Treatment decisions and the employment of breast cancer patients: results of a population-based survey

Introduction

Work is an important source of income, insurance, and social interactions and may be particularly important for individuals with cancer, who may also find that it gives meaning to life, provides a welcome distraction, and improves quality of life.^1,2 Unfortunately, cancer diagnosis and treatment can disrupt patient employment, particularly during active therapy but also in its aftermath. Treatment plans are burdensome and exact a heavy toll on all aspects of quality of life--including physical functioning and emotional well-being--with protracted recovery times in some cases. Financial toxicity, which can develop in part because of lost income, is an important, yet understudied potential threat to patient and family quality of life after diagnosis.

Prior research regarding the impact of breast cancer diagnosis and treatment on employment experiences has yielded variable results, with some studies suggesting limited impact but others suggesting substantial and lasting effects.^3–7 The divergence of prior study results may be explained in part by differences in study settings and population characteristics, the wide variation in relevant policies and culture in different nations, and changes in treatments offered over time.

Our own prior work has shown that many patients with breast cancer are working for pay at time of diagnosis, that most women with breast cancer who are working for pay before diagnosis lose work time during treatment, and that many stop working altogether.^5,6 Furthermore, loss of paid work during treatment can result in permanent and undesirable long-term unemployment.⁷ Thus, it is critical that treatments be no more burdensome than necessary and delivered in ways that minimize disruption for patients.

The growing awareness of the burden of cancer treatment is sparking initiatives to reduce it. Use of chemotherapy in early stage breast cancer is increasingly more selective,⁸ and increased attention to symptom control and management may be reducing avoidable morbidity in those who do receive treatment.⁹ By contrast, trends in surgical management may be increasing patient morbidity--for example, the increasing use of bilateral mastectomy, usually with breast reconstruction, in patients with unilateral cancer.¹⁰ But at the same time there are trends toward less extensive surgery--for example, decreased re-excision¹¹ and use of axillary dissection after lumpectomy. Thus, there is a growing dichotomy in surgical management with major potential impact on patient recovery from treatment. It is essential that we understand how this rapidly evolving treatment context may impact employment of women diagnosed with breast cancer.

Policies regarding employment support for patients with cancer have also evolved in light of growing recognition of the importance of these issues,^12,13 further motivating the need to examine the impact of treatment on employment of patients diagnosed in the United States today. In this transformed landscape of public policy, medical evidence, and treatment options, we sought to document patterns and correlates of missed work in a contemporary population-based sample of women recently diagnosed with breast cancer, with particular focus on associations of employment experiences with primary surgical treatment selected, to inform initiatives to support patients with cancer in their treatment decisions and transitions to survivorship.

Methods

Results

As shown in Figure 1, of the 3930 women diagnosed in 2014–15 we initially selected for our sample based on rapid case ascertainment (which allows earlier survey administration by reducing the time lag from diagnosis to case identification),¹⁸ 258 were subsequently found to be ineligible due to prior breast cancer diagnosis or stage III-IV disease; residing outside the SEER registry area; or being deceased, too ill or unable to complete a survey in Spanish or English. Of 3672 eligible women remaining, 1170 could not be contacted or did not participate, leaving 2502 respondents (68%). Of these, we considered the 1006 women who reported that they had been working before diagnosis for further analysis in this study.

Table 1 shows the characteristics of the analytic sample, which was racially and ethnically diverse (48% white, 19% black, 20% Latina, and 11% Asian). Most patients (62%) received lumpectomy; 16% had unilateral mastectomy (8% with reconstruction); 23% had bilateral mastectomy (19% with reconstruction). One third (33%) received chemotherapy. The vast majority (84%) reported working full time before diagnosis, but only half (50%) had jobs that allowed for paid sick leave, 39% had disability benefits, and 38% had a flexible work schedule.

