- Email: [email protected]
Social determinants of access to minimally invasive hysterectomy: reevaluating the relationship between race and route of hysterectomy for benign disease
Accepted Manuscript Social determinants of access to minimally invasive hysterectomy: Re-evaluating the relationship between race and route of hysterectomy for benign disease Joan T. Price, MD MPH, Lilli D. Zimmerman, MD, Nathan C. Koelper, MPH, Mary D. Sammel, ScD, Sonya Lee, MD, Samantha F. Butts, MD MSCE PII:
S0002-9378(17)30927-4
DOI:
10.1016/j.ajog.2017.07.036
Reference:
YMOB 11791
To appear in:
American Journal of Obstetrics and Gynecology
Received Date: 14 July 2017 Revised Date:
22 July 2017
Accepted Date: 31 July 2017
Please cite this article as: Price JT, Zimmerman LD, Koelper NC, Sammel MD, Lee S, Butts SF, Social determinants of access to minimally invasive hysterectomy: Re-evaluating the relationship between race and route of hysterectomy for benign disease, American Journal of Obstetrics and Gynecology (2017), doi: 10.1016/j.ajog.2017.07.036. 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.
ACCEPTED MANUSCRIPT Price 1 1
Social determinants of access to minimally invasive hysterectomy: Re-evaluating the relationship
2
between race and route of hysterectomy for benign disease
3
Running Title: Social Determinants of Hysterectomy Route
RI PT
4 5
Joan T PRICE MD MPH1; Lilli D ZIMMERMAN MD1; Nathan C KOELPER MPH1; Mary D
6
SAMMEL ScD2; Sonya LEE MD3; Samantha F BUTTS MD MSCE1
8
1
9
Medicine, Philadelphia, PA, USA
SC
7
M AN U
Department of Obstetrics and Gynecology, University of Pennsylvania Perelman School of
10
2
11
Biostatistics, University of Pennsylvania Perelman School of Medicine, Philadelphia, PA, USA
12
3
Department of Biostatistics and Epidemiology, Center for Clinical Epidemiology and
Department of Gynecology, Penn Presbyterian Medical Center, Philadelphia, PA, USA
TE D
13 14
Funding: None.
15
Conflicts of interest: The authors have no conflicts of interest to report.
EP
16 Corresponding author:
18
Joan T Price, MD MPH
19
Department of Obstetrics and Gynecology, UNC-Chapel Hill Medical Center
20
101 Manning Drive, Chapel Hill, NC 27514
21
tel: (919) 843-9546 email: [email protected]
22
Abstract word count: 374
23
Manuscript word count: 2,987
AC C
17
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT Price 2 Condensation
25
Demographic factors are associated with access to various types of minimally invasive
26
hysterectomy independent of body mass index, parity, surgical history, and uterine size.
27 28
Running head: Social Determinants of Hysterectomy Route
29
SC
30 31
M AN U
32 33 34 35
40 41 42 43 44
EP
39
AC C
38
TE D
36 37
RI PT
24
45 46
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT Price 3 ABSTRACT
48
BACKGROUND: Racial and socioeconomic disparities exist in access to medical and surgical
49
care. Studies of national databases have demonstrated disparities in route of hysterectomy for
50
benign indications, but have not been able to adjust for patient-level factors that affect surgical
51
decision-making.
RI PT
47
52
OBJECTIVES: To determine whether access to minimally invasive hysterectomy for benign
54
indications is differential according to race independent of the effects of relevant subject-level
55
confounding factors. The secondary study objective was to determine the association between
56
socioeconomic status and ethnicity and access to minimally invasive hysterectomy.
M AN U
SC
53
57
STUDY DESIGN: A cross-sectional study evaluated factors associated with minimally invasive
59
hysterectomies performed for fibroids and/or abnormal uterine bleeding between 2010 and 2013
60
at three hospitals within an academic university health system in Philadelphia, Pennsylvania.
61
Univariate tests of association and multivariable logistic regression identified factors
62
significantly associated with minimally invasive hysterectomy compared to the odds of treatment
63
with the referent approach of abdominal hysterectomy.
EP
AC C
64
TE D
58
65
RESULTS: Of 1746 hysterectomies evaluated meeting study inclusion criteria, 861 (49%) were
66
performed abdominally, 248 (14%) vaginally, 310 (18%) laparoscopically and 327 (19%) were
67
performed with robot assistance. In univariate analysis, African-American race (OR 0.80, 95%
68
CI 0.65–0.97) and Hispanic ethnicity (OR 0.63, 95% CI 0.39–1.00) were associated with lower
69
odds of any minimally invasive hysterectomy relative to abdominal hysterectomy. In analyses
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT Price 4 adjusted for age, BMI, income quartile, obstetrical and surgical history, uterine weight, and
71
additional confounding factors, African-American race was no longer a risk factor for reduced
72
minimally invasive hysterectomy (OR 0.82, 95% CI 0.61–1.10), while Hispanic ethnicity (OR
73
0.45, 95% CI 0.27–0.76) and Medicaid enrollment (OR 0.59, 95% CI 0.38–0.90) were associated
74
with significantly lower odds of treatment with any minimally invasive hysterectomy. In
75
adjusted analyses, African-American women had nearly half the odds of receiving robot-assisted
76
hysterectomy compared to whites (AOR 0.57, 95% CI 0.39-0.82), while no differences were
77
noted with other hysterectomy routes. Medicaid enrollment (compared to private insurance; OR
78
0.51, 95% CI 0.28–0.94) and lowest income quartile (compared to highest income quartile; OR
79
0.57, 95% CI 0.38–0.85) were also associated with diminished odds of robot-assisted
80
hysterectomy.
M AN U
SC
RI PT
70
81 CONCLUSION:
When accounting for the effect of numerous pertinent demographic and
83
clinical factors, the odds of undergoing minimally invasive hysterectomy were diminished in
84
women of Hispanic ethnicity and in those enrolled in Medicaid but were not discrepant along
85
racial lines. However, both racial and socioeconomic disparities were observed with respect to
86
access to robot-assisted hysterectomy despite the availability of robotic assistance in all hospitals
87
treating the study population. Strategies to ensure equal access to all minimally invasive routes
88
for all women should be explored to align delivery of care with the evidence supporting the
89
broad implementation of these procedures as safe, cost effective, and highly acceptable to
90
patients.
AC C
EP
TE D
82
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT Price 5 INTRODUCTION
92
Minimally invasive hysterectomy (MIH) has advantages over abdominal hysterectomy including
93
shorter hospital stay, improved postoperative pain, fewer complications and lower hospital
94
costs.1 The American College of Obstetrics and Gynecology encourages the application of MIH,
95
particularly vaginal hysterectomy, as safer and more cost effective than abdominal surgery for
96
benign indications.2 Despite established advantages of MIH, racial and socioeconomic disparities
97
exist in access to these approaches.3 The disparate application of minimally invasive procedures
98
among African Americans and socioeconomically disadvantaged groups has been noted across
99
many surgical specialties, including gynecologic surgery.4-7 National investigations of disparities
100
in MIH have found that younger age, white race, higher income and private insurance
101
independently predict laparoscopic over abdominal hysterectomy.8-11
M AN U
SC
RI PT
91
102
The primary aim of this study was to determine the association between race and the odds of
104
MIH for the treatment of fibroids, adenomyosis and abnormal bleeding. The secondary objective
105
was to assess the relationship between indicators of socioeconomic status, ethnicity and the odds
106
of MIH for the above indications. To build on findings from previous reports using national
107
databases, we sought to evaluate subject-level data related to hysterectomies performed at three
108
urban University-affiliated hospitals to more thoroughly control for confounding. Because
109
national databases often do not differentiate between traditional laparoscopic versus robot-
110
assisted routes of hysterectomy, we sought to determine whether certain individual minimally
111
invasive routes were more prone to racial and/or socioeconomic disparities in sub-group
112
analyses. Our hypothesis was that non-white race and low socioeconomic status would be
113
associated with diminished utilization of MIH despite adjusting for confounding due to patient-
AC C
EP
TE D
103
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT Price 6 114
level clinical predictors of route of hysterectomy. Elucidating disparities in access to the
115
recommended hysterectomy approaches is a critical step in improving equity and optimizing
116
outcomes for all women who require major gynecological surgery.
RI PT
117 MATERIALS AND METHODS
119
A cross-sectional study analyzed data from subjects receiving care at the three academic
120
hospitals within the University of Pennsylvania Health System (UPHS) in Philadelphia. Data
121
were gathered on all hysterectomies performed for fibroids, adenomyosis and/or abnormal
122
uterine bleeding at the Hospital of the University of Pennsylvania (HUP), Pennsylvania Hospital
123
(PAH) and Penn Presbyterian Medical Center (PPMC) from January 2010 through December
124
2013. As medical centers within an urban academic health system, each hospital has the
125
resources to perform abdominal, vaginal, laparoscopic, and robot-assisted hysterectomy. The
126
gynecology practice at PPMC is a two-provider high-volume private specialty practice whereas
127
the HUP and PAH sites comprise multiple practice types including private generalist, academic
128
sub-specialist and resident clinic practices that all perform hysterectomies for benign indications.
M AN U
TE D
EP
129
SC
118
Subjects who underwent abdominal, vaginal, laparoscopic and robot-assisted hysterectomy for
131
abnormal uterine bleeding, adenomyosis or fibroids were identified using diagnosis and
132
procedure codes from the International Classification of Disease, Ninth Revision, Clinical
133
Modification (ICD-9-CM). Exclusion criteria were: presence of known or suspected cancer,
134
known endometriosis, and cases performed jointly with another surgical service. Using reports
135
from the Penn Data Store clinical database and verified by electronic medical record review,
136
patient-level data were collected on demographic information, primary indication and mode of
AC C
130
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT Price 7 hysterectomy, self-identified race and ethnicity, median household income quartile (based on zip
138
code of residence),12 and insurance. Insurance status was categorized as (1) private insurance, (2)
139
health maintenance organization / preferred provider organization (HMO/PPO), which
140
comprised only non-Medicaid, non-Medicare managed plans; (3) Medicaid (including managed
141
Medicaid); (4) Medicare (including managed Medicare); and (5) self-pay or uninsured. Both
142
uterine weight and volume were abstracted from surgical pathology reports for each subject. The
143
final multivariable analyses adjusted only for uterine weight to avoid colinearity by including
144
both variables. Uterine weight was chosen as the more objective measure; volume estimates can
145
be imprecise when fibroids distort the uterine contour.
M AN U
SC
RI PT
137
146
The primary study outcome was mode of hysterectomy. Any abdominal hysterectomy including
148
total abdominal and abdominal supracervical hysterectomy was placed in the “abdominal”
149
category. Total vaginal hysterectomy without laparoscopic assistance was its own category
150
whereas
151
hysterectomy was coded as “laparoscopic”. In cases performed with robot assistance, the
152
hysterectomy was categorized as “robotic”. Hysterectomies were categorized under the initial
153
route of hysterectomy regardless of intra-operative conversion in order to reflect pre-operative
154
decision-making and surgical planning.
total
laparoscopic,
or
laparoscopic
supracervical
EP
laparoscopic-assisted,
AC C
155
any
TE D
147
156
Continuous variables were summarized using means and standard deviations, while frequencies
157
were used to describe the categorical variables. Age, BMI, uterine weight, and uterine volume
158
were treated as continuous variables. Uterine weight was transformed into a categorical variable
159
based on contingency tables detecting a frequency of at least two individuals in cross tabulation.
