Simple decompression or anterior subcutaneous transposition for ulnar neuropathy at the elbow: a cost-minimization analysis--Part 2.

Abstract

A randomized controlled trial was reported recently, in which simple decompression and anterior subcutaneous transposition were compared for treatment of ulnaropathy at the elbow. Clinically, both surgical options seem to be equally effective. The objective of this study is to compare the costs, from a societal standpoint, of simple decompression versus anterior subcutaneous transposition in Euros.
Published and unpublished data from a randomized controlled trial that compared simple decompression and anterior subcutaneous transposition were used. The investigators tabulated the actual costs incurred for professional fees, use of the operating room, duration of sick leave, reintervention, and treatment of complications directly related to the surgical treatment.
The total costs per group and per patient were statistically significantly less for those treated with simple decompression. The total median costs per patient were 1124 Euros for simple decompression and 2730 Euros for anterior subcutaneous transposition. The main difference was in the costs related to sick leave, which is significantly shorter for simple decompression. There also was a statistically significant difference in operation time in favor of simple decompression. A remarkable finding was that paid workers in both treatment groups returned to work much later than unpaid workers.
Although clinically equally effective, simple decompression was associated with lower cost than anterior subcutaneous transposition for the treatment of ulnar neuropathy at the elbow. Compared with anterior subcutaneous transposition, we recommend simple decompression for the treatment of idiopathic ulnar neuropathy at the elbow.

Full text

Page 1

CLINICAL STUDIES
SIMPLE DECOMPRESSION OR ANTERIOR SUBCUTANEOUS
TRANSPOSITION FOR ULNAR NEUROPATHY AT THE
ELBOW: A COST-MINIMIZATION ANALYSIS—PART 2
Ronald H.M.A. Bartels,
Ph.D.
Department of Neurosurgery,
University Medical
Center St. Radboud,
Nijmegen, The Netherlands
Evelien H. Termeer, M.D.
Medical Technology Assessment,
University Medical Center,
Faculty of Medicine,
Nijmegen, The Netherlands
Gert Jan van der Wilt,
M.D., Ph.D.
Medical Technology Assessment,
University Medical Center,
Faculty of Medicine,
Nijmegen, The Netherlands
Leo G.M. van Rossum
Medical Technology Assessment,
University Medical Center,
Faculty of Medicine,
Nijmegen, The Netherlands
Jan Meulstee, M.D., Ph.D.
Department of Clinical
Neurophysiology,
Canisius Wilhelmina Hospital,
Nijmegen, The Netherlands
Wim I.M. Verhagen, M.D.,
Ph.D.
Department of Clinical
Neurophysiology,
Canisius Wilhelmina Hospital,
Nijmegen, The Netherlands
J. André Grotenhuis, M.D.,
Ph.D.
Department of Neurosurgery,
University Medical
Center St. Radboud,
Nijmegen, The Netherlands
Reprint requests:
R.H.M.A. Bartels, Ph.D.,
University Medical Center
St. Radboud, Department of
Neurosurgery, R. Postlaan 4,
6500 HB Nijmegen, The Netherlands.
Email: r.bartels@nch.umcn.nl
Received, March 25, 2004.
Accepted, December 13, 2004.
OBJECTIVE: A randomized controlled trial was reported recently, in which simple
decompression and anterior subcutaneous transposition were compared for treatment
of ulnaropathy at the elbow. Clinically, both surgical options seem to be equally
effective. The objective of this study is to compare the costs, from a societal standpoint,
of simple decompression versus anterior subcutaneous transposition in Euros.
METHODS: Published and unpublished data from a randomized controlled trial that
compared simple decompression and anterior subcutaneous transposition were used.
The investigators tabulated the actual costs incurred for professional fees, use of the
operating room, duration of sick leave, reintervention, and treatment of complications
directly related to the surgical treatment.
RESULTS: The total costs per group and per patient were statistically significantly less
for those treated with simple decompression. The total median costs per patient were
€1124 for simple decompression and €2730 for anterior subcutaneous transposition.
The main difference was in the costs related to sick leave, which is significantly shorter
for simple decompression. There also was a statistically significant difference in
operation time in favor of simple decompression. A remarkable finding was that paid
workers in both treatment groups returned to work much later than unpaid workers.
CONCLUSION: Although clinically equally effective, simple decompression was associ-
ated with lower cost than anterior subcutaneous transposition for the treatment of ulnar
neuropathy at the elbow. Compared with anterior subcutaneous transposition, we recom-
mendsimpledecompressionforthetreatmentofidiopathiculnarneuropathyattheelbow.
KEY WORDS: Cost effectiveness, Surgery, Ulnar nerve
Neurosurgery 56:531-536, 2005
U
trolled trial that compared the efficacy of
simple decompression (SD) and anterior sub-
cutaneous transposition (AST) (4). Because
both surgical modalities seem to be equally
effective clinically, cost minimization analysis
may contribute to the discussion as to whether
one of these should be recommended.
DOI: 10.1227/01.NEU.0000154132.40212.6E
www.neurosurgery-online.com
lnar neuropathy at the elbow is com-
mon. Recently, we presented the clin-
ical results of a randomized con-
PATIENTS AND METHODS
The study protocol was approved by the in-
stitutional review boards of the University Med-
ical Center St. Radboud and the Canisius Wil-
helmina Hospital. Data were collected from the
databaseofarecentlyreportedrandomizedcon-
trolled study (1). The study design has been
described elsewhere (4). A group of 152 patients
withidiopathiculnaropathyattheelbowascon-
firmed by electromyography were randomly as-
signed for treatment with SD (n ? 75) or AST (n
? 77). The primary end point was clinical out-
come at 1 year. An additional end point was the
percentage of patients requiring revision sur-
gery within 1 year postoperatively. One year
postoperatively, there were 147 assessable pa-
tients. The remaining patients were lost to
follow-up.
The analysis was conducted from a societal
perspective, implying that health care and
nonhealth care costs were taken into account,
and actual costs rather than tariffs were used
NEUROSURGERY
VOLUME 56 | NUMBER 3 | MARCH 2005 | 531

