Health-related quality of life and work productivity in UK patients with HER2-positive breast cancer: a cross-sectional study evaluating the relationships between disease and treatment stage

This was a cross-sectional, observational study conducted between December 2016 and March 2017. Patients were recruited to attain relatively balanced groups in terms of sample size (planned sample size for each group n = 100) who were each representative of the general population of UK patients with BC in terms of treatment at the relevant stage. Group 1 comprised patients currently undergoing treatment for early BC (patients were selected to ensure adequate representation of patients treated with chemotherapy and targeted HER2 therapy [planned n = 40/100] and those treated with targeted HER2 therapy alone); Group 2 comprised patients with early BC who had completed treatment and were in remission (i.e. no longer receiving loco-regional treatment, chemotherapy or targeted HER2 therapy; patients may still have been receiving hormone therapy); Group 3 comprised patients receiving treatment for metastatic BC (patients were selected to ensure adequate representation of patients receiving first-line treatment [planned n = 50/100] and those receiving second/subsequent lines of treatment for metastatic BC). Two-hundred and ninety-nine patients with HER2-positive BC were enrolled into the study through physician referral using a mixed approach to recruitment (e.g. invited to participate at routine clinic appointments, telephone appointments or via a letter from their physician in order to reach those patients who did not require frequent clinic appointments) from 14 secondary and tertiary care centres across England, UK. Patients were eligible for enrolment at each site if they were aged ≥ 18 years at study start and had been diagnosed with early stage (stage I–III) or metastatic (stage IV) HER2-positive BC (confirmed by the study site physician). HER2-positive was defined as immunohistochemistry-positive and/or in situ hybridisation ≥ 2.0. Patients were excluded if they were unwilling or unable to consent, unable to complete HRQoL questionnaires, or had an Eastern Cooperative Oncology Group Performance Status ≥ 3 [15].

Patients enrolled into the study completed questionnaires, either at the clinic or at home, on socio-demographic characteristics, including education level and work status, as well as several surveys. Surveys completed included the work productivity and activity impairment (WPAI) survey, the EuroQol-5 Dimensions-5 levels (EQ-5D-5L) questionnaire (dimensions: mobility, self-care, usual activities, pain/discomfort, anxiety/depression; visual analogue scale [VAS] score; index values), and the Functional Assessment of Cancer Therapy-Breast (FACT-B) questionnaire (subscale scores: physical wellbeing, functional wellbeing, social wellbeing, emotional wellbeing, BC-specific symptoms; FACT-B total score; FACT-General score; Trial Outcome Index). Clinical staff collected data from participating patients’ medical records related to demographics, medical history, disease and treatment history, and current treatments.

The study was performed in accordance with the Declaration of Helsinki and UK ethical approval from the Health Research Authority Research Ethics Committee (approval 16/EE/0429) was in place prior to study commencement.

Patient groups did not differ significantly with regard to age, level of education, presence of comorbidities or hormone receptor positivity at study enrolment (Table 1 and Additional file 1: Table S1). Mean age of patients was 55.0 (SD: 11.1) years, 57.7 (10.6) years and 55.3 (11.2) years, in Groups 1, 2 and 3, respectively. Median time since diagnosis of early BC was 9.0 months (interquartile range [IQR]: 6.0 months), 45.0 months (32.0) and 79.5 months (82.0) in Groups 1, 2 and 3, respectively; and the median time since diagnosis of metastatic BC in Group 3 was 30.0 (37.0) months (Table 1). In Group 3, over a quarter of patients (26.5%) had de novo metastatic cancer; over three-quarters (75.5%) had visceral metastases and approximately a quarter (24.5%) had central nervous system (CNS) metastases. The treatment status of patients at study enrolment is summarised in Additional file 1: Table S2. Group 2 patients had completed their adjuvant therapy a median (IQR) of 27.5 (30.8) months prior to the study.

Patient-reported employment status significantly differed between groups (χ²(1) = 39.45, p < 0.001; Table 2). Significantly more patients in Group 3, and significantly fewer patients in Group 2 reported inability to work (post hoc p < 0.003 for each comparison, significant after Bonferroni correction) than expected under the assumption of independence. Conversely, significantly fewer patients in Group 3 (post hoc p < 0.001), and marginally more patients in Group 2 (post hoc p = 0.03, not significant after Bonferroni correction) reported being in employment. For patients in part-time employment, the number of part-time hours worked per week reported by patients differed significantly between groups (F(2, 50) = 5.64; p = 0.006); patients in Group 3 reported working significantly fewer hours compared to those in Group 2 (post hoc p = 0.002, significant after Bonferroni correction; Table 2).