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT Price 8 160
Univariate comparisons of each of the potential predictors were performed with the dichotomous
161
outcomes of abdominal versus all MIH, vaginal, laparoscopic, and robotic hysterectomy, using
162
Pearson’s chi-square tests, Student’s t-tests or simple logistic regression as appropriate.
RI PT
163
Four logistic regression models were fit to determine the set of independent variables that best
165
predicted the odds of abdominal compared to any MIH, vaginal, laparoscopic, or robotic
166
hysterectomy. Backward selection was used to generate final models starting with all variables of
167
interest then excluding those that were not significant and successively re-fitting reduced models
168
until all remaining variables were statistically significant at p<0.05. Variables associated with
169
slightly higher p values (p<0.5) were retained if they were known risk factors for abdominal
170
hysterectomy or if the effect estimate for race changed by >10% when the covariate was
171
removed. Likelihood ratio testing was used to screen the robustness of covariates to generate a
172
parsimonious final model and to test for the presence of interactions. Statistical analyses were
173
performed using SAS 9.3 statistical software, Cary, NC.
TE D
M AN U
SC
164
174
The University of Pennsylvania Institutional Review Board granted approval to conduct this
176
study prior to the commencement of data collection; the requirement for informed consent was
177
waived.
AC C
178
EP
175
179
RESULTS
180
A total of 2166 hysterectomies were performed for the indications of fibroids, adenomyosis
181
and/or abnormal uterine bleeding at three UPHS hospitals between January 1, 2010 and
182
December 31, 2013. Of these cases, 420 (19%) were excluded, generating a sample of 1746
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT Price 9 183
eligible cases for analysis (Figure 1). Missing data were rare and addressed using listwise
184
deletion.
185 Most participants self-identified as either African American (57%, n=978) or white (38%,
187
n=659) (Table 1). More than half (55%) resided in zip codes representing neighborhoods in the
188
two lowest median household income quartiles (<$48,000 annually). Most subjects were covered
189
by either private insurance (35%) or managed care (50%).
SC
190
RI PT
186
Among 1746 hysterectomies, 49% (N=861) were performed abdominally, 14% (N=248)
192
vaginally and 37% (N=637) laparoscopically (Table 1). Of the laparoscopic hysterectomies, 327
193
(19% of total) were performed with robotic assistance. The primary indication for hysterectomy
194
was fibroids for 80% of participants.
TE D
195
M AN U
191
Most white subjects (54%) underwent MIH; more than half of those (28% overall) were
197
performed with robotic assistance. Among African-American women, 48% underwent MIH;
198
only 25% of those (12% overall) were robot-assisted. The proportion of whites treated with
199
vaginal hysterectomy was 9% while 18% of African Americans had hysterectomy performed
200
vaginally. Most vaginal hysterectomies (71%) were performed in African Americans.
AC C
201
EP
196
202
Examining the association of race and other clinical or demographic variables revealed that
203
African-American women were slightly younger (p<.001) with modestly higher BMIs compared
204
to white women (p<.001; Table 2). More African-American women than white women had
205
fibroids as their primary diagnosis (p<.001), had surgery performed at PPMC (p<.001), were in
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT Price 10 206
the lower income quartiles (p<.001), were covered by Medicaid (<.001), had a prior vaginal
207
delivery (p<.001), had a prior abdominal surgery (p=.005), and had larger uteri by both weight
208
and volume (p<.001).
RI PT
209
Univariate analysis demonstrated relationships between route of hysterectomy and race,
211
ethnicity, insurance status, and median household income quartile (Table 3). Comparing
212
abdominal hysterectomy to all minimally invasive routes combined, African-American patients
213
had lower odds of undergoing MIH than whites (OR 0.80, 95% CI 0.65–0.97), an association
214
that achieved borderline significance. African Americans had even lower odds than whites when
215
comparing robot-assisted hysterectomy alone to abdominal hysterectomy (OR 0.38, 95% CI
216
0.29–0.50). Conversely, African Americans had greater odds than whites of undergoing a vaginal
217
hysterectomy compared to abdominal hysterectomy (OR 1.67, 95% CI 1.21-2.30).
M AN U
SC
210
TE D
218
Neither household income quartile nor insurance status was associated with MIH over abdominal
220
hysterectomy in univariate analyses. However, women with Medicaid had significantly lower
221
odds of undergoing robot-assisted than abdominal hysterectomy (OR 0.45, 95% CI 0.24–0.86)
222
when compared to women with private insurance. Similarly, women in the lowest median
223
household income quartile had lower odds of treatment with robot-assisted hysterectomy (OR
224
0.35, 95% CI 0.25–0.49) compared to those residing in the highest quartile. Conversely, when
225
compared to those in the highest income quartile, women in the lowest quartile had higher odds
226
of vaginal hysterectomy than abdominal hysterectomy (OR 2.51, 95% CI 1.70–3.69).
AC C
EP
219
227
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT Price 11 In multivariable logistic regression models, we found that adjusting for subject-level factors
229
mitigated many of the univariate associations between route of hysterectomy and race, ethnicity,
230
median household income quartile, and insurance status (Table 4). When compared to white
231
women, African-American women had similar odds of undergoing any mode of MIH as they did
232
abdominal hysterectomy (AOR 0.82; 95% CI 0.61–1.10). In a subgroup multivariable analysis
233
investigating the robot-assisted route alone as the alternate to abdominal hysterectomy, African-
234
American women had significantly lower odds of robot-assisted hysterectomy compared to
235
whites (AOR 0.57, CI 0.39–0.82). Racial disparities in vaginal and laparoscopic hysterectomy
236
compared to abdominal hysterectomy were not statistically significant in adjusted analyses.
237
Compared to women of non-Hispanic ethnicity, self-identified Hispanic women had lower
238
adjusted odds of all MIH (AOR 0.45, 95% CI 0.27–0.76), which was predominantly driven by
239
significantly lower odds of both vaginal (AOR 0.24, 95% CI 0.09–0.64) and laparoscopic
240
hysterectomy (AOR 0.42, 95% CI 0.20–0.89), but not robotic hysterectomy.
SC
M AN U
TE D
241
RI PT
228
Median household income quartile and insurance coverage modestly predicted mode of
243
hysterectomy in multivariable regression. In the primary analysis, income quartile did not
244
predict MIH over abdominal hysterectomy, but women with Medicaid had significantly lower
245
odds of MIH compared to women with private insurance (AOR 0.59, 95% CI 0.38–0.90). In
246
separate subgroup analyses, women in the lowest income quartile (AOR 0.57, 95% CI 0.38–
247
0.85) or with Medicaid (AOR 0.51, 95% CI 0.28–0.94) both had nearly half the odds of robot-
248
assisted over abdominal hysterectomy than women in the highest income quartile or those with
249
private insurance, respectively. Conversely, women in the lowest income quartile had
250
significantly greater odds of undergoing vaginal over abdominal hysterectomy than those in the
AC C
EP
242
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT Price 12 251
highest quartile (AOR 2.52, 95% CI 1.51–4.21). Insurance status did not modify the association
252
between African-American race and mode of hysterectomy described in the multivariable
253
models.
RI PT
254 COMMENT
256
Our analysis demonstrates that racial and socioeconomic disparities in the application of
257
minimally invasive hysterectomy to treat benign gynecological disorders may be mitigated by
258
adjusting for patient-level confounding factors. In contrast to national registry studies, our
259
findings demonstrate that African-American women requiring hysterectomy for fibroids and/or
260
abnormal bleeding had similar odds as white women of treatment with MIH when controlling for
261
BMI, obstetrical and surgical history, and uterine size. By generating multivariable models that
262
adjusted for patient-level factors, we were able to expose the role of confounding on previously
263
described associations between race and access to MIH.4,8-11
M AN U
TE D
264
SC
255
In the United States; African-American race, low income, and underinsurance are all associated
266
with overweight and obesity.13 Obesity limits the application of laparoscopic and robot-assisted
267
surgery; obese women have greater odds of vaginal over other routes of hysterectomy.14
268
Similarly, fibroids disproportionately affect African-American women, who have both larger and
269
more numerous fibroids than whites.15-17 In our analysis, controlling for confounding due to BMI
270
and uterine weight certainly reduced the magnitude and significance of racial disparities noted in
271
unadjusted analyses. However, sensitivity analyses limited to procedures performed among
272
women with uterine weights <400g did not alter the findings of the remaining racial and
273
socioeconomic disparities in access to robotic hysterectomy. Large national databases often do
AC C
EP
265
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT Price 13 274
not report BMI or uterine size such that previous analyses of nationwide disparities in mode of
275
hysterectomy have not been able to account for their effect on surgical route.3,4,8-11
276 In our study population, disparities in access to MIH due to insurance and income status are less
278
stark than in national analyses, given the ability to more comprehensively address confounding
279
due to multiple factors. By using data from three hospitals within one urban hospital system in
280
Philadelphia, our analysis among a relatively uniform population may have biased some
281
socioeconomic associations towards the null. The small proportion of patients covered by
282
Medicaid in our sample could be due to the evolving composition of insurance provider types in
283
an era marked by insurance reforms under the Affordable Care Act. Improving insurance
284
coverage for nonwhite patients in states like Massachusetts has been met with near complete
285
resolution of racial disparities in minimally invasive surgeries.18 Similarly, in an urban setting
286
such as Philadelphia where neighborhoods and zip codes are more economically diverse than in
287
suburban and rural communities, median household income quartile based on zip code of
288
residence is likely an imperfect proxy for individual socioeconomic status. Unfortunately,
289
specific individual economic productivity data were not available for query in our medical
290
system’s database. Despite these limitations, we still observed that having private insurance
291
(compared to Medicaid) was independently associated with greater access to MIH. Being in the
292
highest income quartile (compared to the lowest) and having private insurance (compared to
293
Medicaid) also predicted robot-assisted hysterectomy over abdominal hysterectomy in adjusted
294
subgroup analyses. These findings are comparable to national reports that did not control for
295
subject-level confounding.3,4,8-11
AC C
EP
TE D
M AN U
SC
RI PT
277
296
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT Price 14 Previous analyses have shown racial and socioeconomic disparities in laparoscopic hysterectomy
298
overall but most have not evaluated the role of robot-assisted hysterectomy specifically, as
299
national databases rarely differentiate between laparoscopic surgeries performed with and
300
without robot assistance.4,8-11 In our subgroup analyses, adjusting for confounding in
301
multivariable analysis did eliminate many univariate associations between sociodemographic
302
factors and the outcome of MIH versus abdominal hysterectomy. In specifically investigating
303
robot-assisted hysterectomy, however, we found racial and economic disparities not apparent in
304
other MIH routes, even despite controlling for patient-level factors. The ability to test specific
305
associations that apply to discrete categories of MIH is a unique feature of this investigation. It is
306
plausible that, as a more established and cost-effective route when compared to robot-assisted
307
surgery,19,20 laparoscopic hysterectomy is more equally accessible – both regionally and across
308
diverse communities. The divergent associations of race and income on different modes of MIH
309
may explain why the combined outcome of MIH overall failed to demonstrate significant
310
disparity and supports the investigation of laparoscopic and robot-assisted hysterectomies
311
separately.
