FDA Safety Communication on the Use of Transvaginal Mesh in Pelvic Organ Prolapse Repair Surgery: The Impact of Social Determinants of Health

Abstract

Objective

To examine the impact of the July 2011 FDA safety communication on the use of transvaginal mesh in pelvic organ prolapse (POP) repair by patients’ race and ethnicity.

Methods

We conducted an observational cohort study of women undergoing POP repairs during 2008–2015 in New York State. We examined the changes in transvaginal mesh use in POP repairs before and after the FDA communication by patients’ race and ethnicity. Piecewise logistic regression models were used to assess the trends of mesh use, adjusting for patient characteristics. We performed a subgroup analysis of the trends of transvaginal mesh use by racial groups, stratifying by patients’ neighborhood socioeconomic status (SES).

Results

We included 49,848 women (78% white, 7% black, and 15% Hispanic) with an average age of 60.2(SD: ±13.0) years. Following the safety communication, the use of transvaginal mesh in POP repairs decreased among white women (OR 0.45, 95% CI 0.41–0.50) and African Americans (OR 0.48, 95% CI 0.35–0.67), but remained stable among Hispanic women (OR 0.88, 95% CI 0.70–1.11). Only in the subgroup of patients from high-income areas, there was a trend toward decreasing mesh use among Hispanic patients after 2011 (OR 0.71, 95% CI 0.49–1.04).

Conclusion

The communications related to the safety of transvaginal mesh did not have an equal impact across racial groups. Mesh use decreased among white and African American women but not among Hispanic women following the 2011 FDA safety communication. Particular attention is warranted for patients from disadvantaged groups, especially low-income minorities, when disseminating medical device safety messages.

Summary

Mesh use decreased among white and African American women but not among Hispanic women following the 2011 FDA communication on the safety of transvaginal mesh used in pelvic organ prolapse repair.

Introduction

Polypropylene mesh is used in pelvic organ prolapse (POP) repair to reinforce weakened tissue and was first approved by the US Food and Drug Administration (FDA) in 2002. It came under scrutiny in recent years due to increased adverse event reports related to its short- and long-term complications. After receiving thousands of reports, the FDA released a public health notification (PHN) in 2008 concerning the safety of transvaginal mesh¹ and an updated safety communication in 2011². Transvaginal mesh was eventually removed from the US market in April of 2019 due to the lack of evidence to assure its safety.³ Studies have found that transvaginal mesh use in POP repair was associated with an increased risk of reoperation in both short and long term.^4–6 It was demonstrated that despite the PHN in 2008, mesh use continued to rise between 2008 and 2011.⁴ But following the updated safety communication in 2011, mesh use in POP started to decrease, possibly due to the stronger language and more rigorous evidence being cited in the FDA communication.^7,8

Regulatory safety message, as a form of new knowledge, informs providers and their patients of the risks of a medical treatment, which has been shown to affect the care that patients receive.^7,9,10 However, the emergence of new knowledge often contributes to increased disparities in healthcare.¹¹ Drug safety warnings have been associated with differential decreases in the use of the drug in question across patients’ racial and socioeconomic status (SES) groups.^12,13 Regulatory safety messages help raise patients’ awareness of risks associated with medical devices or pharmaceutical agents. Varying uptake of safety messages related to medical technologies and treatments, if any, represents an important aspect from which health disparities could arise and has downstream implications on patient safety.

It is unknown whether the 2011 FDA communication concerning the safety of transvaginal mesh in POP repair had a differential impact on its use across patient racial groups. In this study, we sought to examine the use of transvaginal mesh in POP repair before and after the 2011 FDA communication by race and ethnicity. We further explored whether residential SES affected the differences in the change of mesh use.

Materials and Methods

Data Source and Study Population

We used data from the New York State Department of Health Statewide Planning and Research Cooperative System (SPARCS),¹⁴ an all age group and all-payer database that collects patient and treatment information for every hospital discharge, outpatient and ambulatory surgery, and emergency department visit in the state. Patients covered by Medicare, Medicaid, and private insurance are all included in the database. SPARCS performs a periodic audit to ensure data quality.

