Ricciardi et al. - 2017 - Patient and Perioperative Variables Affecting 30-D.pdf

Accepted Manuscript
Patient and perioperative variables affecting 30-day readmission for surgical
complications following hip and knee arthroplasty: a matched cohort study
Benjamin F. Ricciardi, MD, Kathryn K. Oi, BA, Steven B. Daines, MD, Yuo-Yu Lee,
MS, Amethia D. Joseph, BA, Geoffrey H. Westrich, MD
PII: S0883-5403(16)30746-X
DOI: 10.1016/j.arth.2016.10.019
Reference: YARTH 55456
To appear in: The Journal of Arthroplasty
Received Date: 20 January 2016
Revised Date: 8 October 2016
Accepted Date: 13 October 2016
Please cite this article as: Ricciardi BF, Oi KK, Daines SB, Lee Y-Y, Joseph AD, Westrich GH, Patient
and perioperative variables affecting 30-day readmission for surgical complications following hip
and knee arthroplasty: a matched cohort study, The Journal of Arthroplasty (2016), doi: 10.1016/
j.arth.2016.10.019.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

M
A
N
U
S
C
R
I
P
T
A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
Patient and perioperative variables affecting 30-day readmission for
surgical complications following hip and knee arthroplasty: a matched cohort study
Benjamin F. Ricciardi, MD
Kathryn K. Oi, BA
Steven B. Daines, MD
Yuo-Yu Lee, MS
Amethia D. Joseph, BA
Geoffrey H. Westrich, MD
Institution Affiliation:
Hospital for Special Surgery
535 East 70th Street
New York, NY 10021
Corresponding Author:
Kathryn K. Oi
Hospital for Special Surgery
535 East 70th Street
New York, NY 10021
(212) 606-1959
oik@hss.edu

M
A
N
U
S
C
R
I
P
T
A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
1
2
3
4
5
6
7
8
9
Patient and perioperative variables affecting 30-day readmission for surgical complications
10
following hip and knee arthroplasty: a matched cohort study
11

M
A
N
U
S
C
R
I
P
T
A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
2
Patient and perioperative variables affecting 30-day readmission for surgical complications following
12
hip and knee arthroplasty: a matched cohort study
13
14
ABSTRACT:
15
Background: Changes in reimbursement for total hip (THA) and knee (TKA) arthroplasty have placed
16
increased financial burden of early readmission on hospitals and surgeons. Our purpose was to
17
characterize factors of 30-day readmission for surgical complications after THA and TKA at a single, high-
18
volume orthopedic specialty hospital.
19
Methods: Patients with a diagnosis of OA and who were readmitted within 30 days of their unilateral
20
primary THA or TKA procedure between 2010 and 2014. Readmitted patients were matched to non-
21
readmitted patients 1:2. Patient and perioperative variables were collected for both cohorts. A
22
conditional logistic regression was performed to assess both the patient and perioperative factors and
23
their predictive value toward 30-day readmission.
24
Results: 21,864 arthroplasties (THA=11,105; TKA=10,759) were performed between 2010 and 2014 at
25
our institution, in which 60 patients (THA=37, TKA=23) were readmitted during this 5-year period. The
26
most common reasons for readmission were fracture (N=14), infection (N=14), and dislocation (N=9).
27
30-day readmission for THA was associated with increased procedure time (p=0.05), LOS shorter than 2
28
days (p=0.04), discharge to a skilled nursing facility (p=0.05), and anticoagulation use other than aspirin
29
(p=0.02). 30-day readmission for TKA was associated with increased tourniquet time (p=0.02), LOS <3
30
days (p<0.01), and preoperative depression (p=0.02). In the combined THA/TKA model, a diagnosis of
31
depression increased 30-day readmission [OR 3.5 (1.4-8.5); p<0.01].
32
Conclusions: Risk factors for 30-day readmission for surgical complications included short LOS,
33
discharge destination, increased procedure/tourniquet time, potent anticoagulation use, and
34
preoperative diagnosis of depression. A focus on risk factor modification and improved risk
35
stratification models are necessary to optimize patient care using readmission rates as a quality
36
benchmark.
37
38
Keywords: 30-day Readmission; Total Knee Arthroplasty; Total Hip Arthroplasty; Comprehensive Care
39
for Joint Replacement; Risk Factors
40
41
42