SC
M AN U
TE D
EP
312
RI PT
297
In contrast to trends in robot-assisted hysterectomy, the odds of undergoing vaginal
314
hysterectomy over abdominal hysterectomy were higher among women in the lowest income
315
quartile compared to those in the highest income quartile and among African Americans
316
compared to whites. Individual practice patterns and patient characteristics at the three hospitals
317
could explain the findings noted despite having controlled for hospital itself. It is conceivable
318
that the individual practices of the highest-volume surgeons at each hospital influenced our
319
findings. At PPMC, vaginal and laparoscopic hysterectomies are exceedingly more common than
AC C
313
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT Price 15 abdominal and robot-assisted hysterectomies and a greater proportion of the patients are low-
321
income and African American than at the other two hospitals. Restricting the analysis to PAH
322
and HUP did not alter the magnitude or significance of the associations with robot-assisted
323
compared to abdominal hysterectomy, but did further increase the odds of vaginal hysterectomy
324
among women in the lowest compared to highest income quartile. The differential application of
325
vaginal hysterectomy over abdominal hysterectomy to low-income women may be due to
326
employment status, less lenient work or childcare schedules, and other economic considerations,
327
which may not only favor vaginal over abdominal hysterectomy, but may also influence the
328
initial decision to perform hysterectomy over more conservative management.21-23 Overall,
329
vaginal hysterectomy trends within our health system appear to be mitigating rather than
330
potentiating disparities apparent in other minimally invasive routes.
M AN U
SC
RI PT
320
331
Although efforts to decrease the proportion of hysterectomies performed abdominally are
333
improving uptake of MIH overall, understanding and minimizing factors that may perpetuate
334
racial and socioeconomic inequities must be prioritized locally and nationally, particularly as
335
vaginal hysterectomies become increasingly rare and robot-assisted hysterectomies become
336
increasingly common.24,25 Vaginal hysterectomy is the preferred route for all patients for whom
337
it is feasible and safe, yet the proportion of hysterectomies performed vaginally remains low in
338
our population and nationwide.1,2 Gynecological training programs should continue to teach
339
vaginal hysterectomy as a low-cost minimally invasive approach that may also promote a more
340
equitable application of MIH to all women.2
AC C
EP
TE D
332
341
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT Price 16 This analysis has limitations. Despite the large, diverse, well-characterized cohort studied, the
343
data represent a retrospective analysis and uncontrolled confounding may exist. We chose to
344
exclude hysterectomies performed with known or suspected endometriosis due to the effect
345
adhesive disease can have on surgical decision-making. While this may limit the generalization
346
of our analysis, most benign hysterectomies in the United States are performed for fibroids or
347
abnormal bleeding.26 Similarly, the generalizability of our findings does not extend to other
348
hospital systems in other geographic locations and controlling for individual providers or
349
provider characteristics within our study was beyond the scope of our analysis. Future research
350
that purposefully selects a robust pool of data from hospitals across the country with diverse
351
patients, providers, and practice types could address confounding factors of both patients and
352
providers to better address these limitations.
M AN U
SC
RI PT
342
353
Our sub-regional analysis of racial and socioeconomic disparities in route of hysterectomy for
355
benign indications suggests that disparities reported in national analyses may be due to
356
uncontrolled patient-level confounding factors that influence surgical decision-making. We must
357
interpret national analyses of disparities with caution as we have shown that adjusting for
358
patient-level factors can mitigate and even eliminate associations found in unadjusted analyses.
359
Still, the persistence of some disparities in access to MIH approaches despite controlling for
360
multiple confounding factors remains a concern. The repeated publication of data revealing
361
racial and socioeconomic disparities in medical and surgical care demands a multidisciplinary
362
effort to carefully elucidate the source and true extent of these inequities, and collaborative
363
strategies to ensure equitable access to all modes of MIH for racial minorities, the underinsured,
364
and low-income women.
AC C
EP
TE D
354
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT Price 17 365
Acknowledgements: The authors would like to thank the Penn Data Store for assistance in
366
assembling information used in this study. We thank Dr. Robert L. Dood for his contribution to
367
preliminary data analysis.
AC C
EP
TE D
M AN U
SC
RI PT
368
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT Price 18 369
REFERENCES
370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412
1.
7.
8.
9. 10. 11.
12. 13. 14.
15.
16.
17.
RI PT
SC
6.
M AN U
5.
TE D
4.
EP
3.
AC C
2.
Aarts JW, Nieboer TE, Johnson N, et al. Surgical approach to hysterectomy for benign gynaecological disease. Cochrane Database Syst Rev. 2015;8:CD003677. ACOG Committee Opinion No. 444: choosing the route of hysterectomy for benign disease. Obstet Gynecol. 2009;114(5):1156-1158. Abenhaim HA, Azziz R, Hu J, Bartolucci A, Tulandi T. Socioeconomic and racial predictors of undergoing laparoscopic hysterectomy for selected benign diseases: analysis of 341487 hysterectomies. J Minim Invasive Gynecol. 2008;15(1):11-15. Jacoby VL, Autry A, Jacobson G, Domush R, Nakagawa S, Jacoby A. Nationwide use of laparoscopic hysterectomy compared with abdominal and vaginal approaches. Obstet Gynecol. 2009;114(5):1041-1048. Varela JE, Nguyen NT. Disparities in access to basic laparoscopic surgery at U.S. academic medical centers. Surg Endosc. 2011;25(4):1209-1214. Robinson CN, Balentine CJ, Sansgiry S, Berger DH. Disparities in the use of minimally invasive surgery for colorectal disease. J Gastrointest Surg. 2012;16(5):897-903; discussion 903-894. Trinh QD, Schmitges J, Sun M, et al. Improvement of racial disparities with respect to the utilization of minimally invasive radical prostatectomy in the United States. Cancer. 2012;118(7):1894-1900. Esselen KM, Vitonis A, Einarsson J, Muto MG, Cohen S. Health Care Disparities in Hysterectomy for Gynecologic Cancers: Data From the 2012 National Inpatient Sample. Obstet Gynecol. 2015;126(5):1029-1039. Lee J, Jennings K, Borahay MA, et al. Trends in the national distribution of laparoscopic hysterectomies from 2003 to 2010. J Minim Invasive Gynecol. 2014;21(4):656-661. Patel PR, Lee J, Rodriguez AM, et al. Disparities in use of laparoscopic hysterectomies: a nationwide analysis. J Minim Invasive Gynecol. 2014;21(2):223-227. Price JT, Lee S. National Trends and Disparities in Minimally Invasive Approaches to Hysterectomy for Benign Indications. Journal of Minimally Invasive Gynecology. 20(6):S11-S12. American Community Survey, 2013 American Community Survey 3-year estimates, Table B19013. http://factfinder2.census.gov. Ogden CLCL. Prevalence of childhood and adult obesity in the United States, 2011-2012. JAMA : the journal of the American Medical Association.311(8):806. Mikhail E, Miladinovic B, Velanovich V, Finan MA, Hart S, Imudia AN. Association between obesity and the trends of routes of hysterectomy performed for benign indications. Obstet Gynecol. 2015;125(4):912-918. Eltoukhi HM, Modi MN, Weston M, Armstrong AY, Stewart EA. The health disparities of uterine fibroid tumors for African American women: a public health issue. Am J Obstet Gynecol. 2014;210(3):194-199. Baird DD, Dunson DB, Hill MC, Cousins D, Schectman JM. High cumulative incidence of uterine leiomyoma in black and white women: ultrasound evidence. Am J Obstet Gynecol. 2003;188(1):100-107. Moore AB, Flake GP, Swartz CD, et al. Association of race, age and body mass index with gross pathology of uterine fibroids. J Reprod Med. 2008;53(2):90-96.
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT Price 19
21. 22. 23. 24. 25. 26.
EP
TE D
27.
RI PT
20.
SC
19.
Loehrer AP, Song Z, Auchincloss HG, Hutter MM. Massachusetts health care reform and reduced racial disparities in minimally invasive surgery. JAMA Surg. 2013;148(12):11161122. Wright JD, Ananth CV, Lewin SN, et al. Robotically assisted vs laparoscopic hysterectomy among women with benign gynecologic disease. JAMA. 2013;309(7):689698. Pasic RP, Rizzo JA, Fang H, Ross S, Moore M, Gunnarsson C. Comparing robot-assisted with conventional laparoscopic hysterectomy: impact on cost and clinical outcomes. J Minim Invasive Gynecol. 2010;17(6):730-738. Marks NF, Shinberg DS. Socioeconomic differences in hysterectomy: the Wisconsin Longitudinal Study. Am J Public Health. 1997;87(9):1507-1514. Materia E, Rossi L, Spadea T, et al. Hysterectomy and socioeconomic position in Rome, Italy. J Epidemiol Community Health. 2002;56(6):461-465. Dharmalingam A, Pool I, Dickson J. Biosocial determinants of hysterectomy in New Zealand. Am J Public Health. 2000;90(9):1455-1458. Committee on Gynecologic P. Committee Opinion No 701: Choosing the Route of Hysterectomy for Benign Disease. Obstet Gynecol. 2017;129(6):e155-e159. Wright JD, Herzog TJ, Tsui J, et al. Nationwide trends in the performance of inpatient hysterectomy in the United States. Obstet Gynecol. 2013;122(2 Pt 1):233-241. Kho KA, Lin K, Hechanova M, Richardson DL. Risk of Occult Uterine Sarcoma in Women Undergoing Hysterectomy for Benign Indications. Obstet Gynecol. 2016;127(3):468-473. Shirlina D, Shirish S. Uterine volume: an aid to determine the route and technique of hysterectomy. J Obstet Gynecol Ind. 2004;54(1):68-72.
M AN U
18.
