Content uploaded by Joseph Donahue
Author content
All content in this area was uploaded by Joseph Donahue on Nov 02, 2020
Content may be subject to copyright.
Administrative Databases Used for Sports Medicine
Research Demonstrate Significant Differences in
Underlying Patient Demographics and Resulting
Surgical Trends
Michelle Xiao, B.S., Joseph Donahue, M.D., Marc R. Safran, M.D., Seth L. Sherman, M.D.,
and Geoffrey D. Abrams, M.D.
Purpose: To discern differences between the PearlDiver and MarketScan databases with regards to patient demographics,
costs, reoperations, and complication rates for isolated meniscectomy. Methods: We queried the PearlDiver Humana
Patient Records Database and the IBM MarketScan Commercial Claims and Encounters database for all patients who had
record of meniscectomy denoted by Current Procedure Terminology 29880 or 29881 between January 1, 2007, and
December 31, 2016. Those that had any other knee procedure at the same time as the meniscectomy were excluded, and
the first instance of isolated meniscectomy was recorded. Patient demographics, Charlson Comorbidity Index, reopera-
tions, 30- and 90-day complication rates, and costs were collected from both databases. Pearson’s
c
2
test with Yate’s
continuity correction and the Student ttest were used to compare the 2 databases, and an alpha value of 0.05 was set as
significant. Results: We identified 441,147 patients with isolated meniscectomy from the MarketScan database (0.36% of
total database), approximately 10 times the number of patients (n ¼49,924; 0.20% of total database) identified from
PearlDiver. The PearlDiver population was significantly older (median age: 65-69) than the MarketScan cohort, where all
patients were younger than 65 (median age: 52; P<.001). Average Charlson Comorbidity Index was significantly lower
for MarketScan (0.172, standard deviation [SD]: 0.546) compared with PearlDiver (1.43, SD: 2.05; P<.001), even when
we restricted the PearlDiver cohort to patients younger than 65 years (1.02, SD: 1.74; P<.001). The PearlDiver <65 years
subcohort also had increased 30- (relative risk 1.53 [1.40-1.67]) and 90-day (relative risk 1.56 [1.47-1.66]) postoperative
complications compared with MarketScan. Overall, laterality coding was more prevalent in the PearlDiver database.
Conclusions: For those undergoing isolated meniscectomy, the MarketScan database comprised an overall larger and
younger cohort of patients with fewer comorbidities, even when examining only subjects younger than 65 years of age.
Level of Evidence: Level III, retrospective comparative study.
From the Department of Orthopedic Surgery, Stanford University School of
Medicine, Stanford, California, U.S.A.
The authors report the following potential conflicts of interest or sources of
funding: G.D.A. reports personal fees and other from Cytonics, personal fees
from Fida Pharma, personal fees from RobiconMD, personal fees from Side-
line Sports Doc, nonfinancial support from Arthrex, and nonfinancial support
from Stryker, outside the submitted work. J.D. reports other from Sta-bilynx,
outside the submitted work. M.R.S. reports personal fees from DJOrthopaedics,
personal fees from Smith & Nephew, personal fees from Stryker, personal fees
from Medacta, personal fees from Anika Therapeutics, personal fees from
Linvatec, personal fees from Biomet Sports Medicine, and other from Bio-
mimedica, outside the submitted work. S.L.S. reports personal fees from
Arthrex, personal fees from Ceterix Orthropaedics, personal fees from
CONMED Linvatec, personal fees from Flexion Therapeutics, personal fees
from GLG Consulting, personal fees from JRF Othro, personal fees from
Moximed, personal fees from Olympus, personal fees from Vericel, personal
fees from RTI Surgical, personal fees from Smith & Nephew, and grants from
DJO, outside the submitted work. Data access for this project was provided by
the Stanford Center for Population Health Sciences Data Core. The PHS Data
Core is supported by a National Institutes of Health National Center for
Advancing Translational Science Clinical and Translational Science Award
(UL1 TR001085) and internal Stanford funding. The funders had no role in
the study design, data collection and analysis, decision to publish, or prepa-
ration of the manuscript. The content is solely the responsibility of the authors
and does not necessarily represent the official views of the National Institutes
of Health. Data and analyses were conducted at Stanford. Full ICMJE author
disclosure forms are available for this article online, as supplementary
material.
Received April 25, 2020; accepted September 9, 2020.
Address correspondence to Geoffrey D. Abrams, M.D., 341 Galvez St., Mail
Code 6175, Stanford, CA 94305. E-mail: gabrams@stanford.edu
Ó2020 by the Arthroscopy Association of North America
0749-8063/20591/$36.00
https://doi.org/10.1016/j.arthro.2020.09.013
Arthroscopy: The Journal of Arthroscopic and Related Surgery, Vol -,No-(Month), 2020: pp 1-8 1
The use of administrative claims databases for con-
ducting orthopaedic surgery research has increased
considerably in the last decade.
1-4
Although databases
such as the National Inpatient Sample and Medicare
claims database have been widely used in the arthro-
plasty and spine literature,
1,4
these databases are not as
conducive to sports medicine studies, where procedures
are outpatient-based and patients are younger.
