TL;DR: A DrChrono practice should automate routine, protocol-eligible refills from intake through a pre-filled approval, and route every controlled-substance request to the prescriber with the chart context already assembled, never auto-approving it. Segment requests by medication class on arrival, apply written criteria to the routine ones, and make the system fail closed: when a request is unclear, it goes to a person.
Why do controlled substances need a different path?
Because the rules, the risk, and the signature requirements are different from a blood pressure refill, and mixing them in one queue with one set of logic is how practices get into trouble.
They're also a bigger share of the queue than most administrators expect. In a California Health Care Foundation evaluation of a refill automation platform at two community health centers, an interviewee said it's not unusual for 15 to 20 percent of refill requests to fall into the controlled-substances category. That's one in five or six requests that need more care than a routine renewal.
Federal law adds structure. Schedule II prescriptions can't be refilled at all: each one needs a new prescription. Schedule III and IV prescriptions can carry a limited number of refills within a limited window. Your state may add its own limits on quantity, early fills, and prescription monitoring checks. This article doesn't try to summarize every state's rules, so confirm yours with your state board and your compliance lead before you set any logic.
How should you segment the queue?
Segment on arrival, before anything else happens. The first decision for every request is which path it belongs on, and that decision depends on the medication.
A workable three-lane structure looks like this:
- Lane 1: Routine maintenance medications. Non-controlled drugs for stable chronic conditions, such as common blood pressure, cholesterol, thyroid, and diabetes medications. These can follow written criteria and be pre-decided.
- Lane 2: Routine with monitoring. Non-controlled drugs that need recent labs or checks, such as certain anticoagulants or medications with required monitoring. These follow criteria but add a monitoring check.
- Lane 3: Controlled substances and high-risk exceptions. Anything scheduled, plus anything your providers decide always needs a clinician. These go to the prescriber with the chart context assembled.
Have your providers sign off on which medications belong in which lane. Put the list in writing, and set an owner for reviewing it. The lane assignment is the single most important rule in the whole workflow, so it shouldn't be left to a software default.
What makes a routine refill safe to pre-approve?
A routine refill is safe to pre-approve when every one of your written criteria is met and the chart data supports each one. Typical criteria include:
- Active medication. The drug and dose match the current medication list.
- Visit recency. The patient has been seen within your protocol's window.
- Monitoring current. Required labs or checks are within range for that drug class.
- Supply within limits. The requested quantity matches the prior prescription and your maximum.
- No flags. No duplicate request, allergy conflict, pharmacy change, or recent hospitalization noted.
If all five pass, the request goes into a one-click approval queue with the evidence attached. If any fails, it routes with the reason. How many of your requests clear the criteria is something your own audit will show. A study of a clinical pharmacist-managed refill clinic found that 85 percent of 5,706 requests were approved and 15 percent denied, with 42 percent needing some form of intervention. That's a reminder that even routine requests need a real exception path.
Visit recency is worth a closer look. The AMA notes that several states, including Idaho, Illinois, Maine, and South Carolina, allow non-controlled maintenance prescriptions to run 15 to 24 months, which can reduce how often a patient needs a visit just to keep a stable medication going. Whether your practice uses that room is a clinical and legal decision for your providers.
How should controlled-substance requests be handled?
Treat them as a clinician-decision lane with automated preparation. The automation does the legwork. The prescriber makes the decision and signs.
For each controlled-substance request, the system should assemble the information a prescriber needs to decide quickly:
- The patient, drug, strength, and quantity requested
- The date and quantity of the last fill and the last visit
- Any early-fill flag, based on the days' supply of the previous prescription
- Notes from the chart on the treatment plan and any agreements in place
- A prompt for whatever your state requires, such as a prescription monitoring program check
Then it routes the request to the prescriber, and stops there.
The signature rule is firm. Under 21 CFR 1311.120, an electronic prescribing application for controlled substances must require two-factor authentication to sign, must limit who can mark prescriptions ready for signature, and must not allow anyone other than the prescribing practitioner to sign. An automation layer that sits next to the EHR can prepare the request, but it isn't a signer, and any design that suggests otherwise should be a red flag in a vendor evaluation.
What about early-refill and duplicate requests?
Both are common, and both should be caught before they reach a clinician.
