Abstract
Background:
The BREAST-Q is the most commonly used patient reported outcome measure (PROM) for breast reconstruction research. However, clinical implementation of this PROM has been impeded by a lack of context for score interpretation. The aim of this study was to define reference values for the BREAST-Q at discrete timepoints following surgery, generating a tool for real-time score interpretation.
Methods:
BREAST-Q scores were prospectively obtained in women who underwent implant or autologous post-mastectomy breast reconstruction preoperatively and for 2 years following surgery at Memorial Sloan Kettering Cancer Center (MSK). Descriptive statistics were used to create reference values for BREAST-Q satisfaction and quality-of-life subscales. Reference values were externally validated by comparing patient characteristics and BREAST-Q scores between the MSK and the Mastectomy Reconstruction Outcomes Consortium (MROC) study cohorts.
Results:
Overall, 3268 MSK patients and 2814 MROC patients were included for analysis. When MSK and MROC cohorts were compared for validation, there were some statistical differences in BREAST-Q scores; however, most of these differences did not meet the minimal clinically important difference of 4 points. Reference values were used to create the BREAST-Q Real-Time Engagement and Communication Tool (REACT).
Conclusions:
Using a large cohort of patients, we have defined BREAST-Q reference values for post-mastectomy breast reconstruction patients for use in clinical practice. The BREAST-Q REACT will help breast reconstruction providers gauge patient wellbeing and satisfaction relative to the “average” breast reconstruction patient and determine which patients may benefit from additional intervention.
INTRODUCTION
Since its development in 2009,1 the BREAST-Q has become both the gold standard and most commonly used patient-reported outcome measure (PROM) for post-mastectomy breast reconstruction. Extensively validated2, 3 and highly reliable, the BREAST-Q has been used in breast reconstruction research over the past decade and has expanded our understanding of how reconstruction technique, timing, adjunct therapies, and patient characteristics impact patient-reported outcomes.4–7 In 2020, the U. S. Food and Drug Administration (FDA) qualified the BREAST-Q Reconstruction Module for breast reconstruction studies as part of the FDA’s Medical Device Development Tool Program.8 However, the BREAST-Q is not commonly implemented in routine clinical practice for breast reconstruction; a systematic review by Cohen et al. reported that 71% of studies were cross-sectional, indicating that most institutions were not administering the BREAST-Q longitudinally.9 Even among studies that have assessed BREAST-Q scores longitudinally—for example, the findings of the Mastectomy Reconstruction Outcomes Consortium (MROC) study—the scores were collected for research rather than for direct patient care.10, 11 Patients and providers in the study were unable to use BREAST-Q scores to guide clinical decision making since, at the time of data collection, there was limited understanding of how BREAST-Q scores should be interpreted in a clinical context.
A major unresolved limitation in the clinical utility of the BREAST-Q for post-mastectomy breast reconstruction has been the lack of context with which to interpret and determine the relevance of BREAST-Q scores for an individual patient. Outside of breast reconstruction, researchers have reported on the difficulty of interpreting PROM data in clinical practice when there are no score cut-offs or thresholds that signal a need for clinical intervention.12–14 One strategy to provide this context is by establishing normalized BREAST-Q scores to serve as reference values against which BREAST-Q data can be compared. Mundy et al. (2017)15 and Klifto et al. (2020)16 recently contributed reference points when they published their normalized BREAST-Q scores, but their studies were cross-sectional and included women who did not have breast cancer. Although these normalized BREAST-Q scores help illustrate general trends in wellbeing and satisfaction for the average woman undergoing breast surgery, they do not reflect the full experience of the post-mastectomy breast reconstruction patient, thereby limiting their applicability as standardized reference values for longitudinal interpretation of BREAST-Q scores in this population.
To add to this growing knowledge base, we report a longitudinal study of BREAST-Q scores collected preoperatively and postoperatively for up to 2 years after initial reconstructive surgery for 3268 post-mastectomy breast reconstruction patients. Our first aim was to determine BREAST-Q reference values for post-mastectomy patients using this large cohort and then to externally validate these reference values through comparison with a multi-center cohort of breast reconstruction patients. We then used these validated reference values to create the BREAST-Q Real-Time Engagement and Communication Tool (REACT), which aids in BREAST-Q score interpretation and can guide clinical decision-making for providers who wish to incorporate the BREAST-Q as a PROM in routine clinical practice.
PATIENTS AND METHODS
Study Design and Population
We conducted a retrospective review of post-mastectomy breast reconstruction patients at Memorial Sloan Kettering Cancer Center (MSK) following approval by the institutional review board (IRB:18-202A). Patients who underwent immediate or delayed autologous or two-stage implant-based breast reconstruction from January 2007 to December 2017 were included. Patients must have completed at least one BREAST-Q assessment preoperatively and up to 2 years postoperatively to be included in this study. BREAST-Q assessments were timed based on the date of their definitive reconstruction (i.e., autologous reconstruction or exchange of tissue expander for permanent implant). Data on demographic characteristics, including age, race, and marital status, as well as clinical characteristics, including smoking status, diabetes, chemotherapy (including neoadjuvant therapy), and radiation therapy, were collected through chart reviews.