Two popular commercial claims databases, Pearl-
Diver
5-14
and MarketScan,
15-20
have emerged within
the field of sports medicine research. Both databases
offer access to a large population of patients in the
United States covered with private health insurance,
which is more representative of the demographic of
patients within an orthopaedic sports medicine practice.
Databases allow access to large sample sizes for identi-
fying trends over time, costs, and adverse events
relating to specific procedures or diagnoses.
3,4,21,22
Nonetheless, these databases could confer differing re-
sults for the same research question due to variations in
insurance coverage, coding accuracy, longitudinal pa-
tient tracking, and regional biases within the databases
themselves. Little evidence exists regarding direct
comparisons between the 2 databases.
Arthroscopic meniscectomy is a high-volume and
commonly performed outpatient orthopaedic proced-
ure performed in the United States
23
and therefore is
suitable for reporting on demographics, trends, and the
incidence of adverse events using large databases. The
purpose of this study was to discern differences be-
tween the PearlDiver and MarketScan databases with
regards to patient demographics, costs, reoperations,
and complication rates for isolated meniscectomy. We
hypothesized that there would be significant differences
between the 2 databases regarding demographics and
rates of adverse events despite the fact that the query
used the same procedure and time periods.
Methods
Databases
We queried 2 separate administrative claims data-
bases: the PearlDiver Humana Patient Records Database
(PearlDiver, Colorado Springs, CO) and the IBM Mar-
ketScan Commercial Claims and Encounters database
(IBM Watson Health, Armonk, NY). Both databases are
commercially available, Health Insurance Portability
and Accountability Actecompliant national databases
of insurance billing records. All data are deidentified
and thus exempt from institutional review board
approval. The PearlDiver Humana database contains
patient record information associated with International
Classification of Diseases, Ninth Revision (ICD-9) codes,
International Classification of Diseases, Tenth Revision (ICD-
10) codes, along with Current Procedure Terminology
(CPT) codes. The PearlDiver database includes records
from approximately 25 million privately insured and
Medicare Advantage patients covered between 2007
through the first quarter of 2017.
The MarketScan database contains individual-level
health insurance claims data across the continuum of
care (e.g., inpatient, outpatient, outpatient pharmacy)
as well as enrollment data from large employers and
health plans across the United States who provide pri-
vate health care coverage for more than 150 million
employees, their spouses, and dependents. This data-
base includes a variety of fee-for-service, preferred
provider organizations, and capitated health plans. The
MarketScan database includes patients covered be-
tween January 1, 2007, through December 31, 2016.
Study Cohorts
In both databases, we identified all patients who had
record of meniscectomy, denoted by CPT-29880 or
CPT-29881 codes, between January 1, 2007, and
December 31, 2016. Within this population, we only
included patients who were continuously enrolled for at
least 6 months before and 6 months after their initial
meniscectomy. To capture patients with an isolated
meniscectomy, those that had any other knee procedure
(Appendix Table 1, available at www.arthroscopy
journal.org) at the same time as the meniscectomy
were excluded. In the isolated meniscectomy cohort, the
first instance of a meniscectomy code for each patient in
the dataset was recorded.
We also identified the subpopulation of patients in
each database who went on to have knee surgery after
their initial meniscectomy. For inclusion into the sur-
gical groups, patients with records indicating the later-
ality of the initial meniscectomy and a subsequent knee
procedure (Appendix Table 1, available at www.
arthroscopyjournal.org) within 1 year with the same
laterality were identified. Because not all records
denoted laterality, we also calculated a separate reop-
eration rate for all patients regardless of the presence of
a laterality code, defined as an initial meniscectomy
followed by any subsequent knee procedure within 1
year. Average costs relating to initial meniscectomy as
well as costs for subsequent knee operations for those in
the reoperation group were calculated for both data-
bases. Occurrence of complications within 30 and 90
days after initial meniscectomy was determined using
ICD-9 diagnostic codes (Appendix Table 2, available at
www.arthroscopyjournal.org). These complications
included infection, deep vein thrombosis, acute kidney
injury (AKI), cardiac arrest, pneumonia, urinary tract
infection (UTI), wound dehiscence, hematoma, and
nerve injury. Patient age at the time of diagnosis, pa-
tient sex, region, and Charlson Comorbidity Index
(CCI)
24
were collected in both databases. The Pearl-
Diver database presents patient age within a 5-year
range, and MarketScan reports patient-level data. As
2M. XIAO ET AL.
the MarketScan database only identified patients
younger than the age of 65 years, we also repeated all
analyses with the subgroup of the PearlDiver patient
population younger than the age of 65 years.
Statistical Analysis
Patient demographic data were compared between
the databases using Pearson’s
c
2
test with Yate’s con-
tinuity correction. The Student ttest was used to
compare CCI and surgical cost data. Prevalence of
complications were calculated and represented as rela-
tive risk (RR) with 95% confidence intervals. All sta-
tistical analyses were 2-tailed with an alpha value of
0.05 set as significant.
Results
We identified 441,147 patients from the MarketScan
database and 49,924 patients from the PearlDiver
database who underwent isolated meniscectomy be-
tween 2007 and 2016. This represents 0.36% of all
MarketScan (123,637,719 total patients) and 0.20% of
all PearlDiver patients (25,034,227 total patients) dur-
ing this timeframe. More than one half (55.6%) of
patients identified from PearlDiver were older than 65
years, whereas all of the patients in the MarketScan
cohort were younger than 65 years. The median age of
PearlDiver patients was significantly older than the
MarketScan cohort (52 vs 65-69; P<.001; Table 1).