Page 2

for unit costs (6). No discounting was used because of the
relatively short analysis period of 1 year. Costs associated with
productivity losses were estimated using the friction-costs
method (8). Standard costs were used as available in accor-
dance with prevailing national guidelines for economic eval-
uation in health care (8). All relevant volumes were recorded
prospectively by use of case record forms and patient diaries.
Protocol-driven costs were ignored. Costs are expressed in
2003 Euros (€1 ? $1.23/£0.67). An incremental approach was
taken to assess differential costs incurred in the two treatment
groups. Unit costs and sources are listed in Table 1.
Operation time is defined as the surgical time (skin to skin).
The operation time is used to calculate the costs of the use of
the operating room, including costs of nursing staff time, debit
of the building and instruments, calculation of the neurosur-
geon’s fee, and so forth. All patients were scheduled for lo-
coregional (plexus) anesthesia. If this was insufficient, either
local infiltration with an anesthetic or general anesthesia was
added. Some patients preferred general anesthesia. For each
reported complication, the costs of treatment were estimated.
According to the protocol of the study, patients who had
experienced no improvement or had worsened at 30 weeks
postoperatively were offered a reintervention. This consisted
of an anterior submuscular transposition.
Patient sick leave was recorded by use of a calendar on which
the patient marked the day that he/she could use the arm nor-
mally and resumed 100% of their occupation. Patients who did
not have a regular job (e.g., retirees) should have regained nor-
mal activities without restriction. For unpaid and paid work,
various costs for 1 day of sick leave were used (8). If sick leave
exceeded 123 days, the remaining days were not included ac-
cording to the friction-cost method. After 123 days, it may be
presumed that the activities of the absentee were assumed by
others without additional cost (8). Missing values were replaced
by the median values of the associated treatment group.
The Student’s t test and Mann-Whitney U test were used for
statistical analysis. A P value of ?0.05 was considered statis-
tically significant.
TABLE 1. Investigated items, definition of units, and costs per unita
Parameter UnitCosts per unit (€)
Operating room (debit operating room, debit
anesthesia instruments, inventory, nurses)
Cost per operationAST: 150
SD: 65.65
Neurosurgeon’s fee Cost per operationb
AST: 24.70
SD: 10.81
Anesthesia
Local anesthesia
Plexus anesthesia
General anesthesia
Plexus and local anesthesia
Plexus and general anesthesia
Total per operationc
Total per operationd
Total per operatione
Total per operationf
Total per operationf
5.95
42.59
41.02
48.54
81.23
Absenteeism from work
Sick leave Absence (d)g
Paid work, 125.67
Unpaid work, 78.34
Reinterventionh
Operating room (see above)
Neurosurgeon’s fee
Cost per operation
Cost per operation
215.65
35.52
Complications
Visit physicianVisitGeneral practitioner, 18.13i
Neurosurgeon, 49.47
0.71 Amoxicillin, 1125 mg daily Daily dose
aAST, anterior subcutaneous transposition; SD, simple decompression.
bBased on a gross monthly salary of €9439 or an hourly salary of €47.35 or €0.79 per minute. This is the same for a neurosurgeon and an anesthesiologist.
cIncluded are anesthetic agent, €1.84 (4); necessities such as needles, syringe, etc., €0.16; anesthesiologist’s fee, 5 minutes: €3.95.
dIncluded are anesthetic agent, €19.71 (4); necessities such as needles, etc., €9.46; anesthesiologist fee, 17 minutes: €13.43.
eIncluded are anesthetic agent, €18.38 (4); necessities such as needles, tube, etc., €12.37; anesthesiologist’s fee, 13 minutes: €10.27.
fThe costs of the respective modes of anesthesia are added.
gA distinction is made between paid and unpaid work. Prices are indexed on 2003 values.
hBased on an operation time of 45 minutes.
iPrices are indexed on 2003 values.
BARTELS ET AL.
532 | VOLUME 56 | NUMBER 3 | MARCH 2005www.neurosurgery-online.com