AC C
413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT
Table 1. Baseline Characteristics of 1746 women who underwent hysterectomy for benign indications, 2010-2013 Abdominal
Vaginal
Mean (±SD)
Mean (±SD)
Mean (±SD)
or % (N)*
or % (N)*
or % (N)*
Total
100 (1746)
49 (861)
14 (248)
Age, y
45.9 (±7.2)
46.3 (±6.7)
BMI, kg/m²
30.4 (±7.2)
30.4 (±7.3)
1 (23)
0 (2)
Missing
or % (N)*
or % (N)*
18 (310)
19 (327)
44.8 (±8.0)
45.1 (±7.2)
46.6 (±7.9)
31.9 (±7.6)
30.3 (±7.0)
29.7 (±6.9)
2 (6)
2 (5)
3 (10)
M AN U
SC
Mean (±SD)
EP
Diagnosis
Mean (±SD)
TE D
Characteristic
Laparoscopic Robot-assisted
RI PT
All
AC C
Bleeding or Adenomyosis
20 (356)
13 (113)
34 (84)
25 (77)
25 (82)
Fibroids
80 (1390)
87 (748)
66 (164)
75 (233)
75 (245)
Social Determinants of Hysterectomy Route
Price 20
ACCEPTED MANUSCRIPT
Price 21
2010
25 (444)
26 (227)
23 (56)
2011
27 (464)
28 (242)
26 (64)
2012
26 (459)
25 (211)
28 (69)
RI PT
Year of Surgery
2013
22 (379)
21 (181)
24 (59)
HUP
40 (699)
39 (339)
PAH
46 (802)
PPMC
14 (245)
20 (65)
22 (68)
28 (90)
24 (74)
32 (105)
23 (72)
21 (67)
28 (70)
47 (146)
44 (144)
58 (502)
25 (63)
18 (56)
55 (181)
2 (20)
46 (115)
35 (108)
1 (2)
M AN U
EP
TE D
Hospital
SC
31 (96)
AC C
Race White
38 (659)
35 (303)
25 (62)
36 (111)
56 (183)
AA
56 (978)
59 (505)
69 (172)
59 (184)
36 (117)
4 (75)
4 (38)
3 (7)
3 (10)
6 (20)
Other
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT
2 (34)
2 (15)
3 (7)
2 (5)
2 (7)
95 (1666)
95 (812)
98 (242)
97 (300)
95 (312)
Hispanic
5 (78)
5 (47)
2 (6)
RI PT
Price 22 Missing
3 (10)
5 (15)
Missing
0 (2)
0 (2)
0 (0)
0 (0)
0 (0)
$64,000+
31 (539)
30 (259)
16 (40)
30 (92)
45 (148)
$48-63,999
14 (239)
13 (113)
13 (31)
13 (41)
17 (54)
$38-47,999
15 (268)
15 (132)
15 (37)
14 (45)
17 (54)
$1-37,999
40 (694)
41 (354)
55 (137)
43 (132)
22 (71)
0 (3)
1 (3)
0 (0)
0 (0)
37 (318)
37 (91)
33 (102)
32 (103)
Insurance Private
EP
0 (6)
AC C
Missing
TE D
Income Quartile
M AN U
Not Hispanic
SC
Ethnicity
35 (614)
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT
49 (862)
47 (403)
42 (105)
53 (163)
58 (191)
Medicaid
8 (140)
10 (82)
11 (27)
6 (19)
4 (12)
Medicare
6 (112)
6 (54)
8 (20)
7 (23)
5 (15)
Missing/Self-pay
1 (18)
0 (4)
2 (5)
1 (3)
2 (6)
Delivery
1.5 (±1.4)
1.3 (±1.4)
2.2 (±1.6)
1.6 (±1.5)
1.3 (±1.4)
0
34 (597)
39 (339)
18 (45)
30 (93)
37 (120)
1-2
42 (742)
41 (356)
40 (98)
45 (141)
45 (147)
3-4
19 (331)
16 (136)
34 (84)
18 (56)
17 (55)
≥5
4 (52)
2 (19)
6 (15)
5 (14)
1 (4)
Missing
1 (24)
1 (11)
2 (6)
2 (6)
0 (1)
SC
HMO/PPO
RI PT
Price 23
TE D
EP
AC C
Prior Cesarean
M AN U
Prior Vaginal
Delivery
0.4 (±0.8)
0.4 (±0.8)
0.2 (±0.6)
0.4 (±0.7)
0.6(±0.9)
0
74 (1298)
75 (644)
85 (210)
74 (229)
66 (215)
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT
14 (240)
14 (118)
8 (21)
15 (46)
17 (55)
2
8 (139)
7 (63)
3 (8)
8 (24)
13 (44)
≥3
3 (48)
3 (27)
2 (4)
2 (5)
4 (12)
Missing
1 (21)
1 (9)
2 (5)
2 (6)
0 (1)
56 (974)
54 (467)
59 (147)
60 (185)
53 (175)
1 (16)
1 (9)
1 (2)
1 (4)
0 (1)
Uterine weight (g)
497 (±610)
716 (±775)
237 (±148)
315 (±299)
288 (±219)
Uterine volume§
579 (±882)
781 (±1057)
316 (±364)
526 (±906)
290 (±292)
SC
1
RI PT
Price 24
TE D
Missing
EP
surgery
M AN U
Prior abdominal
AC C
HUP, Hospital of the University of Pennsylvania; PAH, Pennsylvania Hospital; PPMC, Penn Presbyterian Medical Center; AA, African American; Income Quartile, median household income quartile; HMO/PPO, health maintenance organization/preferred provider organization. * Column totals may not add to 100 percent due to rounding.
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT
Price 25 Uterine Volume, cm³, calculated using the prolate ellipsoid formula of 4/3 x pi x L/2 x W/2 x AP/2 27
AC C
EP
TE D
M AN U
SC
RI PT
§
Social Determinants of Hysterectomy Route
438
ACCEPTED MANUSCRIPT Price 26 Table 2. Clinical and demographic variables according to race White*
AA*
Other*
% or mean
% or mean
% or mean
(±SD)
(±SD)
(±SD)
p
AA vs.
Characteristic
(N=659)
(N=978)
(N=75)
Overall
White
Age (y)
47 (±7)
45 (±7)
46 (±9)
<0.001
<0.001
BMI (kg/m²)
28 (±7)
32 (±7)
27 (±6)
<0.001
<0.001
<0.001
<0.001
0.080
0.370
<0.001
<0.001
<0.001
<0.001
Bleeding or Adenomyosis
26
Fibroids
73
RI PT
85
72
25
15
28
25
32
25
27
28
20
22
25
HUP
44
37
52
PAH
48
45
37
PPMC
7
19
11
59
10
55
2011 2012 2013
AC C
Hospital
†
SC 28
27
EP
2010
15
TE D
Year of Surgery
M AN U
Diagnosis
†
p
Income Quartile $64,000+
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT
$48-63,999
17
11
16
$38-47,999
11
18
9
$1-37,999
12
61
19
Private
33
36
35
HMO/PPO
55
45
57
Medicaid
4
11
4
Medicare
6
7
Missing/Self-pay
2
1
Prior cesarean delivery
26
Prior abdominal surgery
Uterine volume (cm³)
0
59
51
0.018
<0.001
23
27
0.366
0.274
52
59
47
<0.001
0.005
407 (±555)
562 (±647)
460 (±462)
<0.001
<0.001
668 (±905)
726 (±1652)
<0.001
<0.001
EP
Uterine weight (g)
<0.001
4
M AN U
59
TE D
Prior vaginal delivery
<0.001
SC
Insurance
RI PT
Price 27
430 (±690)
AC C
AA, African American; HUP, Hospital of the University of Pennsylvania; PAH, Pennsylvania Hospital; PPMC, Penn Presbyterian Medical Center; Income Quartile, median household income quartile; HMO/PPO, health maintenance organization/preferred provider organization. *Column totals may not add to 100 percent due to rounding †
Tests of significance determined by Pearson’s chi-square for categorical and Type III ANOVA for
continuous variables
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT
Price 28 Table 3. Univariate analyses of race, ethnicity and socioeconomic status on minimally invasive route of hysterectomy versus
All MIH* Variable
Vaginal p‡
OR (95% CI)
OR (95% CI)
p‡
OR (95% CI)
--
Robot-assisted p‡
--
M AN U
--
Laparoscopic
SC
Race White
RI PT
abdominal hysterectomy
OR (95% CI)
p‡
--
0.80 (0.65-0.97)
0.025
1.67 (1.21-2.30)
0.002
0.99 (0.76-1.31)
0.969
0.38 (0.29-0.50) <0.001
Other
0.83 (0.51-1.34)
0.441
0.90 (0.38-2.11)
0.807
0.72 (0.35-1.49)
0.362
0.87 (0.49-1.54)
Ethnicity
0.63 (0.39-1.00)
Income Quartile $64,000+ $48-63,999
--
--
0.049
0.43 (0.18-1.01)
EP
Hispanic
--
AC C
Not Hispanic
TE D
AA
1.03 (0.76-1.40)
0.842
Social Determinants of Hysterectomy Route
-0.054
--
1.78 (1.06-2.98)
0.58 (0.29-1.15)
-0.120
-0.032
1.02 (0.66-1.57)
0.635
0.83 (0.46-1.51)
0.541
-0.923
0.84 (0.57-1.23)
0.357
ACCEPTED MANUSCRIPT
Price 29 0.95 (0.71-1.28)
0.748
1.82 (1.11-2.97)
0.019
0.96 (0.63-1.45)
0.845
0.72 (0.49-1.04)
$1-37,999
0.89 (0.71-1.11)
0.303
2.51 (1.70-3.69)
<0.001 1.05 (0.77-1.43)
0.759
0.35 (0.25-0.49) <0.001
RI PT
$38-47,999
Insurance --
--
--
SC
Private
0.76 (0.52-1.10)
0.146
1.15 (0.70-1.89)
0.580
Medicare
1.15 (0.77-1.73)
0.486
1.29 (0.70-1.89)
0.376
HMO/PPO
1.22 (1.00-1.51)
0.056
0.91 (0.74-2.27)
0.562
-0.233
0.45 (0.24-0.86)
0.010
1.33 (0.78-2.27)
0.307
0.86 (0.46-1.58)
0.620
1.26 (0.95-1.68)
0.113
1.46 (1.10-1.94)
0.007
0.72 (0.42-1.25)
M AN U
Medicaid
0.078
MIH, minimally invasive hysterectomy; OR, odds ratio; CI, confidence interval; HUP, Hospital of the University of Pennsylvania;
TE D
PAH, Pennsylvania Hospital; PPMC, Penn Presbyterian Medical Center; AA, African American; Income Quartile, median household income quartile; HMO/PPO, health maintenance organization/preferred provider organization.
Tests of significance based on Pearson's Chi-square tests
AC C
‡
EP
*All MIH includes vaginal, laparoscopic, and robot-assisted hysterectomies
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT
Price 30
versus abdominal hysterectomy Vaginal
AOR†
95% CI
AOR†
--
--
--
AA
0.82
0.61-1.10
Other
0.90
Variable
95% CI
95% CI
--
--
1.38
0.86-2.21
1.06
0.73-1.56
0.57
0.39-0.82
0.52-1.56
1.65
0.58-4.68
0.99
0.45-2.15
0.82
0.43-1.58
--
--
--
--
--
--
--
--
0.45
0.27-0.76
0.24
0.09-0.64
0.42
0.20-0.89
0.78
0.39-1.53
--
--
--
--
--
TE D
--
Income Quartile $64,000+
AOR†
AC C
Hispanic
95% CI
--
Ethnicity Not Hispanic
Robot-assisted
--
EP
White
AOR†
M AN U
Race
Laparoscopic
SC
All MIH*
RI PT
Table 4. Multivariable logistic regression analysis of race, ethnicity and socioeconomic status on minimally invasive
--
--
Social Determinants of Hysterectomy Route
--
ACCEPTED MANUSCRIPT
$48-63,999
1.15
0.80-1.63
1.72
0.91-3.28
1.05
0.65-1.69
1.12
0.73-1.73
$38-47,999
1.07
0.74-1.52
1.53
0.82-2.85
0.89
0.55-1.45
1.13
0.72-1.75
$1-37,999
1.06
0.79-1.43
2.52
1.51-4.21
1.11
RI PT
Price 31
0.57
0.38-0.85
--
--
--
--
--
--
--
Medicaid
0.59
0.38-0.90
0.69
0.38-1.25
0.54
0.27-1.08
0.51
0.28-0.94
Medicare
0.96
0.60-1.56
0.90
0.43-1.90
0.73
0.37-1.47
1.35
0.73-2.49
HMO/PPO
1.04
0.82-1.32
0.66
0.44-0.99
1.27
0.93-1.74
1.09
0.79-1.50
SC
Private
--
M AN U
Insurance
0.75-1.64
TE D
MIH, minimally invasive hysterectomy; AOR, adjusted odds ratio; CI, confidence interval; AA, African American; Income Quartile, Median household income quartile; HMO/PPO, health maintenance organization/preferred
EP
provider organization.
†
AC C
*All minimally invasive hysterectomy includes vaginal, laparoscopic, and robot-assisted hysterectomies AOR adjusted for age, BMI, procedure year, hospital, obstetrical & surgical history, primary diagnosis, uterine
weight, and all other variables reported (race, ethnicity, median household income quartile, insurance status).
