In my work directing clinical device studies, I sometimes have trials that are highly dependent on laboratory results. These labs may be local or they may be central labs, but most often trials will use a mix of local and central labs.
Often overlooked is just how critical laboratory results are to the trial’s integrity and even the primary clinical endpoint. Poorly managed samples will equal poor results, with the potential to derail years of costly clinical research. As with many things in the medical field, there has been a tremendous shift toward consolidation among medical laboratories with the end result that niche, specialty clinical laboratories no longer exist, and established assays that were less than mainstream are no longer performed. Further, in the rush to cut costs, the specialists required to run and interpret these results are no longer available.
This is generally related to the hospital centric cost/benefit analyses performed by the hospital laboratory management. This process does not really take into account the premium payments paid out of clinical budgets for trial-related samples, the result being a lack of focus toward external, non-trial samples coming from routine patient care. A large clinical lab may also lack the necessary project managers to help guide samples through the labyrinth. Any complication can result in discarded or lost samples.
Site to Site Variability
Other issues may also be that the study coordinators assigned to your protocol are not trained, or have little time and interest in preparing and shipping clinical samples to a central lab. Also, the sample stabilities may not be appropriate to the requirements of the study (or improper use of sample shippers). The study protocol may call for sample preparation that is overly ambitious, or there can be several “central” lab shipments, with the potential for a mix-up.
One question that seems to come up at every trial is whether to go with the standard of care lab test run at the local lab, or to ship the sample out to a central lab. There are advantages inherent in either scenario; local samples may cost only a few dollars to run, have no shipping costs, no lab kit costs, and do not require onsite sample preparation by a skilled (expensive!) study coordinator. However, there are always going to be site to site variability issues when you have multiple local labs analyzing samples. If the sample is really critical, then you may opt for both, with the site being able to use the local lab for patient management, and the central lab sample being used as the major study benchmark.
From the sponsor side, it is all too easy to think of a lab result as something you can obtain by selecting a lab and a test. But it may not be that easy! It is a critical part of an analyte-oriented study for the clinician to learn everything possible about a given assay, and to run a confirmatory assay when possible. This can be similar to running both HbA1c and blood glucose for diabetes, or both creatinine and cystatin C for acute kidney injury markers.
Developing a Comprehensive and Accurate Study Manual
Above all things, continuous training and monitoring of the sample prep team (usually a study coordinator on site) is of the utmost importance. Similarly, developing a comprehensive and accurate study manual that is graphic and easy to follow, along with a supply of proper lab kits is critical. As a clinical trial manager, you may serve your company well by actually going to the central lab (or local lab) and spending time with the lab manager reviewing best and worst practices related to the specific assay.
A few days spent onsite at the lab may mean success or failure of a years’ long study. It will also help develop or review the lab manual to ensure clear, simple, and correct content. The lab is an absolutely critical part of a biomarker/assay dependent study, and it has to be something an overworked coordinator can understand.
Other factors to be aware of include:
- Use of high performance shipping containers with longer dry ice retention times will eliminate some sample loss
- Ensure that samples are clearly and accurately labeled
- When possible, do not ship ambient samples within the U.S. – if ambient-only is available, consider paying the lab to validate frozen shipment methods. Also consider validating extended frozen storage shippers. Get the samples into the -80F freezer for at least an hour prior to dry ice shipment, and of course, make sure that you use sites only that have dry ice on hand, if it is necessary
The Devil is in the Details
When possible, work with a central lab that is known to be clinical trials friendly; this will mean they have project management and the ability to identify and fix issues before samples are written off. Clinical trial assays are lucrative and attractive to most central labs, but this does not mean they have the staff to anticipate a large clinical trial’s needs. Also, try and choose a central lab that has the maximum amount of required tests at their central campus: The more you ship, the greater the costs and risks to the integrity of the samples.
In the current situation of lab consolidation, there can be a lack of interest or infrastructure to deal with specialized clinical study samples that can be valued at multimillion dollar levels (e.g. lab not really set up to take on clinical studies) try and select on that has well qualified project managers, the most on-campus analyses, and willingness of the lab directors to do custom modification of tests to reflect your study’s needs. As an example, a central lab may only have frozen stability for a certain time period for your analyte of interest. But this stability may not be long enough to accommodate real world sample shipments. Ordinarily, this kind of extension validation should be possible (it may cost some money, though) in a lab that is focused on promoting and retaining clinical trial sponsors.
Identify how often and how soon you will need data back from the lab. Many labs batch sample runs for cost reasons, but this may also affect quality (one off samples may not have linear responses). Identify how often sample data needs to be returned to you because it is better to understand positive and negative trends in the study operation. Sample results coming in too late will not allow you to identify these issues in a timely manner.
Finally, work with a lab that has a good dashboard to help you identify where your samples are at all times. Remember: with labs, the devil is always in the details.
Steve Sisk, MSc Pharmacy