Defining Value: The Truth About Value-Based Care
Cleveland Clinic defines value based care on their landing page: “Value-based care is simply the idea of improving quality and outcomes for patients. Reaching this goal is based on a set of changes in the ways a patient receives care. We’re looking to make healthcare proactive instead of reactive, preventing problems before they start. Overall wellness, quality of care and preventive screenings all are key to bringing about better healthcare outcomes.”
This sounds great. But Cleveland Clinic left out what is core to value in this context: cost reduction. Per Sachin Jain, MD, in his recent Forbes article, Value = Quality/Cost. When you start dividing quality by cost, quality tends to go down. Where does true value come from and what does it mean?
Jain, current president and CEO of SCAN Group & Health Plan, takes a critical look at the healthcare industry’s zeal for and love of expressing alignment with value-based care. Jain shines a light on the gap between marketing promises versus real-life results.
Metrics Over Patients
Value-based care is driven by incentives. [LH1] This results in healthcare providers optimizing for a metric rather than optimizing for the best patient care.
Take, for example, Medicare’s Hospital Readmissions Reduction Program (HRRP), a mandate to reduce hospital readmission rates for Medicare recipients. If a patient is readmitted to the hospital within 30 days of discharge, it counts as a readmission event. Hospitals optimized for that metric and reduced the number of readmission events, so HRRP was considered a success. But together with that metric came an increase in mortality for people who were discharged. As with much of value-based care, there is a discrepancy between what’s measured and “better healthcare outcomes.”
Value-based care also incentivizes better documentation over better care. Healthcare administrators create many checklists intended to help healthcare providers streamline patient care. But not every patient fits into the checklist. The amount of time a physician spends with a patient is relatively short. The time they spend on documentation and ticking the boxes on checklists eats into their time with patients- and is often not condensed or harmonized with what is created by clinicians. Add to these processes the fact that much of the information in value-based care systems is derived from ICD codes. Often those codes fail to reflect what is real or relevant about a person’s condition. Ask any physician: patients do not fall neatly into ICD codes.
“There needs to be better ways to characterize, summarize, and represent patients,” says Jung Hoon Son, MD, pulseData’s Director of Informatics. “As an informaticist, I feel the limits of current ICD coding systems. They’re meant for billing and proof of documentation; they don’t really tell the story of a patient well. “ He explains how current healthcare data is structured as an itemized transaction list and often misses the details needed to understand what is happening with people.
“It’s like trying to make an apple pie, but you are just given a receipt of what was bought at the grocery store. You might be making apple pie mixed with a lot of other things. That’s the kind of data we get. It’s a list of ingredients without knowing how much of each to use to get to the right outcome.”
In a value-based system, distilling relevant information about an individual patient is difficult. In many cases, the modern electronic health records (EHR) system doesn’t help, as it also relies on generic documentation. For individual summaries, healthcare providers have to drill down to text-based physician’s notes to learn about the patient. This takes time and effort that is not rewarded or scaleable in a value-based care system.
The right data has to be extracted and summarized in a way that represents a patient fully and truthfully — and that’s missing,” Dr. Son says. “It’s hard to deliver value-based care properly, because we’re not likely to be attaining the understanding of patients needed to address core problems.”
Patients as Phenotypes
Digging deeper is essential to providing care that has true value in patient outcomes. It’s also possible. Rather than using metrics such as ICD codes to extrapolate information at a population level for one-size-fits-all guidelines, we need to extract information from ALL the available data — EHR, claims records, ADT feeds, pharmacy data, clinical laboratory results and complementary data sources — to diagnose, risk stratify, and predict adverse events for individual patients.
Such phenotyping can be accomplished with machine-learning/AI technology — and is necessary to improve value-based care systems. pulseData, for example, aggregates disparate data sets to produce a full, truthful - actionable - picture of the patient. Instead of basing risk scores and metrics solely on financial and population-based guidance, healthcare providers can make better, realistic, holistic decisions around specific patients and their likelihood of adverse outcomes.
This allows patients and providers to team up effectively. It launches a process of leveraging comprehensive datasets to uncover incisive, patient-level understanding. When this happens, clinicians are afforded the time they need to prioritize,act, and improve lives.
The espoused goal of value-based care is compelling - increasing quality while decreasing cost - but the methods often fail to align with results.
Here’s where we are:
The idea of value based care? Great. The execution? Less great. Why?
Turning people into a number fails to bring integrity to what matters to a patient. Unfortunately some of the most common tactics used to track value based care can undermine it. For example- in an effort to document patient care, checklists and metrics often de-center patients themselves.
Better curation of underlying data, (for example phenotyping) can be used to avoid potentially inherent biases and to generate more transparency in the data and algorithms. When people are not buried in a number, the impact is triumphant; trust and understanding emerge as do the people who are at risk and would have gone unseen or misdiagnosed, otherwise. This clarity is core to better care quality, efficiency and impact.
While the compliance needs and metrics of value based care have the potential to cause more harm than good to patients, it’s possible to make a difference now. Aggregating sufficient quantities of high-quality data is a major challenge in health care. That’s because such data often resides in different organizations and its quality varies.
One way to overcome this challenge is to fund the curation and preparation of data libraries-as much as is possible to net a clearer picture. This means better phenotyping of data, and the use of disease tools to predict and avoid adverse events. These work in tandem to make value based care more concrete than conceptual.
The good version of this is possible. With AI, curation and phenotyping- it’s possible to achieve economic outcomes of value based care, generate real alignment and improve the day-to-day lives of more patients and their communities.