Value of Ambulatory Modified Radical Mastectomy

Abstract

Background

Modified radical mastectomy (MRM) still is largely performed in inpatient settings. This study sought to determine the value (expenditures and complications) of ambulatory MRM.

Methods

Health Care Utilization Project (HCUP) state databases from 2016 were queried for patients who underwent MRM. The study examined rates of 30-day readmission for surgical-site infection (SSI) or hematoma, charges by index care setting, and predictors of 30-day readmission.

Results

Overall, 8090 patients underwent MRM: 5113 (63 %) inpatient and 2977 (37 %) ambulatory patients. Compared with the patients who underwent inpatient MRM, those who underwent ambulatory MRM were older (61 vs. 59 years), more often white (66 % vs. 57 %), in the lowest income quartile (28 % vs. 21 %), insured by Medicare (43 % vs. 33 %) and residents in a small metro area (6 % vs. 4 %) (all p < 0.01). Of the 5113 patients treated as inpatients, 126 (2.5 %) were readmitted, whereas 50 (1.7 %) of the ambulatory patients were readmitted (p = 0.02). The adjusted charge for inpatient MRM without readmission was $113,878 (range, $107,355–120,402) compared with $94,463 (range, $86,021–102,907) for ambulatory MRM, and the charge for inpatient MRM requiring readmission was $159,355 (range, $147,142–171,568) compared with $139,940 (range, $125,808–154,073) for ambulatory MRM (all p < 0.01). This difference remained significant after adjustment for hospital length of stay. Adjusted logistic regression showed that the ambulatory setting was protective for readmission (odds ratio, 0.49; 95 % confidence interval, 0.35–0.70; p < 0.01).

Conclusions

The analyses suggest that ambulatory MRM is both safe and less expensive. The findings advocate that MRM, a last holdout of inpatient care within breast surgical oncology, can be transitioned to the ambulatory setting for appropriate patients.

Contemporary surgical quality improvement efforts include minimizing the amount of time patients spend in the hospital after an operation. Surgical care occurs in three main locations: the traditional hospital inpatient setting, outpatient departments within hospitals, and free-standing ambulatory surgical centers. The latter two are considered “same-day surgery” locations, which provide surgical care to patients who are discharged home without inpatient stays, although they may also include patients observed for 23 or fewer hours.

The share of operations performed within ambulatory settings has been steadily increasing since the 1980s. In 1994, 54 % of all operations were performed in the outpatient or ambulatory setting, whereas in 2014 this proportion had risen to 66 %.¹ Enhanced recovery after surgery protocols for colorectal and urologic surgeries, among others, attempt to minimize the hospital length of stay after inpatient operations while improving post-surgical patient outcomes.^2,3 Unfortunately, the COVID-19 pandemic highlighted the value of inpatient hospital beds, and via the Centers for Medicare & Medicaid Services (CMS) section 1135 waiver, ambulatory surgery centers were given the ability to provide 23-h postoperative observation care, expanding the share of procedures they could offer.⁴

Partial mastectomy, with removal of a portion of the breast, is nearly universally performed in an ambulatory surgery setting. In contrast, approximately half of total mastectomy patients still are treated in the inpatient setting despite growing evidence of equivalent safety and improved patient satisfaction with ambulatory mastectomy.^5–9

In this era of optimizing health care utilization, and in light of continued inpatient capacity crises during the COVID-19 pandemic, a study analyzing the value of ambulatory mastectomy might provide the missing piece to facilitate its adoption. Our objective was to compare the values of mastectomy performed in the inpatient versus ambulatory setting by examining expenditures and complications.^10,11 We did this by examining the difference in total charge between inpatient and ambulatory mastectomies, including those requiring readmission and analyzing whether the operative setting is a predictor of 30-day readmission for patients undergoing mastectomy.

Methods

Data Source

We used three 2016 state-level datasets from three states (Florida, New York, and Maryland): the State Inpatient Database (SID), the State Ambulatory Surgery and Services Database (SASD), and the State Emergency Department Database (SEDD) of the Healthcare Cost and Utilization Project (HCUP) Agency for Healthcare Research and Quality.^12–14 These states were chosen based on availability of the HCUP visit link variable, encrypted patient identifiers facilitating linkage between health care encounters within the calendar year, and geographic and economic diversity.

