TL;DR: The ROI of automating refill requests for a mid-to-large independent practice equals staff minutes saved per refill, times monthly refill volume, times loaded staff cost, plus recovered clinician time and fewer patient callbacks, minus software and oversight cost. The formula is simple. The hard part is measuring your own baseline, because published volume and time figures are old, small, and not yours.
How do you calculate the ROI of refill automation?
Start with a formula you can fill in with your own numbers:
Annual savings = (refill requests per year) × (share the automation handles) × (staff minutes saved per request ÷ 60) × (loaded hourly staff cost)
Then add the benefits that don't convert to dollars as cleanly, such as clinician time returned and fewer patient callbacks, and subtract the annual cost of the software plus the time your team spends overseeing it.
Each variable has a trap. Request volume is easy to undercount if you only look at the eRx queue. The share handled depends on your protocol and your channel mix. Minutes saved are usually estimated, not measured. And loaded staff cost should include benefits and payroll taxes, not just the hourly wage. The rest of this article walks through how to fill each one in honestly.
What does the published evidence say about volume and time?
The public numbers are useful for sanity-checking your own, but treat them as ranges, not forecasts.
- A Journal of the American Board of Family Medicine study of five primary care practices found an average of 63 refill requests per physician per week, with a range from 31 to 91. The authors estimated the cost per refill at about $0.25, excluding overhead. That figure is from 2006 and covers only a narrow slice of the work, so it understates what refills cost a modern practice.
- The American Medical Association cites 10 to 25 refill requests per day for a primary care physician, with about 30 minutes a day spent assessing and responding.
- A 2004 MGMA analysis, reported in Physicians Practice, put phone time alone for managing prescriptions at nearly 350 hours per full-time physician per year, about $16,000.
- A study of a clinical pharmacist-managed refill clinic estimated four to five minutes per request and reported physician refill request volume down by up to 60 percent. The authors couldn't directly measure the time effect.
The spread between these figures is the point. Your volume per provider could be a third of the low estimate or double the high one, depending on specialty and patient mix. A cardiology group with many chronic patients will look different from a dermatology group.
How do you measure your own baseline?
Spend two weeks logging every refill request. You need four things: the channel it arrived through, how many minutes of staff time it took from arrival to decision, who touched it, and the outcome. A shared spreadsheet is enough.
Then separate the requests into three buckets: those that would clear a written protocol with no clinical judgment, those that need a clinician decision, and those that are duplicates, pharmacy changes, or otherwise not true refills. Only the first bucket is a candidate for full pre-decision. The second still benefits from the chart being assembled for the prescriber. The third benefits from deduplication.
Measure staff time honestly. Include the chart lookup, the re-keying of faxed or phoned requests, the callback to the patient, and the documentation. People tend to remember the clicking and forget the chasing. If you have more than a few hundred requests a month, ask your staff to time a sample of ten rather than guessing.
What does a worked example look like?
Here's an illustrative example, using assumptions you should replace with your own. These numbers are not benchmarks.
Assume one provider receives 63 refill requests a week, borrowed from the 2006 study as a placeholder. That's 3,276 per year. Assume 70 percent of them are routine enough for the automation to pre-decide, and that doing so saves three minutes of staff time each. Assume a loaded staff cost of $32 per hour.
- Requests handled: 3,276 × 0.70 ≈ 2,293
- Minutes saved: 2,293 × 3 ≈ 6,880, or about 115 hours
- Staff time value: 115 hours × $32 ≈ $3,700 per provider per year
Under those assumptions, the software pays for itself on staff time alone if it costs less than roughly $3,700 per provider per year, before counting anything else. Multiply by the number of providers in your group to see the total. Change any assumption and the answer moves, which is why the baseline matters more than the formula.
Notice what the example leaves out: clinician time, patient callbacks, and error reduction. Those are real, but they're harder to value, so it's better to count them as upside than to bury them in the main case.
What benefits don't show up in the formula?
Several benefits matter to a practice owner even though they don't fit neatly into a dollar figure.
Clinician time. The AMA's figure of about 30 minutes a day per primary care physician is a real burden, even if you can't bill for it. When the chart evidence is attached to each request, a prescriber's review takes less time and fewer interruptions. Whether that time becomes more visits, earlier finishes, or less burnout is your call.
