Physicians' fears of malpractice lawsuits are not assuaged by tort reforms.

Abstract

Physicians contend that the threat of malpractice lawsuits forces them to practice defensive medicine, which in turn raises the cost of health care. This argument underlies efforts to change malpractice laws through legislative tort reform. We evaluated physicians' perceptions about malpractice claims in states where more objective indicators of malpractice risk, such as malpractice premiums, varied considerably. We found high levels of malpractice concern among both generalists and specialists in states where objective measures of malpractice risk were low. We also found relatively modest differences in physicians' concerns across states with and without common tort reforms. These results suggest that many policies aimed at controlling malpractice costs may have a limited effect on physicians' malpractice concerns.

Full text

Page 1

By Emily R. Carrier, James D. Reschovsky, Michelle M. Mello, Ralph C. Mayrell, and David Katz
Physicians’ Fears Of Malpractice
Lawsuits Are Not Assuaged
By Tort Reforms
ABSTRACT Physicians contend that the threat of malpractice lawsuits
forces them to practice defensive medicine, which in turn raises the cost
of health care. This argument underlies efforts to change malpractice
laws through legislative tort reform. We evaluated physicians’ perceptions
about malpractice claims in states where more objective indicators of
malpractice risk, such as malpractice premiums, varied considerably. We
found high levels of malpractice concern among both generalists and
specialists in states where objective measures of malpractice risk were
low. We also found relatively modest differences in physicians’ concerns
across states with and without common tort reforms. These results
suggest that many policies aimed at controlling malpractice costs may
have a limited effect on physicians’ malpractice concerns.
A
coming the subject of a malpractice lawsuit.2
Fear of being sued may compromise physi-
cians’ ability to communicate effectively with
patients, particularly in disclosing medical er-
rors.3Physicians with high malpractice insur-
ance premiums, which reflect a risky liability
environment, have lower career satisfaction
and report more adversarial relationships with
patients than do physicians with lower premi-
ums.4Physicians with high premiums are also
more likely to order diagnostic testing and hos-
pitalize low-risk patients in some settings.5
Federal health reform has heightened con-
cerns about defensive medicine for two reasons.
First, the financial and organizational changes
wrought by health reform have introduced new
sourcesofstressforhealthcareproviders,sharp-
ening their demands for liability reform in ex-
change for their support on other health reform
measures. Second, because it leads to defensive
lthoughanalystsdisagreeaboutthe
scope and cost of defensive medi-
cine,1physicians consistently re-
port that they often engage in
defensive practices and that they
feel intense pressure to do so out of fear of be-
medicine, liability risk is an obstacle to health
reform’s ambition of moving physicians toward
more cost-effective care.6
In this article we report findings concerning
perceptions of malpractice risk among a nation-
allyrepresentativesampleofphysicians.Ourob-
jectives were to assess levels of physician
concern about malpractice, examine associa-
tions between level of concern and physician
practice characteristics, and relate these con-
cerns to objective measures of malpractice risk,
includingstatemedicalmalpracticereformlaws.
We found that individual physicians’ concerns
about their own malpractice risk are pervasive,
vary across specialties in ways that are likely to
reflect underlying malpractice risk, and reflect
objective measures of risk across states to a lim-
iteddegree.Our resultssuggest thatmany popu-
lar tort reforms are only modestly associated
withthelevelof physicians’malpracticeconcern
and their practice of defensive medicine. The
results raise the possibility that physicians’ level
of concern reflects a common tendency to over-
estimatethelikelihoodof“dreadrisks”—rarebut
devastating outcomes—not an accurate assess-
ment of actual risk.
doi:
HEALTH AFFAIRS 29,
NO. 9 (2010): 1585–1592
©2010 Project HOPE—
The People-to-People Health
Foundation, Inc.
10.1377/hlthaff.2010.0135
Emily R. Carrier (ecarrier@
hschange.org) is a senior
health researcher at the
Center for Studying Health
System Change, in
Washington, D.C.
James D. Reschovsky is a
senior health researcher at
the Center for Studying
Health System Change.
Michelle M. Mello is a
professor of law and public
health at the Harvard School
of Public Health, in Boston,
Massachusetts.
Ralph C. Mayrell is a research
assistant at the Center for
Studying Health System
Change.
David Katz is an associate
professor in the Department
of Internal Medicine at the
University of Iowa, in Iowa
City.
September 2010 29:9Health Affairs
1585
Malpractice & Errors