BREAST-Q
The BREAST-Q Reconstruction module (qportfolio.org/breast-q) was used to assess patient-reported outcomes across the following domains: 1) Satisfaction with Breasts, 2) Satisfaction with Outcome, 3) Satisfaction with Nipple Reconstruction, 4) Psychosocial Wellbeing, 5) Physical Wellbeing of the Chest, 6) Physical Wellbeing of the Abdomen, and 7) Sexual Wellbeing. BREAST-Q scores for each subscale are determined by converting raw scores into Q-scores, which range from 0–100, with higher scores indicating better outcomes.
MSK Cohort
We used an online platform to implement the BREAST-Q Reconstruction Module and collect data on quality of life and patient satisfaction. Satisfaction with Breast, Satisfaction with Outcome, Satisfaction with Nipple Reconstruction scores were collected preoperatively and at 6 months, 1 year, and 2 years after their definitive reconstruction (permanent implant placement or autologous reconstruction). Physical Wellbeing of the Chest, Physical Wellbeing of the Abdomen, Psychosocial Wellbeing, Sexual Wellbeing scores were collected longitudinally preoperatively and postoperatively at 2 weeks, 6 weeks, 3 months, 6 months, 1 year, and 2 years after definitive reconstruction.
MROC Cohort
To externally validate the reference values derived from the MSK cohort, patient characteristics, and BREAST-Q scores were collected preoperatively and postoperatively at 3 months, 1 year, and 2 years were compared with those of the MROC cohort—a 5-year prospective study that compared clinical and patient-reported outcomes for post-mastectomy breast reconstruction at 11 institutions in the United States and Canada. MROC participants completed the BREAST-Q Reconstruction Module via an online survey collection tool preoperatively and postoperatively at 1 week, 3 months, 1 year, and 2 years.17 The same inclusion and exclusion criteria used for the MSK cohort were applied to the MROC cohort. MSK was one of the institutions included in the MROC cohort, yet the cohorts were independent, with the MSK patients included in the MROC study not initially being directly included in the MSK patient cohort.
Sample Size Estimate
No protocol is available for determining the appropriate sample size for generating reference values or normative data for BREAST-Q, specifically; however, recommendations for a broadly applicable health-related quality-of-life questionnaire, the SF-36, have used sizes of 200 subjects per cohort group to generate normative data.18 While the recommendations are not an exact match for our study scenario, they are the most appropriate to apply for this study.
Statistical Analysis
Descriptive statistics were used to summarize patient characteristics and BREAST-Q scores for the MSK and MROC cohorts. After visually assessing for normality using violin plots, Chi-square, Fisher’s exact, and Mann-Whitney U tests were used where appropriate to compare the patient characteristics and BREAST-Q scores of the MSK and MROC cohorts. All statistical analyses were performed in R (version 4.0.2, tidyverse, readxl) and visualized using GraphPad Prism (version 8.4.2, GraphPad Software, San Diego, CA). Results were considered statistically significant if p < 0.05.
Additionally, we examined whether statistically significant BREAST-Q score differences also met the minimal clinically significant difference (MCID) of 3–4 points.19 The MCID should be considered in addition to statistical significance since this is the amount of change that patients actually perceive to be important. For the BREAST-Q, the MCID was systematically quantified through distribution-based analysis using the MROC cohort.19
RESULTS
MSK Cohort Characteristics
A total of 3268 patients were included in the MSK cohort: 2932 (89.7%) implant reconstruction patients and 336 (10.3%) autologous reconstruction patients, meeting sample size recommendations for determining reference values. Patient and surgical characteristics are outlined in Table 1.
Table 1.
