1 in 9 Patient Calls Isn't From the Patient — and It's More Urgent

Most analyses of medical-practice phones ask why patients call. This one asks a different question: who is actually on the line? It turns out that's one of the most overlooked signals in the entire phone workflow — and the answer reshapes how a practice should think about the calls it can least afford to drop.

We can ask it because CallMyDoc has handled more than 27 million patient calls over 8+ years across 40 states, with zero breaches and zero lost calls. For this report we took a fully de-identified, aggregated sample of 494,000 recent call transcripts and classified each one — including who placed the call. No patient-identifying information was used or shown. Here is what the caller's identity reveals.

1. One in nine callers isn't the patient

Across 494,000 calls, the breakdown of who is calling looks like this:

• 69% — the patient
• ~11% — a family member, caregiver, or parent calling on the patient's behalf
• ~10% — another provider's office (referrals, records, coordination)
• ~2% — a pharmacy
• The remainder: unknown callers and vendors

The number worth sitting with is the second one. Roughly one in nine inbound calls is a proxy call — an adult child arranging a parent's appointment, a parent calling about a sick kid, a spouse relaying a problem the patient can't describe themselves. That's not a rounding error; in a practice taking 1,000 calls a week, it's more than 100 calls placed by someone other than the person in the chart.

Most phone workflows quietly assume the patient is the caller — identity verification, "what's your date of birth," self-scheduling, portal nudges. One in nine times, that assumption is wrong before the call even starts.

2. The caregiver call is the more urgent call

This is the finding that should change how practices think about who's on the line. When we compared the urgency our research classifier assessed from what each caller actually described, proxy calls were consistently more serious:

• Patient-placed calls: ~16% assessed urgent or emergent.
• Caregiver/family-placed calls: ~26% assessed urgent or emergent — roughly 1.6× higher.

It makes intuitive sense once you see it. People generally manage their own routine business — booking a cleaning, requesting a refill, asking about a bill. But they pick up the phone for someone else when something is wrong and that person can't handle it alone: a parent who's confused, a child with a fever, a spouse who's deteriorating. The act of a third party calling is, statistically, a marker that the situation has escalated past self-service.

To be precise about what this is and isn't: this is an analytical pattern in aggregate, de-identified data — not a clinical judgment about any individual call. The takeaway isn't "diagnose the caregiver." It's the opposite and more important point — the calls least safe to leave in a voicemail box or a press-1 menu are disproportionately the ones where the patient isn't the one calling.

It’s part of a consistent picture in the same dataset: patients rarely flag their own urgent calls — so who is on the line is one of the few urgency-related signals that doesn’t depend on the patient self-reporting.

3. They call for both ends of life

Why are proxy calls more urgent? Look at who they're about. The age distribution of patient-placed vs. proxy-placed calls is strikingly different:

• On patient-placed calls, the patient is almost never a child — about 0.5% are pediatric.
• On proxy-placed calls, 20% are about a child (0–17) and 43% are about someone 65 or older.

In other words, caregiver calls cluster at the two most clinically vulnerable points of life — the very young and the very old — precisely the patients who can't reliably advocate for themselves on a phone tree. The frail elderly and small children don't call their doctor; someone calls for them. A phone system optimized only for the self-sufficient adult in the middle of the age curve is, by construction, worst exactly where the stakes are highest.

4. They're the hardest calls to automate

If proxy calls were just routine errands placed by a helper, automation would handle them fine. They aren't. By the same classification:

• 49% of caregiver/family calls were judged to need a person (vs. 38% of patient calls), and only ~6% fully automatable (vs. ~9%).
• Their reasons skew toward care coordination, not self-service: relaying a message for the provider (12% vs. 7.5%), clinical questions, referrals, and records — rather than the refills and billing patients handle themselves.

This is the operational trap. The proxy call is simultaneously the most likely to be urgent and the least likely to be resolvable without a human. A workflow that funnels everything into self-service automation — or worse, into voicemail — concentrates its failures on exactly this population.

5. After hours is where it peaks

Now combine caller identity with timing, and the pattern sharpens to a point. Proxy callers are disproportionately likely to call when the office is closed:

• Caregiver/family calls are about 50% more likely to fall outside their office's open hours than patient-placed calls.
• And among after-hours proxy calls, our classifier assessed over half (51%) as urgent or emergent — compared with 34% of after-hours patient calls.

Think about what that describes: it's 9 p.m., the office is closed, and the person calling isn't the patient — it's a daughter watching her father get worse, or a parent with a feverish child. More than half of those calls carry an urgent signal. That is the single worst call to drop into a voicemail box with a "we'll get back to you in the morning" — and it's a predictable, recurring share of after-hours volume, not a freak event.

What this means for your practice

Put the five findings together and a clear principle falls out: caller identity is one of the strongest, cheapest signals a practice has — and almost nobody uses it. When the caller isn't the patient, the call is more likely urgent, more likely about a child or a frail elder, less likely to be automatable, and more likely to come after hours. None of that requires diagnosing anything. It just requires not flattening every caller into the same self-service funnel.

The design implication is straightforward: you can't rely on patients — or their caregivers — to self-triage into the right bucket, and you can't leave the closed-office hours to a voicemail box. The safe design is to capture every call, automate the genuinely routine ones, and guarantee a frictionless path to a live person for everyone else — day or night.

That's the problem CallMyDoc is built to solve — during the day, after hours, and across multi-site call centers alike. It captures and answers every call, automates the routine (an average of 47% fully resolved across practices), and guarantees any caller who needs a person — including the caregiver calling at 9 p.m. — reaches one, with the patient verified by date of birth and the call documented back into the chart. It's deliberately not 100% AI: the routine gets handled, and the calls that matter never fall into a voicemail box. Across 27 million calls and 8+ years, that has meant zero breaches and zero lost calls.

Book a demo → and see how your practice handles the calls that aren't from the patient.

Methodology: Findings are aggregated and de-identified, drawn from automated classification of 494,000 recent patient call transcripts handled by CallMyDoc, part of a corpus of 27M+ calls across 40 U.S. states. Caller role, urgency, and reason were produced by an automated research classifier; urgency reflects an AI assessment of call content for analysis, not clinician triage or any live product feature. No patient-identifying information was used or included. Percentages are rounded.