M
A
N
U
S
C
R
I
P
T
A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
3
INTRODUCTION:
43
Healthcare reform in the United States (US) over the past decade has attempted to improve quality of
44
care and accountability of providers while controlling costs for payers. Readmissions in the first 30 days
45
represent a significant burden to the healthcare system, and many of these reforms have targeted a
46
reduction in hospital readmission rates across various medical and surgical diagnoses including total
47
knee (TKA) and total hip (THA) arthroplasty.1
In association with changes implemented as a part of the
48
Affordable Care Act in 2010, the Centers for Medicare & Medicaid Services (CMS) penalizes hospitals
49
financially for high readmission rates relative to the national average and pass on the financial burden of
50
30-day readmissions to hospitals in the form of penalties and non-payments.2
Additionally, the
51
implementation of Comprehensive Care for Joint Replacement (CCJR) by CMS, will place the burden of
52
readmission even further on providers and hospitals.
53
54
Understanding modifiable and non-modifiable risk factors for 30-day readmission after TKA and THA
55
may give providers an opportunity to improve their short-term quality of care and risk stratification
56
within their treatment populations. Previous studies at the institutional level have found payer status,
57
race, sex, medical comorbidities, discharge disposition, and length of stay to be independently
58
associated with 30-day readmission depending on the institution examined.2-6
At the payer level, race,
59
insurance status, hospital volume, discharge disposition, and medical comorbidities have been reported
60
as independent factors for readmission after THA and TKA.7-10
Studies of risk factors for 30-day
61
readmission at tertiary referral institutions for THA and TKA would provide further data on modifiable
62
patient characteristics to help reduce complication rates and non-modifiable characteristics to be used
63
as tools for risk stratification in predictive models for complications. The purpose of our study is to 1)
64
describe the reasons for readmission for surgical complications and 2) characterize patient and
65

M
A
N
U
S
C
R
I
P
T
A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
4
perioperative factors resulting in 30-day readmission for surgical complications after THA and TKA at a
66
single high volume orthopedic specialty hospital.
67
METHODS:
68
Patient Cohort
69
We retrospectively reviewed all patients who underwent primary unilateral THA or TKA with a diagnosis
70
of osteoarthritis (OA) from January 2010 to December 2014 at a single orthopedic specialty hospital that
71
serves as a tertiary referral center institution for TKA and THA. Patients who were readmitted directly to
72
our facility within 30 days of their index procedure were identified from administrative claims data and
73
confirmed through an institutional registry for THA and TKA.
74
75
Exclusion criteria included patients undergoing revision arthroplasty or those undergoing THA or TKA for
76
a diagnosis of hip dysplasia, avascular necrosis, inflammatory disease, inflammatory arthropathy,
77
rheumatoid arthritis, posttraumatic arthritis or fracture. Planned readmissions for unrelated surgery
78
were also excluded, as a planned readmission is part of the same episode of care of the index
79
procedure. Although most patients in the cohort had THA or TKA done on just one side during the study
80
time frame, there were patients who had both sides of the hip or knee replaced at two different time
81
points or had both THA and TKA at two different time points. For these patients who had more than one
82
operation without subsequent readmissions, both operations were included in the cohort. For patients
83
who had more than one operation and had readmission for one of the operations, the second operation
84
was excluded if the readmission was associated with the first operation. If the readmission was
85
associated with the second operation, the first operation was not excluded.
86
87

M
A
N
U
S
C
R
I
P
T
A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
5
Study Variables
88
Patient factors including age, sex, race, BMI, Deyo-Charlson comorbidity index (0, 1-2, 3+), list of
89
Elixhauser's comorbid conditions11
, and characteristics associated with the index procedure were
90
retrieved from electronic health records. Perioperative factors at index procedure were retrieved from
91
our electronic medical record including transfusion rates, tranexamic acid use, anesthesia type, drain
92
use, periarticular injections, and ASA class (Allscripts, Chicago, IL). Comorbidities that were compared
93
included congestive heart failure, valvular disease, peripheral vascular disease, neurological disorders,
94
chronic pulmonary disease, diabetes (with and without chronic complications), hypothyroidism, renal
95
failure, liver disease, coagulopathy, obesity, fluid and electrolyte disorders, anemias, depression and
96
hypertension. Data regarding the readmission was identified using post-discharge records.
97
Matching
98
Readmitted patients were matched 1 to 2 to non-readmitted patients on a set of predefined covariates
99
to control for confounding. The covariates included age (+/-5), sex (exact), Deyo Charlson comorbidity
100
index (exact), and date of surgery (+/- 30 days). Patient characteristics, comorbidities, procedure times,
101
and length of stay (LOS) were compared between the matched readmitted and non-readmitted pairs.
102
TKA and THA patients were analyzed as individual cohorts and a combined cohort against their
103
respective non-readmitted matching cohorts.
104
Statistical Analysis
105
Continuous variables were summarized as means ± standard deviations and compared between
106
readmitted and non-readmitted patients using two-sample t-test. Categorical variables were presented
107
in frequencies and percentages and compared using Chi-square tests or Fisher Exact tests (when
108
expected values n < 5 in any field). Subsequently, a conditional logistic regression was performed to
109