Demographic and Clinical Characteristics of MSK Cohort
| Patient characteristic | Total Cohort n = 3268 |
Implant n = 2932 |
Autologous n = 336 |
|---|---|---|---|
|
| |||
| Age, average (SD) | 49.57 (9.87) | 49.53 (10.05) | 49.92 (8.11) |
|
| |||
| Body Mass Index, average (SD) | 25.66 (5.25) | 25.29 (5.14) | 28.86 (5.13) |
|
| |||
| Race, n (%) | |||
|
| |||
| White, not Hispanic | 2599 (79.5%) | 2369 (80.8%) | 230 (68.5%) |
| Black, not Hispanic | 229 (7.0%) | 186 (6.3%) | 43 (12.8%) |
| Asian | 163 (5.0%) | 148 (5.0%) | 15 (4.5%) |
| Hispanic or Latino | 166 (5.1%) | 145 (4.9%) | 21 (6.3%) |
| Native American/Pacific Islander | 2 (0.06%) | 1 (0.03%) | 1 (0.3%) |
| Unknown | 109 (3.3%) | 83 (2.8%) | 26 (7.7%) |
|
| |||
| Marital Status, n (%) | |||
|
| |||
| Single | 539 (16.5%) | 482 (16.4%) | 57 (17.0%) |
| Married | 2408 (73.7%) | 2159 (73.6%) | 249 (74.1%) |
| Life/Domestic Partner | 13 (0.4%) | 12 (0.4%) | 1 (0.3%) |
| Separated | 38 (1.2%) | 36 (1.2%) | 2 (0.6%) |
| Divorced | 199 (6.1%) | 183 (6.2%) | 16 (4.8%) |
| Widowed | 71 (2.2%) | 60 (2.1%) | 11 (3.3%) |
| Unknown | 0 (0.0%) | 0 (0.0%) | 0 (0.0%) |
|
| |||
| Diabetes, n (%) | 219 (6.7%) | 181 (6.2%) | 38 (11.3%) |
|
| |||
| Smoking, n (%) | |||
|
| |||
| Never Smoker | 2065 (63.2%) | 1860 (63.4%) | 205 (61.0%) |
| Previous Smoker | 932 (28.5%) | 806 (27.5%) | 126 (37.5%) |
| Current Smoker | 183 (5.6%) | 178 (6.1%) | 5 (1.5%) |
| Unknown | 88 (2.7%) | 88 (3.0%) | 0 (0.0%) |
|
| |||
| Chemotherapy, n (%) | 1320 (40.4%) | 1216 (41.5%) | 104 (31.0%) |
|
| |||
| Radiation Therapy, n (%) | |||
|
| |||
| None | 2398 (73.4%) | 2175 (74.2%) | 223 (66.4%) |
| History of prior radiation | 349 (10.7%) | 239 (8.2%) | 110 (32.7%) |
| During or after reconstruction | 521 (15.9%) | 518 (17.7%) | 3 (0.9%) |
BREAST-Q Satisfaction Scores
BREAST-Q scores for the subscales Satisfaction with Breasts, Satisfaction with Outcome, and Satisfaction with Nipple Reconstruction are shown in Figure 1. For Satisfaction with Breasts, the autologous cohort had lower preoperative scores on average than the implant cohort but had higher postoperative scores at 2 years. Mean and median BREAST-Q satisfaction domain scores for each reconstruction type and overall can be found in Table, Supplemental Digital Content 1.
Figure 1.
BREAST-Q scores for satisfaction domains for implant and autologous reconstruction patients in the MSK cohort.
BREAST-Q Quality-of-Life Scores
BREAST-Q scores for Physical Wellbeing of Chest, Physical Wellbeing of Abdomen (autologous reconstruction only), Psychosocial Wellbeing, and Sexual Wellbeing are shown in Figure 2. BREAST-Q scores decreased immediately after surgery, but then trended upwards. Physical Wellbeing and Sexual Wellbeing scores at 2 years approached but did not exceed preoperative scores, while Psychosocial Wellbeing scores at 2 years were greater than those collected preoperatively. Of note, there were no large differences in postoperative scores at 2 years for any of the subscales between the implant and autologous reconstruction cohorts. It is, however, important to note that preoperative scores for the autologous cohort were lower than those for the implant cohort. Mean and median BREAST-Q quality-of-life scores for each reconstruction type and overall can be found in Table, Supplemental Digital Content 2.
Figure 2.
BREAST-Q scores for quality-of-life domains for implant and autologous reconstruction patients in the MSK cohort.
MSK Cohort Validation
The MROC cohort consisted of 2814 patients: 1958 (69.6%) implant reconstruction patients and 856 (30.4%) autologous reconstruction patients. Overall, although statistically significant, differences in proportions between the MSK and MROC cohorts were minimal for most characteristics examined. The largest difference in proportions was found in history of chemotherapy (MSK vs. MROC overall: 40.4% vs. 29.9%, p < 0.001). Of note, 66.4% of the MSK autologous cohort had no radiation treatment, compared with 54.1% of the MROC autologous cohort (p < 0.001). (Table 2) Of the MROC cohort, 78% of patients were from non-MSK institutions.
Table 2.
