Social determinants of health and Patient-Reported Outcomes Following Autologous Breast Reconstruction, Using Insurance as a Proxy

Abstract

Introduction:

Insurance type can serve as a surrogate marker for social determinants of health and can influence many aspects of the breast reconstruction experience. We aimed to examine the impact of insurance coverage on patients reported outcomes with the BREAST-Q in patients receiving deep inferior epigastric artery perforator (DIEP) flap breast reconstruction.

Methods:

We retrospectively examined patients who received DIEP flaps at our institution from 2010–2019. Patients were divided into categories by insurance: commercial, Medicaid, or Medicare. Demographic factors, surgical factors, and complication data were recorded. Descriptive statistics, Fisher’s exact, Kruskal-Wallis rank sum tests, and generalized estimating equations were performed to identify associations between insurance status and five domains of the BREAST-Q Reconstructive module.

Results:

1,285 patients were included, of which 1,011 (78.7%) had commercial, 89 (6.9%) had Medicaid, and 185 (14.4%) had Medicare insurances. Total flap loss rates were significant higher in the Medicare and Medicaid patients as compared to commercial patients; however, commercial patients had a higher rate of wound dehiscence as compared to Medicare patients. With all other factors controlled for, patients with Medicare had lower Physical Well-Being of the Chest (PWBC) than patients with commercial insurance ($\beta$ =−3.1, 95% CI: −5.0, −1.2, p=0.002). There were no significant associations between insurance classification and other domains of the BREAST-Q.

Conclusion:

Patients with government issued insurance had lower success rates of autologous breast reconstruction. Further, Patients with Medicare had lower PWBC than patients with commercial insurance regardless of other factors, while other BREAST-Q metrics did not differ. Further investigation as to the causes of such variation is warranted in larger, more diverse cohorts.

INTRODUCTION

It has become increasingly evident that patient reported outcomes (PROs) are not solely shaped by clinical procedures and medical interventions. Rather, a complex interplay of social determinants of health (SDOH) play a pivotal role in influencing patients’ overall experience and effectiveness of treatment.^1–3 SDOH encompass the environmental factors in peoples’ lives that affect health, functioning, and quality of life.⁴ SDOH can be broken down into the following domains: economic stability, access to education, access to healthcare, neighborhood environment, and social community.⁴ Since SDOH can be difficult to study and quantify given multifactorial inputs, insurance status can serve as a valuable surrogate marker for multiple SDOH, including socioeconomic status (SES), race, education level, and comorbidities.⁵ As examples, a majority of Medicaid beneficiaries are non-White, whereas individuals with more education are more likely to have better insurance and health benefits.^6–9

The Women’s Health and Cancer Rights Act in 1998 mandated that all insurances cover breast reconstruction in patients undergoing mastectomy.¹⁰ Research has since attempted to elucidate whether insurance status continues to impact the reconstructive course.^11–20 Many of these studies have assessed the influence of insurance status on the modality of breast reconstruction as well as clinical outcomes, but little is known about how insurance status — and by extension, SDOH — can affect PROs. Assessing the impact of SDOH on PROs after breast reconstruction is critical to ensure that all patients are satisfied and have optimal quality of life.²¹ BREAST-Q has emerged as one of the most validated tools for examining PROs after breast surgery.^22–26 One study examined the association between socioeconomic status and BREAST-Q scores after breast reconstruction, but this study used Zillow to estimate patients’ incomes rather than having patients report their incomes, which may contribute to inaccuracies.²⁷ Therefore, a more comprehensive examination of SDOH and PROs after breast reconstruction is necessary by using insurance status as a proxy.

This study aims to delve into the critical relationship between PROs in autologous breast reconstruction and SDOH, using insurance type as a pragmatic lens to gain a deeper understanding of these intricate dynamics. Autologous breast reconstruction has been shown to confer the strongest increases in BREAST-Q scores following surgery,^25,28 so our study focused on patients that underwent deep inferior epigastric perforator (DIEP) flaps for breast reconstruction. We hypothesized that, relative to patients with commercial insurance, patients with publicly funded insurance would report worse long-term BREAST-Q scores.

