Factors Associated With Follow-Up Care Among Women With Early-Stage Breast Cancer

Abstract

PURPOSE:

Follow-up guidelines vary widely among national organizations for patients with early-stage breast cancer treated with curative intent. We sought to evaluate the patterns and predictors of provider follow-up care within the first 5 years after diagnosis.

METHODS:

Using the SEER-Medicare linked data set, we evaluated patients who were diagnosed with stage I and II breast cancer who underwent breast-conserving surgery from 2002 to 2007 with follow-up until 2012. We defined discontinuation of follow-up as > 12 months from the previous physician visit without a visit claim from either a surgeon, medical oncologist, or radiation oncologist. We performed a multivariable logistic regression and Cox proportional hazards regression analysis to determine factors associated with the discontinuation of follow-up care.

RESULTS:

Of the 30,053 patients enrolled in our initial cohort, 25,781 (85.8%) saw a medical oncologist and 21,612 (71.9%) saw a radiation oncologist in the first year in addition to a surgeon. Over the 5 years, 6,302 patients (21.0%) discontinued follow-up visits. Discontinuation of physician visits increased with increasing age. Women with stage II cancer (v stage I) were less likely to discontinue follow-up visits (odds ratio, 0.78; 95% CI, 0.73 to 0.83). Time to early discontinuation was greater for patients with hormone receptor–negative tumors (hazard ratio, 1.14; 95% CI, 1.05 to 1.24). Women who were diagnosed more recently were less likely to discontinue seeing any physician.

CONCLUSION:

Twenty-one percent of patients with early-stage breast cancer discontinued seeing any oncology provider over the 5 years after diagnosis. Coordination of follow-up care between oncology specialists may reduce discontinuation rates and increase clinical efficiency.

INTRODUCTION

In patients with early-stage breast cancer treated with curative intent, follow-up is recommended with the purpose of early detection of local or systemic relapse as well as that of a second primary tumor.¹ Wide variation exists in the recommended frequency and duration of follow-up visits, as delineated in guidelines put forth by national and international organizations. For example, ASCO recommends clinical examination every 3 to 6 months for 3 years, every 6 to 12 months for the next 2 years, and then annually.² National Comprehensive Cancer Network guidelines liberalized their recommendations in 2015 to allow more clinician interpretation and flexibility for patient follow-up, but before this recommended interval evaluation every 4 to 6 months for 5 years.³ These guidelines are primarily geared toward medical oncologists and little direction is given to other providers who are also often involved in follow-up care. Whereas it is clear that some sort of follow-up is needed, the optimal frequency of follow-up and which provider the patient should see is not well defined.

It is known that intensive follow-up strategies, such as laboratory testing and imaging at regular intervals, results in higher costs without earlier detection of relapse, improvement in survival, or better quality of life for patients compared with routine clinician visits.^4-7 However, there is currently no evidence that routine clinical examination is of clinical benefit, and this approach may actually be an inefficient way to detect recurrent disease.⁸ Important aims of follow-up care include the detection of early local recurrences or contralateral breast cancer, evaluation of therapy-related complications, and encouraging adherence to hormonal therapy.⁹ The coordination of care and reduction in redundancy of care in the breast cancer survivorship population have become issues of national importance to improve overall quality of care and reduce costs.^10-12

Patients with early-stage breast cancer are often managed by a team of providers, including surgeons, medical oncologists, and radiation oncologists. Follow-up care guidelines for these patients are not evidence based and have an unknown effect on cancer outcomes. The aims of the current study were to use a large population-based database to evaluate the patterns and predictors of provider follow-up care within the first 5 years after diagnosis and determine if follow-up is concordant with current guidelines.