Our study included adult women undergoing pelvic organ prolapse repair in inpatient and outpatient surgery settings between January 2008 and September 2015 in New York State. Procedures were identified based on ICD-9-CM (International classification of diseases, ninth revision, clinical modification) and CPT-4 (Current Procedural Terminology coding system, fourth edition) codes (Appendix Table).⁴ Mesh was defined as any augmenting material, including synthetic and biologic materials, which was consistent with our previous study. We excluded sacrocolpopexy because the FDA communication targeted transvaginal mesh use only. We further restricted to non-Hispanic white women, African American women, and Hispanic women for the analysis. Other racial groups (e.g. non-Hispanic Asian, multiracial) were small. Grouping them would lead to a large heterogeneity in the “other race” group. Therefore, they were not included in the analysis.

Key Variables

The primary outcome of interest was the use of transvaginal mesh in POP repair surgeries. Mesh use was defined by procedure codes, which were outlined in the appendix Table.⁴ The primary predictors of interest were time before and after the 2011 FDA communication and patients’ race and ethnicity (non-Hispanic white, non-Hispanic African American, and Hispanic). We analyzed time by year for the descriptive analysis and by quarter in the regression analysis.

Patient-level covariates examined were patient age, insurance payer (Medicare, Medicaid, commercial, and other), procedure setting (inpatient and outpatient), previous POP surgery, concurrent hysterectomy and sling procedures, the use of apical prolapse repair, comorbidities, and residential SES. Common comorbidities were identified using ICD algorithms validated by Elixhauser.¹⁵ Patients’ residential SES variables were obtained by linking the 5-digit zip code to the 2015 American Community Survey.¹⁶ SES variables analyzed in this study were median household income and proportion of residents graduating high school by zip code.

Statistical Analysis

We examined baseline characteristics at the time of the procedure by patients’ race and ethnicity. For the comparison of patient characteristics between groups, we used one-way ANOVA for age and Chi-square tests for categorical variables. We compared the use of transvaginal mesh in 2008, 2011, and 2015 across groups using Chi-square tests. We examined the unadjusted trends in the use of mesh in POP repairs over time by patients’ race and ethnicity. We used Cochran-Armitage trend tests to examine the crude trends in mesh use before and after 2011, respectively.

Piecewise multivariable logistic regression models were used to assess the trends in transvaginal mesh use in POP surgery before and after the 2011 FDA communication, and to compare the trends between racial groups. Time was modeled as a continuous variable in the unit of quarters in the regression analysis. A piecewise regression model is a spline model with one inflexion point, where the two continuous segments join. Because the FDA communication was released in July 2011, the inflexion point for our piecewise modeling was determined to be the third quarter of 2011. We estimated the change in mesh use from the beginning of the study to the second quarter of 2011 (pre-communication period) and the change from the third quarter of 2011 to the end of the study (post-communication period). The multivariable model adjusted for patient age, comorbidities, and procedural characteristics (listed in Table 1).

Table 1.

Cohort characteristics by patient race and ethnicity.

