# Which Vendors Keep Running When Your EHR Goes Down

An electronic health record (EHR) outage does not stop referrals arriving. Faxes keep coming, patients keep calling, and documents keep landing somewhere. What your other vendors do with that work during those hours is almost never written down anywhere you can read it.

**Category:** Listicles
**Written by:** Zain Qayyum
**Last updated:** Nov 13, 2026
**Source:** https://signals.medsender.com/listicles/vendors-that-keep-running-in-ehr-downtime

---

An electronic health record (EHR) outage does not stop referrals arriving. Faxes keep coming, patients keep calling, and documents keep landing somewhere. What your other vendors do with that work during those hours is almost never written down anywhere you can read it.

Every downtime guide published in this category is written for the EHR itself. This one is about everything else in the stack.

## TL;DR

Among nine vendors read on 12 November 2026, Medsender and eight competitors, none publishes a service level agreement with a stated remedy.

One publishes a public status page with inspectable incident history.

None describes what happens to work in progress during an EHR outage.

A published uptime percentage is a target, not a commitment, and the two are routinely confused.

Vendors frequently publish different uptime figures on different pages, which is worth checking before you rely on either.

The real risk arrives on recovery, when the backlog lands all at once.

## Why This Is a Vendor Question, Not an Information Technology Question

Downtime planning usually stops at the EHR, and that leaves the larger half of the problem unexamined.

### The Inbound Side Does Not Pause

Referrals, records and calls arrive on their own schedule rather than on your system's. An outage quietly converts an automated queue into a manual one, usually without anybody announcing it, and the practice discovers the change when somebody notices the fax tray filling.

### The Risk Arrives on Recovery

Most practices cope for a few hours. The difficulty comes when the system returns and several hours of held work lands at once, which is precisely when the same document gets filed twice by two different people working the same backlog from opposite ends.

## A Published Uptime Number Is Not a Service Level Agreement

The difference between a target and an agreement decides how much any published figure is worth, and most vendors benefit from it staying blurry.

### What a Percentage Tells You

A target, usually self-measured and self-reported. Scheduled maintenance is commonly excluded from the calculation, and scheduled maintenance is where a meaningful share of real unavailable hours actually sits.

### What an Agreement Adds

A remedy, normally a service credit, plus a definition of what counts as down and who decides. Without a remedy there is nothing to enforce and nothing to measure against, which makes the number a marketing figure rather than a commitment.

### Check Whether the Figure Is Even Consistent

Vendors publish different uptime figures on different pages more often than you would expect, usually because copy written years apart was never reconciled. A number that changes between a security page and a cancellation page is not a commitment. This applies across the category and Medsender is not an exception, which is why the comparison below records consistency as its own criterion rather than taking any single figure at face value.

## How These Were Ranked

Five criteria, applied to the public pages of Medsender and eight competitors on 12 November 2026.

Is an uptime figure published, where, and is it stated consistently

Is there a service level agreement with a stated remedy, or only a target

Is there a public status page with incident history

Does the vendor say what happens to work in progress during an EHR outage

Does the vendor say how the backlog is handled on recovery

## The Vendors, and What Each One Publishes

Six vendors, ranked on disclosure rather than on resilience. None of them publishes enough to judge resilience from outside, which is itself the result.

### Updox: The Only Public Status Page in This Comparison

Updox publishes a status page at partnerstatus.updox.com, and no other vendor compared here publishes one. That matters more than any percentage, because a status page carries history a buyer can inspect rather than a figure a buyer has to accept. Twelve months of incident records tell you about recovery behaviour, communication during an outage and frequency, none of which a target can convey. Updox also holds Electronic Healthcare Network Accreditation Commission (EHNAC) DirectTrust accreditation for secure messaging, which is an independent assessment rather than a self-declaration. Against that, it publishes no service level agreement with a remedy, no uptime target, and nothing about what happens to documents in flight when a practice EHR becomes unreachable. It is a unified communications and charge capture platform rather than a processing one, so inbound documents largely sit until a person handles them regardless, which changes the shape of the downtime question rather than removing it.

### Medsender: Publishes a Figure, Without a Remedy or a Status Page

Medsender publishes an uptime figure, but it appears differently on different pages of its own site, so this comparison records it as published but not stated consistently and does not reproduce either number. There is no service level agreement with a stated remedy, no public status page, and nothing published about what happens to documents during an EHR outage. That is the same position most of this list occupies. What Medsender does publish and can be credited with: hosting on Google Cloud with United States based servers and no on-premise footprint, System and Organization Controls (SOC) 2 Type II with the report available through its trust centre on request, a business associate agreement (BAA) included free with every account, 256-bit encryption in transit, and role-based user permissions. Its one genuinely relevant downtime behaviour, exception routing to a named referral coordinator, is published, though not framed as a downtime feature.

### Linear Health: Strong on Process, Silent on Availability

Linear Health publishes more about how work flows than anyone here, stating that routine work is automated and exceptions escalated at roughly eighty per cent automated with the team in the loop on the rest. For downtime purposes that matters, because a practice at least knows that a human queue exists and who staffs it. It publishes no uptime figure, no service level agreement and no status page, and says nothing about behaviour during an EHR outage. It does publish pricing from $2,000 a month flat per workflow with nothing charged until the workflow is live, and it holds SOC 2 Type 2 with a BAA executed for every customer. Its native athenahealth write-back is relevant here in an unhelpful way: the deeper the integration, the more there is to queue when the destination system is unavailable. A practice evaluating it should ask what the escalation queue looks like during an outage specifically, since that is the one surface staff can still work from.

