Content uploaded by Ziv Carmon
Author content
All content in this area was uploaded by Ziv Carmon on Apr 06, 2015
Content may be subject to copyright.
In response to Dr Browning, I did not lay the problems
of general surgery “at the feet of primary care.” Having been
involved in an effort to undo the flawed sustainable growth
rate system for the past decade, I know that the system is
broken. In a rational world, specialties would unite to cor-
rect this flawed system, which seems cynically designed to
pit specialties against each other.
The calculations for physician compensation used in
the cited study of internal medicine reimbursement
3
cut
off in 2004 while the increases in evaluation and manage-
ment codes took place in 2004 through 2007. For
example, evaluation and management codes increased in
work relative value units from 2006 to 2007 (code 99213:
increase from 0.67 to 0.92 [37%]; code 99214: 1.10 to
1.42 [29%]; and code 99215: 1.77 to 2.00 [12.9%]). Fur-
thermore, this article has been rebutted.
4
The primary
care–specialty income gap has largely been corrected, at
an estimated cost of $4 billion.
5
I appreciate the enthusiasm of Dr Maa and colleagues
for the surgical hospitalist, particularly in emergency and
trauma care. However, I do not believe this is the answer
to the shortage of general surgeons, who not only do
emergency and trauma care but often perform endoscopy
and other general surgical operations in critical access
hospitals.
Philosophically, I have difficulty with medical or surgi-
cal hospitalists. My objection to the medical hospitalist
system is that patients may perceive that they are being
abandoned by their primary care physician at a time of
their most dire need—when they are sufficiently ill to
require hospitalization. I would hope that this sense of
abandonment is not present in a surgical hospitalist sys-
tem. When we train “physicians who operate,” we try to
instill judgment about the need for surgery. But if opera-
tion is required, the operator must be thoroughly compe-
tent. I believe that such decisions are more informed
when the physician/surgeon is familiar with the patient.
The surgical hospitalist system may aid trauma and emer-
gency care in urban centers but will not solve the access
problem in small rural hospitals.
Josef E. Fischer, MD
jfische1@bidmc.harvard.edu
Department of Surgery
Harvard Medical School
Boston, Massachusetts
Financial Disclosures: None reported.
1. Zuckerman R, Doty B, Gold M, et al. General surgery programs in small rural
New York state hospitals. J Rural Health. 2006;22(4):339-342.
2. Smart DR, ed. Physician Characteristics and Distribution in the US, 2007. Chi-
cago, IL: American Medical Association; 2007.
3. Bodenheimer T, Berenson R, Rudolf P. The primary care-specialty income gap:
why it matters. Ann Intern Med. 2007;146(4):301-306.
4. Rich WL III. The primary care-specialty income gap. Ann Intern Med. 2007;
146(12):895-896.
5. Centers for Medicare & Medicaid Services (CMS). Medicare program: revi-
sions to payment policies, five-year review of work relative value units, changes
to the practice expense methodology under the physician fee schedule, and other
changes to payment under part B. Fed Regist. 2006;71(231):69623-70251.
RESEARCH LETTER
Commercial Features of Placebo
and Therapeutic Efficacy
To the Editor: It is possible that the therapeutic efficacy of
medications is affected by commercial features such as lower
prices. Because such features influence patients’ expecta-
tions,
1
they may play an unrecognized therapeutic role by
influencing the efficacy of medical therapies, especially in
conditions associated with strong placebo responses.
2,3
To
investigate this possibility, we studied the effect of price on
analgesic response to placebo pills.
Methods. In 2006 we recruited 82 healthy paid volunteers
in Boston, Massachusetts, using an online advertisement. Each
participant was informed by brochur eabout a (purported) new
opioid analgesic approved by the Food and Drug Administra-
tion; it was described as similar to codeine with faster onset time,
but it was actually a placebo pill. After randomization, half of
the participants were informed that the drug had a regularprice
of $2.50 per pill and half that the price had been discounted to
$0.10 per pill (no reason for the discount was mentioned). All
participants received identical placebo pills and were paid $30.
Participants were blinded to the study purpose, and research-
ers were blinded to groupassignment. The study was approved
by the Massachusetts Institute of Technology institutional re-
view board, and all participants providedwritten informed con-
sent and were debriefed after the study.
