between a habit that holds and one that lapses.
Who the record belongs to
A point worth settling early, because it shapes how records get written: the patient generally has a right to see what is written about them.
The details vary by jurisdiction, but the direction is consistent. Patients can usually request their records, and practices are expected to provide them within a defined period. This is not an adversarial event in most cases. People request records because they are moving, because an insurer asked, or because they want to understand what was found.
Two practical consequences follow. The first is about tone. A record written in the knowledge that its subject may read it tends to be more careful and more factual, which is also what makes it better clinically. Anything you would want to rephrase before handing it over is worth rephrasing now.
The second is about retrievability. A request means producing a complete record, which is straightforward when everything sits on one profile and laborious when clinical notes are in one place, radiographs in another and payment history in a third. Practices discover the cost of fragmentation at exactly this moment.
Decide in advance who handles such a request and what the practice's process is. It happens rarely enough that improvising each time is the norm, and often enough that improvising is avoidable.
Auditing your own records
Most practices assume their records are complete and have never checked. A small audit answers it in an hour and is worth repeating yearly.
Take twenty patient records at random from those seen in the last year, and check five things. Is there a medical history, and when was it last confirmed. Is there a current contact number that has been updated at least once since registration. Does the most recent clinical entry explain why, not only what. Is there a recorded referral source. Does the balance, if any, match what the patient believes.
Twenty is enough. If three or more fail on the same item, that is a systematic gap rather than individual oversight, and systematic gaps have process fixes. If failures are scattered across different items, that is normal variation and probably not worth chasing.
Repeat it on the same twenty items rather than redesigning the audit each year. The value is in the comparison, and an audit that changes shape annually produces numbers that cannot be read against each other. Keep the five checks, keep the sample size, change nothing but the records.
Do the audit yourself rather than delegating it, at least the first time. The value is partly in the numbers and largely in reading twenty of your own records consecutively, which surfaces patterns nobody would report to you.
Retention, and what not to keep
Records have to be kept for periods set where you practise, often for years after the last treatment and longer for patients who were children at the time. Those rules are local, they change, and they are worth checking against a current source rather than memory.
The less discussed question is what not to record. A patient record should contain what is needed for the four purposes above, and accumulating beyond that creates obligations without benefit. Informal notes about a patient's personality, opinions about their circumstances, or detail about their family relationships are rarely necessary, and the test is whether you would be comfortable with the patient reading it, since in many jurisdictions they may.
The same caution applies to copies of things you do not need. Correspondence about a patient that arrives unsolicited, or documents belonging to another practice, should be considered rather than filed automatically. A record that accumulates everything sent to the clinic is harder to read, harder to produce on request, and holds material the practice has no reason to be responsible for.
Be similarly careful with information about third parties that arrives incidentally. What a patient mentions about a relative's health is their relative's information, not a clinical finding about the patient in front of you.
If you want to see how identity, history, plans and payments sit together as one record rather than several, the feature overview shows the structure, and you can book a demo to run the twenty record audit against your own.