A radiologist reading remotely is not short of pixels.

The RealityOne worklist or several is the whole design

⚠️

The Problem

The study arrives without its history.

Current images route cleanly, priors do not, and the reading radiologist is comparing against nothing while a report deadline runs.

🔧

The Solution

Route priors with the study, or make them fetchable.

Decide per modality what history follows a study, and give the reader a way to pull anything else without a phone call.

✅

The Resolution

Reads that are as safe remotely as on site.

Comparison available, indication visible, worklist ordered by clinical urgency rather than by arrival.

Image transfer and what actually moves

Start by being precise about what has to travel. A study is images plus a header identifying the patient, the accession, the modality and the parameters. What frequently does not travel is the order indication, the prior report and the priors themselves.

Two architectures dominate. Push, where the originating site sends studies to the reading destination on a rule. Pull, where the reader queries a shared archive and retrieves what they need. Push is simpler and fails by omission — what the rule did not send does not arrive. Pull is more flexible and fails by latency, because a radiologist waiting on a retrieval is a radiologist not reading.

Most working setups are a hybrid: push the current study and the most recent relevant prior, make everything else pullable. Deciding what “relevant” means per modality is the configuration that matters, and it is a radiologist’s decision rather than an engineer’s.

Prior studies are the hard part

This is the point most teleradiology projects underestimate, and it is worth stating plainly: a read without the relevant prior is a different and worse read.

Priors fail to arrive for mundane reasons. The patient was registered under a different identifier at the other site. The prior predates the current archive and sits in a migrated system. The routing rule sends one prior and the clinically relevant one is three studies back. None of those produce an error message; they produce a report written without comparison.

So build the fallback deliberately. The reading radiologist needs a way to see that priors exist and retrieve them without leaving the reading application, and somebody has to own the identity reconciliation between sites. Where several archives and modalities are involved, a dedicated interface layer gives you one place to fix a mapping rather than several.

Transmission security: what the rule requires and what it does not

Security here is frequently overstated in one direction and understated in another, so it is worth reading the rule rather than the marketing.

The HIPAA Security Rule’s technical safeguards cover access control, audit controls, integrity controls, authentication and transmission security, and require technical measures guarding against unauthorised access to electronic protected health information transmitted over a network. Encryption is an addressable implementation specification rather than a required one — which does not mean optional. It means you implement it, or you document why an equivalent alternative achieves the same purpose.

In practice, for images crossing between sites over any network you do not wholly control, encryption is the reasonable and appropriate measure and the documentation exercise is short. For working through what the standards mean concretely, NIST’s cybersecurity resource guide for implementing the Security Rule, revision 2 of SP 800-66 published in February 2024, maps each standard to Cybersecurity Framework subcategories and SP 800-53 controls.

Reading workflow across sites

A worklist is a queue, and the design question is how many queues there are and who sees which.

One shared worklist across all sites balances load well and loses local knowledge — the radiologist who knows that a particular referrer’s chest films always come with a specific question. Separate worklists preserve that and create the situation where one site is backed up while another reader is idle. The usual resolution is a shared pool with site affinity: readers see their own site first and can draw from the pool.

Whatever the shape, ordering by clinical urgency rather than by arrival time is the change with the largest safety return, and it depends on the indication travelling with the study. Getting that right is usually part of rebuilding the ordering workflow rather than a routing change. Which brings the design back to the same place: what moves alongside the images is what determines everything else.

Turnaround time and the measure that misleads

Turnaround is the metric every teleradiology arrangement is judged on, and a single average of it hides the thing worth knowing.

Measure from study completion to report availability, not from when the study reached the reading system, because the second measure excludes exactly the routing delays a remote arrangement introduces. Then report the median and a count of studies beyond a threshold rather than the mean, because the tail is where the clinical risk sits and an average will not show it.

In our experience the outliers cluster — around a modality, a site, an hour of the night, or studies whose priors failed to arrive. That clustering is the useful finding, and it is invisible in an average.

Where the standards are heading

Imaging interoperability is an active area rather than a settled one. ONC issued a request for information on diagnostic imaging interoperability standards and certification, noting that patients and providers still face challenges specific to access, exchange and use of diagnostic images across care settings, and seeking views on whether standards and certification criteria would help.

That request names no specific standard and commits to no rulemaking, and we are not going to imply otherwise. What it signals is that the gap between how well clinical data moves and how well images move is recognised at policy level.

The practical implication for a practice setting up teleradiology today is modest but real: prefer arrangements built on documented, non-proprietary exchange over ones that lock studies into a vendor’s own transport, because the direction of travel favours the former.

• Decide what priors follow each modality. A radiologist’s decision, not an engineer’s.
• Give readers a retrieval path. Visible priors, fetchable without a phone call.
• Encrypt in transit and write the one-page rationale. Addressable is not optional.
• Measure completion to report, median plus tail. Never the mean alone.

Qventive works as a teleradiology workflow setup consultant for practices across New Jersey, as part of our radiology EHR and IT support and a broader radiology practice management engagement. Because studies cross organisational boundaries, the design belongs in your HIPAA risk analysis, and the networks and archives underneath it usually warrant a wider cybersecurity compliance review.

Teleradiology setup: common questions