Page 2

Study Data And Methods
Data Physician data were obtained from the
2008CenterforStudyingHealthSystemChange
(HSC) Health Tracking Physician Survey, a na-
tionally representative mail survey of U.S. physi-
cians who provide at least twenty hours of direct
patientcareper week. The surveywas sponsored
by the Robert Wood Johnson Foundation. The
sampleof physicians was drawnfrom the Ameri-
can Medical Association (AMA) Physician Mas-
terfile and included active, nonfederal, office-
and hospital-based physicians. Residents and
fellows were excluded, along with radiologists,
anesthesiologists, and pathologists.
The survey had a response rate of 62 percent
(N ¼ 4;720). It asked a broad array of questions
regarding physicians’ demographic and practice
characteristics, as well as subjective questions
dealing with such issues as career satisfaction
and concerns about malpractice.7
To assess the association between malpractice
concernsandstate-leveldataonmalpracticerisk
and malpractice premiums, we used secondary
data from the National Practitioner Data Bank,
available on the Kaiser Family Foundation Web
site;8,9the Medical Liability Monitor;10market
share reports published by the National Associ-
ation of Insurance Commissioners;11and the
AMA Physician Masterfile, obtained from the
Kaiser Family Foundation Web site.12Malprac-
ticepremiumdataforobstetricsandgynecology,
general surgery, and internal medicine from the
Medical Liability Monitor were weighted by mar-
ket share data from the National Association of
Insurance Commissioners. Information on state
tort reforms affecting malpractice litigation was
obtained from the database of state tort law re-
forms, developed by Ronen Avraham.13Each re-
form was considered separately.
With cross-sectional data, it is difficult to infer
a causal association between specific laws and
physicians’ malpractice concerns. Some states
may have adopted multiple laws that changed
the way malpractice claims are addressed, in-
cluding caps on various types of damages, as a
way to respond to existing high levels of overall
malpractice risk. To capture the temporal rela-
tionshipbetweenstates’policiesand physicians’
concerns, we used data on medical malpractice
laws in effect in 2007, one year before the 2008
physician survey. (See the Appendix for a de-
scription of state policies.)14
Assessment Of Concerns The survey in-
cluded questions from a malpractice concerns
scale developed and validated by Kevin Fiscella
andcolleagues.15–19Thequestionsaskedrespond-
ents to indicate how strongly they agreed with
the following statements based on a five-point
Likert scale, ranging from “strongly disagree” to
“stronglyagree”:(1)IamconcernedthatIwillbe
involved in a malpractice case sometime in the
next ten years. (2) I feel pressured in my day-to-
day practice by the threat of malpractice litiga-
tion. (3) I order some tests or consultations sim-
ply to avoid the appearance of malpractice.
(4) Sometimes I ask for consultant opinions pri-
marily to reduce my risk of getting sued. (5) Re-
lying on clinical judgment rather than on
technology to make a diagnosis is becoming
risky because of the threat of malpractice suits.
We computed the percentage of statements
with which each respondent agreed or strongly
agreed, across the five statements. The resulting
compositescoreisreportedonascaleof0to100.
We compared regression-adjusted means of
thecompositescoreacrossrespondentswithdif-
ferent individual and practice characteristics, as
well as across physicians in different groups of
states as defined by values on various measures
of malpractice risk, including enacted tort re-
forms. We also used regression-adjusted means
to compare composite scores between specialty
groupsandtocomparephysiciansacrosstertiles
(thirds) of statewide malpractice risk.
We controlled for differences in the character-
istics of physicians, practices, and patient pan-
els. Those characteristics included physician’s
sex, years in practice, and practice type; number
of physicians in practice; percentage of practice
revenue from Medicare and from Medicaid; per-
centage of patients who suffer from chronic dis-
eases; and percentage of patients who are
members of racial and ethnic minority groups.
Generally, adjusted means differed little from
unadjusted ones.
We further report the results of two distinct
subscales representing malpractice concerns
(statements 1, 2, and 5 on the malpractice con-
cerns scale) and defensive medicine (statements
3 and 4 on the scale). All analyses used survey
weights to adjust for probability of selection and
differential survey nonresponse.
Limitations Our study has limitations. Our
measure of malpractice insurance premiums is
at the state level and does not reflect the pre-
mium burden experienced by individual re-
spondents. Similarly, we do not have any
informationonindividualphysicians’awareness
of individual tort reforms intended to limit mal-
practice claims.
We have no measure of claims that are closed
butdidnotresultinpayment,whichnonetheless
might cause distressand professional and finan-
cial loss to physicians. Performing a statistical
adjustment used in previous studies to approxi-
mate the number of closed claims did not reveal
new significant associations with tort reforms.2
Our sample population excludes radiologists
Malpractice & Errors
1586
Health Affairs September 201029:9