Page 3

RESULTS
As previously demonstrated (4), operative time is signifi-
cantly less for SD compared with AST. SD lasted 13.7 ? 4.9
minutes, whereas AST required 31.3 ? 7.0 minutes. This dif-
ference was highly statistically significant (P ? 0.0001). There-
fore, the costs related to the operation (operating room and
neurosurgeon fee) were less for SD than for AST (Table 2). The
ratio of operative time required for SD compared with AST
(0.43) was the same as the ratio of the surgery-related costs of
both treatments.
Because the mode of anesthesia was not affected by the
duration of the operation, these costs were nearly equivalent.
The total costs related to anesthesia were €3197.73 for SD and
€3382.34 for AST. This small difference is nonexistent if calcu-
lated on a per-patient basis: €42.64 for SD and €43.93 for AST.
The costs generated by reintervention were nearly equiva-
lent on a per-patient basis: €27.42 for SD and €22.89 for AST.
The costs attributable to the treatment of complications were
€4.31 per patient for SD, and €4.01 for AST, which was not
statistically significant.
The mean duration of sick leave was 18.7 ? 22.7 days for the
SD group and 28.0 ? 23.5 for the AST group. For four patients
from the SD group and two of the AST group, information
regarding return to work was missing. If a distinction is made
between paid and unpaid work, the mean duration of sick
leave for unpaid work was 9.3 ? 8.0 days for the SD group and
20.0 ? 21.2 days for the AST group. This difference reached
statistical significance (P ? 0.008). Comparison of paid work
resulted in a statistically significant difference (P ? 0.008): 23.9
? 26.4 days for the SD group and 31.6 ? 23.7 days for the AST
group. The duration of sick leave fro one patient in the AST
group with paid work exceeded 123 days (134 d). According to
the friction-costs method, the actual value is replaced by 123
days for calculation of the costs. Considering paid or unpaid
work within a treatment group, a statistically significant differ-
TABLE 2. Calculation of costs for simple decompression and anterior subcutaneous transpositiona
Parameter
SDAST
No. of
units
Total costs
(€)
Subtotal
(€)
No. of
units
Total costs
(€)
Subtotal
(€)
Operation
Operating room
Neurosurgeon’s fee
75
75
4923.75
810.96
77
77
11,550
1902.195734.7113,452.19
Anesthesia
Local anesthesia
Plexus anesthesia
General anesthesia
Plexus and local anesthesia
Plexus and general anesthesia
2 11.90
2470.22
246.06
388.32
81.23
—
61
12
1
3
—
58
6
8
1
2597.99
492.12
48.54
243.69 3197.73 3382.34
Work absenteeism from
Sick leave, paid
Sick leave, unpaid
23.9 ? 48b
9.3 ? 27b
144,409.91
19,671.17
31.4 ? 53b
20.0 ? 24b
209,339.83
37,527.99 164,081.08246,867.82
Reintervention
Operating room
Neurosurgeon’s fee
Plexus anesthesia
7
7
7
1509.55
248.64
298.13
6
6
6
1293.90
213.12
255.542056.32 1762.56
Complications
Visit physician
General practitioner
Neurosurgeon
Amoxicillin
1
6
18.13
296.82
8.52
3
5
54.39
247.35
7.1012323.47 10 308.84
Total 175,393.31 265,773.75
Total per patient 2338.583451.61
aSD, simple decompression; AST, anterior subcutaneous transposition.
bThe mean duration of absence in days ? the number of patients is calculated. For the six missing values, the median value of the duration of sick leave of the
involved group was used.
COST-MINIMIZATION ANALYSIS FOR ULNAR NERVE TREATMENT
NEUROSURGERY
VOLUME 56 | NUMBER 3 | MARCH 2005 | 533