METHODS

Data Source

We analyzed data from the SEER-Medicare database.¹³ SEER contains tumor registry data for 28% of the US population and includes pathologic, staging, treatment, demographic, and survival information.¹⁴ SEER is linked with the Medicare database, which contains claims for covered health care services for inpatients and outpatients, diagnoses, and billed services from the time of a person’s Medicare eligibility until death.¹⁵

Cohort Selection

We identified all women age ≥ 65 years who were diagnosed with stage I and II breast cancer who underwent breast-conserving surgery within 6 months of diagnosis, between January 1, 2002, and December 31, 2007, and were observed through December 31, 2012. Surgery was determined by examination of current procedural terminology (CPT) procedure codes with the corresponding date of the procedure. Only patients who were alive for 5 years after diagnosis were included. Stage III patients were excluded because of the high rate of cancer recurrence in this population. Patients who underwent mastectomy were excluded as they are not routinely seen by a radiation oncologist. We excluded patients who were not covered by Medicare Parts A and B or who were enrolled in health maintenance organizations from 12 months before the date of diagnosis to 5 years after the date of diagnosis. We also required that breast cancer be the patient’s first cancer diagnosis. If a patient developed a second malignancy less than 6 months after breast cancer diagnosis, the patient was excluded.

Clinical and Demographic Characteristics

We evaluated the following variables: age at diagnosis (65 to 69, 70 to 74, 75 to 79, and ≥ 80 years), race/ethnicity (black, white, Hispanic, other), marital status (unmarried, married, unknown), tumor stage, area of residence (urban, rural), tumor grade (low, high, unknown), hormone receptor status (positive, negative, unknown), and year of diagnosis. We used the Klabunde adaptation of the Charlson comorbidity index, which yields an overall comorbidity score on the basis of 15 comorbid disease categories, calculated from International Classification of Diseases, 9th Revision, codes, Healthcare Common Procedure Coding System codes, and surgery codes in the Medicare database and expressed as a score of 0, 1, or ≥ 2.¹⁶ Socioeconomic status was evaluated from the SEER-Medicare database using ZIP codes and median annual household income to create an aggregate socioeconomic score that was then stratified into quintiles.¹⁷

Primary End Point: Discontinuation of Follow-Up

We defined discontinuation of follow-up care as > 12 months from the previous physician visit without a visit claim from either a surgeon, medical oncologist, or radiation oncologist. Physician visits were identified using Medicare National Claims History and CPT codes that included physician claims for inpatient and outpatient visits. The Outpatient Standard Analytic File database was also used to gather office visit data and includes a CPT procedure code with accompanying date. Surgeon visits were defined as a visit to the same surgeon who performed the original lumpectomy. Medical oncology and hematology/oncology were included to define a medical oncology visit. Radiation oncologists were identified as physicians who were listed with a specialty of radiation oncology or diagnostic radiologist. As described previously,¹⁸ not all patients who receive radiation therapy are captured under a radiation oncologist and we considered only physicians who had Medicare claims with a radiation therapy CPT code as practicing radiation oncology.

Statistical Analysis

We used multivariable logistic regression models to analyze the association between patient demographic and tumor-specific factors and discontinuation of follow-up care over the 5 years after diagnosis. Results are reported as odds ratios (ORs) with 95% CIs, with an OR > 1 indicating an increased likelihood of discontinuation. We developed a Cox proportional hazards regression model to determine the effect of each variable on the time to early discontinuation of physician follow-up visits. Patients who initiated chemotherapy > 1 year after diagnosis or patients with a diagnosis of a later stage of breast cancer > 6 months after diagnosis were censored because of a concern about recurrence of cancer and having a higher risk tumor. Those patients who developed a second malignancy > 6 months after their breast cancer diagnosis were also censored. We estimated the probability of early discontinuation at various time points using the Kaplan-Meier method and compared early discontinuation between various groups. All analyses were performed using SAS (SAS/STAT User’s Guide, Version 9.4; SAS Institute, Cary, NC ). All statistical tests were two sided, with an α of .05.

