Automated Data Exchange: Approaches to Vitalize your Cancer Registry

Q & A Transcript

Q – If automation helps lead to concurrent abstracting and reporting, I wonder if it would benefit my clinical trials and research departments?

A – Most definitely.  Any effort to decrease the latency of data collection would address the concerns of researchers and other stakeholders who currently believe they can’t use the registry because it is not current.

Q – I currently just use a casefinding import from Epic. Could I utilize a pathology import too?

A – Yes. There is no limitation with regard to how many automated interfaces can be executed at one time. There are data distinctions that can be set within the abstract to identify one data feed vs another, even if they run on the same day. Stacking automated interfaces can provide not only needed data elements but also a mechanism to create a more complete casefinding solution.

Q – What are some examples of additional data that can be extracted from the EMR other than just for casefinding?

A – Depends on the software of course, but we know that certain EMRs carry discrete information such as AJCC staging, elements and TNM values. Some EMR’s have developed oncology modules that carry treatment data that can be retrieved but also we’ve seen EMRs work to standardize data collection and the initial data capture. But, standardization of data often leads to more discrete data retrieval opportunities.

Q – Our facility is thinking about using Oncolens as their software tool for Tumor Board management. Would data integration with CRStar be beneficial? 

A – Integration with tumor board software solutions will offer a unique opportunity to process cancer data on multiple levels all at once. Beyond the integration of the actual meeting notes, the committees speak to the patient engagement across the abstract spectrum. You can derive casefinding opportunities, as well as diagnostic and staging information, and you can get planned courses of treatment.

Q – Is ERS working to automate data exchange between pathology or surgical synoptic reports into the ERS data collection software?

A – Yes, once we’ve addressed some of the technical challenges involved with pathology interfaces that we talked about earlier, we can automate that data into CRStar. There are considerations in making synoptic report data elements discrete, which have their own challenges, but we do currently have the ability to bring in the complete path report, synoptic report and all, as a text component of the casefinding record. We know that the synoptic reports add an additional layer of valuable cancer registry information on top of the pathology. We’re working with registrars to ensure we make the best use of that information.

Q – Won’t most of these integrations still require cancer registrars to review the data for things like multiple primary histology rules and coding errors?

A – Yes, we don’t want anybody to think that this effort in integration is a method of taking registrars and the registry work that they’ve been doing out of the picture. That’s not the case at all. They will be critical with regard to knowing how to massage the data, how to validate the data, how to make the best use of the data. They will be integral with regard to pulling the levers of these automations at the right time – to say we’ve gotten the right data in and now we’re going to see this data show up in it’s complete format, available across the health network. We really see the registrar’s role as stewards of this data, it always has been and it always will be no matter what automation solution comes into play. Registrar’s are going to be needed more so than ever, even with these integrations.

Q – Is NLP used now or something ERS would like to implement?

A – ERS has been evaluating and testing NLP technologies for almost a year. At this point, we are ready to implement this at a few pilot sites.

Q – What is the value of interfacing with the RadOnc system?

A – The value in an interface with a radiation system can be seen in multiple steps. First, it can speak to the common level that registrars collect radiation today, as most of the popular RadOnc systems measure radiation in the same manner that the cancer registry 2018 regulatory changes shift too. Second, there’s an opportunity to collect records for casefinding that may not necessarily be found in a pathology interface. Third, it can also provide ongoing followup on the patient if radiation treatment extends into boost treatments and other subsequent radiation.

Q – Does the data get imported as a coded value or does the CTR need to do this?

A – By and large, the goal of these interfaces is to provide as many coded values as possible. We find that even when external data sources have coded values, they are not using the cancer registry standards of coding. So in those instances, a crosswalk can be done during the time of import. With an NLP solution, the software intelligence is set to further standardize data which provides a mechanism to also codify the data.

Q – How do interfaces help with Precision Medicine?

A – Cancer registry is a great resource of a variety of different types of information.  There’s a depth of information pertaining to population health – looking at how patient phenotypes and profiles and seeing how they reacted to specific treatment profiles. This is historical.  It could also be a source for identifying patients for specific treatment protocols – particularly if the information is kept reasonably current.