Incorporating Patient‐Reported Outcome Measures into Breast Surgical Oncology: Advancing Toward Value‐Based Care

Abstract

Value in health care is defined as the health outcome achieved per unit of cost. For health care systems, improving value means achieving better outcomes at lower costs. Improving outcomes, including patient‐reported outcomes (PROs), as well as more established metrics such as mortality and complication rates, ensures high‐quality care. This is particularly true in breast cancer surgery, where survival and recurrence rates are comparable across different surgical approaches. Outcomes reflecting survivorship quality may therefore better inform decision making regarding surgical approaches. PROs can be assessed using validated instruments known as patient‐reported outcome measures (PROMs). They are obtained directly from patients reflecting their health‐related quality of life (HRQOL). Ongoing initiatives strive to define PROMs that accurately reflect HRQOL and demonstrate value, with the goal of establishing benchmarks for quality of care. Clinicians caring for patients with breast cancer are well positioned to be involved in defining meaningful measures of value‐based breast cancer care. This article reviews value‐based breast cancer care in the context of locoregional therapy, with attention paid to the work done by the International Consortium of Health Outcome Measures in which a “standard set” of value‐based patient‐centered outcomes for breast cancer for international use is defined. In addition, an overview is provided of relevant PROMs and previously reported scores. Recommendations and future challenges for implementation of routine collection of PROs are also discussed.

Implications for Practice

Opportunity exists to act as early adopters of the routine collection of longitudinal patient‐reported outcome data for breast cancer, allowing transition of current care to value‐based cancer care.

Short abstract

The focus of this article is on the role of patient‐reported outcomes in advancing toward value‐based breast cancer care in the context of locoregional therapy, with attention to the work of the International Consortium of Health Outcome Measures, which defines a standard set of value‐based patient‐centered outcomes for breast cancer for international use.

Introduction

In value‐based health care (VBHC), the value of care is defined by the achieved health outcomes per total unit of cost (e.g., dollars spent) 1. The numerator of this equation is focused on outcomes that are defined by both provider‐reported outcomes and patient‐reported outcomes (PROs). PROs are assessed directly from patients, typically using validated patient‐reported outcome measures (PROMs), which aim to reflect their quality of life or function. In contrast to the current focus at most health care systems, these outcomes are not evaluating processes or volumes but are designed to assess the patient's own perspective of the results of the systems in place. It is expected that by using these outcomes to capture and evaluate the care delivered, higher value can be achieved 1. The VBHC initiative is expected to pave the way for shared decision making in treatment decisions. The key to this initiative is the systematic incorporation of PROMs in current (breast) cancer care.

Substantial efforts have been made by the International Consortium of Health Outcome Measures (ICHOM) to develop a “standard set” of value‐based patient‐centered outcomes for breast cancer 2, as well as a variety of other medical conditions, including lung cancer 3 and prostate cancer 4. The goal of the ICHOM initiative is to develop a standard set of outcomes felt to accurately assess the quality of life and that would be feasible to implement internationally and in a range of clinical settings. The goal of this standardization is to improve clinical care of patients, provide quality outcomes at the individual and institutional levels, and allow for comparative effectiveness research. The proposed standard set includes outcomes related to survival (overall and disease‐free), disutility of care (defined as adverse effects or toxicity of treatment), and long‐term health‐related quality of life (HRQOL). The ICHOM breast cancer outcome set consists of 75% PRO data 2. In an online survey, a recent study used PROMs (as proposed in the ICHOM set) within a Dutch breast cancer population sample and evaluated the participants’ satisfaction with and expectations of PROMs 5. We found statistically significant different results for PROs after different types of breast surgery, which are important for future clinical decision making. Most participants reacted positively to the idea of use of PROMs in clinical practice.

The health outcomes component of value must reflect all the care delivered to patients. To define the total cycle of care, a hierarchical order can be made 6. Outcomes defined within any VBHC system are disease‐specific, and multiple outcomes are needed to capture the total care cycle. To capture all these outcomes, ambulatory and hospital care must be integrated. Although current clinical care of patients with breast cancer generally involves a multidisciplinary approach administered in multiple settings, capturing the impact of this effort is a challenging aspect of the VBHC initiative. Surgical procedures and their long‐term impact on HRQOL are recognized as important aspects of this care cycle.