METHODS

This was a retrospective study examining all patients receiving unilateral or bilateral DIEP flaps at Memorial Sloan Kettering Cancer Center between January 2010 and December 2019. Institutional Review Board approval was obtained for this study. Patients were included if they were 18 years of age or older. Exclusion criteria included other types of autologous breast reconstruction and implant-based reconstruction.

Demographic characteristics of interest included age, race, ethnicity, marital status, body mass index (BMI), smoking status, median household income, history of psychiatric diagnosis, and receipt of either chemotherapy or radiation. Mean household income was determined based on patients’ zip-code. Insurance payers were categorized into four categories: commercial, Medicare, and Medicaid. Medicaid is generally for patients with limited income and resources while Medicare is for patients 65 and older and/or with certain disabilities and conditions.²⁹ Surgical characteristics included donor site laterality, contralateral or bilateral prophylactic mastectomy, concurrent axillary lymph node dissection, and postoperative length of stay.

We examined postoperative complications to see if these could influence patient-reported outcomes. Complications of interest included abdominal wall hernia, infection/cellulitis, delayed healing, mastectomy skin flap necrosis, total flap loss, seroma, wound dehiscence, hematoma, and systemic complications (deep venous thrombosis, pulmonary embolism, and pneumonia). We also examined presence of any complications as well as readmissions or reoperations related to the index procedure within 90 days of reconstruction.

BREAST-Q scores were examined preoperatively and at 2 weeks, 3 months, 6 months, 1 year, and 2 years postoperatively for Physical Well-being of the Chest (PWBC) and Physical Well-being of the Abdomen (PWBA) domains. The 2-week postoperative survey was not administered for the other domains. Thus, for Satisfaction with Breasts, Psychosocial Well-being, and Sexual Well-being, scores were examined preoperatively and at 3 months, 6 months, 1 year, and 2 years postoperatively. A minimal clinically important difference (MCID) of 4 points was considered clinically important.³⁰

Patient demographics and outcomes were compared between each of the insurance categories overall using Fisher’s exact and Kruskal-Wallis rank sum tests. Pairwise comparisons of BREAST-Q scores between individual insurance categories were conducted using Wilcoxon rank-sum tests. The generalized estimating equation method was used to identify confounders of differences in patient-reported outcomes across insurance groups. Multiple imputation was used to account for missing BREAST-Q scores, median household income, and postoperative length of stay. All analyses were conducted in R (version). Statistical significance was set at an alpha of 0.05.

Results

1,285 patients were included in the database, of which 1,011 (78.7%) had commercial insurance, 89 (6.9%) had Medicaid, and 185 (14.4%) had Medicare (Table 1). Overall, the median age of the patients was 50 (Interquartile Range: 44, 57) and majority of the patients were White (69%) and not Hispanic (86%). There were significant differences in age at surgery, race, ethnicity, smoking status, smoking status, marital status, and median household income between the three cohorts. Proportion of donor site laterality, prophylactic mastectomy, and axillary lymph node dissection were significantly different among the insurance types (Tables 1).

Table 1.