The PearlDiver <65 years cohort had a similar age
range as the MarketScan cohort (50-54 vs 52). The
majority of patients (59.5%) in the PearlDiver cohort
were from the South (Table 1). Of those who under-
went isolated meniscectomy, 8.46% in the MarketScan
cohort and 5.46% in the PearlDiver went on to have
another knee procedure within 1 year. This percentage
increased to 9.8% having record of subsequent
surgeries when we restricted the cohort of PearlDiver
patients to only include those younger than 65
(Table 1). When laterality was accounted for, the
overall number of defined reoperations decreased 19-
fold in the MarketScan cohort and 4-fold in the Pearl-
Diver <65 years subcohort (Table 1). There was a sig-
nificant difference in the reoperation rates between the
MarketScan and PearlDiver cohorts (P<.001; Table 1).
The average CCI for the MarketScan cohort was
significantly lower compared with both the PearlDiver
cohort and PearlDiver <65 years subcohort (P<.001;
Table 1). The overall 30-day and 90-day complication
rates were greater in the PearlDiver cohort compared
with the MarketScan cohort (30-day RR 1.81; 90-day
RR 1.93; Tables 2 and 3). Individual rates of complica-
tions for infection (P<.001), AKI (P<.001), pneu-
monia (P<.001), UTI (P<.001), wound dehiscence
(P<.046), and hematoma (P<.001) were also greater
for the PearlDiver cohort 30 days after surgery
(Table 2). All 90-day complication rates were signifi-
cantly greater in the PearlDiver cohort, and the RR of
having complications of AKI (RR ¼11.09), cardiac
Table 1. Patient Demographics and Reoperation Rates for Those Having an Isolated Meniscectomy Procedure Between 2007 and
2016 as Recorded in 2 National Claims Databases
MarketScan (n ¼441,147) PearlDiver (n ¼49,924) PearlDiver <65 (n ¼22,149)
n(%) n(%) n(%)
Sex
Female 194,833 (44.17) 27,103 (54.29) 10,437 (47.12)
Male 246,314 (55.83) 22,821 (45.71) 11,712 (52.88)
Age, y
<25 30,705 (6.96) 1024 (2.05) 1024 (4.62)
25-34 20,455 (4.64) 839 (1.68) 839 (3.79)
35-44 66,203 (15.01) 2695 (5.40) 2695 (12.17)
45-54 152,062 (34.47) 7174 (14.37) 7174 (32.39)
55-64 171,722 (38.93) 10,417 (20.87) 10,417 (47.03)
65-74 e20,968 (42.00) e
>75 e6807 (13.63) e
Median 52 65-69 50-54
Region
Northeast 81,281 (18.42) 979 (1.96) 243 (1.10)
North Central/Midwest 115,978 (26.29) 13,931 (27.90) 6318 (28.52)
South 160,265 (36.33) 29,689 (59.47) 13,681 (61.77)
West 74,760 (16.95) 5325 (10.67) 1907 (8.61)
Reoperations
No laterality code 37,324 (8.46) 2724 (5.46) 2170 (9.80)
With laterality code 1954 (0.83) 960 (2.59) 542 (3.23)
CCI (SD) 0.172 (0.546) 1.43 (2.05)*1.02 (1.74)*
Median age is only presented as a range in the PearlDiver database.
CCI, Charlson Comorbidity Index.
*P<.001 versus MarketScan.
ADMINISTRATIVE DATABASE DIFFERENCES 3
arrest (RR ¼4.80), pneumonia (RR ¼2.46), UTI
(RR ¼2.34), and hematoma (RR ¼2.31) all were
greater than 2.0 for PearlDiver compared with Mar-
ketScan (Table 3). Even when only patients younger
than 65 years were included within the PearlDiver
subcohort, the PearlDiver <65 years subcohort still had
increased 30- (RR 1.53 [1.40-1.67]) and 90-day (RR
1.56 [1.47-1.66]) postoperative complications
compared with MarketScan (Tables 2 and 3).
The cost billed for first time isolated meniscectomy
procedures as reported in the MarketScan cohort was
an average of $2853.37, whereas costs for the Pearl-
Diver cohorts were significantly less (P<.001), at
$1,418.36 for those younger than 65 years and
$1,121.64 for the entire cohort. The cost of a reopera-
tion procedure was significantly greater in both Pearl-
Diver cohorts as compared with MarketScan (P<.001).
Discussion
We found the MarketScan meniscectomy cohort to
have approximately 10 times the number of patients
versus the PearlDiver database. Even when we limited
analyses to include only patients younger than the age
of 65 years, PearlDiver patients had more comorbidities,
were more likely to be from the Southern region of the
United States, and have significantly increased 30- and
90-day complication rates versus those included in the
MarketScan database. The PearlDiver database, how-
ever, did have improved laterality coding for proced-
ures versus the MarketScan database (Table 4).
PearlDiver has been an established database used in
sports medicine research, accounting for 62% of all
database studies through 2015 in one journal.
