ANATOMIC PATHOLOGY
Original Article
Value-Based Pathology
A Cost-Benefit Analysis of the Examination of Routine and
Nonroutine Tonsil and Adenoid Specimens
JULIE C. NETSER, MD,
1
ROBERT A. ROBINSON, MD, PhD,
1
RICHARD
J.
SMITH, MD,
2
AND
STEPHEN
S.
RAAB, MD
1
To study the cost-effectiveness of the histologic examination of tonsil
and adenoid specimens, the histologic diagnoses for all routine
(2,700) and nonroutine (71) tonsil and adenoid specimens during a
10-year period were reviewed. There were 27 routine cases (1%) and
56 nonroutine cases (79%) with a diagnosis other than normal, tonsil-
litis,
or hyperplasia. Twelve of the 27 routine cases did not have a sig-
nificant clinical history, and a potentially significant pathologic diag-
nosis was made in only 6 cases; in none of these cases did the
As efforts persist to obtain quality care at an afford-
able cost, cost-effectiveness has become a dominant
issue for deciding optimal medical care, including
diagnostic testing.
1
The traditional goals of the patho-
logic examination have been the following: (1) to pro-
vide information to guide patient care; (2) to serve as
a quality control measure; and (3) to teach residents
and fellows. These goals now must be weighed
against the cost of examination.
Applying cost-benefit analysis to anatomic pathol-
ogy is complicated because the utility depends on the
type of tissue examined. All tissues do not provide
equal information to guide patient care, and speci-
mens simplistically may be classified as routine or
nonroutine. Routine specimens, such as hernia sacs,
gallbladders, elective joint replacement tissue, and
disk material, rarely provide diagnostic information
that affects patient care.
2-7
Nonroutine specimens,
From the Departments of ^Pathology and
2
Otolaryngology,
University of
Iowa
Hospitals and Clinics, Iowa City, Iowa.
This study was the recipient of the Stowell-Orbison runner-up
award at the United States and Canadian Academy of Pathology
Meeting, March 23-29,1996, Washington, DC.
Manuscript received October 9,1996; revision accepted January
24,1997.
Address reprint requests to Dr Raab: University of Iowa
Hospitals and Clinics, Department of Pathology, 200 Hawkins Dr
6235 RCP, Iowa City, IA 52242-1009.
pathologic diagnosis affect patient care. In all nonroutine cases, the
pathologic diagnosis affected patient care. The average charge per
case to detect potentially significant disease in routine and nonrou-
tine cases was $64,718 and $525, respectively. We conclude that histo-
logic examination of nonroutine cases is cost-effective, whereas in
most routine cases with adequate clinical history, histologic examina-
tion is not cost-effective. (Key words: Cost; Pathology; Head and
neck; Tonsils; Adenoids) Am
J
Clin Pathol 1997;108:158-165.
such as those from biopsies and most resections, are
removed primarily to determine the pathologic basis
of disease, and the pathologic diagnosis usually
affects patient care.
Some specimens, such as tonsils and adenoids
(T&A), are routine or nonroutine, depending on the
clinical scenario.
8
In North America, there are not
uniform laboratory practice parameters for examin-
ing routine T&A specimens.
9
In a national survey of
pediatric otolaryngologists, Dohar and Bonilla
10
found that only approximately half (56%) of those
responding stated their T&A specimens had under-
gone microscopic analysis. Some pathology laborato-
ries histologically examine all routine T&A speci-
mens;
others use an age cutoff below which routine
specimens undergo gross examination, and still oth-
ers do not perform a histologic examination on any
specimen. These practices are guided by the follow-
ing contrasting beliefs: (1) Unless all routine T&A
specimens are examined histologically, a rare case of
a clinically significant disease may be missed; and (2)
Clinically significant disease is so rare that no histo-
logic examination is warranted.
11,12
The incidence of
clinically significant disease and the utility of the his-
tologic examination never have been studied in rou-
tine T&A specimens. In comparison with these speci-
mens,
nonroutine T&A specimens always are
examined histologically, although the value of this
examination also has not been studied.
158
Downloaded from https://academic.oup.com/ajcp/article/108/2/158/1756930 by U.S. Department of Justice user on 16 August 2022
NETSER ET AL
159
Utility of the Tonsil
o
We used cost-benefit analysis to study the inci-
dence of clinically significant disease and the utility of
the histologic examination of routine and nonroutine
T&A specimens.