Page 4

ence also was found (P ? 0.001 for SD; P ? 0.042 for AST).
Because patients from the SD group regained full activities and
workmuchmorequickly,thecostsrelatedtosickleavearemuch
lower in that group. On a per-patient basis, costs associated with
sick leave were €2187.75 for SD and €3206.08 for AST. This
difference reached statistical significance (P ? 0.001).
The total of calculated costs were €175393.31 for the patients
treated via SD and €265773.75 for those who underwent AST.
On a per-patient basis, the total costs were €2338.58 for SD and
€3451.61 for AST. The difference between the two means was
highly significant (P ? 0.001) (Fig. 1). The costs are not dis-
tributed normally; therefore, the median costs also were cal-
culated (Fig. 2). The difference in the median costs also
reached statistical significance (P ? 0.001).
DISCUSSION
After carpal tunnel syndrome, ulnaropathy at the elbow is
the most frequent nerve compression syndrome in the arm.
Since its first description in 1816, surgical treatment of ulnar
neuropathy at the elbow has received much attention (2, 3).
Despite the long-standing interest, only recently was a large
randomized controlled study reported that compared the clin-
ical effectiveness of SD with that of AST. With the exception of
the complication rate, both surgical options seemed to be
equally effective clinically in the treatment of idiopathic ulnar
palsy. A good outcome was achieved in 67.1% of the SD group
and 73% of the AST group (4).
When clinical effectiveness is equivalent, costs related to a
treatment can be an argument to recommend one treatment
over another. Cost-minimization analysis is an analytic tool to
compare the costs of programs or treatments that achieve the
same result (7). Because of the inclusion of sick leave in our
analysis, the cost minimization analysis was conducted from a
societal viewpoint.
It was clearly shown that SD generated statistically signifi-
cantly less costs than AST. Two major reasons can be defined.
First, sick leave accrued for paid as well as for unpaid work
was much shorter after SD. Therefore, the costs were propor-
tionately lower. Whether they underwent a SD or AST, all
patients were encouraged to regain their normal and working
activities as soon as possible. Restrictions were not given.
Probably because of the amount of tissue dissection, SD is less
painful, and patients use their arms sooner. It is remarkable
that this is not reflected in the clinical results after 6 weeks
postoperatively; these results were equivalent for SD and AST.
We assume that the reason for this discrepancy is because, in
both groups, most patients had regained their activities before
FIGURE 1. Histograms showing the total costs per patient and the num-
ber of patients. A, SD; B, AST. Neither is normally distributed (continu-
ous curve).
FIGURE 2. Graph of median costs (solid line) and values of the 25%
and 75% quartiles (upper and lower limits of boxes), and the upper
and lower costs (lines above and below stems). Extremes also are repre-
sented (?).
BARTELS ET AL.
534 | VOLUME 56 | NUMBER 3 | MARCH 2005 www.neurosurgery-online.com