25
Pearl-
Diver currently offers access to the Humana Claims
Database, Medicare Provider Utilization and Payment
Data, Medicare Standard Analytic Files, and the Na-
tional Inpatient Sample. Data are queried using a
program-specific bucket language, which affords a
simplified platform to gather and filter claims data
related to an ICD or CPT code. The large samples sizes
and smaller learning curve have made PearlDiver use-
ful in studying trends in surgical procedures or injuries
Table 2. 30-Day Complication Rates Following Isolated Meniscectomies Identified in MarketScan and PearlDiver
30-Day Complications
MarketScan PearlDiver PearlDiver <65 Years
n (%) n (%) RR (95% CI) PValue n (%) RR (95% CI) PValue
Infection 798 (0.18) 150 (0.30) 1.66 (1.40-1.98) <.001 69 (0.31) 1.72 (1.35-2.20) <.001
DVT 3072 (0.70) 381 (0.76) 1.10 (0.99-1.22) .096 175 (0.79) 1.13 (0.97-1.32) .112
AKI 65 (0.01) 101 (0.20) 13.73 (10.06-18.75) <.001 31 (0.14) 9.50 (6.19-14.57) <.001
Cardiac arrest 13 (<0.01) <11 ee<11 ee
Pneumonia 402 (0.09) 140 (0.28) 3.08 (2.54-3.73) <.001 53 (0.24) 2.63 (1.97-3.50) <.001
UTI 1826 (0.41) 553 (1.11) 2.68 (2.43-2.94) <.001 158 (0.71) 1.72 (1.47-2.03) <.001
Wound dehiscence 119 (0.03) 22 (0.04) 1.63 (1.04-2.57) .046 11 (0.05) 1.84 (0.99-3.41) .078
Hematoma 279 (0.06) 72 (0.14) 2.28 (1.76-2.95) <.001 31 (0.14) 2.21 (1.53-3.21) <.001
Nerve injury 34 (0.01) <11 ee<11 ee
Any complication 6456 (1.46) 1321 (2.65) 1.81 (1.71-1.92) <.001 495 (2.23) 1.53 (1.40-1.67) <.001
Bolded values reached statistical significance.
AKI, acute kidney injury; CI, confidence interval; DVT, deep-vein thrombosis; UTI, urinary tract infection.
Table 3. 90-Day Complication Rates Following Isolated Meniscectomies Identified in MarketScan and PearlDiver
90-Day Complications
MarketScan PearlDiver PearlDiver <65
n (%) n (%) RR (95% CI) PValue n (%) RR (95% CI) PValue
Infection 1462 (0.33) 288 (0.58) 1.74 (1.53-1.97) <.001 120 (0.54) 1.63 (1.36-1.97) <.001
DVT 3967 (0.90) 564 (1.13) 1.26 (1.15-1.37) <.001 243 (1.10) 1.22 (1.07-1.39) .003
AKI 165 (0.04) 207 (0.41) 11.09 (9.04-13.60) <.001 66 (0.30) 7.97 (5.99-10.60) <.001
Cardiac arrest 35 (0.01) 19 (0.04) 4.80 (2.74-8.38) <.001 <11 ee
Pneumonia 1256 (0.28) 349 (0.70) 2.46 (2.18-2.76) <.001 138 (0.62) 2.19 (1.84-2.61) <.001
UTI 5793 (1.31) 1534 (3.07) 2.34 (2.21-2.47) <.001 473 (2.14) 1.63 (1.48-1.78) <.001
Wound dehiscence 199 (0.05) 38 (0.08) 1.69 (1.19-2.39) <.001 19 (0.09) 1.90 (1.19-3.04) .01
Hematoma 406 (0.09) 106 (0.21) 2.31 (1.86-2.86) .004 48 (0.22) 2.35 (1.75-3.17) <.001
Nerve injury 95 (0.02) 20 (0.04) 1.86 (1.15-3.01) <.001 <11 ee
Any complication 12950 (2.94) 2823 (5.65) 1.93 (1.85-2.00) .016 1014 (4.58) 1.56 (1.47-1.66) <.001
Comparisons in complication rates for the PearlDiver cohort and the cohort of PearlDiver patients younger than 65 were made against the
MarketScan cohort and reported as relative risk (RR). Bolded values reached statistical significance.
AKI, acute kidney injury; CI, confidence interval; DVT, deep-vein thrombosis; UTI, urinary tract infection.
4M. XIAO ET AL.
such as rotator cuff repair,
26
distal biceps tendon rup-
tures,
27
ulnar collateral ligament reconstruction,
8
articular cartilage lesions,
11
and meniscus procedures.
5
In PearlDiver, the majority of the deidentified bucket
breakdowns are presented in an aggregate form, so it is
more difficult to track patients longitudinally. Adverse
events such as reoperations and complications after
surgery can be tracked, however, and these adverse
events have been studied in the shoulder,
10,14,28
knee,
29
and hip.
13,30
The current investigation used
the Humana claims database within PearlDiver, which
also includes Medicare Advantage patients, explaining
why the majority of the PearlDiver cohort was older
than 65 years.
In comparison, the MarketScan database has become
more widely used for sports medicine research in recent
years, with studies ranging from identifying trends in the
treatment of SLAP lesions,
16
ACL injuries,
18
and shoul-
der instability,
31
to investigating opioid use after shoul-
der arthroscopy
17
and adverse events following
preoperative shoulder injections.