Patient Demographics

Patient Demographics	Overall (n=1,285)	Commercial (n=1,011)	Medicaid (n=89)	Medicare (n=185)	p-value2
Age at Surgery1	50 (44, 57)	49 (44, 55)	46 (41, 54)	63 957, 66)	<0.001
Race					<0.001
White	886 (69%)	708 (70%)	44 (49%)	134 (72%)
Asian	111 (8.6%)	95 (9.4%)	7 (7.9%)	9 (4.9%)
Black	164 (13%)	121 (12%)	17 (19%)	26 (14%)
Other	124 (9.6%)	87 (8.6%)	21 (24%)	16 (8.6%)
Ethnicity					0.001
Not Hispanic	1,106 (86%)	884 (87%)	63 (71%)	159 (86%)
Hispanic	132 (10%)	92 (9.1%)	21 (24%)	19 (10%)
Unknown	47 (3.7%)	35 (3.5%)	5 (5.6%)	7 (3.8%)
Marital Status					<0.001
Married	941 (73%)	773 (76%)	43 (48%)	125 (68%)
Separated	121 (9.4%)	78 (7.7%)	16 (18%)	27 (15%)
Single	223 (17%)	160 (16%)	30 (34%)	33 (18%)
Body Mass Index (BMI)1	26.2 (23.4, 29.7)	26.2 (23.4, 29.7)	26.5 (23.9, 29.8)	26.9 (23.3, 29.8)	0.8
Smoking Status					0.024
Never	948 (74%)	757 (75%)	68 (76%)	123 (66%)
Current	32 (2.5%)	23 (2.3%)	5 (5.6%)	4 (2.2%)
Former	305 (24%)	231 (23%)	16 (18%)	58 (31%)
Median household income	96,606 (73,514, 127,680)	102,763 (77,044, 130,121)	66,332 (56,870, 99,572)	87,397 (69,353, 1223,492)	<0.001
History of psychiatric comorbidities	718 (56%)	547 (54%)	58 (65%)	113 (61%)	0.04
Chemotherapy
Adjuvant/Neoadjuvant	998 (78%)	777 (77%)	78 (88%)	143 (77%)	0.064
Radiation
Adjuvant/Neoadjuvant	87 (6.8%)	67 (6.6%)	8 (9.0%)	12 (6.5%)	0.7
Donor Site Laterality					0.007
Bilateral	628 (49%)	517 (51%)	35 (39%)	76 (41%)
Unilateral	657 (51%)	494 (49%)	54 (61%)	109 (59%)
Prophylactic					0.012
None	855 (67%)	650 (64%)	69 (78%)	136 (74%)
Contralateral prophylactic	391 (30%)	326 (32%)	20 (22%)	45 (24%)
Bilateral prophylactic	39 (3.0%)	35 (3.5%)	0 (0%)	4 (2.2%)
Axillary lymph node biopsy	763 (59%)	611 (60%)	63 (71%)	89 (48%)	<0.001
Postoperative length of stay (LOS) in days1	3 (3,5)	3 (3,5)	4 (3,6)	4 (3,6)	0.2

¹Median (IQR); n(%),

²Kruskal-Wallis rank sum test; Fischer’s Exact Test for Count Data with simulated p-value

Complications

There were significant differences in the proportion of patients who developed total flap loss and wound dehiscence. Total flap loss occurred most frequently in patients with Medicaid (4.5%) compared to 0.8% in commercially insured patients and 1.6% in Medicare patients (p=0.014); wound dehiscence occurred in 5.2% of commercial, 4.5% of Medicaid, and 1.1% of Medicare patients (p=0.026). No significant differences in insurance type were observed for other complications or the presence of any complication (Table 2). Given the low complication events for these three complications, especially in the Medicare cohort, it was determined that it would be difficult to reliably estimate a beta coefficient through logistic regression models.

Table 2.

Complications

All Patients	Overall (n=1,285)1	Commercial (n=1,011)1	Medicaid (n=89)1	Medicare (n=185)1	p-value2
Abdominal wall - hernia	8 (0.6%)	6 (0.6%)	0 (0%)	2 (1.1%)	0.6
Infection/cellulitis	174 (14%)	128 (13%)	16 (18%)	30 (16%)	0.2
Delayed healing	63 (4.9%)	49 (4.8%)	2 (2.2%)	12 (6.5%)	0.3
Mastectomy Skin Flap/Nipple Necrosis	175 (14%)	134 (13%)	9 (10%)	32 (17%)	0.2
Total Flap Loss	15 (1.2%)	8 (0.8%)	4 (4.5%)	3 (1.6%)	0.014
Seroma	58 (4.5%)	52 (5.1%)	3 (3.4%)	3 (1.6%)	0.085
Wound Dehiscence	59 (4.6%)	53 (5.2%)	4 (4.5%)	2 (1.1%)	0.026
Hematoma	94 (7.3%)	70 (6.9%)	5 (5.6%)	19 (10%)	0.2
Systemic Comorbidities (DVT, PE, pneumonia)	12 (0.9%)	8 (0.8%)	1 (1.1%)	3 (1.6%)	0.4
Any Complication	655 (51%)	509 (50%)	47 (53%)	99 (54%)	0.7
Readmission within 90 days	138 (11%)	100 (9.9%)	14 (16%)	24 (13%)	0.13
Reoperation within 90 days	558 (43%)	440 (44%)	45 (51%)	73 (39%)	0.2