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT Price 32 FIGURE LEGEND
AC C
EP
TE D
M AN U
SC
RI PT
Figure 1. Flow diagram of study participants
Social Determinants of Hysterectomy Route
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
S0002-9378(17)30927-4
DOI:
10.1016/j.ajog.2017.07.036
Reference:
YMOB 11791
To appear in:
American Journal of Obstetrics and Gynecology
Received Date: 14 July 2017 Revised Date:
22 July 2017
Accepted Date: 31 July 2017
Please cite this article as: Price JT, Zimmerman LD, Koelper NC, Sammel MD, Lee S, Butts SF, Social determinants of access to minimally invasive hysterectomy: Re-evaluating the relationship between race and route of hysterectomy for benign disease, American Journal of Obstetrics and Gynecology (2017), doi: 10.1016/j.ajog.2017.07.036. 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.
ACCEPTED MANUSCRIPT Price 1 1
Social determinants of access to minimally invasive hysterectomy: Re-evaluating the relationship
2
between race and route of hysterectomy for benign disease
3
Running Title: Social Determinants of Hysterectomy Route
RI PT
4 5
Joan T PRICE MD MPH1; Lilli D ZIMMERMAN MD1; Nathan C KOELPER MPH1; Mary D
6
SAMMEL ScD2; Sonya LEE MD3; Samantha F BUTTS MD MSCE1
8
1
9
Medicine, Philadelphia, PA, USA
SC
7
M AN U
Department of Obstetrics and Gynecology, University of Pennsylvania Perelman School of
10
2
11
Biostatistics, University of Pennsylvania Perelman School of Medicine, Philadelphia, PA, USA
12
3
Department of Biostatistics and Epidemiology, Center for Clinical Epidemiology and
Department of Gynecology, Penn Presbyterian Medical Center, Philadelphia, PA, USA
TE D
13 14
Funding: None.
15
Conflicts of interest: The authors have no conflicts of interest to report.
EP
16 Corresponding author:
18
Joan T Price, MD MPH
19
Department of Obstetrics and Gynecology, UNC-Chapel Hill Medical Center
20
101 Manning Drive, Chapel Hill, NC 27514
21
tel: (919) 843-9546 email: [email protected]
22
Abstract word count: 374
23
Manuscript word count: 2,987
AC C
17
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT Price 2 Condensation
25
Demographic factors are associated with access to various types of minimally invasive
26
hysterectomy independent of body mass index, parity, surgical history, and uterine size.
27 28
Running head: Social Determinants of Hysterectomy Route
29
SC
30 31
M AN U
32 33 34 35
40 41 42 43 44
EP
39
AC C
38
TE D
36 37
RI PT
24
45 46
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT Price 3 ABSTRACT
48
BACKGROUND: Racial and socioeconomic disparities exist in access to medical and surgical
49
care. Studies of national databases have demonstrated disparities in route of hysterectomy for
50
benign indications, but have not been able to adjust for patient-level factors that affect surgical
51
decision-making.
RI PT
47
52
OBJECTIVES: To determine whether access to minimally invasive hysterectomy for benign
54
indications is differential according to race independent of the effects of relevant subject-level
55
confounding factors. The secondary study objective was to determine the association between
56
socioeconomic status and ethnicity and access to minimally invasive hysterectomy.
M AN U
SC
53
57
STUDY DESIGN: A cross-sectional study evaluated factors associated with minimally invasive
59
hysterectomies performed for fibroids and/or abnormal uterine bleeding between 2010 and 2013
60
at three hospitals within an academic university health system in Philadelphia, Pennsylvania.
61
Univariate tests of association and multivariable logistic regression identified factors
62
significantly associated with minimally invasive hysterectomy compared to the odds of treatment
63
with the referent approach of abdominal hysterectomy.
EP
AC C
64
TE D
58
65
RESULTS: Of 1746 hysterectomies evaluated meeting study inclusion criteria, 861 (49%) were
66
performed abdominally, 248 (14%) vaginally, 310 (18%) laparoscopically and 327 (19%) were
67
performed with robot assistance. In univariate analysis, African-American race (OR 0.80, 95%
68
CI 0.65–0.97) and Hispanic ethnicity (OR 0.63, 95% CI 0.39–1.00) were associated with lower
69
odds of any minimally invasive hysterectomy relative to abdominal hysterectomy. In analyses
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT Price 4 adjusted for age, BMI, income quartile, obstetrical and surgical history, uterine weight, and
71
additional confounding factors, African-American race was no longer a risk factor for reduced
72
minimally invasive hysterectomy (OR 0.82, 95% CI 0.61–1.10), while Hispanic ethnicity (OR
73
0.45, 95% CI 0.27–0.76) and Medicaid enrollment (OR 0.59, 95% CI 0.38–0.90) were associated
74
with significantly lower odds of treatment with any minimally invasive hysterectomy. In
75
adjusted analyses, African-American women had nearly half the odds of receiving robot-assisted
76
hysterectomy compared to whites (AOR 0.57, 95% CI 0.39-0.82), while no differences were
77
noted with other hysterectomy routes. Medicaid enrollment (compared to private insurance; OR
78
0.51, 95% CI 0.28–0.94) and lowest income quartile (compared to highest income quartile; OR
79
0.57, 95% CI 0.38–0.85) were also associated with diminished odds of robot-assisted
80
hysterectomy.
M AN U
SC
RI PT
70
81 CONCLUSION:
When accounting for the effect of numerous pertinent demographic and
83
clinical factors, the odds of undergoing minimally invasive hysterectomy were diminished in
84
women of Hispanic ethnicity and in those enrolled in Medicaid but were not discrepant along
85
racial lines. However, both racial and socioeconomic disparities were observed with respect to
86
access to robot-assisted hysterectomy despite the availability of robotic assistance in all hospitals
87
treating the study population. Strategies to ensure equal access to all minimally invasive routes
88
for all women should be explored to align delivery of care with the evidence supporting the
89
broad implementation of these procedures as safe, cost effective, and highly acceptable to
90
patients.
AC C
EP
TE D
82
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT Price 5 INTRODUCTION
92
Minimally invasive hysterectomy (MIH) has advantages over abdominal hysterectomy including
93
shorter hospital stay, improved postoperative pain, fewer complications and lower hospital
94
costs.1 The American College of Obstetrics and Gynecology encourages the application of MIH,
95
particularly vaginal hysterectomy, as safer and more cost effective than abdominal surgery for
96
benign indications.2 Despite established advantages of MIH, racial and socioeconomic disparities
97
exist in access to these approaches.3 The disparate application of minimally invasive procedures
98
among African Americans and socioeconomically disadvantaged groups has been noted across
99
many surgical specialties, including gynecologic surgery.4-7 National investigations of disparities
100
in MIH have found that younger age, white race, higher income and private insurance
101
independently predict laparoscopic over abdominal hysterectomy.8-11
M AN U
SC
RI PT
91
102
The primary aim of this study was to determine the association between race and the odds of
104
MIH for the treatment of fibroids, adenomyosis and abnormal bleeding. The secondary objective
105
was to assess the relationship between indicators of socioeconomic status, ethnicity and the odds
106
of MIH for the above indications. To build on findings from previous reports using national
107
databases, we sought to evaluate subject-level data related to hysterectomies performed at three
108
urban University-affiliated hospitals to more thoroughly control for confounding. Because
109
national databases often do not differentiate between traditional laparoscopic versus robot-
110
assisted routes of hysterectomy, we sought to determine whether certain individual minimally
111
invasive routes were more prone to racial and/or socioeconomic disparities in sub-group
112
analyses. Our hypothesis was that non-white race and low socioeconomic status would be
113
associated with diminished utilization of MIH despite adjusting for confounding due to patient-
AC C
EP
TE D
103
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT Price 6 114
level clinical predictors of route of hysterectomy. Elucidating disparities in access to the
115
recommended hysterectomy approaches is a critical step in improving equity and optimizing
116
outcomes for all women who require major gynecological surgery.
RI PT
117 MATERIALS AND METHODS
119
A cross-sectional study analyzed data from subjects receiving care at the three academic
120
hospitals within the University of Pennsylvania Health System (UPHS) in Philadelphia. Data
121
were gathered on all hysterectomies performed for fibroids, adenomyosis and/or abnormal
122
uterine bleeding at the Hospital of the University of Pennsylvania (HUP), Pennsylvania Hospital
123
(PAH) and Penn Presbyterian Medical Center (PPMC) from January 2010 through December
124
2013. As medical centers within an urban academic health system, each hospital has the
125
resources to perform abdominal, vaginal, laparoscopic, and robot-assisted hysterectomy. The
126
gynecology practice at PPMC is a two-provider high-volume private specialty practice whereas
127
the HUP and PAH sites comprise multiple practice types including private generalist, academic
128
sub-specialist and resident clinic practices that all perform hysterectomies for benign indications.
M AN U
TE D
EP
129
SC
118
Subjects who underwent abdominal, vaginal, laparoscopic and robot-assisted hysterectomy for
131
abnormal uterine bleeding, adenomyosis or fibroids were identified using diagnosis and
132
procedure codes from the International Classification of Disease, Ninth Revision, Clinical
133
Modification (ICD-9-CM). Exclusion criteria were: presence of known or suspected cancer,
134
known endometriosis, and cases performed jointly with another surgical service. Using reports
135
from the Penn Data Store clinical database and verified by electronic medical record review,
136
patient-level data were collected on demographic information, primary indication and mode of
AC C
130
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT Price 7 hysterectomy, self-identified race and ethnicity, median household income quartile (based on zip
138
code of residence),12 and insurance. Insurance status was categorized as (1) private insurance, (2)
139
health maintenance organization / preferred provider organization (HMO/PPO), which
140
comprised only non-Medicaid, non-Medicare managed plans; (3) Medicaid (including managed
141
Medicaid); (4) Medicare (including managed Medicare); and (5) self-pay or uninsured. Both
142
uterine weight and volume were abstracted from surgical pathology reports for each subject. The
143
final multivariable analyses adjusted only for uterine weight to avoid colinearity by including
144
both variables. Uterine weight was chosen as the more objective measure; volume estimates can
145
be imprecise when fibroids distort the uterine contour.
M AN U
SC
RI PT
137
146
The primary study outcome was mode of hysterectomy. Any abdominal hysterectomy including
148
total abdominal and abdominal supracervical hysterectomy was placed in the “abdominal”
149
category. Total vaginal hysterectomy without laparoscopic assistance was its own category
150
whereas
151
hysterectomy was coded as “laparoscopic”. In cases performed with robot assistance, the
152
hysterectomy was categorized as “robotic”. Hysterectomies were categorized under the initial
153
route of hysterectomy regardless of intra-operative conversion in order to reflect pre-operative
154
decision-making and surgical planning.
total
laparoscopic,
or
laparoscopic
supracervical
EP
laparoscopic-assisted,
AC C
155
any
TE D
147
156
Continuous variables were summarized using means and standard deviations, while frequencies
157
were used to describe the categorical variables. Age, BMI, uterine weight, and uterine volume
158
were treated as continuous variables. Uterine weight was transformed into a categorical variable
159
based on contingency tables detecting a frequency of at least two individuals in cross tabulation.
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT Price 8 160
Univariate comparisons of each of the potential predictors were performed with the dichotomous
161
outcomes of abdominal versus all MIH, vaginal, laparoscopic, and robotic hysterectomy, using
162
Pearson’s chi-square tests, Student’s t-tests or simple logistic regression as appropriate.