Page 3

and anesthesiologists—specialists known to
have high levels of concern about malpractice.
Finally, our survey measures cannot be inter-
preted as a direct measurement of defensive
medical practices. Rather, our aim was to mea-
sure physicians’ level of fear or concern about
liability, which is a subjective construct.
Study Results
Concern About Malpractice Liability Con-
cern about malpractice liability is pervasive
among physicians: 60–78 percent of them ex-
pressed agreement or strong agreement with
each of the five statements (Exhibit 1). Physi-
cians agreed most strongly with the statement
that it is becoming increasingly risky to rely on
clinical judgment, rather than diagnostic test-
ing; 78 percent expressed agreement or strong
agreement with that statement. Only 11 percent
did not agree with any of the statements.
Variations In Concern Across Specialties
Malpractice concern varied considerably by spe-
cialty. Although we lack objective data on
malpracticeriskorpremiumsbyspecialty,physi-
cians in specialties generally thought to be at
higher risk for malpractice claims—such as
emergency physicians and obstetrician-gynecol-
ogists—expressed greater concern (Exhibit 2).
Physicians who disagreed or strongly disagreed
with the five statements were more likely to be
psychiatrists or general pediatricians.
This pattern was similar for the malpractice
concern and defensive medicine subscales.
There was some variation among specialties,
but it does not appear to be consistent.
Variation Across Other Physician And
Practice
Exhibit 3, the level of malpractice concern was
associated with several physician and practice
characteristics (only characteristics for which
there were significant differences are shown in
the exhibit; the full set of results is available in
Appendix Exhibit 2).14Physicians with fewer
thanfiveyearsofpracticeexperiencehadsignifi-
cantlygreatermalpracticefear—averageconcern
score of 70.4 points—than physicians with more
than ten years of experience—average concern
score of 64.4 points.
Practicing in a larger group was associated
with greater malpractice concern. Physicians
in practices with eleven to fifty doctors (with
an average concern score of 68.8 points) ex-
pressed higher levels of concern than did physi-
cians in solo or two-person practices (with an
average concern score of 65.1 points). Group/
staff health maintenance organization (HMO)
physicians had average concern scores of only
60.9 points. However, physicians in this kind of
practice reported undertaking defensive medi-
cine practices that were not significantly differ-
ent from those used by any other group.
The proportion of patients with a chronic
illness affected levels of malpractice concern.
Physicians whose practices were predominantly
(more than 50 percent) patients with chronic
illnesses had average concern scores of 66.8
points,whilephysicianswhosepracticeshadrel-
atively few such patients (less than 10 percent)
reported average concern scores of 60.8 points.
The use of health information technology (IT)
wasnotassociatedwithsignificantdifferencesin
malpractice concern. This was true whether
health IT use was determined by the use of an
Characteristics
As shownin
EXHIBIT 1
Physicians’ Level Of Agreement With Items In The Malpractice Concerns Scale, 2008
Strongly disagree Disagree Neutral Agree Strongly agree
Relying on clinical judgment rather than on technology to make a
diagnosis is becoming risky because of the threat of malpractice suits
Sometimes I ask for consultant opinions primarily to reduce
my risk of being sued
I order some tests or consultations simply to avoid
the appearance of malpractice
I feel pressured in my day-to-day practice by the threat of
malpractice litigation
I am concerned that I will be involved in a malpractice case
sometime in the next ten years
Percent
SOURCE Center for Studying Health System Change (HSC) Health Tracking Physician Survey, 2008.
September 2010 29:9 Health Affairs
1587

