A great mix of standards and great visions when CDISC met in Vienna

The second visit to Vienna in two month for the author offered a quite different atmosphere than the first visit. The subject of both meetings where about data and how to share and make use of it on a larger scale but the CDISC conference which was the objective this time does this by defining standards so that data can be pooled and compared while the Open PHACTS meeting focusses on linkage of disparate data sources, the goal is though the same - innovation for better and safer cure!

Relating this to the hosting city of Vienna, with its famous composers and orchestras, the different data sources that ship in could be seen as different instruments, all important parts of the symphony but without the rules (the notes) which CDISC defines, it would not sound very nice at all! The picture below shows the Musikverein from where the Wiener Philharmonic Orchestra give us the New Year Concert!

The first day of the CDISC meeting was very much about the standards themselves but the key note was quite inspiring and different. It was held by Dr Amanda Rojek. She told the story about how a clinical trial is done on the field. The field in this story was not the ordinary clinical trial setting but on the ground in the Ebola chaos in Sierra Leone in beginning of 2015! It really set the stage for the conference reminding us all that there is a reality behind the data.

The CDISC campaign UnlockCures.org was presented and a number of nice words of wisdom where shared, two of those where:

"Billions of dollars are spent on clinical research, but when data isn’t shared, results get buried. Along with patients needing cures."

"When Clinical research Data doesn´t speak the same language, the conversation can turn deadly"

A lot of focus was around SHARE. SHARE is a web based Meta Data repository for the CDISC standards and controlled vocabularies. Today it can be used to download all the standards but CDISC is working on an API, slightly delayed, that will make it possible to actually build applications on top of the API, guaranteeing access to the latest updated standards and code lists. A vision delivered by Peter van Reusel was that there in the future would be like a web shop or browser on top of the SHARE API where you could pick the pieces needed for a trial, from design, via CRFs all the way to analysis metadata and algorithms, all interactively and ready to use.

There where updates on the implementation of regulatory frameworks from EMA, FDA and PMDA, all of which will soon require electronic submissions of Clinical Trial data on CDISC SDTM format!

So what is there to come? The SHARE API is of course important but equally important is for example the work on semantic technologies as presented during the second day of the conference. SHARE can already today be delivered as RDF, Both code lists and the standard (the implementation guide for SDTM) itself. Chris Price demonstrated how Roche where actually designing the study data collection tools using RDF. The next big step would be to also store and eventually submit data as RDF. By doing so there would be no need for SDTM or ADaM, it would all be available using for example style sheets based on the RDF data itself.

Semantics could also be used in other areas as shown by William Stevens. He demonstrated how Adverse Event groupings and analysis could be achieved using OWL and reasoning. The approach he showed for analysis, bundled with study data delivered itself as RDF and with the vision of having also the WHODrug as linked data, would open up fantastic analysis possibilities!

FHIR is a hot topic, is it the saviour or just another abbreviation that will die? What is clear is that HL7 has identified version 3 to be slightly too complex for its own best and hard to implement and get right. What FHIR does is to focus on entities of data, not entire documents. You only ask for what you are interested or sends what you have collected. This is especially interesting in the world of mobile devices, wearables and “insidables” where we look at streams of data, not the traditional forms and questionnaires used as the standard in data collection today.

Another interesting presentations was made by Dorina Bratfalean representing CDISC. She presented how to analyse interconnected data sets using Graphs where she used Neo4j as graph database and visualization tool.

The Clinical Trial register is an ongoing standardization work where CDISC has taken the lead. They try to make it align with requirements from the different authorities around the world. The intention is that it should harmonize with both ClinicalTrials.gov, Eudra CT as well as PMDA and WHO requirements – yes, certainly a big undertaking! The big challenge for the future is how to incorporate also trial result data in the register. This would really make the register interesting for research. One option for this would be to use the Data Cube vocabulary as a basis as represented by W3C and the RDF standard.

One of the final presentations was by Anders Helmø Larsen. He turned the focus back to standards as such. The key standards to keep an eye on for the near future, according to Anders, is IDMP, CDISC and OMOP and there will be a big Standard Bang around 2020 when for example IDMP will be ready, I though personally think that the uncertainties around IDMP is still too big to expect it to change the world as envisioned by the presenter. IDMP is still far away from being the spine around which everything else spins.

The meeting was of course not only about presentations. A vital part of the meeting are the interactions taking place during breaks and the evening event where the delegates can meet under more relaxed circumstances. This year the evening event was held in the Piaristenkeller. There was Austrian music, a hat parade, wine tasting and a very nice dinner with Austrian delicacies like Wienerschnitzel and Apfelstrudel.

Again a very interesting and good CDISC meeting. There are fewer exhibitors and less focus on the exhibition in Europe compared to US but the content of the sessions is far more interesting, at least in the humble opinion of the author.