RI PT
163
Four logistic regression models were fit to determine the set of independent variables that best
165
predicted the odds of abdominal compared to any MIH, vaginal, laparoscopic, or robotic
166
hysterectomy. Backward selection was used to generate final models starting with all variables of
167
interest then excluding those that were not significant and successively re-fitting reduced models
168
until all remaining variables were statistically significant at p<0.05. Variables associated with
169
slightly higher p values (p<0.5) were retained if they were known risk factors for abdominal
170
hysterectomy or if the effect estimate for race changed by >10% when the covariate was
171
removed. Likelihood ratio testing was used to screen the robustness of covariates to generate a
172
parsimonious final model and to test for the presence of interactions. Statistical analyses were
173
performed using SAS 9.3 statistical software, Cary, NC.
TE D
M AN U
SC
164
174
The University of Pennsylvania Institutional Review Board granted approval to conduct this
176
study prior to the commencement of data collection; the requirement for informed consent was
177
waived.
AC C
178
EP
175
179
RESULTS
180
A total of 2166 hysterectomies were performed for the indications of fibroids, adenomyosis
181
and/or abnormal uterine bleeding at three UPHS hospitals between January 1, 2010 and
182
December 31, 2013. Of these cases, 420 (19%) were excluded, generating a sample of 1746
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT Price 9 183
eligible cases for analysis (Figure 1). Missing data were rare and addressed using listwise
184
deletion.
185 Most participants self-identified as either African American (57%, n=978) or white (38%,
187
n=659) (Table 1). More than half (55%) resided in zip codes representing neighborhoods in the
188
two lowest median household income quartiles (<$48,000 annually). Most subjects were covered
189
by either private insurance (35%) or managed care (50%).
SC
190
RI PT
186
Among 1746 hysterectomies, 49% (N=861) were performed abdominally, 14% (N=248)
192
vaginally and 37% (N=637) laparoscopically (Table 1). Of the laparoscopic hysterectomies, 327
193
(19% of total) were performed with robotic assistance. The primary indication for hysterectomy
194
was fibroids for 80% of participants.
TE D
195
M AN U
191
Most white subjects (54%) underwent MIH; more than half of those (28% overall) were
197
performed with robotic assistance. Among African-American women, 48% underwent MIH;
198
only 25% of those (12% overall) were robot-assisted. The proportion of whites treated with
199
vaginal hysterectomy was 9% while 18% of African Americans had hysterectomy performed
200
vaginally. Most vaginal hysterectomies (71%) were performed in African Americans.
AC C
201
EP
196
202
Examining the association of race and other clinical or demographic variables revealed that
203
African-American women were slightly younger (p<.001) with modestly higher BMIs compared
204
to white women (p<.001; Table 2). More African-American women than white women had
205
fibroids as their primary diagnosis (p<.001), had surgery performed at PPMC (p<.001), were in
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT Price 10 206
the lower income quartiles (p<.001), were covered by Medicaid (<.001), had a prior vaginal
207
delivery (p<.001), had a prior abdominal surgery (p=.005), and had larger uteri by both weight
208
and volume (p<.001).
RI PT
209
Univariate analysis demonstrated relationships between route of hysterectomy and race,
211
ethnicity, insurance status, and median household income quartile (Table 3). Comparing
212
abdominal hysterectomy to all minimally invasive routes combined, African-American patients
213
had lower odds of undergoing MIH than whites (OR 0.80, 95% CI 0.65–0.97), an association
214
that achieved borderline significance. African Americans had even lower odds than whites when
215
comparing robot-assisted hysterectomy alone to abdominal hysterectomy (OR 0.38, 95% CI
216
0.29–0.50). Conversely, African Americans had greater odds than whites of undergoing a vaginal
217
hysterectomy compared to abdominal hysterectomy (OR 1.67, 95% CI 1.21-2.30).
M AN U
SC
210
TE D
218
Neither household income quartile nor insurance status was associated with MIH over abdominal
220
hysterectomy in univariate analyses. However, women with Medicaid had significantly lower
221
odds of undergoing robot-assisted than abdominal hysterectomy (OR 0.45, 95% CI 0.24–0.86)
222
when compared to women with private insurance. Similarly, women in the lowest median
223
household income quartile had lower odds of treatment with robot-assisted hysterectomy (OR
224
0.35, 95% CI 0.25–0.49) compared to those residing in the highest quartile. Conversely, when
225
compared to those in the highest income quartile, women in the lowest quartile had higher odds
226
of vaginal hysterectomy than abdominal hysterectomy (OR 2.51, 95% CI 1.70–3.69).
AC C
EP
219
227
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT Price 11 In multivariable logistic regression models, we found that adjusting for subject-level factors
229
mitigated many of the univariate associations between route of hysterectomy and race, ethnicity,
230
median household income quartile, and insurance status (Table 4). When compared to white
231
women, African-American women had similar odds of undergoing any mode of MIH as they did
232
abdominal hysterectomy (AOR 0.82; 95% CI 0.61–1.10). In a subgroup multivariable analysis
233
investigating the robot-assisted route alone as the alternate to abdominal hysterectomy, African-
234
American women had significantly lower odds of robot-assisted hysterectomy compared to
235
whites (AOR 0.57, CI 0.39–0.82). Racial disparities in vaginal and laparoscopic hysterectomy
236
compared to abdominal hysterectomy were not statistically significant in adjusted analyses.
237
Compared to women of non-Hispanic ethnicity, self-identified Hispanic women had lower
238
adjusted odds of all MIH (AOR 0.45, 95% CI 0.27–0.76), which was predominantly driven by
239
significantly lower odds of both vaginal (AOR 0.24, 95% CI 0.09–0.64) and laparoscopic
240
hysterectomy (AOR 0.42, 95% CI 0.20–0.89), but not robotic hysterectomy.
SC
M AN U
TE D
241
RI PT
228
Median household income quartile and insurance coverage modestly predicted mode of
243
hysterectomy in multivariable regression. In the primary analysis, income quartile did not
244
predict MIH over abdominal hysterectomy, but women with Medicaid had significantly lower
245
odds of MIH compared to women with private insurance (AOR 0.59, 95% CI 0.38–0.90). In
246
separate subgroup analyses, women in the lowest income quartile (AOR 0.57, 95% CI 0.38–
247
0.85) or with Medicaid (AOR 0.51, 95% CI 0.28–0.94) both had nearly half the odds of robot-
248
assisted over abdominal hysterectomy than women in the highest income quartile or those with
249
private insurance, respectively. Conversely, women in the lowest income quartile had
250
significantly greater odds of undergoing vaginal over abdominal hysterectomy than those in the
AC C
EP
242
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT Price 12 251
highest quartile (AOR 2.52, 95% CI 1.51–4.21). Insurance status did not modify the association
252
between African-American race and mode of hysterectomy described in the multivariable
253
models.
RI PT
254 COMMENT
256
Our analysis demonstrates that racial and socioeconomic disparities in the application of
257
minimally invasive hysterectomy to treat benign gynecological disorders may be mitigated by
258
adjusting for patient-level confounding factors. In contrast to national registry studies, our
259
findings demonstrate that African-American women requiring hysterectomy for fibroids and/or
260
abnormal bleeding had similar odds as white women of treatment with MIH when controlling for
261
BMI, obstetrical and surgical history, and uterine size. By generating multivariable models that
262
adjusted for patient-level factors, we were able to expose the role of confounding on previously
263
described associations between race and access to MIH.4,8-11
M AN U
TE D
264
SC
255
In the United States; African-American race, low income, and underinsurance are all associated
266
with overweight and obesity.13 Obesity limits the application of laparoscopic and robot-assisted
267
surgery; obese women have greater odds of vaginal over other routes of hysterectomy.14
268
Similarly, fibroids disproportionately affect African-American women, who have both larger and
269
more numerous fibroids than whites.15-17 In our analysis, controlling for confounding due to BMI
270
and uterine weight certainly reduced the magnitude and significance of racial disparities noted in
271
unadjusted analyses. However, sensitivity analyses limited to procedures performed among
272
women with uterine weights <400g did not alter the findings of the remaining racial and
273
socioeconomic disparities in access to robotic hysterectomy. Large national databases often do
AC C
EP
265
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT Price 13 274
not report BMI or uterine size such that previous analyses of nationwide disparities in mode of
275
hysterectomy have not been able to account for their effect on surgical route.3,4,8-11
276 In our study population, disparities in access to MIH due to insurance and income status are less
278
stark than in national analyses, given the ability to more comprehensively address confounding
279
due to multiple factors. By using data from three hospitals within one urban hospital system in
280
Philadelphia, our analysis among a relatively uniform population may have biased some
281
socioeconomic associations towards the null. The small proportion of patients covered by
282
Medicaid in our sample could be due to the evolving composition of insurance provider types in
283
an era marked by insurance reforms under the Affordable Care Act. Improving insurance
284
coverage for nonwhite patients in states like Massachusetts has been met with near complete
285
resolution of racial disparities in minimally invasive surgeries.18 Similarly, in an urban setting
286
such as Philadelphia where neighborhoods and zip codes are more economically diverse than in
287
suburban and rural communities, median household income quartile based on zip code of
288
residence is likely an imperfect proxy for individual socioeconomic status. Unfortunately,
289
specific individual economic productivity data were not available for query in our medical
290
system’s database. Despite these limitations, we still observed that having private insurance
291
(compared to Medicaid) was independently associated with greater access to MIH. Being in the
292
highest income quartile (compared to the lowest) and having private insurance (compared to
293
Medicaid) also predicted robot-assisted hysterectomy over abdominal hysterectomy in adjusted
294
subgroup analyses. These findings are comparable to national reports that did not control for
295
subject-level confounding.3,4,8-11
AC C
EP
TE D
M AN U
SC
RI PT
277
296
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT Price 14 Previous analyses have shown racial and socioeconomic disparities in laparoscopic hysterectomy
298
overall but most have not evaluated the role of robot-assisted hysterectomy specifically, as
299
national databases rarely differentiate between laparoscopic surgeries performed with and
300
without robot assistance.4,8-11 In our subgroup analyses, adjusting for confounding in
301
multivariable analysis did eliminate many univariate associations between sociodemographic
302
factors and the outcome of MIH versus abdominal hysterectomy. In specifically investigating
303
robot-assisted hysterectomy, however, we found racial and economic disparities not apparent in
304
other MIH routes, even despite controlling for patient-level factors. The ability to test specific
305
associations that apply to discrete categories of MIH is a unique feature of this investigation. It is
306
plausible that, as a more established and cost-effective route when compared to robot-assisted
307
surgery,19,20 laparoscopic hysterectomy is more equally accessible – both regionally and across
308
diverse communities. The divergent associations of race and income on different modes of MIH
309
may explain why the combined outcome of MIH overall failed to demonstrate significant
310
disparity and supports the investigation of laparoscopic and robot-assisted hysterectomies
311
separately.
