Analyzable Data is Submittable Data: Interim or Final SEND (Live Webinar)
Live Industry Discussion – how analysis-ready data, in short, leads to submission-ready data.
October 19 2021
From the way described, the CRO has to actively present the study data for upload (at the required
frequency) – I assume this must be a file-neutral format – has the CRO had to devise ways to extract that
from their LIMS – how is the study structure made available? – has PointCross assisted in that? I assume
the CROs are charging for this data provision.
A1. PointCross: The CROs load the data (in whatever native format is allowed by their LIMS system) directly into the secured and encrypted Virtual Data Room (VDR), which is a space made for the CRO or the Sponsor customer. LIMS systems are computers with the ability to export selected data for a study in digital format including their native data model, columnar formats including CSV, XML , XPT, EXCEL or other; and human readable tabulations in PDF format.
Because interim study data comes from CROs as well as external labs or research facilities, the disparate data is characterised as raw, as-collected, and oftentimes unstructured and must be mapped sequentially – a process PointCross assists through our Data Concierge Service.
The first time a CRO is brought on line to provide a Sponsors’ interim study data, a sample study data upload is made -so that the automated transformations to the Universal Data Model (UDM) of the interim and study repository – can be mapped for re-use in future uploads from that CRO. This takes a few hours, and the subsequent mapping and automation renders the raw data to be made available for visualization and analysis in near-real time.
CROs usually know how to generate the data exports. If their staff are unfamiliar with the controls, PointCross aids the data transfer process and can help enable CROs to automatically upload data from specific ongoing studies at a pre-defined frequency, and if or when necessary, automation scripts will be provided. The CRO and Sponsor are kept on the same lines of communication at all times.
By “Study Structure” we assume the question is referring to the way the Study Design is coded. Some LIMS systems attempt to generate SEND-like domains for Trial Design. These work most of the time for simple parallel studies. However, the automatically generated Trial Designs in some of the LIMS systems tend to be incorrect. Since the Sponsor is intimately aware of the Protocol, a simple tool for generating the Trial Design domains is provided to the sponsor and they will establish the trial design domain before the first load is obtained. Having an accurate Trial Design is key to being able to analyze data efficiently and with confidence.
Dessi: We haven’t had any requests for additional fees from our CRO’s for uploading interim data. Prior to PointCross, the CRO would send us data every few days for a study in excel or PDF format. When we brought PointCross on board, we established a weekly data upload using XPTs from their data collection system (which is typically Provantis or Pristima) and/or from LIMs. As such, the process for the CRO has become more efficient and less cumbersome in getting data to us for interim study monitoring. Additionally, they use the VDR for study documents (protocols, amendments, etc) which has also proven to be easier and more efficient for them. In general, the process has streamlined our communications and data transfers with the CRO reducing theirs and our overall workload.
Dragomir: We have different arrangements with our CRO partners. One of them is charging the same rate for daily and weekly extracts and since everything on our end is fully automated, we receive daily extracts from their test sites. Another CRO charges based on the number of extracts for a study, so unless required by the Project Team specifically, we go with weekly extracts from them. The timing of the [weekly] extracts can be optimized, so that relevant data (e.g. lab findings) become available for review as soon as possible (typically two days after blood collection for the lab data).
Is it very cumbersome for the CRO to upload the data? Does it add much cost in your experience?
A2. PointCross: CROs and labs have a LIMS system for the digital data they collect. Data extracts from the LIMS systems can be directly loaded into the encrypted Virtual Data Room (VDR) of Xbiom™. In the past few years most of these LIMS systems have a machine readable column format in XPT, CSV, XML or even Excel. This data can be “dragged and dropped”, or selected for upload into the secure VDR in the allocated folder structure organized by Study. Automated upload is configured for repeat customers, as this is simpler and less effort for the CRO.
When data is not sourced from a LIMS system, we accept printed (e.g. PDF, or hand-written) and annotated lab notes, and PointCross digitizes and enters the data into Xbiom on behalf of the Sponsor or CRO, as a Service.
This is not an expensive or cumbersome process. PointCross charges, for example, a fixed monthly subscription for supporting interim study monitoring on a fixed number of studies per month. Additionally, multiple studies that have come in in non-English language have successfully (and quickly) been loaded into Xbiom™.
Dessi: See my answer above, it can also apply to this question.
Dragomir: Once the data transfer process is tested and established (your first study upload[s]), typically there no issues afterwards. Nowadays, this is pretty routine practice for the major CROs (might be a different story for the small, niche CROs).
Does PointCross need to be involved in the interim data uploads, or is the CRO in direct contact with
A3. PointCross: The sponsor can engage with the CRO independent of PointCross. PointCross’s Data Concierge Service is designed to “plug-in” with the Sponsors’ CRO to ingest the interim data load, process and map the interim data, and render the data available in Xbiom™ for visualization and analysis.
This Data Concierge service is leveraged by sponsors with limited internal data science resources or Sponsors who prefers to have the CRO and PointCross deliver analyzable data. In such cases, a tri-partite collaboration is maintained with PointCross only involved in making sure that the arriving data and its integration is checked for completeness and integration; the sponsor is entirely responsible for the interaction with the CRO.
If the CRO data is in SEND .XPT format, it is seamless, and the CRO can directly load the data into Xbiom without any involvement by PointCross’ data standardization team. However, non-SEND, unstructured, or alternative formatted LIMS extracts require data curation and conversion to a universal data model to enable the data to be visualized and analyzed in Xbiom, and this normally takes only a few hours upon receipt of the raw data.
What are the implications of using SEND data that is not considered GLP to interpret a study?
A4. PointCross: A study report in a GLP study is considered GLP because the data in the study report is traceable back to the LIMS system, and it is audited before signature. During interim monitoring, the only data that is used is what comes from the CRO using their LIMS system as the source.
SEND, on the other hand, is a separate process done by separate people who take the GLP LIMS data and reconstruct it according to the SENDIG model and CT to generate SEND. There is no traceability required, or possible, and it is entirely up to the SEND preparer to confirm that it is capable of regenerating the same summary data as reported in the GLP Study Report.
Because a 100% consistency check of thus prepared SEND against the Study Report is considered expensive by many SEND preparers – they resort to a spot-check. PointCross does a 100% check by comparing it against the SRR, which is a digital extraction of the Study Report.
The implication, in our perspective is that the user of SEND data must be “aware” of whether the data has been 100% checked against GLP or not before making interpretations or decisions based upon that data.
Dragomir: According to the GLP, the Study Director is responsible for all aspects of the study, including the study results interpretation. The individual and summary results as well as some statistical analyses as defined in the Study Protocol (Plan) are typically created within the validated LIMS at the Test Facility. SEND like data (extracts from the LIMS) while study is ongoing and the SEND packages created by the CRO SEND teams or third party are not audited and there is no regulatory requirement for these to be GLP-compliant. This data is used by the Sponsors and FDA to generate their own opinion on the study outcomes.