The SID was queried for modified radical mastectomies (MRMs) performed on the basis of inpatient and inpatient re-admissions within 30-days after the index procedure, the SASD for modified radical mastectomies (MRMs) performed in the ambulatory setting, and finally, the SEDD for emergency room visits within 30-days after the index procedure regardless of procedure setting.

We chose to study MRM in an effort to ensure a homogeneous population (i.e., patients with cancer), whereas total or simple mastectomy could include cosmetic procedures. We did not include patients who underwent reconstruction. The SID and SASD databases were narrowed to patients undergoing MRM. Then, using the HCUP visit link for patients who underwent MRM, we queried the SEDD and SID for emergency department visits and hospital readmissions for SSI and postoperative hematoma within 30 days.

Definition of Variables

The study included patients with a breast cancer diagnosis as defined by the International Classification of Diseases (ICD) 10 diagnoses codes C50* and D05* who underwent MRM Current Procedural Terminology (CPT) procedure code 19307, with or without modifier 50; ICD10 0HTT, 0HTU, 0HTV) without reconstruction in 2016. We defined “ambulatory” operations as those performed in hospital-based outpatient surgical departments or free-standing ambulatory surgical centers, which may still include a short 23-h observation period. We defined “same-day” operations as those without any prolonged postoperative observational stay.

We limited complications of interest to SSI and postoperative hematomas (ICD10 codes T8130XA, T8131XA, T8131XD, T814XXA, T814XXD, T8189XA, T819XXA, and L7622) to avoid confounders, specifically medical conditions that may lead to readmission but would have influenced the surgeon’s choice of setting for the index operation in the first place (e.g., heart failure or an exacerbation of chronic obstructive pulmonary disease).

The payer variable within HCUP is the expected primary payer, including Medicare, Medicaid, private insurance, self-pay, no charge, or other. The zip income quartile provides a quartile classification of the estimated median household income of residents in the patient’s ZIP code. The exact value changes depending on the year. For 2016, the first quartile was $1.00 to $42,999, the second quartile was $43,000 to 53,999, the third quartile was $54,000 to 70,999, and the fourth quartile was $71,000+.

Statistical Analysis

The patient and hospital characteristics in this final cleaned and merged dataset were sorted by location of index operation (inpatient vs. ambulatory operation). Unadjusted differences in charge by index care setting with and without readmission were estimated using a linear regression model. Charge by index care setting with and without readmission was adjusted for patient age, rurality, payer, setting, and state, then calculated using a linear regression model. A fully adjusted model included hospital length of stay (LOS) together with all the variables in the adjusted model. We then used an unadjusted and adjusted multivariable (age, rurality, payer, LOS, setting, and state) logistic regression to identify predictors of 30-day readmission for patients who underwent MRM.

Statistical analysis was conducted using Stata version 14 (StataCorp LLC, College Station, TX). All t tests were two-tailed, and an alpha of 0.05 was considered significant. Oversight for this project was provided by the Mass General Brigham Institutional Review Board.

Results

As shown in Table 1 in 2016, 8090 patients underwent MRM in Florida, New York, and Maryland. Overall, 5113 (63 %) of the 8090 patients had their operation in an inpatient setting, and 2977 (37 %) patients had their operation in an ambulatory setting. The average LOS after inpatient MRM was 2.1 days versus 0.8 days for ambulatory treatment (p < 0.01). In the inpatient group, 53 patients (0.01 %) had a same-day discharge, whereas 904 patients (30.6 %) in the ambulatory group had a same-day discharge. Compared with the inpatient group, the group that underwent ambulatory MRM were older (61 vs. 59 years), more often had Medicare (43 % vs. 33 %), were white (66 % vs. 57 %), were in the first (lowest) income quartile (28 % vs. 21 %), and lived in a small metro area (6 % vs. 4 %) (all p < 0.01). Most of the ambulatory MRMs (71 %, 2114/2977) were performed in Florida, whereas 61 % (3144/5113) of all the inpatient MRMs were performed in New York (p < 0.01). Of the 5113 patients who underwent inpatient MRM, 126 had a readmission within 30 days (2.5 %), whereas 50 of the 2977 ambulatory MRMs resulted in a readmission (1.7 %) (p = 0.02).

Table 1.