Comparison of Demographic Characteristics between MSK and MROC Cohorts
| MSK Cohort | MROC Cohort | p-values (MSK vs. MROC) | |||||||
|---|---|---|---|---|---|---|---|---|---|
|
| |||||||||
| Total n = 3268 | Implant n = 2932 | Autologous n = 336 | Total n = 2814 | Implant n = 1958 | Autologous n = 856 | Overall | Implant | Autologous | |
|
| |||||||||
| Age, average (SD) | 49.57 (9.87) | 49.53 (10.05) | 49.92 (8.11) | 49.57 (10.20) | 48.58 (10.53) | 51.85 (8.99) | p = 0.999 | p = 0.002 | p < 0.001 |
|
| |||||||||
| BMI, average (SD) | 25.66 (5.25) | 25.29 (5.14) | 28.86 (5.13) | 26.69 (5.76) | 25.69 (5.46) | 29.00 (5.77) | p < 0.001 | p = 0.012 | p = 0.698 |
|
| |||||||||
| Race, n (%) | p < 0.001 | p < 0.001 | p < 0.001 | ||||||
|
| |||||||||
| White, not Hispanic | 2599 (79.5%) | 2369 (80.8%) | 230 (68.5%) | 2324 (82.6%) | 1624 (82.9%) | 700 (81.8%) | |||
| Black, not Hispanic | 229 (7.0%) | 186 (6.3%) | 43 (12.8%) | 162 (5.8%) | 115 (5.9%) | 47 (5.6%) | |||
| Asian | 163 (5.0%) | 148 (5.0%) | 15 (4.5%) | 134 (4.8%) | 88 (4.5%) | 46 (5.4%) | |||
| Hispanic or Latino | 166 (5.1%) | 145 (4.9%) | 21 (6.3%) | 144 (5.1%) | 106 (5.4%) | 38 (4.4%) | |||
| Native American/Pacific Islander | 2 (0.06%) | 1 (0.03%) | 1 (0.3%) | 29 (1.0%) | 12 (0.6%) | 17 (2.0%) | |||
| Unknown | 109 (3.3%) | 83 (2.8%) | 26 (7.7%) | 21 (0.7%) | 13 (0.7%) | 8 (0.9%) | |||
|
| |||||||||
| Marital Status, n (%) | p < 0.001 | p < 0.001 | p < 0.001 | ||||||
|
| |||||||||
| Single | 539 (16.5%) | 482 (16.4%) | 57 (17.0%) | 233 (8.3%) | 172 (8.8%) | 61 (7.1%) | |||
| Married | 2408 (73.7%) | 2159 (73.6%) | 249 (74.1%) | 2050 (72.9%) | 1440 (73.5%) | 610 (71.3%) | |||
| Life/Domestic Partner | 13 (0.4%) | 12 (0.4%) | 1 (0.3%) | 136 (4.8%) | 83 (4.2%) | 53 (6.2%) | |||
| Separated | 38 (1.2%) | 36 (1.2%) | 2 (0.6%) | 57 (2.0%) | 32 (1.6%) | 25 (2.9%) | |||
| Divorced | 199 (6.1%) | 183 (6.2%) | 16 (4.8%) | 250 (8.9%) | 172 (8.8%) | 78 (9.1%) | |||
| Widowed | 71 (2.2%) | 60 (2.1%) | 11 (3.3%) | 69 (2.5%) | 43 (2.2%) | 26 (3.0%) | |||
| Unknown | 0 (0.0%) | 0 (0.0%) | 0 (0.0%) | 19 (0.7%) | 16 (0.8%) | 3 (0.4%) | |||
|
| |||||||||
| Diabetes, n (%) | 219 (6.7%) | 181 (6.2%) | 38 (11.3%) | 117 (4.2%) | 56 (2.9%) | 61 (7.1%) | p < 0.001 | p < 0.001 | p = 0.025 |
|
| |||||||||
| Smoking, n (%) | p < 0.001 | p < 0.001 | p = 0.381 | ||||||
|
| |||||||||
| Never Smoker | 2065 (63.2%) | 1860 (63.4%) | 205 (61.0%) | 1816 (64.5%) | 1311 (67.0%) | 505 (59.0%) | |||
| Previous Smoker | 932 (28.5%) | 806 (27.5%) | 126 (37.5%) | 896 (31.8%) | 572 (29.2%) | 324 (37.9%) | |||
| Current Smoker | 183 (5.6%) | 178 (6.1%) | 5 (1.5%) | 74 (2.6%) | 50 (2.6%) | 24 (2.8%) | |||
| Unknown | 88 (2.7%) | 88 (3.0%) | 0 (0.0%) | 28 (1.0%) | 25 (1.3%) | 3 (0.4%) | |||
|
| |||||||||
| Chemotherapy, n (%) | 1320 (40.4%) | 1216 (41.5%) | 104 (31.0%) | 841 (29.9%) | 616 (31.5%) | 225 (26.3%) | p < 0.001 | p < 0.001 | p = 0.008 |
|
| |||||||||
| Radiation Therapy, n (%) | p < 0.001 | p < 0.001 | p < 0.001 | ||||||
|
| |||||||||
| None | 2398 (73.4%) | 2175 (74.2%) | 223 (66.4%) | 1910 (67.9%) | 1447 (73.9%) | 463 (54.1%) | |||
| History of prior radiation | 349 (10.7%) | 239 (8.2%) | 110 (32.7%) | 360 (12.8%) | 110 (5.6%) | 250 (29.2%) | |||
| During or after reconstruction | 521 (15.9%) | 518 (17.7%) | 3 (0.9%) | 544 (19.3%) | 401 (20.5%) | 143 (16.7%) | |||
The BREAST-Q scores for all subscales for the MSK and MROC cohorts were compared preoperatively and postoperatively at 3 months, 1 year, and 2 years after initial reconstructive surgery (Figure 3 and Table, Supplemental Digital Content 3). The minimal clinically important difference of 4 points was not met for most score comparisons, either at baseline or postoperatively. At baseline, scores of MSK implant patients did not differ from those of MROC implant patients in a clinically significant manner, but MSK and MROC autologous patients had clinically significant differences in Satisfaction with Breast (MSK – MROC = −7.6 points, p<0.001). Post-operatively, clinically important differences were most prevalent at 3 months after surgery. At 3 months, Physical Wellbeing scores did not differ clinically for both implant and autologous patients, but differences existed for Psychosocial (Implants: MSK – MROC = +9.4 points, p<0.001; Autologous: MSK – MROC = +4.7 points, p=0.019) and Sexual Wellbeing (Implants: MSK – MROC = +6.1 points, p<0.001). By 1 year after surgery, differences between MSK and MROC cohorts diminished, with only Physical Wellbeing of the Abdomen at 2 years having both a statistically significant and clinically significant difference (MSK – MROC = +4.9 points, p=0.003).
Figure 3.
BREAST-Q score differences between MSK and MROC cohorts. Values displayed were determined by subtracting the median BREAST-Q score for the MROC cohort from the median BREAST-Q score of the MSK cohort and taking the absolute value of that difference.
* indicates statistical significance with p < 0.05.
Dashed line indicates the minimal clinically important difference for the BREAST-Q
BREAST-Q Real-Time Engagement and Communication Tool (REACT)
The data utilized to generate the current reference values were used to create the BREAST-Q REACT, a score interpretation tool, for each modality of breast reconstruction (Figure 4). The BREAST-Q score trajectory from baseline to 2 years after surgery for each domain is displayed for patients at the 25 percentile, 50th percentile and 75th percentile of BREAST-Q scores.
Figure 4.
BREAST-Q REACT score interpretation tool for a implant-based reconstruction and b autologous reconstruction
DISCUSSION
In this study, we defined and externally validated BREAST-Q reference values for post-mastectomy breast reconstruction patients using a large, diverse cohort in North America. We then used these reference values to develop the BREAST-Q REACT, which allows the patient and provider to interpret BREAST-Q scores in real-time for individualized patient care.
Importance of Reference Values for BREAST-Q Utilization in Clinical Practice
A single BREAST-Q score provides little clinically actionable information when examined in isolation. Providers must also have access to the expected range of scores for post-mastectomy breast reconstruction patients, which may explain why the BREAST-Q has not quickly gained traction in the clinical setting. Another consideration is that a patient’s BREAST-Q score changes throughout her recovery. The expected range of BREAST-Q scores should also then change over time so that providers are using the appropriate temporal context in which to interpret scores. In developing the BREAST-Q REACT, we did not generate one reference value, as seen in other studies on normalized values15, 16; rather, we created timepoint-based values as well as reference value curves to enhance interpretation in the postoperative setting.
Our goal was to create reference value curves that would be broadly applicable to various clinical practice types with diverse breast reconstruction patient populations. Large, longitudinal cohorts for the study of patient-reported outcomes require significant time and financial investment, yet this is the only method of determining reference values specific enough for any clinical practice. We compared our institution’s 10-year BREAST-Q post-mastectomy breast reconstruction cohort to a large, longitudinal, multi-center and multi-regional BREAST-Q post-mastectomy breast reconstruction cohort for external validation. BREAST-Q minimum clinically important differences (MCID) were derived based on Voineskos et al., which used distribution-based methods to determine that a difference of 4 points on each domain was what patients perceived to be important. We used the MCID for our study to verify that MSK and MROC patients were experiencing the same level of satisfaction and wellbeing in a clinically significant, rather than statistically significant, way.19 We identified few clinically meaningful differences in terms of patient characteristics and BREAST-Q scores particularly in the postoperative period, indicating that the reference values described in our study are applicable to most breast reconstruction patients in North America.
As implant and autologous breast reconstruction patients have different BREAST-Q trends,4, 10, 11 we generated separate BREAST-Q REACT tools for each reconstruction modality to provide patients and providers with individualized, patient-centric predictions of breast reconstruction outcomes to help manage patient expectations. BREAST-Q REACT could be adapted further based on other patient characteristics known to impact patient-reported outcomes, such as race, laterality or postoperative radiation therapy.4 However, the recognition of differential outcomes based on patient characteristics should potentially indicate a need to examine current clinical practices to better serve patients with poorer outcomes, rather than to adjust our reference values for what is standard for these patients. As more longitudinal BREAST-Q data is collected over time, future research can examine how BREAST-Q reference values should be refined or modified, if at all, based on these characteristics. For this study, we chose to create a generalized framework for BREAST-Q score interpretation, since the intention is not to predict exact outcomes but rather to give meaningful reference to a numerical value. In most medical assessments, reference values provide a wide range of expected or normal values, and abnormal values should be interpreted with clinical context in mind. Given the wide variation of scores seen in our cohort of patients, the BREAST-Q score is no different and provides a single, but important, aspect of the overall clinical picture. Physicians should use the BREAST-Q REACT as one additional tool to enhance decision-making.
