To clear a refill request backlog in CureMD without adding clinical staff, write down your refill protocol by medication class, sort incoming requests into three lanes (prepare automatically, provider review, visit required), and automate the routine lane so nurses and providers only touch exceptions. Then track turnaround time weekly so the backlog stays gone. The work is mostly rules and routing, not headcount.
Why does the refill backlog keep growing in the first place?
Refill backlogs grow because every request costs a person the same few minutes whether it is routine or complicated, and request volume does not slow down when your staffing does. A backlog is simply the point where arrivals outrun handling capacity for long enough that the queue stops emptying overnight.
The volume is large. In one health system study of primary care out-of-visit work, 178,160 of 472,411 tasks were refill requests. Without scribe support, providers took about three hours on average to close a refill task, and that number reflects queue time as much as work time.
The hidden cost is that most refills are not hard. They are lookups: Is the patient due for a visit? Are labs current? Is this on our approved list? Spending a trained nurse's time on a yes-or-no answer a rule could give is the part you can remove.
Step 1: Measure the backlog you actually have
You cannot fix what you have not counted, so spend a few days measuring before you change anything. Pull four numbers from your CureMD inbox and message logs:
- Weekly volume. How many refill requests arrive per week, split by source (pharmacy, portal, phone, fax).
- Open count. How many are sitting unresolved right now, and how old the oldest one is.
- Turnaround. The median and the 90th percentile time from request to closure. The 90th percentile is the number patients complain about.
- Touch time. A rough estimate of how many minutes a request takes from start to finish. A sample of 30 to 40 requests is enough.
For a reference point on touch time, a Geisinger rheumatology group timed 40 refills and found about two minutes each, against a volume of more than 1,200 requests a month. Two minutes sounds trivial until you multiply it. Your own numbers will differ, which is why you measure them.
Write the results on one page. You will use them again in the last step to show whether the backlog really shrank.
Step 2: Write your refill protocol by medication class
A protocol is a short, written list of the conditions under which a request can be renewed without a provider investigating it. This is the step that makes automation possible, and it is the step most practices skip.
Start with your top 20 to 30 requested medications, which usually cover most of the volume. For each class, define:
- Who can approve. Provider only, or nursing staff under a standing order.
- Visit requirement. Often a visit within the past 12 months.
- Monitoring requirement. Labs or vitals that must be current, such as a TSH for thyroid medication.
- Supply limits. How many days or refills you issue, and what a bridge supply looks like for an overdue patient.
- Hard stops. Conditions that always send the request to a provider.
You do not have to invent this from scratch. Published examples exist. One family practice's standing-order protocol lets nurses refill listed medications the office has previously prescribed, requires a normal TSH and a visit within 12 months for thyroid refills, and sends overdue patients a 30-day supply only after an appointment is booked. It also states plainly that no controlled substance prescriptions may be sent by LPNs or RNs.
Have your medical director or lead providers sign off. A protocol the providers did not agree to will be overridden within a month.
Step 3: Split requests into three lanes
Once the rules are written, every request falls into one of three lanes. This is the structural change that clears the backlog.
Lane one: prepare automatically. The medication is on the approved list, the patient has been seen recently, monitoring is current, and there are no flags. The system drafts the renewal and queues it for provider confirmation, or for nurse completion if your standing orders allow it.
Lane two: provider review. The request is plausible but needs judgment: a dose change in the last few months, a specialist-managed medication, a missing lab, a conflicting note. It goes to a provider with the reason it was flagged.
Lane three: visit required. The patient is overdue, or the medication needs an assessment before another supply. The request routes to scheduling, a bridge supply goes out if your protocol allows it, and the patient is told what to do next.
Controlled substances never enter lane one. Federal rules at 21 CFR 1311.120 require the prescribing practitioner to sign electronic controlled-substance prescriptions with two-factor authentication, so those requests always go to a prescriber.
The goal is not to put every request in lane one. It is to make sure staff time goes to lanes two and three, where people add value.