RESULTS

We identified 30,053 women age ≥ 65 years who were diagnosed with stage I and II breast cancer who underwent breast-conserving surgery within 6 months of diagnosis. In addition to the surgeon, 25,781 patients (85.8%) saw a medical oncologist and 21,612 (71.9%) saw a radiation oncologist in the first year. Sixty-six percent of patients visited all three providers over the first year. The characteristics of patients who saw a medical oncologist were similar to those of patients who did not. The mean number of total physician visits for years 2 to 5 were 4.2, 3.1, 2.5, and 2.1, respectively. The overall mean number of surgeon visits was 2.1, mean number of oncology visits was 9.0, and mean number of radiation oncology visits was 2.2 over years 2 to 5.

Over the five years, 6,302 women (21.0%) discontinued follow-up visits—that is, went > 12 months without seeing any provider. The number of patients who saw only a medical oncologist increased each year—22.8% in year 2 to 32.7% in year 5—whereas radiation oncology and surgeon visits decreased. Visits to a radiation oncologist only decreased from 2.4% in year 2 to 1.5% in year 5. The percentage of patients with a visit only to a surgeon decreased from 9.4% to 5.4% over years 2 to 5. Patients were less likely to see all three physicians annually, over time—15.1% in year 2 and 2.9% in year 5 (Appendix Fig A1, online only).

We evaluated patients who visited a provider two or more times per year as that is the minimum recommended visit frequency.³ For each oncology specialist, the proportion of patients with two or more physician visits decreased annually. The percentage of patients who saw a surgeon at least biannually decreased from 27.0% in year 2 to 7.4% in year 5. Similarly, the number of patients who saw a radiation oncologist at least biannually decreased from 22.6% in year 2 to 4.2% in year 5. Each year, medical oncologist visits were notably higher than those of the other two providers. In year 2, 65.5% of patients saw a medical oncologist at least biannually, whereas in year 5, 44.1% saw the same provider (Appendix Fig A1). After year 2, the majority of patients saw any oncology provider fewer than two times per year. However, there was a subset of patients who saw an oncology specialist five or more times over years 2 to 5 after diagnosis (Fig 1). By year 5 after diagnosis, the majority of patients had fewer than two physician visits with an oncology provider.

Fig 1.

Percentage of patients with a visit to an oncology provider per year.

Factors Associated With Early Discontinuation

Demographic variables and tumor characteristics associated with early discontinuation are shown in Table 1. In a multivariable model evaluating factors that resulted in any discontinuation over the 5-year period, patients were more likely to discontinue physician visits with increasing age. Conversely, women with higher-stage cancer were less likely to discontinue follow-up (OR, 0.78; 95% CI, 0.73 to 0.83). Patients with low-grade tumor were more likely to discontinue follow-up compared with those with high-grade cancer (OR, 1.09; 95% CI, 1.02 to 1.18). Women who were diagnosed in later years were less likely to discontinue seeing any of the three physicians (OR, 0.97; 95% CI, 0.95 to 0.98). Race, socioeconomic status, and marital status were not associated with discontinuation of physician follow-up. A proportional hazards model evaluating time to discontinuation demonstrated similar trends with regard to the association between age, diagnosis year, stage, and grade with early discontinuation. Patients with hormone receptor–negative breast cancer were more likely to discontinue follow-up (HR, 1.14; 95% CI, 1.05 to 1.24).

TABLE 1.

Multivariable Analysis of Factors Associated With Discontinuation of Follow-Up Care

Time to Early Discontinuation

Figure 2 shows Kaplan-Meier curves of time to early discontinuation among patients who saw each provider the first year after diagnosis. Among those patients who were seen by medical oncology the first year after diagnosis, we evaluated the time to early discontinuation from medical oncology. Patients who saw all three providers the first year (top curve) had the lowest likelihood of discontinuation with any provider compared with the other groups. Compared with those patients who saw medical oncology or all three providers the first year, patients with a visit to radiation oncology or a surgeon had a higher probability of having a 12-month gap in seeing that provider at the end of 5 years.

Fig 2.

Time to early discontinuation of follow-up visits by specialty.