¹n (%),

²Kruskal-Wallis rank sum test; Fisher’s exact test; Pearson’s Chi-squared test

BREAST-Q: PWBC Analysis

BREAST-Q scores across the insurance cohorts at different timepoints are listed in Table 3 along with the number of patient scores available. Differences in PWBC scores were observed across the three insurance cohorts at 6-months, 1-year, and 2-year postoperatively (Table 3). Pairwise comparison showed that commercially insured patients had statistically higher median preoperative PWBC scores than Medicare patients at 6-months and 1-year postoperatively (Table 4). At 6-months, commercially insured patients had a median score of 72 (IQR: 60, 85) while Medicare patients had a score of 69 (56, 81) (p=0.016) (Figure 1). The difference was clinically meaningful at 1-year: commercial: 76 (64, 92) versus Medicare: 68 (56, 80) (p=0.005) (Figure 2). At 2-years, commercially insured patients scored 12 points higher than Medicaid patients (76 [64, 92] versus 64 [55, 83], p=0.031) (Figure 3).

Table 3.

BREAST-Q Scores

All Patients	N	Commercial (n=1,011)1	N	Medicaid (n=89)1	N	Medicare (n=185)1	p-value2
Physical Well-being of the Chest
Preoperative	512	76 (60, 91)	28	73 (60, 91)	89	71 (55, 85)	0.059
2-weeks	578	60 (50, 71)	46	60 (41, 72)	94	64 (45, 76)	0.5
3-months	487	68 (59, 80)	27	68 (58, 83)	81	64 (50, 77)	0.3
6-months	575	72 (60, 85)	48	68 (55, 80)	110	69 (56, 81)	0.016
1-year	644	76 (64, 92)	44	70 (57, 85)	122	68 (56, 80)	0.005
2-year	413	76 (64, 92)	27	64 (55, 83)	80	76 (55, 85)	0.031
Physical Well-being of the Abdomen
Preoperative	482	83 (69, 100)	26	100 (71, 100)	83	76 (63, 100)	0.8
2-weeks	572	49 (39, 58)	44	47 (39, 58)	92	52 (43, 60)	0.2
3-months	482	61 (52, 72)	26	61 (47, 71)	78	66 (51, 76)	0.7
6-months	567	66 (56, 81)	46	64 (51, 75)	104	66 (54, 81)	0.4
1-year	629	69 (58, 81)	42	66 (56, 80)	120	66 (56, 88)	0.8
2-year	409	69 (58, 88)	26	76 (58, 88)	77	69 (58, 81)	0.7
Satisfaction with Breasts
Preoperative	501	48 (33, 58)	26	41 (30, 52)	88	44 (34, 58)	0.7
3-months	464	59 (52, 69)	25	61 (55, 67)	76	58 (46, 71)	0.8
6-months	568	62 (53, 75)	48	59 (52, 71)	102	58 (48, 71)	0.08
1-year	634	64 (53, 78)	45	58 (49, 75)	119	61 (48, 75)	0.09
2-year	412	65 (53, 78)	26	64 (53, 75)	79	65 (52, 75)	0.5
Psychosocial Well-being
Preoperative	500	60 (47, 74)	27	62 (43, 75)	88	60 (48, 75)	0.9
3-months	466	65 (52, 83)	25	69 (52, 79)	78	64 (51, 82)	>0.9
6-months	564	66 (53, 83)	47	64 (50, 85)	106	61 (52, 79)	0.2
1-year	635	69 (56, 87)	43	60 (52, 70)	119	65 (53, 83)	0.1
2-year	410	69 (56, 87)	27	64 (52, 94)	80	68 (58, 86)	0.7
Sexual Well-being
Preoperative	478	43 (29, 59)	24	41 (24, 55)	80	43 (24, 62)	0.8
3-months	445	48 (34, 59)	22	53 (30, 64)	68	46 (29, 59)	0.9
6-months	544	48 (36, 63)	42	48 (25, 59)	96	43 (31, 59)	0.2
1-year	612	52 (39, 66)	40	42 (24, 57)	106	48 (32, 63)	0.049
2-year	400	53 (39, 66)	24	43 (32, 61)	68	48 (36, 66)	0.8

Table 4.

Pairwise Comparisons, BREAST-Q Median Scores

All Patients	Commercial versus Medicaid (p-value)	Commercial versus Medicare (p-value)	Medicaid versus Medicare (p-value)
Physical Well-being of the Chest
6-months	72 vs 68 (0.06)	72 vs 69 (0.039)	68 vs 69 (0.7)
1-year	76 vs 70 (0.064)	76 vs 68 (0.003)	70 vs 68 (>0.9)
2-year	76 vs 64 (0.032)	76 vs 76 (0.10)	64 vs 76 (0.4)
Satisfaction with Breasts
1-year	64 vs 58 (0.059)	64 vs 61 (0.040)	58 vs 61 (0.7)
Sexual Well-being
6-months	48 vs 48 (0.3)	48 vs 43 (0.019)	48 vs 43 (0.7)
1-year	52 vs 42 (0.007)	52 vs 48 (0.045)	42 vs 48 (0.3)

Figure 1.

Box Plots of Physical Well-being of the Chest at 6-Months by Insurance Type

Figure 2.

Box Plots of Physical Well-being of the Chest at 1-Year by Insurance Type

Figure 3.

Box Plots of Physical Well-being of the Chest at 2-Year by Insurance Type

Generalized estimation equation modeling showed that when controlling for all other factors, postoperative PWBC scores at 2-weeks and 3-months were significantly lower than the preoperative score. With all other factors controlled for, including the time of BREAST-Q administration, patients with Medicare have significantly lower PWBC than patients with commercial insurance ($\beta$ =−3.1, 95% CI: −5.0, −1.2, p=0.002). Patients who identified as Asian, Black, or Other had significantly lower PWBC scores than White patients, as did patients with a history of psychiatric diagnosis and longer postoperative stay (Table 5).

Table 5.

GEE Models for BREAST-Q Scores Over Time

	Physical Well-being of the Chest		Physical Well-being of the Abdomen		Satisfaction with Breasts		Psychosocial Well-being		Sexual Well-being
Patient Demographics	Beta (95% CI)	p-value2	Beta (95% CI)	p-value2	Beta (95% CI)	p-value2	Beta (95% CI)	p-value2	Beta (95% CI)	p-value2
Time
Preop	—		—		—		—		—
2-weeks	−14 (−16, −12)	<0.001	−26 (−28, −23)	<0.001	X		X		X
3-months	−4.8 (−6.9, −2.7)	<0.001	−12 (−14, −9.2)	<0.001	15 (13, 18)	<0.001	9.4 (6.7, 12)	<0.001	4.3 (1.4, 7.1)	0.004
6-months	−1.2 (−3.3, 0.85)	0.2	−7.7 (−10, −5)	<0.001	18 (16, 20)	<0.001	10 (7.8, 12)	<0.001	6.2 (3.5, 8.9)	<0.001
1-year	1.0 (−1.0, 3.1)	0.3	−5.3 (−7.6, −2.9)	<0.001	19 (17, 22)	<0.001	12 (9.9, 14)	<0.001	9.0 (6.3, 12)	<0.001
2-year	2.1 (−0.69, 4.9)	0.14	−5.3 (−8.0, −2.6)	<0.001	20 (18, 23)	<0.001	12 (9.7, 15)	<0.001	11 (7.7, 14)	<0.001
Insurance
Commercial	—		—		—		—		—
Medicaid	−1.8 (−4.8, 1.1)	0.2	−0.31 (−3.4, 2.7)	0.8	−0.34 (−3.7, 3.1)	0.8	(−2.2, 4.4)	0.5	−1.5 (−5.4, 2.5)	0.5
Medicare	−3.1 (−5.0, −1.2)	0.002	0.40 (−1.4, 2.2)	0.7	−1.1 (−3.3, 1.1)	0.3	−0.62 (−2.7, 1.5)	0.6	−1.3 (−4.0, 1.3)	0.3
Race
White	—		—		—		—		—
Asian	−3.9 (−6.4, −1.5)	0.002	−2.7 (−5.7, 0.37)	0.084	−0.79 (−3.5, 1.9)	0.6	−4.2 (−7.1, −1.2)	0.007	−2.6 (−5.8, 0.062)	0.11
Black	−3.2 (−5.3, −1.1)	0.004	−3.8 (−5.9, −1.7)	<0.001	−1.3 (−3.8, 1.1)	0.3	−1.6 (−3.9, 0.75)	0.2	0.26 (−2.7, 3.2)	0.9
Other	−4.5 (−6.8, −2.3)	<0.001	−2.0 (−4.7, 0.71)	0.15	−4.7 (−7.1, −2.2)	<0.001	−6.0 (−9.0, −3.0)	<0.001	−4.9 (−8.1, −1.6)	0.004
History of psychiatric comorbidities	−4.1 (−5.5, −2.6)	<0.001	−6.0 (−7.4, −4.5)	<0.001	−5.8 (−7.2, −4.3)	<0.001	−10 (−12, −8.4)	<0.001	−10 (−12, −8.6)	<0.001
Median household income	0.000004 (−0.000016, 0.000024)	0.7	0.000008 (−0.00001, 0.00003)	0.4	0.00002 (−0.000003, 0.00004)	0.093	0.00003 (0.000008, 0.00005)	0.008	0.000029 (0.000003, 0.00006)	0.031
Chemotherapy	−0.45 (−2.0, 1.1)	0.6	4.3 (2.3, 6.2)	<0.001	1.4 (−0.23, 3.1)	0.091	2.8 (0.77, 4.7)	0.007	1.7 (−0.53, 3.9)	0.13
Radiation	−1.7 (−4.3, 0.90)	0.2	2.0 (−0.92, 4.8)	0.2	1.0 (−1.9, 3.9)	0.5	3.1 (0.28, 5.9)	0.031	3.0 (−0.73, 6.8)	0.11
Donor site laterality
Bilateral	—		—		—		—		—
Unilateral	0.79 (−0.89, 2.5)	0.4	2.6 (0.63, 4.5)	0.010	−1.1 (−3.0, 0.78)	0.3	1.0 (−0.96, 3.0)	0.3	0.79 (−1.4, 3.0)	0.5
Prophylactic
Bilateral	—		—		—		—		—
Not prophylactic	−1.4 (−4.8, 2.1)	0.4	−1.9 (−6.1, 2.2)	0.4	−1.3 (−5.2, 2.6)	0.5	−0.14 (−4.3, 4.1)	>0.9	−0.46 (−4.6, 3.7)	0.8
Contralateral	−0.67 (−3.9, 2.6)	0.7	−1.1 (−4.7, 2.6)	0.6	0.22 (−3.5, 3.9)	>0.9	0.97 (−3.0, 4.9)	0.6	0.82 (−3.1, 4.7)	0.7
Postoperative length of stay	−0.0061 (−0.0096, −0.0025)	0.001	−0.0074 (−0.012, −0.003)	<0.001	−0.0037 (−0.008, 0.0007)	0.10	−0.0041 (−0.008, 0.0002)	0.064	−0.0054 (−0.01, −0.0005)	0.032
Complications	−0.61 (−1.9, 0.69)	0.4	−0.05 (−1.7, 1.6)	>0.9	−1.6 (−3.1, −0.10)	0.037	−0.98 (−2.4, 0.42)	0.2	−2.0 (−4.2, 0.24)	0.079

BREAST-Q: PWBA Analysis

We observed no differences in physical wellbeing of the abdomen at any point preoperatively or postoperatively when comparing all three insurance cohorts at once.

Generalized estimation equation modeling showed that when all other factors were controlled for, all postoperative PWBA scores over time were lower than the preoperative score, although it improved over time (p<0.001). It also showed that insurance type was not an independent predictor of PWBA. Patients who identified as Black (versus White), have a history of psychiatric diagnosis, and longer postoperative stay had significantly lower PWBA scores while those with a receipt of chemotherapy and unilateral donor site (vs bilateral) had higher PWBA scores (Table 5).

Satisfaction with Breasts

Although not significant, notable differences were seen in Satisfaction with Breasts at 6-months and 1-year across the three insurance types (p=0.08, p=0.09, respectively). Pairwise comparison demonstrated that commercially insured patients had significantly higher Satisfaction with Breasts than Medicare patients at 1-year (64 [53, 78] versus 61 [46, 75], p=0.04). This, however, was not clinically significant.

