Dictation software for locum radiologists

A locum tenens radiologist reads for a hospital this month, a specialist imaging center next quarter, and covers a colleague’s leave at a private practice after that. Each site has its own Picture Archiving and Communication System (PACS), its own reporting interface, and its own way of getting images from the scanner to the screen.

According to staffing vendor Medicus Healthcare, roughly 29% of radiologists have worked locum tenens positions. That share is growing: a Journal of the American College of Radiology study cited by Medicus found that job-switching rates among U.S. radiologists rose by more than 60% over the past decade. The Royal College of Radiologists’ 2025 census reports a 32% consultant workforce shortfall in the UK, which is driving similar patterns of radiologists covering multiple sites.

The dictation problem that comes with this flexibility rarely gets discussed. Enterprise radiology dictation software was built for institutions: a single PACS vendor, an IT team, a long-term contract. When your workplace changes every few months, you quickly find that the mainstream tools were not built for you.

This guide explains what locum and moonlighting radiologists actually need from dictation software, and where the main products succeed or fall short. For a broader comparison of the market, see the medical dictation software guide for radiologists.

Why does multi-site reading break most dictation software?

Enterprise radiology dictation is built around deep integration with one institution’s infrastructure. Nuance PowerScribe connects to a hospital’s PACS and RIS via its PowerConnect interface. It is provisioned by that institution’s IT team, configured for that site’s reporting workflow, and stores voice profiles on servers that belong to that organization. It is excellent software for a hospital radiology department. For a locum who starts a new contract on Monday, it is often simply unavailable.

Dragon Medical One handles multi-site work better than PowerScribe in one respect: a portable voice profile follows you across devices and logins, which means your voice model travels with you. Dragon Medical One has won Best in KLAS for clinical speech recognition six years running, and that portability is a real advantage. The problem is that it is a cloud subscription product, and every word you speak is processed on Microsoft Azure. You need a live, stable internet connection at every site, every session. When you are reading from a rural hospital with unreliable bandwidth or a clinic whose network has restrictions you cannot control, that dependency becomes a real obstacle.

Dolbey Fusion Narrate and Augnito have the same structural issue: audio goes to a vendor’s cloud server, a subscription bill follows the license, and you pay whether you are actively working locum or between contracts.

What do locum radiologists actually need from dictation software?

The requirements are straightforward when you spell them out.

Works in any app. A locum radiologist may read into Epic at one site, a browser-based PACS reporting tool at the next, and a standalone RIS at a third. Dictation software that only works inside its own portal, or that has optimized integrations for a handful of named systems, cannot follow you across all of these. Cursor-based dictation solves this: the software acts like a keyboard, typing your spoken words wherever your cursor is, in any text field, in any application.

Works offline. Not every reading room or on-call setup has fast, reliable internet. A tool that processes every word in the cloud exposes you to connectivity failures mid-dictation. On-device speech processing removes that dependency entirely.

Self-installs in minutes. When you arrive at a new site, you may have no IT support and no time to wait for provisioning. Dictation software that requires enterprise IT rollout, vendor onboarding, or coordination with the host institution’s system administrator is a practical barrier. A self-contained installer that does not need admin rights or server configuration fits the locum model.

Radiology vocabulary. General medical dictation handles common clinical terms, but radiology reports use specific anatomy, modality names, measurement conventions, and structured report phrasing. A tool trained on radiology language reduces correction time and improves throughput from the first session.

No per-device license complexity. Locum radiologists read on their own machines across multiple sites. Dictation software that ties a license to a single device, or that requires per-device activation through a vendor portal, adds friction every time you set up on a new workstation.

Does the subscription model make sense for locum work?

The subscription question deserves a direct answer.

Dragon Medical One is widely cited at approximately $79-99 per provider per month (reseller-cited estimates; Nuance does not publish pricing officially). At $99 a month, you pay $1,188 per year. If you work locum for eight months and have four months between contracts, you still pay for all 12. The subscription does not adjust to your work schedule.

Dolbey Fusion Narrate is published at $850 per user per year, plus a $500 onboarding fee. That is a one-year minimum commitment before you have seen a single contract.

Enterprise tools that are provided by the host institution solve this problem for you: you use the site’s license, you pay nothing, and you walk away when the contract ends. But those tools often require IT provisioning time, may not be available to short-term locums on a given site’s license, and do not travel with you when you move to the next engagement.

A one-time payment removes the mismatch between your billing cycle and your work pattern. You pay once, own the tool, and use it whenever you work, at any site, without a recurring bill. See the PowerScribe alternatives guide for a full cost comparison across the major radiology dictation tools.

What about patient privacy when reading on a personal laptop?

Locum radiologists often use their own machines. A MacBook Pro or a Windows laptop travels with them from site to site. When that machine runs cloud dictation software, patient audio is captured locally and then sent to a vendor’s cloud server for processing.

This is worth thinking about carefully. The audio that reaches a cloud server contains the content of radiology reports: patient details, clinical history, findings, and impressions. A Business Associate Agreement covers the vendor’s handling of that data, but the audio still leaves your device and transits a vendor’s infrastructure.

On-device dictation processes speech entirely on the local machine. No audio leaves the device, ever. This is privacy by architecture, not by policy. For a locum reading on a personal laptop that connects to untrusted hotel or clinic networks, removing that cloud transmission is a meaningful reduction in risk.

What about radiology residents who moonlight?

Many radiology residents and fellows take on moonlighting shifts at outside hospitals or urgent care facilities during their senior training years. The dynamics are similar to locum work: new site, different PACS, limited IT access, and a need for software that installs quickly and works in whatever reporting interface the site uses.

RadMyk is free for radiology trainees for the entire length of their training program, so the cost question does not apply. Setup takes the same few minutes it would for any attending. The free trial and trainee access details are at radmyk.com/pricing/.

Where do enterprise tools genuinely win?

It is worth being direct about this.

PowerScribe offers structured reporting, laterality and gender conflict checks, peer review modules, and deep PACS/RIS integration via PowerConnect and PowerCast. These are real capabilities that matter in a hospital quality assurance workflow. A locum radiologist typically cannot access them on a short-term contract anyway, but the honest summary is: if the site has PowerScribe and wants you to use it, use it.

Dragon Medical One has a portable voice profile that follows clinicians across devices and facilities without re-training. For locums who work in Dragon-enabled institutions, that is a genuine convenience. Compared to a setup that requires installing a local engine on each new machine, the cloud profile is a simpler experience when the network cooperates.

Is RadMyk the right fit for locum and moonlighting work?

For the specific needs of multi-site and locum reading, the fit is strong.

RadMyk runs on macOS (Apple Silicon, M1 or newer) and Windows. It installs without IT involvement. After a one-time download of the on-device speech engine, it runs fully offline. It types at the cursor in any application: a PACS report field, a RIS text box, a browser-based reporting tool, a Citrix or remote desktop session. Anywhere you can place a text cursor, you can speak.

The speech model is tuned for radiology language, with 96.1% word accuracy measured out of the box at approximately 220 words per minute. A guided calibration step tunes the model to your voice and microphone on the first run.

One license installs on two devices and can be active on one at a time, which matches how most locum radiologists work: a primary machine at home and a backup laptop for travel.

RadMyk is not a structured reporting platform, does not have peer review, and does not integrate with PACS at the reporting-workflow level. For front-end dictation, which is what most locum and moonlighting radiologists need, it covers the job.

The current launch price is $199 one-time for the first 100 users, rising after that. A 28-day free trial is available with no credit card required. Start the trial or buy outright at radmyk.com/pricing/, or learn more about how RadMyk compares to the full range of dictation options at radmyk.com.