Bivariable associations between employment experiences and patient characteristics are provided in Supplementary Table 1. On multivariable analysis, including treatment and clinical factors alone, several factors were significantly correlated with missing more than a month of work or stopping work altogether as compared to missing up to 30 days (Table 2). Those with poorer health (vs. excellent health) were overall less likely (p<0.001) to miss work (OR for stopping work altogether 2.5, 95% CI 1.2–5.1). Chemotherapy receipt also correlated with stopping work (OR 1.3, 95% CI 0.8–2 for missing over one month and OR 3.9, 95% CI 2.6–5.8 for stopping work altogether). Surgical treatment was strongly correlated with missing over one month of work (OR 7.8, 95% CI 4.5–13.4, for bilateral mastectomy with reconstruction, compared to lumpectomy) and with stopping work altogether (OR 3.1, 95% CI 1.6–5.9, for bilateral mastectomy with reconstruction, compared to lumpectomy). Race was correlated with missed work (p=0.01). For blacks versus whites, the OR for missing over one month was 2.0 (95% CI 1.3–3.2) and for stopping work altogether was 1.7 (95% CI 1.1–2.8). Those with paid sick leave were less likely to stop working altogether (OR 0.5, 95% CI 0.3–0.7). Those with flexible work schedule were less likely to stop working altogether (OR 0.3, 95% CI 0.2–0.5) or to miss more than a month of work (OR 0.7, 95% CI 0.5–1). Conversely, women with disability benefits were more likely to stop working (OR 1.6, 95% CI 1–2.4) or miss over a month of work (OR 2.7, 95% CI 1.8–3.9). Also significant were study site (with patients from Georgia less likely to miss over a month of work, OR 0.6, 95% CI 0.4–0.8), household income (with the highest income group with income ≥$90,000 having an OR of 0.6, 95% CI 0.3–1, for stopping work), and number supported by family income (with those whose household income supported ≥4 persons being less likely to stop working altogether, OR 0.4, 95% CI 0.2–0.8). Of note, 7% of patients (13% of those receiving radiotherapy) were still receiving radiotherapy at the time of survey; excluding these patients did not affect the significance of any covariates in the model.

Figure 2 shows adjusted rates of missed work, by surgical treatment received. Patients receiving lumpectomy were far less likely to miss over a month of work or stop working altogether, as compared to women receiving mastectomy.

Those who missed more work also reported losing greater amounts of income due to time off from work since breast cancer diagnosis (p<0.001), as shown in Table 3. Specifically, among those who missed 0–30 days, 74% lost $0–$500 and only 6% lost >$5000. Among those who missed more than 30 days, 40% lost 0–$500 and 29% lost >$5000. Among those who stopped working altogether, 17% lost 0–$500 and 54% lost >$5000. The median patient reported losing 3.6% of their annual household income due to time off from work, and 19% of patients reported losing 10% or more of their annual household income

Of the women in our analytic sample, all of whom had been employed before diagnosis, 65% reported that their current employment status at time of survey was full-time employment, 15% reported part-time employment (including 38 of the 844 women who had been working full-time before diagnosis), 3% were unemployed and looking for work, 6% were temporarily laid off or on sick or other leave, 4% were disabled, 4% were retired, 1% were students, and 2% were homemakers.

Discussion

In this large, modern and diverse cohort of patients newly diagnosed with breast cancer, we observed striking variations in rates of missed work by type of surgery received, along with findings consistent with prior research regarding the impact of chemotherapy, sociodemographic factors, and the employment context. These findings are important because never before have women with breast cancer faced such a wide range of choices for surgical management, nor has the dichotomy in surgical treatment options been more dramatic. Some women receive breast conservation, while others receive bilateral mastectomy for exactly the same condition, often also with reconstruction. Understanding the employment effects of different surgical decisions is critically important to the many patients who consider more aggressive surgical treatments than medically necessary to treat their cancer. Surgeons who treat patients with breast cancer can now provide compelling evidence that women who receive mastectomy experience considerably higher risks of missed work than those who receive breast conserving therapy. Moreover, the current data allow quantification of the financial impact of this missed work—data that may be very useful in helping patients understand the full impact of treatment decisions.