Cost, the denominator of the value equation, encompasses both direct (e.g., costs for procedures or admission rates) and indirect costs (e.g., days of sick leave). Because most health care systems do not currently capture these outcomes or costs, opportunities and challenges arise to affect both the numerator and denominator when moving toward a more VBHC system. Because of challenges in obtaining cost information and lack of substantial data on this topic, cost is outside the scope of our discussion.

PROs are important to all members of the multidisciplinary team caring for a patient with breast cancer. We focus here on the potential opportunities for surgeons to become early adopters of the outcome measures selected to ensure accurate reflection of care delivered, given the role of breast cancer surgeons over the entire care cycle for patients with breast cancer and this article's focus on local therapy. As emphasized by Tevis and colleagues in their recent publication 6, PROMs are expected to be a valuable addition to the more traditional outcome metrics. Thoughtful implementation of PROMs with institutional support and engagement of patients is necessary 6. The focus of this article is on the role of PROs in advancing toward value‐based breast cancer care in the context of locoregional therapy, with attention paid to the work done by the ICHOM in which a standard set of value‐based patient‐centered outcomes for breast cancer for international use is defined.

Value‐Based Breast Cancer Care

Refinement in surgical and radiation techniques allows for less invasive treatment strategies for breast cancer. The result is multiple locoregional therapy options that all result in excellent oncologic outcomes to discuss with patients 7, 8, 9, 10, 11. For patients with early stage breast cancer, 10‐year overall survival rates of greater than 90% have been shown for patients undergoing either breast‐conserving therapy or mastectomy 12. The result of these outstanding survival rates is a long‐term survivorship period, such that HRQOL becomes increasingly important. PRO data on long‐term HRQOL are thus needed to guide patients as they make decisions about treatment.

One challenge in implementing PROMs is defining which instruments are most applicable to patients with breast cancer. Another challenge is determining whether it is most beneficial to have a small number of generic instruments (i.e., broad‐based PROMs that can be used in a variety of patient populations) or more condition (or disease)–specific tools. This choice affects cost, personnel requirements, and complexity of implementation of a PRO program. We would assert that inclusion of PROMs that specifically assess local therapy outcomes in breast cancer is essential to detect differences in outcome and measure quality for patients with breast cancer.

Multiple societies and nonprofit organizations have committed to advancing current breast cancer care to a value‐based care system, including the American College of Surgeons (ACS) 13 and ICHOM. Incorporating meaningful PROMs to evaluate the care delivered is an important component of their efforts 2, 14, 15, 16, 17. The ACS is supporting collection of PROs through its Surgeon Specific Registry as a quality measure for the Centers for Medicare and Medicaid Services (CMS) Merit‐Based Incentive Payment System. ICHOM, together with health care institutions worldwide, has developed standard sets of outcome measures (including PROMs) for different medical conditions. The set proposed for patients with breast cancer, which includes locoregional‐specific outcomes, was one of the first developed 2.

The ICHOM breast cancer set focuses on meaningful outcomes to patients, with PROs composing 75% of the standard set 2. The PROs proposed (supplemental online Table 1) are applicable either to all patients with breast cancer or specific to the type of treatment administered. To evaluate the outcome of all patients, the European Organization for Research and Treatment of Cancer (EORTC) core 30‐question quality of life questionnaire (QLQ‐C30) and the EORTC breast cancer–specific 23‐question quality of life questionnaire (QLQ‐BR23) are proposed. To evaluate HRQOL related to surgery and/or radiation, the EORTC QLQ‐C30, EORTC QLQ‐BR23, and the BREAST‐Q “Satisfaction with breasts” domain are proposed. EORTC QLQ‐BR23, EORTC 21‐question liver metastases colorectal module quality of life questionnaire (QLQ‐LMC21), and the Functional Assessment of Cancer Therapy—Endocrine Subscale (FACT‐ES) are recommended for evaluation of patients receiving systemic therapy.