SC
M AN U
TE D
EP
312
RI PT
297
In contrast to trends in robot-assisted hysterectomy, the odds of undergoing vaginal
314
hysterectomy over abdominal hysterectomy were higher among women in the lowest income
315
quartile compared to those in the highest income quartile and among African Americans
316
compared to whites. Individual practice patterns and patient characteristics at the three hospitals
317
could explain the findings noted despite having controlled for hospital itself. It is conceivable
318
that the individual practices of the highest-volume surgeons at each hospital influenced our
319
findings. At PPMC, vaginal and laparoscopic hysterectomies are exceedingly more common than
AC C
313
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT Price 15 abdominal and robot-assisted hysterectomies and a greater proportion of the patients are low-
321
income and African American than at the other two hospitals. Restricting the analysis to PAH
322
and HUP did not alter the magnitude or significance of the associations with robot-assisted
323
compared to abdominal hysterectomy, but did further increase the odds of vaginal hysterectomy
324
among women in the lowest compared to highest income quartile. The differential application of
325
vaginal hysterectomy over abdominal hysterectomy to low-income women may be due to
326
employment status, less lenient work or childcare schedules, and other economic considerations,
327
which may not only favor vaginal over abdominal hysterectomy, but may also influence the
328
initial decision to perform hysterectomy over more conservative management.21-23 Overall,
329
vaginal hysterectomy trends within our health system appear to be mitigating rather than
330
potentiating disparities apparent in other minimally invasive routes.
M AN U
SC
RI PT
320
331
Although efforts to decrease the proportion of hysterectomies performed abdominally are
333
improving uptake of MIH overall, understanding and minimizing factors that may perpetuate
334
racial and socioeconomic inequities must be prioritized locally and nationally, particularly as
335
vaginal hysterectomies become increasingly rare and robot-assisted hysterectomies become
336
increasingly common.24,25 Vaginal hysterectomy is the preferred route for all patients for whom
337
it is feasible and safe, yet the proportion of hysterectomies performed vaginally remains low in
338
our population and nationwide.1,2 Gynecological training programs should continue to teach
339
vaginal hysterectomy as a low-cost minimally invasive approach that may also promote a more
340
equitable application of MIH to all women.2
AC C
EP
TE D
332
341
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT Price 16 This analysis has limitations. Despite the large, diverse, well-characterized cohort studied, the
343
data represent a retrospective analysis and uncontrolled confounding may exist. We chose to
344
exclude hysterectomies performed with known or suspected endometriosis due to the effect
345
adhesive disease can have on surgical decision-making. While this may limit the generalization
346
of our analysis, most benign hysterectomies in the United States are performed for fibroids or
347
abnormal bleeding.26 Similarly, the generalizability of our findings does not extend to other
348
hospital systems in other geographic locations and controlling for individual providers or
349
provider characteristics within our study was beyond the scope of our analysis. Future research
350
that purposefully selects a robust pool of data from hospitals across the country with diverse
351
patients, providers, and practice types could address confounding factors of both patients and
352
providers to better address these limitations.
M AN U
SC
RI PT
342
353
Our sub-regional analysis of racial and socioeconomic disparities in route of hysterectomy for
355
benign indications suggests that disparities reported in national analyses may be due to
356
uncontrolled patient-level confounding factors that influence surgical decision-making. We must
357
interpret national analyses of disparities with caution as we have shown that adjusting for
358
patient-level factors can mitigate and even eliminate associations found in unadjusted analyses.
359
Still, the persistence of some disparities in access to MIH approaches despite controlling for
360
multiple confounding factors remains a concern. The repeated publication of data revealing
361
racial and socioeconomic disparities in medical and surgical care demands a multidisciplinary
362
effort to carefully elucidate the source and true extent of these inequities, and collaborative
363
strategies to ensure equitable access to all modes of MIH for racial minorities, the underinsured,
364
and low-income women.
AC C
EP
TE D
354
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT Price 17 365
Acknowledgements: The authors would like to thank the Penn Data Store for assistance in
366
assembling information used in this study. We thank Dr. Robert L. Dood for his contribution to
367
preliminary data analysis.
AC C
EP
TE D
M AN U
SC
RI PT
368
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT Price 18 369
REFERENCES
370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412
1.
7.
8.
9. 10. 11.
12. 13. 14.
15.
16.
17.
RI PT
SC
6.
M AN U
5.
TE D
4.
EP
3.
AC C
2.
Aarts JW, Nieboer TE, Johnson N, et al. Surgical approach to hysterectomy for benign gynaecological disease. Cochrane Database Syst Rev. 2015;8:CD003677. ACOG Committee Opinion No. 444: choosing the route of hysterectomy for benign disease. Obstet Gynecol. 2009;114(5):1156-1158. Abenhaim HA, Azziz R, Hu J, Bartolucci A, Tulandi T. Socioeconomic and racial predictors of undergoing laparoscopic hysterectomy for selected benign diseases: analysis of 341487 hysterectomies. J Minim Invasive Gynecol. 2008;15(1):11-15. Jacoby VL, Autry A, Jacobson G, Domush R, Nakagawa S, Jacoby A. Nationwide use of laparoscopic hysterectomy compared with abdominal and vaginal approaches. Obstet Gynecol. 2009;114(5):1041-1048. Varela JE, Nguyen NT. Disparities in access to basic laparoscopic surgery at U.S. academic medical centers. Surg Endosc. 2011;25(4):1209-1214. Robinson CN, Balentine CJ, Sansgiry S, Berger DH. Disparities in the use of minimally invasive surgery for colorectal disease. J Gastrointest Surg. 2012;16(5):897-903; discussion 903-894. Trinh QD, Schmitges J, Sun M, et al. Improvement of racial disparities with respect to the utilization of minimally invasive radical prostatectomy in the United States. Cancer. 2012;118(7):1894-1900. Esselen KM, Vitonis A, Einarsson J, Muto MG, Cohen S. Health Care Disparities in Hysterectomy for Gynecologic Cancers: Data From the 2012 National Inpatient Sample. Obstet Gynecol. 2015;126(5):1029-1039. Lee J, Jennings K, Borahay MA, et al. Trends in the national distribution of laparoscopic hysterectomies from 2003 to 2010. J Minim Invasive Gynecol. 2014;21(4):656-661. Patel PR, Lee J, Rodriguez AM, et al. Disparities in use of laparoscopic hysterectomies: a nationwide analysis. J Minim Invasive Gynecol. 2014;21(2):223-227. Price JT, Lee S. National Trends and Disparities in Minimally Invasive Approaches to Hysterectomy for Benign Indications. Journal of Minimally Invasive Gynecology. 20(6):S11-S12. American Community Survey, 2013 American Community Survey 3-year estimates, Table B19013. http://factfinder2.census.gov. Ogden CLCL. Prevalence of childhood and adult obesity in the United States, 2011-2012. JAMA : the journal of the American Medical Association.311(8):806. Mikhail E, Miladinovic B, Velanovich V, Finan MA, Hart S, Imudia AN. Association between obesity and the trends of routes of hysterectomy performed for benign indications. Obstet Gynecol. 2015;125(4):912-918. Eltoukhi HM, Modi MN, Weston M, Armstrong AY, Stewart EA. The health disparities of uterine fibroid tumors for African American women: a public health issue. Am J Obstet Gynecol. 2014;210(3):194-199. Baird DD, Dunson DB, Hill MC, Cousins D, Schectman JM. High cumulative incidence of uterine leiomyoma in black and white women: ultrasound evidence. Am J Obstet Gynecol. 2003;188(1):100-107. Moore AB, Flake GP, Swartz CD, et al. Association of race, age and body mass index with gross pathology of uterine fibroids. J Reprod Med. 2008;53(2):90-96.
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT Price 19
21. 22. 23. 24. 25. 26.
EP
TE D
27.
RI PT
20.
SC
19.
Loehrer AP, Song Z, Auchincloss HG, Hutter MM. Massachusetts health care reform and reduced racial disparities in minimally invasive surgery. JAMA Surg. 2013;148(12):11161122. Wright JD, Ananth CV, Lewin SN, et al. Robotically assisted vs laparoscopic hysterectomy among women with benign gynecologic disease. JAMA. 2013;309(7):689698. Pasic RP, Rizzo JA, Fang H, Ross S, Moore M, Gunnarsson C. Comparing robot-assisted with conventional laparoscopic hysterectomy: impact on cost and clinical outcomes. J Minim Invasive Gynecol. 2010;17(6):730-738. Marks NF, Shinberg DS. Socioeconomic differences in hysterectomy: the Wisconsin Longitudinal Study. Am J Public Health. 1997;87(9):1507-1514. Materia E, Rossi L, Spadea T, et al. Hysterectomy and socioeconomic position in Rome, Italy. J Epidemiol Community Health. 2002;56(6):461-465. Dharmalingam A, Pool I, Dickson J. Biosocial determinants of hysterectomy in New Zealand. Am J Public Health. 2000;90(9):1455-1458. Committee on Gynecologic P. Committee Opinion No 701: Choosing the Route of Hysterectomy for Benign Disease. Obstet Gynecol. 2017;129(6):e155-e159. Wright JD, Herzog TJ, Tsui J, et al. Nationwide trends in the performance of inpatient hysterectomy in the United States. Obstet Gynecol. 2013;122(2 Pt 1):233-241. Kho KA, Lin K, Hechanova M, Richardson DL. Risk of Occult Uterine Sarcoma in Women Undergoing Hysterectomy for Benign Indications. Obstet Gynecol. 2016;127(3):468-473. Shirlina D, Shirish S. Uterine volume: an aid to determine the route and technique of hysterectomy. J Obstet Gynecol Ind. 2004;54(1):68-72.
M AN U
18.