DISCUSSION

In this analysis of patients age ≥ 65 years with early-stage breast cancer who underwent lumpectomy, we found that 21.0% of women discontinued follow-up visits with either medical oncology, radiation oncology, and surgery in the first 5 years after diagnosis. Discontinuation of clinical follow-up decreased with each subsequent year of diagnosis. In addition, older patients and those with hormone receptor–negative cancer and low-grade tumors were more likely to discontinue follow-up visits. Patients with higher-stage disease were less likely to discontinue seeing a physician.

Whereas numerous guidelines exist for the follow-up of patients with breast cancer after initial treatment, there is vast heterogeneity among them, and evidence in the literature is lacking as to the best approach. We found that patients were more likely to follow-up with a medical oncologist compared with a radiation oncologist or surgeon as time progressed. In a SEER-Medicare analysis of patients with early-stage breast cancer, Neuman and colleagues¹⁹ showed that medical oncologists are more likely to observe patients who would be eligible for systemic therapy, but patients who are poor candidates for systemic therapy—that is, because of age or comorbidities—are instead more likely to receive follow-up care by a surgeon, radiation oncologist, and/or primary care provider. By delving into why different types of oncologists participate in follow-up care and their perceived roles, Neuman et al found that nonmedical oncologists were more selective in the patients they observed, focusing on those who would benefit most from their specific skillset of locoregional assessment.²⁰ Our findings are consistent with prior work that demonstrated that, nationally, surgeons provide less breast cancer survivorship care compared with medical oncologists and primary care providers.²¹

There is a paucity of data that evaluate factors that are associated with the early discontinuation of follow-up care in patients with cancer. In the primary care setting, Goldman et al²² demonstrated that certain variables had independently significant associations with future appointment-keeping behavior, including the patient’s age, race, and physician-cited psychological or social problems. We found that discontinuation of follow-up visits increased as age increased. In a prior SEER-Medicare study of patients with early-stage breast cancer, younger age was shown to be associated with more frequent follow-up visits with an oncology specialist.²³ Older survivors of breast cancer represent a potentially vulnerable population, as they may have functional limitations²⁴ that could limit access to suitable health care and appropriate follow-up may help reduce adverse outcomes.

The primary reason for follow-up care in patients with early-stage breast cancer who have undergone curative treatment is to detect a locoregional or distant recurrence of the tumor. Prior studies have established that intensive testing is not recommended as it does not improve the chances of survival and increases cost.^4,25 Since the 1990s, ASCO has published recommendations against the routine assessment of tumor markers for the detection of recurrence of disease.^26-28 Although there is no evidence from randomized controlled trials that the earlier detection or treatment of asymptomatic, metastatic breast cancer recurrence improves clinical outcomes, some suggest that patients with early and limited metastatic recurrence may be cured.^29,30 Follow-up with an oncology provider and clinical examination per guidelines may be useful to detect recurrences early, but whether this improves survival outcomes has not been proven.

Not all patients have an equal risk of developing locoregional recurrences. Those with high-risk characteristics, such as larger tumors or more lymph node involvement, correspond to higher locoregional recurrence rates.³¹ Despite having a higher rate of recurrence and a more aggressive clinical course,³² we found that patients with hormone receptor–negative tumors who are ineligible for hormone therapy are more likely to discontinue follow-up (HR, 1.14; 95% CI, 1.05 to 1.24). An important aspect of follow-up is monitoring for adverse effects and adherence to endocrine therapy, which is recommended for at least 5 years.³³ Adverse effects from hormonal treatment are common and can frequently result in nonadherence to therapy.^34,35 Future studies may assess the association between loss of follow-up and hormone therapy adherence. In a multivariable analysis of patients with hormone receptor–positive disease, we found that similar factors were associated with discontinuation of follow-up care compared with our overall cohort. In our patient cohort, by year 5, the majority of patients had more than two visits with a medical oncologist compared with a surgeon or radiation oncologist. Unlike patients with hormone receptor–positive breast cancer, those patients who are ineligible for hormone therapy may not be motivated to follow-up with medical oncology or other oncology providers.