The protocol followed an established approach for studying
pain.
4
Electrical shocks to the wrist were calibrated to each par -
ticipant’s pain tolerance. After calibration, participants received
the test shocks, rating the pain on a computerized visual ana-
log scale anchored by the labels “no pain at all” and “the worst
pain imaginable.” Participants received all possible shocks in
2.5-V increments between 0 V and their calibrated tolerance.
Stimulation at each intensity level was carried out twice for each
participant (before and after taking the pill), and the change
in reaction to the stimulation was assessed. V isual analog scale
ratings were converted to a 100-point scale, the postpill score
for each voltage was subtracted from the prepill score, and the
mean of these differences was calculated for each participant.
The per centage of participants experiencing a mean score re-
duction vs increase was compared between the 2 groups using
a 2-tailed
2
test. Because stronger pain may be associated with
stronger placebo responses,
5
we also compared r esults for the
50% most painful shocks for each participant. In addition, mean
differ ences at each voltage between the 2 groupswer e compar ed
overall with a sign test and individually with F tests. A P value
of .05 was considered statistically significant. Analyses were per-
formed using SPSS version 15 (SPSS Inc, Chicago, Illinois).
Results. Patient characteristics are shown in the T
ABLE.In
the regular-price group, 85.4% (95% confidence interval [CI],
74.6%-96.2%) of the participants experienced a mean pain re-
duction after taking the pill, vs 61.0% (95% CI, 46.1%-75.9%)
in the low-price (discounted) group (P =.02). Similar results
LETTERS
1016 JAMA, March 5, 2008—Vol 299, No. 9 (Reprinted) ©2008 American Medical Association. All rights reserved.
at INSEAD, on March 4, 2008 www.jama.comDownloaded from
occurredwhen analyzing only the 50% most painful shocks for
each participant (80.5% [95% CI, 68.3%-92.6%] vs 56.1% [95%
CI, 40.9%-71.3%], respectively; P =.03).
Considering all voltages tested, pain reduction was greater
for the regular-price pill (P ⬍ .001). In addition, for 26 of
29 intensities (from 10 to 80 V), mean pain reduction was
greater for the regular-price pill (F
IGURE).
Comment. These results are consistent with described phe-
nomena of commercialvariables affectingquality expectations
1
and expectations influencing therapeutic efficacy.
4
Placebo re-
sponses to commer cial featureshavemany potential clinical im-
plications. For example, they may help explain the popularity
of high-cost medical therapies (eg, cyclooxygenase 2 inhibitors)
over inexpensive, widely available alternatives (eg, over-the-
counter nonsteroidalanti-inflammatory drugs) and why patients
switchingfr om branded medicationsmayr eportthattheir generic
equivalents are less effective. Studies of real-worldeffectiveness
may be more generalizable if they reflect how medications are
sold in addition to how they are formulated. Furthermore, cli-
nicians may be able to harness quality cues in beneficial ways,
6
for example, by de-emphasizing potentially deleterious com-
mercial factors (eg, low-priced, generic).
These findings need to be replicated in broader popula-
tions and clinical settings to better understand how commu-
nicating quality cues with patient populations can maximize
treatment benefits and patient satisfaction.
Rebecca L. Waber, BS
Massachusetts Institute of Technology
Cambridge, Massachusetts
Baba Shiv, PhD
Stanford University
Stanford, California
Ziv Carmon, PhD
INSEAD
Singapore
Dan Ariely, PhD
ariely@mit.edu
Massachusetts Institute of Technology
Author Contributions: Dr Ariely had full access to all of the data in the study and
takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Waber, Shiv, Carmon, Ariely.
Acquisition of data: Waber.
Analysis and interpretation of data: Waber, Ariely.
Drafting of the manuscript: Waber, Shiv, Ariely.
Critical revision of the manuscript for important intellectual content: Waber, Shiv,
Carmon, Ariely.
Statistical analysis: Waber, Ariely.
Obtained funding: Ariely.
Administrative, technical, or material support: Waber.
Study supervision: Ariely.
Financial Disclosures: None reported.
Funding/Support: This study was funded by the Massachusetts Instituteof Technology.
