Data-driven healthcare. What’s next and how do we get there?

Claus Puhlmann, PhD

Cofounder and Chief Scientific Officer of The Medical Algorithms Co.

Data-driven healthcare. What’s next and how do we get there?

8 June 2020 | 8min

Quick Takes

  • Data must be integrated and standardized in order for value-based healthcare to thrive

  • Standardized data can be used to identify gaps in care and make improvements

  • Decisions that drive future care should be made based on actionable evidence

Global healthcare is in crisis. Nearly every healthcare delivery model is financially unsustainable, eating up a greater and greater percentage of GDP each year. And it’s especially true in the United States, where healthcare spending is projected to account for 20% of the GDP by 2026. Global trends are shifting towards value-based, and to make this shift, health systems must adopt new care models and exhibit new team-based behaviors. 

Healthcare, however, is infamously change-resistant. That’s where data comes into the picture. Not only can data help identify the most wasteful care, it can help care teams adhere to clinical best practices.

“Physicians chose to become physicians for the right reasons: to help people,” says Dr. Scott Weingarten, Chief Executive Officer of Stanson Health. “If you approach them with an opportunity to improve their practice that is founded on evidence, they’ll support the initiative.”

What does data-driven healthcare look like? 

The future of data-driven and evidence-based care is multifaceted, with a variety of dynamics to consider:

  • Data will be used to identify the costliest, highest-risk patients 
  • Data will shed light on gaps in care
  • Actionable data will notify frontline care team members in real time 
  • Patients will access portals to learn about their conditions, risk factors, and action items to maintain their health 

Together, these factors will help populate an integrated care plan, so that patients and providers can work together. Payers, too, will be part of this mix. 

“Payers increasingly want to partner with providers for population health,” says Dr. Philip Chen, the Chief Strategy Officer of Sonic Healthcare USA. “They want to see data proving quality outcomes. That’s why nearly every contract now includes value-based metrics.”

Data needs to be integrated

Much of the technical infrastructure needed for data-driven care transformation, like electronic medical records (EMRs), is in place. Care is being digitally documented by care teams – clinical data repositories and data warehouses have collected massive stores of data – yet despite this aggregation, a significant amount of data remains fragmented.

“Data integration remains a challenge,” acknowledges Dr. Bruce Muma, the Chief Medical Officer, President, and CEO of the Henry Ford Physician Network. “It is still siloed. This makes assembling a complete patient record difficult. It is hard to identify gaps in care when you have gaps in data.”  

Patients often have lab testing in different places across an organization, and sometimes outside the organization. And when you consider that the majority of a patient’s record is comprised of lab data, it’s easy to see the challenges that can arise.

“Fragmented lab data is a real problem,” says Dr. Chen. “How do we share data with payers if it is not in the record? How do we get paid to perform a lab test if there is no record of it in the patient’s chart?”

Data needs to be standardized 

Despite massive amounts of aggregated data, and the fact that a portion of it remains fragmented, another subset of data remains non-discrete, or unstructured. This makes it difficult to digitize, capture, and act on. A lot of the information that could be used to guide better patient care, better quality, and lower-cost care is trapped as free text.

“We need real-time data coming from the EMR that is discrete, so that we can identify trends and triggers that are likely to happen,” Dr. Muma says. “Unfortunately, a lot of the data that we need is non-discrete at this point in time.”

“Data standardization is critical,” adds Dr. Weingarten. “There are some studies that say about 35% of valuable data is discrete and about 65% is non-discrete or free text information. Interpreting free text information in the EHR is very important.”

Efforts to simplify the standardization of lab data have had mixed results. LOINC (Logical Observation Identifiers Names and Codes) allows labs to standardize test naming nomenclatures across vendor methodologies. However, as the number of tests has grown, so, too, have the number of LOINC codes. This has created an explosion of complexity for labs who are trying to accurately code their tests for billing. Making matters more challenging are select laboratory system manufacturers who don’t provide LOINC codes.  

“Ordering physicians use more than one lab and must submit their results to payers,” Dr. Chen explains. “If they don’t have a complete set of LOINC codes, it can impact their performance metrics, because test results could be missing.”

Clinical practice needs to be standardized on data

One of the biggest drivers of poor outcomes is care variation. Reducing it on an enterprise scale requires that clinical practice become standardized in focusing on evidence. 

“The challenge is that a significant percentage of clinical practice today does not reflect the evidence,” says Dr. Weingarten. “We’ll be moving clinical care practice to be more evidence-based.”