There are several potential adverse effects of frequent follow-up of patients with early-stage breast cancer. This includes possible financial burdens of traveling to appointments and missed work. Quality of life, including psychological distress and anxiety with more intensive follow-up, is also a concern; however, in one randomized controlled trial that compared intensive with standard surveillance follow-up strategies in patients with early-stage breast cancer, no difference in health-related quality of life was detected between the two groups.⁶ In our patient cohort, a subset of women saw providers more than two times each year in the years after diagnosis. Whereas the majority of patients visited medical oncology, there was a percentage of patients who saw radiation oncology and surgery more frequently than what is recommended by guidelines. By year 5, approximately 11% of patients were seeing an oncology specialist five or more times. Given the limitations of the SEER-Medicare data set, we are unable to delineate the reasons behind frequent follow-up. Improved definition and guidance regarding a particular provider’s role during follow-up is needed to reduce redundancy and improve the effectiveness of follow-up care.

Although the SEER-Medicare registry data are comprehensive, this study has a few limitations. Our study includes a restricted patient sample, specifically, nondisabled patients older than age 65 years. Whereas findings cannot be generalized to a younger population, more than 50% of patients with breast cancer are older than age 65 years³⁶; therefore, our results are relevant to the majority of women. We are also unable to determine the reason for discontinuation of follow-up. It is difficult to distinguish whether patients were not advised to follow-up with other providers or whether they discontinued follow-up of their own volition. As such, there may be reasons for discontinuation that are not captured with claims data. Prior research has demonstrated that follow-up with a primary care physician who has been provided guidelines on follow-up results in similar outcomes.^37,38 Current guidelines recommend follow-up care with an oncology specialist, and our goal was to assess compliance with those guidelines. Future studies should better ascertain how primary care physicians and midlevel providers can participate in the follow-up care of patients with breast cancer as trends in practice continue to evolve.

In summary, significant variation exists in the follow-up care of older patients with breast cancer treated with curative intent. Current practice guidelines are directed toward oncology specialists and suggest frequent follow-up with a provider but are not explicit in their recommendations with regard to the level of care that should be provided during follow-up visits. Coordination of follow-up care between oncology specialists and other providers may reduce discontinuation rates as well as the redundancy of visits, thereby increasing clinical efficiency. Identifying patients who are at risk for early discontinuation of follow-up will eventually allow for the promotion of public health initiatives to improve access to care. More research is needed to determine the optimal follow-up for maintaining adherence to therapy, reducing over-testing, and decreasing cost.

APPENDIX

Fig A1.

Percentage of physician visits by specialty (left) and percentage of patients with two or more physician visits per year (right). All MDs, medical oncologist, radiation oncologist, and surgeon; MO, medical oncologist; RO, radiation oncologist.

AUTHOR CONTRIBUTIONS

Conception and design Farah F. Quyyumi, Dawn L. Hershman

Financial support: Dawn L. Hershman

Administrative support: Grace C. Hillyer

Provision of study materials or patients: Dawn L. Hershman

Collection and assembly of data: Donna Buono, Cynthia W. Law

Data analysis and interpretation: Farah F. Quyyumi, Dawn L. Hershman, Jason D. Wright, Melissa K. Accordino, Donna Buono, Alfred I. Neugut

Manuscript writing: All authors

Final approval of manuscript: All authors

Accountable for all aspects of the work: All authors

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

Factors Associated With Follow-Up Care Among Women With Early-Stage Breast Cancer

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/jop/site/ifc/journal-policies.html.

Jason D. Wright

Consulting or Advisory Role: Clovis Oncology, Tesaro

Melissa K. Accordino

Honoraria: Sermo, Sermo (I), M3

Alfred I. Neugut

Stock and Other Ownership Interests: Stemline Therapeutics

Consulting or Advisory Role: Pfizer, Otsuka, United Biosource Corporation, EHE International

Expert Testimony: Hospira

No other potential conflicts of interest were reported.