In the first and second installments of this series, we covered the opportunities that lie in health data as well as the challenges that preclude us from reaping greater value from data. In this final blog, we will propose our early ideas on how cross-industry leaders can use their data better to create shared opportunities for their organizations and patients.
Health care executives are bought into the premise that data is critical to achieving their goals. They are also eager to solve the industry's legacy data challenges that have barred progress. The federal government is taking a leading role in this effort: The Office of the National Coordinator of Health Information (ONC), for example, is spearheading efforts to improve data interoperability and prevent information blocking.
Executives must play a leading role as well. Through our research, we have identified three hypotheses for how industry leaders can draw greater value from the health care data.
Let's first contextualize those hypotheses in the constraints the industry faces:
1. Leaders don't want to place additional burden on the clinician.
Many data access and quality challenges could be solved by simply requiring clinicians to follow more stringent rules on what data they must capture and in what format. Industry leaders often avoid doing this, and rightly so—it's no secret that provider burnout is among the industry's greatest challenges.
Nonetheless, standardizing and expanding data collection without adding new burdens to those collecting the data necessarily will add complexity to potential solutions.
2. Leaders opt for limited (but attainable) workarounds rather than transformative (but difficult) solutions to root causes.
In part because leaders avoid further burdening clinicians, they tend to tackle health care data challenges by building workaround to challenges rather than solving their root causes. For example, natural language processing (NLP) is commonly viewed as a way to standardize clinical notes so they can be analyzed at scale. Yet NLP, while an important and worthy tool, does nothing to solve the root cause problem of unstandardized clinical terminology.
The reality is that the industry's model for data collection is so deeply engrained in legacy billing model that there is little incentive to tackle root causes heads on because they would require uprooting the billing process altogether.
3 ways to get more out of health care data
Health care needs new approaches for how to draw more value out of its data. Our hypotheses for how to do so are far from fully baked—rather, they are meant to be starting points for future conversation and progress. To that end, we have included a few of our own open questions on each. Here they are:
1. Focus on data terminology, not just data interoperability.
Data interoperability will indeed unlock greater value from health care data. But as leaders pursue interoperability, they must also work further upstream by tackling one root cause of poor interoperability: unstandardized clinical terminology. Creating standardized terminology can help ensure that clinical data is as usable to researchers and administrators as it is to frontline clinicians.
- How can leaders standardize terminology without disrupting the workflows of the clinicians capturing the data?
- Insofar as standardizing terminology requires placing more burden on clinicians, how can leaders ensure that clinicians share in the benefits of their increased labor?
2. Be judicious and targeted with the data you share with clinicians.
Clinicians will likely need to play some role in standardizing clinical data, meaning their buy-in is crucial. Yet they are fatigued by vendors pitching them data-driven tools promised to reap outsized rewards on clinical and business success—and by how often they fail. Administrators, executives, and vendors alike must be more judicious when choosing what data they share with frontline clinicians and in what format. They also must target their data sharing toward metrics that are most useful and actionable for clinical decision-making.
- What does "actionable" data mean in a clinical context?
- How can data insights be delivered to clinicians in a way that clinicians are receptive to them?
3. Design business models that incentivize data sharing across organizations.
Traditional business and payment models have tended to disincentivize data sharing—under fee-for-service payment, for example, providers are averse to aggregating claims data with other providers out of fear of exposing their negotiated reimbursement rates.
While these competitive constraints are unlikely to dissipate, leaders can still drive greater data sharing by creating partnerships that incentivize member organizations to pool their data for shared benefit.
Truveta, for example, has amassed a clinical database covering 16% of all U.S. clinical care by promising to generate more representative insights on patient care needs and treatment opportunities that will benefit its member organizations and the population as a whole.
- At what point are carrots not enough to incentivize data sharing—and the industry will need to turn to sticks?
- Will data aggregation efforts be more productive led by the private or public sectors?
Health care needs to partner to capitalize on its data
Health care leaders are right that data is a competitive asset but are wrong to shield their data in a competitive silo. To draw the full value that lives in the industry's data, organizations must be willing to pool their data with that of others. In the process, they can reap rewards for their businesses while simultaneously improving care for their patients.