High-Need, High-Cost Patients
There is consensus that healthcare costs are too high, which we have considered
from a number of perspectives. Here, I want to look at the small number of patients who
use a disproportionate share of the resources—the high-need, high-cost patient. If you can
identify such patients prospectively, care management resources could be deployed cost-
effectively, as most studies of care management methods in patients selected on the basis
of a single diagnosis yield disappointing results. Recently, Berkman and associates
published an extensive review of multiple databases and a literature review to examine the
state of the art. 1
The authors point out there is no single definition of this group of patients. Previous
studies have looked at multiple chronic diagnoses, mental illness, patient demographics,
patient beliefs about their health, (and also their beliefs about the efficacy of medical
interventions.) Previous utilization has been examined, but has not proved as robust as
hoped as a way to predict future utilization. In their study they found an association with
the number of chronic conditions, amplified by behavioral health conditions and “social
Savitz and colleagues published an editorial critique of the paper. 2 They agree with
mixing both quantitative and qualitative data, as was done in the review, and particularly
cite the emphasis on patient-specific items such as activities of daily living, medication
adherence, behavioral health and social risks. They have found similar results looking at
their own health plan (UPMC.) They disagreed with the observation that demographic data
had no impact on ED or in-patient use, which they have found to be important predictors.
In discussing the difference, it seems to me the issue is how well demographics can be used
as a substitute for what used to be called socio-economic factors, or now social
determinants of health. They also note they include a look-back at utilization over 12
months, where this study used 6 months.
I find it ironic the editorial is called a real-world view, as none of the authors in
either paper is a clinician. So, let me provide a few observations based on clinical care of
dialysis patients, who start out as high-cost patients, but even then, there are different
patterns of “high-costs.” First, the high costs may be associated with an acute medical event
such as a heart attack, a car wreck, or pneumonia. These events are predictable from an
actuarial perspective, but no case management method is likely to prove useful for
individual patients. Furthermore, these events probably won’t occur again anytime soon, so
they are not good predictors of future needs and costs.
Second, the high costs may be associated with serious mental health issues or, more
commonly, behavioral issues and substance use disorders. I was in conversation last month
with a colleague concerning a patient with progressive chronic kidney disease, who was
also psychotic. I commented that my experience trying to provide dialysis to psychotic
patients was dismal unless the psychiatrist was able to find effective medications the
patient would take. He agreed. But the odds are this patient will end up in the ED at some
1 Berkman ND, Chang E, Siebert J, Ali R. Characteristics of High-Need, High-Cost Patients: A “Best Fit”
Framework Synthesis. Ann Intern Med 2022;175:1728-1741. doi:10.7326/M21-4562.
2 Savitz LA, Rice Williams K, Swayze D. Identifying High-Need, High-Cost Patients: A Real-World Perspective.
Ann Intern Med 2022;1751767-1768. doi:10.7326/M22-3161.
point and someone will try dialysis. Should psychosis be equated to metastatic cancer in
Behavioral issues are different in that they are hard to identify prospectively, and
most are not treatable. Furthermore, many are rational from the patient’s perspective. I am
reminded of a conversation I came into on dialysis rounds one day. The RN was frustrated
because she had tried to convince a patient to stop behaviors that were hurting his overall
health. She was young, with small children, and could not grasp that he was middle-aged,
with little social support, so getting another year on dialysis did not have any real appeal to
him. Certainly, it did not appeal enough for him to give up current “negative” behavior.
This also brings up the issue of whose behavior is the real issue. At a conference
some years ago, I was presented the case of a patient who skipped a lot of treatments and
went out binge-drinking the day his social security check came in. He would then show up
in the ER in respiratory distress and the staff would rush him to dialysis to “save him.” I
suggested they needed to know some day he was going to wait 30 minutes too long to show
up. If they recognized he had a fatal illness, they would do what they could, but might see it
as palliative, not curative, care. It would not make his care “better,” but it would make it
less stressful for all concerned.
Substance use disorders are also a special case. First, they are common—about 12-
15% of the population. Second, they are hard to treat. I can think of only one patient with
an active drug use problem, who got sober and stayed that way. If the patient has not
gotten sober by the time they get to end-stage disease, there is not much motivation for the
patient to change behaviors.
This leaves the third group, where knowledge deficiencies, or lack of resources is
causing high utilization. This is the group where case-management, or care-management is
able to “bend the cost curve.” Unfortunately, this is likely only about 10% of any given
population. That turns out to be a lot of people, and, again, identifying them in advance is
difficult. And the intervention that turns the tide may not be medical. I am reminded of a
patient who transferred to us already labeled as a fluid abuser. Educational interventions
by the staff were proving ineffective. But one day, after the “lecture,” an older woman in the
next chair leaned over and said: “I used to be hard-headed, too, but then I woke up in the
hospital with a tube in my throat, so now I measure out a quart jar’s worth of water and
keep it in the refrigerator. When it’s gone, I quit.” He took her advice and became a model
patient, even passing on his story to other patients.
Is there a moral here? When dealing with the specificity of individual patients, you
can predict high-need, high-cost patients about as well as the formulas, but it often does not
lead to an effective intervention. The issue is “two-handed.” On the one hand we have the
care processes, on the other hand we have the patient. Of course, in many of these cases,
continued medical intervention is not going to result in long periods of calm. It might be
reasonable, and certainly cost-effective, to move to palliative care. But for that to happen,
we must re-think a lot of our assumptions about what people really want from the health
care system. Not everyone thinks consumption of medical care at “maximal strength” is an
unmitigated good. Maybe we would be better off if we gave up the idea, too.
Activating Patients - The Achilles Heel of Healthcare Reform?
Studies show 25% of the population is not involved in their healthcare, but reform efforts assume wider application of evidence-based medicine is the key to better value. Perhaps it is the Achilles' Heel?
Changing Physician Behavior
A consideration of the interactions of patient preferences, evidence-based medicine and peer review.
A new study suggests only 5% of Medicare spending in 2012 was preventable, much of it in frail, elderly patients. Is this good news or bad?
Risk, Reward, and Other Reasons Patients Don't Follow Medical Advice
Patients often don't do what their doctors recommend. The problem is important and contributes to "bad" outcomes, yet we have little insight into the problem.