AC C
413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT
Table 1. Baseline Characteristics of 1746 women who underwent hysterectomy for benign indications, 2010-2013 Abdominal
Vaginal
Mean (±SD)
Mean (±SD)
Mean (±SD)
or % (N)*
or % (N)*
or % (N)*
Total
100 (1746)
49 (861)
14 (248)
Age, y
45.9 (±7.2)
46.3 (±6.7)
BMI, kg/m²
30.4 (±7.2)
30.4 (±7.3)
1 (23)
0 (2)
Missing
or % (N)*
or % (N)*
18 (310)
19 (327)
44.8 (±8.0)
45.1 (±7.2)
46.6 (±7.9)
31.9 (±7.6)
30.3 (±7.0)
29.7 (±6.9)
2 (6)
2 (5)
3 (10)
M AN U
SC
Mean (±SD)
EP
Diagnosis
Mean (±SD)
TE D
Characteristic
Laparoscopic Robot-assisted
RI PT
All
AC C
Bleeding or Adenomyosis
20 (356)
13 (113)
34 (84)
25 (77)
25 (82)
Fibroids
80 (1390)
87 (748)
66 (164)
75 (233)
75 (245)
Social Determinants of Hysterectomy Route
Price 20
ACCEPTED MANUSCRIPT
Price 21
2010
25 (444)
26 (227)
23 (56)
2011
27 (464)
28 (242)
26 (64)
2012
26 (459)
25 (211)
28 (69)
RI PT
Year of Surgery
2013
22 (379)
21 (181)
24 (59)
HUP
40 (699)
39 (339)
PAH
46 (802)
PPMC
14 (245)
20 (65)
22 (68)
28 (90)
24 (74)
32 (105)
23 (72)
21 (67)
28 (70)
47 (146)
44 (144)
58 (502)
25 (63)
18 (56)
55 (181)
2 (20)
46 (115)
35 (108)
1 (2)
M AN U
EP
TE D
Hospital
SC
31 (96)
AC C
Race White
38 (659)
35 (303)
25 (62)
36 (111)
56 (183)
AA
56 (978)
59 (505)
69 (172)
59 (184)
36 (117)
4 (75)
4 (38)
3 (7)
3 (10)
6 (20)
Other
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT
2 (34)
2 (15)
3 (7)
2 (5)
2 (7)
95 (1666)
95 (812)
98 (242)
97 (300)
95 (312)
Hispanic
5 (78)
5 (47)
2 (6)
RI PT
Price 22 Missing
3 (10)
5 (15)
Missing
0 (2)
0 (2)
0 (0)
0 (0)
0 (0)
$64,000+
31 (539)
30 (259)
16 (40)
30 (92)
45 (148)
$48-63,999
14 (239)
13 (113)
13 (31)
13 (41)
17 (54)
$38-47,999
15 (268)
15 (132)
15 (37)
14 (45)
17 (54)
$1-37,999
40 (694)
41 (354)
55 (137)
43 (132)
22 (71)
0 (3)
1 (3)
0 (0)
0 (0)
37 (318)
37 (91)
33 (102)
32 (103)
Insurance Private
EP
0 (6)
AC C
Missing
TE D
Income Quartile
M AN U
Not Hispanic
SC
Ethnicity
35 (614)
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT
49 (862)
47 (403)
42 (105)
53 (163)
58 (191)
Medicaid
8 (140)
10 (82)
11 (27)
6 (19)
4 (12)
Medicare
6 (112)
6 (54)
8 (20)
7 (23)
5 (15)
Missing/Self-pay
1 (18)
0 (4)
2 (5)
1 (3)
2 (6)
Delivery
1.5 (±1.4)
1.3 (±1.4)
2.2 (±1.6)
1.6 (±1.5)
1.3 (±1.4)
0
34 (597)
39 (339)
18 (45)
30 (93)
37 (120)
1-2
42 (742)
41 (356)
40 (98)
45 (141)
45 (147)
3-4
19 (331)
16 (136)
34 (84)
18 (56)
17 (55)
≥5
4 (52)
2 (19)
6 (15)
5 (14)
1 (4)
Missing
1 (24)
1 (11)
2 (6)
2 (6)
0 (1)
SC
HMO/PPO
RI PT
Price 23
TE D
EP
AC C
Prior Cesarean
M AN U
Prior Vaginal
Delivery
0.4 (±0.8)
0.4 (±0.8)
0.2 (±0.6)
0.4 (±0.7)
0.6(±0.9)
0
74 (1298)
75 (644)
85 (210)
74 (229)
66 (215)
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT
14 (240)
14 (118)
8 (21)
15 (46)
17 (55)
2
8 (139)
7 (63)
3 (8)
8 (24)
13 (44)
≥3
3 (48)
3 (27)
2 (4)
2 (5)
4 (12)
Missing
1 (21)
1 (9)
2 (5)
2 (6)
0 (1)
56 (974)
54 (467)
59 (147)
60 (185)
53 (175)
1 (16)
1 (9)
1 (2)
1 (4)
0 (1)
Uterine weight (g)
497 (±610)
716 (±775)
237 (±148)
315 (±299)
288 (±219)
Uterine volume§
579 (±882)
781 (±1057)
316 (±364)
526 (±906)
290 (±292)
SC
1
RI PT
Price 24
TE D
Missing
EP
surgery
M AN U
Prior abdominal
AC C
HUP, Hospital of the University of Pennsylvania; PAH, Pennsylvania Hospital; PPMC, Penn Presbyterian Medical Center; AA, African American; Income Quartile, median household income quartile; HMO/PPO, health maintenance organization/preferred provider organization. * Column totals may not add to 100 percent due to rounding.
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT
Price 25 Uterine Volume, cm³, calculated using the prolate ellipsoid formula of 4/3 x pi x L/2 x W/2 x AP/2 27
AC C
EP
TE D
M AN U
SC
RI PT
§
Social Determinants of Hysterectomy Route
438
ACCEPTED MANUSCRIPT Price 26 Table 2. Clinical and demographic variables according to race White*
AA*
Other*
% or mean
% or mean
% or mean
(±SD)
(±SD)
(±SD)
p
AA vs.
Characteristic
(N=659)
(N=978)
(N=75)
Overall
White
Age (y)
47 (±7)
45 (±7)
46 (±9)
<0.001
<0.001
BMI (kg/m²)
28 (±7)
32 (±7)
27 (±6)
<0.001
<0.001
<0.001
<0.001
0.080
0.370
<0.001
<0.001
<0.001
<0.001
Bleeding or Adenomyosis
26
Fibroids
73
RI PT
85
72
25
15
28
25
32
25
27
28
20
22
25
HUP
44
37
52
PAH
48
45
37
PPMC
7
19
11
59
10
55
2011 2012 2013
AC C
Hospital
†
SC 28
27
EP
2010
15
TE D
Year of Surgery
M AN U
Diagnosis
†
p
Income Quartile $64,000+
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT
$48-63,999
17
11
16
$38-47,999
11
18
9
$1-37,999
12
61
19
Private
33
36
35
HMO/PPO
55
45
57
Medicaid
4
11
4
Medicare
6
7
Missing/Self-pay
2
1
Prior cesarean delivery
26
Prior abdominal surgery
Uterine volume (cm³)
0
59
51
0.018
<0.001
23
27
0.366
0.274
52
59
47
<0.001
0.005
407 (±555)
562 (±647)
460 (±462)
<0.001
<0.001
668 (±905)
726 (±1652)
<0.001
<0.001
EP
Uterine weight (g)
<0.001
4
M AN U
59
TE D
Prior vaginal delivery
<0.001
SC
Insurance
RI PT
Price 27
430 (±690)
AC C
AA, African American; HUP, Hospital of the University of Pennsylvania; PAH, Pennsylvania Hospital; PPMC, Penn Presbyterian Medical Center; Income Quartile, median household income quartile; HMO/PPO, health maintenance organization/preferred provider organization. *Column totals may not add to 100 percent due to rounding †
Tests of significance determined by Pearson’s chi-square for categorical and Type III ANOVA for
continuous variables
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT
Price 28 Table 3. Univariate analyses of race, ethnicity and socioeconomic status on minimally invasive route of hysterectomy versus
All MIH* Variable
Vaginal p‡
OR (95% CI)
OR (95% CI)
p‡
OR (95% CI)
--
Robot-assisted p‡
--
M AN U
--
Laparoscopic
SC
Race White
RI PT
abdominal hysterectomy
OR (95% CI)
p‡
--
0.80 (0.65-0.97)
0.025
1.67 (1.21-2.30)
0.002
0.99 (0.76-1.31)
0.969
0.38 (0.29-0.50) <0.001
Other
0.83 (0.51-1.34)
0.441
0.90 (0.38-2.11)
0.807
0.72 (0.35-1.49)
0.362
0.87 (0.49-1.54)
Ethnicity
0.63 (0.39-1.00)
Income Quartile $64,000+ $48-63,999
--
--
0.049
0.43 (0.18-1.01)
EP
Hispanic
--
AC C
Not Hispanic
TE D
AA
1.03 (0.76-1.40)
0.842
Social Determinants of Hysterectomy Route
-0.054
--
1.78 (1.06-2.98)
0.58 (0.29-1.15)
-0.120
-0.032
1.02 (0.66-1.57)
0.635
0.83 (0.46-1.51)
0.541
-0.923
0.84 (0.57-1.23)
0.357
ACCEPTED MANUSCRIPT
Price 29 0.95 (0.71-1.28)
0.748
1.82 (1.11-2.97)
0.019
0.96 (0.63-1.45)
0.845
0.72 (0.49-1.04)
$1-37,999
0.89 (0.71-1.11)
0.303
2.51 (1.70-3.69)
<0.001 1.05 (0.77-1.43)
0.759
0.35 (0.25-0.49) <0.001
RI PT
$38-47,999
Insurance --
--
--
SC
Private
0.76 (0.52-1.10)
0.146
1.15 (0.70-1.89)
0.580
Medicare
1.15 (0.77-1.73)
0.486
1.29 (0.70-1.89)
0.376
HMO/PPO
1.22 (1.00-1.51)
0.056
0.91 (0.74-2.27)
0.562
-0.233
0.45 (0.24-0.86)
0.010
1.33 (0.78-2.27)
0.307
0.86 (0.46-1.58)
0.620
1.26 (0.95-1.68)
0.113
1.46 (1.10-1.94)
0.007
0.72 (0.42-1.25)
M AN U
Medicaid
0.078
MIH, minimally invasive hysterectomy; OR, odds ratio; CI, confidence interval; HUP, Hospital of the University of Pennsylvania;
TE D
PAH, Pennsylvania Hospital; PPMC, Penn Presbyterian Medical Center; AA, African American; Income Quartile, median household income quartile; HMO/PPO, health maintenance organization/preferred provider organization.
Tests of significance based on Pearson's Chi-square tests
AC C
‡
EP
*All MIH includes vaginal, laparoscopic, and robot-assisted hysterectomies
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT
Price 30
versus abdominal hysterectomy Vaginal
AOR†
95% CI
AOR†
--
--
--
AA
0.82
0.61-1.10
Other
0.90
Variable
95% CI
95% CI
--
--
1.38
0.86-2.21
1.06
0.73-1.56
0.57
0.39-0.82
0.52-1.56
1.65
0.58-4.68
0.99
0.45-2.15
0.82
0.43-1.58
--
--
--
--
--
--
--
--
0.45
0.27-0.76
0.24
0.09-0.64
0.42
0.20-0.89
0.78
0.39-1.53
--
--
--
--
--
TE D
--
Income Quartile $64,000+
AOR†
AC C
Hispanic
95% CI
--
Ethnicity Not Hispanic
Robot-assisted
--
EP
White
AOR†
M AN U
Race
Laparoscopic
SC
All MIH*
RI PT
Table 4. Multivariable logistic regression analysis of race, ethnicity and socioeconomic status on minimally invasive
--
--
Social Determinants of Hysterectomy Route
--
ACCEPTED MANUSCRIPT
$48-63,999
1.15
0.80-1.63
1.72
0.91-3.28
1.05
0.65-1.69
1.12
0.73-1.73
$38-47,999
1.07
0.74-1.52
1.53
0.82-2.85
0.89
0.55-1.45
1.13
0.72-1.75
$1-37,999
1.06
0.79-1.43
2.52
1.51-4.21
1.11
RI PT
Price 31
0.57
0.38-0.85
--
--
--
--
--
--
--
Medicaid
0.59
0.38-0.90
0.69
0.38-1.25
0.54
0.27-1.08
0.51
0.28-0.94
Medicare
0.96
0.60-1.56
0.90
0.43-1.90
0.73
0.37-1.47
1.35
0.73-2.49
HMO/PPO
1.04
0.82-1.32
0.66
0.44-0.99
1.27
0.93-1.74
1.09
0.79-1.50
SC
Private
--
M AN U
Insurance
0.75-1.64
TE D
MIH, minimally invasive hysterectomy; AOR, adjusted odds ratio; CI, confidence interval; AA, African American; Income Quartile, Median household income quartile; HMO/PPO, health maintenance organization/preferred
EP
provider organization.
†
AC C
*All minimally invasive hysterectomy includes vaginal, laparoscopic, and robot-assisted hysterectomies AOR adjusted for age, BMI, procedure year, hospital, obstetrical & surgical history, primary diagnosis, uterine
weight, and all other variables reported (race, ethnicity, median household income quartile, insurance status).
Social Determinants of Hysterectomy Route
ACCEPTED MANUSCRIPT Price 32 FIGURE LEGEND
AC C
EP
TE D
M AN U
SC
RI PT
Figure 1. Flow diagram of study participants
Social Determinants of Hysterectomy Route
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT