Improving accessibility to breast reconstruction for women undergoing surgical treatment of breast cancer has been a grassroots effort for more than 20 years. This effort, in part, was codified into national health care policy with the Women's Health and Cancer Rights Act of 19981 which helped establish insurance coverage for reconstructive services2 and encourage discussion of available options and preference.3 Building on this, Breast Reconstruction Awareness days across the country and the Breast Cancer Patient Education Act4 passed through Congress now seek to ensure that women and health-care providers are informed about a woman's reconstructive options.
Although insurance coverage and educational campaigns are a prerequisite to improving access, neither are sufficient to improve overall access to reconstructive options. First, recent studies have found that 55 million people in the United States live in areas with fewer than 1 plastic surgeon per 100,000 population and an established direct correlation exists between plastic surgeon availability and receipt of breast reconstruction.5,6 Insurance coverage and patient education alone will not improve access for women in this setting. Second, the presence of a plastic surgeon does not guarantee reconstructive services or the availability of reconstructive options (ie, implant-based versus autologous reconstruction). When only a single type of breast reconstruction is offered, some women may choose not to undergo reconstruction or undergo the available surgery with a suboptimal experience or outcome. These experiences have been cited by women opting to “go flat,”7 another term for not having breast reconstruction after mastectomy.
If patients undergoing mastectomy for cancer have limited reconstructive options, then additional policies are warranted to improve access to care beyond insurance coverage and patient education. Therefore, we conducted this study to evaluate whether variation in breast reconstruction modality exists across health service areas (HSAs) as a potential marker for a woman's available reconstructive options. Because previous research has noted an influence between plastic surgeon density and breast reconstruction rates, we similarly assessed whether plastic surgeon density was correlated with a broader mix of reconstructive options. We hypothesized that some HSAs would favor implant-based reconstructions while others would display a broader mix of reconstructive procedures.
We conducted a cross-sectional study using the 2009–2012 Arkansas, Florida, Nebraska, and New York state inpatient databases and the 2009–2011 California state inpatient databases. These are available through the Agency for Healthcare Research and Quality's Healthcare Cost and Utilization Project.8 These data are a census of hospital discharges from acute care, nonfederal, community hospitals. Each discharge abstract includes patient demographic, anticipated payer, discharge disposition information, and up to 25 International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) procedure and diagnostic codes. These states were selected for inclusion because of their geographic diversity and large populations which accounted for 24% of the United States' adult population in 2010. Additionally, these years were selected because of the quality and usability of the plastic surgeon workforce data required for the secondary analysis.
Patient Selection Process
From the state inpatient databases, we identified discharges for women at least 18 years of age who underwent mastectomy (ICD-9-CM 85.4x, 85.33–85.36) for a diagnosis of breast cancer (ICD-9-CM 174.x, 233.0) with or without concurrent implant (ICD-9-CM 85.33, 85.35, 85.5x, 85.95) or autologous tissue-based (ICD-9-CM 85.6x, 85.7x, 85.8x, 85.33, 85.35, 85.84, 85.85, 85.89) breast reconstruction.
Primary Outcome: Determining Breast Reconstruction Modality
The primary outcome for this study was the modality used for breast reconstruction defined as follows: 1) implant or 2) autologous tissue-based reconstruction. Implant-based reconstruction included patients who underwent either first-stage expander placement or direct to implant reconstruction at the time of mastectomy. We defined autologous tissue reconstruction as any pedicled or free flap with or without implant reconstruction at the time of mastectomy. For patients who underwent a combination of procedures (ie, pedicled latissimus with concurrent expander placement), we classified these patients as having autologous reconstruction. For subsequent subgroup analyses, we also identified patients who underwent free tissue transfer with or without concurrent implant placement.
Covariates for Risk Standardization
We defined several patient variables for subsequent risk standardization that may impact the ability of providers to offer immediate reconstruction. This included a patient's age, diagnosis (in situ disease, invasive cancer, or metastatic cancer), degree of medical comorbidity as defined by the enhanced Elixhauser algorithm9,10 and the patient's smoking status. For each of the 31 medical conditions identified by Elixhauser, we considered a condition present if it was a listed diagnosis during the hospitalization for mastectomy or at any hospital admission in the 24 months preceding surgery. The Elixhauser index score was then calculated for each patient based on previous research11 In a similar manner, we assessed whether the patient had a history of tobacco use (ICD-9-CM 305.1).