M
A
N
U
S
C
R
I
P
T
A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
6
account for matching nature of the data in identifying risk factors of readmission. Variables that were
110
significantly different between the readmitted and non-readmitted patients in the univariate analysis
111
were included as covariates in the subsequent multivariate analysis to identify the risk factors of
112
readmission. Due to the low sample size of this study, we chose the most parsimonious model according
113
to the Akaike’s information criterion (AIC), a measure that accounts for both model fit and model
114
complexity parameters.12
All analyses were performed using SAS v9.3 (SAS Institute Inc., Cary, NC). All
115
tests were two-sided with a significance level of α=0.05.
116
117
RESULTS:
118
Patient Demographics
119
A total volume of 21,864 patients underwent primary total hip or knee arthroplasty at our institution
120
(THA=11,105; TKA=10,759) between 2010 and 2014. The entire cohort was 58.8% female (n=12,866),
121
the average age at time of surgery was 65.8+/-10.4 years and a BMI average of 29.3+/-6.1kg/m2
. 73.8%
122
reported 0 comorbidities, 23.6% reported 1-2 and 2.6% reported 3 or more (see tables 1 and 2 for
123
complete summary of basic demographics and comorbidities for which we screened).
124
60 patients were admitted to our institution within the first 30 days of their index procedure (THA=37,
125
TKA=23). The readmitted cohort averaged 66.1 +/-9.4 years of age and was 53.3% (n=32) female. Two-
126
thirds (n=40) reported no comorbidities and one third reported 1-2 comorbidities. The most common
127
reasons for readmission from this small subset were fracture (THA=12, TKA=2), infection (THA=7,
128
TKA=7), and dislocation (THA=9, TKA=0). See Table 3 for complete list of reasons for 30-day readmission.
129
Matching
130

M
A
N
U
S
C
R
I
P
T
A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
7
After matching, a balanced cohort was created with the readmitted patients (N=60) as the observational
131
group and the non-readmitted patients as the control group (N=120). No significant differences were
132
found between groups on the matching variables (age, number of comorbidities, sex, laterality, and
133
joint), mitigating any potential confounding variables between the cohorts in our analysis (see Table 4).
134
Significant Variables
135
THA
136
Patients readmitted following primary THA had increased procedure times (77.7 +/-18.7 min vs 71.3 +/-
137
14.7 min; p=0.05). There were more readmitted patients with LOS<2 days (89.2% vs 98.6%; p=0.04).
138
Patients readmitted following hip arthroplasty were less likely to have been treated with aspirin as their
139
anticoagulant (45.9% vs 24.3%; p=0.02) and more likely to be discharged to a nursing facility (18.9% vs
140
6.8%; p=0.05). [Table 5] Chronic pulmonary disease was more common in the readmitted group (18.9%
141
vs 5.4%; p=0.04). There was a non-significant trend of depression in those readmitted (18.9% vs 6.8%;
142
p=0.10). [Table 6]
143
TKA
144
Patients readmitted following primary TKA had both longer procedure and tourniquet times. Procedure
145
time for those readmitted averaged 95.3 +/-29.1 minutes versus 81.1 +/- 13.5 minutes (p=0.04).
146
Tourniquet times for readmitted TKA patients averaged 64.6 +/- 38.7 minutes compared to their non-
147
readmitted TKA patients’ average of 43.6 minutes +/- 15.5 (p=0.02). There were more readmitted
148
patients with LOS less than 3 days (78.3% vs 100%; p<0.01). [Table 7] There was a non-significant trend
149
of hypothyroidism in those readmitted (34.8% vs 13%; p=0.06). Depression was more prevalent in those
150
readmitted (34.8% vs 10.9%; p=0.02). [Table 8]
151
THA and TKA Combined
152

M
A
N
U
S
C
R
I
P
T
A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
8
Readmitted patients had significantly longer procedure time (84.4 minutes +/- 24.6 vs 75.0 minutes +/-
153
15.0; p=0.01) and more likely to have a LOS less than two days (93.3% vs 99.2%; p=0.04). A pre-existing
154
diagnosis of depression was also associated with 30-day readmission (25% vs 8.3%; p<0.01). There was a
155
trend toward increased use of an anticoagulant other than aspirin in the readmitted patients (61.7% vs
156
47.5%); however, the differences did not reach statistical significance (p=0.08). No differences were
157
found between readmitted and non-readmitted cohorts in the following perioperative factors:
158
transfusion rates, tranexamic acid use, anesthesia type, drain use, periarticular injections, and ASA class.
159
[See tables 9 and 10]
160
Conditional Logistic Regression
161
The results from the conditional logistic regression showed, for TKA and THA patients combined,
162
depression was a significant predictor for readmission – patients with a diagnosis of depression were 3.5
163
times more likely to be readmitted [OR 3.5 (1.4-8.5); p<0.01]. Depression trended toward increased
164
readmission risk in TKA patients [OR 3.4 (0.7-15.6); p=0.10] and in THA patients [OR 3.2 (0.9-10.8);
165
p=0.07] when analyzed individually. TKA patients with increased tourniquet times were more likely to
166
be readmitted – 3.5% increase in likelihood of being readmitted for every minute increase in the
167
tourniquet time [OR 1.04 (1.0-1.7); p=0.02]. [See Table 11]
168
DISCUSSION:
169
Many studies have attempted to characterize risk factors influencing 30-day readmission after TKA and
170
THA at both the institutional and payer level.2-10, 13-16
Such efforts provide targets for improved protocols
171
and procedures that may reduce readmissions and improve patient outcomes. In our single orthopedic
172
specialty center cohort, the most common reasons for 30-day readmission for surgical complications
173
after THA or TKA were periprosthetic fracture, periprosthetic infection, and dislocation. Risk factors for
174
30-day readmission after THA were longer operative times, shorter length of stay (<2 days),
175

