Can avoidance of complications lead to biased healthcare decisions?

Judgment and Decision Making, Vol. 1, No. 1, July 2006, pp. 64–75
Can avoidance of complications lead to biased healthcare decisions?
Jennifer Amsterlaw3,5, Brian J. Zikmund-Fisher1,2,3, Angela Fagerlin1,2,3, and Peter A. Ubel?1,2,3,4
1VA Health Services Research & Development Center of Excellence, VA Ann Arbor Healthcare
System, Ann Arbor, MI
2Division of General Internal Medicine, University of Michigan, Ann Arbor, MI
3Center for Behavioral and Decision Sciences in Medicine, Ann Arbor, MI
4Department of Psychology, University of Michigan, Ann Arbor, MI
5Institute for Learning & Brain Sciences, University of Washington, Seattle WA
Abstract
Imagine that you have just received a colon cancer diagnosis and need to choose between two different surgical
treatments. One surgery, the “complicated surgery,” has a lower mortality rate (16% vs. 20%) but compared to the
other surgery, the “uncomplicated surgery,” also carries an additional 1% risk of each of four serious complications:
colostomy, chronic diarrhea, wound infection, or an intermittent bowel obstruction. The complicated surgery dominates
the uncomplicated surgery as long as life with complications is preferred over death.
In our ?rst survey, 51% of a sample (recruited from the cafeteria of a university medical center) selected the domi-
nated alternative, the uncomplicated surgery, justifying this choice by saying that the death risks for the two surgeries
were essentially the same and that the uncomplicated surgery avoided the risk of complications. In follow-up surveys,
preference for the uncomplicated surgery remained relatively consistent (39%–51%) despite (a) presenting the risks in
frequencies rather than percents, (b) grouping the 4 complications into a single category, or (c) giving the uncomplicated
surgery a small chance of complications as well. Even when a pre-decision “focusing exercise” required people to state
directly their preferences between life with each complication versus death, 49% still chose the uncomplicated surgery.
People’s fear of complications leads them to ignore important differences between treatments. This tendency appears
remarkably resistant to debiasing approaches and likely leads patients to make healthcare decisions that are inconsistent
with their own preferences.
Keywords: risk communication, medical decisions, cognitive biases
1 Introduction
erty of performing a mastectomy (Lerner, 2001). These
practices are unheard of today. Patients with cancer di-
Over the past several decades, there has been a revolution
agnoses are told about their diagnoses and are often in-
in healthcare decision making, with much more recogni-
volved in important treatment decisions: deciding for ex-
tion among healthcare practitioners that patients deserve
ample, whether to opt for surgical therapies versus radi-
a role in their healthcare decisions. A few decades ago,
ation. Healthcare practitioners are involving patients in
oncologists frequently withheld cancer diagnoses from
these decisions out of recognition that many of these de-
patients out of fear that patients could not handle this in-
cisions are not purely medical judgments but also include
formation (Novack et al., 1979). It was not uncommon
value judgments that only patients themselves can make
around this time for a woman to wake up from a breast
(Gafni & Whelan, 1998). It is the patient who needs to
biopsy procedure to learn not only that she had breast
consider tradeoffs between the bene?ts of treatment and
cancer, but that the surgeon had already taken the lib-
the potential complications treatments entail. The “right
?During the conduct of this research, Dr. Ubel was recipient
choice” for any speci?c patient therefore often depends
of a Presidential Early Career Award for Scientists and Engineers
on that patient’s preferences or attitudes about possible
(PECASE) and Dr. Zikmund-Fisher was supported by an HSR&D
outcomes.
Post-Doctoral Fellowship from the U.S. Department of Veterans Af-
fairs. This work was also supported by grants from the National In-
At the same time as the medical community has been
stitutes of Health (RO1 CA87595, P50 CA101451, R01 HD40789).
moving toward greater patient involvement in healthcare
Correspondence to: Peter Ubel, MD, University of Michigan, 300
North Ingalls Building, Room 7C27, Ann Arbor, MI 48109–0429,
decisions, decision-making research has identi?ed a host
paubel@med.umich.edu. http://www.cbdsm.org.
of circumstances in which people don’t seem to make the
64