20
The MarketScan
Commercial Database contains claims data for inpatient
and outpatient services, pharmaceutical claims, and
enrollment information for non-Medicare patients
covered with employee-sponsored insurance plans. Each
enrollee is assigned a unique record identifier number
that allows for robust longitudinal tracking over time.
Data are housed in various tables and are queried using a
structured query language. MarketScan provides
patient-level data down to the patient birth year and
provider ZIP Code and variables are similar to those
offered in PearlDiver. The database also provides sepa-
rate datasets including employee-sponsored Medicare
supplemental coverage, Medicaid, and laboratory
records.
Previous database studies using PearlDiver have shed
light on trends in surgical management of meniscus
pathologies, showing significant increases in the num-
ber of meniscus repairs performed between 2005 and
2011, coinciding with an increased emphasis on
meniscus preservation and improved surgical tech-
niques during that time frame.
5
Essilfie et al.
32
also
found that the number of meniscectomies performed in
patients older than 50 years significantly decreased
between 2010 and 2015. Degenerative meniscus tears
increase with age, and up to 56% of men aged 70 to 90
years have incidental findings on magnetic resonance
imaging.
33
Multiple randomized controlled trials have
concluded that outcomes after arthroscopic meniscec-
tomy are no better than sham surgery
34,35
or physical
therapy
36-38
for these degenerative tears. While the
demographic data between MarketScan and PearlDiver
are not similar in large part due to the exclusion of
non-Medicare patients in MarketScan, the patient
population in MarketScan may be more conducive for
studying certain sports medicine injuries or procedures.
A systematic review reported the short-term (0-4
years) reoperation rate following isolated partial
meniscectomies was 1.4%.
39
However, their definition
of a reoperation was a meniscus procedure only,
whereas the current study included more arthroscopic
knee procedures in our definition of a reoperation. We
found vastly differing reoperation rates depending on
the database and if laterality was coded in both the initial
operation and subsequent reoperation. This finding
highlights that reoperations calculated from database
studies may be inaccurate, and MarketScan has a more
limited capacity for reporting re-operations. Only about
5% of patients in the MarketScan database who under-
went a reoperation within 1 year had record of laterality
for the initial meniscectomy and a subsequent proced-
ure, which likely underestimates the true reoperation
percentage. Using MarketScan, we reported an 8.46%
reoperation rate without considering laterality, but this
percentage dropped down to 0.83% calculated with
laterality coding. In PearlDiver, 35% of patient reoper-
ation records included laterality coding, so reoperation
reporting may be more accurate. This difference could be
due to stricter coding guidelines required for supple-
mental Medicare reimbursement compared with private
insurance companies.
To assess rare complications for arthroscopic proced-
ures, a large sample size is necessary. An analysis of
prospectively collected hospital data from the National
Health Service in England investigated the adverse out-
comes following arthroscopic partial meniscectomy in
699,965 patients within a 10-year time frame.
40
The
authors found that serious complications, defined as
Table 4. Comparisons of Strengths and Limitations Between the MarketScan and PearlDiver Databases
Comparison MarketScan PearlDiver
Population size More than 150 million patients 25 million patients
Age All patients younger than 65 years Majority of patients 65 years and older
Region More representative of the United States Heavily favors the Southern United States
Insurance coverage Non-Medicare employee-sponsored plans Humana, including Medicare Advantage
Difficulty Relational database, requires knowledge of SQL, more difficult analysis Simple bucket language, easier analysis
Longitudinal tracking More robust, using patient-level data More difficult, using aggregate data
Laterality coding Very limited, difficult to define reoperations More procedures include laterality coding
Cost $5000-$20,000 per study cohort $25,000-$50,000 per year
SQL, structured query language.
ADMINISTRATIVE DATABASE DIFFERENCES 5
either pulmonary embolism, myocardial infarction,
stroke, infection requiring surgery, fasciotomy, neuro-
vascular injury, or death, within 90-days occurred in
0.317% of patients. This is less than what was found in
either database in the current study, although we
included less serious complications as well. The disparate
rates could be due to different health care systems,
reimbursement incentives, and culture between the
United States and England. Nonetheless, PearlDiver
reported greater complication rates compared with
MarketScan, even when restricting the cohort to those
younger than 65 years. These differences were especially
evident for AKI, pneumonia, and hematoma, where the
RR was greater than 2.
Although the median age for the MarketScan cohort
and PearlDiver <65 years subcohort were comparable,
the average CCI for PearlDiver was significantly greater
than MarketScan. Existing comorbidities increase the
risk for postoperative complications, which may explain
the greater complication rates seen in PearlDiver. In
addition, geographic variations in health show that the
southern region of the United States has a greater prev-
alence of obesity
41
and cardiovascular disease,
42
which
also may account for the increased complications in the
PearlDiver cohort. Humana insurance is headquartered
in Louisville, KY, which may explain the over-
representation of patients from the southern United
States in the PearlDiver cohort. In comparison, the
MarketScan database contains claims from a wide vari-
ety of health insurance plans. Martin et al.
43
studied 30-
day complication rates following any knee arthroscopy
procedure using the prospective National Surgical
Quality Improvement Program database. They found the
overall incidence of having any complication was 1.6%,
comparable with 30-day complications reported from
our MarketScan and PearlDiver <65 years cohorts.
When selecting a claims database to investigate a
sports medicine research question, it is important to
keep in mind the target age demographic. PearlDiver
includes an older patient population due to the inclu-
sion of patients covered under Medicare Advantage
plans, whereas MarketScan contains a larger patient
population, all younger than the age of 65 years.
Further, MarketScan affords better longitudinal patient
tracking capabilities and more patient-level data,
although calculating reoperation rates is highly limited.
The PearlDiver query language may provide an easier
method for gathering and analyzing data compared to
MarketScan. Access to both these databases can also be
costly, with PearlDiver running approximately $25,000
to $50,000 per year and MarketScan costing between
$5,000 to $20,000 per study cohort.
3
Limitations
There are several limitations to the current investiga-
tion. For both databases, individual patient demo
graphics such as height, weight, body mass index,
symptoms, activity level, and occupation were not
available. Further, physical function and patient re-
ported outcomes were unable to be gathered. These
factors may play a role in the decision-making process for
proceeding with isolated meniscectomy. With all claims
databases, the quality of results relies on the accuracy of
diagnosis and procedure code reporting. Neither data-
base includes the entire population of the United States,
and the exclusion of Medicare patients within the Mar-
ketScan cohort limited our available claims population.
Laterality coding was not present for the majority of
patients in both databases. As such, the reoperation rates
calculated may be inaccurate. In addition, there is a large
regional bias in the PearlDiver population, as 59% are
from the Southern region and only about 2% are from
the Northeast. Therefore, the US population may not be
as accurately represented by this database. In either
database, we could not account for the patients who
dropped out of their respective health plans, but we did
ensure to only include patients who were enrolled
continuously for 1 year. Finally, with large sample sizes
in these databases, statistical significance is able to be
reached with great power. However, statistically
significant does not necessarily translate into clinical
significance, and it is unknown which database confers
more accurate re-operation and complication rates.
Conclusions
For those undergoing isolated meniscectomy, the
MarketScan database comprised an overall larger and
younger cohort of patients with fewer comorbidities,
even when we examined only subjects younger than
65 years of age.
References
1. Karlson NW, Nezwek TA, Menendez ME, Tybor D,
Salzler MJ. Increased utilization of american administra-
tive databases and large-scale clinical registries in ortho-
paedic research, 1996 to 2016. J Am Acad Orthop Surg Glob
Res Rev 2018;2:e076.
2. Pugely AJ, Bozic KJ. Editorial Commentary: Rising in-
terest in “big data”in arthroscopy: Is the juice worth the
squeeze? Arthroscopy 2017;33:232-233.
3. Pugely AJ, Martin CT, Harwood J, Ong KL, Bozic KJ,
Callaghan JJ. Database and registry research in ortho-
paedic surgery: Part I: Claims-based data. J Bone Joint Surg
Am 2015;97:1278-1287.
4. Bohl DD, Singh K, Grauer JN. Nationwide databases in
orthopaedic surgery research. J Am Acad Orthop Surg
2016;24:673-682.
5. Abrams GD, Frank RM, Gupta AK, Harris JD,
McCormick FM, Cole BJ. Trends in meniscus repair and
meniscectomy in the United States, 2005-2011. Am J
Sports Med 2013;41:2333-2339.
6. Cancienne JM, Brockmeier SF, Carson EW, Werner BC.
Risk factors for infection after shoulder arthroscopy in a
6M. XIAO ET AL.
large medicare population. Am J Sports Med 2018;46:
809-814.
7. Kew ME, Cancienne JM, Christensen JE, Werner BC. The
timing of corticosteroid injections after arthroscopic
shoulder procedures affects postoperative infection risk.
Am J Sports Med 2019;47:915-921.
8. Erickson BJ, Nwachukwu BU, Rosas S, et al. Trends in
medial ulnar collateral ligament reconstruction in the
United States. Am J Sports Med 2015;43:1770-1774.
9. Iglinski-Benjamin KC, Xiao M, Safran MR, Abrams GD.
Increased prevalence of concomitant psychiatric diagnoses
among patients undergoing hip arthroscopic surgery.
Orthop J Sports Med 2019;7:232596711882245.
10. Leroux TS, Saltzman BM, Sumner SA, et al. Elective
shoulder surgery in the opioid naïve: Rates of and risk
factors for long-term postoperative opioid use. Am J Sports
Med 2019;47:1051-1056.
11. McCormick F, Harris JD, Abrams GD, et al. Trends in the
surgical treatment of articular cartilage lesions in the
United States: An analysis of a large private-payer database
over a period of 8 years. Arthroscopy 2014;30:222-226.
12. Qin C, Curtis DM, Reider B, Shi LL, Lee MJ, Athiviraham A.
Orthopaedic shoulder surgery in the ambulatory surgical
center: Safety and outcomes. Arthroscopy 2019;35:
2545-2550.e2541.
13. Truntzer JN, Hoppe DJ, Shapiro LM, Abrams GD,
Safran M. Complication rates for hip arthroscopy are
underestimated: A population-based study. Arthroscopy
2017;33:1194-1201.
14. Xiao M, Abrams GD. Increased reoperation rates among
patients undergoing shoulder arthroscopy with concomi-
tant biceps tenodesis. JSES Open Access 2019;3:344-349.