Aggregation to the HSA Level and Supplementary Workforce Data
The National Center for Health Statistics defines HSAs12 as a single county or cluster of counties, which are “self-contained” in respect to hospital care. The National Cancer Institute subsequently modified these areas by dividing HSAs that straddled state lines so that all HSAs are contained within state boundaries. The state inpatient databases contain county-level identifiers for the location where treatment is provided. Using these identifiers, we first aggregated all clinical data to the county and then to the HSA level. To this clinical dataset, we merged plastic surgeon workforce data (plastic surgeons per 100,000 population within each HSA) obtained from the 2013–2014 Area Health Resource File as previously described.5,6
First, we calculated risk-standardized, immediate breast reconstruction rates (all modality) for each HSA using a 2-level (patient and HSA), hierarchical generalized linear model. In this model, breast reconstruction was the binary outcome, and the independent variables are those described previously. Models also included HSA random intercepts to account for clustering of patients within HSAs and permit separation of the within and between HSA variation in breast reconstruction rates after accounting for patient characteristics. The predicted-to-estimated ratio obtained from these models was then multiplied by the mean, unadjusted immediate breast reconstruction rate among all HSAs included in the study to yield the risk-standardized rates. Next, we calculated risk-standardized autologous (pedicled and free) and free flap (free flap only) reconstruction rates for each HSA. Finally, to assess the relationship between plastic surgeon density and the risk-standardized, immediate breast, autologous, or free flap breast reconstruction rates, we calculated volume-weighted correlation coefficients.
The final sample included 67,185 women across 103 HSAs in 5 states. The average age of the cohort was 58.5 years; most women had either private insurance (53.5%) or Medicare (32.9%), and surgery was most often a unilateral mastectomy for invasive cancer (Table 1).
Patient Level Outcomes
At the patient level, 48.7% (n = 32,702) of women underwent immediate breast reconstruction. Implant-based reconstruction was the most common reconstructive modality (n = 23,811; 72.8%), followed by autologous tissue (n = 5,440; 16.6%) or a combination of both modalities (n = 3,451; 10.6%). The implant-based cohort was predominantly composed of first-stage expander placement (87.1%) rather than direct-to-implant (12.9%) reconstruction. Among the autologous tissue cohort, deep inferior epigastric perforator or free transverse rectus abdominis flaps (65.4%) were more common than pedicled transverse rectus abdominis flaps (19.0%). When tissue was used in combination with an implant, a pedicled latissimus dorsi flap (35.1%) was the most common named flap.
Variation in Breast Reconstruction Rates Across HSAs
At the HSA level (n = 103 HSAs), the median, risk-standardized, immediate breast reconstruction rate was 48.9%. Although all HSAs included provided mastectomy, the breast reconstruction rate varied widely from 27.1% in the lowest quartile to 75.9% in the highest quartile (interquartile range [IQR] = 48.8%). When breast reconstruction was performed, implant-based options predominated with low-risk standardized, autologous tissue (median = 11.5%; IQR = 15.8%) and free flap (5.5%; IQR = 6.6%) reconstruction rates noted. When viewed another way, 30 HSAs (29.1%) performing mastectomy for cancer did not perform any immediate breast reconstruction procedures. Of the HSAs performing breast reconstruction (n = 73), 13.7% (n = 10) did not perform any pedicled or free tissue reconstruction, whereas 23.3% (n = 17) did not perform any free tissue reconstructions (Figs. 1–3).
Correlation Between Plastic Surgeon Workforce and Breast Reconstruction Rates
Plastic surgeon density varied across HSAs (median = 1.3 plastic surgeons per 100,000 population; IQR = 2.7) with 34 HSAs (33.0%) appearing to not have a plastic surgeon. There was a direct and significant relationship between an HSA's plastic surgeon density and risk-standardized breast reconstruction (correlation coefficient [r] = 0.675; P < 0.001), autologous tissue (r = 0.410, P < 0.001), and free flap (r = 0.44, P < 0.001) reconstruction rates.
There is substantial geographic variation in both immediate breast reconstruction rates and the modalities used for breast reconstruction across HSAs. Additionally, nearly 30% of HSAs performing mastectomy for cancer did not perform any breast reconstruction. When reconstruction was performed, implant-based modalities were either the predominate method of reconstruction or the only form of reconstruction performed. Taken together, this raises significant concerns regarding a woman's options for breast reconstruction overall and available types of reconstruction simply based on the area where treatment is provided.
Substantial geographic variation in the use of a specific breast reconstruction modality may signify underutilization or overutilization of a given procedure for numerous reasons. Some reasons may be patient driven. For example, patient preference may vary across regions whereby a patient informed of all options may preferentially chose implant-based reconstruction,13,14 or geographic variation may appear because of regionalization of autologous reconstructive care15 with the goal of improving outcomes.16 However, other causes may not be patient centric. In this scenario, variation may be related to lack of access, or referral, to a plastic surgeon. Identifying the underlying cause is important because not all women are excellent candidates for both procedures, nor do all women want the same form of breast reconstruction. The shepherding of women into a single form of breast reconstruction can lead to poor outcomes and the decision to opt for no reconstruction7 instead and, more broadly, contributes to a negative view of plastic surgery in this setting.