---

Judgment and Decision Making, Vol. 1, No. 1, July 2006
Avoidance of complications
65
right choice. For example, people make different choices
struction, or a colostomy), and is unable to cure 16%
when their options are framed as gains or losses, prefer-
of patients, who therefore die of the cancer. Another
ring a surgical procedure with a 90% survival rate to one
surgery (the “uncomplicated surgery”) cures 80% of pa-
with a 10% mortality rate, even though the two proce-
tients without complications and is unable to cure 20% of
dures are identical (McNeil, Pauker, & Tversky, 1988).
patients, who therefore die of the cancer. In a pilot study,
They prefer different healthcare providers when evaluat-
we determined that the vast majority of people believe
ing each possible doctor separately versus when consid-
that living with each of the four surgical complications
ering the set of possible physicians all at once (Zikmund-
is better than being dead. For those people, the compli-
Fisher, Fagerlin, & Ubel, 2004). Their choices are unduly
cated surgery is the choice that best ?ts these preferences.
in?uenced by uninformative anecdotes (Ubel, Jepson, &
Yet the sheer number and graphicness of these four com-
Baron, 2001). A recurrent theme in this line of research
plications might nevertheless be enough to compel these
is that people’s preferences are often inconsistent or eas-
people to choose the uncomplicated surgery.
ily overridden by subtle cognitive processes (Fischhoff,
Our current research has two goals. In Study 1, we
1991).
tested how people respond in a decision-making situa-
On one side, then, is a push to give patients more in-
tion where their ‘best’ treatment option carries a risk of
formation so they can make decisions that are consis-
several unpleasant complications. Using the colon can-
tent with their personal preferences, while on the other
cer case just described, we asked people which treatment
side is a growing psychological literature revealing peo-
they would choose. We found that many people prefer the
ple’s tendency to make choices that are in fact inconsis-
uncomplicated surgery — that is, they opt for the treat-
tent with their own preferences (Ubel, 2002). These two
ment with a higher risk of death just to avoid the possi-
worlds are in the process of colliding, as clinicians and
bility of complications. In Studies 2 through 4, we var-
researchers become aware of the likelihood that patients,
ied how we presented the scenario information to inves-
even when given comprehensible information about im-
tigate the stability of this choice and to investigate un-
portant healthcare tradeoffs, will make irrational or in-
derlying processes. Our results con?rmed that for most
consistent choices because of the way they process the
people, choice of the uncomplicated surgery was actu-
information they are given.
ally inconsistent with their stated preferences and values.
Take, for example, a common rule of thumb about the
Nevertheless, even when we made people’s own prefer-
kind of information healthcare providers are expected to
ences transparent, many continued to make inconsistent
give patients before enrolling them in research trials or
choices. We conclude that avoidance of treatments car-
before consenting them for invasive surgical procedures.
rying small risks of unpleasant complications leads to in-
Clinicians are expected to inform patients about any treat-
consistencies between stated preferences and people’s de-
ment complication that is reasonably likely to occur. Al-
cisions, and that these inconsistencies are pervasive and
though there is no absolute cutoff for how likely a compli-
dif?cult to eliminate.
cation must be for clinicians to tell patients about it, most
experts feel that clinicians should tell patients about any
moderately severe complication that occurs at least 1% of
2 Study 1
the time, and should inform them about serious complica-
tions that occur even less often than that. This approach to
In Study 1, we presented participants with the colon can-
risk communication is re?ected in the Food and Drug Ad-
cer scenario described above (see Appendix for full text).
ministration’s Guidance for Industry regarding the con-
Accompanying the scenario was a simple table summa-
tent and format of prescription drug labels (“Guidance
rizing the treatments (Table 1).
for industry: adverse reactions section of labeling for hu-
It should be noted: in none of our studies did we refer
man prescription drug and biological products — content
to the treatment options as “the complicated surgery” or
and format,” 2006).
“the uncomplicated surgery,” but instead referred to them
Many clinicians recognize the potential problems cre-
throughout as Surgery 1 and Surgery 2. However, for pur-
ated by this rule of thumb. A long list of potential compli-
poses of presentation, we will refer to them throughout
cations may scare patients away from what otherwise ap-
this manuscript as the complicated and the uncomplicated
pears to be their best treatment option. Imagine a patient
surgery.
with colon cancer who faces a choice between the fol-
The choice set up a tradeoff between mortality risk and
lowing two hypothetical surgical procedures: one surgery
risk of serious surgical complications: the complicated
(the “complicated surgery”) cures 80% of patients with-
surgery had a lower mortality rate (16% vs. 20%) but by
out complications, cures 4% of patients with one of four
the same margin of difference carried an additional 1%
complications (leaving them with either chronic diarrhea,
risk of each of four complications: colostomy, chronic
a slow healing wound infection, an intermittent bowel ob-
diarrhea, wound infection, or intermittent bowel obstruc-