Patient and facility characteristics according to index surgical care setting

Characteristic	Inpatient	Ambulatory	p Value
	(n = 5113) n (%)	(n = 2977) n (%)
Mean age: years (95 % CI)	59 (58–59)	61 (61–62)	< 0.01
Mean LOS: days (95 % CI)	2.1 (2.1–2.2)	0.8 (0.8–0.8)	< 0.01
Same-day discharge	53 (0.01)	904 (30.6)
Payer			< 0.01
Medicare	1662 (0.33)	1291 (0.43)
Medicaid	788 (0.15)	246 (0.08)
Private	2494 (0.49)	1323 (0.44)
Self-pay	64 (0.01)	28 (0.01)
No charge	17 (<0.01)	26 (0.01)
Other	87 (0.02)	63 (0.02)
Race			< 0.01
White	2927 (0.57)	1961 (0.66)
Black	905 (0.18)	440 (0.15)
Hispanic	584 (0.11)	399 (0.13)
Asian/Pacific Islander	294 (0.06)	71 (0.02)
Native American	≤ 11 (< 0.01)	≤ 11 (< 0.01)
Other	383 (0.07)	91 (0.03)
Zip code income quartile			< 0.01
1	1085 (0.21)	827 (0.28)
2	1070 (0.21)	841 (0.28)
3	1132 (0.22)	694 (0.23)
4	1780 (0.35)	580 (0.19)
Urban/rural stats			< 0.01
Large/medium metro area	4721 (0.92)	2645 (0.89)
Small metro area	188 (0.04)	192 (0.06)
Micropolitan	120 (0.02)	81 (0.03)
Rural	75 (0.01)	55 (0.02)
State			< 0.01
Florida	1464 (0.29)	2114 (0.71)
Maryland	505 (0.10)	288 (0.10)
New York	3144 (0.61)	575 (0.19)
Readmissions	126 (0.025)	50 (0.017)	0.02

CI, confidence interval; LOS, hospital length of stay

The unadjusted total charge for inpatient MRM without readmission was $56,077 (95 % confidence interval [CI], $54,881–57,274 [reference]) compared with $49,723 (95 % CI, $46,661–52,914; p < 0.01) for ambulatory MRM without readmission (Table 2). The unadjusted total charge for inpatient MRM requiring readmission was $101,467 (95 % CI, $44,414–109,127; p < 0.01) compared with $95,113 (95 % CI, $85,536–104,768; p < 0.01) for ambulatory MRM requiring readmission. After adjustment for age, rurality, payer, setting, and state, the charge for inpatient MRM without readmission was $113,878 (95 % CI, $107,355–$120,402 [reference]) compared with $94,463 (95 % CI, $86,021–102,907; p < 0.01) for ambulatory MRM without readmission. The charge for inpatient MRM requiring readmission was $159,355 (95 % CI, $147,142–171,568, p < 0.01) compared with $139,940 (95 % CI, $125,808–154,073; p < 0.01) for ambulatory MRM requiring readmission.

Table 2.

Total charge by index care setting and readmission status

Unadjusted charge by index care setting		p Value
Index care setting and readmission status	Mean total charge: U.S. dollars (95 % CI)
Inpatient not requiring readmission	56,077 (54881–57274)	Reference
Ambulatory not requiring readmission	49,723 (46611–52914)	< 0.01
Inpatient requiring readmission	101,467 (44414–109127)	< 0.01
Ambulatory requiring readmission	95,113 (85536–104768)	< 0.01
Adjusted charge by index care setting (adjusted for age, rurality, payer, setting, state)
Inpatient not requiring readmission	113,878 (107355–120402)	Reference
Ambulatory not requiring readmission	94,463 (86021–102907)	< 0.01
Inpatient requiring readmission	159,355 (147142–171568)	< 0.01
Ambulatory requiring readmission	139,940 (125808–154073)	< 0.01
Fully adjusted model (adjusted model plus length of stay)
Inpatient not requiring readmission	99,533 (93,533–105,534)	Reference
Ambulatory not requiring readmission	90,178 (84,236–96,120)	< 0.01
Inpatient requiring readmission	141,996 (134,139–149,852)	< 0.01
Ambulatory requiring readmission	132,640 (124,787–140,493)	< 0.01

In the fully adjusted model including LOS, the charge for inpatient MRM without readmission was $99,533 (95 % CI, $93,533–105,534 [reference]). The charge for ambulatory MRM without readmission was $90,178 (95 % CI, $84,236–96,120; p < 0.01). For MRM requiring readmission, the inpatient charge was $141,996 (95 % CI, $134,139–149,852; p < 0.01), and the ambulatory charge was $132,640 (95 % CI, $124,787–140,493; p < 0.01).