Page 4

electronic medical record; the use of an elec-
tronic record with clinical decision support; or
the use of an electronic record with automated
reminders, e-prescribing, decision support, and
other features. Variation in response patterns
across the subscales was minimal.
State Liability Environment We compared
levels of malpractice concern in states with vary-
ing levels of medical liability, as represented by
several different measures (Exhibit 4). Results
that were not significant are not shown. Full
results are in Appendix Exhibit 3.14
There is wide state-to-state variation in physi-
cians’ risk of incurring a malpractice claim—
through either a settlement or a trial verdict—
as well as the average size of paid claims.20The
average actual malpractice risk in the one-third
of states with the highest values (as defined by
the number of paid claims multiplied by the size
of the awards) is more than three times that
found in the third of states with the lowest val-
ues—$5,081 versus $1,662 per physician. Aver-
age actual malpractice risk is defined as the rate
of malpractice claims per 1,000 physicians that
providers or their designees must pay, multi-
plied by the average dollar amount of the award.
Althoughphysicians’malpracticeconcernwas
positively and significantly associated with aver-
age malpractice risk, the relationship is fairly
weak in light of the more than threefold differ-
ence in objective measures of risk. Physicians in
the highest-risk states had survey composite
scores only 4.3 percentage points higher than
those practicing in the third of states with the
lowest risk: 67.8 percent versus 63.5 percent
(p < 0:01). These general trends also apply to
the components of malpractice risk—the paid
claims rate and average award size. However,
only the comparison between the highest and
lowest one-third is statistically significant.
Thesamepatternappliestomalpracticeinsur-
ance premiums. There is nearly a threefold dif-
ferencebetween average
malpractice premiums in the bottom and top
thirds of states. Yet physician survey composite
scores in the third of states with the highest
premiums were 66.2 percent—just 5.4 percent-
age points higher than comparable scores in the
third of states with the lowest premiums, where
they were 60.8 percent (p < 0:01).
We examined the relationship of malpractice
concern to several state tort reforms (Appendix
Exhibit 1).14Empirical research has demon-
strated that a few reforms—most notably, caps
on noneconomic damages—can affect liability
insurance premiums and the use of services con-
sidered to be indicative of defensive medicine.21
Overall, physicians’ malpractice concerns ap-
pear to be relatively insensitive to their states’
malpractice reforms, including caps on noneco-
specialty-adjusted
EXHIBIT 2
Adjusted Percentage Of Items In The Malpractice Concerns Scale With Which Physicians Agreed Or Strongly Agreed,
By Specialty, 2008
Specialty
All physicians
Emergency physicians
Obstetrician-gynecologists
Surgical specialists
Adult primary care physicians
(reference group)
Pediatric specialists
Adult cognitive specialistsd
Adult procedural specialistsd
General pediatricians
Psychiatrists (adult and pediatric)
% of
physicians
100.0
5.8
6.6
21.3
Composite
scorea
65.4
82.0***
77.2***
71.4***
Defensive medicine
subscoreb
62.0
77.3***
68.5
65.5
Malpractice concern
subscorec
67.7
83.3***
81.0***
75.3***
31.1
2.1
10.3
8.2
7.6
6.9
66.4
59.6
59.0***
58.6***
57.4***
51.4***
66.5
50.8***
55.8***
51.5***
56.3***
43.8***
65.8
62.5
61.7**
65.8
56.3***
54.2***
SOURCE Center for Studying Health System Change (HSC) Health Tracking Physician Survey, 2008. NOTES Adjusted for physician’s sex,
years in practice, and practice type; number of physicians in practice; percentage of revenue from Medicare and from Medicaid;
percentage of patients with chronic illnesses; and percentage of patients who are members of a racial or ethnic minority group.
Percentages may not add to 100 percent because of rounding.
strongly agreed.
or strongly agreed.
strongly agreed.
uncertainty or recommend a course of treatment. Procedural specialists’ primary role involves performing a technical procedure
to aid diagnosis, cure a condition, identify and prevent new conditions, or palliate symptoms. See Forrest C. A typology of
specialists’ clinical roles. Arch Intern Med. 2009;169(11):1062–8. **p ≤ 0.05 ***p ≤ 0.01
aPercentage of statements with which physicians agreed or
bPercentage of statements related to defensive ordering of tests or consultations with which physicians agreed
cPercentage of statements related to overall concern regarding malpractice with which physicians agreed or
dCognitive specialists’ primary role involves providing diagnostic or therapeutic advice to reduce clinical
Malpractice & Errors
1588
Health Affairs September 201029:9