---

Judgment and Decision Making, Vol. 1, No. 1, July 2006
Avoidance of complications
66
Table 1: Treatment summary table presented with scenario in Study 1.
Surgery 1
Surgery 2
Possible outcome
(complicated)
(uncomplicated)
Cure without complication
80%
80%
Cure with colostomy
1%
Cure with chronic diarrhea
1%
Cure with intermittent bowel obstruction
1%
Cure with wound infection
1%
No cure (death)
16%
20%
tion. (Note that the scenario did not explicitly describe
reported being af?liated with the medical/health profes-
the outcomes as mutually exclusive, and in this respect
sion; these individuals did not make signi?cantly differ-
the information is consistent with how side effects and
ent surgery choices (39% vs. 58%, ?2(1) = 3.06, n.s.).
complications often occur.) In pilot testing, we deter-
There were also no differences as a function of prior ex-
mined that the vast majority of people (>90%) thought
perience with colon cancer. While 44% of participants
that each of the four surgical complications was prefer-
reported that they or a close friend or family member had
able to dying of colon cancer. Thus, for most people, the
had colon cancer, these individuals chose the uncompli-
complicated surgery should be preferred to the uncom-
cated surgery at the same rate as other participants (54%
plicated surgery. However, we expected that the desire to
vs. 48%, ?2(1) = .315, n.s.). Respondents who choose
avoid the complications associated with the complicated
the uncomplicated surgery tended to be older than respon-
surgery might lead some of these people to prefer the un-
dents who chose the complicated surgery (M age = 43.3
complicated surgery.
vs. 35.1 years, t(84) = 3.10, p < .01).
Most participants (87%) provided a written explana-
2.1 Methods
tion for their surgery choice. Of these, 94% of partic-
ipants selecting the uncomplicated surgery cited a de-
2.1.1 Participants
sire to avoid complications as the reason for their choice,
with 60% simply reporting that they did not want compli-
Participants were volunteers recruited from a university
cations and 34% also explicitly mentioning the tradeoff
of?ce building and the cafeteria of a university medical
with mortality risk. Two additional people mistakenly be-
center. A total of 87 participants completed the ques-
lieved the uncomplicated surgery improved their chances
tionnaire. The sample was 91% Caucasian and 65%
of survival. In contrast, 100% of participants selecting
female. The mean age of participants was 39.2 years
the complicated surgery cited its higher survival rate as
(SD = 12.9) and 43% of participants had completed at
the reason for their choice, with 68% simply mentioning
least a college degree.
its better survival rate and 32% also describing the trade-
off between death and complications.
2.1.2 Procedure
Participants received the scenario as part of a written
2.3 Discussion
questionnaire that also contained other questions about
unrelated medical decision-making topics. Demographic
Results from Study 1 showed that regardless of back-
questions about participants’ age, race, education, pro-
ground factors such as gender, race, education, medi-
fession, and personal experience with colon cancer were
cal af?liation, or previous experience with colon cancer,
also included. Participants completed the questionnaire
many people preferred the uncomplicated surgery to the
at their own pace.
complicated surgery. Written responses indicated that
this preference stemmed from a desire to avoid compli-
2.2 Results
cations. Most participants appeared to have understood
the information presented, with well over half the sample
Out of 87 participants, 44 (about 51%) selected the un-
explicitly acknowledging the tradeoff between mortality
complicated surgery, the dominated alternative. Anal-
risk and risk of complications. However, those who chose
yses showed no signi?cant effects of gender, race, or
the uncomplicated surgery were unwilling to risk the pos-
education level on surgery choice. 42% of our sample
sibility of serious complications to improve their overall

---

Judgment and Decision Making, Vol. 1, No. 1, July 2006
Avoidance of complications
67
Table 2: Treatment summary table presented with Study 2 “Complications Added” version.
Surgery 1
Surgery 2
Possible outcome
(complicated)
(uncomplicated)
Cure without complication
80%
80%
Cure with colostomy
1%
0.25%
Cure with chronic diarrhea
1%
0.25%
Cure with intermittent bowel obstruction
1%
0.25%
Cure with wound infection
1%
0.25%
No cure (death)
16%
19%
chance of survival. Why would this be the case? Bar-
feel worse to be bombed nightly than to be bombed less
ring the possibility that people actually would prefer to
often. However, people in the suburbs were signi?cantly
die than live with complications (a possibility we exam-
more likely to develop stomach ulcers than city dwellers,
ine in Study 4), selecting the treatment with the higher
because they were so stressed out by the uncertainty of
mortality risk seems irrational. We next explored reasons
when they would be exposed to bombing raids (Frederick
for this apparent inconsistency.
& Loewenstein, 1999).
We hypothesized that the uncomplicated surgery, de-
3 Study 2
spite having a higher death rate than the complicated
surgery, would feel less uncertain to subjects, and there-
Study 2 investigated the possibility that choice was af-
fore more desirable. To test this, we developed a new ver-
fected by the number of possible outcomes listed for
sion of the scenario in which the uncomplicated surgery
each treatment option. One possibility was that people
was now described as carrying a small risk of complica-
were attracted to the uncomplicated surgery because it
tions (Table 2). In this “Complications Added” version,
had only two outcomes associated with it — total cure
the uncomplicated surgery now carried a 0.25% risk of
and death — while the complicated surgery had six possi-
each of the four complications described for the com-
ble outcomes, four of which were ambiguous “cured with
plicated surgery. To compensate for this change while
complications” outcomes. People may simply have been
preserving the same basic probability information, we re-
averse to this kind of uncertainty, leading them to choose
duced the death rate for the uncomplicated surgery from
the surgery with fewer and less ambiguous possible out-
20% to 19%. If people simply preferred the uncompli-
comes.
cated surgery because it had less ambiguous outcomes,
People are notoriously averse to uncertainty. For ex-
they should now prefer it less (even though the decrease
ample, in the 1980s, back when HIV infection was a
in the death rate actually makes this option more attrac-
death sentence, studies showed that men were happier af-
tive). With both treatments now carrying the potential for
ter they received HIV test results than while they were
the same set of unpleasant outcomes, people should be
waiting for the results no matter what their test results
more inclined to choose the option that maximizes their
revealed! The uncertainty of not knowing their HIV sta-
survival.
tus was harder to cope with than the certainty of a rapid
The “Complications Added” version also tests an
demise (Sieff, Dawes, & Loewenstein, 1999). The same
explanation for choosing the uncomplicated surgery
phenomenon was demonstrated among people undergo-
grounded in the non-linear probability weighting function
ing genetic testing for Huntington’s Disease, a devastat-
of Kahneman and Tversky’s Prospect Theory (Kahneman
ing, hereditary neurologic illness that causes uncontrolled
& Tversky, 1979). Prospect Theory holds that very small
spasms, dementia and death (Wiggins et al., 1992). Ra-
probability events are overweighted in decisions. As a
tionally speaking, it should be easier to live with a 50%
result, the change in the probability of a complication
chance of Huntington’s Disease than a 100% chance. But
(e.g., colostomy) from 0% to 1% may in?uence choice far
it is dif?cult for people to cope with the uncertainty of
more than the equivalent 1% reduction in the death rate
a 50% chance of illness. Uncertainty is so stressful that
from 20% to 19%. However, the very small (0.25%) risks
it can create paradoxical situations. For example, during
of each complication added to the uncomplicated surgery
World War II, people living in London were deluged by
in the “Complications Added” version should be simi-
nightly bombing raids, while those living in the suburbs
larly overweighted, and thus Prospect Theory would pre-
were raided sporadically. Objectively speaking, it should
dict that people should shift their preferences towards the

