1. Why general EHRs fail at this

General-purpose EHRs are organized around the episodic visit: a patient arrives with a problem, a history is taken, a note is written, a plan is made. That model describes primary care well. It does not describe a hormone, weight-loss, or IV therapy clinic, where the same patient receives the same injection at the same dose every week for years, often without seeing a provider at all.

Forced into the visit model, a weekly injection becomes one of two things: a full clinical note nobody wants to write fifty-two times a year, or nothing at all. In practice it becomes a third thing — a line in a paper binder, written in ten seconds by whoever gave the shot. Paper wins on speed, which is exactly why it survives. It loses on everything afterwards: you can't search it, you can't report on it, you can't hand it to an auditor, and you can't answer a recall with it.

The fix isn't to make providers write more. It's to give a repeat administration its own lightweight record, separate from the clinical note, that takes seconds to file and produces structured data.

2. What the record has to capture

An injection record is a clinical and — for testosterone and other controlled substances — a regulatory document. A checkbox that says "injection given" is not one. At minimum it needs:

  • Medication, dose, and unit — 100 mg, 0.5 mL, 500 units. Free text won't group into a dose history or a trend.
  • Route — IM, SubQ, IV, intradermal. It changes absorption and it changes what the dose means.
  • Injection site — and it should not pre-fill from last time, because sites are supposed to rotate. Everything else can default; this one shouldn't.
  • Lot number and expiration of the vial used — see below. This is the field most systems omit entirely.
  • Who administered it, recorded from the signed-in user rather than typed.
  • Date and time, with the ability to backdate — because a busy Monday morning gets documented Monday afternoon, and forcing "now" on the timestamp just produces records everyone knows are wrong.
  • Quantity drawn, so stock moves as you work instead of during a monthly count.

Everything on that list is information the person giving the injection already has in their hand. The design goal is capturing it in two taps, not asking for more.

3. Lot tracking is the part people skip

Compounded testosterone, semaglutide, peptides, and B12 all come from outsourcing facilities and compounding pharmacies, and recalls happen. When one does, the question you'll be asked is simple and unforgiving: which of your patients received a dose from this lot?

With a paper binder that's an afternoon of page-turning across however many months the lot was in use, and a genuine chance of missing someone. With lot numbers recorded against each administration it's one lookup and a list of names you can start calling.

Lot capture pays for itself in a second, quieter way. When a patient reports a reaction, the chart tells you which vial they were on, and whether anyone else on that vial reported anything similar. That's a question you simply cannot reconstruct from memory three weeks later.

If you dispense compounded product, ask vendors about lot capture explicitly and don't accept "you can type it in the notes." A lot number buried in free text is not searchable across patients, which defeats the entire purpose.

4. Logging a dose should move inventory

The moment you keep vials on site, you're running a small pharmacy — and the gap between what you bought and what went out the door is pure margin leakage. Most clinics discover the gap during a monthly count, by which point nobody can explain it.

Administration and inventory should be the same action. Selecting the lot at the point of injection and recording the volume drawn means the vial draws down in real time, partial vials are tracked by volume rather than as whole units, and expiring stock surfaces before it's wasted rather than after. It also means your controlled-substance reconciliation is a report you run, not a spreadsheet someone maintains.

Ask specifically whether the system decrements by volume. Plenty of inventory modules only understand whole units, which is useless for a 10 mL multi-dose vial that serves twenty patients.

5. The history is the point, not the entry

Logging a dose is only half the value. What a provider actually needs, with the patient in the chair, is the pattern: how many doses in the last thirty days, the average interval between them, the current dose, and when the last one was.

A patient who has quietly drifted from every seven days to every eleven is a clinical signal, and it usually shows up in the interval long before it shows up in symptoms or labs. Someone whose dose was raised six weeks ago and hasn't been back since is a retention problem you'd rather catch now. Neither is visible in a paper log unless someone sits down and counts, and nobody sits down and counts.

