The NEW ENGLA ND JOURNAL of MEDICINE

Perspective december 23, 2010

What Is Value in Health Care?

Michael E. Porter, Ph.D.

I n any field, improving performance and account-

ability depends on having a shared goal that
unites the interests and activities of all stakehold-
value is a central challenge. Nor
is value measured by the process
of care used; process measure-
ment and improvement are im-
ers. In health care, however, stakeholders have portant tactics but are no sub-
stitutes for measuring outcomes
myriad, often conflicting goals, Value — neither an abstract and costs.
including access to services, prof- ideal nor a code word for cost Since value is defined as out-
itability, high quality, cost con- reduction — should define the comes relative to costs, it encom-
tainment, safety, convenience, framework for performance im- passes efficiency. Cost reduction
patient-centeredness, and satis- provement in health care. Rigor- without regard to the outcomes
faction. Lack of clarity about ous, disciplined measurement and achieved is dangerous and self-
goals has led to divergent ap- improvement of value is the best defeating, leading to false “sav-
proaches, gaming of the system, way to drive system progress. Yet ings” and potentially limiting
and slow progress in performance value in health care remains large- effective care.
improvement. ly unmeasured and misunder- Outcomes, the numerator of
Achieving high value for pa- stood. the value equation, are inherently
tients must become the over- Value should always be de- condition-specific and multidi-
arching goal of health care de- fined around the customer, and mensional. For any medical con-
livery, with value defined as the in a well-functioning health care dition, no single outcome cap-
health outcomes achieved per system, the creation of value for tures the results of care. Cost,
dollar spent.1 This goal is what patients should determine the the equation’s denominator, re-
matters for patients and unites rewards for all other actors in fers to the total costs of the full
the interests of all actors in the the system. Since value depends cycle of care for the patient’s
system. If value improves, patients, on results, not inputs, value in medical condition, not the cost
payers, providers, and suppliers health care is measured by the of individual services. To reduce
can all benefit while the eco- outcomes achieved, not the vol- cost, the best approach is often
nomic sustainability of the health ume of services delivered, and to spend more on some services
care system increases. shifting focus from volume to to reduce the need for others.

n engl j med 363;26  nejm.org  december 23, 2010 2477

The New England Journal of Medicine
Downloaded from nejm.org on January 31, 2011. For personal use only. No other uses without permission.
Copyright © 2010 Massachusetts Medical Society. All rights reserved.
PERSPE C T I V E What Is Value in Health Care?