---

Judgment and Decision Making, Vol. 1, No. 1, July 2006
Avoidance of complications
68
Table 3: Treatment summary table presented with Study 2 “Grouped Complications” version.
Surgery 1
Surgery 2
Possible outcome
(complicated)
(uncomplicated)
Cured without complications
80%
80%
Cured, but with one of the following complications: colostomy, chronic
4%
diarrhea, intermittent bowel obstruction, or wound infection
No cure (death)
16%
20%
survival maximizing outcome, the complicated surgery,
identical result when compared to Study 1, ?2(1) = 0.01,
when compared to the results from Study 1.
N = 167, n.s.). In the Grouped Complications version,
Another possible source of the inconsistency was that
40 out of 100 participants (40%) selected the uncom-
people were relying on a simple tallying strategy to de-
plicated surgery, which also did not differ signi?cantly
cide on the best option, for example by counting up the
from the proportion obtained with the original scenario,
treatments’ “wins” and “losses” in each outcome cate-
?2(1) = 2.10, N = 187, n.s.) Thus, preference for
gory. In that case, the complicated surgery may have
the uncomplicated surgery did not depend simply on the
looked like a bad option because it had ?ve “losses” to
number of outcomes possible for each treatment or on
the uncomplicated surgery (for each of the four compli-
whether the complications were presented as four sepa-
cations outcomes and the death outcome) and only one
rate outcomes or as one.
“win” (for death rate). One potential way to increase pref-
Participants’ explanations for their choice of the un-
erence for the complicated surgery, then, would be to re-
complicated surgery again re?ected a desire to avoid
duce its “losses” to the uncomplicated surgery by present-
complications. Across both versions, about 80% of par-
ing the complications outcomes as a single outcome with
ticipants provided explanations for their choice. Of these,
a 4% risk, rather than as four separate outcomes each with
23 out of 32 participants (72%) who selected the uncom-
a 1% risk. This is what we did in the “Grouped Compli-
plicated surgery in the Complications Added version said
cations” version of the scenario (Table 3). We predicted
they wanted to avoid complications. Two mistakenly re-
that if a tallying strategy was in use, this change would
ported that the survival rate was better for the uncom-
increase preference for the complicated surgery.
plicated surgery, while seven (22%) gave a non-speci?c
explanation, such as “so I’ll have a better chance.” In
3.1 Methods
the Grouped Complications version, 26 out of 30 (87%)
explained their choice of the uncomplicated surgery by
3.1.1 Participants
saying they wanted to avoid complications. Three gave
non-speci?c “better chance” explanations and one gave
Participants were volunteers recruited from the cafete-
an ambiguous response.
ria of a university medical center, two university of?ce
Upon closer examination, three main types of expla-
buildings, and a local shopping center. A total of 80
nations that expressed a desire to avoid complications
participants completed the Complications Added version,
emerged. Some people simply said they wished to avoid
and 100 participants completed the Grouped Complica-
complications; some referred to a tradeoff, saying that
tions version. The sample was 84% Caucasian and 57%
they realized that the uncomplicated surgery’s death rate
female. The mean age of participants was 40.6 years
was higher but were willing to take that risk to be free of
(SD = 16.3), and 47% of participants had completed
complications; and some in effect bypassed the tradeoff
a college degree.
by stating that the difference between the surgeries’ mor-
tality rates was too small to be meaningful. Across both
3.1.2 Procedure
versions of the scenario, responses of these types consti-
tuted 24%, 21%, and 16% of all explanations for choice
The procedure was identical to that used in Study 1. The
of the uncomplicated surgery, respectively.
only difference was that the probability information was
changed, as described above.
3.2 Discussion
3.1.3 Results
Even when signi?cant changes in the scenario were
In the Complications Added version, 41 out of 80 partici-
made, a signi?cant minority of people maintained pref-
pants (about 51%) selected the uncomplicated surgery, an
erence for the uncomplicated surgery. This leads us re-

