You may have thought from the title of this post that I was going to post some vampire fan fiction. While this wouldn’t be the first time someone thought I was a vampire (that happened years ago collecting blood at night in Haiti for a lymphatic filariasis survey), that’s not really my thing. Last time I talked a bit about the differences between clinical data collected on the Case Report Form (CRF) and non-CRF laboratory data. For today’s post I’m going to walk you through the life-cycle of a specimen and how my team ensures that every specimen possible can be used for testing and subsequent analysis.
The life of a specimen starts at a local lab when the study protocol indicates that a sample is needed for particular testing at that specific visit. The vast majority of this decided ahead of time when the protocol is being finalized. There are specific tests that need to be run at specific time points, either before and/or after treatment or vaccination. For example, at the peak immunogenicity time point post-vaccination, there are specific immunological assays that have to be run to determine if the vaccine has elicited an immune response. For the sake of brevity for this post, I’ll defer discussions on what immunological assays are run for another post. Try not to be too overcome with anticipation.
The tube, or tubes, of blood collected at the clinic are sent along to a local lab to be processed and to have some safety labs run. You’ll remember from a previous post that the type of lab data that I will be opining/educating about is the non-safety lab data for clinical trials. Accompanying the vial(s) of blood is often a written form that includes an inventory of the vials in that shipment and some metadata surrounding the vial, including participant ID, visit number, visit date, specimen type, etc. Now, I want you to pay particular attention to this seemingly minute detail. Because now we have metadata for that specimen entered in the CRF (the lab tech had to check off in the CRF that the specimen was collected and that check produces metadata around the participant ID, specimen type, visit number, data and time collected for the specimen, and all of that is recorded and retained in the clinical database). We also have that metadata on the physical sheet that goes along with the specimen to the processing/local lab. One of the tenets of data management is that if the same information is entered in multiple places, there will likely be errors.
Right now our specimen (i.e. vial of blood) is at the local lab or processing lab to be processed into plasma or serum or cell pellets. Those blood products are aliquoted out and stored either at the local lab or often, at a repository. Now don’t think that all those little tubes are sitting in freezer boxes all nameless. All that metadata that was entered into the CRF and transferred to the lab form is now entered into a Laboratory Information Management System (LIMS). LIMS systems are used to manage all the information around specimens and assay results. If you’re keeping track of our specimen metadata, we now have metadata for the specimens in the CRF, on a physical form and in the LIMS. And every little aliquot (tube) that was derived from the single specimen has that same metadata associated with it.
Now a testing lab is ready to perform testing on a designated aliquot, as outlined in the protocol. The specimens are shipped to the lab with a shipping manifest that contains an inventory of the specimens in the shipment. The specimens’ bar codes are scanned into the receiving lab’s LIMS system and now the fun can begin. For those of you keeping score, the metadata around the specimen now resides in: 1) the CRF, 2) the lab form, 3) the LIMS installation at the processing lab, (4 the LIMS installation at the repository (if one is being used), 5) the LIMS installation at the central or endpoint lab…and a partridge in a pear tree. As you can imagine, having the specimen metadata replicated in all these different places can lead to errors occuring as a consequence of data transfers and being perpetuated through all the downstream locations. This is where my team comes in. We programmatically compare the specimen metadata in the CRF to the metadata in LIMS. The goal is to identify and correct all errors before the specimens are shipped out to the labs peforming the testing. In order to accomplish the daring feat of data management, we have a crack team of programmers supporting us and creating and maintaining the code that does the comparison and spits out reports with errors on it.
Of course, nothing is ever as simple as “generate a report and be done”. The lab data managers on my team work very closely with clinical sites and labs to determine the source of the error and what the definitive source of any given metadata is and to ensure that changes are made in all places where the metadata may be incorrect.
So way all this effort to ensure that a visit date for a specimen is correct? Does that really make a difference in the grand scheme of a whole trial? Channeling our inner consultants, let’s unpack that assumption. Due the complexities of participants that are on PrEP or the fact that HIV vaccines illicit anti-HIV antibodies, HIV diagnosis for clinical trials follows a testing algorithm where specific tests are dictated by the results of previous tests (confirmatory testing) or vist type in the study (i.e. before or after vaccination). This is actually done for HIV testing outside of clinical trials as well. There is a required confirmatory test if you test positive by a rapid test, the same way a woman would go to the doctor for a confirmatory pregnancy test. https://www.cdc.gov/hiv/testing/laboratorytests.html But I digress, as I mentioned the HIV diagnostic testing algorithms can differ by visit. If the wrong algorithm is run on a specimen because the visit number was incorrect in the metadata, it could lead to the wrong result for the participant. That’s obviously not something anyone wants to happen.
While that example is on the extreme end of the spectrum of what ifs, metadata errors for other values can lead to the incorrect testing being performed for other tests, which would lead to incorrect data ending up in the dataset for analysis. If the lab data are being used to evaluate study endpoints, the quality of the lab data is paramount. One of the main goals of my group is to make sure that the lab data used for analysis is as clean as possible and that each data point is a valid data point.
From an ethical standpoint, ensuring that each specimen collected from a participant can be used is critical. Clinical trial participants are a special breed of people who are willing to be part of these studies, sometimes not for immediate benefit to themselves but for the advancement of the science toward a cure. The whole study team is dedicated to guaranteeing that a participant’s involvement in a trial isn’t for naught. Our small contribution to that guarantee to try and make sure that any specimen they give as part of the trial is tested and that data used for analysis and that participants aren’t brought back for additional specimens uneccesarily because no one can find their initial specimen.
I hope that I have convinced you that specimen management is a vital part of the clinical trial process. Please add a comment if you have any questions about the process or why we’ve invested so much time and energy into it.
Up next time…I get back to my “how to run a team” posts with an update of a team retreat we just had.