M
A
N
U
S
C
R
I
P
T
A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
9
anticoagulation other than aspirin, and discharge to skilled nursing facility. After TKA, longer procedure
176
times, increased tourniquet times, shorter length of stay, and a diagnosis of depression preoperatively
177
were associated with 30-day readmissions. In our combined model of both TKA and THA, the
178
preoperative diagnosis of depression was associated with 30-day readmission in patients undergoing
179
THA or TKA.
180
181
Studies of TKA or THA outcomes have also found periprosthetic fracture, surgical site infection, and hip
182
dislocation to be common causes of early readmission. In the Kaiser Permanente Total Joint
183
Replacement Registry, periprosthetic infection and hip dislocation were the most common readmissions
184
in their health system after THA.9
Kurtz et al. had similar findings in the Medicare population with
185
dislocation and surgical site infection being the most common causes of readmission.8
At an
186
institutional level, Schairer et al. found that dislocation and infection were also significant contributors
187
to surgical readmissions in their THA population.15
In our study, medical complications such as cardiac
188
disease or pulmonary complications were not reasons for readmission. Medical readmissions for
189
conditions such as myocardial infarction, venothromboembolic disease or pulmonary related
190
complications tend to occur at outside institutions with more broad treatment capabilities than our
191
specialty hospital, which limits our data in assessing the risk factors for medical readmission. Previous
192
authors have found surgical complications to be more common than medical causes of readmissions
193
after TKA and THA, however, suggesting that addressing surgical readmissions would result in a
194
substantial decrease in overall 30-day readmission rates.8, 15
195
196
Previous studies at the institutional and payer levels have described a number of modifiable and non-
197
modifiable risk factors for 30-day readmission after THA or TKA. The most consistent risk factors are
198
discharge destination, significant medical comorbidities, and longer length of stay at index admission.3-
199

M
A
N
U
S
C
R
I
P
T
A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
10
5,15
Our data is consistent with many of these findings. Discharge to a skilled nursing facility (SNF) after
200
THA appears to be a major risk factor for readmission across most studies and may represent a
201
modifiable risk factor. Some of this association may be due to more frail patients being discharge to a
202
SNF postoperatively, however, Bini et al. found an odds ratio for 90-day readmission of 1.6 to 1.9 in
203
healthy patients discharged to a skilled nursing facility when controlling for age, sex, and ASA.13
The
204
performance of SNFs are variable, and associations between SNF quality and hospital readmission may
205
suggest that characteristics directly associated with the SNF can have some influence on these findings
206
across institutions.17
Increasing rates of home discharge when safe for the patient may reduce 30-day
207
readmission rates. It is important to note that length of stays shorter than our anticipated rapid
208
recovery pathways (<2 days for THA and <3 days for TKA) were associated with 30-day readmission. The
209
use of rapid recovery pathways has not found to increase complications rates in THA or TKA across may
210
institutions, however, it is possible that our current pathways have not been optimized for lengths of
211
stay less than expected.18, 19
The balance between a safe home discharge and reducing length of stay
212
needs to be reached through refinement of clinical pathways and patient-physician interaction and
213
continued studies are needed to optimize this process. Non-modifiable factors such as race, insurance
214
status, hospital/surgeon volume have been associated with 30-day readmissions at a payer level and
215
could not be assessed in our study due to small sample sizes and a more uniform patient population at
216
our institution.8-10
217
218
Other potentially modifiable risk factors that were associated with 30-day readmission were depression,
219
tourniquet time for TKA, and use of anticoagulation other than aspirin in THA. Depression has been
220
identified as a risk factor for complications after THA and TKA.8, 16, 20
Gold et al. found depression to
221
increase the risk of 90-day readmission following THA or TKA independent of other comorbid
222
conditions.16
Complications such as periprosthetic facture, infection, and dislocation were increased in
223

M
A
N
U
S
C
R
I
P
T
A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
11
patients with psychiatric diagnoses such as depression.20
Depression may directly increase rates of
224
readmission through factors such as inability to appropriately care for oneself at home or be associated
225
with other factors that have not been accounted for in these multivariable models. It is unclear if it is a
226
modifiable risk factor preoperatively, but at the very least, incorporation into risk stratification models
227
would be warranted. The impact of tourniquet use in TKA is controversial. Rathod et al. found that using
228
a tourniquet only for cementation resulted in decreased early complication rates in TKA primarily
229
because of lower manipulation rates and pulmonary emboli.21
Other randomized controlled trials have
230
found a possible association with delayed quadriceps recovery with tourniquet use, however, total
231
blood loss and complications were not increased.22
It is also possible that increased tourniquet time
232
represented a more complex procedure or intraoperative difficulty that predisposed the patients to 30-
233
day readmission. Further studies are needed to assess the relationship of tourniquet time and
234
readmission. In THA, readmission was more common when using anticoagulants other than aspirin.
235
The downside of potent anticoagulant use includes increased rates of surgical site hematoma and major
236
bleeding.23
Risk stratification protocols that identify and treat only the highest risk patients with potent
237
anticoagulation and utilize aspirin for lower risk patients can keep the rate of complications from deep
238
vein thrombosis low while avoiding some of the complications of overtreating lower risk patients.24, 25
239
The use of anticoagulation at our hospital is at the discretion of the operating surgeon, and some
240
surgeons follow standardized risk stratification protocols while others prefer more potent
241
anticoagulation for all patients. It is also possible that the use of potent anticoagulation increased
242
readmissions due to an association with increased patient comorbidities receiving these agents. Further
243
studies with greater sample size are necessary to assess this possibility.
244
245