Once a patient reaches the plastic surgeon, access to all forms of postmastectomy breast reconstruction is not a guarantee. First, nearly 50 million people live in areas with fewer than 1 plastic surgeon per 100,000 population. Women who live in these areas undergo breast reconstruction less often than those in areas with a greater density of plastic surgeons. Second, identifying a plastic surgeon is not a guarantee that they provide breast reconstruction services in general or all forms of reconstruction specifically. Previous studies have demonstrated that the services offered by plastic surgeons vary between surgeons and throughout one's career. This may be related to the time commitment related to autologous reconstruction, lack of reimbursement relative to this time commitment for autologous reconstruction, or uneasiness with the technical components of the procedures. This may be reflected in the current study where some HSAs had recorded no immediate breast reconstructions, or when breast reconstruction was offered, no autologous methods were used.
Several options exist for increasing accessibility to breast reconstruction in all its forms. First, some plastic surgeons may be uncomfortable offering autologous forms of reconstruction because of lack of exposure during training or practicing in a location where free tissue transfer would be challenging within their hospitals. In these cases, efforts to improve training opportunities within residency and afterward may be beneficial. Furthermore, refining and popularizing autologous reconstructive techniques which do not require microvascular anastomoses may increase accessibility to autologous reconstruction. Second, plastic surgeons who perform multiple breast reconstruction modalities including pedicle or free tissue-based measures should be granted additional reimbursement for both the time and resource investment. Although reimbursement may be higher initially, this may reduce costs in the long term for patients who would rather have an autologous reconstruction. Third, if accessibility to breast reconstruction or the modality of choice for the patient is unavailable in her treatment area, regional referral centers should be established to channel patients to alternative locations where reconstructive services are available. Further advocacy efforts are needed to ensure women have access to the breast reconstruction options that current advocacy efforts are designed to inform them of.
This study should be considered in the context of several important limitations. First, the presented data were collected from a total of 5 states which may not be representative of the country as a whole. The large population size and geographic diversity should aid in lessening this concern. Second, this study relies on the accuracy of administrative coding to properly form the study cohort and identify those who received breast reconstruction. Although previous studies have called into question the sensitivity and specificity of breast reconstruction coding, this was primarily for more granular definitions (ie, deep inferior epigastric perforator, SIEA), not breast reconstruction in general. For this reason, we focused on the larger groups of autologous and implant reconstruction. Finally, the data in this study were collected from 2009 to 2012 and may not represent current trends in these states. However, significant shifts in practice patterns whereby a broader mix of reconstructive options is available to patients in the form of autologous options in the intervening years are unlikely.
In conclusion, the location where a woman receives a mastectomy for breast cancer may directly impact her likelihood of receiving immediate breast reconstruction overall and the type of breast reconstruction performed.
2. Albornoz CR, Cohen WA, Razdan SN, et al. The impact of travel distance on breast reconstruction
in the United States. Plast Reconstr Surg
3. Buckley EJ, Zahnd WE, Rea DJ, et al. Impact of rural–urban status on survival after mastectomy without reconstruction versus mastectomy with reconstruction. Am J Surg
5. Bauder AR, Gross CP, Killelea BK, et al. The relationship between geographic access to plastic surgeons and breast reconstruction
rates among women undergoing mastectomy for cancer. Ann Plast Surg
6. Bauder AR, Sarik JR, Butler PD, et al. Geographic variation in access to plastic surgeons. Ann Plast Surg
9. Elixhauser A, Steiner C, Harris DR, et al. Comorbidity measures for use with administrative data. Med Care
10. Quan H, Sundararajan V, Halfon P, et al. Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data. Med Care
11. van Walraven C, Austin PC, Jennings A, et al. A modification of the Elixhauser comorbidity measures into a point system for hospital death using administrative data. Med Care
12. Centers for Disease Control and Prevention. National Center for Health Statistics. Updated July 12, 2018. Available at: https://www.cdc.gov/nchs/index.htm
. Accessed June 9, 2018.
13. de Ligt KM, van Bommel ACM, Shreuder K, et al. The effect of being informed on receiving immediate breast reconstruction
in breast cancer
patients. Eur J Surg Oncol
14. Ashraf AA, Colakoglu S, Nguyen JT, et al. Patient involvement in the decision-making process improves satisfaction and quality of life in postmastectomy breast reconstruction
. J Surg Res
15. Iskandar ME, Dayan E, Lucido D, et al. Factors influencing incidence and type of postmastectomy breast reconstruction
in an urban multidisciplinary cancer center. Plast Reconstr Surg
16. Tuggle CT, Patel A, Broer N, et al. Increased hospital volume is associated with improved outcomes following abdominal-based breast reconstruction
. J Plast Surg Hand Surg