Healthcare needs Smart Data not Big Data

Big Data is an infuriating buzz phrase in healthcare. In reality we need to focus on giving clinicians the right data, at the right time and in the right place. Here's how.

What’s the purpose of the quantified self movement? What’s the purpose of digital technology in healthcare? To me healthcare has always been about humans. About the covenant of trust and respect that exists between clinicians and their patients. Technology can empower all sides to be more transparent, responsible and accountable more than ever before but we need to give it context first.

Big Data is one of these hyped sectors that lacks context right now. Here are my thoughts about how we can turn it into something that's tangibly useful in clinical practice.

1. Emphasise information

The wanton collection of data is, by itself, one of the most dangerous acts of 21st century technology. It’s awe inspiring and overwhelming at the same time. We collect because we can. Do we know why we’re collecting it? Do we know what it even means? Who’s going to use it? Does it actually change anything?

When I was practicing medicine I remember one of my professors stating the reasonably obvious, ‘It’s not about testing because we can. It’s about doing the right test when it’s needed and when it will change the plan.’

That statement is as true today as it ever has been. Healthcare professionals deal with mountains of data on a daily basis and they don’t need more volume. They need to collect data that can have impact, that can inform, insight and affect decision making and management. And by the way, that’s the same for patients as well as clinicians. So if you're an IT or 'digital health' vendor make sure you know what it will take to maximise the benefit of all the data you are collecting.

2. Listen instead of talking

Vendors are always questioning how they can extract commercial value from the significant amounts of data that they are accumulating whether it's from remote monitoring devices or electronic health record (EHR) systems. Asking those questions internally will only get you so far (though having a CMIO/CCIO helps) so instead you need to be engaging clinicians and hospital executives about what's best to do with all this data.

Beyond that are the issues of integration and use. Key questions I always bring up include:

• Do you understand what data clinicians use on a daily basis and in what priority?
• Do you know how they access data?
• Are you assuming mobility has been built into existing clinical IT systems?
• Do you know how clinicians interact with their data (i.e. workflow) and does your solution map it?

When it comes to patients remote monitoring devices are are great in theory but not always in reality. How many USBs are lost in washing machines every year? What happens to skin patches when people have a shower? How can you collect the right data in the most user friendly way possible? Ultimately, listening helps.

3. Be built for humans

Healthcare is about humans. It’s about trust and empathy as much as it is about MRI scans, blood tests and surgery. You might be thinking, ‘What’s this got to do with big data?’.

It’s got everything to do with it. As technology gets more sophisticated in its ability to collect data and detect patterns then teams of healthcare professionals who use it will have greater responsibilities than ever before to interpret and manage the delivery of all this data as information and knowledge for their patients. Patients’ will be empowered to ask more questions and search for answers but when they’re receiving this raw data it might be so new that no-one has the right answers.

Be considerate. As a vendor think about how your technology is going to change and, hopefully, enhance the clinician-patient relationship as well as create new anxieties. How can you make this process easier practically?

4. Be context sensitive

A big data solution isn't useful if the sum total of its capabilities is the ability to mine, store and organise data. It needs to offer intelligence that makes the process of accessing and analysing that data more intuitive. Different clinicians will look for different patterns and correlations and in different contexts. Whether it's the analysis of data stacks for clinical research or the analysis of blood result patterns in the acute setting a big data solution needs to map the needs of clinicians throughout the broad range of their clinical workflows.

5. Think less is more

There are a lot of tools and features that can be offered with a big data analytics solution. These may make perfect sense to an academic statistician or data scientist but not so much for a clinician or patient. There are some specific ways that clinicians interpret information and a solution that excels in delivering data and information in a way that fits the clinical user experience will excel far beyond those that overload the user with too many bells and whistles.

For a long time, to me, big data just meant the collection and organisation of volumes of data in an automated manner. That might sound basic but it's the perception of many clinicians so if you're a vendor start engaging your end-users and finding out how to turn big data into smart data through smart solutions.

Dr Saif Abed (@Saif_Abed)

Founding Partner

AbedGraham Healthcare Strategies Ltd.

www.abedgraham.com

(Adapted from original article published at www.hitconsultant.net)