Modeling for Satisfaction with Breasts showed that all postoperative timepoints were associated with higher Satisfaction with Breasts scores relative to preoperative, holding all else constant. No association was observed between Satisfaction with Breasts and insurance type. Patients who identified as ‘Other’ (versus White), have a history of psychiatric diagnosis, and experienced complications had significantly lower Satisfaction with Breasts scores.

BREAST-Q: Psychosocial Well-Being Analysis

No significant differences were seen for Psychosocial Well-being at any timepoint.

GEE models for Psychosocial Well-being showed no significant differences by insurance type. Patients who identified as Asian or “Other” (versus White) and have a history of psychiatric diagnosis had significantly lower Psychosocial Well-being scores. Patients with a receipt of chemotherapy, radiation, or have higher median household income significantly higher Psychosocial Well-being.

BREAST-Q: Sexual Well-Being Analysis

Sexual Well-being was significantly different among the insurance types at 1-year (p=0.049). Pairwise analysis indicated that there were significant differences between commercial (48 [36, 63]) and Medicare (43 [31, 59]) patients at 6-months (p=0.019), meeting the MCID of 4. The difference was also observed at 1-year, in which commercially insured patients scored 4-points higher on Sexual Well-being than Medicare patients (52 [39, 66] versus 48 [32, 63], p=0.045). Patients with commercial insurance also scored 10 points higher than patients with Medicaid at this timepoint (52 [39, 66] versus 42 [24, 57], p=0.007).

Generalized estimation equation modeling for Sexual Well-Being showed that, holding all other factors constant, all postoperative scores were significantly higher than preoperative scores (p<0.001). Insurance status was not an independent predictor of Sexual Well-Being scores. Patients who identified as “Other” (versus White), have a history of psychiatric diagnosis, and stayed in the hospital for longer had significantly lower Sexual Well-being. Median household income was positively correlated with Sexual Well-being (Table 5).

Discussion

Insurance status can be used as a proxy for many social determinants of health, including socioeconomic status (SES), race, education level, and comorbidities.⁵ In this study, we examined how insurance type correlates with complication rates and PROs following DIEP breast reconstruction. Total flap loss and wound dehiscence were significantly different across the insurance types. We also found that insurance type most profoundly influences PWBC, with Medicare patients having significantly lower PWBC scores than commercially insured patients.

When examining all surgical specialties, including general surgery, orthopedic, gynecology, and plastic surgery, Medicare and Medicaid patients experience higher rates of complications than commercial payers.³¹ In implant- and autologous-based breast reconstruction, patients under Medicaid and Medicare are more likely to develop infections, cellulitis, and tissue necrosis than those who are commercially insured.³² In our focused examination of DIEP patients, we found that the overall presence of complications did not differ by payer type, but we did find that total flap loss and wound dehiscence were significantly different. Notably, Medicaid patients had the highest and commercial patients had the lowest rates of flap loss. This may be related to how Medicaid patients in our study have higher proportion of current smokers, relative to commercial and Medicare patients. Smoking is a well-known independent risk factor for complications in ABR as it can cause vasoconstriction and reduced blood flow.^33–35 We also found that commercial patients had a significantly higher rate of wound dehiscence as compared to Medicare patients. It’s important to highlight that due to the minimal occurrence of complication events, particularly within the Medicare and Medicaid cohorts, interpreting these findings should be approached with caution. Additional studies with greater sample sizes are necessary to control for such confounding variables and to assess whether insurance type truly impacts complication rates after DIEP flap reconstruction.

The Women’s Health and Cancer Rights Act in 1998 mandated that all breast reconstruction after mastectomy is covered, yet health disparities still exist in accessing reconstructive services. Medicare patients are least likely to undergo breast reconstruction compared to Medicaid and commercial patients and publicly insured patients are less likely to undergo ABR than commercially insured patients.^16,36 Our research extends the current body of literature by demonstrating that Medicare patients have significantly lower PWBC than commercially insured patients. The pairwise tests showed that Medicaid patients also have lower PWBC than commercially insured patients at 2-years. The differences may be for several reasons. PWBC scale may be reflective of general, overall health as it asks patients about symptoms including breast pain, breast tightness, and difficulty moving arms.³⁷ Given that Medicare services patients aged 65 or older or those with certain comorbidities,³⁸ it is logical that these patients would have a lower PWBC. We also found that patients with Medicare have higher rates of psychiatric diagnosis, which has been previously associated with lower PWBC scores.²⁶ Furthermore, although not significant, patients with Medicaid experienced higher rates of radiotherapy compared to patients with commercial insurance. Radiation has been associated with increased chronic postsurgical pain and chest/upper body discomfort in patients undergoing breast reconstruction by causing skin retraction and tightness.^39–41 Lastly, physical therapy is often advised for patients undergoing breast reconstruction to alleviate postsurgical pain.⁴² However, patients with public insurances may be less inclined to attend physical therapy and other rehabilitative programs, and may not realize the health benefits associated with physical therapy.⁴³ Providers should recognize the potentially lower PWBC in publicly insured patients and provide necessary interventions to optimize their postoperative function and mobility.⁴⁴