---

Judgment and Decision Making, Vol. 1, No. 1, July 2006
Avoidance of complications
69
Table 4: Treatment summary table presented with Study 3 “Reframing” version.
Surgery 1
Surgery 2
Possible outcome
(complicated)
(uncomplicated)
Cured of colon cancer 840
800
800 cured without complications
800 cured without complications
40 cured with one of the following
0 cured with one of the following
complications:
complications:
• colostomy
• colostomy
• chronic diarrhea
• chronic diarrhea
• intermittent bowel obstruction
• intermittent bowel obstruction
• wound infection
• wound infection
No cure (death)
160
200
ject to two plausible hypotheses about the predominant
this underlines the point that the risk of unpleasant com-
source of the bias. In general, people did not avoid
plications — even when small – looms inordinately large
the complicated surgery simply because of the uncer-
in people’s decision-making in a way not captured by lin-
tainty associated with its multiple and ambiguous op-
ear probability weighting.
tions, nor were they relying on a simple tallying strategy
that was overwhelmed by these multiple possible side ef-
fects. Written explanations for choosing the uncompli-
4 Study 3
cated surgery again indicated strong aversion to compli-
cations. A closer analysis of these explanations yielded
When people ignore important probability differences in
two discernable subtypes of responses: A small number
mortality risk across options, they effectively bypass the
of people appeared to prefer death rather than risk life
tradeoff that is inherent in the choice. When the cure rates
with complications, and a larger number believed the dif-
and death rates are both seen as equivalent across surg-
ference in mortality rates was too small to be signi?cant,
eries, only complication rates remain to distinguish be-
leaving complication rate as the deciding factor.
tween the two treatments, and the uncomplicated surgery
is a clear winner. The goal of Study 3 was to explore
People’s tendency to equate the 16% and 20% the mor-
ways of heightening people’s sensitivity to meaningful
tality risks is problematic. Why should people view a
probability differences across options and to the neces-
4% difference in complications rates as signi?cant, but
sary tradeoff between mortality risk and risk of compli-
dismiss an equally-sized difference in death rates? The
cations. We created two new versions of the scenario with
asymmetry is perhaps most obvious in the Complications
this in mind.
Added case, where both surgeries have the same set of
In the ?rst of these, the “Explicit Tradeoff” version
possible outcomes. In that case, people appeared to view
(Table 4), we made two major changes. First, we pre-
a 3% difference in the total complications rate as signif-
sented the outcome information in terms of frequencies
icant, but not a 3% difference in death rate, even though
rather than percents. To emphasize the additional 4% of
both surgery options included small risks of complica-
people whose lives could potentially be saved by the com-
tions. The ?nding is consistent with prior research in psy-
plicated surgery, we chose to present the information in
chology (Baron, 1997; Fetherstonhaugh, Slovic, Jonhn-
terms of the likely outcomes for 1000 people undergoing
son, & Friedrich, 1997; Jenni & Loewenstein, 1997) and
each treatment. If participants could see that 40 addi-
medicine (Bobbio, Demichelis, & Giustetto, 1994; For-
tional people would be saved by the complicated surgery,
row, Taylor, & Arnold, 1992; Malenka, Baron, Johansen,
this might reduce their tendency to dismiss the mortal-
Wahrenberger, & Ross, 1993) showing that people often
ity rates as equivalent. Second, we divided the death
think about risks in relative, rather than absolute terms.
outcome for the uncomplicated surgery into two separate
Speci?cally, equal sized changes in risk may be per-
outcomes — death from colon cancer (the same as in all
ceived as greater when they represent a larger fraction
previous scenarios) and death from “scar tissue in?am-
of the baseline risk level. In our scenario, people may
mation,” a new fatal complication of the uncomplicated
have focused on the 1% to 4% increase in complications
surgery. The probability of death from scar tissue in?am-
risk from the uncomplicated surgery to the complicated
mation under the uncomplicated surgery (4%) was pre-
surgery because it represents a greater (relative) change
cisely equal to the probability of being cured with compli-
in risk than the 19% to 16% decrease in mortality. Again,
cations under the complicated surgery. With this change,

