How Much Health Data is Too Much?
Stephen Beller, PhD
Prosocial Entrepreneur, Clinical Psychologist, Psychotherapist, Cognitive Scientist, Software Architect and Model Builder, Knowledge System Inventor, Consultant, Writer, Futurist
How much healthcare data is needed to continually improve healthcare outcomes?
Conventional wisdom is to collect as little patient data as possible to support healthcare decisions. After all, data collection costs time and money.
Continually improving healthcare outcomes requires knowledge of patients' conditions, physiological (biomedical) and psychological strengths and weaknesses/vulnerabilities, available treatment options and previous treatments, preferences and motivations, adverse social determinants of health, etc. If this knowledge is adequate, care teams are able to gain a deep understanding of what the patient needs and wants, and the patients are able to participate in shared decision-making and self-manage their health with competence. Decisions based on this knowledge and understanding are likely to be better clinical outcomes than is otherwise possible.
Question is: How much data are needed to generate information that builds this kind of knowledge and understanding?
As the image above shows, the plethora of health data is huge and accelerating. Social Determinants of Health (SDH or SDoH) data are adding to it. As such, I attempt to answer this question using SDH data as a pertinent example.
The SDH Data Predicament
I've posted several articles about SDH on LinkedIn, each of which addressed aspects of the data requirements issue:
- Standardizing Social Determinants of Health -- bit.ly/2RsIacp
- Comparing SDH Assessment Tools and Data Sets: bit.ly/2SSgVoo
- Social Determinants of Health: A Silver Bullet: bit.ly/2McvHEb
The American Medical Association (AMA) has also been exploring the SDH data; its latest conversation is here.
The Problem
Many individuals and organizations are contributing to the definition of SDH and the offering suggestions about what should be considered as SDH data. There is indisputable evidence that non-medical factors often have a major impact on a person's physical wellness, as well as, of course, on one's psychological well-being. The problem is establishing a universal definition of those influential SDH factors, describing how to measure them, and using that information to improve healthcare outcomes.
One of my comments to the discussion on the AMA site addresses this problem. I indicated that we should start by compiling the most expansive set of SDH data possible. Despite the challenges involved, I recommended that this would be a first step in implementing a rational solution.
A Rational Solution
My comment made the case that we should learn as much as possible about the non-medical factors that influence (cause or exacerbate) biomedical/physiological symptoms and disorders. The more we can learn, I reasoned, the greater our knowledge and understanding about what should and could be done to improve overall healthcare outcomes.
I stated that we need an adequate quantity and variety useful data to support research, analytics, and AI that provides meaningful decision support. This decision support would enable health IT tools to present a comprehensive biopsychosocial view of the whole person and the social and physical environments in which people live.
What we don’t want, however, is to drown in an ocean of redundant or irrelevant data that increases information overload, results in unnecessary work and workflow changes, and causes avoidable stress.
A practical solution could be for the AMA (or other organizations) to establish a task force that does the following five things:
- Compiles an in-depth list of all available SDH assessment items (question-answer sets) from existing tools/questionnaires and from other relevant sources
- Organizes those data into meaningful categories
- Removes redundancies
- Identifies and fills in gaps
- Offers the resulting data set to clinicians, researchers, and informaticists who would work to reduce that expansive data set. Their work would focus on extracting the essential, valid, reliable, and useable core data set that would then be piloted and revised/evolved as necessary.
Requirements
For this solution to work, there needs to be adequate resources: people, money, and digital information tools (e.g., for knowledge management, data analytics, and interdisciplinary collaboration).
Another requirement is a social network whereby people adversely affected by negative SDH factors receive the help they need to cope with their situations.
If we make this a top priority, we have the potential to radically improve the health and well-being of populations around the world!