15. Bonazza NA, Homcha B, Liu G, Leslie DL, Dhawan A.
Surgical trends in arthroscopic hip surgery using a large
national database. Arthroscopy 2018;34:1825-1830.
16. Dougherty MC, Kulenkamp JE, Boyajian H, Koh JL,
Lee MJ, Shi LL. National trends in the diagnosis and repair
of SLAP lesions in the United States. J Orthop Surg
2020;28:230949901988855.
17. Gil JA, Gunaseelan V, Defroda SF, Brummett CM, Bedi A,
Waljee JF. Risk of prolonged opioid use among opioid-
naïve patients after common shoulder arthroscopy pro-
cedures. Am J Sports Med 2019;47:1043-1050.
18. Herzog MM, Marshall SW, Lund JL, Pate V, Mack CD,
Spang JT. Trends in incidence of ACL reconstruction and
concomitant procedures among commercially insured
individuals in the United States, 2002-2014. Sports Health
2018;10:523-531.
19. Jacobs CA, Hawk GS, Jochimsen KN, et al. Depression
and anxiety are associated with increased health care costs
and opioid use for patients with femoroacetabular
impingement undergoing hip arthroscopy: Analysis of a
claims database. Arthroscopy 2020;36:745-750.
20. Traven SA, Brinton D, Simpson KN, et al. Preoperative
shoulder injections are associated with increased risk of
revision rotator cuff repair. Arthroscopy 2019;35:706-713.
21. Alluri RK, Leland H, Heckmann N. Surgical research using
national databases. Ann Transl Med 2016;4:393.
22. Patel AA, Singh K, Nunley RM, Minhas SV. Administra-
tive databases in orthopaedic research: Pearls and pitfalls
of big data. J Am Acad Orthop Surg 2016;24:172-179.
23. Molina CS, Thakore RV, Blumer A, Obremskey WT,
Sethi MK. Use of the National Surgical Quality Improve-
ment Program in orthopaedic surgery. Clin Orthop Rel Res
2015;473:1574-1581.
24. Charlson ME, Pompei P, Ales KL, Mackenzie CR. A new
method of classifying prognostic comorbidity in longitu-
dinal studies: Development and validation. J Chronic Dis
1987;40:373-383.
25. Weinreb JH, Yoshida R, Cote MP, O’Sullivan MB,
Mazzocca AD. A Review of Databases Used in Orthopae-
dic Surgery Research and an Analysis of Database Use in
Arthroscopy: The Journal of Arthroscopic and Related
Surgery. Arthroscopy 2017;33:225-231.
26. Zhang AL, Montgomery SR, Ngo SS, Hame SL, Wang JC,
Gamradt SC. Analysis of rotator cuff repair trends in a
large private insurance population. Arthroscopy 2013;29:
623-629.
27. Kelly MP, Perkinson SG, Ablove RH, Tueting JL. Distal
biceps tendon ruptures. Am J Sports Med 2015;43:
2012-2017.
28. Erickson BJ, Basques BA, Griffin JW, et al. The effect of
concomitant biceps tenodesis on reoperation rates after
rotator cuff repair: A review of a large private-payer
database from 2007 to 2014. Arthroscopy 2017;33:
1301-1307.e1301.
29. Cancienne JM, Kew ME, Smith MK, Carson EW,
Miller MD, Werner BC. The timing of corticosteroid in-
jections following simple knee arthroscopy is associated
with infection risk. Arthroscopy 2019;35:1688-1694.
30. Khazi ZM, An Q, Duchman KR, Westermann RW. Inci-
dence and risk factors for venous thromboembolism
following hip arthroscopy: A population-based study.
Arthroscopy 2019;35:2380-2384.e2381.
31. Bonazza NA, Liu G, Leslie DL, Dhawan A. Trends in
surgical management of shoulder instability. Orthop J
Sports Med 2017;5:232596711771247.
32. Essilfie A, Kang HP, Mayer EN, Trasolini NA, Alluri RK,
Weber AE. Are orthopaedic surgeons performing fewer
arthroscopic partial meniscectomies in patients greater
than 50 years old? A national database study. Arthroscopy
2019;35:1152-1159.e1151.
33. Englund M, Guermazi A, Gale D, et al. Incidental meniscal
findings on knee MRI in middle-aged and elderly persons.
N Engl J Med 2008;359:1108-1115.
34. Sihvonen R, Paavola M, Malmivaara A, et al. Arthroscopic
partial meniscectomy versus placebo surgery for a degen-
erative meniscus tear: A 2-year follow-up of the rando-
mised controlled trial. Ann Rheum Dis 2018;77:188-195.
35. Sihvonen R, Paavola M, Malmivaara A, et al. Arthroscopic
partial meniscectomy versus sham surgery for a degen-
erative meniscal tear. N Engl J Med 2013;369:2515-2524.
36. Kise NJ, Risberg MA, Stensrud S, Ranstam J, Engebretsen L,
Roos EM. Exercise therapy versus arthroscopic partial
meniscectomy for degenerative meniscal tear in middle
aged patients: randomised controlled trial with two year
follow-up. BMJ 2016:i3740.
37. Katz JN, Brophy RH, Chaisson CE, et al. Surgery versus
physical therapy for a meniscal tear and osteoarthritis.