MATERIALS AND METHODS
The final diagnosis text for all specimens coded
according to SNOMED (standardized nomenclature
of medicine) as "tonsil or adenoid" during a 10-year
8-month period from January 1,1985, to September 1,
1995,
at the University of Iowa Hospitals and Clinics
were reviewed retrospectively. During this period, the
records for 2,700 patients who had undergone bilat-
eral tonsillectomy, bilateral tonsillectomy and ade-
noidectomy, or adenoidectomy alone were identified.
These patients were considered to have routine speci-
mens.
Patients fulfilling these criteria, but having the
procedure performed as part of a multiple biopsy
panendoscopy procedure were excluded. Of the 2,700
patients, 1,033 (38.3%) had tonsillectomy, 898 (33.3%)
had tonsillectomy and adenoidectomy, and 769
(28.5%) had adenoidectomy alone. Of the cases, 2,500
(92.6%) were not performed in conjunction with any
other procedure; 125 cases (4.6%) were performed as
part of a uvulopalatopharyngoplasty procedure; and
75 cases (2.8%) were performed in conjunction with
other procedures, such as operations on the sinuses,
nevi removal, or submandibular gland excision. To
generate an age distribution, pathology reports for the
first 20 to 25 cases accessioned during each calendar
year (270 cases) were reviewed, and the age and type
of procedure were recorded.
Of the patients with a routine specimen, 27 (1%)
had a diagnosis other than normal, acute or chronic
tonsillitis, or lymphoid or follicular hyperplasia.
The medical chart, pathology requisition form,
pathology final report, and histologic slides were
reviewed for each of these patients. From the med-
ical charts, the age, clinical impression, and follow-
up data were recorded. From the pathology requisi-
tion form, the clinical history was recorded as
significant or nonsignificant. A significant clinical his-
tory consisted of a reported clinical diagnosis other
than normal, tonsillitis, or hyperplasia. A recorded
clinical diagnosis of normal, tonsillitis, or hyperpla-
sia or no recorded clinical diagnosis was considered
nonsignificant. The diagnosis from the pathology
report was compared with the diagnosis on retro-
spective histologic review, and the number of dis-
crepant diagnoses was recorded. The slides were
reviewed by one of us (R.A.R.) who was unaware of
Vol.'
1
Adenoid Examination
the original diagnosis and history. The pathologic
diagnoses were reclassified as consequential and non-
consequential. Nonconsequential diagnoses (eg,
papilloma) did not affect patient care, whereas con-
sequential diagnoses (eg, granulomatous disease)
could affect patient care.
To determine the percentage of potentially misdi-
agnosed routine cases, the histologic slides of 2% (54
cases) of the cases with a diagnosis of normal, tonsilli-
tis,
or hyperplasia were randomly reviewed, and the
number of discordant diagnoses was recorded.
During the same 10-year 8-month period, 71
patients had a unilateral tonsillectomy (18 patients) or
a tonsillar biopsy (53 patients) that was not performed
as part of a panendoscopy procedure. These patients
were considered to have nonroutine specimens. The
medical chart, pathology requisition form, pathology
final report, and histologic slides were reviewed for
each patient.
The clinical utility of performing the gross and
histologic examination for routine and nonroutine
specimens was performed using charge-benefit
analysis. A pathologic examination was assumed to
be beneficial if the diagnosis determined by this
examination had a potential or an actual effect on
patient care. The benefits of quality assurance and
teaching were not assessed.
For routine specimens, if no cases had undergone
pathologic examination, there would have been no
changes in patient care because disease detected by
pathologic examination would have gone undetected.
A potential effect on patient care existed under the
following conditions: (1) the pathologic diagnosis was
not normal, tonsillitis, or hyperplasia; (2) the clinical
history was insignificant; and (3) the pathologic diag-
nosis was consequential. By the first criterion, 2,673
patients were excluded, leaving the set of 27 patients
previously discussed. The second and third criteria
further excluded patients. Thus, the group of patients
in whom the pathologic diagnosis had a potential
effect on patient care had to have no reported clinical
diagnosis or a benign clinical diagnosis and a patho-
logic diagnosis that was unsuspected and of enough
import to warrant possible intervention. In only a
subgroup of these patients was there an actual effect
on patient care as determined by chart review.
For all nonroutine specimens, we assumed that
patient care depended on the pathologic diagnosis;
these specimens had been removed primarily to
obtain diagnostic information. In a subset of these
patients, there was an actual effect on patient care,
which was determined by chart review.