Applying the BREAST-Q REACT to Clinical Practice
BREAST-Q REACT adds immediate, practical value to the BREAST-Q for clinical care, as it can potentially be used as an objective tool at the point of care for evaluating the need for additional surgical and non-surgical interventions. Sustained utilization of and engagement with PROMs is challenging when patients and providers feel the tools do not have clinical utility, which can occur when it is unclear how to interpret these measures or how to apply them to patient care.20, 21 With our tool, the BREAST-Q becomes more interpretable and therefore more applicable and relevant to any clinical practice, enabling more patient-centered care for breast reconstruction patients. Surgeons can use BREAST-Q REACT like a pediatric growth chart; a patient’s BREAST-Q scores over time can be charted on the tool to determine where her scores lie relative to the rest of the post-mastectomy breast reconstruction population. A single BREAST-Q score can also be plotted on the tool and compared to the reference values if the time from surgery is known. Ideally, patients will complete their BREAST-Q prior or during clinical encounters so it can be scored and charted onto REACT.
At our institution, we collect BREAST-Q scores as part of routine clinical care preoperatively and postoperatively. Patients with extremely low or falling BREAST-Q scores are identified during clinical visits and assessed for potentially beneficial interventions. As a hypothetical example, consider a 58-year-old patient who received autologous reconstruction for a bilateral mastectomy with axillary lymph node dissection. At 6 months after surgery, we notice that her Physical Wellbeing of the Chest domain score is 50, much below the 25th percentile for this domain in autologous reconstruction patients. During her next clinical encounter, we initiate discussion regarding her score, and she reveals limited arm mobility and pain, leading to a new referral to physical therapy to address her symptoms. Adding reference values to BREAST-Q interpretation may also reveal deficiencies beyond what is readily apparent. Patients who fall off their expected BREAST-Q score trajectory may warrant further investigation and may still benefit from interventions that optimize their clinical care, even if their scores appear acceptable. As another example, at 1 year, we notice that the same patient has a Satisfaction with Breasts score of 60, unchanged from her previous 6-month score when we expect that it should have improved. At her next visit, discussion about this reveals lingering dissatisfaction with the contour of her breast, leading us to recommend a revision procedure with fat grafting to potentially improve Satisfaction with Breast scores. The availability of reference values is an important first step in improving the clinical utility of the BREAST-Q. Future prospective or randomized controlled studies that identify which score trajectories warrant intervention and which interventions are most effective for addressing concerning BREAST-Q scores would improve this PROM’s clinical value further. We believe that the 25th percentile, graphically represented by the red line on the REACT, is a reasonable cutoff for initiating discussion and consideration of potential intervention. The benefits of this type of PROM-driven care have not yet been quantified, but some research has indicated that PROMs have a positive impact when used to guide clinical care.22–25
While the BREAST-Q REACT addresses one important concern for routine BREAST-Q utilization—the lack of reference values for score interpretation—it does not address other issues with implementation of BREAST-Q for clinical practice. One concern with the BREAST-Q could be its length, which may deter patients from completing the BREAST-Q and make it challenging to administer in clinic. At our institution, we have attempted to address this concern by administering the BREAST-Q through an online patient portal prior to the clinic visit, so patients can complete the BREAST-Q at home in their own time, rather than using clinic time. This also increases the likelihood that the BREAST-Q will be used to guide clinical decision-making, as patients and providers will have access to BREAST-Q results prior to the visit. Additionally, we do not administer every module of the BREAST-Q at every timepoint; rather, patients receive only quality-of-life modules during the immediate postoperative period, with satisfaction modules added later on. This way, patients are not burdened by modules that may not be as relevant to them earlier on in the recovery process. Computerized adaptive testing may also reduce survey burden on the BREAST-Q; it has been shown to reduce the number of questions needed by 38–75%.26
Strengths and Limitations
Our study is one of the first to generate reference values for post-mastectomy breast reconstruction patients using the BREAST-Q Reconstruction module, which are more applicable to clinical practice than normalized values proposed in previous studies. Mundy et al. (2017)15 utilized BREAST-Q scores from non-breast reconstruction and non-breast cancer patients in the Army of Women cohort, and Klifto et al. (2020)16 from gynecology clinics at a single center, basing their normalized BREAST-Q scores on the average woman rather than the average post-mastectomy breast reconstruction patient. Both important studies found that, in breast reconstruction patients, Satisfaction with Breast scores were higher and Physical Wellbeing of Chest scores were lower than normalized values at 1 year after initial reconstructive surgery. Therefore, normalized BREAST-Q scores may be limited benchmarks for post-mastectomy breast reconstruction patients. If applied as reference values in clinical practice, they may lead to unrealistic expectations or undue stress when patient scores do not match normalized scores. Additionally, these studies were cross-sectional and do not consider that breast reconstruction is a journey for most patients and not a single event. On the other hand, our reference values are the next step in this understanding and implementation to help inform patients and providers on what to reasonably expect at each stage of the reconstructive process.