M
A
N
U
S
C
R
I
P
T
A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
12
There are some limitations to our study. As an orthopedic specialty, tertiary referral hospital that
246
frequently sees patients from a wide geographic region, some of our patients may reside at a distance
247
from our institution. Additionally, our location in a tristate area makes it difficult to track readmissions
248
in neighboring states outside of the Medicare population. Most patients readmitted to our hospital are
249
admitted for surgical complications such as infection, hip dislocation or fracture. Medical readmissions
250
for conditions such as myocardial infarction, or pulmonary related complications tend to occur at
251
outside institutions with a more broad treatment scope than our specialty hospital. This makes our data
252
reflective of readmission for surgical complications, but not medical readmissions. Our results may not
253
be readily generalizable to other dissimilar institutions. CMS reported our 30-day readmission rate for
254
patients in the Medicare population as better than national average, however, many patients in our
255
study are not in the Medicare population and readmissions to outside institutions is impossible to
256
quantify. Despite these limitations, our findings on 30-day readmissions for surgical complications are
257
reflective of other institutions’ experiences and confirm previous studies providing further evidence for
258
modifiable risk factors to target for quality improvement and risk stratification.
259
260
CONCLUSION:
261
30-day readmission for THA was associated with increased procedure time, short LOS, discharge to a
262
skilled nursing facility, and anticoagulation use other than aspirin. 30-day readmission for TKA was
263
associated with increased tourniquet time, short LOS, and preoperative depression. In the combined
264
THA/TKA model, a diagnosis of depression increased 30-day readmission. A focus on risk factor
265
modification and improved risk stratification models are necessary to optimize patient care using
266
readmission rates as a quality benchmark.
267
268

M
A
N
U
S
C
R
I
P
T
A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
13
269
270
References
271
1. Jencks SF, Williams MV and Coleman EA. Rehospitalizations among patients in the Medicare fee-
272
for-service program. N Engl J Med. 2009;360:1418-28.
273
2. Clement RC, Derman PB, Graham DS, Speck RM, Flynn DN, Levin LS and Fleisher LA. Risk factors,
274
causes, and the economic implications of unplanned readmissions following total hip arthroplasty. J
275
Arthroplasty. 2013;28:7-10.
276
3. Schairer WW, Sing DC, Vail TP and Bozic KJ. Causes and frequency of unplanned hospital
277
readmission after total hip arthroplasty. Clin Orthop Relat Res. 2014;472:464-70.
278
4. Mesko NW, Bachmann KR, Kovacevic D, LoGrasso ME, O'Rourke C and Froimson MI. Thirty-day
279
readmission following total hip and knee arthroplasty - a preliminary single institution predictive model.
280
J Arthroplasty. 2014;29:1532-8.
281
5. Tayne S, Merrill CA, Smith EL and Mackey WC. Predictive risk factors for 30-day readmissions
282
following primary total joint arthroplasty and modification of patient management. J Arthroplasty.
283
2014;29:1938-42.
284
6. Bosco JA, 3rd, Karkenny AJ, Hutzler LH, Slover JD and Iorio R. Cost burden of 30-day
285
readmissions following Medicare total hip and knee arthroplasty. J Arthroplasty. 2014;29:903-5.
286
7. Kurtz SM, Lau EC, Ong KL, Adler EM, Kolisek FR and Manley MT. Which Hospital and Clinical
287
Factors Drive 30- and 90-Day Readmission After TKA? J Arthroplasty. 2016.
288
8. Kurtz SM, Lau EC, Ong KL, Adler EM, Kolisek FR and Manley MT. Hospital, Patient, and Clinical
289
Factors Influence 30- and 90-Day Readmission After Primary Total Hip Arthroplasty. J Arthroplasty. 2016.
290