Page 5

the first postoperative control examination at 6 weeks. If we
had included a control examination at 3 weeks, a difference
probably could have been demonstrated.
The difference in regaining activities was greatly affected by
whether work was paid. Unpaid workers (i.e., housewives) re-
sumed their (work) activities much sooner than paid workers.
Part of this could be explained because the group of paid work-
ers included some industrial employees exposed to heavy labor,
although housekeeping is not a light activity. In our opinion, our
social security system, which provides very good reimbursement
in case of sick leave, should be held most responsible. The drive
to regain work is less when income is guaranteed.
Second, the difference in operative time is an important
factor in evaluating the difference in costs. As SD is a simpler
method than AST, it is logical that the difference in operative
time between SD and AST is statistically significant (4). Be-
cause the costs for surgery are directly related to operative
time, these will be proportionally higher for AST. This was
made evident when the ratio of the costs for surgery of SD and
AST per patient and the ratio of the operation times per
patient were compared; they were equivalent.
Some points should be addressed. Although the study had
a high internal validity, the results cannot be extrapolated to
other countries without caution. For example, the social insur-
ance system is very well organized in the Netherlands. Pa-
tients are not forced to start working as soon as possible
because they are compensated during their sick leave. In other
countries with a less developed social insurance system, pa-
tients probably will resume work much sooner, which would
reduce the costs of both treatments accordingly. Even when
the costs of sick leave were excluded from analysis, however,
a significant difference in costs remained.
This difference was based mainly on the different operative
times. We calculated the fee per minute. In most countries, a
fixed price per operation would be used rather than hourly
compensation. This would not affect our results unless the fee
for SD is more than 50% less than that for AST. We are not
aware of any such a situation. In most instances, the fee is the
same or nearly the same for SD and AST. Thus, our calculation
remains valid because two SDs can be performed in the time
required to perform one AST.
The calculation of operative time is not a true representation
of all the time neurosurgeons spend with their patients. They
must speak with the patient and/or family members before
and after surgery and conduct perioperative management
tasks (putting in orders, etc.). For some surgeons, additional
staff members (e.g., residents, physician assistants) assist with
the operation and perioperative management, which may save
the surgeon time but also add cost to the procedure.
We estimated rather than measured precisely the time spent
in perioperative management such as including pre- and post-
operative visits with the patient, administration, and phone
calls. Because the number of pre- and postoperative visits did
not differ and, because they are related to the institution, we
do not think that incorporating an estimation of the costs
related to this aspect of treatment makes a major contribution
to the discussion.
Despite these caveats, we conclude that for treatment of
idiopathic ulnaropathy at the elbow, SD is the method of
choice compared with AST. SD is equally effective as AST
clinically, it is a less-demanding surgical procedure, it has a
lower complication rate (4), and it generates less cost.
CONCLUSION
In the Netherlands, costs associated with SD were less than
those associated with AST for the treatment of idiopathic
ulnaropathy at the elbow. In combination with our former
clinical results, SD is the method of choice for the treatment of
idiopathic ulnar neuropathy at the elbow compared with AST.
REFERENCES
1. http://statline.cbs.nl. Accessed January 31, 2005.
2. Bartels RHMA: History of the surgical treatment of ulnar nerve compression
at the elbow. Neurosurgery 49:391–400, 2001.
3. Bartels RHMA, Menovsky T, van Overbeeke JJ, Verhagen WIM: Surgical
management of ulnar nerve compression at the elbow: An analysis of the
literature. J Neurosurg 89:722–727, 1998.
4. Bartels RHMA, Verhagen WIM, van der Wilt GJ, Meulstee J, van Rossum
LGM, Grotenhuis JA: Prospective randomized controlled study comparing
simple decompression versus anterior subcutaneous transposition for idio-
pathic neuropathy of the ulnar nerve at the elbow: Part 1. Neurosurgery
56:522–530, 2005.
5. Deleted in proof.
6. Drummond MF, O’Brien B, Stoddart GL, Torrance GW: Methods for the
Economic Evaluation of Health Care Programmes. Oxford, Oxford University
Press, 1997.
7. Gold MR, Siegel JE, Russel LB, Weinstein ME: Cost-effectiveness in Health and
Medicine. New York, Oxford University Press, 1996.
8. Oostenbrink JB,Koopmanschap MA,
kostenonderzoek, methoden voor richtlijnprijzen voor ecomische evaluaties in de
gezondheidszorg. Amstelveen, College voor zorgverzekeringen, 2000.
RuttenFFH:
Handleidingvoor
Acknowledgments
This study was supported by Grant 99112/601/8 from the National Health
Insurance Board and the National Society of University Medical Centres.
COMMENTS
T
(AST) for ulnar neuropathy at the elbow reiterates the fact that
of the two procedures, SD was the more favorable approach,
this time from a cost standpoint. SD generated statistically less
cost than AST per procedure, which in United States dollars
was $1382.52 for SD versus $3357.90 for AST (2003 rate of
conversion).
The reasons for the cost difference included sick leave, with
a much work leave associated with SD versus AST. Sick leave
is paid in the Netherlands, and patients are not forced to start
working as soon as possible. In addition, difference in oper-
ating room time was an important factor in the cost difference:
SD is simpler than AST and is performed more quickly.
his second part of the comparison of simple decompres-
sion (SD) versus anterior subcutaneous transposition
COST-MINIMIZATION ANALYSIS FOR ULNAR NERVE TREATMENT
NEUROSURGERY
VOLUME 56 | NUMBER 3 | MARCH 2005 | 535