Groups differed significantly on the WPAI subscales of absenteeism (F(2, 71.218) = 6.23; p = 0.003), overall work impairment (F(2, 80.107) = 4.04; p = 0.021), and activity impairment (F(2, 277.666) = 13.81; p < 0.001), but not on presenteeism (F(2, 97) = 0.25; p = 0.781) (Fig. 1, Additional file 1: Table S3). In particular, employed patients in Group 2 reported a significantly lower proportion of absenteeism than those in Group 1 (post hoc p < 0.001), and a significantly lower proportion of overall work impairment compared to those in Group 1 (post hoc p = 0.015). Conversely, considering both employed and non-employed patients, patients in Group 3 reported a significantly higher proportion of activity impairment than those in Groups 1 (post hoc p = 0.004) and 2 (post hoc p < 0.001), which did not differ significantly from one another.

Higher activity impairment and overall work impairment were significantly associated with poorer HRQoL, as assessed by EQ-5D-5L VAS, and FACT-B total and FACT-G scores (all p < 0.001), and the strongest associations were observed for activity impairment (Additional file 1: Table S6).

The correlations between FACT-B domains and absenteeism and presenteeism are presented in Additional file 1: Table S7. Poorer physical and functional wellbeing and breast cancer-specific symptoms were significantly associated with higher impairment in work productivity as measured by levels of absenteeism and presenteeism (Pearson’s correlations p < 0.001); while social wellbeing was not observed to be associated with either domain of work productivity. In a multiple linear regression model, physical wellbeing, functional wellbeing and breast cancer specific symptoms significantly predicted absenteeism (p < 0.001), and collectively explained 24% of the variance in absenteeism (Additional file 1: Table S8). Of these predictors, only functional wellbeing was a significant independent predictor. Physical wellbeing, emotional wellbeing, functional wellbeing and breast cancer specific symptoms significantly predicted presenteeism (p < 0.001), and collectively explained 53% of the variance in presenteeism. Of these predictors, both physical and functional wellbeing were significant independent predictors.

Metastatic disease was found to impact employment status, with approximately one quarter of patients with metastatic disease (Group 3) reporting being unable to work, compared with fewer than one in ten patients with early BC (Groups 1 and 2). When in part-time employment, patients in Group 3 also worked fewer hours than patients in Groups 1 and 2. In a recent meta-analysis, breast cancer survivors remained at a significantly higher risk of unemployment compared to the general population, which was not the case for other cancers [23]. This may be because the other cancers (blood cancers and testicular cancers) are more common in younger patients, and that perhaps patients with BC were opting for early retirement. However, it is notable that despite lower rates of employment in patients with metastatic BC, of the patients in employment, there were no significant differences in terms of presenteeism between the three groups, suggesting that patients who were able to work, even part-time, were able to be productive. Patients with metastatic BC had similar levels of overall work impairment compared to patients with early BC on treatment. The main impact on productivity was from increased absenteeism in both groups, suggesting that loss of working days due to treatment was an important reason for impaired productivity and would decrease once the patients entered remission. Two studies conducted in North America have reported on loss of working days in women with BC: in the US an average of 22 days were missed from work, and 40 days in patients at more advanced stage of disease [24]; and in Canada patients took almost 6 months off work on average [25]. Further studies would be warranted to better understand absenteeism in patients with breast cancer and how the impact of treatment and stage of disease on work attendance and productivity may be minimised.

Levels of overall work impairment and activity impairment were below 50% in all groups in our study. This is relatively lower than previously reported in Sweden and the Netherlands where levels of work-productivity impairment among women currently receiving treatment for BC were 72% and 69%, respectively, and levels of activity impairment were 62% and 55%, respectively [12]. However, our estimates are similar to those of a US study in patients with advanced BC where activity impairment was 30% and work-productivity impairment ranged between 20 and 40% for the different WPAI scores in employed patients [2]. This emphasises the importance of determining country-specific measurements for estimating the impact of BC on productivity. Patients in remission in our study were significantly less impaired than patients on treatment for subscale scores of work productivity (absenteeism, overall work impairment and activity impairment, not significantly for presenteeism), consistent with the previous study from Sweden and the Netherlands [12]. However, even in patients in remission, absence due to disease and impairment whilst at work remained at levels impacting productivity, emphasising the need for long-term support for women going back to work after treatment for BC.

Patients with early BC experienced similar levels of HRQoL, whether they were in remission or on treatment, and significantly higher HRQoL than patients with metastatic BC. Similarly, in a meta-analysis in Asian women, patients with BC with comorbidities and those treated with chemotherapy had poorer HRQoL compared to women at earlier stages of disease [22]. In a study of US women with HER2-positive metastatic BC, levels of HRQoL increased over time, with women living longer with the disease experiencing an improvement in HRQoL compared to women diagnosed more recently [21]. Levels of HRQoL, as indicated by the EQ-5D-5L and FACT-B summary scores, were similar to that previously reported in other countries, confirming that patients with BC have relatively high levels of HRQoL [2, 13, 26]. In a previous study in Sweden and the Netherlands mean utility scores for stable and progressive diseases were 0.81 and 0.61, respectively [12], similar to utility scores in our sample as estimated with the England tariff. However, it is notable that the FACT-B breast cancer-specific subscale scores were similar between the three groups, suggesting this subscale is not sensitive to differences in breast cancer stage or treatment, consistent with results of the original validation paper that demonstrated no association of this subscale with extent of disease [27].