---

Judgment and Decision Making, Vol. 1, No. 1, July 2006
Avoidance of complications
70
Table 5: Treatment summary table presented with Study 3 “Explicit Tradeoff” version.
Surgery 1
Surgery 2
Possible outcome
(complicated)
(uncomplicated)
Cured without complications
800
800
Cured, but with one of the following complications: colostomy, chronic
40
0
diarrhea, intermittent bowel obstruction, or wound infection
No cure, death from scar tissue in?ammation within 2 years
0
40
No cure, death from colon cancer within 2 years
160
160
we hoped people would now more clearly see the trade-
the uncomplicated surgery, again not signi?cantly differ-
off they had to make — either incur the risk of surviving
ent from the original scenario, ?2(1) = 2.52, N = 175,
with complications or incur the risk of dying from one.
n.s.).
In the second version, the “Reframing” version (Table
5), we again presented outcome information in terms of
4.3 Discussion
frequencies rather than percents. In addition, we changed
how the information about complications outcomes was
Although participants in both versions used in Study 3
presented. In previous versions, we had always presented
tended to pick the uncomplicated surgery slightly less of-
“cured with complications” as an outcome distinct from
ten than in the Study 1 base case, in neither version did
being cured without complications. In the Reframing ver-
we observe a signi?cant difference in behavior. Neither
sion, we now presented it as a subset of the larger group
making the tradeoff between the risks of death and life
of people cured of their colon cancer, thereby giving that
with complications explicit, nor reframing life with com-
outcome a more positive spin. We hypothesized that this
plications as a subset of the cured population, was effec-
approach would increase preference for the complicated
tive in encouraging substantially more study participants
surgery by making the complications outcomes seem less
to select the option that maximized survival. Of note: in
negative and also by highlighting its higher cure rate.
determining the sample sizes for our studies, we set out to
?nd large differences between versions, looking for phe-
nomena that explained all or most of the bias. Since our
4.1 Methods
manipulations of both the risk statistics and the format of
4.1.1 Participants
their presentation yielded no dramatic behavior changes,
we next considered the possibility that the observed se-
Participants were recruited from the cafeteria of a uni-
lection of the uncomplicated surgery was actually a true
versity medical center, a local bus station, and the lo-
re?ection of people’s preferences between life with com-
cal public library. A total of 76 participants completed
plications and death.
the Explicit Tradeoff version of the questionnaire, and 88
completed the Reframing version. The sample was 67%
Caucasian and 54% female. The mean age was 35.3 years
5 Study 4
(SD = 15.0), and 52% had completed a college degree.
Selecting the surgery with the higher mortality risk just to
avoid possible side effects seems irrational. But is it? The
4.1.2 Procedure
18th Century philosopher, David Hume, said, “It is not
contrary to reason to prefer the destruction of the whole
The procedure was identical to earlier versions. Probabil-
world to the scratching of my little ?nger.” That is, the ra-
ity information in the scenario was changed as described
tionality of a given choice depends on whether that choice
above.
is consistent with one’s goals and values. The uncompli-
cated surgery is a bad choice only if people would prefer
4.2 Results
to live with complications rather than die. If people actu-
ally would rather die than live with any of the complica-
For the Explicit Tradeoff version, 31 out of 78 partici-
tions, we cannot call them irrational for choosing the un-
pants (40%) chose the uncomplicated surgery. This dif-
complicated surgery. On the other hand, if a person thinks
ference was not signi?cant when compared to results of
that living with complications is better than being dead,
Study 1, ?2(1) = 1.95, N = 165, n.s.). For the Re-
then he should prefer the complicated surgery. If such a
framing version, 34 out of 88 participants (39%) chose
person nevertheless chooses the uncomplicated surgery,

---

Judgment and Decision Making, Vol. 1, No. 1, July 2006
Avoidance of complications
71
he has made an irrational choice — a choice inconsis-
the opportunity to change their original treatment choice
tent with his own preferences. In Study 4, we investi-
if they desired.
gated whether preference for the uncomplicated surgery
is in fact an irrational choice, or whether it simply re-
?ects people’s underlying beliefs about the value of life
with complications relative to death. We did so by hav-
5.2 Results
ing each subject complete a rating exercise in which they
directly compared life with each of the four possible com-
plications versus death.
Participants’ responses on the rating exercise revealed
A second goal of Study 4 was to examine whether
an overwhelming preference for life with complications
having participants make such ratings would affect their
over death. Looking just at the ratings of participants in
surgery choices. Thus we compared the surgery choices
the Rate-Before-Choice condition, 71 out of 77 partic-
of people who performed the rating exercise before ver-
ipants (92%) indicated that life with complications was
sus after making their surgery choice. If people are
preferable to death for all four complications listed. On
choosing the uncomplicated surgery because they pre-
the opposite extreme, only one participant indicated a
fer death over life with complications, then undertaking
preference for death in all four cases. Five participants
the rating exercise before making a choice should not
(7%) preferred death over life with complications in some
in?uence people’s decisions. However, if participants
cases but not others, with three people reporting that they
do prefer to live with complications rather than die –
would rather die than live with a colostomy bag and two
but nonetheless make treatment choices inconsistent with
reporting that they would rather die than live with inter-
these preferences — expressing a preference for life in
mittent bowel obstruction. Preference for life with com-
a pre-decision rating exercise could focus their decision-
plications was somewhat lower in the Rate-After-Choice
making around those priorities, leading to greater prefer-
condition, suggesting that the choice activity affected par-
ence for the complicated surgery.
ticipants’ rating responses. However, 61% of participants
in the Rate-After-Choice condition still indicated that life
with complications was preferable to death for all four
5.1 Methods
complications listed. Roughly 7% believed death was
5.1.1 Participants
preferable in all four cases, with the remaining 32% re-
porting a preference for complications over death only in
Participants were recruited from the cafeteria of a univer-
some cases.
sity medical center. A total of 154 participants completed
In many cases, participants’ surgery choices were not
the questionnaire, with half receiving the “Rate-Before-
consistent with these preferences. In the Rate-Before-
Choice” version of the questionnaire and half receiving
Choice condition (in which 92% of participants indicated
the “Rate-After-Choice” version. The sample was 89%
that they preferred life with complications over death in
Caucasian and 62% female. The mean age of participants
all cases), 49% of participants went on to select the un-
was 42.2 years (SD = 16.5), and 52% of participants had
complicated surgery, thus selecting a surgery which did
completed a college degree. Participants’ demographic
not re?ect their stated beliefs. In the Rate-After-Choice
characteristics (gender, race, age, education level) did not
condition, 45% selected the uncomplicated surgery even
differ across the two versions of the questionnaire.
though 60% then indicated that they preferred complica-
tions over death in all cases. Stated another way, 49%
5.1.2 Procedure
of people who chose the uncomplicated surgery in the
Participants were presented with the same basic scenario
Rate-Before-Choice and 32% of people who chose the
used in Study 1. In addition, either before or after their
uncomplicated surgery in the Rate-After-Choice condi-
choice, participants received the rating exercise. This
tion made choices inconsistent with their stated prefer-
consisted of four questions that asked “What would be
ences. Further, when participants were given the chance
better, being dead or living with
?” for each of the pos-
to change their initial choice in the Rate-After-Choice
sible complications: colostomy, chronic diarrhea, inter-
condition, few did so. Only three switched from the un-
mittent bowel obstruction, and wound infection. Partici-
complicated surgery to the complicated surgery, and four
pants could respond by selecting either “death would be
actually switched in the opposite direction.
better” or “living with
would be better.” Descriptions
Thus, when people were asked directly about whether
of the four complications (identical to the descriptions
they thought it was better to be dead or to live with com-
provided in the original scenario) accompanied the rat-
plications, most said that they would rather live with com-
ing exercise. In addition, in the Rate-After-Choice condi-
plications than die. However, these preferences were fre-
tion, the last page of the questionnaire gave participants
quently not re?ected in their surgery choices.

