.jpg)
Mira Gwehn Revilla
- Confirmations, directions, form status, and refill questions move off the phone.
- Threads route by department, so billing never sits behind scheduling.
- Any authorized agent at any site can claim a thread and close it.
- Curogram clients confirm more than 75% of visits by text, based on our internal data.
- Call volume drops by as much as 50% once routine traffic moves to te
A patient calls with one question. Is the appointment at 2:00 or 2:30? An agent answers, verifies a date of birth, opens the chart, reads the slot, says the time, and logs the call. Four minutes spent on a ten-second answer.
Multiply that by the 80-plus calls a day a single location fields, then again by the number of sites feeding one call center. Hiring more agents scales the cost of that math without changing it.
eClinicalWorks holds everything an agent needs: schedule, chart, task list, billing record. It gives patients no way to ask a small question without occupying a phone line and a person. So the phone stays the only front door.
Refill status, parking, "did you get my form," "can I bring my daughter," all of it arrives as a ring. Your best agents spend the shift working as a human IVR, and they know it.
Call deflection changes which door those questions come through. Move the routine traffic to text and the volume that never needed a voice stops competing with the volume that does.
Curogram clients confirm more than 75% of visits by text, based on our internal data, which pulls the biggest routine call driver out of the queue before it forms.
Not every call can move, and pretending otherwise is how these projects fail. Clinical triage, angry escalations, and anything needing a real decision still belong on the phone. Sorting the two is the job. One week of call logs will hand you the split.
The Ringing Wall: Why eCW Front Desks Still Drown in Routine Calls
Watch one call from pickup to wrap-up. Greeting and identity check run about 45 seconds. Pulling the right chart in eCW takes another 20 to 30, longer when a patient has a common last name or two open encounters. The answer itself takes ten seconds.
Documentation, disposition coding, and the small pause before the next call eat the rest. Every step there exists for a reason. Skip the identity check and you have a compliance problem. Skip the note and you break the audit trail.
Cost comes from running a four-minute process against a ten-second question, forty times a shift. Stack those into one agent's day and the picture gets ugly fast.
Why Multi-Location Groups Feel It Hardest
Front desk phone tag in a multi-location group carries an extra hop that single sites never deal with.
A patient calls the number on the appointment card, reaches the central queue, and asks something only the Fontana office can answer. The agent takes a message. Fontana calls back at 3:40, gets voicemail, and the loop restarts tomorrow.
Two agents, two sites, one patient, and the question is still open. eCW centralized scheduling calls tend to cluster around three things: which location, which provider, and what time. All three are lookups a patient could settle by reading a text on a lunch break.
Callbacks carry a cost that phone reports rarely surface. Each one occupies a seat on the outbound side, so a queue that looked overstaffed at 9:00 a.m. is underwater by 2:00 p.m.
What Patients Do When They Can't Get Through
Patients don't wait patiently, and your phone report only shows part of what they do instead.
A caller sits on hold for four minutes, hangs up, then calls the location directly ten minutes later. Two calls now exist in your system for one question, and only one of them shows as answered.
Some skip the call and just show up early, which pushes the waiting room load onto the same front desk that was already covering the phones. Others no-show because they never confirmed the time they were unsure about.
Atlas Medical Center ran a 14.20% no-show rate before switching that confirmation step to text, and brought it to 4.91% in three months, based on our internal data.
Abandonment is the number worth pulling first. It tells you how many questions your queue never even got the chance to answer.
The Staffing Fix That Wears Off
Most groups answer the wall by hiring. Two more agents, a deeper phone tree, a callback queue, maybe an after-hours service. Volume absorbs the headcount inside a quarter, because none of those moves reduce how many questions arrive.
Turnover then does its own damage. A seat that spends the day repeating appointment times is a seat people leave, and every departure costs weeks of eCW training before the replacement is useful.
eCW call center average handle time climbs during that ramp, hold times follow it up, and patients who gave up the first time call again.
Practices trying to reduce call volume for an eClinicalWorks front desk often start with a phone tree. Patients press 3, wait, press 3 again, then hang up and dial the site directly. That's how a central queue ends up with a shadow queue at every location.
The Network Switchboard: How Call Deflection Works on Top of eCW
Deflection begins with a sorting exercise. Pull one week of call logs, tag every call by reason, and the list separates into three buckets on its own.
|
Call type |
Example |
Where it belongs |
|
Pure lookup |
"What time is my appointment?" |
Automated text, no agent |
|
Short exchange |
"Can I move to Thursday?" |
2-way text thread |
|
Real conversation |
New symptom, disputed balance |
Phone, with a person |
Buckets one and two are your deflection opportunity. In most groups they make up the bulk of daily volume, which lines up with the 50% call reduction Curogram clients report, based on our internal data. Bucket three is why you still staff a phone queue, and it should stay staffed.
Do this sort before any vendor call. A demo will show you a pretty inbox. Your own reason codes will show you whether the inbox is worth buying.