A strength of this study is the use of our large and diverse cohort of post-mastectomy breast reconstruction patients with long-term BREAST-Q follow-up, making our BREAST-Q reference values applicable to a broader patient population. Nevertheless, a limitation to applicability is that our cohort is from a single, academic, tertiary medical center, which may support the use of the multi-center MROC cohort instead as the frame of reference for BREAST-Q reference values. To account for this limitation, we compared patient characteristics and BREAST-Q scores between our MSK cohort and the MROC cohort and found minor differences. Another strength is that we provide greater granularity than the MROC study, specifically in terms of BREAST-Q assessment timepoints, which may benefit providers wishing to use the BREAST-Q to guide clinical practice. Although it may be sufficient to assess satisfaction outcomes at time points further from surgery, as patients require time to recover or to undergo additional reconstructive procedures, we measure quality-of-life outcomes at 2 weeks, 6 weeks, and 3 months to assess for pain or other symptoms that require immediate treatment. Crucially, our MSK cohort is constantly growing, which allows us to refine reference values over time. BREAST-Q reference values may have to be refined as we modify our practices in terms of surgical technique and postoperative management. Since 2017, we have almost doubled the number of BREAST-Qs collected, giving us a robust database for further development of reference values. On the other hand, the MROC cohort, completed study recruitment in 2016. The MROC cohort is therefore well-suited for establishing the external validity of our MSK cohort, but reference values based on MROC may become outdated and cannot be further refined.
Our study is potentially limited by lower sample sizes for early postoperative time points. This may explain greater differences in MSK and MROC scores during these earlier time points, as response bias may influence median scores. Reference values for early postoperative time points may be particularly affected by patients who experience complications, especially if these patients are more likely to respond to BREAST-Q. However, major complications such as flap or implant loss are relatively rare occurrences, so they should not dramatically impact median BREAST-Q scores. Nonetheless, future research should aim to confirm these short-term reference values with a larger cohort. Additionally, our study does not generate reference values for the tissue expansion process; patients undergoing tissue expansion should receive quality-of-life domains of the BREAST-Q as this can inform providers of how to improve postoperative care for these patients. Our reference values may also not apply to patients who underwent different procedural pathways for their reconstruction, such as those who received direct-to-implant reconstruction. Further research should examine BREAST-Q scores for these reconstruction types and refine the BREAST-Q REACT accordingly. Finally, MSK was one of the institutions that participated in the MROC study, which impacts the external validity of the MSK cohort. However, we determined that approximately 80% of patients in the MROC cohort were from non-MSK institutions. This indicates that our external validation, while not perfect, still adequately compares our MSK cohort to North American patients overall.
CONCLUSIONS
We defined BREAST-Q reference values using a large longitudinal cohort of post-mastectomy breast reconstruction patients, and we created the BREAST-Q REACT to incorporate these reference values into a tool for clinical use. These values can be used to compare patient wellbeing and satisfaction over time to that of a “average” breast reconstruction patient and identify patients who may require additional interventions. The availability of these reference values will transform the BREAST-Q from largely a research tool to a patient care tool with widespread application in clinical practice and lead to greater incorporation of the BREAST-Q in standard breast reconstruction care. Utilization of the BREAST-Q in this way can provide patient-specific feedback at the point of care to improve patient education, manage expectations, and, ultimately, improve our clinical care and outcomes.
Supplementary Material
SYNOPSIS:
We created clinical reference values and BREAST-Q score interpretation tool based on BREAST-Q scores from 3268 breast reconstruction patients at Memorial Sloan Kettering Cancer Center. These reference values were externally validated using 2814 patients in the Mastectomy Reconstruction Outcomes Cohort.
ACKNOWLEDGEMENTS
This research was funded in part through the NIH/NCI Cancer Center Support Grant P30CA008748. Dagmar Schnau provided editorial assistance.
Footnotes
Declaration of Interest: Andrea Pusic is a codeveloper of the BREAST-Q, which is owned by Memorial Sloan Kettering Cancer Center. She receives a portion of the licensing fees (royalty payments) when the BREAST-Q is used in industry-sponsored clinical trials. Babak Mehrara is a consultant for PureTech Corporation. The remaining authors declare no conflicts of interest
REFERENCES
- 1.Pusic AL, Klassen AF, Scott AM, et al. Development of a new patient-reported outcome measure for breast surgery: the BREAST-Q. Plast Reconstr Surg 2009;124(2):345–53. [DOI] [PubMed] [Google Scholar]
- 2.Cano SJ, Klassen AF, Scott AM, et al. The BREAST-Q: further validation in independent clinical samples. Plast Reconstr Surg 2012;129(2):293–302. [DOI] [PubMed] [Google Scholar]
- 3.Fuzesi S, Cano SJ, Klassen AF, et al. Validation of the electronic version of the BREAST-Q in the Army of Women study. Breast 2017;33:44–49. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Nelson JA, Allen RJ, Jr., Polanco T, et al. Long-term patient-reported outcomes following postmastectomy breast reconstruction: an 8-year examination of 3268 patients. Ann Surg 2019;270(3):473–483. [DOI] [PubMed] [Google Scholar]
- 5.Nelson JA, Sobti N, Patel A, et al. The impact of obesity on patient-reported outcomes following autologous breast reconstruction. Ann Surg Oncol 2020;27(6):1877–1888. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Coriddi M, Shenaq D, Kenworthy E, et al. Autologous breast reconstruction after failed implant-based reconstruction: evaluation of surgical and patient-reported outcomes and quality of life. Plast Reconstr Surg 2019;143(2):373–379. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Razdan SN, Cordeiro PG, Albornoz CR, et al. Cost-effectiveness analysis of breast reconstruction options in the setting of postmastectomy radiotherapy using the BREAST-Q. Plast Reconstr Surg 2016;137(3):510e–517e. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.U. S. Food and Drug Administration. Medical Device Development Tools 10/20/2020, 2020. Available at: https://www.fda.gov/medical-devices/science-and-research-medical-devices/medical-device-development-tools-mddt. Accessed 10/30/2020.