N Engl J Med 2013;368:1675-1684.
38. Yim J-H, Seon J-K, Song E-K, et al. A comparative study
of meniscectomy and nonoperative treatment for
ADMINISTRATIVE DATABASE DIFFERENCES 7
degenerative horizontal tears of the medial meniscus. Am
J Sports Med 2013;41:1565-1570.
39. Paxton ES, Stock MV, Brophy RH. Meniscal repair versus
partial meniscectomy: A systematic review comparing
reoperation rates and clinical outcomes. Arthroscopy
2011;27:1275-1288.
40. Abram SGF, Judge A, Beard DJ, Price AJ. Adverse out-
comes after arthroscopic partial meniscectomy: A study of
700 000 procedures in the national Hospital Episode Sta-
tistics database for England. Lancet 2018;392:2194-2202.
41. Ward ZJ, Long MW, Resch SC, et al. Redrawing the
US obesity landscape: Bias-corrected estimates of
state-specific adult obesity prevalence. PLoS One
2016;11:e0150735.
42. Gebreab SY, Davis SK, Symanzik J, Mensah GA,
Gibbons GH, Diez-Roux AV. Geographic variations in
cardiovascular health in the United States: Contributions
of state- and individual-level factors. J Am Heart Assoc
2015;4:e001673-e001673.
43. Martin CT, Pugely AJ, Gao Y, Wolf BR. Risk factors for
thirty-day morbidity and mortality following knee
arthroscopy: A review of 12,271 patients from the na-
tional surgical quality improvement program database.
J Bone Joint Surg Am 2013;95:e98 91-10.
8M. XIAO ET AL.
Appendix
Appendix Table 1. CPT Codes and Descriptions for All Knee Procedures Used to Analyze Reoperations and to Establish Isolated
Meniscectomy Patients in This Study
CPT Code Description
27310 Arthrotomy, knee, with exploration, drainage, or removal of foreign body (e.g., infection)
29871 Arthroscopy, knee, surgical; for infection, lavage, and drainage
27570 Manipulation of knee joint under general anesthesia
27457 Osteotomy, proximal tibia, including fibular excision or osteotomy; after epiphyseal closure
27418 Anterior tibial tubercleplasty (e.g., Maquet-type procedure)
29866 Arthroscopy, knee, surgical; osteochondral autograft(s) (e.g., mosaicplasty) (includes harvesting of the autograft(s))
29867 Arthroscopy, knee, surgical; osteochondral autograft(s) (eg, mosaicplasty)
29868 Arthroscopy, knee, surgical; meniscal transplantation (includes arthrotomy for meniscal insertion), medial or lateral
29870 Arthroscopy, knee, diagnostic, with or without synovial biopsy (separate procedure)
29873 Arthroscopy, knee, surgical; with lateral release
29874 Arthroscopy, knee, surgical; for removal of loose body or foreign body (e.g., osteochondritis dissecans fragmentation, chondral
fragmentation)
29875 Arthroscopy, knee, surgical; synovectomy, limited (e.g., plica or shelf resection) (separate procedure)
29876 Arthroscopy, knee, surgical; synovectomy, major, 2 or more compartments (e.g., medial or lateral)
29877 Arthroscopy, knee, surgical; debridement/shaving of articular cartilage (chondroplasty)
29879 Arthroscopy, knee, surgical; abrasion arthroplasty (includes chondroplasty where necessary) or multiple drilling or microfracture
29880 Arthroscopy, knee, surgical; with meniscectomy (medial AND lateral, including any meniscal shaving) including debridement/
shaving of articular cartilage (chondroplasty), same or separate compartment(s), when performed
29881 Arthroscopy, knee, surgical; with meniscectomy (medial OR lateral, including any meniscal shaving) including debridement/
shaving of articular cartilage (chondroplasty), same or separate compartment(s), when performed
29882 Arthroscopy, knee, surgical; with meniscus repair (medial OR lateral)
29883 Arthroscopy, knee, surgical; with meniscus repair (medial AND lateral)
29884 Arthroscopy, knee, surgical; with lysis of adhesions, with or without manipulation (separate procedure)
29888 Arthroscopically aided anterior cruciate ligament repair/augmentation or reconstruction
29889 Arthroscopically aided posterior cruciate ligament repair/augmentation or reconstruction
27405 Repair, primary, torn ligament and/or capsule, knee; collateral
27427 Ligamentous reconstruction (augmentation), knee; extra-articular
Appendix Table 2. International Classification of Diseases 9th
Revision (ICD-9) Codes for Complications
Event ICD-9 Code
Infection 686.8, 696.9, 730.86, 730.89, 730.96,
730.99,996.60, 996.67, 998.59, 998.51,
996.66
Urinary tract infection 599
Pneumonia 480.0-486.0
Acute kidney injury 584.50-583.90
Hematoma 998.11, 998.12, 998.13
Deep-vein thrombosis 451.19, 453.82, 453.2, 453.3, 453.40-453.42
Cardiac arrest 427.41, 427.50
Wound dehiscence 998.30, 998.31, 992.32, 998.33
Nerve injury 355.3, 355.4, 907.5, 907.9, 956.2, 956.3,
956.4, 956.5, 956.8, 956.9
ADMINISTRATIVE DATABASE DIFFERENCES 8.e1