M
A
N
U
S
C
R
I
P
T
A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
14
9. Paxton EW, Inacio MC, Singh JA, Love R, Bini SA and Namba RS. Are There Modifiable Risk
291
Factors for Hospital Readmission After Total Hip Arthroplasty in a US Healthcare System? Clin Orthop
292
Relat Res. 2015;473:3446-55.
293
10. Oronce CI, Shao H and Shi L. Disparities in 30-Day Readmissions After Total Hip Arthroplasty.
294
Med Care. 2015;53:924-30.
295
11. Quan H, Sundararajan V, Halfon P, Fong A, Burnand B, Luthi JC, Saunders LD, Beck CA, Feasby TE
296
and Ghali WA. Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data.
297
Med Care. 2005;43:1130-9.
298
12. Burnham KP and Anderson DR. Mathematical models for nonparametric inferences from line
299
transect data. Biometrics. 1976;32:325-36.
300
13. Bini SA, Fithian DC, Paxton LW, Khatod MX, Inacio MC and Namba RS. Does discharge disposition
301
after primary total joint arthroplasty affect readmission rates? J Arthroplasty. 2010;25:114-7.
302
14. Vorhies JS, Wang Y, Herndon J, Maloney WJ and Huddleston JI. Readmission and length of stay
303
after total hip arthroplasty in a national Medicare sample. J Arthroplasty. 2011;26:119-23.
304
15. Schairer WW, Vail TP and Bozic KJ. What are the rates and causes of hospital readmission after
305
total knee arthroplasty? Clin Orthop Relat Res. 2014;472:181-7.
306
16. Gold HT, Slover JD, Joo L, Bosco J, Iorio R and Oh C. Association of Depression With 90-Day
307
Hospital Readmission After Total Joint Arthroplasty. J Arthroplasty. 2016.
308
17. Rahman M, McHugh J, Gozalo PL, Ackerly DC and Mor V. The Contribution of Skilled Nursing
309
Facilities to Hospitals' Readmission Rate. Health Serv Res. 2016.
310
18. Stambough JB, Nunley RM, Curry MC, Steger-May K and Clohisy JC. Rapid recovery protocols for
311
primary total hip arthroplasty can safely reduce length of stay without increasing readmissions. J
312
Arthroplasty. 2015;30:521-6.
313

M
A
N
U
S
C
R
I
P
T
A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
15
19. Khan SK, Malviya A, Muller SD, Carluke I, Partington PF, Emmerson KP and Reed MR. Reduced
314
short-term complications and mortality following Enhanced Recovery primary hip and knee arthroplasty:
315
results from 6,000 consecutive procedures. Acta Orthop. 2014;85:26-31.
316
20. Klement MR, Bala A, Blizzard DJ, Wellman SS, Bolognesi MP and Seyler TM. Should We Think
317
Twice About Psychiatric Disease in Total Hip Arthroplasty? J Arthroplasty. 2016.
318
21. Rathod P, Deshmukh A, Robinson J, Greiz M, Ranawat A and Rodriguez J. Does Tourniquet Time
319
in Primary Total Knee Arthroplasty Influence Clinical Recovery? J Knee Surg. 2015;28:335-42.
320
22. Dennis DA, Kittelson AJ, Yang CC, Miner TM, Kim RH and Stevens-Lapsley JE. Does Tourniquet
321
Use in TKA Affect Recovery of Lower Extremity Strength and Function? A Randomized Trial. Clin Orthop
322
Relat Res. 2016;474:69-77.
323
23. Sharrock NE, Gonzalez Della Valle A, Go G, Lyman S and Salvati EA. Potent anticoagulants are
324
associated with a higher all-cause mortality rate after hip and knee arthroplasty. Clin Orthop Relat Res.
325
2008;466:714-21.
326
24. Raphael IJ, Tischler EH, Huang R, Rothman RH, Hozack WJ and Parvizi J. Aspirin: an alternative
327
for pulmonary embolism prophylaxis after arthroplasty? Clin Orthop Relat Res. 2014;472:482-8.
328
25. Nam D, Nunley RM, Johnson SR, Keeney JA, Clohisy JC and Barrack RL. The Effectiveness of a Risk
329
Stratification Protocol for Thromboembolism Prophylaxis After Hip and Knee Arthroplasty. J
330
Arthroplasty. 2016;31:1299-306.
331
332
333
334
335

M
A
N
U
S
C
R
I
P
T
A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
Table 1 Demographics of Entire Cohort (THA and TKA)
OVERALL
N=21864
Distribution by year
2010 4418 (20.2%)
2011 4321 (19.8%)
2012 4636 (21.2%)
2013 4469 (20.4%)
2014 4020 (18.4%)
Joint
Hip 11,105 (50.8%)
Knee 10,759 (49.2%)
Age at surgery (years) 65.8 +/- 10.4
BMI (kg/m2
) 29.3 +/- 6.1
Sex
Male 8,998 (41.2%)
Female 12,866 (58.8%)
Comorbidity
0 16,134 (73.8%)
1-2 5,168 (23.6%)
3+ 562 (2.6%)
Race
White 18,933 (86.6%)
Black 1,196 (5.4%)
Asian 298 (1.4%)
American Indian 25 (0.1%)
Hispanics 173 (0.8%)
Others 607 (2.8%)
N/A 632 (2.9%)
Table 2 Incidence of Comorbidities Existing at Time of Index Surgery
Comorbidity N (%)
Congestive heart failure 246 (1.1%)
Valvular disease 1,404 (6.4%)
Pulmonary circulation disease 332 (1.5%)
Peripheral vascular disease 248 (1.1%)
Paralysis 42 (0.2%)
Other neurological disorders 799 (3.7%)
Chronic pulmonary disease 2,296 (10.5%)
Diabetes w/o chronic complications 2,263 (10.4%)