- 9.Cohen WA, Mundy LR, Ballard TN, et al. The BREAST-Q in surgical research: A review of the literature 2009–2015. J Plast Reconstr Aesthet Surg 2016;69(2):149–62. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Santosa KB, Qi J, Kim HM, et al. Long-term patient-reported outcomes in postmastectomy breast reconstruction. JAMA Surg 2018;153(10):891–899. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Pusic AL, Matros E, Fine N, et al. Patient-reported outcomes 1 year after immediate breast reconstruction: results of the Mastectomy Reconstruction Outcomes Consortium study. J Clin Oncol 2017;35(22):2499–2506. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Shi Q, Mendoza TR, Cleeland CS. Interpreting patient-reported outcome scores for clinical research and practice: definition, determination, and application of cutpoints. Med Care 2019;57 Suppl 5 Suppl 1:S8–S12. [DOI] [PubMed] [Google Scholar]
- 13.Reeve BB, Wyrwich KW, Wu AW, et al. ISOQOL recommends minimum standards for patient-reported outcome measures used in patient-centered outcomes and comparative effectiveness research. Qual Life Res 2013;22(8):1889–905. [DOI] [PubMed] [Google Scholar]
- 14.Schunemann HJ, Akl EA, Guyatt GH. Interpreting the results of patient reported outcome measures in clinical trials: the clinician’s perspective. Health Qual Life Outcomes 2006;4:62. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Mundy LR, Homa K, Klassen AF, et al. Breast cancer and reconstruction: normative data for interpreting the BREAST-Q. Plast Reconstr Surg 2017;139(5):1046e–1055e. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Klifto KM, Aravind P, Major M, et al. Differences between breast cancer reconstruction and institutionally established normative data using the BREAST-Q Reconstruction Module: a comparative study. Plast Reconstr Surg 2020;145(6):1371–1379. [DOI] [PubMed] [Google Scholar]
- 17.Wilkins EG. Mastectomy Reconstruction Outcomes Consortium (MROC) Study (NCT01723423) July 13, 2017. Available at: https://clinicaltrials.gov/ct2/show/NCT01723423. Accessed 10/30/2020.
- 18.Gandek B, Ware JE Jr. Methods for validating and norming translations of health status questionnaires: the IQOLA Project approach. International Quality of Life Assessment. J Clin Epidemiol 1998;51(11):953–9. [DOI] [PubMed] [Google Scholar]
- 19.Voineskos SH, Klassen AF, Cano SJ, et al. Giving meaning to differences in BREAST-Q scores: minimal important difference for breast reconstruction patients. Plast Reconstr Surg 2020;145(1):11e–20e. [DOI] [PubMed] [Google Scholar]
- 20.Hancock SL, Ryan OF, Marion V, et al. Feedback of patient-reported outcomes to healthcare professionals for comparing health service performance: a scoping review. BMJ Open 2020;10(11):e038190. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Ovretveit J, Zubkoff L, Nelson EC, et al. Using patient-reported outcome measurement to improve patient care. Int J Qual Health Care 2017;29(6):874–879. [DOI] [PubMed] [Google Scholar]
- 22.Boyce MB, Browne JP, Greenhalgh J. The experiences of professionals with using information from patient-reported outcome measures to improve the quality of healthcare: a systematic review of qualitative research. BMJ Qual Saf 2014;23(6):508–18. [DOI] [PubMed] [Google Scholar]
- 23.Rivera SC, Kyte DG, Aiyegbusi OL, et al. The impact of patient-reported outcome (PRO) data from clinical trials: a systematic review and critical analysis. Health Qual Life Outcomes 2019;17(1):156. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Holmes MM, Lewith G, Newell D, et al. The impact of patient-reported outcome measures in clinical practice for pain: a systematic review. Qual Life Res 2017;26(2):245–257. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Salek S, Roberts A, Finlay AY. The practical reality of using a patient-reported outcome measure in a routine dermatology clinic. Dermatology 2007;215(4):315–9. [DOI] [PubMed] [Google Scholar]
- 26.Young-Afat DA, Gibbons C, Klassen AF, et al. Introducing BREAST-Q computerized adaptive testing: short and individualized patient-reported outcome assessment following reconstructive breast surgery. Plast Reconstr Surg 2019;143(3):679–684. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.