•
No.
2
Downloaded from https://academic.oup.com/ajcp/article/108/2/158/1756930 by U.S. Department of Justice user on 16 August 2022
160
ANATOMIC PATHOLOGY
Original Article
The charge benefit for the pathologic examination of
routine specimens was the total pathology charge associ-
ated with all routine specimens divided by the number
of specimens in which the pathologic diagnosis had a
potential or an actual effect on patient care. The resulting
value was the charge per case of potential effect on
patient care or the charge per case of actual effect on
patient care. In other words, this was the charge per case
to detect a potentially or actually significant unexpected
disease, assuming all routine cases were processed.
The charge benefit for the pathologic examination
of nonroutine specimens was the total pathology
charge associated with all nonroutine specimens
divided by the total number of specimens. This was
the charge per nonroutine case to determine a patho-
logic diagnosis that affected patient care.
All case charges consisted of a technical and a pro-
fessional charge. The technical charge consisted of a
charge per block submitted, whereas the professional
charge consisted of a charge per part examined.
Charge data were obtained from the University of
Iowa Hospitals and Clinics pathology billing office
and are reported in 1996 US dollars. To determine the
total charge for routine specimens, the average num-
ber of blocks submitted per case (2 blocks per bilateral
tonsillectomy, 1 block per adenoidectomy, and 2.4
blocks per bilateral tonsillectomy and adenoidec-
tomy) was used. The total pathology charge for a rou-
tine specimen was $78.25 per slide.
For nonroutine specimens, the average number of
blocks per case was 1.7 (range, 1-8 blocks). The total
pathology charge was $135.25 per slide. In some cases,
ancillary studies, such as immunoperoxidase, direct
immunofluorescence, gene rearrangement, and other
special stain charges were assessed. Immunoperoxidase
was performed in 5 cases (average charge, $515). Direct
immunofluorescence was performed in 11 cases (aver-
age charge, $1,309). Gene rearrangement studies were
performed in 5 cases (average charge, $445).
RESULTS
Patients With Routine Specimens
The average age was 24.2 years (range, 2-66 years;
median, 19.5 years), and the male-to-female ratio was
1.7:1. Of the 270 patients included, 161 were younger
than 9 years old. Of the 27 patients with a routine
specimen and a pathologic diagnosis other than
benign, tonsillitis, or hyperplasia, 12 had a nonsignifi-
cant clinical history, and 15 had a significant clinical
history. In patients younger than 16 years of age, a
diagnosis other than benign, tonsillitis, or hyperplasia
was made in one of every 1,015 routine specimens
examined. A similar diagnosis was given for one of
every 168 patients older than 16 years. Retrospective
histologic review of the 27 cases showed 100% concor-
dance with the original histologic diagnoses.
Patients with nonsignificant clinical history—The indi-
cations for operation in these 12 patients were sleep
apnea (4 patients), recurrent tonsillitis (6 patients), or
middle ear disease (2 patients). The average age was
26.6 years (range, 2-66 years; median, 22 years), and the
male-to-female ratio was 1.4:1. Six patients had a conse-
quential pathologic diagnosis (Table 1) and six patients
TABLE 1. PATIENTS WITH ROUTINE SPECIMENS, INSIGNIFICANT CLINICAL HISTORY,
AND CONSEQUENTIAL PATHOLOGIC DIAGNOSES
Case No.
1
2
3
4
5
6
Age/Sex
2/M
31/F
20/
F
4/M
19/F
29/M
History
Chronic serous otitis media,
adenoid hyperplasia*
Chronic tonsillitis
Recurrent tonsillitis
Persistent middle ear effusion
Chronic tonsillitis
Chronic tonsillitis
Histologic Diagnosis
Atypical lymphoid proliferation suggestive
of posttransplant lymphoproliferative disorder
Noncaseating granulomas, infectious cause
vs sarcoidosis considered
Ulcerative tonsillitis with follicular and
immunoblastic hyperplasia consistent
with viral mononucleosis
Lymphoid hyperplasia and reactive
immunoblastic hyperplasia suggestive
of a viral infection
Follicular hyperplasia with granulomas
and yeast consistent with Histoplasma capsitlatum
Reactive hyperplasia with some atypical features
Change
None
None
None
None
None
None
Follow-up
4 y, no evidence
of disease
None
None (Epstein-
Barr virus
titers equivocal)
2y
None
8d
'History not provided;
2
years after organ transplantation.
