Diagnostic technology: be ready for the next five years

In the veterinary industry, our diagnostic needs are always changing.

We have changing patterns of ownership, changing farming practices and an ever-advancing understanding of non-human pathophysiology, meaning there is always something new to detect. And changing business models, practice guidelines and public attitudes meaning there’s always new constraints on that detection.

So, it’s clear that, alongside pandemics, new measurands and workflows might drive unpredictable change for veterinary diagnostics in the next few years. But some things will stay the same: diagnostics must always consist of data gathering, calculation and decision-making. And, since lives and livelihoods depend on it, we will always want to do those three things faster, more easily and with fewer errors.

Possible future scenarios

As we saw when looking at communications technology, skilled professionals (especially vets) place a lot of belief in the efficacy of our skills. When things go well, we take pride. When they don’t, we take responsibility.

For people like us, if someone tells us to make things faster, slicker and better, our response is to work harder and harder with the same old tech. Sometimes that will work. And sometimes it really won’t. Consider the scenarios in the panel below.

In those possible future situations, and thousands of others like them, the vet cannot just deduce the right answer by working harder or being smarter. The future looks like it needs new diagnostics technology.

Tech requirements

But which diagnostic technology is required? The exact spec is determined by the exact needs facing the practices.

In the first scenario, we’re being asked to detect something new. This might be newly discovered biomarkers, or very subtle variations in physical signs. It might also be a synchrony in markers that has previously been overlooked. In any case, the technological need is primarily for increased sensitivity and specificity. It’s important that we realise that, as with all sensing, smarter data-processing software can offer the same benefits as higher sensitivity or specificity by cleaning up our data so we can make better sense of it.

In the second, we need to detect something at a distance. This time, sensor communication is important (see the last article; June issue), but so is sufficient automation to allow any remote sensors to be placed or used by owners without supervision.

In scenario three, there’s also a distance involved, but it’s temporal not spatial. The tech needed here is rigorous storage of a larger pool of parameters: Keeping track of the animal’s weight at each consult is useful. RR and HR are useful. But the ability to rapidly and reliably capture, for example, a short video clip, could enable us to rapidly scan back and compare for changes in appearance, posture, gait, behaviour and myriad other clues. Combining with genetics information is a no brainer at this stage, too.

Finally, in the fourth scenario, we combine a little bit of all three: the technology required to be able to compete with that practice would need to be able to automatically and continually search for pathological trends in the data streams of every animal signed up to the service. Those data streams would need to include at least some element of real-time information from a worn or in-home sensor and would need comparison to stored information from consults past. It is also likely that trends across multiple synchronous markers would prove important.

These four scenarios are in no way exclusive – they could all easily occur in the same future. In fact, the sort of predictive analytics that would be required to survive scenario four are exactly what’s now being used to drive online advertising. They’re not outlandish scare stories.

Luckily, over the next few years, our practices will be offered hundreds of new technologies to help us deal with scenarios like this. Choosing between them will be a major challenge.

Selecting technologies

Clearly, we can’t know exactly which diagnostic challenges you’re going to face in five years’ time. So, here are some things to look for in new tech.

The trio of faster, easier and more accurate will always be important. Saving time and preventing errors when a patient’s life is in the balance is critical (and helps your work-life balance). Also, don’t underestimate the effect of choosing easier testing protocols; they’ll drive better uptake among colleagues and help machines pay for themselves faster.

The right measurands will also be critical in getting maximum clinical utility out of your investments. Keep an eye on the veterinary press for these, but only consider new hardware if you’re certain that you can’t get the same information out of your existing equipment by using it in a new way, potentially with new software. Be aware that location will change what some measurands mean. For example, high blood pressure in a consult room and high blood pressure while sleeping may indicate completely different pathologies. Remote measuring is very possible.

Talking of software, we expect that the four pillars of good diagnostic software in veterinary medicine will match those in human medicine.

Trend-tracking

Reporting on trends in readings over time, rather than absolute values, has a lot more relevance to prognoses and treatment plans. Clearly, this is done only within certain absolute ranges.

Automation – the triggering of (and correct completion of) recording and reporting without physician intervention enables assessment at time and in places that would otherwise be impractical.

Decision support – in the next five years, we’re not expecting veterinary diagnostics software to progress beyond very simple triage (for example, warnings to check on certain patients, or confirmation that patients are within their normal-for-them ranges), but if you’re not familiar with what Babylon Health is poised to do to human medicine, now might be the time to look it up?

Storage – if diagnostic tools are going to be automated and to learn from the data they gather, then it will pay to let them gather a lot. Physical storage media have never been cheaper (per bit), but the right cloud storage not only offloads the physical bulk, but also the legal liability in case of a breach. Ensure any diagnostic tool you consider will integrate with your chosen storage solution.

Which just leaves two strategic considerations:

Open standard – diagnostic technology makers have worked hard to lock us into their particular machines in the past. As the power moves away from the hardware to the software, we should expect some degree of “format war” (for example, HDDVD versus BluRay); an “open standard” (such as USB) might emerge as a safer bet.

Cost – as with all technology purchases, more efficient ways will exist to make a practice profitable in the short term.

Wrap up

We’ll keep this short because we’re already over word count.

While the core of what is needed to make a diagnosis will always remain the same, major, unpredictable change is both possible and likely in veterinary diagnostics over the next five years.

Expect the biggest advance in diagnostics over this time to come from better software. And get your business ready to make use of at-home diagnostics; continuous recording; trend tracking; triage; and faster, better answers.