An important aspect of incorporating PROMs into outcomes assessment is to measure PROs longitudinally over the course of treatment. This allows for monitoring and detection of changes in physical or psychosocial function during the care cycle. Targeted supportive care may then be provided based on PRO scores, ultimately improving efficiency and patient satisfaction with care 18, 19, 20, 21. This has also been noted to improve physician satisfaction 22. Data suggest that collection of PROs may even improve overall survival 23. To most effectively affect patient care, information gained from utilization of these instruments is dependent on real‐time and accurate interpretation of the scores. Transitioning toward value‐based breast cancer care therefore depends on an increased understanding of PROMs, their validity to measure the outcomes intended, and their reference scores. Involving clinicians who treat breast cancer in the first initiatives to implement and evaluate the proposed PROs would generate much‐needed knowledge on their clinical utility.

ICHOM's PROs for Locoregional Therapy of Patients with Breast Cancer

As indicated above, ICHOM has recommended five PROMs for use in patients with breast cancer. The measures most relevant for those involved in the locoregional treatment of breast cancer are detailed below. Understanding of these PROMs and their history can guide the clinician in assessing their utility to his or her clinical practice.

EORTC QLQ‐C30 and BR23

The EORTC QLQ‐C30 is a generic oncologic questionnaire containing 30 questions with six single‐item scores, nine multiple‐item scales, three symptom scales, and an additional global health status/quality of life scale 24, 25, 26. It was developed and validated by the EORTC Quality of Life group between 1995 and 1997. Normative scores (obtained in a general population sample) and reference scores (i.e., scores obtained in patients with breast cancer) are available for all domains (Fig. 1) 26, 27, 28, 29, 30, 31, 32, 33, 34. Scores are based on the item response theory enabling comparisons between subjects and different timepoints within a subject. For the EORTC QLQ‐C30, minimally important differences are established 35, 36, 37, comparisons are made between paper‐based and electronic versions showing minor 38 or no differences 39, and working groups have been initiated for computerized adaptive testing 15, 40. The EORTC QLQ‐BR23 is a breast cancer–specific questionnaire by the EORTC Quality of Life group that contains 23 questions making up eight multiple‐item scales and is considered an addition to the EORTC QLQ‐C30 specifically for patients with breast cancer. In the ICHOM set, the EORTC QLQ‐C30 and the EORTC QLQ‐BR23 “Sexual functioning” and “Body image” domains are proposed as a PROM for all patients. The “Arm symptoms” and “Breast symptoms” of the EORTC QLQ‐BR23 are proposed to evaluate patients undergoing locoregional treatment.

Figure 1.

Mean scores with corresponding standard deviation of the European Organization for Research and Treatment of Cancer (EORTC) core 30‐question quality of life questionnaire (QLQ‐C30) domains in different populations. Scale (0–100) for “Pain” and “Fatigue”: higher scores represent more symptoms/poorer quality. Normative (EORTC Quality of Life group 27): general population sample (n = 7,802); countries involved were Germany, Norway, Austria, Denmark, U.S.; published July 2008. Reference (EORTC Quality of Life group 27): female patients with breast cancer (n = 2,702), published July 2008. Arraras et al. 28: study performed 2011–2014 in Spain, n = 243, mean age 52.2 ± 6.8 years, mean time since surgery 9.8 ± 4.0 years. Chu et al. 29: study performed 2007 in France, n = 187, mean age 58.4 ± 11.9 years, mean time since surgery 7.0 months (range 6–60 months). Høyer et al. 30: study performed 2007–2008 in Sweden, n = 1,076, mean age 61.8 ± 12.6 years, mean time since surgery 4.0 months (range 1–8 months). Xia et al. 31: year study performed unknown, China, n = 1,421 (MAS, n = 982; BCT, n = 439), mean age 56.9 ± 5.5 years, mean time since surgery unknown. Hadi et al. 32: year study performed unknown, Iran, n = 287 (MAS, n = 160; BCT, n = 127), mean age 48.9. ± 10.5 years and 47.0 ± 10.8 yeras, mean time since surgery 23 and 16 months (range 6–60 months), respectively, for patients with mastectomy and BCT. Moro‐Valdezata et al. 33: study performed 2003–2007 in Spain, n = 364, median scores only, mean age 59.0 ± 13.0 years, mean time since surgery 1 year. Tsai et al. 34: study performed 2011 in Taiwan, n = 544 (MAS, n = 327; BCT, n = 217), mean age 52.8 ± 9.4 years, mean time since surgery 1 year.Abbreviations: BCT, breast‐conserving therapy; MAS, mastectomy; PRO, patient‐reported outcome.