Department Routing Keeps Threads in the Right Queue
One shared inbox with one queue fails fast. Billing questions pile up behind scheduling requests, and an agent trained on copays ends up reading a refill message at 4:50 p.m.
Curogram routes by department. Billing threads land with billing, scheduling with scheduling, records requests with records. Each queue gets its own owners, its own response target, and its own view.
Routing also closes the multi-location gap that phone systems never handled. A thread tagged to Fontana is visible to Fontana staff and to any authorized agent in the central call center. Whoever is free takes it. No message pad, no callback promise, no second call tomorrow.
Set response targets per queue rather than one number for everything. Scheduling can run on a 15-minute target. Records requests can sit at four hours without anyone complaining.
Threads That Write Back to the eCW Chart
Side documentation is how a good deflection tool turns into a bad one. When an agent settles a question in a separate texting app and the chart never sees it, the next person to open eCW has no idea the conversation happened.
Curogram writes threads back to the patient chart in eCW. An agent on a live call can read Tuesday's text exchange without opening a second system, and nobody retypes anything into a telephone encounter note.
That one behavior decides whether staff use the channel at all. Teams abandon tools that create double work, no matter how good the inbox looked in a demo.
From Human IVR to High-Value Work
Take a 4-site group running one central call center on eCW. A reschedule that used to cost two calls and a voicemail now runs like this:
- Monday, 7:00 a.m. Reminders send for Wednesday's schedule. Most patients confirm by reply, and those never reach an agent.
- Monday, 9:15 a.m. One patient replies asking to come Thursday instead. The thread drops into the scheduling queue, tagged to the Rialto site.
- Monday, 9:17 a.m. A central agent claims it, checks Thursday openings in eCW, and offers two slots in a single message.
- Monday, 11:40 a.m. The patient picks 10:20. The agent books it in eCW and closes the thread.
- Total agent time: under three minutes, spread across a morning, with no hold and no callback.
Between 9:17 and 11:40 that agent worked six other threads. Asynchronous handling is what makes the gap free instead of expensive. On the phone, the same reschedule needs both people available at the same moment, twice.
The Numbers to Put on the Supervisor Dashboard
Phone systems report on calls. Deflection needs a second column beside them, and four figures cover most of it.
|
Measure |
What it tells you |
|
Threads handled per site |
Where the volume actually moved |
|
First response time |
Whether patients wait less than they did on hold |
|
Confirmation rate |
How much reminder traffic never became a call |
|
Abandonment rate |
The questions your queue used to lose |
Review these per location, not as a group average. One site with a slow scheduling queue will hide inside a four-site rollup and stay hidden for a quarter.
Pair each measure with a baseline from before go-live. Without that, a director looking at 900 threads a month has no way to say whether 900 calls disappeared or 900 new conversations appeared on top of the old volume.
Curogram Highlight: 2-Way Texting Built for Call Center Queues
2-Way HIPAA Texting turns a one-question call into a thread any agent can claim and close. It sits next to eCW instead of replacing anything inside it.
Threads route by department. Billing, scheduling, and records each get a queue with named owners. Agents work from quick replies for the questions that repeat. Every exchange writes back to the patient chart in eCW, so there's no side file to keep.
Consent and opt-out handling runs in the background. A patient who replies STOP is removed, and the audit trail stays with the thread.
The payoff shows up in numbers your supervisors already track. Curogram clients confirm more than 75% of visits by text, based on our internal data. That takes confirmation calls out of the queue for good.
Covina Arthritic Clinic confirms more than 1,100 visits a month this way. Atlas Medical Center brought no-shows from 14.20% to 4.91% in three months, and each one it prevented was also a rescheduling call nobody had to make.
Supervisors see deflection metrics per location: threads handled, first response time, and calls avoided. Those roll up across sites. A director can compare Fontana against Rialto without exporting a phone report.
Training runs about ten minutes for most staff. The inbox works like the texting app already on their phone, which is why adoption rarely stalls at week two.
Curogram signs a BAA and manages message retention, so the review with your privacy officer stays short.
Conclusion: Deflect the Routine, Keep the Human
Most of what rings your phone today is a lookup, and lookups read better than they sound.
The phone line stays exactly where it is. Curogram takes the traffic that never needed a voice and moves it somewhere a patient can read at a red light and an agent can close in one tap. That leaves the line open for the mother calling about a fever and the patient disputing a $400 balance.
Before you evaluate any vendor, run the counting exercise yourself. Pull one week of call logs. Tag each call with a reason code, then add up the ones a single sentence would have answered. That total is your deflection opportunity, and it's usually bigger than the call center director expects.
Then look at what the recovered hours buy. Covina Arthritic Clinic confirms more than 1,100 visits a month by text. Atlas Medical Center cut no-shows from 14.20% to 4.91% in three months. Both numbers started as calls somebody used to make by hand.
Request a demo and we'll model your deflection rate location by location, using your own call mix and site count. You'll leave with a staffing impact estimate you can take to your board.