---

Judgment and Decision Making, Vol. 1, No. 1, July 2006
Avoidance of complications
72
5.3 Discussion
complicated surgery rates (Study 2: “Grouped Complica-
tions” and the two versions of Study 3 — see Table 6) all
These results demonstrate that many people who chose
grouped the risk of the four possible complications into a
the uncomplicated surgery chose it despite a clear pref-
single category. This suggests that some fraction of peo-
erence for life with complications over death. That is,
ple choosing the uncomplicated surgery were in?uenced
their choices contradict their own preferences. The re-
by the sheer number of categories under consideration.
sults from the Rate-Before-Choice condition are espe-
Still, we note that our studies were powered to detect rel-
cially striking: People’s choices were inconsistent with
atively large effects — that is, relatively “common” pat-
their preferences even though they had expressed those
terns of choice — and in none of these versions did the
preferences only minutes before.
percentage of people picking the uncomplicated surgery
ever drop below 39% of the sample, so the impact of this
6 General Discussion
issue is moderate at best. Most likely, a variety of factors
contribute to inconsistent decision-making in these cases.
The goal of informed consent discussions in healthcare
We hope to clarify these further in our future research, by
practice is to help patients decide which alternatives best
conducting larger studies with the power to reveal subtle
?t their individual preferences. A choice between two
processes contributing to — if not completely explaining
surgical operations may hinge, for example, on how par-
— the effect, and by conducting studies that explicitly
ticular patients weigh the relevant risks and bene?ts of
assess the role of affect in people’s decisions.
the two procedures. The same kind of weighing, of pros
Nevertheless, it is worth speculating about potential
and cons, helps people make all kinds of healthcare deci-
causes of this bias. Our initial intuition was that the sheer
sions, such as whether to enter clinical trials, or whether
number of complications of the complicated surgery was
to undergo risky treatments for serious illnesses. And be-
the source of the bias. However, in the “Complications
cause people’s attitudes toward risks and bene?ts differ,
Added” survey from Study 2, both of the surgical alter-
the right choice for any one person will depend on his/her
natives have the same number of complications, yet 51%
values. Consequently, experts contend that patients de-
of people still chose the dominated alternative. This sub-
serve to receive comprehensible medical information and
study proves that it is not simply the number of compli-
the freedom to choose among available alternatives.
cations that leads to the bias.
Our study reveals one problem with the way informed
We also predicted that the number of complications,
consent is currently obtained. As our study shows, when
and the graphic way each complication was described,
people receive comprehensible information about their
might make it dif?cult for people to perceive the dom-
treatment alternatives, they do not always make choices
inance relationship. Perhaps people were just unaware
that ?t their own preferences. This in itself is not a new
that one surgery dominates the other. We no longer be-
?nding, as people have been shown to be susceptible to
lieve this explanation, however. To begin with, neither
a whole host of biases when making healthcare deci-
education nor measures of people’s mathematical ability
sions (Redelmeier, Rozin, & Kahneman, 1993). How-
were signi?cantly associated with treatment choice in any
ever, our study is signi?cant for two important reasons.
of our studies. Moreover, we presented this scenario to
First, we have demonstrated a decision-making inconsis-
a random sample of 119 primary care physicians in the
tency that is particularly relevant to healthcare decisions:
U.S., and asked them what they would choose for them-
lists of graphic complications can drive people away from
selves. Almost 40% chose the uncomplicated surgery,
treatments, even when the same people acknowledge that
suggesting that neither medical training nor relevant deci-
these treatments are preferable to other alternatives in
sion experiences prevent biased choices. Finally, in Study
terms of expected outcomes. Second, we have shown
4, we asked people to state whether they preferred death
just how persistent this inconsistency can be. Even when
or any of the four surgical complications prior to asking
people’s preferences are completely transparent — even
them to choose between the two surgical alternatives. We
after people have seen the inconsistencies of their own
believe that this method should have clari?ed the domi-
views — people still make choices that don’t map onto
nance relationship, and yet many study participants were
their own preferences. It is as if many of our subjects told
still willing to choose the dominated alternative.
us “the complicated surgery is better than the uncompli-
We are now in the process of exploring other mecha-
cated surgery, but I prefer to receive the uncomplicated
nisms that could explain the source of the bias we have
surgery.”
demonstrated in this paper. One possibility is that the
Although none of our manipulations signi?cantly re-
bias results from processes similar to those documented
duced the percentage of participants selecting the uncom-
in studies of betrayal aversion. Research shows that peo-
plicated surgery from that observed in the base case, the
ple are bothered by bad outcomes when the cause of those
three versions yielding the lowest preference for the un-
outcomes is perceived as some kind of betrayal. For ex-

