Understanding Pattern Formation in Clinical Care – who will
benefit, who will be harmed, and for whom does treatment represent
waste?
For essentially every test and treatment we have in health care, there
are basically three subpopulations of patients who undergo a test or
receive a treatment. First, there is a group that benefits from the test
or treatment, but there is also a group that does not benefit (this is
waste in our system), and finally, there is a group of people who are
harmed as a result of that test or treatment.
Until now, our simplistic thinking has allowed us to rationalize that
the waste and harm was just a necessary evil to help those patients who
benefit from a test or treatment. Who could argue that e.g. a few
unnecessary mammograms/PSA tests/screening tests for hypertension or
diabetes are justified to save a patient’s life? But systems science
principles argue, and the data from many decades of screening have
shown, that it is not so simple, and we are perpetrating a degree of
waste and harm in patient care that is not sustainable and not ethical
(the problem has now received Cochrane recognition with the formation of
the “sustainable healthcare” group -
https://sustainablehealthcare.cochrane.org/ ).
To understand how to define the subpopulations that are harmed and those
who do not benefit (waste), a more complete understanding of systems
science and data analysis is required. The ultimate goal of systems
science is to improve outcomes that measure the value of care for any
definable, whole patient process. This is achieved by discovering the
patient factors and treatment factors that most impact outcomes that
measure value and applying insight from the analysis of data to improve
these outcomes.
Over time, the analysis of data can produce algorithms to identify these
subpopulations. With insight from multiple feedback loops, the clinical
team can implement ideas for improvement which can result in lowered
costs while outcomes are improved over time. For example, when enough
data is accumulated and analyzed appropriately, a subpopulation can be
identified who would likely be harmed and another subpopulation that
would have no benefit from screening or interventions.
At the same time, another subpopulation of people could be identified
who would benefit, some of whom would not otherwise be receiving a test
or treatment. With these subpopulations better identified, the costs for
disease screening and treatment could plummet while other quality
measures would improve, resulting in better value for patients and the
system-as-a-whole.