Does the growing complexity of medical data lead to better outcomes?

The number of available pathology tests only grows. This is a good thing – we can delve into the genetics of cancers to better match treatments, predict diseases like Alzheimer’s disease more cheaply and quickly, and identify risk factors for our biggest public health problems like cardiovascular disease. But the introduction of new tests is rarely coupled with removal of old ones, and the way pathology laboratories are geared to report tests is additive.

Every new test we offer increases the complexity of the output to the requesting clinician, providing more data to integrate with their own to come to a provisional diagnosis or assess a trend. Artificial intelligence (AI) could go some way to enabling more streamlined delivery of care, but deficiencies in the clinical evidence base underpinning decision-making remain.

In pondering this topic, I was keen to draw a parallel to a frequent frustration – that despite the improvement in computing power by orders of magnitude over the decades, Microsoft Word still takes just as long to load on a contemporary desktop computer as it did three decades ago. So, I self-consciously asked ChatGPT, as I frequently do on a range of subjects: “Am I the only one who feels this way?”

Apparently not, and the phenomenon has a name – Wirth’s law. Put simply, this law states that software is getting slower more rapidly than hardware becomes faster. Yes, we can do many more things on our computers than we used to, but the user experience continues to frustrate and disappoint.

It seems our ability to generate diagnostic data is increasing more quickly than our ability to make that information clinically intelligible. Systems for deriving enhanced meaning and insights are mostly clunky, and have limited interoperability with other systems.

This is not an argument against innovation, nor a suggestion that every introduction of a new test needs to see the retirement of an old one. Just because we have a new blood test for Alzheimer’s disease doesn’t mean we don’t need to measure serum sodium.