---

Judgment and Decision Making, Vol. 1, No. 1, July 2006
Avoidance of complications
73
Table 6: Participants’ surgery choices across all scenario versions. Statistics compare proportions against those from
Study 1.
Percent selecting
N dominated alternative ?2
p
Study 1
Original scenario
87
51%
Study 2
Complications added to uncomplicated surgery
80
51%
0.01 .931
Grouped complications for complicated surgery 100
40%
2.10 .147
Study 3
Explicit tradeoff
78
40%
1.95 .163
Reframing
88
39%
2.52 .112
Study 4
Focusing rating before choice
77
49%
0.00 .992
ample, people believe it is worse to be killed in an auto
Hsee, & Welch, 2001). Along these lines, people may
accident by a faulty airbag than by other malfunctions in
know that the complicated surgery is better than the un-
a car, because the airbag is supposed to protect people
complicated surgery, but it might feel like the uncompli-
from injury. (Koehler & Gershoff, 2003) Medical inter-
cated surgery is a better option. We plan a series of follow
ventions are supposed to improve people’s health, and so
up studies, in which we will try to in?uence the emotional
the fact that the complicated surgery might cause other
salience of the four surgical complications, as well as the
(lesser) harms might be perceived as kind of betrayal, re-
emotional salience of death, to see how that in?uences
sulting in aversion to that choice. Similar tendencies may
people’s choices.
contribute to well-known omission biases. For example,
Is consistency something people should strive for when
people are reluctant to get vaccinated if the vaccine car-
making important decisions? We think so. Consistency
ries a risk of health side effects, even if the risk/bene?t
is hardly the hobgoblin it is often made out to be. When
pro?le of the vaccine is better than the risk/bene?t pro?le
confronted with inconsistencies, most people, in most cir-
of remaining unvaccinated (Ritov & Baron, 1990).
cumstances, do not merely shrug off the inconsistency.
The bias may also result from the affective salience of
They try to understand why they have made an incon-
the surgical complications. Intermittent diarrhea may not
sistent choice, or discover some consistency lying under-
have much impact on people’s quality of life, and may not
neath the surface of their choice. The huge ?eld of re-
come close to being perceived as being as bad a death.
search on cognitive dissonance is a testament to the im-
But diarrhea is icky. So is the thought of a colostomy
portance people place on achieving some type of internal
or a wound infection. People are much more sensitive
consistency in their lives. Of course, consistency is not
to the probability of emotionally mundane events than
always desirable, nor is inconsistency always troubling.
to more emotionally salient events, when making deci-
People change their minds over time, for example, and
sions (Rottenstreich & Hsee, 2001). As a result, because
such inconsistencies, if they can be even called that, can
the complicated surgery includes a risk of four affectively
be a sign of growth or open-mindedness.
loaded complications, people may feel strong aversion to
Yet in this article we are not exploring reasonable in-
that surgery despite the low probability of each complica-
consistencies occurring over the course of people’s lives.
tion, and despite the fact the complications are preferable
Instead, across the span of two minutes, after earlier
to the alternative of being dead. By contrast, our simple
stating a preference for the complications over death,
description of the risk of death may lack the emotional
many people had no problem choosing the uncomplicated
salience of the graphically described complications. In-
surgery over the complicated surgery. This is a dilemma.
deed, a number of decision making theories postulate that
If the complicated surgery is better than the uncompli-
anticipatory emotions play a large role in people’s deci-
cated surgery, then people should choose it, and if it is
sions, and can skew the relationship between probabil-
not, then their preferences should re?ect this view. We
ity, utility, and decisions (Damasio, 1994; Finucane, Al-
have shown that even when people receive easily com-
hakami, Slovic, & Johnson, 2000; Loewenstein, Weber,
prehensible information, and when their own preferences

---