BREAST‐Q

The development of the BREAST‐Q is an excellent example of early adaption to a changing health care system. First published in 2009, the BREAST‐Q was designed 16 to evaluate PROs after breast surgery 41. It was developed with extensive qualitative input from patients and Rasch psychometric measures 16. Postoperative modules were later developed to evaluate outcomes after mastectomy without reconstruction 42 and breast‐conserving therapy 43, 44, 45. Normative scores are available based on the BREAST‐Q reconstructive module (Fig. 2) 43, 45, 46, 47, 48, 49, 50, 51, 52. Within the ICHOM set, the BREAST‐Q is proposed to measure “Satisfaction with breast” after breast cancer surgery. Multiple‐item scales are also available to evaluate “Psychosocial well‐being,” “Sexual well‐being,” “Physical well‐being,” and “Satisfaction with care” 16. Since its validation, the BREAST‐Q has been translated and linguistically validated in over 40 languages and used extensively for comparative effectiveness research worldwide 53, 54, 55, 56, 57. CMS has approved the BREAST‐Q as a quality performance measure.

Figure 2.

Mean scores with corresponding standard deviation of BREAST‐Q domains in different populations. Scale 0–100; higher scores represent higher quality/function. Normative (Mundy et al. 46): study performed 2017 in the U.S., general population sample (n = 1,200), mean age 54 ± 13 years. Zhong et al. 47: study performed 2009–2010 in Canada; n = 106 (delayed implant reconstruction, n = 76; immediate implant reconstruction, n = 30); age distribution, ≤49 years (68%), 50–59 years (28%), and ≥60 years (13%); median time since surgery, preoperative and 12 months postoperative. Rosson et al. 48: study performed 2008–2009 in the U.S., n = 170 preoperative patients (scheduled for immediate reconstruction, n = 117; delayed, n = 21; revision surgery, n = 32), mean age 50.3 ± 9.5 years. O'Connell et al. 43: study performed 2010–2016 in the United Kingdom, patients with breast‐conserving therapy only (n = 200), mean age 60.0 ± 11.1 years, time since surgery 35.5 ± 17.8 months. Barone et al. 49: study performed 2005–2016 in Italy, n = 144 (unilateral REC‐I, n = 121; bilateral REC‐I, n = 23), mean age 54 years, median time since surgery 16 months. Hwang et al. 50: study performed 2016 in the U.S., survey (n = 3,977) of patients with CPM (n = 1,598; CPM no REC, n = 328 vs. CPM with REC, n = 1,270) vs. no CPM (n = 2,379), mean age 57 ± 9.7 years, median time since surgery 4.6 ± 6.5 years. Koslow et al. 51: study performed 2000–2007 (CPM) and 2008–2010 (no CPM) in the U.S., n = 294 (CPM with REC‐I, n = 112 vs. no CPM, n = 182), mean age CPM 46.6 years (range 33–75 years) vs. no CPM 50.2 years (23–76 years), mean time since surgery CPM 51.9 months vs. no CPM 52.7 months. Ng et al. 52: study performed 2000–2010 in Australia, n = 143 (mastectomy, n = 79 vs. reconstruction, n = 69), mean age 54.5 ± 12.9 years, median time since surgery unknown.Abbreviations: bi, bilateral surgery; CPM, contralateral prophylactic mastectomy; del, delayed reconstruction; imm, immediate reconstruction; MAS, mastectomy; post, postoperative; pre, preoperative; PRO, patient‐reported outcome; REC, reconstruction; REC‐I, implant reconstruction; uni, unilateral surgery.