Strikingly, the magnitude of risk to employment with more aggressive surgery observed in this study was similar to the risks associated with chemotherapy, which has traditionally been the major target of efforts to reduce the burden of cancer care. Indeed, most prior studies of immediate treatment impact on breast cancer patients’ employment^19,20 have focused on chemotherapy. Several studies have suggested that patients who received chemotherapy were most likely to experience disruptions in employment²¹ and prolonged absences.²² Although evidence has been more mixed regarding the long-term effects of treatment on employment, with some studies suggesting that breast cancer treatment and particularly adjuvant chemotherapy might not delay or prevent ultimate return to work,^3,4 there is reason to believe that the adverse effects of chemotherapy on employment may be long lasting. In our group’s prior work, adjuvant chemotherapy receipt was associated with long-term job loss among survivors at four years, and many of these women were actively seeking employment, suggesting that this was involuntary.⁷ Women who lacked employment support (sick leave or flexible hours) were most vulnerable. Studies in other settings, including ones with greater employment support, have also documented greater rates of job discontinuation or decreased work time among breast cancer survivors who received chemotherapy.²³ Moreover, recent research has highlighted how certain women, including those with low income, may be particularly vulnerable to the risk of not returning to work in the months and years after treatment.²⁴

As Hassett and colleagues noted in relevant prior work, these findings “reinforce the need to assess the impact of treatments, especially new treatments, on patient-centered outcomes such as employment.” Of note, at the time of most prior studies, rates of mastectomy overall were considerably lower than in the current era, and bilateral mastectomy was rarely used, so it was not evaluated separately from unilateral mastectomy in terms of impact on employment. However, in recent years, in the wake of celebrity disclosures and growing patient interest, rates of mastectomy overall and particularly in combination with contralateral prophylactic mastectomy have surged: more than one in five patients in the current sample of working patients had bilateral mastectomy. Although some women with early-stage breast cancer are not candidates for breast conservation, most are. Therefore, it is crucial to ensure that patients are fully informed of the risks of treatment, including the potential for impact on employment (a critical component of financial toxicity) to optimize the true goals of shared decision-making. With the growing use of mastectomy, further research is necessary to monitor whether the short-term impact of more aggressive surgery that was observed in the current study will also translate into longer term consequences for these women’s employment and well-being.

Prior research has emphasized the importance of workplace accommodations in promoting return to work.^25,26 In our current study, flexible work arrangements were associated with substantial decreases in missing over 30 days of work or stopping altogether, although disability benefits were found to have the converse association, suggesting that some missed work may reflect the ability of a patient to take the time she needs to recover. Nevertheless, even after accounting for flexibility and other workplace policies, treatments, and particularly more aggressive surgery, had a strong effect that merits note.

Although our study has numerous strengths, including its large and recent sample drawn from population-based registries, it also has limitations that merit consideration. First, as in any observational study, correlation may not imply causation. Still, there is little reason to believe that those who selected more aggressive treatments were predisposed to missing or stopping work, after adjusting for multiple sociodemographic and employment factors. Second, not all missed work is necessarily concerning; voluntary time off might benefit patients by giving them a chance to cope with diagnosis and treatment. Further research is necessary to determine whether the short-term impact we observed translates into long-term challenges, particularly among the youngest patients, with the greatest years of potential employability, who most often selected the most aggressive surgical options. Third, this study was intended to assess associations between treatment and employment outcomes; therefore, it included only patients diagnosed with breast cancer and not non-cancer patients from the population. Although studies evaluating the employment effects of cancer diagnosis in patients as compared to healthy controls are important, the inclusion of healthy controls was not necessary to study the treatment effects that we sought to evaluate here. Fourth, because patients were surveyed relatively after diagnosis to minimize recall bias, a minority were still completing adjuvant therapy, and the full impact of such treatments (and particularly radiotherapy) might not be appreciated given the timing of survey administration. Fifth, to minimize respondent burden in the context of a larger study evaluating breast cancer treatment decisions, only select employment-related factors were evaluated. We hope to conduct follow-up research with this cohort as they proceed further into the survivorship phase, which will allow us to capture long-term, detailed measures of employment-related constructs of importance to patients. Finally, our study was conducted in two large U.S. areas; the results should be generalized with caution to other U.S. settings and not at all to countries with markedly dissimilar employment support policies or culture.

Supplementary Material