### Assort Health: Voice Changes the Downtime Question

Assort Health publishes no uptime figure, no service level agreement, no status page and nothing about EHR outage behaviour. Its trust centre exists but is not a status page with inspectable history. For a voice product the downtime question works differently, and fairly so: a call that cannot be completed is recoverable because the patient is still on the line and can be told to call back, whereas a document that fails to file leaves no such signal. Assort covers scheduling around the clock, intake, medication refills, lab requests, waitlist management, payment resolution and outbound outreach, and names only athenahealth among a claimed fifteen or more platforms. The question worth asking is what the agent tells a patient when it cannot reach the practice system, because that script is the outage experience from the patient's side. A graceful message and a callback promise are a very different outcome from a silent failure or a dropped line.

### Valerie Health: Compliance Stated, Availability Not

Valerie Health states SOC 2 Type II, which three vendors here do not, and that is worth crediting because it indicates controls were tested over a period rather than asserted. On availability it publishes nothing at all: no uptime figure, no service level agreement, no status page and no description of outage behaviour. It names no EHR or practice management integrations either, which compounds the difficulty, since you cannot reason about what happens when a system is unreachable without knowing which systems are reached in the first place. It sells Referral Automation, Fax and Document Automation, Automated Scheduling and Growth Analytics. Valerie Health also describes itself as certified under the Health Insurance Portability and Accountability Act (HIPAA), and HIPAA operates no certification scheme, so that claim is worth a question about what was actually assessed. Pair that question with one about availability, since neither is answerable from the site today.

### Tennr and Brevium: Nothing Published on Availability

Neither publishes an uptime figure, a service level agreement, a status page or any description of outage behaviour. Tennr publishes no pricing, named integrations or security claims either, so the silence is consistent rather than selective, and its enterprise buyers typically extract these answers through procurement rather than from a website. It carries a $101M Series C with more than two hundred stated providers. Brevium names Nextech, athenahealth, ModMed and eClinicalWorks, claims more than fifty systems through its Stitch layer, and describes a BAA-backed, HITRUST-aligned environment; Brevium's HITRUST-aligned is not the same as HITRUST certified, and SOC 2 is not mentioned. For both, availability is a procurement conversation rather than a published fact. That is workable for a buyer with a security review process and leverage to demand answers in writing. It is considerably harder for a small practice signing a standard order form, which is the practical reason published disclosure matters more at the smaller end of this market than at the larger one.

## How to Decide What to Require

Decide what you actually need before asking, because the honest answer for most practices is that a status page matters more than a percentage. A target without a remedy changes nothing on the day, whereas twelve months of published incidents tell you how a vendor behaves when things break and how quickly it says so.

Then put six questions in writing. Is your uptime figure backed by a credit, and what triggers it. Does that figure appear the same way everywhere you publish it. Is scheduled maintenance counted in it. Where is your status page and how far back does it go. If our EHR is unreachable, what still works. When it returns, what stops the backlog filing twice. Receiving, queueing and ordering can continue without the EHR; deciding what to do with a held document still needs a person; and returning the result to the chart happens if your EHR supports it. Our guide to[how automation handles documents, duplicates and downtime](/articles/ai-fax-automation-documents-duplicates-downtime)covers the recovery problem,[vetting a healthcare AI vendor](/articles/vet-healthcare-ai-vendor-baa-trust-center)covers the wider security questions, and[what vendors publish about data export and residency](/listicles/healthcare-vendors-data-export-residency-retention)covers the adjacent disclosure gap.

## FAQs

Common questions about vendor behaviour during an EHR outage.

### What Happens to Incoming Faxes if the EHR Goes Down?

Transmission and receipt continue, because they do not depend on your practice system. What stops is anything that writes a result into a chart. Those items queue somewhere, and the question worth asking is whether staff can see that queue while the outage is ongoing.

### Is a Published Uptime Figure the Same as a Service Level Agreement?

No. A figure is a target, typically self-measured and often excluding scheduled maintenance. An agreement adds a remedy, usually a service credit, plus a definition of what counts as down. Among the nine vendors read for this comparison, none published an agreement with a remedy.

### What Should a Healthcare Vendor's Agreement Include?

A measurable availability target, a definition of unavailability, how it is measured and by whom, whether scheduled maintenance counts, and a remedy when the target is missed. Without that final element the rest is descriptive rather than binding, and nothing happens when the number is not met.

### How Do Practices Avoid Duplicate Filing After an Outage?

By identifying duplicates before filing rather than afterwards. Ask what attribute the software matches on, whether detection happens automatically on recovery, and whether a person reviews the backlog before it releases. Two staff working the same backlog from opposite ends is the usual cause.

### Should a Practice Require a Public Status Page?

A status page tells you more than any other disclosure available here, because it provides history rather than a target. Only one vendor compared here publishes one. Requiring it outright would rule out most of the category, so treat its absence as a question about incident communication instead.

### Does Scheduled Maintenance Count Against Uptime?

Commonly not, and that exclusion is where a meaningful share of real unavailable hours sits. Ask how maintenance windows are scheduled, how much notice you receive, whether they fall inside clinic hours, and whether the published figure includes or excludes them.