Role of the Sponsor: The sponsor had no role in the design or conduct of the study;
the collection, management, analysis, or interpretation of data; or the preparation,
review, or approval of the manuscript.
Additional Contributions: Taya Leary, MS, Tom Pernikoff, BS, and John Keefe, BS, all
with the Massachusetts Institute of Technology at the time of this study, provided as-
sistance in datacollection. Mr Keefe receivedcompensation for thisrole. Andrew Lippman,
PhD, Massachusetts Instituteof Technology, provided logisticalsupport and MarkVangel,
PhD, Massachusetts General Hospital, provided statistical assistance. Neither received
compensation for these roles.
1. Rao AR, Monroe KB. The effect of price, brand name, and store name on buy-
ers’ perceptions of product quality. J Marketing Res. 1989;26(3):351-357.
2. Benedetti F. How the doctor’s words affect the patient’s brain. Eval Health Prof.
2002;25(4):369-386.
3. Koshi EB, Short CA. Placebo theory and its implications for research and clini-
cal practice. Pain Pract. 2007;7(1):4-20.
4. Berns GS, Chappelow MC, Zink CF, Pagnoni G, Martin-Skurski ME. Neurobio-
logical substrates of dread. Science. 2006;312(5774):754-758.
5. Price DD, Fields HL. The contribution of desire and expectation to placebo an-
algesia: implications for new research strategies. In: Harrington A, ed. The Placebo
Effect: An Interdisciplinary Exploration. Cambridge, Massachusetts: Harvard Uni-
versity Press; 1999:118-119.
6. Gracely RH, Dubner R, Deeter WR, Wolskee PJ. Clinicians’ expectations influ-
ence placebo analgesia. Lancet. 1985;1(8419):43.
Table. Comparison of Participants Assigned to Regular-Price Placebo
vs Low-Price (Discounted) Placebo
Regular Price
(n = 41)
Low Price
(n = 41)
P
Value
Women, No. (%) 27 (65.9) 24 (58.5) .50
Age, mean (SD), y 30.9 (12.4) 30.0 (11.4) .74
Calibrated maximum tolerance,
mean (SD), V
51.8 (18.7) 54.9 (23.3) .50
Shocks received, No. (SD) 18.2 (7.2) 18.6 (9.1) .80
Change in pain scores
a
All shocks,
No. (%) [95% CI]
Pain reduction 35 (85.4)
[74.6-96.2]
25 (61.0)
[46.1-75.9]
.02
b
Pain increase 6 (14.6)
[3.8-25.5]
16 (39.0)
[24.1-54.0]
Highest-intensity shocks only,
No. (%) [95% CI]
c
Pain reduction 33 (80.5)
[68.3-92.6]
23 (56.1)
[40.9-71.3]
.03
b
Pain increase 8 (19.5)
[7.4-31.6]
18 (43.9)
[28.7-59.1]
Abbreviation: CI, confidence interval.
a
Comparison of participants experiencing a mean reduction in pain after vs before the
placebo pill was administered (visual analog scale point reduction between 0.01 and
48.4) and those experiencing a mean increase in pain (visual analog scale point in-
crease between 0 and 29.2).
b
Two-tailed
2
test.
c
Highest 50% of shocks by intensity.
Figure. Pain Ratings by Voltage Intensity
35
10
25
20
15
30
5
0
–5
No.
Regular price
Low price
Shock Intensity, V
Mean Difference
10
41
Placebo price
Regular
Low
15 20 25 30 35 40 45 50 55 60 65 70 75 80
41 41 40 37 31 27 23 21 20 18 14 12 9 8
41 41 41 40 38 31 29 27 24 19 17 11 7 5 4
Mean difference in pain ratings, after vs before placebo, by voltage intensity. Higher
value indicates greater pain reduction. The table depicts the intensity of the shocks
and the number of observations in the regular-price and low-price conditions. P
value is less than .05 for the shock intensities 27.5 V through 30.0 V, 35.0 V through
75.0 V, and 80.0 V.
LETTERS
©2008 American Medical Association. All rights reserved. (Reprinted) JAMA, March 5, 2008—Vol 299, No. 9 1017
at INSEAD, on March 4, 2008 www.jama.comDownloaded from