Caveats for #DigitalHealth by Bharat Gera, @bgera

Everyone from Marc Andreessen to the angel next door is talking about the disruptive forces of digital technology knocking the doors of healthcare industry. Nothing is more appealing than a cure for cancer or at least a possible early diagnosis. Potential for solving the resource crunch by replacing doctors with AI appears to be on the horizon.

Now, let us switch to reality. It’s much more thornier than the rosy pictures put out by the leading technology leaders. Eric Topol writes in Deep Medicine that “The greatest opportunity offered by AI is not reducing errors or workloads, or even curing cancer: it is the opportunity to restore the precious and time-honored connection and trust—the human touch—between patients and doctors.”

Based on my experience in healthcare, I offer 3 Caveats to those attempting Digital Health – based on my engagement with the ecosystem as the CIO of a teaching hospital and advisor to early stage ventures in digital health. Too often, there will be a temptation to ignore these caveats as the vision of a brave new world is enticing to us. Beware, too many have fallen at these hurdles for the last couple of decades in healthcare.

Caveat 1 – Do not underestimate the hurdle of legacy systems

Healthcare IT has been a laggard in technology for a long time, most of the existing systems used are at least 10-15 years old in their technology architecture. Forget about devops and dockers, most do not have standard APIs to integrate and the User Interface is ancient text based forms in most applications used by hospitals.

Getting rid of these systems is not a trivial job, in spite of all their technology limitations – these systems run the operations of hospitals. Business operations of a hospital are a curious mix of retail and services, with several considerations that are not relevant in other industries.

Often, most digital health ventures fall at this first hurdle. Unable to integrate with the legacy system leads to frustration and a long sales cycle, something that an early stage venture cannot afford to bear.

Leading hospitals have realized this constraint, they are working with technology companies to set up platforms that offer easy integration with legacy systems. Still far from plug-and-play reality of consumer products, only a step ahead in open integration with digital world.

Caveat 2 – Do not increase the cost of healthcare with digital by claiming it replaces doctors.

Hospitals in the USA spend over a trillion dollars, nearly 1/3 of the total spend on healthcare that has risen to 18% of GDP. On the other hand, countries like India spend less than $200 per capita. Neither of them can afford an increase in healthcare costs due to digital technologies.

Asking hospitals to spend more on cancer treatment using AI is not going to work, even if the performance is nearly as good as a clinician. Saving clinician time or increasing productivity are promised but fail to deliver.

Being realistic that the benefit of digital health is going to take some time to kick-in, digital health companies will be more successful if they can replace existing technology with lower cost options or uncover new methods of treatment that bring down costs.

Before meeting the prospect for your game changing digital health technology, make sure that you have a business case for reducing costs for the customer. If you can relate the sepsis algorithm to reduced ALOS (average length of stay) in the ICU or help treat that knee without surgery – you may have a winner 🙂

Don’t argue that the hospital is going to run without doctors and nurses, it’s still a distant dream!!

Caveat 3 – Do not treat the patient as data for digital health, start with empathy.

Patients are not data, they are human beings needing care like you and me. Starting with empathy for a person with cancer is difficult unless we have experienced it personally or seen a close one go through it. Too often, we are tempted to look at them as data to develop an algorithm.

Digital technology in healthcare is creating petabytes of data, from novel imaging techniques to sensors – we can collect a huge volume and variety of data for each individual throughout their life. Fitbit and smart watches can track every step we take on a daily basis. Obviously, we are drawn towards the potential of using this stream of data to make clinical and wellness decisions.

Unfortunately, this deluge of data is not accompanied with advances in the science of medicine and wellness. Most of the times, the individual and caregiver is overwhelmed with the deluge of data. Beyond simple advice, like eat less carbohydrates and walk more often – there is not much more that can be done with the abundant data – nor does it vastly improve clinical decisions.

Some of us are hoping that the data being made available will help in discovering a better science by analyzing patterns using AI. Often, these hopes are unscientific – analyzing image patterns with AI will not replace the work of a pathologist. Connecting the baby to more monitors in NICU to collect more data will add more burden to the mother when she needs to provide that life saving hug.

Instead of building a case for discovering patterns and algorithms from patterns, focus on solving a problem using relevant data captured with the least intrusive device. Always put the human first in healthcare, the algorithm can not rule and treat the person as a lab rat for the supposed benefits using AI.

Hope these caveats help you with taking Digital Health ahead without the obvious pitfalls many have fallen into in previous attempts!!

All the best 🙂

This article was first published on the Author’s LinkedIn Pulse Blog, its been republished here with the author’s permission

Author