In an unadjusted logistic regression, the odds ratio (OR) of 30-day readmissions after an index operation in an ambulatory setting was 0.68 (95 % CI, 0.49–0.94; p = 0.02) compared with an inpatient index operation (Table 3). In the adjusted model, state was significantly predictive of 30-day readmission (Florida [reference], New York [OR, 0.45]; Maryland [OR, 1.84]; p < 0.01), as was ambulatory setting (OR, 0.49; 95 % CI, 0.35–0.70; p < 0.01). Payer, income quartile, and patient race were found to be co-linear, so payer was chosen as the variable to keep within the model because we believed it was the most accurate indicator of access to care. In contrast, payer was not a significant predictor of 30-day readmission (OR, 0.99; 95 % CI, 0.98–1.01; p = 0.39). Hospital LOS also was found to be a statistically significant predictor of readmission (OR, 1.04; 95 % CI, 1.01–1.08; p < 0.01).

Table 3.

Logistic regression analyses identifying predictors of 30-day readmission

	OR	95 % CI	p Value
Unadjusted
Index care setting
Inpatient	Reference
Ambulatory	0.68	0.49–0.94	0.02
Adjusted
Age (years)	0.99	0.98–1.01	0.47
Payer
Medicare	Reference
Medicaid	1.36	0.76–2.42	0.30
Private	0.94	0.59–1.49	0.79
Self-pay	0.71	0.09–5.32	0.74
No charge	3.29	0.94–11.48	0.06
Other	2.45	1.11–5.39	0.03
LOS	1.04	1.01–1.08	0.01
Urban/rural status: n (%)
Large metro area	Reference
Small metro area	0.95	0.46–1.96	0.89
Micropolitan	1.62	0.70–3.74	0.26
Rural	1.58	0.57–4.35	0.38
State: n (%)
Florida	Reference
New York	0.42	0.29–0.62	< 0.01
Maryland	1.81	1.21–2.70	< 0.01
Index care setting
Inpatient	Reference
Ambulatory	0.54	0.37– 0.77	< 0.01

Discussion

Using pooled data from three states representing a geographically, economically and racially diverse set of patients, our analysis showed that for appropriately selected patients MRM in the ambulatory setting is possible with low complication and readmission rates. Ambulatory MRM had significantly fewer readmissions for SSI and hematoma than inpatient MRM.

Ambulatory MRM also appeared to have lower total charges, both with and without readmission, than inpatient MRM of the same readmission status. In the adjusted analysis, ambulatory MRM not requiring readmission was significantly less expensive than inpatient MRM, by nearly $20,000 per episode of care. These differences remained significant after adjustment for LOS, but were attenuated, suggesting that extra hospital nights are driving some of the difference in expense of MRM in the inpatient setting.

In this study, the patients undergoing MRM as an ambulatory procedure were one third less likely to undergo 30-day readmission in the unadjusted model, and half as likely to experience a 30-day readmission in the adjusted model. These findings are consistent with prior study demonstrating that same-day simple mastectomy has lower rates of 30-day readmissions and reoperations than inpatient simple mastectomy.^15–17 Although these findings may be attributable to underlying patient differences, investigators at Kaiser Northern California found no increase in emergency department visits, reoperations, or readmission rates for similar patients at their institution after implementation of a home-based recovery program after mastectomy.⁸

Overall, this study saw very few readmissions for surgical complications after MRM, whether the surgery was performed in the ambulatory or inpatient setting, yet very few patients underwent same-day MRM in either group. Given the very low complication and readmission rates, there may be opportunity to increase the proportion of patients discharged immediately after MRM.

Our findings differ from those of other studies in the surgical safety literature, which show that patient-level factors are indeed a significant predictor of postoperative readmissions.^18,19 In the breast surgical oncology literature, black race and higher American Society of Anesthesiologists class are shown to be significant factors of readmission.⁵ In colorectal surgery, minority-serving hospitals have higher readmission rates than non-minority-serving hospitals, and patient-level factors account for 65 % of this difference.²⁰ These inconsistent findings may be explained by the fact that race and ethnicity are non-modifiable patient-level factors that are potential proxies for implicit bias and systemic racism and often confounded with other social risk factors.