Page 5

nomic and punitive damages. Again, variation
across subscales was minimal.
States that had established caps on total dam-
ages or abolished joint-and-several liability22
were associated with modestly lower levels of
physicianmalpracticeconcern.Differencesasso-
ciatedwithothertortreforms,suchascollateral-
source rule reform and periodic payment re-
form,23were not statistically significant. Two re-
forms, split recovery and patient compensation
funds,24
wereassociated
higher levels of concern.
withsignificantly
Discussion
This study of a nationally representative sample
of physicians found high levels of concern about
the risk of malpractice litigation among physi-
cians across a range of specialties, practice set-
tings, and geographic areas. Physicians in
specialties generally considered to be at highest
risk for costly malpractice claims, such as emer-
gency medicine, expressed the greatest concern.
The relationship between physicians’ level of
malpractice concern and some objective mea-
sures of the riskiness of the state liability envi-
ronment, such as malpractice premium levels
and the risk of incurring a paid malpractice
claim, was statistically significant. But the mag-
nitude of these associations was very modest.
To put our results in perspective, the largest
difference inphysician
tertiles—or thirds—of malpractice risk was 5.4
points on a 100-point scale. This is roughly
equivalent to the observed difference in concern
betweenanaveragegeneralsurgeonandanaver-
age primary care provider, or one-third of the
differencebetweentheaverageemergencyphysi-
cian and the average primary care provider.
For other measures, such as the number of
paid claims and the average amount paid per
claim, physicians with twice the objective mea-
concernacross
EXHIBIT 3
Adjusted Percentage Of Items In The Malpractice Concerns Scale With Which Physicians Agreed Or Strongly Agreed,
By Physician And Practice Characteristics, 2008
Characteristic
All physicians
Years in practice
Fewer than 5 (ref)
5–10
More than 10
Sex
Male (ref)
Female
Percent of patients with a chronic illness
<10% (ref)
10%–49%
≥50%
Practice type/number of physicians
1–2 physicians (ref)
3–10 physicians
11–50 physicians
≥51 physicians
Group/staff HMO
Hospital, CHC, or other
Percent of physicians
100.0
Composite scorea
65.4
8.2
21.1
70.7
70.4
66.9
64.4**
72.5
27.4
67.2
60.7***
9.9
28.4
61.7
60.8
64.1
66.8***
32.0
24.2
9.7
6.1
3.5
24.5
65.1
67.4
68.8**
67.6
60.9
62.7
SOURCE Center for Studying Health System Change (HSC) Health Tracking Physician Survey, 2008. NOTES Excluding the characteristic
of interest, reported malpractice concern scores are adjusted for physician specialty, sex, and years in practice; practice type and
number of physicians; percentage of revenue from Medicare and from Medicaid; percentage of patients who suffer from chronic
disease; and percentage of patients who are members of racial or ethnic minority groups. This exhibit omits characteristics for
which no significant differences were found at the 0.05 level. The omitted characteristics are percentage minority patients, use
of health information technology with clinical decision support, routine use of full electronic medical record, and routine use of
full electronic medical record with decision support. We also omitted urbanicity of practice location. We tested for differences
between urban areas with a population of one million or more—the reference group—and urban areas with a population of less
than one million and nonurban areas. The only difference we detected compared to the reference group was in nonurban areas
(p < 0:05) on the malpractice concern score. Full results are available in Appendix Exhibit 2, which can be accessed by clicking on
the Appendix link in the box to the right of the article online. Percentages may not add to 100 percent because of rounding.
HMO is health maintenance organization. CHC is community health center.
agreed or strongly agreed. Defensive medicine and malpractice concern subscores are in Appendix Exhibits 2 and 3, available
online as above. **p ≤ 0.05 ***p ≤ 0.01
aPercentage of statements with which physicians
September 201029:9 Health Affairs
1589