In our study, disease stage did not explain the level of anxiety and depression, which was not a concern for approximately 40–50% of patients. This is in line with a study demonstrating a majority of women with advanced disease remaining below clinical thresholds for depression [14], and another study reporting high levels of emotional wellbeing in patients in long-term remission [13]. In a meta-analysis, symptoms of depression were elevated in patients in remission 1-year after treatment, and decreased over the ensuing years; and anxiety was not raised in any year post-treatment [28]. In contrast, in our study, anxiety and depression in patients with early BC were similar in those on treatment (diagnosed on average 6 months prior to study enrolment) and in patients in remission (diagnosed on average 2.7 years prior to study enrolment). However, it should be acknowledged that poorer emotional wellbeing, as assessed by FACT-B, was observed in patients with metastatic BC compared with those with early BC. While our study was not adequately powered to further investigate the impact of time in remission on anxiety and depression, our results suggest that, although treatment for BC impacts on various aspects of daily life, including emotional wellbeing and productivity, the majority of women at all stages of BC maintain good mental health in terms of anxiety and depression, even in the presence of metastatic disease.

Poorer HRQoL was associated with greater work and non-work impairment: physical and functional wellbeing, along with BC specific symptoms, were correlated with absenteeism and presenteeism, and emotional wellbeing was associated with presenteeism only. This is in line with the findings of a previous study, which showed that overall quality of life in patients with BC in remission was associated with economic aspects such as work productivity, quality of work, absenteeism, and change in spouse/partner’s income [29]. A recent study in patients with HER2-positive metastatic BC observed a temporal adjustment to the impact of disease, with patients having lived longer with the disease experiencing higher levels of HRQoL and also reporting lower productivity impairment than those recently diagnosed [21]. Taken together, this evidence suggests that maintaining high levels of HRQoL in patients with BC may have a positive economic impact, and conversely limiting the economic burden in patients with BC may translate into improved HRQoL, at all stages of disease and treatment.

Our study has several strengths. Characteristics of patients in our study were similar to UK patients with HER2-positive BC as reported nationally with regard to age, ethnicity, and educational status [30, 31]. A quarter of patients in Group 3 were diagnosed de novo with metastatic BC. This is within the range reported in other studies, including a study in the Netherlands which observed 19% of patients diagnosed with de novo metastatic BC [32], and a study in US patients which reported 33% with de novo metastatic BC [33].

This study is subject to limitations. Selection bias may have been introduced by including consenting patients only, as they may have differed from patients refusing consent. Two thirds of the patients included in our study were not in paid employment, the majority of which were retired, and therefore some analyses in employed patients will have been conducted in small samples. This was a cross-sectional study and therefore the temporal relationship between disease/treatment stage and impact on productivity and HRQoL could not be assessed. In particular, it was not possible to conclude whether worse HRQoL and productivity for patients with metastatic BC (three quarters of which were not de novo metastatic and had progressed from early BC) compared to patients with early BC was due to differences between individuals rather than to worsening of patient’s condition with disease progression. Similarly, the better HRQoL in patients with early BC in remission compared to those still on treatment could be attributable in part to sample variation rather than temporal improvement. Furthermore, we did not evaluate the relationship between the time since diagnosis and HRQoL. In the population of patients in the study with metastatic disease, approximately one quarter had central nervous system (CNS) disease. This is consistent with the reported incidence of brain metastases in HER-2 positive breast cancer [34]. The presence of CNS disease would be expected to have a disproportionate impact on HRQoL and work productivity compared to extracranial disease sites. Furthermore, the study relied on the completeness and quality of medical records and the patients’ answers to questionnaires.

Whilst a longitudinal study design could identify temporal changes in HRQoL and work productivity throughout the course of disease, significant loss to follow-up would be expected during a prolonged period of observation, which would limit the interpretation of results, supporting the cross-sectional approach.

It should be noted that analyses reported were not corrected for potential confounders because groups did not differ on demographic variables or comorbidities and other covariate variables were considered to be either intrinsically linked to group selection criteria (e.g. treatment or disease history), or to the variables being assessed (employment status is somewhat dependent on HRQoL). However, we cannot exclude that other factors not evaluated in the present study (e.g. time since diagnosis) may have influenced the study results. The results of this study should be interpreted in light of these limitations.