Health care delivery involves cused factories” concentrating on structure also helps explain why
numerous organizational units, narrow groups of interventions, physicians fail to accept joint re-
ranging from hospitals to physi- we need integrated practice units sponsibility for outcomes, blam-
cians’ practices to units provid- that are accountable for the to- ing lack of control over “out-
ing single services, but none of tal care for a medical condition side” actors involved in care (even
these reflect the boundaries with- and its complications. those in the same hospital) and
in which value is truly created. Because care activities are in- patients’ compliance.
The proper unit for measuring terdependent, value for patients The concept of quality has it-
value should encompass all ser- is often revealed only over time self become a source of confu-
vices or activities that jointly de- and is manifested in longer-term sion. In practice, quality usually
termine success in meeting a set outcomes such as sustainable means adherence to evidence-
of patient needs. These needs recovery, need for ongoing in- based guidelines, and quality
are determined by the patient’s terventions, or occurrences of measurement focuses overwhelm-
medical condition, defined as an treatment-induced illnesses.2 The ingly on care processes. For ex-
interrelated set of medical circum- only way to accurately measure ample, of the 78 Healthcare Ef-
stances that are best addressed value, then, is to track patient out- fectiveness Data and Information
in an integrated way. The defini- comes and costs longitudinally. Set (HEDIS) measures for 2010,
tion of a medical condition in- For patients with multiple the most widely used quality-
cludes the most common associ- medical conditions, value should measurement system, all but 5 are
ated conditions — meaning that be measured for each condition, clearly process measures, and
care for diabetes, for example, with the presence of the other none are true outcomes.3 Process
must integrate care for condi- conditions used for risk adjust- measurement, though a useful
tions such as hypertension, renal ment. This approach allows for internal strategy for health care
disease, retinal disease, and vas- relevant comparisons among pa- institutions, is not a substitute
cular disease and that value should tients’ results, including compar- for measuring outcomes. In any
be measured for everything in- isons of providers’ ability to care complex system, attempting to
cluded in that care.1 for patients with complex condi- control behavior without measur-
For primary and preventive tions. ing results will limit progress to
care, value should be measured The current organizational incremental improvement. There
for defined patient groups with structure and information sys- is no substitute for measuring ac-
similar needs. Patient populations tems of health care delivery tual outcomes, whose principal
requiring different bundles of pri- make it challenging to measure purpose is not comparing pro-
mary and preventive care services (and deliver) value. Thus, most viders but enabling innovations
might include, for example, providers fail to do so. Providers in care. Without such a feedback
healthy children, healthy adults, tend to measure only what they loop, providers lack the requisite
patients with a single chronic dis- directly control in a particular information for learning and im-
ease, frail elderly people, and pa- intervention and what is easily proving. (Further details about
tients with multiple chronic con- measured, rather than what mat- measuring value are contained
ditions. ters for outcomes. For example, in a framework paper, “Value in
Care for a medical condition current measures cover a single Health Care,” in Supplementary
(or a patient population) usually department (too narrow to be Appendix 1, available with the full
involves multiple specialties and relevant to patients) or outcomes text of this article at NEJM.org.)
numerous interventions. Value for for a whole hospital, such as in- Measuring, reporting, and com-
the patient is created by provid- fection rates (too broad to be paring outcomes are perhaps the
ers’ combined efforts over the relevant to patients). Or they most important steps toward rap-
full cycle of care. The benefits of measure what is billed, even idly improving outcomes and mak-
any one intervention for ultimate though current reimbursement ing good choices about reducing
outcomes will depend on the ef- practices are misaligned with costs.4 Systematic, rigorous out-
fectiveness of other interventions value. Similarly, costs are mea- come measurement remains rare,
throughout the care cycle. sured for departments or billing but a growing number of exam-
Accountability for value should units rather than for the full ples of comprehensive outcome
be shared among the providers care cycle over which value is measurement provide evidence of
involved. Thus, rather than “fo- determined. Faulty organizational its feasibility and impact.

2478 n engl j med 363;26  nejm.org  december 23, 2010

The New England Journal of Medicine

Downloaded from nejm.org on January 31, 2011. For personal use only. No other uses without permission.
Copyright © 2010 Massachusetts Medical Society. All rights reserved.
PERSPECTIVE What Is Value in Health Care?

Determining the group of rel-

evant outcomes to measure for
any medical condition (or patient
population in the context of pri- Survival
mary care) should follow several Tier 1
principles. Outcomes should in- Health status
achieved
clude the health circumstances or retained
most relevant to patients. They Degree of health or recovery
should cover both near-term and
longer-term health, addressing a
period long enough to encom-
pass the ultimate results of care.
And outcome measurement should
include sufficient measurement of Time to recovery and time to return
risk factors or initial conditions to normal activities
to allow for risk adjustment.
Tier 2
For any condition or popula-
Process
tion, multiple outcomes collec- of recovery Disutility of care or treatment process
tively define success. The com- (e.g., diagnostic errors, ineffective care,
plexity of medicine means that treatment-related discomfort, compli-
cations, adverse effects)
competing outcomes (e.g., near-
term safety versus long-term func-
tionality) must often be weighed
against each other.
The outcomes for any medi-
cal condition can be arrayed in a
Sustainability of health or recovery
three-tiered hierarchy (see Figure and nature of recurrences
Recurrences
1), in which the top tier is gen- Tier 3
erally the most important and Sustainability
lower-tier outcomes involve a pro- of health
Long-term consequences of therapy Care-induced
gression of results contingent on (e.g., care-induced illnesses) illnesses
success at the higher tiers. Each
tier of the framework contains
two levels, each involving one or
more distinct outcome dimen- Figure 1. The Outcome Measures Hierarchy.
sions. For each dimension, suc-
cess is measured with the use of
one or more specific metrics. heavily. The second level in Tier secondary process measure, as
Tier 1 is the health status 1 is the degree of health or re- some believe. Delays in diagnosis
that is achieved or, for patients covery achieved or retained at the or formulation of treatment plans
with some degenerative condi- peak or steady state, which nor- can cause unnecessary anxiety.
tions, retained. The first level, mally includes dimensions such Reducing the cycle time (e.g.,
survival, is of overriding impor- as freedom from disease and rel- time to reperfusion after myo-
tance to most patients and can evant aspects of functional status. cardial infarction) can improve
be measured over various periods Tier 2 outcomes are related to functionality and reduce compli-
appropriate to the medical condi- the recovery process. The first level cations. The second level in Tier 2
tion; for cancer, 1-year and 5-year is the time required to achieve is the disutility of the care or
survival are common metrics. recovery and return to normal or treatment process in terms of dis-
Maximizing the duration of sur- best attainable function, which comfort, retreatment, short-term
vival may not be the most im- can be divided into the time complications, and errors and
portant outcome, however, espe- needed to complete various phases their consequences.
cially for older patients who may of care. Cycle time is a critical Tier 3 is the sustainability of
weight other outcomes more outcome for patients — not a health. The first level is recur-