M
A
N
U
S
C
R
I
P
T
A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
Diabetes w/ chronic complications 137 (0.6%)
Hypothyroidism 3,286 (15.0%)
Renal failure 528 (2.4%)
Liver disease 220 (1.0%)
Peptic ulcer Disease x bleeding 1 (0.0%)
Acquired immune deficiency
syndrome
4 (0.0%)
Lymphoma 72 (0.3%)
Metastatic cancer 12 (0.1%)
Solid tumor w/out metastasis 87 (0.4%)
Rheumatoid arthritis/collagen vas 165 (0.8%)
Coagulopthy 703 (3.2%)
Obesity 4,645 (21.2%)
Weight loss 20 (0.1%)
Fluid and electrolyte disorders 2,281 (10.4%)
Chronic blood loss anemia 4 (0.0%)
Deficiency Anemias 2,267 (10.4%)
Alcohol abuse 95 (0.4%)
Drug abuse 38 (0.2%)
Psychoses 232 (1.1%)
Depression 2,456 (11.2%)
Hypertension 11,518 (52.7%)
Table 3 Reasons for Readmission
Readmission
Reason
Overall THR TKR
(n=60) (n=37) (n=23)
Fracture 14 12 2
Infection 14 7 7
Dislocation 9 9 0
Cellulitis 7 1 6
DVT/PE 4 2 2
Pain 4 1 3
Hematoma 3 3 0
Incisional Drainage 2 0 2
Anemia 1 1 0
Instability 1 1 0
Other 1 0 1
Table 4 Matching Criteria

M
A
N
U
S
C
R
I
P
T
A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
Matching
variables
Non-readmission
group
Readmission group
N=120 N=60 P-value
Age at surgery 66.3 +/- 8.5 66.1 +/- 9.4 0.90
Comorbidity 1.0
0 80 (66.7%) 40 (66.7%)
1-2 40 (33.3%) 20 (33.3%)
Laterality 0.18
Unilateral 115 (95.8%) 54 (90.0%)
Bilateral 5 (4.2%) 6 (10.0%)
Sex 1.0
Male 56 (46.7%) 28 (46.7%)
Female 64 (53.3%) 32 (53.3%)
Joint 1.0
Hip 74 (61.7%) 37 (61.7%)
Knee 46 (38.3%) 23 (38.3%)
Year of surgery 1.0
2010 16 (13.3%) 8 (13.3%)
2011 32 (26.7%) 16 (26.7%)
2012 30 (25.0%) 15 (25.0%)
2013 28 (23.3%) 14 (23.3%)
2014 14 (11.7%) 7 (11.7%)
Table 5 THA Perioperative Factors
Non-readmission
group
Readmission
group
P-value
N=74 N=37
Procedure time (min) 71.3 +/- 14.7 77.7 +/- 18.7 0.05
Transfusion 18 (39.1%) 11 (47.8%) 0.49
LOS>2 73 (98.6%) 33 (89.2%) 0.04
Aspirin 56 (75.7%) 20 (54.1%) 0.02
TXA 7 (9.5%) 3 (8.1%) 1.0
Disposition 0.05
Home 56 (75.7%) 20 (54.1%)
SNF 5 (6.8%) 7 (18.9%)
Rehab 13 (17.6%) 10 (27.0%)
Anesthesia (spinal v. other) 4 (5.4%) 4 (10.8%) 0.44
Drain 31 (41.9%) 19 (51.4%) 0.35

M
A
N
U
S
C
R
I
P
T
A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
Periarticular injection 16 (21.6%) 8 (21.6%) 1.0
ASA 0.62
1-2 59 (79.7%) 28 (75.7%)
3-4 15 (20.3%) 9 (24.3%)
Table 6 Comorbidities Existing at Index Procedure (THA Cohort)
Non-
readmission Readmission P-value
(N=74) (N=37)
Congestive heart failure 1 (1.4%) 1 (2.7%) 1.00
Valvular disease 6 (8.1%) 3 (8.1%) 1.00
Peripheral vascular disease 1 (1.4%) 1 (2.7%) 1.00
Other neurological disorders 0 1 (2.7%) 0.33
Chronic pulmonary disease 4 (5.4%) 7 (18.9%) 0.04
Diabetes w/o chronic
complications
12 (16.2%) 2 (5.4%) 0.14
Diabetes w/ chronic
complications
1 (1.4%) 0 1.00
Hypothyroidism 7 (9.5%) 5 (13.5%) 0.53
Renal failure 3 (4.1%) 0 0.55
Coagulopthy 4 (5.4%) 3 (8.1%) 0.68
Obesity 9 (12.2%) 9 (24.3%) 0.10
Fluid and electrolyte disorders 3 (4.1%) 3 (8.1%) 0.40
Deficiency Anemias 5 (6.8%) 6 (16.2%) 0.18
Depression 5 (6.8%) 7 (18.9%) 0.10
Hypertension 33 (44.6%) 20 (54.1%) 0.35
Table 7 TKA Perioperative Factors
Non-readmission
group
Readmission
group
P-value
N=46 N=23
Procedure time (min) 81.1 +/- 13.5 95.3 +/- 29.1 0.04
Tourniquet time (min) 43.6 +/- 15.5 64.6 +/- 38.7 0.02
Transfusion 18 (39.1%) 11 (47.8%) 0.49
LOS>3 46 (100.0%) 18 (78.3%) <0.01
Aspirin 7 (15.2%) 3 (13.0%) 1.00
TXA 8 (17.4%) 3 (13.0%) 0.74

