0:03
A useful framework for understanding
different payment model was proposed in 2009 by Harold Miller.
Miller focused on per capita health care costs.
That is simply health care cost divided by population
as a mathematical product of five different ratios.
We begin with the first ratio on the far right,
the prevalence of health conditions in the population.
For example, how many people in the population suffer from heart disease?
The next ratio is the number of episodes of care people require per condition.
For example, how many heart attacks a person with a heart disease has?
The next ratio is the number and types of
health care services a person receives in each episode.
For example, when a person has a heart attack,
does he or she receives CABG surgery,
a stent, angioplasty, or medical management?
The next ratio is the number and types of activities,
devices, and drugs involved in each service.
For example, the type of stent the heart attack patient receives.
Finally, the fifth ratio is
the average cost of each of those individual activities, devices, and drugs.
For simplicity, let's assume that the cost per activity is given.
Out of the remaining four ratios,
where does the physician have the most influence?
Well, physicians can alter the number and types of activities,
devices, and drugs involving each service.
They can also affect the number and types of
health care services a person receives in each episode.
But when it comes to the number of episodes per condition,
the span of influence of the physician will depend on
their ability to effectively engage in population health management.
That is, to collect and analyze data on segments of
their patient population to manage specific diseases within that population.
This combination of analysis and action uncovering gaps in care,
as well as feeling those gaps,
require resources, time, manpower, and technology.
Therefore, the ability of physicians to influence this ratio is not trivial.
Finally, while effective prevention and
patient education can affect the prevalence of disease at least to some extent,
this ratio falls largely outside of the physician's control.
As we move to the right,
the physician's control over these ratios become weaker and weaker.
Now, let's use this formula to discuss how we pay providers.
The most common way of paying physician is called fee-for-service,
a pre-determined amount paid for each discrete service provider.
In other words, fee-for-service is
a payment model where services are unbundled and paid for separately.
In this framework, payments put the providers at
risk for the number and cost of activities within each service.
But there is no limit on the number of services,
and provider gets paid regardless of quality or outcomes.
Therefore, fee-for-service gives an incentive for physicians to provide
more services because payment is dependent on the quantity of care,
rather than quality of care.
Insurance companies address these issues by micromanaging providers using
requirements like prioritization for certain services and pay for performance schemes.
But these activities can be distortive and lead to unintended consequences.
But none of these changes,
the fundamental problem, that fee-for-service arrangements,
lead to supply and reduce demand.
On the other extreme, we have traditional computation,
a payment made on a per member base.
This applies to providers who sign a contract with
a health plan to cover the care of their patient panels.
The basic concept is that a physician receives
a single payment to cover all of the services their patients
need during a specific period of time regardless of
how many or few episodes of care their patients experience.
This is how most health maintenance organizations or HMO pay physicians.
The payment is usually determined on a per member per month basis.
The problem is that physicians are taking
financial risk if they receive the same payment for a patient on
their panel who is not using any care
and for patients who suffer from multiple conditions.
This is why payments are typically a risk adjusted.
That is, payments are differentiated based on
the characteristics of the enrollees in each provider panel.
Common risk adjustment factors include age, gender,
health status, and prior health care utilization,
as well as socio-demographic factors.
Regardless of the nuances of capitation,
it provides the opposite incentive as fee-for-service.
This is, to do too little as the payment
to physician is unrelated to the services they provide.
This would also be the case for physicians who are salaried by health care system.
Their salary is independent of the way they practice medicine.
Between these two extremes of capitation and fee-for-service,
there are two payment models that help balance over and underutilization.
The first is episode-based payments,
a single pre-determined payment for
all of the services needed by a patient during an entire episode of care.
Episode-based payment gives the provider responsibility
for one additional factor in the health care cost equation,
the number and types of services within an episode.
Examples of episode of care payment include DRG payment for hospitals and more recently,
bundled payments, which extend the episode of care,
post-discharge into other care providers.
This calls for greater coordination of care among providers.
The second Miller ground approach is comprehensive care payment,
also known as conditional adjusted capitation.
Here, the physician is at risk for the number of episodes of
care per condition but not for the prevalence of the condition.
Like in the case of traditional computation,
a physician would receive a single payment to cover
all of the services their patients need during the course of a year.
However, this payment will be adjusted based on the health of the patients.
For example, a provider would receive a higher payment if
he or she has more patients with hypertension or diabetes.
The physician is paid more for taking care of
sicker patients but not for providing more services,
as payment would not depend on what kind of treatment these patients receive.
Under this model, the physician is asked to take more risk compared with
episode-based care as they are asked to manage
their patients condition in a high quality and efficient manner.
Nevertheless, this model involves less risk compared to
traditional capitation as the physician is not asked to take on the insurance risk.
Each payment method has advantages and drawbacks,
and some are more appropriate in some settings than others.
But payment models are not sufficient in producing high value care for patients.
Ezekiel J. Emanuel, MD, PhDDiane v.S. Levy and Robert M. Levy University Professor
Guy David, PhDGilbert and Shelley Harrison Associate Professor of Health Care Management