	Non-Hispanic White (N=38820)	African American (N=3749)	Hispanic (N=7279)	P value
Age
Mean(SD)	61.2(12.9)	58.6(12.6)	56.0(13.2)	<.001
Insurance				<.001
Medicare	15062(38.8%)	1197(31.9%)	1990(27.3%)
Medicaid	1903(4.9%)	599(16.0%)	2289(31.5%)
Commercial	20883(53.8%)	1816(48.4%)	2696(37.0%)
Other	956(2.5%)	137(3.7%)	303(4.2%)
Procedural
POP procedure type				<.001
Mesh	10758(27.7%)	728(19.4%)	1373(18.9%)
Non-mesh	28062(72.3%)	3021(80.6%)	5906(81.1%)
Outpatient procedure	15132(39.0%)	1020(27.2%)	2388(32.8%)	<.001
Previous POP surgery	2814(7.2%)	135(3.6%)	361(5.0%)	<.001
Concurrent Hysterectomy	17522(45.1%)	1937(51.7%)	3180(43.7%)	<.001
Concurrent Sling	16164(41.6%)	1254(33.4%)	3400(46.7%)	<.001
Apical procedure	22997(59.2%)	2384(63.6%)	3723(51.1%)	<.001
Comorbidities
CAD	1660(4.3%)	136(3.6%)	213(2.9%)	<.001
Hypertension	14336(36.9%)	1846(49.2%)	2341(32.2%)	<.001
CHF	243(0.6%)	48(1.3%)	29(0.4%)	<.001
Diabetes	3485(9.0%)	716(19.1%)	995(13.7%)	<.001
CPD	4141(10.7%)	407(10.9%)	804(11.0%)	.61
Obesity	2046(5.3%)	384(10.2%)	322(4.4%)	<.001
Anemia	1237(3.2%)	290(7.7%)	278(3.8%)	<.001
PVD	420(1.1%)	25(0.7%)	35(0.5%)	<.001
Depression	2899(7.5%)	119(3.2%)	308(4.2%)	<.001
CKD	301(0.8%)	64(1.7%)	36(0.5%)	<.001
No. Comorbidities				<.001
0	18820(48.5%)	1428(38.1%)	3946(54.2%)
1	12202(31.4%)	1114(29.7%)	1864(25.6%)
2+	7798(20.1%)	1207(32.2%)	1469(20.2%)
Residential SES*
Median income				<.001
Low	15790(40.7%)	2527(67.4%)	4848(66.6%)
High	21820(56.2%)	1166(31.1%)	2321(31.9%)
% High school graduate				<.001
Low	16173(41.7%)	2654(70.8%)	5434(74.7%)
High	21437(55.2%)	1039(27.7%)	1735(23.8%)

Abbreviations: POP=Pelvic Organ Prolapse, CAD=Coronary Artery Disease, CHF=Congestive Heart Failure, CPD=Chronic Pulmonary Disease, PVD=Peripheral Vascular Disease, CKD=Chronic Kidney Disease

^*2.8% of the patients were not New York State residents and did not have residential SES matched. Residential median household income and proportion of residents graduating high school were categorized into high and low based on median (income: $43,000; high school graduate: 50%)

To explore whether SES factors were related to the differential use of mesh across racial groups after 2011, we also conducted a subgroup analysis to assess the trends in mesh use in POP surgeries by patients’ race and ethnicity within categories of residential SES. The trends in mesh use were examined among patients from areas with high and low zip code income and education levels, respectively. Residential median household income and proportion of residents graduating high school were categorized into high and low for the subgroup analysis based on median (income: $43,000; high school graduate: 50%) to maximize power. Separate regression models were fit for each group. A p-value of <0.05 was determined statistically significant. All analyses were performed using SAS v9.3 (SAS Institute Inc., Cary, NC).

Results

Cohort Characteristics

Our cohort consisted of 49,848 women (78% non-Hispanic white, 7% African American, and 15% Hispanic) who underwent surgery for POP repair in New York State between Jan 2008 and Sept 2015. The average age at surgery was 60.2 (±13.0) years, and half of the patients had commercial insurance. Transvaginal mesh was used in 22.3% of the POP procedures at the beginning of the study period in 2008, 31.6% of the procedures at the time of the FDA communication in 2011, and 18.9% of the procedures at the end of the study in 2015.

When examining cohort characteristics by patients’ race and ethnicity, African American women and Hispanic women undergoing POP repairs were younger, more likely to have Medicaid insurance, and less likely to receive the procedure in outpatient settings (Table 1). African American patients were more likely to undergo a concurrent hysterectomy, but less likely to have a concurrent sling, when compared with white and Hispanic women. African American and Hispanic women were more likely to reside in low SES areas.

Trends of mesh use

In 2008, 24.0% of non-Hispanic white women, 20.0% of African American women, and 12.1% of Hispanic women received mesh-based POP repairs (p<0.001, Figure 1). In 2011, 34.1% of non-Hispanic whites, 23.4% of African Americans, and 22.5% of Hispanic women underwent mesh-based POP repairs. From 2008 to 2011, the use of transvaginal mesh increased among non-Hispanic white and Hispanic women (both trend p<0.001), but not among African American women (p=0.22). From 2011 to 2015, the use of mesh-based POP repair decreased among non-Hispanic white women to 18.5% and among African American women to 14.2% (trend p<0.001), but remained unchanged among Hispanics (21.5% in 2015, p=0.18).

Figure 1.

Unadjusted trends in use of transvaginal mesh in pelvic organ prolapse surgeries over the years by race and ethnicity.

After adjusting for patient characteristics, there was a significant interaction between patients’ race/ethnicity and the trend in transvaginal mesh use in POP repair after Q3 2011 (p=0.02). The decrease in mesh use during the post-communication period was significant among non-Hispanic white women (slope p<0.001, Q3 2015 vs. Q3 2011 OR 0.45, 95% CI 0.41–0.49) and African American women (slope p<0.001, Q3 2015 vs. Q3 2011 OR 0.47, 95% CI 0.34–0.66) (Table 2). There was no significant change in mesh use in POP repair among Hispanic women following the 2011 FDA communication (slope p=0.22, Q3 2015 vs. Q3 2011 OR 0.87, 95% CI 0.69–1.09).

Table 2.

Piecewise logistic regression of trends in the use of transvaginal mesh in pelvic organ prolapse surgeries (quarterly analysis) by race/ethnicity.

	Unadjusted		Adjusted for patient characteristics
	OR (95% CI)	P value	OR (95% CI)	P value
2008 Q1 – 2011 Q3
White	1.75(1.61–1.90)	<.001	1.78(1.64–1.94)	<.001
African American	1.28(0.94–1.74)	.34	1.17(0.85–1.61)	.28
Hispanic	2.02(1.60–2.55)	<.001	1.89(1.49–2.41)	<.001
2011 Q3 – 2015 Q3
White	0.39(0.36–0.43)	<.001	0.45(0.41–0.50)	<.001
African American	0.40(0.29–0.55)	<.001	0.48(0.35–0.67)	<.001
Hispanic	0.81(0.65–1.01)	.29	0.88(0.70–1.11)	.22

Time was modeled as a continuous variable with a piecewise logistic regression. The OR was estimated based on the slope and interpreted as the odds of mesh use at the end of the specified time-period compared with the odds at the beginning for the racial group.

Subgroup analysis

In the subgroups stratified by patients’ residential SES, the trends in transvaginal mesh use in POP repair across different racial and ethnic groups were mostly consistent with that in the main analysis (Figure 2, Table 3). Regardless of the SES stratification, non-Hispanic white patients were the most likely to receive mesh-based POP repair at the beginning of the study. Mesh-based POP surgery among white and African American patients decreased after Q3 2011. Only among patients from high-income areas, there was a trend toward decreasing mesh use among Hispanic patients after Q3 2011 (slope p=0.08, Q3 2015 vs. Q3 2011 OR 0.71, 95% CI 0.49–1.04). In other subgroups, the use of mesh-based POP surgery among Hispanic women mostly remained unchanged from Q3 2011 to Q3 2015.

Figure 2.

Unadjusted trends in use of transvaginal mesh in pelvic organ prolapse surgeries over the years by race/ethnicity, stratified by zip code median income [(A) low and (B) high] and by zip code proportion of high school graduate [(C) low and (D) high].

Table 3.

Multivariable piecewise logistic regression of trends in the use of transvaginal mesh in pelvic organ prolapse surgeries (quarterly analysis) by race/ethnicity, stratified by zip code SES categories. (shown OR with 95% CI)

	Low income	High income	Low education attainment	High education attainment
2008 Q1 – 2011 Q3
White	1.80(1.57–2.07)	1.77(1.58–1.98)	1.82(1.59–2.09)	1.75(1.56–1.96)
African American	0.81(0.54–1.21)	2.11(1.23–3.60)	0.96(0.66–1.41)	1.89(1.03–3.46)
Hispanic	1.76(1.30–2.39)	2.05(1.36–3.08)	1.81(1.36–2.41)	2.01(1.27–3.17)
2011 Q3 – 2015 Q3
White	0.36(0.30–0.42)	0.52(0.46–0.59)	0.42(0.36–0.49)	0.48(0.43–0.55)
African American	0.62(0.41–0.95)	0.32(0.18–0.56)	0.51(0.34–0.76)	0.39(0.21–0.71)
Hispanic	0.98(0.73–1.32)	0.71(0.49–1.04)	0.91(0.69–1.20)	0.88(0.57–1.35)

Time was modeled as a continuous variable with a piecewise logistic regression. The OR was estimated based on the slope and interpreted as the odds of mesh use at the end of the specified time-period compared with the odds at the beginning for the racial group.

