Transformative Changes in Healthcare

Transformative Changes in Healthcare

By Justin Pestrue, Administrative Director of Quality Analytics, University of Michigan Health System

Justin Pestrue, Administrative Director of Quality Analytics, University of Michigan Health System

It will not come as a controversial statement to say information and its supporting technology will be transformative in healthcare over the next decade. Everyone is eager for the practical applications of machine learning and artificial intelligence to solve intractable healthcare problems like cost, access and quality. But I am more excited about a healthcare future that will ask analytics and technology to deliver care at the scale of n of 1. N of 1 is typically used to describe a trial where 1 patient is the entire sample, essentially a case study. In this instance I mean to describe a healthcare delivery system where the interventions and care paths are designed specifically for an individual patient taking into account the variation that makes each patient unique from the next. This is the scale at which we will be making clinical, financial and operations decisions for every patient, every day.

"Areas like genomics and social determinants research have clearly demonstrated that patients with the identical conditions are anything but identical"

I have been in the technology space for over 20 years, half of that in healthcare. I have watched healthcare systems be pulled in by the forces of standardization. The need to reduce costs and increase efficiency has naturally pushed systems to embrace economies of scale, streamline processes, implement evidence-based pathways and embrace reductions variation. The prevailing question is how to deliver cost effective, high quality, consistent care every time. At its heart this sounds right, but this model belies one important fact – that a system’s consistent care might not be the best or even right care for every patient at any time. 

In subtle ways the pendulum is swinging away from standardization. Areas like genomics and social determinants research have clearly demonstrated that patients with the identical conditions are anything but identical. We now realize that the obstacles patients face differ across genetic, clinical, social, economic, geographic, historic and many other dimensions of life. Understanding this difference means optimal treatment and care delivery should be altered to accommodate patient individuality. Underpinning this move are complex models, large scale datasets and software like EMRs that will deliver recommendations at n of 1 scale.

Take the case of a patient being readmitted to the hospital as an example. In my early years in healthcare hospitals were summarizing30-day readmission rates and saying we could do better. More recently we have begun risk stratifying patients to deliver additional resources to those patients at risk. Future evolution will have us predict, segment, a/b test and stratify care and interventions to smaller and smaller populations being more effective at each level of specificity.  Ultimately we will hit a limit where individual patients will have their interventions customized specifically for them. 

One of the sure signs of this evolution will be the great disappearing risk score. Current expressions of patient individuality take the form of a risks core or group categorization. For easy consumption these scores are heuristically group into categories like high, medium, low. The distinction between a risk score of 87 and 89 is effortlessly lost even if the underlying difference are that patient A lives alone in an urban center with little access to healthcare and patient B has 3 chronic conditions but family support and resources to spare. In the end they are both just high risk. As we get closer to an n of 1 care delivery system these scores will disappear. In their place care givers will be presented options or guided decisions that were crafted by analytic models that have the complexity to understand and express simply key distinctions.

Another sure sign of this evolution towards an n of 1 will be the reorientation of hospitals from a place of consistency to a place of embraced variety. Change management methods and teams will value complexity and develop language around additions as well as reductions. There will be growing talk of patient reported outcomes, social determinants and mobile device data in daily care supplementing the understanding of the lives of patients and their families. 

There are a number of challenges to making this world a reality. One of the clearest challenges is the quality of the data in transactional systems. If we can’t clearly understand the dimensions that make a patient unique we will struggle to implement solutions that are fueled by them. NLP, missing data algorithms and more advanced models of signal detection should begin to fill in these gaps but the distance is still a long one. Other technical challenges exist like rules engines that can execute at the scale of n at the speed of care delivery or effectively utilizing data engineers and scientists to crack the nut of these difficult problems. The breadth and scope of the technology to build this reality continues to be a challenge for executives to understand and rally behind.

And yet maybe the greatest challenge to making n of 1 care delivery a reality is the healthcare industry itself. Transitioning from a model of what is most efficient to deliver to a model of what is most effective for a patient to receive, means that healthcare will have to change its identity. It will need to make care outcomes, not care delivery, its guide. It will need to embrace technology and analytics that will guide this transition. Providers will need to trust systems and orient towards this future while implementation specialists across the industry need to make systems easy to trust, differentiating between signal and noise.

The technology to make this healthcare delivery world a reality largely exists. The speed and capacity offered by cloud solutions remove the technical barriers for most. New payment models that incentivize outcomes over delivery are growing. Then of 1 care delivery will be born when a system can exploit those technological advancements and marry it with an organization ready for change. Patients will embrace this new world and we may yet deliver on the promise of technology in healthcare.

Read Also

How Technology is Disrupting the Pharma and Life Sciences Industry

How Technology is Disrupting the Pharma and Life Sciences Industry

Colin Boatwright, CIO, Ironshore Pharmaceuticals Inc
Communication-Driven Workflow for Better Patient Care Delivery

Communication-Driven Workflow for Better Patient Care Delivery

Katie Myers, Director-Clinical Workflow Management, RiverMend Health

Weekly Brief

Top 10 Pharma Machinery Solution Companies - 2019