n engl j med 363;26  nejm.org  december 23, 2010 2479

The New England Journal of Medicine
Downloaded from nejm.org on January 31, 2011. For personal use only. No other uses without permission.
Copyright © 2010 Massachusetts Medical Society. All rights reserved.
PERSPE C T I V E What Is Value in Health Care?

Primary Acute Knee Osteoarthritis

Breast Cancer Dimensions Requiring Replacement

Survival rate (1-yr, 3-yr, 5-yr, longer) Survival Mortality rate (inpatient)

Remission Functional level achieved

Functional status Degree of health or recovery Pain level achieved
Breast preservation Extent of return to physical activities
Breast-conservation-surgery outcomes Ability to return to work

Time to remission Time to treatment

Time to recovery and time to return
Time to achievement of functional Time to return to physical activities
to normal activities
and cosmetic status Time to return to work

Nosocomial infection Pain

Nausea or vomiting Length of hospital stay
Febrile neutropenia Disutility of care or treatment process Infection
Limitation of motion (e.g., diagnostic errors, ineffective care, Pulmonary embolism
Breast reconstruction discomfort or treatment-related discomfort, compli- Deep-vein thrombosis
complications cations, adverse effects) Myocardial infarction
Depression Immediate revision
Delirium

Cancer recurrence Maintained functional level

Sustainability of health or recovery
Consequences of recurrence Ability to live independently
and nature of recurrences
Sustainability of functional status Need for revision or reoperation

Incidence of second primary cancers Loss of mobility due to inadequate

Brachial plexopathy rehabilitation
Premature osteoporosis Risk of complex fracture
Long-term consequences of therapy
Susceptibility to infection
(e.g., care-induced illnesses)
Stiff knee due to unrecognized
complication
Regional pain syndrome

Figure 2. Outcome Hierarchies for Breast Cancer and Knee Osteoarthritis.

rences of the original disease or more timely, reducing discomfort, Improving one outcome di-
longer-term complications. The and minimizing recurrence. mension can benefit others. For
second level captures new health Each medical condition (or example, more timely treatment
problems created as a consequence population of primary care pa- can improve recovery. However,
of treatment. When recurrences tients) will have its own outcome measurement can also make ex-
or new illnesses occur, all out- measures. Measurement efforts plicit the tradeoffs among out-
comes must be remeasured. should begin with at least one come dimensions. For example,
With some conditions, such as outcome dimension at each tier, achieving more complete recov-
metastatic cancers, providers may and ideally one at each level. As ery may require more arduous
have a limited effect on survival experience and available data in- treatment or confer a higher
or other Tier 1 outcomes, but they frastructure grow, the number of risk of complications. Mapping
can differentiate themselves in dimensions (and measures) can these trade­offs, and seeking
Tiers 2 and 3 by making care be expanded. ways to reduce them, is an essen-

2480 n engl j med 363;26  nejm.org  december 23, 2010

The New England Journal of Medicine

Downloaded from nejm.org on January 31, 2011. For personal use only. No other uses without permission.
Copyright © 2010 Massachusetts Medical Society. All rights reserved.
PERSPECTIVE What Is Value in Health Care?

