A particular genre of healthcare technology article reappears every few years: the declaration that fax in healthcare is finally dead, or dying, or will be dead within 18 months. The argument usually points to EHR adoption rates, the rise of care coordination platforms, the broader digitization of healthcare workflows, and sometimes regulatory pressure to eliminate fax as a primary communication channel.
Each of these articles is wrong in approximately the same way. Fax volume in specialty care, particularly for referral communications between referring practices and specialty providers, has not declined meaningfully. In several subspecialty areas including infusion therapy and durable medical equipment, it has grown as care delivery has fragmented across more provider settings. The people writing those articles are correct about the direction of technology development. They are wrong about the timeline and about the forces that keep fax entrenched in specific workflow contexts.
We built Coral to handle fax-transmitted referral documents. That decision required a clear-eyed assessment of whether fax would actually remain the dominant referral channel for the providers we serve. Here is what we found.
The Interoperability Gap That Fax Fills
The persistence of fax in healthcare is not primarily a technology conservatism problem. It is an interoperability problem. Electronic health records from different vendors do not share records natively in most practice settings. A referring physician's EHR cannot directly send a structured referral to a specialty provider's EHR without a specific integration being in place between those two systems. Given the number of EHR vendors, the variation in implementation, and the different EHR systems used by different practices in a geographic area, those direct integrations exist for a small fraction of referring relationships.
In the absence of a direct EHR-to-EHR integration, practices need a communication mechanism that works regardless of what software the receiving end uses. Fax is that mechanism. Every practice has a fax number. Every practice can receive a fax. The information encoded in the fax document is not constrained by the receiving system's data schema. A referring physician using System A can send a complete referral to a specialty provider using System B, C, or no EHR system at all, through fax, without any prior arrangement.
Direct secure messaging (DirectTrust) and interoperability initiatives like the Da Vinci Project aim to change this, but implementation is uneven. Large health systems with dedicated health IT teams are further along. Independent practices with two to five physicians are much further behind. The small specialty infusion center and DME provider market sits squarely in the less-connected segment of the provider ecosystem.
The Network Effect Problem
Even if a better technology exists, adoption requires both sides of the communication pair to use it. A specialty infusion center could adopt a modern referral intake portal and still receive faxes because the majority of the physicians who refer to them have not adopted the same portal. The specialty provider has limited leverage to force referring practices to change their workflow. Referring physicians send to many different specialty providers. Asking each of them to use a different portal is a large ask with no obvious benefit to the referring practice.
This is why "just switch to a portal" does not work as a complete strategy. The specialty provider can offer a portal, but they cannot guarantee that referring practices will use it. In practice, many specialty providers who build or adopt referral portals find that they receive a mix of portal referrals from high-volume referring practices and fax referrals from everyone else. The fax channel does not disappear; it remains the fallback for the long tail of referring practices.
The implication for a referral intake tool is that it needs to handle fax as a primary channel, not as a legacy edge case. A tool designed around portal submissions that handles fax as an exception will serve the long tail of referrals poorly, which in most practices is a substantial fraction of total volume.
What the Data on Fax Volume Actually Shows
Healthcare fax volume data is harder to find than you might expect, partly because fax infrastructure is fragmented across thousands of practices and partly because no single entity tracks it systematically. The picture that emerges from the available sources suggests that while overall fax volume in healthcare has changed over the past decade, the rate of change is much slower than the "fax is dead" narrative implies.
Virtual fax services, which route fax to digital inboxes rather than physical machines, have grown significantly. This is sometimes cited as evidence of fax's decline, but it is more accurately a modernization of fax infrastructure rather than its replacement. The documents still travel as fax-encoded transmissions; the receiving end is a software inbox rather than a thermal paper printout. The document processing challenge, reading and extracting information from fax-quality images, is essentially unchanged.
For infusion centers and DME providers specifically, the referral fax is often accompanied by clinical documentation from referring practices that is itself a faxed scan of a printed clinical note or a printed EHR export. Even when the referring practice uses an EHR, the documentation format that accompanies referrals is frequently a PDF that was printed and faxed, not a structured electronic record. The content is digital at the source, but it arrives as an image.
What This Means for Automation Tools
Building document extraction tools that assume high-quality, structured digital input is building for the wrong channel. Referral faxes are lower-resolution images than scanned documents produced in a controlled environment. They vary in quality based on the sending fax machine, the scan quality at the source, and the transmission quality. They arrive in layouts determined by the referring practice's forms and letterhead, not by a standard the receiving practice controls.
Coral is specifically built for this environment. The extraction pipeline handles fax-quality image inputs as the baseline, not as the degraded case. We have worked through a substantial range of document layouts from different referring practice types, including layouts that are common in the Southeast where we are focused. The confidence scoring system is calibrated on the real-world distribution of fax document quality, not on idealized inputs.
We are not saying fax will never be replaced. The regulatory and interoperability trajectory is pointing toward more structured electronic exchange. The CMS FHIR API requirements are accelerating that direction for Medicare Advantage. But "eventually" and "by 2027" are not the same as "now." The providers who need help with referral intake today are receiving referral faxes today, and they will be receiving them for several more years while the interoperability infrastructure catches up.
Building automation tools that only work after the fax problem is solved means building tools that cannot help anyone until the problem they most urgently need help with has already been solved by other means. We made a different choice: build for the channel that exists, with the quality that exists, and let the transition to better channels happen at its own pace without disrupting the clinics that are waiting for the technology to catch up to the ambition.