AJCP • August 1997
Downloaded from https://academic.oup.com/ajcp/article/108/2/158/1756930 by U.S. Department of Justice user on 16 August 2022
NETSER ET AL
161
Utility of the Tonsil and Adenoid Examination
TABLE 2. PATIENTS WITH ROUTINE SPECIMENS, INSIGNIFICANT CLINICAL HISTORY,
AND INCONSEQUENTIAL PATHOLOGIC DIAGNOSES
Case No.
7
8
9
10
11
12
Age/Sex
22/F
22/M
52/F
66/M
5/M
47/M
History
Recurrent pharyngitis
Recurrent tonsillitis
Excessively loud snoring,
redundant mucosa, and
normal tonsils
Sleep apnea
Sleep apnea and
adenotonsillar hypertrophy
None
Histologic Diagnosis
Follicular hyperplasia and papilloma
Lymphoid hyperplasia and
fungal aggregates
Follicular hyperplasia and papilloma
Follicular hyperplasia and papilloma
Lymphoid hyperplasia and papilloma
Lymphoid hyperplasia and papilloma
Change
None
None
None
None
None
None
Folloiv-up
None
6 y, no evidence
of disease
None
1 y, no evidence
of disease
None
2 wk, no evidence
of disease
had an inconsequential pathologic diagnosis (Table 2).
The six patients with a consequential pathologic diagno-
sis had a clinical history of recurrent tonsillitis (4
patients) or middle ear disease (2 patients). The average
age of these patients was 17.5 years (range, 2-29 years),
and the male-to-female ratio was 1:1. Two patients were
younger than 16 years, and 4 were older. These patients
were considered to have a pathologic diagnosis that
potentially affected patient care. However, based on
chart review, the pathologic diagnosis did not affect clin-
ical management in any case. Indeed, nowhere in the
clinical notes were the histologic diagnoses mentioned.
The most concerning diagnosis, "atypical lymphoid pro-
liferation worrisome for PTLD [posttransplantation lym-
phoproliferative disorder]," was made in case 1. This
case should have been considered nonroutine, or at least
routine with a significant history, because the patient
had undergone organ transplantation 2 years earlier and
was at risk for such a process. Thus, the subset of
patients in whom the pathologic diagnosis had an actual
effect on management was empty.
The six patients who had an inconsequential patho-
logic diagnosis had a clinical history of sleep apnea (4
patients) or recurrent tonsillitis (2 patients). The aver-
age age was 36 years (range, 5-66 years), and the male-
to-female ratio was 2:1. Five patients had papillomas,
and one had fungal aggregates identified in the tonsil-
lar crypts. There was no tissue invasion by the hyphae,
and it was believed that the fungi were only colonizers.
In no case was there an effect on patient care.
Patients with significant clinical history—Data for
the 15 patients who had significant clinical history
are given in Table 3. In all cases, the clinical and
pathologic diagnoses were congruent. In 13 cases, the
pathologic diagnosis was consequential, and in 2
cases,
the pathologic diagnosis was inconsequential
(squamous papilloma).
Patients with normal diagnoses—Retrospective his-
tologic review of the 54 cases with a pathologic diag-
nosis of normal, tonsillitis, or hyperplasia showed
100%
concordance with the original diagnosis. In this
group, the average patient age was 9.3 years (range, 9
months to 60 years).
Nonroutine Specimens
The clinical and pathologic diagnoses of the 71
patients who had a nonroutine specimen are shown in
Table 4. The average patient age was 48.6 years
(range, 6-87 years; median, 46.5 years), and the male-
to-female ratio was 3:2. Three patients were 16 years
of age or younger. A diagnosis other than normal,
tonsillitis, or hyperplasia was made in 56 (79%) of the
71 patients. The overall concordance between the clin-
ical and pathologic diagnoses was 89%. There was a
high clinical suspicion of malignancy in 31 cases,
which was confirmed by pathologic examination in 29
cases;
in the other 2 cases, the pathologic diagnosis
was atypical lymphoid proliferation. In six cases the
clinical diagnosis was squamous cell carcinoma, and
the pathologic diagnosis was another malignancy. The
clinical impression was "favor benign" in 40 cases,
and a compatible histologic diagnosis was made in 39
cases.
In the discordant case, the patient had a history
of acquired immunodeficiency syndrome and an ulcer
that was believed, based on clinical data, to be benign;
Vol.
