Voice input in a medical information system: how to dictate an appointment protocol and how it differs from a medical dictation tool

Potok transcribes Russian speech; this site has been translated into English for convenience.

There are three ways to fill in an appointment protocol by voice. Specialized integrations like Voice2Med are built right into a clinic's medical information system (MIS), a decision at the level of the medical organization, not for personal installation. Western medical dictation tools do not state Russian. And an ordinary app like Potok works on top of any program: it is not built into the MIS, but it does not require arranging anything with it.

Three ways to dictate a protocol instead of typing

A doctor who wants to give up the keyboard at least partly really has three different classes of tools, and they solve different tasks. The first is a specialized integration built right into the clinic's medical information system. The second is an international medical dictation tool designed for the English-speaking market. The third is a universal voice input app that is installed on the computer separately from the MIS and works in any of its text fields the same way as in email or a messenger. I will go through each in order and say honestly where Potok (which belongs to the third class) has a limit.

Voice2Med and similar integrations: voice input built right into the MIS

In Russia such a solution exists and has been working for several years. For example, the Voice2Med service from the CRT group of companies (SpeechPro). This is not a separate app that a doctor installs on their own, but an integration: the system is connected to its MIS by the clinic itself, usually through a contract with the vendor.

According to CRT, the service supports voice input for nine specialties: from internal medicine and pediatrics to CT and MRI diagnostics, ultrasound, endoscopy, cardiology and surgery. Moscow radiologists, as reported by ComNews, filled in more than 210 thousand protocols by voice. The vendor states a noticeable saving of time on filling in a protocol; look at its materials for exact figures.

The main difference of such systems is specialized post-processing of terms for a specific specialty. For example, the phrase "dvadtsat pyat yedinits Haunsfilda" (twenty-five Hounsfield units) is automatically turned into "25 HU". Such integrations usually go through SEMD (the Russian structured electronic medical documents) and are tied to a specific infrastructure, for example the unified digital healthcare platform of Moscow.

The flip side: this is a solution at the level of the medical organization, not a personal choice of the doctor. You cannot connect it yourself, without the clinic and without a contract with the vendor: an integration with exactly that MIS which is on the workplace is needed, and usually this is a separate project with the clinic's budget.

Dragon Medical and foreign medical dictation tools: why this is not an option in Russian

Dragon Medical One (a Nuance product, now part of Microsoft) is one of the best-known specialized medical dictation tools in the world: cloud speech recognition, templates for clinical documentation, integration with foreign EHR systems.

The problem for a doctor in Russia is simple: Russian is not stated among the languages supported by Dragon Medical One. The product is aimed at the English-speaking and Western European market, so even a purchased license will not recognize Russian speech: this is a matter of the language model, not of settings.

Potok: dictation on top of any MIS, without an integration

The third way is not to be built into the MIS but to work on top of it, at the operating system level. This is how Potok works. The app is installed on the doctor's computer, and then everything is simple: hold the hotkey (Ctrl+Space), say the text, release, and it appears in the field where the cursor was. It makes no difference which program it is: the examination protocol window in the clinic's MIS, the Complaints field in an outpatient card or an ordinary text editor.

The strong side. No one's integration is needed. Potok is not tied to a specific MIS: it works the same in any system, because it does not know at all that in front of it is an MIS and not email. One doctor can install it alone, without approval from the IT department and without the clinic's contract with a vendor. Recognition is built on its own model trained for Russian speech, including names, terms and a mix of Russian with Latin names and abbreviations.

The honest limit. Potok is not medical software, is not connected to SEMD and does not do special post-processing of format like "Hounsfield units": it simply turns speech into text, and then this text lives by the rules of the MIS into which it was inserted. The personal dictionary in the Pro subscription lets you teach it your own terms and abbreviations, but ready templates for a specific specialty, as in Voice2Med, are not there out of the box.

Table: how the three ways differ in practice

What suits whom

Voice2Med and similar integrations: if the decision is made by a clinic or a network, is ready to invest in a project, and the specialization is specific (radiology, ultrasound, endoscopy), where automatic post-processing of the format of measurements matters.

Dragon Medical is in fact not an option for Russian-language practice: it is worth looking at only if you work in an English-speaking jurisdiction.

Potok: if the decision is made by the doctor (or a small clinic with no budget for an integration), you need to start today and not after approval with the IT department, and it is enough for the text simply to appear in the MIS field and not be automatically formatted for a specific specialty.

Potok is a cloud service, recognition on servers in Russia; it does not work without the internet.

Frequently asked questions

Built-in dictation and Potok

Primary sources: support.microsoft.com · support.apple.com