Page 6

This second article was necessarily separated from the first
because, as the authors stated in their submission letter, “the
first article is a pure clinical study and the other a manuscript
addressing the costs of two different treatments. Combining
the two articles would distract the attention from each sepa-
rate message and weaken the studies, whereas as the separate
articles reinforce each other.” The two articles complement
each other.
Daniel H. Kim
Judith A. Murovic
Stanford, California
T
ysis comparing SD with AST in 152 patients surgically treated
for idiopathic ulnar neuropathy. The authors made a reason-
able effort to estimate the cost associated with each procedure,
which included professional fees, operating room time, dura-
tion of sick leave, reintervention, and operative complications.
Some of the estimations may not be entirely accurate. For
example, the authors used 14 minutes for calculation of the
neurosurgeon’s fee for an SD. Although a neurosurgeon could
perform an SD from skin to skin in an average of 14 minutes,
neurosurgeons may spend much longer time on an average
procedure, e.g., the time spent talking to the patient and
family members before and after surgery, perioperative man-
agement, dictating operative reports, and so forth.
It seems that the cost associated with sick leave accounted
for more than 93% of the average total cost of two procedures
(€2703 per patient). A less morbid procedure such as SD with
his article is the second part of a prospective, randomized
controlled study in which the authors performed cost anal-
a faster return to normal use of the arm thus had a signifi-
cantly decreased cost associated with sick leave. Although the
cost for operation differed significantly between the two pro-
cedures (€119 for SD versus €218 for AST), it accounted for
almost a negligible portion of the total cost per patient (?6%).
The same is true for the cost associated with operative com-
plications (average, €4 per patient, ?0.2% of the total cost). As
the authors correctly observe, the conclusions of this study
should not be extrapolated to other countries without caution
relative to the significant differences in insurance structure,
reimbursement system, and tort systems.
Jason H. Huang
Eric L. Zager
Philadelphia, Pennsylvania
T
tive than AST. Although the reoperation rate and complication
rate was lower in SD than AST, data were only extended out
1 year. In our experience, most ASTs performed elsewhere are
successful, and perhaps only 1 of 20 patients requires reop-
eration during a 5-year period. However, although the need
for acute reoperation is less for SD than AST, my impression
is that during a longer period of time, these figures even out or
even slightly favor AST, so a case for AST can still be made.
he authors provide an interesting cost analysis of SD ver-
sus AST. It is not surprising that SD was more cost effec-
David G. Kline
New Orleans, Louisiana
Image from Exercitationes de Motu Cordis et Sanguinis (drawing in pen and ink, 1628), by Willian Harvey (from, facsimile of The Ana-
tomical Exercises of Dr. William Harvey. New Haven, Yale University Medical Historical Library, 1987).
BARTELS ET AL.