Review of Literature on PROs for Locoregional Therapy

The number of studies reporting on PROs in patients with breast cancer has increased over the last decade (Fig. 3), demonstrating increased recognition of the importance of these data. We conducted a search of the published literature for large (n > 100 patients), noninterventional, clinical studies (supplemental online File A) evaluating ICHOMs’ proposed PROs for patients with breast cancer undergoing locoregional treatment (i.e., the EORTC QLQ‐C30, EORTC QLQ‐BR23, and the BREAST‐Q) and identified 76 studies published from 2007 to 2018 (supplemental online File B). As shown in Figure 1 and Figure 4, the EORTC QLQ‐C30 and EORTC QLQ‐BR23 have been used extensively 26, 27, 28, 29, 30, 31, 32, 33, 34, 58. A selection of the studies that evaluate BREAST‐Q scores for reconstructive procedures 47, 48, 50, 51, 52, 59 or those comparing reconstructive procedures with mastectomy 42, 49 or breast‐conserving therapy 43 patients is shown in Figure 2.

Figure 3.

Overview of numbers of publications per year. Horizontal axes represent years 2007–2017. This figure includes published articles until December 31, 2017.

Figure 4.

Mean scores with corresponding standard deviation of the European Organization for Research and Treatment of Cancer (EORTC) breast cancer–specific 23‐question quality of life questionnaire (QLQ‐BR23) domains in different populations. Scale (0–100) for “Arm symptoms” and “Breast symptoms”: higher scores represent more symptoms/poorer quality. Reference (EORTC Quality of Life group 27): female patients with breast cancer (n = 2,702), published July 2008. Arraras et al. 28: study performed 2011–2014 in Spain, n = 243, mean age 52.2 ± 6.8 years, mean time since surgery 9.8 ± 4.0 years. Chu et al. 29: study performed in France, n = 187, mean age 58.4 ± 11.9 years, mean time since surgery 7.0 months (range 6–60 months). Xia et al. 31: year study performed unknown, China, n = 1,421 (MAS, n = 982; BCT, n = 439), mean age 56.9 ± 5.5 years, mean time since surgery unknown. Hadi et al. 32: year study performed not given, Iran, n = 287 (MAS, n = 160; BCT, n = 127), mean age 48.9. ± 10.5 years and 47.0 ± 10.8 years, mean time since surgery 23 and 16 months (range 6–60 months), respectively, for patients with mastectomy and BCT. Moro‐Valdezata et al. 33: study performed 2003–2007 in Spain, n = 364, median scores only, mean age 59.0 ± 13.0 years, mean time since surgery 1 year. Tsai et al. 34: study performed 2011 in Taiwan, n = 544 (MAS, n = 327; BCT, n = 217), mean age 52.8 ± 9.4 years, mean time since surgery 1 year.Abbreviations: BCT, breast‐conserving therapy; MAS, mastectomy; PRO, patient‐reported outcome.

Establishing reference scores represents an opportunity to advance PROs from a research setting toward a meaningful clinical setting, which is essential for VBHC initiatives. Although the current literature reflects some similarities in reported outcomes, most PROs in these studies were collected cross‐sectionally, hampering conclusion on causality and introducing bias. Therefore, use of these scores as a reference should be done with caution. The literature furthermore shows that in multivariate analyses, many clinical and sociodemographic variables are associated with PRO scores. For example, time since treatment 32, 34, 60, age 34, 61, 62, response shifts (changes in internal standards that may affect how one responds to questions about HRQOL over time), and/or recall biases 63, 64 were independently associated with PRO scores and frequently were more significant than clinical factors 30, 31, 32, 33, 34, 55, 65.

Ideally, PRO data would be collected in a prospective fashion with baseline assessment before surgery (or initiation of treatment) and with adjustment for demographics and clinically relevant covariables. Several studies have done this and provide high‐level evidence to guide treatment decision and patient counselling 34, 48, 56, 59, 66. A recent study of patients followed for 2 years after mastectomy with reconstruction revealed higher scores in satisfaction with breasts as well as psychosocial and sexual well‐being in patients who underwent autologous reconstruction versus implants 57. Prospective studies are important for establishing reference values and identifying potential areas for intervention in clinical care. Only a few trials have evaluated the utility of PROs as clinical tools 67. One trial reported improved communication by both physicians and patients when PROs were used to guide clinic visits 20. Another trial demonstrated an improvement in PRO scores when they were used to guide supportive care interventions 68. Longitudinal data collection is needed to define adequate reference values to determine the clinical utility of assessing PROs and providing patients with PRO data related to surgical outcomes early in the surgical decision‐making process for patients with breast cancer undergoing local regional therapy.