Turnaround and patient experience. Faster refill decisions mean fewer patient calls asking where the refill is. A California Health Care Foundation evaluation of a refill automation platform at two health centers reported same-day refill completion rising from 57 to 68 percent at one site, and estimated 21 and 77 physician hours saved per month at the two sites using the vendor's two-minutes-per-refill benchmark. Both are estimates from a small, vendor-linked study.
Fewer gaps in therapy. A refill that sits in a queue for three days can mean a patient runs out of a maintenance medication. That's a clinical and quality issue, and in value-based contracts it can affect measured performance.
Staff retention. Refill work is repetitive and interruptive. Taking the repeatable part away is one of the cheaper ways to make a medical assistant's day better.
What costs should you include?
Be as thorough on the cost side as you are on the benefit side, or the number won't survive a skeptical partner.
- Subscription or usage fees. Get a written quote that covers all providers and all channels you plan to use.
- Implementation time. Count your own team's hours spent writing the protocol, configuring rules, and running a pilot.
- Integration work. Ask what connecting to your EHR involves and who does it.
- Ongoing oversight. Someone owns the exception queue and reviews the audit log. Count a few hours a week.
- Training. Staff and providers need to learn the new queue.
Then run the sensitivity test. What happens to the ROI if the automation handles only half as many requests as you assumed? What if it saves one minute less per request? If the case still works at the pessimistic end, it's strong enough to bring to your partners.
How should you present the business case to your partners?
Owners and physician partners tend to ask the same three questions, so build the case around them.
What's the number, and how sure are you? Show a base case and a pessimistic case side by side, with your measured baseline behind both. A range with honest assumptions earns more trust than a single impressive figure.
What could go wrong? Name the risks before they do: a controlled-substance request handled incorrectly, a rule that's too loose, a vendor that's hard to turn off. Describe how the design fails closed, how controlled substances stay with the prescriber, and how the audit trail works. Under 21 CFR 1311.120, only the prescribing practitioner can sign an electronic controlled-substance prescription, and a good tool never tries to change that.
How will we know it's working? Commit to three measures before you start: median hours from request to decision, the share of requests that needed a staff chart lookup before clinician approval, and the exception rate. Set a review date after a pilot, and agree in advance on what result would lead you to expand, adjust, or stop.
A pilot is the cleanest way to settle the argument. Pick a narrow set of routine medications, run it for a month or two, and let your own data replace the assumptions in this article.
When is the ROI negative?
It's worth saying plainly: automation doesn't pay in every practice. The ROI is weak when refill volume per provider is low, when most requests are already electronic and cleared quickly, when your protocol isn't written down, or when the real bottleneck is clinician sign-off rather than staff preparation. In those cases, the better move is often a written standing-order protocol or a nurse-led refill process, and the numbers from the pharmacist-managed clinic study suggest those can work well too.
How does Honey Health's Refill Management agent fit the math?
In the example above, Honey Health's Refill Management agent is the automation layer: it pulls requests from multiple channels into one queue, runs the chart check against your written protocol, and gives your staff a labeled list of what's ready to approve and what needs attention. The savings come from the first two items in the formula, the share handled and the minutes saved per request.
Ask any vendor for the inputs you need: what share of requests it handled in comparable practices, how it measured minutes saved, and what it costs at your provider count. Then substitute your own baseline. A vendor's average is a starting point, not your answer.
Frequently Asked Questions
How long does it take for refill automation to pay back?
It depends on volume and cost. Under the illustrative assumptions above, the break-even is a per-provider annual cost of roughly $3,700. Calculate it with your own numbers, including implementation time, rather than relying on a general payback period.
What's a realistic time savings per refill?
Published estimates range from about two to five minutes per request, and most are estimates, not measurements. Time a sample of your own requests to get a number you can defend.
Should I count clinician time in the ROI?
Count it as upside, not in the base case. Clinician time is real but hard to value, and including it can make a business case look inflated to a skeptical partner.
Does a larger practice get a better return?
Usually, because volume per provider is multiplied across the group and a written protocol is easier to justify at scale. But multi-site groups also face more variation in rules, which needs to be settled first.
What if my refill volume is mostly electronic?
Then the native EHR queue may already cover much of the work, and the return from a separate tool will be smaller. Look at where staff time goes before you buy anything.
