How will data-driven diagnostics lead to value-based care?

Maneesh Juneja

Digital health futurist

How will data-driven diagnostics lead to value-based care?

25 November 2019 | 7min

Quick Takes

  • Diagnostics have the potential to play a bigger role in value-based care, but only if data is used to drive decision making

  • New sources of data are presenting an array of opportunities and challenges

  • People, policy and process will need to move in tandem in order to foster a data-driven culture

For many in healthcare, the notion of value-based care still seems like a distant dream, rather than something attainable here and now. With the cost of healthcare rising around the globe, the value attributable to those costs is becoming of greater interest to decision makers. 

Every aspect of healthcare is being looked at with greater scrutiny. In the USA, 72% of physicians said lab data informs most clinical decisions they make for their patients.1 However, are all of these tests always providing value? 

Naturally, we all want a world where we allocate resources based upon patient experience and outcomes. Transitioning to this future requires building a data-driven culture in our organizations. Some are more advanced in this journey than others. 

So, how can we accelerate this transition with the use of data-driven diagnostics?

Leveraging real world data

Leveraging real world data for value-based diagnostics

Data collected outside of clinical trials, such as electronic medical records (EMR), health insurance claims data or even a patient’s social media posts fall within the realm of real world data (RWD), and are being used to generate real world evidence (RWE). Traditionally used within drug development and drug safety, the use of RWD has recently helped pharmaceutical companies gain approvals for new drugs.2

In the United States, the 21st Century Cures Act required the FDA to create a program for evaluating the use of RWE data to support regulatory decisions about the effectiveness of drugs and biological products. This recent policy change highlights the ongoing efforts by both regulation bodies and the healthcare industry to support the proper generation and use of RWE in approval processes. 

There is a role for RWD in helping diagnostics become data-driven. In 2018, research was published which applied machine learning to EMR data from Stanford Hospital, USA to systematically predict laboratory pre-test probabilities of being normal under different conditions.3 This research was proposing a data-driven method to identify cases where the incremental value of testing is worth reconsidering, but leaving the clinician to make the final decision about ordering the test.

In 2017, there was another interesting piece of research published that used EMR data from the University of California, San Francisco Medical Center, USA to create ethnicity-specific reference intervals for lab tests.4 Given that normal test values can differ between ethnicities, this kind of research is an important step towards ensuring that the clinical value of lab tests is optimal for everyone in society.

Personalized healthcare is something many groups of people are working on; a future where we really can personalize every aspect of medicine to be as precise as possible. As this research continues and one day gets implemented into clinical practice, imagine that we will look back and wonder how we put up with one size fits all medicine for so long. 

Moving from system-centered care to person-centered care

Moving from system-centered care to person-centered care

In addition to personalized healthcare, is the ongoing cultural shift where instead of building everything around the needs of the healthcare system we build everything around the needs of the person who is seeking care.

This year, LabCorp, announced the option for consumers to order their own blood tests online and receive the results online too, a week after the blood sample has been collected.5 At the moment, 25 different packages are available, covering screening for diseases such as colorectal cancer and diabetes. In recent years, direct to consumer (DTC) genetic tests have become more popular, promising insights into risks of specific diseases. 

Given how consumers have convenience in so many other sectors such as retail and banking, why shouldn’t diagnostics be easy to access? Debate is ongoing about the value of some of these tests, and the implications of incorrect interpretation by consumers or even inaccuracies in some of these genetic tests. 

Countries such as France and Germany have actually prohibited these tests. There are also privacy concerns surrounding test results, and who might use your test results in the future. Testing companies have responded, such as Nebula Genomics, by offering anonymous genetic testing.6 

Some are also proposing that genetic testing be conducted by the healthcare system at birth to help screen for rare diseases. In fact, the UK has proposed just that, with a pilot project starting in 2020.7 Whilst this seems like a valuable initiative, there are quite a few ethical challenges, particularly since the results of this screening could be added to a child’s medical records. 