10
Downloaded from https://academic.oup.com/ajcp/article/108/2/158/1756930 by U.S. Department of Justice user on 16 August 2022
r
162
ANATOMIC PATHOLOGY
Original Article
TABLE 3. PATIENTS WITH ROUTINE SPECIMENS AND SIGNIFICANT CLINICAL HISTORY
Case No. Age/Sex History
Histologic Diagnosis
Follow-up
13
14
15
16
17
18
19
20
21
22
23
24
25
5/M
4/M
60/F
34/M
18/F
27/M
19/F
34/M
29/F
19/M
15/M
3/M
5/F
26
27
17/F
44/M
Obstructive sleep apnea
and Hurler's syndrome
Obstructive sleep apnea and
type 6 mucopolysaccharidosis
Tonsillitis with left peritonsillar abscess
Acute tonsillitis and anterior tonsillar bulge
consistent with abscess
Peritonsillar abscess
Acute tonsillitis and peritonsillar abscess
Papillomatous lesion
None
Huge tonsils in patient with mononucleosis
and airway obstruction
Mononucleosis and streptococcal tonsillitis
Necrotic oozing tonsils, hemorrhagic tonsillitis,
agranulocytosis, and hemocytophagia
secondary to Epstein-Barr virus
Obstructive tonsillitis and necrotic tonsils;
rule out mononucleosis
Hepatic transplantation; differential diagnosis,
lymphoid hyperplasia vs posttransplantation
lymphoproliferative disorder
Tonsillectomy 4 to 5 weeks earlier with no
adenoids seen; then obstructive adenoids;
workup for lymphoproliferative disease
Acute myelomonocytic leukemia (M5),
with airway obstruction
Follicular hyperplasia and histiocytes
consistent with storage disorder
Histiocytes consistent with storage disorder
Follicular hyperplasia and peritonsillar abscess
Lymphoid hyperplasia and wall of abscess
Tonsillitis and peritonsillar abscess
Lymphoid hyperplasia and peritonsillar abscess
Squamous papilloma
Squamous papilloma
Features consistent with mononucleosis
Marked lymphoid proliferation consistent
with mononucleosis
Left tonsil, exudate containing yeast consistent
with Candida organisms; right tonsil,
lymphoid depletion
Necrotizing tonsillitis with florid
immunoblastic hyperplasia
Follicular and immunoblastic hyperplasia
Follicular and interfollicular hyperplasia;
consider mononucleosis
Leukemic infiltrate
2y
6 mo
None
1 mo
2y
None
None
1 wk
1 mo
None
5y
None
2y
1 mo
1 V2 years,
died of
disease
histologic examination showed an atypical lymphoid
proliferation that, on follow-up, behaved aggressively.
Based on the medical chart review, in all cases, the
pathologic diagnosis affected patient care. The retro-
spective histologic diagnosis in all cases was concor-
dant with the original histologic diagnosis.
Charge-Benefit Analysis
Data from the charge-benefit analysis are shown in
Table 5. The total pathology charge for the examina-
tion of all routine T&A specimens was $390,482, or
$36,608 per year. The average charge for the examina-
tion of a routine case was $145. The average charge
per routine case to detect a diagnosis other than
benign, tonsillitis, or hyperplasia was $14,462. The
average charge per routine case to detect potentially
clinically significant disease was $64,718; this figure
excludes patients with significant clinical history (15
patients). There was no charge determined per case
for actual effect on patient care, because in no cases
was there actually a change in patient care based on
the pathologic examination.
The total pathology charge for the examination of
all nonroutine T&A specimens was $37,269. The aver-
age charge per nonroutine case was $525; this figure
also was the charge to detect potentially clinically sig-
nificant disease.
DISCUSSION
For routine T&A specimens, the incidence of clini-
cally significant disease was remarkably low. In only
27 (1%) of 2,700 cases from a period somewhat longer
than 10 years was the pathologic diagnosis something
other than benign, tonsillitis, or hyperplasia. After
excluding the cases in which the clinical history was
significant or in which the pathologic diagnosis was
not significant, there were only six cases in which the
pathologic diagnosis had the potential to affect patient
care.
The charge per case to examine all routine T&A
specimens to detect these six cases was $64,718. In
AJCP • August 1997
Downloaded from https://academic.oup.com/ajcp/article/108/2/158/1756930 by U.S. Department of Justice user on 16 August 2022