Discussion

We found that transvaginal mesh use in POP repair prior to 2011 was higher among non-Hispanic white women. Following the 2011 FDA safety communication, the use of mesh in POP repair decreased among non-Hispanic white women and African Americans but remained unchanged among Hispanic patients. In the subgroup analysis, there was a trend toward decreasing mesh use among Hispanic patients only among those from high-income areas.

Our findings of a differential decrease in mesh use after the FDA communication indicated that the dissemination of regulatory messages related to medical products might not have an equal impact across different social groups. When the FDA released warnings against antidepressants use in children and adolescents, it was found that the decrease in antidepressant use was different for white children and for minority children.¹³ In that study, the decrease in antidepressant use was less pronounced among children of minority races when compared with that among white children. These findings in the pharmaceutical context were consistent with our results in the medical device context. The difference in safety message diffusion could represent an important mechanism of rising disparities, as disadvantaged groups are less likely to be protected from the potential risks of the medical products.

We also found that among the subgroup of patients from high-income areas, there was a trend toward decreasing mesh use among Hispanic patients. In the aftermath of the FDA warning related to antidepressant use in children and adolescents, another research found that the decrease in antidepressant use after the warning was less obvious among children whose parents’ education level was lower.¹² These findings showed that not only is SES closely associated with new safety knowledge acquirement, it might also modify the impact of safety message diffusion in minority racial groups. As a result, minority patients from lower SES groups may be least likely to be protected from potential adverse events that these safety messages intended to address.

Our study demonstrates that when a new safety message is disseminated, there could be rising racial disparities due to unequal diffusion of the message. While transvaginal mesh for POP repair has been removed from the market, this evidence has implications for future scenarios. For a commonly performed procedure like POP repair, a 10% change in treatment choice in one year could mean that tens of thousands of patients are removed from exposure to the risk. These results highlighted the gap we need to cover when disseminating regulatory messages regarding medical device adverse events. Low SES minority patients were most vulnerable. High SES Hispanic patients tended to have decreased use of transvaginal mesh after the safety communication, but the decrease was not as pronounced as that among white women. These results suggest that particular attention is warranted for patients from disadvantaged groups when it comes to safety message dissemination, which could involve multiple aspects such as patients’ access to new safety information and cultural and language background.

There are limitations to consider in interpreting our study findings. First, we relied on administrative data for the current study. Administrative data do not contain clinical details such as the severity of prolapse, which cannot be adjusted for in this study. We attempted to address this by adjusting for the use of apical support. In addition, it could not be ascertained in administrative data regarding who drove the change in treatment decisions. Further qualitative research is needed to unravel this. Second, ICD-9 codes for mesh were not available prior to 2008. Thus we were not able to assess the trends in transvaginal mesh adoption in POP surgery for the earlier time period. Third, one of our variables of interest was residential SES, which may be different from individual SES in some cases, and should be interpreted accordingly. Results cannot be used to interpret the effect of individual-level education on the change of mesh use. Lastly, we only evaluated the State of New York. Proportionally, New York State has a bigger minority population than the entire country. The proportion of residents who are non-English speakers is also higher in New York State than in the entire country.¹⁷ The magnitude of disparities in the change of mesh use may thus vary across states as the language barrier patients encounter when seeking healthcare might be different.

Conclusions

Regulatory safety messages are intended to help raise awareness of the potential risks associated with medical products among providers and their patients. Our results demonstrated that the communications related to the safety of transvaginal mesh did not have an equal impact across racial groups. Transvaginal mesh use decreased among white and African American women but not among Hispanic women following the 2011 FDA safety communication.