The value in value-based care with diagnostics is finding the optimum balance between convenience and benefit for consumers within the context of the constraints of the healthcare ecosystem.

As technology enables patients to become empowered and to take control of their health, this also has implications for laboratory-based physicians, in terms of the role they will play and the training they will need to help them adapt to their new role.8 The transformation is not just technological, but cultural too. 

Predicting the future

Predicting the future of data-driven value-based diagnostics

Predictive analytics is all around us. It’s an online store like Amazon, for example, that predicts what we are most likely to order and serves up a list of recommended products. What if that is the future of data-driven diagnostics? 

How might data help patients and their physicians to order a diagnostic test before the patient even shows symptoms? Could it be more RWD such as data from their Fitbit (i.e. step count that has gone down month after month), or from the sensor in their bed (i.e. sleep quality or duration that has declined over time) that flags up that someone needs testing? 

“The best path forwards is the one where we avoid creating digital silos across the healthcare ecosystem.”

Facial expression recognition technology is being developed that can detect if someone is nervous or confused.9 Other companies are working on the use of artificial intelligence (AI) in a car’s cabin to detect cognitive and emotional states of both the driver and the passengers.10 Researchers are working on algorithms that one day might be able to detect signs of depression just from analyzing someone’s voice.11  The emergence of these digital biomarkers provides even better opportunities for data-driven diagnostics. 

How soon before sensors embedded in the devices in our everyday lives are continuously monitoring different aspects of our health and making decisions autonomously such as booking doctor appointments and ordering diagnostic tests? Would consumers really be willing to hand control over to machines? 

A recent global survey suggests that some might actually want such a service. In the Ericsson 2019 Consumer Trends survey (which asked consumers around the globe), 57% want a smartphone that knows when they are becoming ill before they notice themselves, and 43% would like a virtual assistant that decides when they should visit a doctor.12

The question wasn’t asked in the survey, but I wonder if a similar proportion of consumers would be willing to let a machine automatically schedule a biopsy or a blood test? Who will know best about deciding when to order a diagnostic test, and which one to order: Physician, patient or algorithm? The value here is the ability to detect disease much earlier than possible today.

That value can only be fully realized if data are linked together. One of the biggest challenges in this future we are racing towards is that the data will be owned and controlled by different entities and it requires changes in policy, process and partnerships in order to create a 360° view of a consumer’s life. 

This may require partnering with entities you may not have dealt with before and new ways of solving problems. The best path forwards is the one where we avoid creating digital silos across the healthcare ecosystem. Underpinning this future is human-centered design, as without it, you risk adding layers of digital veneer to your products and services. 

With pricing in diagnostics under pressure, those organizations prepared to look at emerging sources of data are set to pioneer new models of care, and in my opinion, are more likely to survive and prosper during the 21st century. Just providing diagnostics alone will not be enough to stay relevant.

Maneesh Juneja is a digital health futurist who explores the convergence of emerging technologies to see how they can make the world a healthier and happier place. He looks at these technologies in the context of socio-cultural, political and economic trends, helping organizations around the world to think differently about the future.


  1. Stephens. (2019). Article available from [Accessed November 2019]
  2. Helfand. (2019). Article available from [Accessed November 2019]
  3. Roy et al. (2019). AMIA Jt Summits Transl Sci Proc 2018, 217–226
  4. Rappoport et al. (2018) J Appl Lab Med 3, 366–377 
  5. Rappleye. (2019). Article available from [Accessed November 2019]
  6. Nebula Genomics. Company website available from [Accessed November 2019]
  7. Reynolds. (2019). Article available from [Accessed November 2019]
  8. Orth et al. (2019) J Clin Path 72, 191-197
  9. Fujitsu Laboratories. (2019) Press release available from [Accessed November 2019]
  10. Affectiva. Company website available from [Accessed November 2019]
  11. Canary Speech. Company website available from [Accessed November 2019]
  12. Ericsson. Article available from [Accessed November 2019]