Although our data did not reflect the impact of patient-level factors on 30-day readmission after MRM, we did find that patient-level factors were significantly associated with the site of care at which patients underwent their MRM. We saw that the majority of ambulatory MRMs were provided to high-income, white patients with Medicare or private insurance. This reflects study in the surgical quality and care access literature showing that non-white and lower-income patients are less likely to undergo surgical care in ambulatory and outpatient settings.^21–23 As we accrue increasing evidence that ambulatory surgical care is high-value care, race- or income-based differentials in access to ambulatory surgery represents an equity problem for provision of high-quality surgical care.

The pursuit of high-value surgical care frequently indicates the need to transition well-chosen patients from inpatient to ambulatory surgical care, even to a same-day, surgical care setting.²⁴ Studies show that ambulatory surgical care, once reserved for only the lowest-acuity procedures, are not only safe, but also a higher-value method of providing surgical care across an increasingly complex portfolio of surgical procedures across diverse surgical specialties.^25–27

For shoulder arthroplasty, the average charges for ambulatory surgical care were 40 % lower than for inpatient care.²⁸ Ambulatory surgery centers likely expend less due to leaner teams of providers, judicious use of consumables, and same-day discharge protocols. In other surgical disciplines, previous study also has found lower rates of surgical complications when procedures are performed in an ambulatory setting versus a hospital setting, demonstrating that with current methods of patient selection, surgeries in an ambulatory setting meet a similar standard of safety.^29–31 Mastectomy with prosthetic reconstruction also has been shown to have equivalent readmission rates for ambulatory versus hospital-based procedures, and recent evidence from the COVID-19 pandemic has demonstrated that mastectomy with reconstruction can be performed on an ambulatory/outpatient basis without compromising safety or the patient experience.^{7,15,16,18,32–34} Indeed, this ongoing experience demonstrates no difference in complications when mastectomy is performed in an ambulatory setting with or without prosthetic reconstruction.

The COVID-19 pandemic demonstrated that preserving inpatient bed capacity for truly necessary admissions also represents an important value consideration for the health care system as a whole. As hospital bed and staffing shortages continue even outside pandemic surges, uncoupling the ability to provide surgical care from inpatient bed capacity is an important step for ensuring reliable access to surgical care.

Breast surgical oncology has long been at the forefront of moving operative care to the ambulatory setting. Shown to be equally safe and less expensive by our study and others, the current challenge is more universal implementation of same-day discharge after mastectomy. To this end, the Patient Safety Quality Committee of the American Society of Breast Surgeons has formed a working group to investigate barriers to implementation of ambulatory mastectomy, and to help centers transition to same-day discharge after simple mastectomy.³⁵

This study had several limitations. Although we made every effort to include a strict population, studying MRM in an effort to enroll only cancer patients, including only SSI and hematoma-related readmissions and performing adjusted analyses, this was a retrospective review with inherent selection bias. State-level HCUP databases do not allow for adjustment by patient comorbidities, nor do they capture intangible confounders such as surgeon intuition about whether a patient will do well after a same-day discharge. Presumably, surgeons who operate in both inpatient and ambulatory settings decide whether to offer patients inpatient or ambulatory MRM based on their clinical judgment of a patient’s health and appropriateness for same-day discharge. Our analyses, in this context, suggested that if a surgeon assesses a patient and determines the patient is healthy enough to undergo a same-day operation, performance of an MRM in an ambulatory setting is a safe and appropriate choice.

Conclusions

Breast surgical oncology has been a forerunner in moving clinically appropriate care from the inpatient to the outpatient setting. Partial mastectomy has been thoroughly transitioned to same-day discharge, and simple mastectomy currently is the focus of high-value, high-quality implementation efforts. Our analyses showed that modified radical mastectomy—the most extensive operation regularly performed as a part of breast surgical oncology care—seems safe and less expensive when performed in the ambulatory setting. Although important differences likely exist between patients currently undergoing MRM in the inpatient setting and those undergoing MRM in the ambulatory setting, the rates of complications and readmissions in this study were low regardless of the operative setting. These findings suggest that MRM also should be transitioned to the ambulatory setting for clinically appropriate patients.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors

Disclosures

Anna Weiss reports an institutional research agreement with Myriad Laboratories Inc. at affiliation no. 1. Christina Minami reports research funding from the National Institute on Aging and the American College of Surgeons. David F. Friedlander reports research funding from an American Urological Association Research Scholars Award. Molly P. Jarman reports research funding from the National Institute on Aging and the Department of Defense. The remaining authors have no conflicts of interest.