Medical imaging on desktop
Case Study

Hidden Efficiencies in a Radiology Workflow

RTAT & Non-interpretative Tasks

The time that it takes an imaging study to be completed can be broken down into discrete intervals. Our focus here is the time between a study is performed and the radiologist’s report is made available to the clinician. We define this interval as the report turnaround time, or RTAT.

Workflow representation.

We monitored over ten thousand imaging studies across all modalities and anatomies. For each study, we measured the percent of RTAT that was spent waiting in a worklist before a radiologist’s interpretation. Around 25% of studies spent over half of their turnaround queued in a worklist. This long tail represents an opportunity to improve workflow by reducing the amount of time these STAT studies spend waiting to be read.


Automatically Identifying Priority Levels

RadiLens approached this as a supervised machine learning problem. This model derives features from the imaging study’s metadata such as patient location, study modality, and the reason for the study from the ordering clinician.


Now that we have reliably translated the human intuition of priority labeling into a statistical model, we can deploy it to make predictions in real-time.  RadiLens standardized the long tail of queued minutes so that all reports have a better turnaround time. 

We have focused this case study from the perspective of the radiologist.  This automated prioritization helps them select the next best study to complete to improve their report turnaround time.  However, this automation is valuable for other stakeholders as well. 

From the perspective of a clinician ordering an imaging study, automated prioritization garners trust that their studies will be returned in an appropriate timeframe. 

From the perspective of a technologist, automated prioritization removes administrative hurdles they face to appropriately label studies with complex rules. 

From the perspective of a hospital, automated prioritization ensures the best turnaround for clinical care and minimizing time wastage among their teams.

Machine Learning Study Prioritization standardizes the long tail of turnaround times

RadiLens’ worklist intelligence product looks at each imaging study’s metadata such as patient location, study modality, and the reason for the study from the ordering clinician, to automatically apply a priority level and ultimately present the right study for you to read next. Reach out to us below to learn more & download the full version of this White Paper here:

Learn more about hidden efficiencies in a Radiology workflow

Enter your email address and we'll send you the complete whitepaper as a PDF document.

Request a demo

Enter your email address and a member of our team will be in touch soon.