How much does a data query cost?
We all hear about the negative impact data queries have in clinical trials. Their abundance, which has been described and written about widely, results in a lot of hand wringing and declarations of hope that we can do away with most of them.
The questions I am most interested in understanding are:
- What causes data queries?
- Are they the result of negligence and lack of care on the part of investigator sites?
- Is what we are asking sites to do so hard that, inevitably, there will be queries for most of their submissions?
- Who stands to benefit from data queries?
In imaging trials a significant number of queries are due to the fact that the specific imaging charter was not properly followed when the subject was scanned. This imaging charter, which is discussed at length during investigator meetings (IM) at the start of the clinical trial, and in which trial coordinators must be trained, is reflected in specific, long and arcane manuals which every participant gets in the IM. With the exception of some cardiovascular projects, the specifications are then forwarded to the radiology department or imaging center technologists that will be carrying out patient scans in support of the trial. These specifications describe how to configure the scanners to ensure that all images being submitted to the core lab are taken using uniform methods. In an ideal world, imaging parameters could be somehow transmitted from investigator meeting to individual scanners and when a subject arrives for scanning, a single press of the proverbial “scan” button would generate perfect data. Alas, this is not likely to happen any time soon.
The first line of defense against possible data queries comes from the trial coordinator.
As we have said before, coordinators are often the unsung heroes of successful clinical trials. Submitting data is but one of a myriad of tasks they have to perform, most of which have to do with patient care. If there is one person involved in trial logistics who suffers from data queries, it is the trial coordinator. As far as they’re concerned, they want to complete their task and move on. One and done!
Having to deal with data queries after the fact is difficult, inefficient and undesirable. Until now, the clinical trial industry has seen core laboratories as the main line of defense in data compliance. This after-the-fact method of quality control is generally a failing proposition. It has been enshrined in our field for two reasons: first, there had not been a way to automatically check for data quality at the source, i.e., the site, prior to submissions; second, for many (though not all!) core labs, data queries (a large component of the “data management” budget) are a source of revenue they don’t want to part with. One need only look at a trial budget to see that the line items describing Quality Control activities are significant, and are typically only estimates. Most trials charge per-query, and they more often than not exceed the estimated budget. It doesn’t matter much that query management is a low-margin business. In fact, the industry has immortalized the practice with the slogan: “Any data is better than no data.” When you are in the business of selling man/hours, every little bit counts.
Until recently, automating QC of image submissions to occur at the site, was just a panacea. That has changed.
Not only can image submissions be checked at the source, but the trial coordinator need not become a trained radiologist to complete the task. The process is automated, is part of the data submission process, and provides the site with oportunities to correct problems, inform core labs and sponsors of why the data they are sending is non compliant (and electronically sign this explanation to, in effect, in a Part 11- tractable manner, open and close a query at the same time the submission is being made). This capability is as powerful as it is disruptive. It takes the industry closer to reducing queries and simultaneously avoiding 100% source data verification, both of which are stated goals of sponsors and CROs everywhere.
The answers to my questions are clear.
Data queries are the result of mistakes that can occur either because of lack of training or oversight, and not by negligence or lack of care. What we are asking our trial coordinators to do could not be achieved if they didn’t have easy to use tools that perform technical tasks that are beyond their functional realm. This is very common in modern times where tools like Excel enable people with little training to perform complex math.
Finally, as with all technological evolution in business, there is an inherent conflict of interest between the best available solution and the incumbent provider. It has happened in going from coal to oil, oil to natural gas and natural gas to solar or wind power. Resisting evolution in order to preserve revenues fueled by inefficiency and preventable errors is short-sighted and ultimately futile.