We can count almost everything – but should everything count?

If you have an email address, Twitter account and Facebook page, or some combination thereof, you know that the amount of information that is readily accessible about almost any topic has expanded exponentially in the past decade. The same is true in science. What used to take considerable effort – combing through paper records and medical files – can now be done with the push of a button. It is not hard to predict that in the very near future, especially after electronic health records are fully phased in, the amount of potential information available for research will be remarkable in scope and detail.

This should be good news to anyone who wants to advance a complex system like the clinical trials process. After all, the more we know, the better job we can do pinpointing improvement. Knowledge is power, and data fuels knowledge, right?

One of the great advantages of applying systems thinking to any process, including clinical trials, is that we can examine how our expanding data sources will interact across a number of dimensions and subsystems. For example, while additional data seems to be a theoretical improvement, does that data serve the needs of decision-makers, such as the FDA (which ultimately approves drugs for marketing) and stakeholders (such as patients and their families)? Systems thinking, the art and science of making inferences and predictions based on a deeper understanding of structure and function, thus leads us to ask questions about how the pieces of the clinical trials process interact and respond when a bolus of new data comes their way.

In our work, we have learned that more data is not a universal balm. First, it is clear from our interviews with some decision-makers that very often too much data is available, and the data fog it creates can actually impede, rather than improve, evaluations of new drug products. Second, we have learned that it is imperative to recognize that stakeholders value their privacy, and rebel against the unfettered use of individual information. Not surprisingly, there is a tension between data disclosure and data protection, and laws and policies have been put in place to try to balance these important imperatives. However, it is very possible that growing the data base about adverse drug reactions could lead to insights that might prevent loss of life and disabilities. Again, our research indicates that this part of the clinical trials system is not as information rich as it needs to be.

Our ability to collect and collate information continues to grow. By studying clinical trials through a systems lens, we can most effectively calibrate how and when to deploy this data and also decide when exploitation of data should be limited.

Paul Locke