From Research Tool to the Clinical Arena, Advancing Toward Value‐Based Breast Cancer Care

Incorporation of the value‐based breast cancer care initiative into clinical practice or more broadly into a health care system requires generation of sufficient supporting data, which demands routine PRO collection. This process may significantly affect clinical workflow and requires careful consideration of timing and location of PROMs administration as well as the personnel needed to collect and interpret the data. Robust data collection methods and display tools for PROs are critical for successful implementation into clinical practice. A data collection tool should securely store data and ultimately be integrated into the electronic health record to facilitate the real‐time user‐friendly review of the data that has been shown to improve communication 19, symptom detection 69, and satisfaction of both patients and providers 18. Experience with electronic PRO collection tools is increasing 70, 71. Integration of these data collection tools in the electronic health record also facilitates aggregation of data such as treatment characteristics and associated costs for quality work and research.

Routine PRO collection will facilitate much‐needed insight into reference scores, minimally important differences in scores, and clinical and sociodemographic factors associated with PROs. Reference scores applicable to different international populations can be obtained only through sufficient case‐mixed analyses accounting for all the different factors associated with PRO scores. Large numbers of baseline and longitudinal PROs are needed to evaluate these models so that current reference scores can be adapted. International initiatives such as Patient‐Reported Outcomes Measurement Information System (PROMIS) 14 are excellent examples of ongoing efforts to gain insight in PRO reference scores.

Conclusion

Advancing toward value‐based breast cancer care will require a major transformation for the current health care system 6, and efforts are underway. Transformations of this scale take time and need champions to succeed. A transition period is needed in which health care providers can work toward advancing value‐based breast cancer care without disrupting current care and without financial consequences to establish what metrics best reflect quality clinical care. Only when PROs are shown to have adequate validity to evaluate certain outcomes in breast cancer care, such as those in the ICHOM standard set, should they be used as an indicator within the payment system 72. Involving relevant stakeholders, particularly patients, and engaging clinicians to champion this effort at this early stage is pivotal to define and validate disease‐specific sets of outcomes measures. This early involvement should also initiate parallel efforts to accurately capture data on direct and indirect costs made when providing care.

Clinicians treating breast cancer have an opportunity to act as early adopters of the routine collection of longitudinal PRO data. In this article, we focused on how we envision these efforts will facilitate conversations between patients and providers regarding locoregional therapy decisions and improve quality. It is hoped that this may serve as a framework for other multidisciplinary groups and guide implementation of PRO collection to improve quality of care across disease sites. These efforts are of significant importance within the broader multidisciplinary team, allowing for implementation of value‐based breast cancer care over the entire cycle of care, ideally improving all elements of the care delivered 18, 20, 69.

Author Contributions

Conception/design: Mirelle Lagendijk, Elizabeth Mittendorf, Laura S. Dominici

Collection and/or assembly of data: Mirelle Lagendijk, Laura S. Dominici

Data analysis and interpretation: Mirelle Lagendijk, Elizabeth Mittendorf, Tari A. King, Christopher Gibbons, Andrea Pusic, Laura S. Dominici

Manuscript writing: Mirelle Lagendijk, Elizabeth Mittendorf, Tari A. King, Christopher Gibbons, Andrea Pusic, Laura S. Dominici

Final approval of manuscript: Mirelle Lagendijk, Elizabeth Mittendorf, Tari A. King, Christopher Gibbons, Andrea Pusic, Laura S. Dominici

Disclosures

Tari A. King: Genomic Health (H); Andrea Pusic: BREAST‐Q (IP). The other authors indicated no financial relationships.

(C/A) Consulting/advisory relationship; (RF) Research funding; (E) Employment; (ET) Expert testimony; (H) Honoraria received; (OI) Ownership interests; (IP) Intellectual property rights/inventor/patent holder; (SAB) Scientific advisory board

Supporting information

See http://www.TheOncologist.com for supplemental material available online.

Supplemental Appendices

Supplemental Table