tial part of the care-innovation times even counterproductive. slowed innovation, led to ill-
process. Today, health care organizations advised cost containment, and
Figure 2 illustrates possible out- measure and accumulate costs encouraged micromanagement of
come dimensions for breast cancer around departments, physician physicians’ practices, which im-
and acute knee osteoarthritis re- specialties, discrete service areas, poses substantial costs of its
quiring knee replacement. Most and line items such as drugs and own. Measuring value will also
current measurement efforts fail supplies — a reflection of the permit reform of the reimburse-
to capture such comprehensive organization and financing of ment system so that it rewards
sets of outcomes, which are need- care. Costs, like outcomes, should value by providing bundled pay-
ed to fully describe patients’ re- instead be measured around the ments covering the full care cycle
sults. No organization I know of patient. Measuring the total costs or, for chronic conditions, cover-
systematically measures the en- over a patient’s entire care cycle ing periods of a year or more.
tire outcome hierarchy for the and weighing them against out- Aligning reimbursement with val-
medical conditions for which it comes will enable truly struc- ue in this way rewards providers
provides services, though some tural cost reduction, through steps for efficiency in achieving good
are making good progress. (Fur- such as reallocation of spending outcomes while creating account-
ther details, including risk adjust- among types of services, elimi- ability for substandard care.
ment, are addressed in a frame- nation of non–value-adding ser- Disclosure forms provided by the author
work paper, “Measuring Health vices, better use of capacity, are available with the full text of this arti-
cle at NEJM.org.
Outcomes,” in Supplementary Ap- shortening of cycle time, provision
pendix 2, available at NEJM.org.) of services in the appropriate From Harvard Business School, Boston.
The most important users of settings, and so on.
This article (10.1056/NEJMp1011024) was
outcome measurement are pro- Much of the total cost of car- published on December 8, 2010, at NEJM
viders, for whom comprehensive ing for a patient involves shared .org.
measurement can lead to sub- resources, such as physicians,
1. Porter ME, Teisberg EO. Redefining health
stantial improvement.5 Outcomes staff, facilities, and equipment. care: creating value-based competition on
need not be reported publicly to To measure true costs, shared re- results. Boston: Harvard Business School
benefit patients and providers, and source costs must be attributed Press, 2006.
2. Institute of Medicine. Performance mea-
public reporting must be phased to individual patients on the basis surement: accelerating improvement. Wash-
in carefully enough to win pro- of actual resource use for their ington, DC: National Academies Press, 2006.
viders’ confidence. Progression to care, not averages. The large cost 3. National Committee for Quality Assurance
(NCQA). HEDIS and quality measurement:
public reporting, however, will ac- differences among medical con- technical resources. NCQA Web site. (http://
celerate innovation by motivating ditions, and among patients with www.ncqa.org/tabid/1044/Default.aspx.)
providers to improve relative to the same medical condition, reveal 4. Porter ME. Defining and introducing val-
ue in health care. In: Evidence-based medi-
their peers and permitting all additional opportunities for cost cine and the changing nature of health care:
stakeholders to benefit fully from reduction. (Further aspects of cost 2007 IOM annual meeting summary. Wash-
outcome information. measurement and reduction are ington, DC: Institute of Medicine, 2008:161-
72.
Current cost-measurement ap- discussed in the framework pa- 5. Porter ME, Baron JF, Chacko JM, Tang RJ.
proaches have also obscured value per “Value in Health Care.”) The UCLA Medical Center: kidney transplan-
in health care and led to cost- The failure to prioritize value tation. Harvard Business School Case 711-
410. Boston: Harvard Business School Pub-
containment efforts that are in- improvement in health care de- lishing, 2010.
cremental, ineffective, and some- livery and to measure value has Copyright © 2010 Massachusetts Medical Society.

Putting the Value Framework to Work

Thomas H. Lee, M.D.

“V alue” is a word that has

long aroused skepticism
among physicians, who suspect
ing number of health care deliv-
ery organizations, including my
own, now describe enhancement
cepts developed by Michael Porter
(see pages 2477–2480, and the
framework papers in Supplemen-
it of being code for “cost reduc- of value for patients as a funda- tary Appendixes 1 and 2 of that
tion.” Nevertheless, an increas- mental goal and are using con- article) to shape their strategies.

n engl j med 363;26  nejm.org  december 23, 2010 2481

The New England Journal of Medicine
Downloaded from nejm.org on January 31, 2011. For personal use only. No other uses without permission.
Copyright © 2010 Massachusetts Medical Society. All rights reserved.