Early refills. Compare the request date with the days' supply of the last fill. If the patient is asking well ahead of schedule, flag it with the numbers and route to the prescriber. For controlled substances, an early request is never a routine approval. For non-controlled drugs, your protocol can allow a small early window, for example for travel, with a documented reason.
Duplicates. The same refill often arrives from the pharmacy's automated request and the patient's call. Match on patient, drug, and date, and close the second request with a note pointing to the first. Duplicate controlled-substance requests deserve extra attention, because a duplicate that slips through can become a double fill.
Both checks are cheap for software and tedious for people, which is a good reason to automate them. The JABFM study of refill management found that error rates in refill decisions are generally higher than physicians think they are, which is an argument for consistent, rules-based checks rather than relying on memory at the end of a long day.
Why should automation fail closed?
Failing closed means that when the system is unsure, it doesn't act. It routes to a person. In a refill workflow, that's the right default because a wrongly delayed routine refill is an inconvenience, while a wrongly approved controlled-substance refill is a patient safety and compliance problem.
In practice, design for these behaviors:
- Missing data stops the line. If the last visit date, the medication list, or a required lab result can't be found, the request goes to staff with the gap named.
- Ambiguous matches stop the line. If the patient or the drug can't be matched with confidence, a human confirms.
- New or unusual requests stop the line. A dose change, a new pharmacy, or a drug not on the medication list gets reviewed.
- Every decision gets a logged reason. The audit trail should say what rule passed or failed and what data it used, and it should be written back to the chart.
- The system can be paused. If something looks wrong, you can switch the automation off and fall back to the manual queue.
Ask any vendor, Honey Health included, to show you each of these behaviors in a test, not a slide.
How do you roll this out and keep it safe?
Roll out in stages, and give each stage an exit test. A sequence that works for most practices:
Stage 1: Observe. Run the segmentation and the criteria checks without acting on them. Compare what the system would have done with what your staff did, for two weeks. Every disagreement is either a rule to fix or a case that needed judgment.
Stage 2: Recommend. Let the system pre-fill decisions on Lane 1 medications only, with a person confirming each one. Track how often staff change the recommendation. If they change it often, your criteria are off.
Stage 3: Expand carefully. Add Lane 2 medications once monitoring checks prove reliable. Leave Lane 3 on the clinician path permanently, with automated preparation only.
Throughout, assign one named owner for the exception queue and one for the lane list. Review the audit log weekly for the first month, then monthly. Look specifically for controlled-substance requests that touched the routine lane by mistake. The target is zero, and a single occurrence is a reason to pause and fix the segmentation rule before going further.
Also keep your compliance and legal advisers in the loop. They don't need to approve every setting, but they should see the lane list, the criteria, and a sample of the audit log before you go live. That review is faster and cheaper before launch than after an incident, and it gives your providers the confidence to trust the workflow.
How does Honey Health's Refill Management agent handle this?
Honey Health's Refill Management agent is built to keep clinicians in control of high-risk decisions. It reads incoming requests across channels, sorts them by lane, applies your written criteria to routine refills, and routes controlled substances and exceptions to the prescriber with the chart context attached. Prescribers continue to sign in the EHR.
As with any tool in this space, the quality of the result depends on your protocol and your lane assignments. Start with a narrow set of routine medications, run in recommend-only mode, and expand once the exception data looks clean.
Frequently Asked Questions
Can a refill workflow ever approve a controlled substance automatically?
It shouldn't. Electronic controlled-substance prescriptions must be signed by the prescribing practitioner under federal rules, and the clinical judgment involved isn't something to hand to software. Automation can prepare the request and route it.
Which refills are safe to automate first?
Routine, non-controlled maintenance medications for established patients with clear visit and lab rules. Start with a narrow list, and add medication classes as your exception data supports it.
What should the system do with an early-refill request?
Flag it with the numbers, including the last fill date and days' supply, and route it to a clinician. For controlled substances, treat early requests as exceptions every time.
Do state rules change how I set this up?
Yes. States differ on prescription duration, controlled-substance quantity limits, and monitoring checks. Confirm your state's requirements with your compliance lead before writing your criteria.
What if the chart is missing information?
The request should route to staff with the gap named, not get approved or denied on partial data. Missing information is a reason to stop, and the reason should appear in the audit log.