This is the argument for structured capture that has nothing to do with compliance. The compliance case is about the day something goes wrong; the interval case pays out every single week.

6. At-home doses are a different record

Many clinics have patients who self-administer between visits — most GLP-1 programs work this way, and plenty of TRT patients inject at home. Those doses belong on the same timeline, because a provider wants one answer to "when did this person last have their shot."

But they should never be silently merged. A patient tapping taken in an app is a self-report. A clinician documenting an administration is a witnessed clinical record. Summing them into a single adherence percentage gives both the same evidentiary weight — which is not a distinction you want to blur on a Schedule III medication, and not one you want to explain to a board later.

The right design shows both on one timeline, visually distinct, and keeps them separable in any report. Patient-reported adherence is genuinely useful clinical information. It just isn't the same kind of information.

7. Corrections, voids, and the audit trail

Someone will log a dose on the wrong patient. It happens in every clinic, and how the software handles it matters more than it sounds.

A medication record that can simply be deleted looks like a cover-up to anyone reviewing the chart afterwards, because there's no way to distinguish an honest fix from a removed inconvenience. The correct behaviour is a void with a reason: the original entry stays visible, struck through, with who voided it and why. Inventory should reverse at the same time, or your stock counts drift every time somebody fixes a typo.

This is also the cheapest compliance feature in the entire system, and its absence tells you something about how seriously the vendor took the rest.

8. Questions to ask a vendor

Take these to any demo. Ask them to do it live on a chart rather than describe it:

  • Can I log a repeat injection in two taps, pre-filled from the patient's last dose?
  • Does the record capture lot number — and can I search every patient who received a given lot?
  • Does logging a dose decrement inventory automatically, by volume rather than by whole vial?
  • Can I see dose count, doses in the last 30 days, and average interval without running a report?
  • Can a medical assistant log an administration, or does it require a provider login?
  • Can I backdate an entry for a shot given this morning?
  • Are in-clinic and patient-reported at-home doses distinguishable on the chart?
  • Can a mistaken entry be voided with a reason rather than deleted — and does inventory reverse?

If a vendor's answer to most of these is "you'd put that in the note," you'll be keeping the binder.

9. How Aminova handles it

Aminova records in-clinic administrations directly on the patient chart. You pick the medication and the specific lot from your own inventory, enter dose, route, and site, and the vial draws down by the volume used as you log. Repeat doses pre-fill from the patient's last administration — except the site, which stays blank so it rotates.

The history above the entry form shows total doses, doses in the last 30 days, the current dose, and the average interval computed from the actual gaps between administrations. Doses patients log themselves at home appear on the same timeline, clearly marked as self-reported and never merged into the clinical count. Mistakes are voided with a reason, stay visible on the record, and reverse the inventory movement.

It exists because a clinic owner told us plainly that this was the one thing keeping him on paper — and that nobody he'd demoed had a real answer for it.

10. Frequently asked questions

Is an injection log the same as an eMAR? Conceptually yes — it's a medication administration record. The difference is scale and shape: hospital eMARs are built around scheduled inpatient rounds, while an outpatient clinic needs a fast, walk-in-friendly record attached to a recurring protocol.

Do I need this if I only prescribe and patients inject at home? Less urgently, but you still want the at-home doses on the chart — adherence is the single best predictor of whether a patient stays on protocol, and it's invisible without it.

Does recording lot numbers slow the front desk down? Only if it's typed. Selecting an open lot from your inventory list is one tap, and it's the same lot for every patient that day.

Can medical assistants log injections? They should be able to — they're usually the ones giving them. What matters is that the record captures who administered it, and that scope-of-practice rules for your state are respected in who can be selected as the administering staff member.

What about IV therapy, not just injections? Same record, different route. IV pushes and drips want the same structure — product, lot, volume, site, administering staff — plus start and stop times for infusions.