While we no longer observed statistical significance in other BREAST-Q domains by insurance type, race and income were significantly associated with these domains. Patients who identified as Asian, Black, or Other had lower BREAST-Q scores than White patients. In the existing literature, Hanson et al found that American Indian/Alaska Native patients have lower BREAST-Q scores while Oskar et al did not observe any postoperative BREAST-Q score differences across racial groups; however, sample sizes in these two studies were smaller.^45,46 Race is intricately tied with insurance coverage, but the expansion of Medicaid has helped to reduce racial disparities in accessing health care.^8,47 Our results indicate that BREAST-Q domains, except for PWBC, may be driven by inherent, racial and cultural differences rather than insurance coverage. Moreover, we found that patients with higher income had improved Psychosocial and Sexual Well-being. Another study also observed that higher SES was tied to better Satisfaction with Breasts, Psychosocial Well-being, and Sexual Well-being, implying that insurance type and income may not always be correlated.^27,45 Patients with lower SES, for instance, may have commercial insurance because of employer-sponsored plans while patients with higher SES can still qualify for Medicaid due to preexisting disabilties.^48,49 As such, while SES and race are critical SDOH to take into account, our research underscores the importance of evaluating insurance status independently as distinct SDOH factors may exert varying influences on outcomes. Further work is warranted to address existing SDOH in breast reconstruction and to minimize the impact of these non-medical factors that influence health outcomes.

The strengths of this study include a long follow-up time, up to 2 years postoperatively, and a rigorous statistical analysis to control for outside influences that may confound the effect of insurance coverage on BREAST-Q outcomes. Nonetheless, there are several limitations. First, this study is retrospective, which inherently limits the data we can collect and variables for analysis. It is feasible that there are other unknown variables which, if included in our modeling, could impact our results, but we are unable to control for all possible confounding factors. For example, Dinis et al demonstrated that publicly insured patients are less likely to undergo perforator flaps than commercially insured patients, indicating that selection bias may play a role in our findings.¹³ Another limitation is the limited cohort size in the publicly insured cohorts. For example, we were not able to conduct logistic regression models to highlight the actual impact of insurance status on complications given the low complication events across the insurance types. Given that not all patients completed BREAST-Q at every timepoint, we utilized multiple imputation to account for missing scores; still, the limited number of available scores in the Medicare and Medicaid cohorts is a limitation to our study. There may be additional significant differences that we were not able to observe due to the limited sample size. Lastly, our patient cohort was largely White, not Hispanic, and had a median income that is higher than the majority of the United States. This may contribute to the lower number of patients in the publicly insured cohorts and our findings may not be generalizable to other patient populations; they must be interpreted with caution. Nonetheless, as a cancer center that performs countless autologous breast reconstruction, we provide robust results on the impact of insurance on patient-reported outcomes. Future studies should continue to explore the relationship between insurance type and PROs in breast reconstruction in different patient populations as well as the factors that modify the effect of insurance status on patient-reported outcomes.

Conclusion

In this study, we used insurance status as a proxy for social determinants of health in breast reconstruction. Total flap loss and wound dehiscence were significantly different across the insurance types. Medicare patients have significantly lower Physical Well-being of the Chest than commercial payers regardless of other factors, while Physical Well-being of the Abdomen, Sexual Well-Being, Psychosocial Well-Being, and Satisfaction with Breasts do not display a correlation with insurance status. Future work should be done to further delineate how to reduce the discrepancies that may draw from differential insurance coverage.