Step 4: Automate lane one inside your existing CureMD workflow
Lane one is where software earns its keep. The tasks are repetitive, rule-based, and high-volume, which is exactly what automation does well.
An automated refill workflow reads the incoming request, matches it to the patient and the active medication in the chart, runs your protocol checks, and drafts the renewal. The provider or nurse then confirms with one review instead of a five-step investigation. Nothing leaves the building without a licensed person approving it.
Look for three things when you evaluate options. First, it should work with the chart data CureMD already holds rather than asking staff to re-enter information. Second, it should show its work, so a reviewer can see which checks passed. Third, it should route exceptions by your rules, not its own.
Honey Health's refill agent is built around this lane structure: it prepares routine renewals with the supporting chart context and routes the rest to the right person with the reason attached. Whether you use that or another tool, run the first few weeks in a review-everything mode so your team can compare the system's decisions to their own before it handles volume.
Step 5: Set pharmacy and patient communication rules
A surprising share of refill work is not the renewal itself but the follow-up: the pharmacy that calls to ask where the request is, the patient who phones because they never heard back. Fixing that clears backlog by reducing duplicate requests.
Decide in advance what happens at each outcome. When a renewal is sent, does the patient get a message? When a visit is required, who contacts them and by when? When a request is denied, how does the pharmacy find out so it stops re-sending? Write these rules into the same protocol.
Duplicate requests are a quiet source of inflation. A patient who submits through the portal, then calls, then lets the pharmacy fax, can generate three tasks for one need. A workflow that recognizes and merges duplicates removes work without touching a single clinical decision.
Step 6: Track results weekly and fix the exceptions
After launch, measure the same four numbers from step one every week for the first couple of months. Add two more: the share of requests handled in lane one, and the share of prepared renewals that providers changed.
The Geisinger group offers a useful benchmark for what a clear protocol can do. After shifting routine refills to a pharmacist-led process, physicians handled about 26% of refills, and the group reported roughly 41.5 hours a month returned to its rheumatologists. That was a human-run protocol, so treat it as proof that protocol design recovers time, not as a promise about any software.
When providers change a prepared renewal, find out why. If the same reason keeps appearing, it belongs in the protocol. Once a month, review the lane two and lane three requests and ask which of them could have been lane one with a better rule.
What can go wrong?
Three failure modes show up repeatedly. The first is an unwritten protocol: if providers disagree about the rules, automation will expose that disagreement and staff will lose trust in it. The second is messy chart data. Labs stored in scanned documents or free-text notes cannot be checked reliably, so some requests will always need a person until that data is cleaned up.
The third is skipping the pilot. Switching on automation across all medication classes in one go makes errors harder to trace. Start narrow, prove accuracy on maintenance medications, and expand from there.
Frequently Asked Questions
How long does it take to clear a refill backlog?
Expect two to six weeks, depending on how quickly your providers agree on a protocol. The measurement and protocol steps take the most time. Once lane one is running, a stale backlog usually clears within days because the routine requests drain fast.
Do we need to hire another nurse or MA to handle refills?
Not necessarily. Many backlogs come from process, not headcount. If most requests are routine lookups, rules and automation can absorb them. If lane two and lane three volume is still too high afterward, you will at least have the data to justify a hire.
Can nurses refill medications without a provider seeing each request?
In many practices, yes, under written standing orders or protocols, but rules vary by state and by medication. Controlled substances are excluded. Check your state's scope-of-practice rules and have your medical director approve the protocol.
What should we do about patients who are overdue for a visit?
Send them to lane three. A common approach is a short bridge supply once an appointment is booked, plus a clear message telling the patient what to do. Your protocol should state the supply length and who contacts the patient.
How do we know the backlog is actually shrinking?
Track median turnaround, the 90th percentile, and the open count every week. If the median drops but the 90th percentile stays high, a category of requests is still getting stuck, and that category is where to focus next.
