Adherence Analytics

Overview

Once your historical data is successfully parsed, we will populate the dashboard view with a visualization of your imaging exams and the recommendations made within those reports.

Closed-Loops Rate

This shows what the practice's total closed-loop's rate is including the other recommendation status's such as

• Scheduled Not Completed
• Rx Received

And is calculated as follows:

Adherent + Scheduled Not Completed + Rx Received / (Adherent + Overdue + Coming Due + Non Compliant + Scheduled Not Completed + Rx Received )

Historical Adherence

As a baseline, we will populate a historical adherence graph for you that shows counts of your recommendations and where the recommendation's status was relative the due date and the month selected.

The adherence rate is calculated based on the recommendation counts in each of the columns as

Adherent / (Adherent + Overdue + Coming Due + Non Compliant)

*Exclusion - we show you the Non-mature count in the graph however it is NOT included in the adherence calculation

Filters

This graph is filterable by a number of dimensions that you can combine including:

• Radiologist
• Recommended Modality
• Recommended Anatomy
• Timeframe Extracted

There is a hover state that will give you the inputs for that month and individual adherence rate. If you click that month if will also bring you to a bar graph representation of that month.

Aggregate Charts

There's a few core components of the analytics overview that help give insights into how your follow up process is performing.

1. Recommendations Histogram - A time series graph that shows your imaging exams and how many of those exams we tagged that have follow ups recommended at an aggregate level.
2. Recommended Modalities - If a report suggests a follow up with a specific modality, it will show up within this aggregate chart, by default we show the top 5 most recommended and bucket the rest into an others category which you can see in more detail within the Follow up List.
3. Recommended Anatomies - If a report contains a recommended anatomy, it will show up within this aggregate chart, by default we show the top 5 most recommended and bucket the rest into an others category which you can see in more detail within the Follow up List.
4. Initial Modalities - This shows the modality breakdown of your reports with a capability to filter around only the modalities that create the most recommendations
5. Reco Rate Funnel - We tag each report with associated timelines and assign a status value detailed further below and in this view show an aggregate value of those recommendation status'

📘

What do we consider a recommendation?

At a high level, we track sentences within the radiology report that specifically recommend a follow-up imaging exam (e.g. “CT”, “MRI”, “Mammogram”). Our Natural Language Processing, tags each report and extracts key information around recommendations such as:

1. Is there a follow up recommended?
2. What modality is the recommendation for?
3. What anatomy is the recommendation for?
4. When should the patient come in for that recommended exam?

What counts?

• Include cases that specifically recommend follow-up imaging
• Include cases that recommend interventional procedures (e.g. FNA, ultrasound guided)
• Multiple recommendations - we track each recommendation separately if a report includes more than one recommendation

What don't we count?

• Exclude cases that don’t recommend follow-up imaging (e.g. “per guidelines”, “per protocol”)
• Exclude follow-up recommendations for direct visualization, biopsies, or other specialty visits
• Exclude cases with a minimum follow-up timelines greater than 12 months

Recommendation Timeframe

For the recommendation timeframe, we assign some basic rules to what we find in the report text. Ideally, there is a complete timeframe recommendation in the report, but just in case there's not, here's our approach.

Recommendation Status Definitions

After the timeframe is extracted and other data is gleaned from the report, we then assign a status to each of the recommendations based on the time it is prescribed to occur.

Calculating Adherence & Closing the Loop

📘

How do we calculate adherence you ask?

Each recommendation we identify is converted with the time ranges as noted above with either the actual value or the imputed value. On a historical basis, we are able to determine the Baseline Adherence and a probability used for future monthly reporting of adherence

Recommendation Action

Our mission is to enable open and effective dialogue about the clinical necessity, risk and benefit of radiology recommendations. In the triumvirate of the radiologist, referring doctor, and patient relationship, we enable the collection of 1st party feedback through messaging. The intent of those messages then effectively **closes-the-loop** based on that dialogue and is updated automatically (or manually) in the Recommendation Action Status