To do so, industry leaders recommend taking these three steps:

Step 1: Identify high-impact opportunities

Chronic disease management holds great potential for its ability to cost-effectively treat patients at home. In all cases, follow a systematic approach to identifying the best opportunities:

  • Review the evidence
  • Identify:
    • Conditions with high morbidity and mortality
    • Avoidable costs
    • Overuse of care that causes harm
    • Evidence that care can be improved
    • Tests that need to be performed
    • Gaps that can be closed

Step 2: Approving and supporting data-driven care initiatives 

Make sure that any recommended initiative aligns with your organization’s strategic objectives, and look for signs of a return on investment (ROI) as you consider signing off. However, keep in mind that you might not always be presented with a concrete ROI, because that would require something that is very hard to quantify: avoided costs. 

“How do you measure the heart attack that didn’t happen?” asks Dr. Muma. “How do you measure the hospitalization that did not happen? I think ROI is still our biggest challenge in terms of value-based care.”

As we transition into this next phase of data driven care transformation, not everything will be as financially cut-and-dry as you’re used to. For example, it’s easy to show an ROI on something like an MRI system that can be filled up with people to directly generate cash. But initiatives that are designed to get ahead of costly exacerbations, on the other hand, are often best justified with better patient outcomes than they are with dollars and cents. If the ROI isn’t immediately apparent, look to the clinical evidence and consider the avoided costs that will logically follow. 

Step 3: Deliver actionable information at the point of care

As value-based care becomes more prevalent around the world, so does the pressure placed on providers to drive it. This is causing clinician burnout. In the United States, the Institute for Healthcare Improvement (IHI) is elevating awareness of it by adding “Improved Clinical Experience” to its Triple Aim, now making it the Quadruple Aim. 

Clinical decision support (CDS) tools like pop-ups, alerts, and rules are designed to reduce cognitive overload. But for some, they are contributing to it. Physicians are responding by tuning those alerts out, proving that simply delivering more information is not the answer. Delivering information that engages people and inspires them to favorably act in real time is.  

“It’s really about cleaning house and removing or improving the low performing alerts and replacing them with high performing alerts that improve care,” Dr. Weingarten says.

“We started a cleanup process at Henry Ford in partnership with Stanson Analytics,” says Dr. Muma. “We identified the nurses and pharmacists receiving the highest volume of alerts and asked, ‘Do you really need this?’ In four months on this project, we reduced alert fatigue by 10%, and we are optimistic that we are on the right path.” 

Data-driven healthcare trends are underway ..

It’s easy to forget how much investment and effort has been devoted to digitally transforming healthcare. Not long ago, paper charts hung from the ends of patient beds and physicians placed orders by pen. Transcription errors and adverse drug events were routine. 

Today, EMR systems have replaced the paper chart and computerized physician order entry has replaced the pen. Providers who once believed that patients shouldn’t have access to their records are now welcoming their real-time digital access to it. 

“We have the ability, in real time, in the cloud, to query discrete data elements, labs, problem lists, demographics – all of the information that is codified – as well as free text information like doctors notes, and compare that with evidence-based guidelines, and provide recommendations,” Dr. Weingarten says. “This is a significant advance that has occurred recently.”

Dr. Muma compares data-driven care transformation to a 9-inning baseball game. Before the game can begin, there needs to be a playing field. 

We had to build the stadium first, and I think the stadium is mostly built,” he says. “It’s built enough for us to start playing.”

…but much more needs to be done

For all that has been accomplished, we are still in the early stages of what needs to be done.  

“We’re in the first inning,” Dr. Muma says, referring back to his baseball analogy. “Even though we’ve done a tremendous amount of work in building this infrastructure, we still have a long way to go to really leverage it in terms of data and analytics.” 

It’s clear the future of healthcare will revolve around data-driven care transformation – a  future filled with data scientists, collaborative care teams, new roles, and greater expectations. And despite the burden of these constant changes, it’s an exciting time for the industry. 

“We’ve already begun extracting actionable information,” says Dr. Chen. “The next step is putting it into action to directly impact care. It’s not a linear process. It’s a continuous feedback loop. I see it as a long road ahead, but a very exciting one.”

Claus Puhlmann, PhD is a former research scientist with Max-Planck-Institute, has extensive experience in cancer biology and immunology, and over 25 years experience in publishing scientific and medical information for professionals, content management, and patient-focused, internet-based medical information and health record systems. Claus is also author of more than 200 medical review articles for patients and an award winning photographer. His primary focus is on media structure, quality control, and online content.