M
A
N
U
S
C
R
I
P
T
A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
Disposition 0.66
Home 24 (52.2%) 14 (60.9%)
SNF 5 (10.9%) 3 (13.0%)
Rehab 17 (37.0%) 6 (26.1%)
Drain 34 (73.9%) 15 (65.2%) 0.32
ASA 0.24
1-2 32 (69.6%) 19 (82.6%)
3-4 14 (30.4%) 4 (17.4%)
Table 8 Comorbidities Existing at Index Procedure (TKA Cohort)
Non-
readmission
group
Readmission
group
P-value
(N=46) (N=23)
Peripheral vascular disease 1 (2.2%) 0 1.00
Other neurological disorders 3 (6.5%) 1 (4.3%) 1.00
complications
5 (10.9%) 4 (17.4%) 0.47
Liver disease 0 1 (4.3%) 0.33
Obesity 12 (26.1%) 4 (17.4%) 0.42
Depression 5 (10.9%) 8 (34.8%) 0.02
Hypertension 27 (58.7%) 13 (56.5%) 1.00
Table 9 THA/TKA Combined Perioperative Factors
Non-
readmission
group
Readmission
group
Non matching variables N=120 N=60 P-value
Procedure time (min)
75.0 +/-
15.0
84.4 +/- 24.6 0.01
Transfusion 28 (23.3%) 16 (26.7%) 0.89
LOS>2 days 119 (99.2%) 56 (93.3%) 0.04
LOS>3 days 98 (81.7%) 44 (73.3%) 0.20
PCA/PCEA
102
(85.0%)
54 (90.0%) 0.35

M
A
N
U
S
C
R
I
P
T
A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
Aspirin 63 (52.5%) 23 (38.3%) 0.08
TXA 15 (12.5%) 6 (10.0%) 0.62
Disposition 0.20
Home 80 (66.7%) 34 (56.7%)
SNF 10 (8.3%) 10 (16.7%)
Rehab 30 (25.0%) 16 (26.7%)
Drain 65 (54.2%) 34 (56.7%) 0.33
ASA 0.71
1-2 91 (75.8%) 47 (78.3%)
3-4 29 (24.2%) 13 (21.7%)
Table 10 Comorbidities Existing at Index Procedure (THA/TKA Combined Cohort)
Non-
readmission
group
Readmission
group
Non matching variables N=120 N=60 P-value
Congestive heart failure 1 (0.8%) 1 (1.7%) 1.00
Peripheral vascular disease 2 (1.7%) 1 (1.7%) 0.47
Other neurological disorders 3 (2.5%) 2 (3.3%) 0.47
complications
17 (14.2%) 6 (10.0%) 0.43
Diabetes w/ chronic
complications
1 (0.8%) 0 1.00
Hypothyroidism 13 (10.8%) 13 (21.7%) 0.05
Renal failure 4 (3.3%) 0 0.30
Liver disease 0 1 (1.7%) 0.33
Rheumatoid arthritis/collagen vas 1 (0.8%) 1 (1.7%) 1.00
Coagulopathy 5 (4.2%) 3 (5.0%) 1.00
Obesity 21 (17.5%) 13 (21.7%) 0.50
Depression 10 (8.3%) 15 (25.0%) <0.01
Hypertension 60 (50.0%) 33 (55.0%) 0.53

M
A
N
U
S
C
R
I
P
T
A
C
C
E
P
T
E
D
ACCEPTED MANUSCRIPT
Table 11 Conditional Logistic Regression Analysis
Odds Ratio P-value
THA Readmissions
LOS LOS>3 vs LOS≤3 1.33 (0.43-4.16) 0.62
Depression 3.15 (0.92-10.77) 0.07
Discharge disposition Rehab vs Home 3.02 (1.10-8.30) 0.03
TKA Readmissions
Depression 3.35 (0.72-15.59) 0.12
Tourniquet time (min) 1.04 (1.00-1.07) 0.02
THA/TKA Readmissions
combined
LOS LOS>3 vs LOS≤3 0.58 (0.22-1.53) 0.27
Depression 3.48 (1.43-8.51) 0.01

Ricciardi et al. - 2017 - Patient and Perioperative Variables Affecting 30-D.pdf

Recommended

Recommended

More Related Content

Similar to Ricciardi et al. - 2017 - Patient and Perioperative Variables Affecting 30-D.pdf

Similar to Ricciardi et al. - 2017 - Patient and Perioperative Variables Affecting 30-D.pdf (20)

Recently uploaded

Recently uploaded (20)

Ricciardi et al. - 2017 - Patient and Perioperative Variables Affecting 30-D.pdf